EPA/600/R-07/064
                                 September 2007
   Considerations for Developing a
  Dosimetry-Based Cumulative Risk
Assessment Approach for Mixtures of
     Environmental Contaminants
          National Center for Environmental Assessment
            Office of Research and Development
            U.S. Environmental Protection Agency
                Cincinnati, OH 45268

-------
                                   NOTICE
      The U.S. Environmental Protection Agency through its Office of Research and
Development funded and managed the research described here under contract no. EPA
3C-R102-NTEX to Colorado State University.  It has been subjected to the Agency's
peer and administrative review and has been approved for publication as an EPA
document. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
                                 ABSTRACT
      For most of its history, the U.S. Environmental Protection Agency (U.S. EPA) has
assessed risks based on individual contaminants and has often focused on one source,
pathway or adverse effect. But in reality, the public is exposed to multiple contaminants
from a variety of sources, and tools are needed to understand the resulting combined
risks. In keeping with its continuing effort of developing science-based risk assessment,
a major mission goal and challenge for the U.S. EPA has been underway for the
development of cumulative risk assessment. The Office of Pesticide Programs (OPP),
U.S. EPA, took the lead and conducted cumulative risk assessment on
organophosphorus (OP) pesticides under the Congressional mandate of the Food
Quality Protection Act (FQPA), addressing  only chemicals with the same
mode/mechanism of action. The present work is complementary to that of OPP. This
report develops a framework to guide decisions whether to incorporate physiologically-
based pharmacokinetic (PBPK) modeling into the cumulative risk assessment process.
It is not restricted to FQPA applications and intends to guide the risk assessor through
choosing to conduct and conducting a dosimetry based cumulative risk assessment,
addressing the level of toxicant in the target tissue, rather than relying on exposure
concentrations and the  hazard index  approach. This report includes data and values for
several compounds to demonstrate the proposed approach by means of example.  It  is
not intended to present an analysis of the internal exposures or risks from these
mixtures. Some of the example chemicals (trichloroethylene, tetrachloroethylene) are
the subject of ongoing U.S. EPA risk  assessments, and risk values for other chemicals
are those developed by ATSDR, rather than by the U.S. EPA.  The main objectives are
(1) to improve the science and risk assessment by moving the measure of dose from
concentration in the environmental contact medium to inside the body (i.e., tissue dose
rather than exposure dose); (2) to achieve this objective is via PBPK modeling; (3) to
incorporate PBPK  modeling into the cumulative risk assessment process and (4) to
make the optimal use of in vitro techniques, including the use and application of human
tissues.  The proposed  approach retains the 10-step process developed by the OPP;
these are: (1) Identify Common Mechanism Group (CMC); (2) Identify Potential
Exposures; (3) Characterize and Select Common Mechanism Endpoint(s); (4)
Determine The Need For a Dosimetry-Based Cumulative Risk Assessment; (5)
Determine Candidate Cumulative Assessment Group (CAG); (6)  Conduct Dose-

-------
Response Analyses and Determine Relative Potency and Points of Departure; (7)
Develop Detailed Exposure Scenarios for All Routes and Durations; (8) Establish
Exposure Input Parameters; (9) Conduct Final Cumulative Risk Assessment; and (10)
Conduct Characterization of Cumulative Risk.  In the present effort, however, the first
five steps were grouped under Phase I (Initial Analysis) and the next five steps were
grouped under Phase II (Dosimetry-Based Cumulative Risk Assessment).  PBPK
models are proposed to be developed for the components that are considered in
Phase II, i.e., those that are in the CAG.  This approach stresses the importance of
initial analysis to eliminate those situations that do not warrant a PBPK-based approach
to cumulative risk assessment thereby reducing the unnecessary expenditure of
resources.  The development of this framework/approach utilizes two model sets of
chemical mixtures:  Mixture 1 consists of 6 OP pesticides (methyl-parathion, parathion,
chlorpyrifos, fenthion, diazinon, and fenitrothion) which share the same mode of action
of inhibiting acetylcholinesterase (AChE); Mixture 2 consists of four chlorinated
hydrocarbon solvents or volatile organic chemicals (trichloroethylene,
tetrachloroethylene, 1,1,1-trichloroethane, and chloroform) which have different modes
of action on a variety of toxic endpoints. The wealth of information available for these
chemicals was a prominent factor in their choice as two model chemical mixtures for
this effort. The advantages of utilizing PBPK modeling in cumulative risk assessment
and the incorporation of credible human tissue studies in PBPK modeling, as well as the
methodologies involving the incorporation of interactive PBPK models are discussed.
The intent of this document is to serve as a logical framework upon which to integrate
information to decide whether to embark on technical and resource-intensive PBPK
approaches to cumulative risk assessment.
Preferred citation:
U.S. EPA.  2007. Considerations for Developing a Dosimetry-Based Cumulative Risk Assessment
Approach for Mixtures of Environmental Contaminants.  U.S. Environmental Protection Agency, National
Center for Environmental Assessment, Cincinnati, OH. EPA/600/R-07/064.

-------
                         TABLE OF CONTENTS
TABLE OF CONTENTS	iv
LIST OF TABLES	vi
LIST OF FIGURES	vi
LIST OF ABBREVIATIONS	vii
GLOSSARY	ix
LIST OF AUTHORS, CONTRIBUTORS AND REVIEWERS	xi
ACKNOWLEDGMENTS	xii

1.    INTRODUCTION	1

     1.1.   BACKGROUND	1
     1.2.   STRATEGY: INCORPORTAION OF PBPK MODELING AND
           HUMAN TISSUE STUDIES IN CUMULATIVE RISK ASSESSMENT	3
     1.3.   OBJECTIVES	4
     1.4.   TWO MODEL MIXTURES OF DRINKING WATER CONTAMINANTS	5
     1.5.   PHARMACOKINETICS vs. PHARMACODYNAMICS	5
     1.6.   INTERACTIONS: CUMULATIVE RISK AND PHARMACOKINETICS 	6

2.    THE  PROPOSED APPROACH	13

     2.1.   OVERVIEW	13
     2.2.   PHASE I: INITIAL ANALYSIS	16

           2.2.1. Step 1.  Identify Common Mechanism Group (CMC)	16
           2.2.2. Step 2.  Identify Potential Exposures	18
           2.2.3. Step 3.  Characterize and Select Health Endpoint(s) for
                Evaluation	20
           2.2.4. Step 4.  Determine the Need for a Dosimetry-Based
                Cumulative Risk Assessment	22
           2.2.5. Step 5.  Identify Chemicals in the Candidate Cumulative
                Assessment Group (CAG) for Further Evaluation	36

     2.3.   PHASE II: DOSIMETRY-BASED CUMULATIVE RISK
           ASSESSMENT	37

           2.3.1. Step 6.  Characterize Dose-Response, Point of Departure and
                Determine Relative Potency from Tissue Doses	38
           2.3.2. Step 7.  Develop Detailed Exposure Scenarios for All
                Routes and Durations	49
           2.3.3. Step 8.  Quantify Parameters for Exposure	49
                                  IV

-------
                     TABLE OF CONTENTS cont.
          2.3.4. Step 9. Conduct Dosimetry-Based Cumulative Risk
               Assessment	50
          2.3.5. Step 10. Characterize Cumulative Risk via Pharmacokinetic
               Analysis	52

3.    CONCLUSIONS	54

4.    REFERENCES	57

APPENDIX A: A GENERAL REFERENCE ON PHYSIOLOGICALLY-BASED
PHARMACOKINETIC/PHARMACODYNAMIC (PBPK/PD) MODELING	A-1

APPENDIX B: A REFERENCE ON HOW TO  INCORPORATE CREDIBLE
HUMAN ENZYME STUDIES INTO THE PHYSIOLOGICALLY-BASED
PHARMACOKINETIC (PBPK) MODELING AND RISK ASSESSMENT
PROCESS	B-1

APPENDIX C: INFORMATION ON CHEMICAL MIXTURES AND THEIR
COMPONENT CHEMICALS	C-1

APPENDIX D: EXTERNAL REVIEW COMMENTS	D-1

-------
                              LIST OF TABLES






1     Screening Hazard Index Approach	23



2     Derivation of TTD Values for Chloroform	25



3     Derivation of TTD Values for Tetrachloroethylene	26



4     Derivation of TTD Values for Trichloroethylene	27



5     Derivation of TTD Values for 1,1,1 -Trichloroethane	28



6     Target Organ Toxicity Doses, Hazard Quotients and Hazard Indices	30






                              LIST OF FIGURES






1     Data Evaluation for Dosimetry-Based Cumulative Risk Assessment	15



2.     A Generalized Metabolic Scheme for OP Pesticides	32



3     Metabolic Pathway for Solvents Indicating Potential Sites for Interactions	33



4     A Preliminary PBPK Model Structure for OPs	44



5     A Pharmacodynamic Submodel for CNS/PNS Compartment	46



6     A Schematic for Integrating Data for Cumulative Risk Assessment	51
                                     VI

-------
                         LIST OF ABBREVIATIONS

AChE       Acetylcholinesterase
AL         Acceptable level
AUC       Area-under the concentration vs. time curve
BMD       Benchmark dose
BMDL      Lower bound benchmark dose
BMDS      Benchmark dose software
CAG       Cumulative Assessment Group
CE         Carboxylesterase
CHF       Chloroform
CMC       Common Mechanism Group
CMS       Central nervous system
CRA       Cumulative Risk Assessment
DBCRA    Dosimetry-Based Cumulative Risk Assessment
CYP       Cytochrome P450
E          Exposure
FQPA      Food Quality Protection Act
GST       Glutathione S-transferase
HI         Hazard Index
HQ         Hazard Quotient
IRIS       Integrated Risk Information System
LOAEL     Lowest-observed-adverse-effect level
MC         1,1,1-Trichloroethane
NCEA      National Center for Environmental Assessment
NOAEL    No-observed-adverse-effect level
OP         Organophosphorus
OPP       Office of Pesticide Programs
PBO       Piperonyl butoxide
PBPK      Physiologically-based pharmacokinetic
PD         Pharmacodynamic
PERC      Tetrachloroethylene
PK         Pharmacokinetic
POD       Point of departure
PNS       Peripheral nervous system
RfC        Reference concentration
                                    VII

-------
                       LIST OF ABBREVIATIONS cont.

RfD        Reference dose
TCE       Trichloroethylene
TTD       Target Organ Toxicity Dose
U.S. EPA   U.S. Environmental Protection Agency
VOCs      Volatile organic chemicals
WOE       Weight of evidence
                                    VIM

-------
                                 GLOSSARY
Acceptable Level: A measure of the acceptable level of human exposures.  For
cumulative risk assessment, reference values are typically used.

Aggregate Exposure: The combined exposure of an individual (or defined population)
to a specific agent or stressor via relevant routes, pathways, and sources.

Aggregate Risk: The risk resulting from aggregate exposure to a single agent or
stressor.

Chemical Mixture: Any set of multiple chemical substances that may or may not be
identifiable, regardless of their sources, that may jointly contribute to toxicity in the
target population.  May also be referred to as a whole mixture"  or as the "mixture of
concern."

Components: Single chemicals that make up a chemical mixture that may be further
classified as systemic toxicants, carcinogens,  or both.

Common Mechanisms Group: The group of chemicals under evaluation that induce a
common effect via the same mechanism of toxicity. The Initial  Analysis begins with the
identification of this group of chemicals.

Critical Organ/Effect: The effect or the organ that responds first as dose increases.
The dose-response for the critical organ serves as the basis for establishing reference
values (e.g., RfD Values).

Cumulative Assessment Group: The group  of chemicals surviving the Initial Analysis,
for which a dosimetry-based cumulative Risk Assessment will be conducted.

Cumulative Risk: The combined risks from aggregate exposures to multiple agents or
stressors.

Cumulative Risk Assessment: An analysis, characterization and possible
quantification  of the combined risks to health or the environment from multiple agents or
stressors.

Dosimetry-Based Cumulative Risk Assessment: Cumulative risk assessment
undertaken using tissue concentrations, or internal measures of exposure, rather than
external measures of exposure.

Hazard Quotient: The ratio of human exposure to reference values used to estimate
the potential for  non-cancer health effects. HQ = E/AL, where E is exposure and AL is
an acceptable level of exposure (e.g., RfD value).

Hazard Index: The summed Hazard Quotient values for a given set of chemicals or
responding tissues, depending on the analysis.
                                      IX

-------
                              GLOSSARY cont.

Index Chemical: The chemical selected as the basis for standardization of toxicity of
components in a mixture. The index chemical must have a clearly defined dose-
response relationship.

Initial Analysis: The first 5 steps of the process of developing a dosimetry-based
cumulative risk assessment. This initiates with the identification of a common
mechanism group and culminates with the analysis of pharmacokinetic data and a
Hazard index analysis.

Interactions: Interactions among mixture components are demonstrated by responses
that, when measured, differ from those predicted by additivity. Chemical interactions
may occur due to commonalities in pharmacokinetic or pharmacodynamic processes.

Pharmacokinetics: The (study of the) distribution of chemical dosimetry in the body.

Pharmacodynamics: The process of developing a biological response to a chemical.

Point of Departure: The dose-response point that marks the beginning of a low-dose
extrapolation. This point can be the lower bound on dose for an estimated incidence or
a change in response level from a dose-response model (BMD), or a NOAEL or LOAEL
for an observed incidence, or change in level of response.

Target Organ: Any organ adversely affected by chemical exposures.

Target Organ Toxicity Dose Approach: A cumulative risk approach quite similar to
the Hazard  Index Approach. The TTD approach differs in that includes dose-response
data to estimate AL in the E/AL format for all responding organs, not just the critical
organ.

Target Organ Toxicity Dose Values: Hazard quotient values developed for organs
that are not the critical organ in the IRIS Assessment, or for which a reference value has
not been formally established.

Weight of Evidence: A qualitatively useful adjunct to the Hazard Index approach that
incorporates information available in binary combinations of chemicals. This approach
is used to account for interactions among mixture components.

Definitions obtained from U.S. EPA (2000a, 2003c) and Mumtaz and Durkin (1992).

-------
            LIST OF AUTHORS, CONTRIBUTORS AND REVIEWERS

AUTHORS

John C. Lipscomb
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH

Raymond S.H. Yang
Colorado  State University
Center for Environmental Toxicology and Technology
Fort Collins, CO

James Dennison
Colorado  State University
Center for Environmental Toxicology and Technology
Fort Collins, CO


INTERNAL REVIEWERS

Rob DeWoskin
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC

Rick Hertzberg
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH


EXTERNAL REVIEWERS

Richard J. Bull, Ph.D.
MoBull Consulting
Richland,  WA

Harvey Clewell (Chair)
CUT Centers for Health Research
Research Triangle Park, NC
                                    XI

-------
          LIST OF AUTHORS, CONTRIBUTORS AND REVIEWERS cont.

Gary L. Ginsberg, Ph.D.
Connecticut Department of Public Health
Hartford, CT

Margaret MacDonell,  Ph.D.
Argonne National Laboratory
Argonne, IL

MoizM. Mumtaz,  Ph.D.
Agency for Toxic Substances and Disease Registry (ATSDR)
Chamblee, GA

Clifford P. Weisel, Ph.D.
Environmental & Occupational Health Sciences Institute (EOHSI)/UMDNJ
Piscataway, NJ
ACKNOWLEDGMENTS


      This project was a team effort by members of the Quantitative and Computational
Toxicology Group at the Center for Environmental Toxicology & Technology (CETT),
Colorado State University. The literature reviews of component chemicals of the two
chemical mixtures were carried out by many of the graduate students working toward
their Ph.D.s in Toxicology and Chemical Engineering at CETT; they are: Ms. Amanda
Ashley, Mr. Sun Ku Lee, Mr. Ken Liao,  Mr. Manupat Lohitnavy, Ms. Ornrat Lohitnavy,
Mr. Yasong Lu, and Mr. Damon Perez.  Their collective effort is gratefully
acknowledged. Dr. Anna Lowit of OPP, U.S. EPA read the Draft Approach and kindly
provided valuable comments and suggestions.
                                     XII

-------
                              1. INTRODUCTION

1.1.   BACKGROUND
      On July 3, 1997, the U.S. Environmental Protection Agency (U.S. EPA)
Administrator, Carol M. Browner, and Deputy Administrator, Fred Hansen, jointly issued
a memorandum entitled "Cumulative Risk Assessment Guidance - Phase I Planning
and Scoping" to top U.S. EPA officials.  The content of this memo, quoted below,
provided the essence of the reasoning for cumulative risk assessment.


            As you are aware, the processes that EPA and others follow to
      assess environmental risk are of great interest to environmental
      professionals and to the public,  and growing attention is being given to the
      combined effects of multiple environmental stressors. Consistent with
      this, EPA and others are asking more questions about the wider and more
      complex issues that define a cumulative approach to risk assessment.
      Today, we are providing guidance for all EPA offices on cumulative risk
      assessment. This guidance directs each office to take into account
      cumulative risk issues in scoping and planning major risk assessments
      and to consider a  broader scope that integrates multiple sources, effects,
      pathways, stressors and populations  for cumulative risk analyses in all
      cases for which relevant data are available.  This assures a more
      consistent and scientifically complete Agency-wide approach to
      cumulative risk assessments in  order to better protect public health and
      the environment.
            This approach provides a platform for significant advances in our
      scientific approach to assessing environmental risks. For most of our
      history, EPA has assessed risks and  made environmental  protection
      decisions based on individual contaminants—such as lead, chlordane,
      and DDT—with risk assessments for  these chemicals often focused on
      one source, pathway or adverse effect. Today, better methods and data
      often allow us to describe and quantify the risks that Americans face from
      many sources of pollution, rather than by one pollutant at  a time.  We are
      increasingly able to assess not simply whether a population is at risk, but
      how that risk presents itself.  In  addition, we are better able in many cases
      to analyze risks by considering any unique impacts the risks may elicit due
      to the gender, ethnicity, geographic origin, or age of the affected
      populations.  Where data are available, therefore, we may be able to
      determine more precisely whether environmental threats pose a greater
      risk to women, children, the  elderly, and other specific populations, and
      whether a cumulative exposure to many contaminants, in  combination,
      poses a greater risk to the public.
            Of particular importance are the right-to-know implications of this
      guidance, which requires that we build opportunities for citizens and other

-------
      stakeholders to understand our ongoing risk assessments, and to provide
      us with their comments. Our goal is to ensure that citizens and other
      stakeholders have an opportunity to help define the way in which an
      environmental or public health problem  is assessed, to understand how
      the available data are used in the risk assessment, and to see how the
      data affect decisions about risk management.
            Some Regions and Programs within the Agency are already
      making significant efforts to use integrated or cumulative risk assessment
      techniques, and this guidance both reflects those practices and makes
      them consistent across the Agency. The scope of integrated risk
      assessments often  involves coordination across many program offices
      and statutory mandates for risk analysis; for example, those called for
      under the new safe drinking water and food safety laws.  Therefore, this
      guidance calls for ongoing communication among risk assessors, risk
      managers, economists, engineers,  and  other technical experts within the
      Agency.
            While we can more consistently take into account many new factors
      in this approach to risk assessment, many other potentially important
      factors are more difficult to include  in our analyses, particularly the social,
      economic, behavioral or psychological factors  that also may contribute to
      adverse health effects.  These include,  among others, such factors as
      existing health conditions, anxiety,  nutritional status, crime and
      congestion. Assessment of these factors is often hampered by a lack of
      data to establish plausible cause-and-effect relationships; difficulties in
      measuring exposure,  incidence and susceptibilities related to these risks;
      and few methods for assessing or managing these risks. This guidance
      does not address these factors.  We expect, nonetheless, that this
      guidance will be updated as our understanding and experience develop;
      and, the Agency is focusing its research to improve our ability to
      incorporate these broader concerns into our cumulative risk assessments
      as new data and methods are brought forward.
            Please take the steps needed to ensure that all major risks
      assessments undertaken in your area embrace this cumulative approach,
      so that we can better advise all citizens about  the environmental and
      public health risks they face, and improve our  ability to protect the
      environment and public health for the nation.


      During his tenure as U.S. EPA Science Advisor and Chair, Science  Policy
Council, Dr Paul Oilman underscored the  importance of cumulative risk assessment to
the U.S. EPA in the following memo entitled "Framework for Cumulative Risk
Assessment" to senior managers at U.S. EPA.

-------
            I am very pleased to present EPA's Framework for Cumulative Risk
      Assessment. For most of our history, EPA assessed risks based on
      individual contaminants and often focused on one source, pathway or
      adverse effect.  But in reality, the public is exposed to multiple
      contaminants from a variety of sources, and tools are needed to
      understand the resulting combined risks.  The Framework represents an
      important milestone for EPA in expanding our focus from an individual
      chemical-based approach to a community or population-based approach
      for multiple stressors.
            Development of the Framework is one of a series of Agency
      activities to better address combined risks from multiple stressors.
      Several  EPA Programs and Regions have growing experience in
      conducting cumulative risk assessments, such as in evaluating cumulative
      risk for pesticides with a similar mode of action.  In  addition,  the EPA
      Science Policy  Council (SPC) has sponsored several efforts designed to
      lead to Agency-wide guidance on planning and scoping for cumulative risk
      assessments.  As a next step, a technical panel of the Risk Assessment
      Forum has now completed the Framework, which builds on prior efforts by
      identifying the basic elements and definitions for cumulative risk
      assessment. It also serves as the foundation for future efforts, such as
      evaluating past and emerging case studies in relation to the approach
      outlined in the document. I offer my thanks to this cross Agency panel of
      experts for their efforts to further advance the thinking in this area.
            Cumulative risk assessment is a major challenge for the Agency.
      This Framework moves us closer to achieving our goal of producing the
      most scientifically rigorous and realistic evaluation of cumulative risk that
      the state-of-the-science can accommodate.  The Administrator and I
      encourage Agency personnel to incorporate the thinking embodied in this
      document in the development of cumulative risk assessments that
      address risk management needs.
1.2.   STRATEGY: INCORPORATION OF PBPK MODELING AND HUMAN TISSUE
      STUDIES IN CUMULATIVE RISK ASSESSMENT

      The present report follows the vision outlined in the above memos by proposing
an approach which integrates Physiologically-Based Pharmacokinetic (PBPK) modeling,
particularly the inclusion of credible human tissue studies, into the cumulative risk
assessment of chemical mixtures. There are several reasons for utilizing the PBPK
modeling approach:


   1.  As indicated in the above memos, U.S. EPA continues to improve its risk
      assessment process by incorporating the state-of-the-art science and
      technology. PBPK modeling, though advancing continuously in the last 30 years

-------
      or so, is a new computational toxicology technology. From a different
      perspective, the state-of-the-science is such that PBPK modeling is ready to be
      incorporated into cumulative risk assessment.  In this regard, it is gratifying to
      note that an inter-office endeavor on "Physiologically-Based Pharmacokinetic/
      Pharmacodynamic Modeling: Preliminary Evaluation and Case Study for the
      N-Methyl Carbamate Pesticides: A Consultation" at the U.S. EPA was reviewed
      by the FIFRA Science Advisory Panel in December 2003.
                                                     However, due to the lack of
      availability of a PBPK model for all but one chemical, a PBPK modeling approach
      was not taken in the final assessment (U.S. EPA, 2007a). Many advantages,
      which coincide with what's presented here, were noted in the "Consultation"
      above document (U.S. EPA, 2003a) and the subsequent FIFRA SAP meeting
      minutes (U.S. EPA, 2003b) for utilizing PBPK modeling in cumulative risk
      assessment.
   2.  Multiple chemical exposure and interaction in our body is an extremely complex
      phenomenon.  PBPK modeling is a powerful tool to integrate the various routes
      and modes of exposure  and potential biological interactions with multiple
      chemicals to help resolve the best target tissue dose metrics for risk assessment.
   3.  PBPK models support hypothesis testing through experimentation of evaluation
      using in silico methods and serves as a platform upon which to integrate existing
      data, and extend the efficient use of the available data. A well developed  PBPK
      model is tested for accuracy in simulating existing study results, and if sufficiently
      able to do so, can  be used to simulate other needed data to minimize animal
      usage that would otherwise be required for "unnecessary experiments". This is
      particularly  relevant for the many animal toxicology experiments that would be
      required to  evaluate multiple chemical interactions in a cumulative risk
      assessment.
   4.  PBPK models can extrapolate dosimetry across dose level, route, species, age,
      or gender, and can characterize the uncertainty in simulation results (which are
      derived from the accuracy of the model in simulating existing data). A PBPK
      model may even provide useful information for the dose-response relationship at
      low doses where experimental studies are impossible to conduct.
   5.  PBPK modeling offers a reliable tool for use  in exploring the relationship between
      external dose and  tissue response in the growing area of systems biology.


1.3.   OBJECTIVES
      The main objectives/emphases are: (1) to improve the science and risk
assessment by moving the measure of dose from concentration in the environmental
contact medium to inside the body (i.e., tissue dosimetry rather than  exposure dose); (2)
to achieve the objective of tissue dosimetry determination  via PBPK modeling; (3) to
incorporate the best and most efficient scientific approach, PBPK modeling, into
cumulative risk assessment; and (4) to make the best use of in vitro techniques;

-------
including the use and application of human tissues. The intent of this document is to
serve as a logical framework upon which to integrate information to decide whether to
embark on technical and resource-intensive PBPK approaches to cumulative risk
assessment.

1.4.   TWO MODEL MIXTURES OF DRINKING WATER CONTAMINANTS
      It should be emphasized that this effort is for establishing a framework/approach
for integrating PBPK modeling into the cumulative risk assessment  process, not the
actual conduct of cumulative risk assessment.  Indeed, two example component
chemicals, trichloroethylene and tetrachloroethylene, are the subject of ongoing U.S.
EPA risk assessments.  The actual conduct of a cumulative risk assessment is
understandably much more resource-intensive.  As shown below, this report provides a
description of a recommended approach for integrating PBPK modeling into cumulative
risk assessment for two groups of drinking water contaminants.  The first group is a
mixture of six organophosphorous (OP)  insecticides (methyl-parathion, parathion,
chlorpyrifos, fenthion, diazinon, and fenitrothion) which share the same mode of action
of inhibiting acetylcholinesterase (AChE); the second group is a mixture of four
chlorinated hydrocarbon solvents or volatile organic chemicals (trichloroethylene,
tetrachloroethylene,  1,1,1-trichloroethane,  and chloroform) which have different modes
of action for a variety of toxic endpoints. We understand that chlorpyrifos and diazinon
household usage have been cancelled; however, all these OP pesticides can still be in
drinking water and/or as food contaminants because of residues from earlier
applications.  Furthermore, we use these two model mixtures as a test set for
demonstrating the establishment of a framework/approach. The risk assessment
approach is basically similar to the 10-step approach recommended by the  U.S. EPA for
cumulative risks (U.S. EPA, 2002a,b).

1.5.   PHARMACOKINETICS vs. PHARMACODYNAMICS
      Pharmacokinetics (PK) and pharmacodynamics (PD) are inseparable in the
sense that they formulate a continuum of a toxicological process from exposure, to
tissue dose metrics,  to  molecular interaction(s) and finally to toxic effect(s).  This
document will address the grouping of chemicals according to similarities in
pharmacokinetics, pharmacodynamics (mode/mechanism of action) and target organ.
These are mutually separable groupings: chemicals may attack the same target organ
through different modes of action and have markedly different pharmacokinetic profiles
with different physiological and or biochemical functions determining the

-------
pharmacokinetic profile. Alternately, chemicals may share the same target organ, and
may have similar pharmacokinetic profiles, but act on different cellular or organ function
via different modes or mechanisms of action.  Presently, efforts are advancing to guide
the assessment of chemicals grouped by mode or mechanism of action, and mature
guidance exists for grouping chemicals according to common target organ (ATSDR,
2004;  U.S. EPA, 1989, 2000a).  Because of the rate at which pharmacokinetic analyses
are becoming directly and quantitatively incorporated into risk assessments, it seemed
valuable to develop a discourse on some considerations that should be undertaken prior
to embarking on a complex and resource-intensive pharmacokinetic analysis to support
a cumulative risk assessment.  For those reasons  and because of the state of
development of the area of PBPK modeling, we place more emphasis on PBPK
modeling in this effort.  Thus, at the present time, we would like to focus on
incorporating PBPK modeling into the cumulative risk assessment process first.  As the
science advances and more relevant information becomes available, we will consider
PBPD modeling in cumulative risk assessment as  well.
      Because cumulative risk assessment will impact upon our society at large, when
this process is carried out, it should be transparent and inclusive of all stakeholders
much the same way as Office of Pesticide Programs (OPP) did it for OP pesticides
(U.S. EPA, 2002b, 2003c).

1.6.   INTERACTIONS: CUMULATIVE RISK AND PHARMACOKINETICS
      The term "interaction" has different meanings regarding cumulative risk and
pharmacokinetics. In cumulative risk parlance, interaction describes the risk outcome
(toxicity, organism response) when exposure to a chemical mixture results in a degree
of response not predicted by additivity (e.g., antagonism, synergism, potentiation). If
the outcome is over or under-predicted by additivity, then the chemicals are said to have
an "interaction." With respect to applying an increased understanding of tissue
dosimetry in risk assessment, assessing interactions becomes problematic when the
observed response is related to the external dose  (mg/kg) and the observation (i.e.,
enzyme levels in blood indicative of heart damage) is made at the level of the intact
organism. The outcomes are  influenced by absorption, distribution, metabolism and
elimination processes (pharmacokinetics) on production/reduction of the toxic chemical
species and its delivery to the target organ/tissue,  as well as the inherent
responsiveness of the tissue to the insult (toxicodynamics).
      Later, this document will present a grouping of chemicals into a Common
Mechanism Group, which is done so that a cumulative risk assessment can be

-------
performed on that group. Recently, the organophosphate insecticides have been
grouped according to their common inhibition of acetylcholinesterase (AChE). This is
the key step in their toxic effect—for which the mode of action might be an accumulation
of acetylcholine.  The distinction between mechanism of action and mode of action can
be important and has been treated elsewhere:

      The U.S. EPA Guidance for Conducting Health Risk Assessment of
      Chemical Mixtures (U.S. EPA,  2000) defines mode of action (MoA) as a
      series of 'key events' and processes starting with interaction of an agent
      with a cell, and proceeding through operational and anatomical changes
      causing disease formation. A 'key event', as defined in the 2005  U.S.
      EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005), is an
      empirically observable precursor step that Is Itself a necessary element of
      the mode of action or is a biologically based marker for such an element.
      In contrast, mechanism of toxic action implies a more detailed
      understanding and description of events, often at the molecular and
      cellular level (U.S. EPA, 1999,  2003a).  The U.S.  EPA 2005 and U.S. EPA
      2000 Guidelines both emphasize MoA over mechanism, indicating that
      MoA is a critical determinant for evaluation of risk posed by environmental
      hazards. For well defined modes of action,  mechanistic data may merely
      provide more detail in support of identified key events.  However,  for less
      well defined toxic modes of action, mechanistic data may lead to
      identification of previously unknown obligatory steps in the causal pathway
      leading to toxicity (Lambert and Lipscomb, 2007).

      What is most important is whether chemicals in a mixture can  interact, altering
the toxicity of the mixture. Later, this  section presents several examples of chemical
interactions that would not have been anticipated based only on data describing either
the mode of action or the mechanism of action.
      From a pharmacokinetic standpoint, chemicals have the potential to interact in
either passive (i.e., distribution phenomena;  tissue partitioning) or active (i.e., energy-
requiring processes; active transport,  metabolism)  processes. Available evidence
indicates that chemicals partition into  tissues independent of one another, when
encountered at concentrations that do not disrupt tissue integrity. However, it is widely
accepted that chemicals (e.g., metabolic substrates) can, and do compete against one
another for metabolism, and that the likelihood for this competition increases with
increasing levels of exposure (dose).  When substrates are metabolized by the same
enzyme, or require the same cofactor, interaction is possible.  This interaction may be
based on a suicide inhibition of an enzyme (mechanism  based inhibitors) or a simple
competition for an enzyme (competitive inhibition). Mechanism based inhibitors (suicide

-------
substrates; this type inhibition is demonstrated as non-competitive) have effects that
cannot be overcome by changing the ratio of substrates.  When these substrates are
metabolized, an irreversible complex is formed between product (or intermediate) and
enzyme, irreversibly binding the active site and inhibiting catalysis. However,
competitive inhibition is more common.  Even chemicals (competing substrates) that
share the same enzyme may not demonstrate inhibition under some conditions (levels)
of exposure. Studies have demonstrated that there are thresholds of exposure, below
which chemicals that do share a common enzyme do not interfere with the metabolism
of one another.  In the case of chemicals thought to have a metabolic (competitive)
interaction, the total exposure to the mixture, rather than the exposure to single
components should be evaluated.  Dobrev et al. (2001) constructed an analysis to
estimate interaction thresholds in the rat for three commonly found environmental
contaminants, trichloroethylene (TCE), tetrachloroethylene (PERC) and
1,1,1-trichloroethane (methyl chloroform; MC). PBPK models were constructed for each
chemical  using published values for parameters, including Km values.  PBPK modeling
of gas uptake data was applied, and interactions were characterized as concentrations
of PERC  and MC required to increase blood TCE concentrations by 10%.  These
solvents were chosen because of their ubiquitous occurrence as environmental
contaminants, as well as their dependence on cytochrome P450 (CYP)-mediated
oxidation  as a primary metabolic process.  Interaction models were constructed for
competitive, non-competitive and uncompetitive inhibition. Although all models fit the
data to some extent, the similarity between model-predicted (optimized) values for the
inhibition  constant (Kj) and Km values indicated that the competitive model best fit the
data.  This model was implemented to interpret gas uptake data for binary and ternary
mixtures.   In the model, TCE exposure was maintained at the TLV of 50 ppm and
interactions from  binary and ternary exposures were examined.  Under binary
exposures to TCE and either MC or PERC, the threshold for interaction was 175 ppm
(MC) and there was no interaction when PERC exposure was at 25 ppm. However,
when the model simulated a ternary exposure to TCE (50 ppm), PERC (25 ppm) and
MC, interaction was noted at MC concentrations of 130 ppm (compared to 175 ppm
when encountered in a binary mixture).  These results demonstrate the value of PBPK
modeling in refining the descriptions of toxicologically important (metabolic,
toxicokinetic) interactions among components of environmentally-important mixtures.
      The above interaction is metabolic, or pharmacokinetic in nature. When a
mixture or cumulative exposure results in a response different from that predicted from
additivity  and there are no data on tissue dosimetry,  it is not possible to determine
                                      8

-------
whether the interaction is based on toxicokinetic or toxicodynamic interactions, or both.
From an "interactions" perspective, it is possible that pharmacokinetic interactions may
explain chemical mixture (cumulative risk) interactions, that may not be "interactions" as
defined by cumulative risk; such departures from additivity may only be departures
when considered at the level of the external or applied dose, and may be strictly
additive when "dose"  is expressed as "dose metric"—the concentration of the
toxicologically active chemical species in the target tissue.
      For example, consider two chemicals, A and B, that share a common metabolic
pathway. The response from chemical A is dependent on a metabolite, whereas the
response from chemical B is mediated by the parent compound.  When the animal is
concomitantly exposed to these chemicals at a level where metabolic interactions do
occur—chemical A out-competes chemical B for metabolism; chemical A is metabolized
to a greater extent than chemical  B, the response from a mixtures exposure is much
higher than the response predictions made on the assumption  of additivity applied to
single chemical toxicity results. An explanation may include the metabolic interaction,
and when the "dose" is expressed in terms of the biologically active dose (a metabolite
of chemical A and the parent form of chemical B), the response may, indeed, be
predicted by additivity. Appreciating, developing and communicating the distinction, as
well as its basis, will decrease the level of uncertainty associated with assessments of
cumulative  risk.
      Pharmacodynamics is concerned with the development  of the response,
ultimately at the level of the cell. Processes that determine or modify delivery of the
toxicologically active chemical species to the cell are addressed as pharmacokinetics.
The biological response is ultimately mediated at the cellular level and key events may
be identified, and sorted into those defining the mechanism of toxicity and those
defining the mode of action. However, when chemicals act on  the same biological
and/or biochemical cellular processes, a toxicodynamic interaction may occur,
regardless whether the shared biological processes are involved in mode or mechanism
of toxic action.  These are not  trivial, and interactions may occur in biochemical
pathways not directly involved in the critical toxic event for a chemical. Piperonyl
butoxide (PBO) is a synergist in some insecticides. PBO is not especially toxic, and it
inhibits enzymes that are important in detoxicating insecticides. Under a single-
chemical exposure to PBO,  inhibition of these enzymes is not especially important,
though it does represent a toxicodynamic effect—it represents  an adverse interaction of
the toxic agent with a biomolecule. However, when PBO is co-exposed with an
insecticide, the result is rather marked potentiation of insecticidal activity. The

-------
interaction is toxicokinetic in this regard, because the potency (when measured as effect
per applied dose) of the insecticide is increased.  However, if the measure of dose is
expressed relative to the internal dose of the insecticide, the potency (expressed per
mg/kg of internally circulating insecticide) is not altered.  An example of a strictly
toxicodynamic interaction may occur when two chemicals share a common biological or
biochemical event.  If both chemicals alter cellular communication resulting in an
outgrowth of damaged cells, then this type of interaction is likely to demonstrate an
effect characterized by additivity.  This is the type of interaction that is anticipated for
mixtures of organophosphate insecticides. However, if two chemicals act on different
biochemical processes in the same response pathway, a response other than additive
might occur. Such would be the case if one chemical caused a decrease in the efficacy
of DMA repair and/or a release of apoptotic control of cellular growth, and a second
chemical caused damage to DMA. Under the event of a co-exposure, then the net
result might be the outgrowth of cells in which DMA damage was responsible for toxicity.
The effect resulting from the DMA damage might be a further loss of control  over cell
growth leading to tumorigenesis, or may be manifest as the production of a protein
whose function has been adversely affected, resulting in increased cellular toxicity.
      Whereas some of the above interactions results in potentiation or synergism,
such is not  always the case.  Mehendale and colleagues evaluated the interaction
between thioacetamide and carbon tetrachloride (reviewed in Mehendale, 2005). Both
compounds are known to induce liver injury.  However, it was demonstrated that  low
doses of thioacetamide given in the hours before carbon tetrachloride administration
actually protected against CCU-induced hepatotoxicity. Ultimately, the effect was
demonstrated to reside in the rebound effect of cellular repair following thioacetamide
exposure. Here, the pre-exposure resulted in a stimulation of cellular repair processes;
these processes were already active when CCU exposure  and injury occurred. The
interaction was significant to the extent that thioacetamide  exposure could spare
animals from doses of CCU that would otherwise have been lethal (reviewed in
Mehendale, 2005).
      The application of pharmacokinetic analyses to chemical  interactions can
suggest the basis for interactions (determine whether and to what extent departures of
response from additivity can be attributed to pharmacokinetic interactions).  An example
is provided  by the joint exposure of kepone and carbon tetrachloride.  Mehendale and
colleagues  have demonstrated a significant interaction between these chemicals,
resulting in  an approximate 67-fold induction of CCU-induced hepatotoxicity. A PBPK
model was  developed to study the interactions between CCU and kepone in rats
                                       10

-------
(EI-Masri et al., 1996a). Previous experimental findings demonstrated that kepone
co-exposure amplified the CCU-induced lethality 67-fold. In an additional study, CCU
was administered i.p. to rats fed a control diet or a diet containing Kepone. Data
describing the pathology and the exhalation of CCL4 (indicative of metabolism) were
collected and a pharmacodynamic model was developed and incorporated that
characterized the injury, repair and death sequence for cells and animals.  The results
of this PBPK-PD model demonstrated an approximate 61 to 142-fold increase in CCU
lethality with kepone treatment. However, when the results were adjusted for PK
differences in CCU metabolism, the increase was approximately 4-fold.  These results
suggest that the kepone-CCU interaction is primarily based on a TK interaction, but may
also rely to some extend on a TD interaction.
      Another example of pharmacokinetic interactions altering response can be
provided by the organophosphate insecticides. While these compounds have been
compiled into the same common mechanism group based on inhibition of AChE (the
toxicodynamic determinant of CMS toxicity), a pharmacokinetic  interaction among  OPs
may also influence response.  These compounds are  degraded by carboxylesterase
(CE) enzymes. When exposure to OPs includes some compounds with relatively  low
toxic potency (e.g., fenitrothion) as well as more potent compounds (like parathion), the
lower potency compounds may compete against the other compounds for detoxication
via CE enzymes. The net result can be that higher levels of internal dose  per unit
external dose are developed than would have been predicted on the basis of single
chemical studies (see Chambers et al., 1991; Cohen,  1984).  In this example,
interaction would occur at a level that may not have been predicted based on the results
of toxicity studies.  In this regard, this case is similar to that demonstrated  by piperonyl
butoxide: metabolic interactions alter internal dosimetry and are responsible for toxic
interactions.
      This example highlights another important consideration—that of timing. Timing
may relate to pharmacokinetics, where dose timing can be different when characterized
at the level of the external (encountered) dose or the tissue dose. For some chemicals,
a prolonged tissue exposure can result from a brief environmental contact. In those
instances, toxicant concentrations in tissues can remain appreciable hours, days or
even weeks after the cessation of external exposures. Here, a joint exposure might
occur, even if the person is not concomitantly "exposed" to multiple chemicals. Another
mechanism through which timing may be complicated is when the effects of an
exposure persist following cessation of exposure and  elimination of the compound from
the body. An example of such an effect is provided by the enzyme inducing ability of
                                      11

-------
some compounds. Many compounds can and do induce hepatic cytochrome P450
enzymes, and the previous administration of these compounds can modify the toxic (or
therapeutic) response to subsequently-encountered xenobiotics via a modification of
their internal dosimetry.  The metabolism of a compound may represent a bioactivation
or a detoxication process, and so the effect of enzyme induction (or inhibition) on toxic
outcome must be evaluated on a chemical by chemical basis.
                                     12

-------
                         2. THE PROPOSED APPROACH

2.1.   OVERVIEW
      The existing U.S. EPA guidance for addressing the cumulative risk for pesticide
mixtures acting through a common mode of action (U.S. EPA, 2002a)1 suggests a
series of evaluations that first lead to the determination of the Cumulative Assessment
Group.  The U.S. EPA's Office of Pesticide Programs has undertaken cumulative risk
assessments for four groups of compounds (organophosphates, N-methyl carbamates,
triazines and chloroacetanilides), which are available at
                                                                  The reader of
this report should not confuse the Cumulative Assessment Group (CAG) with the
Common Mechanism Group (CMC). A cumulative risk assessment can include
chemicals for which the mechanism(s) of toxicity may be similar (common) or
independent. For the purpose of this document, the CAG is the group of chemicals that
will undergo full risk analysis (the Dosimetry-Based Cumulative Risk Assessment;
DBCRA) and may not include all chemicals originally considered.  For some chemicals,
data on exposures and/or dose-response  may be so limiting, or a comparison of
anticipated exposures to risk levels may reveal such a gulf that these chemicals would
be "de-selected" for inclusion in  a DBCRA. In some instances, this situation may be
relevant to the entire mixture. For those chemicals or mixtures, only the conduct of a
screening level analysis may be warranted.  Therefore, steps prior to the determination
of the CAG may be thought of as an "Initial Analysis" to determine whether the
exposures to a chemical warrant the DBCRA.  DBCRA is an analysis that includes
evaluation of and inclusion of tissue or organ-specific levels of active toxic agent.
Tissue levels of toxicant are employed in the determination of toxicokinetic interactions,
i.e., metabolic competition, and tissue levels of toxicant are employed in the prediction
of toxic  events (toxicodynamics). In some cases,  the Initial Analysis may be sufficient.
For instance, the likelihood of a  positive response predicted by additivity my not be of
concern, and there may appear  to be no reason to suspect an interaction (for example,
if the Target Organ Toxicity Dose-based Hazard Index is appreciably below unity). If
1 In 2002, the U.S. EPA, Office of Pesticide Programs (OPP) finalized guidance that would be relied upon
to conduct a Congressionally-mandated assessment of the cumulative risks of organophosphate and
other pesticides, including the risks to children.  The guidance specifically addressed the development of
an FQPA Safety Factor for the mixture of pesticides.  The guidance document itself indicates that the
intent of the document was not to serve as a guide for the conduct of all cumulative risk assessment
approaches, within or without OPP. Specific provisions were included so that risk assessors may choose
to depart from specific guidance when data and circumstances warrant doing so.  This guidance has
been subjected to Agency and Public review and is freely available on the World Wide Web.

                                       13

-------
there is insufficient evidence to further support a DBCRA, then the results of the Initial
Analysis may be sufficient, and embarking on a DBCRA may not be deemed necessary.
Alternately, the screening hazard index approach—that of assessing cumulative risks
on the  basis of external (instead of internal) concentrations/doses, may apply (U.S.
EPA, 2000a). Thus, the hazard analysis of chemicals in a mixture during the "Initial
Analysis" requires not only a determination of the population risk presented by the
chemical alone, but any additional risks that would occur due to interactions in the
mixture. The 10-step process communicated in this document (Figure 1) is divided into
the two broad activities, Initial Assessment and the conduct of the Dosimetry-Based
Cumulative Risk Assessment.  In this process, the CAG is determined in step 5; steps
6-10 describe the specific activities that comprise the DBCRA.
      A biologically based analysis of risks including evaluations of tissue dosimetry
can be very resource-intensive and this is the main reason why the "Initial Analysis" is
carried out to avoid investing resources into analyses that are not worthwhile.  Likewise,
development of PBPK models, particularly interactive PBPK models (PBPK models
capable of addressing mixtures of chemicals, where additive and non-additive
interactions at the level of pharmacokinetics and/or pharmacodynamics can  be
addressed) for chemical mixtures, would require additional resources.  Thus, it is
important to apply a screening method (initial analysis) to eliminate from further
consideration and development of a DBCRA (a CRA whose exposure and dose
response evaluations are based on tissue concentrations of toxicant, rather than on
externally encountered concentrations/doses) those chemical mixtures anticipated not
to cause significant cumulative risk and/or those mixtures for which the available data
would not support a credible DBCRA. Accordingly, we recommend that the cumulative
risk assessment (here, for drinking water contaminants) be conducted in a two-phase
framework (1) Phase I: Initial Analysis; and (2) Phase II: Dosimetry-Based Cumulative
Risk Assessment. PBPK models would be developed for the components that are
considered in Phase II, i.e., those that are in the CAG. This approach stresses the
importance of the initial analysis to eliminate those situations that do not warrant a full-
scaled  cumulative risk assessment thereby reducing the burden of the Public in
conducting dosimetry-based cumulative risk assessment.
      The 10 steps described in the U.S. EPA guidance document on cumulative risk
assessment (U.S. EPA, 2002a) from the Office of Pesticide Programs remained
unchanged under our proposed two-phase approach because: (1) the steps have been
established, reviewed, and appear useful as a starting point, (2) we would
                                      14

-------
                                                Initial Analysis
       1. Group chemicals
         by similar target
         organ,
         pharmacokinetics
         and mechanism
         (CMC)
         (Section 2.2.1)
 , Identify
  potential  	^
  exposures
  (Section 2.2.2)
3. Characterize
  health
  endpoints for
  evaluation
  (Section 2.2.3)
4. Determine need for    5. Identify
  dosimetry-based    	
  cumulative risk
  assessment
   • Hazard index
   • Target organ
     toxicity dose
   • Weight of evidence
  (Section 2.2.4)
chemicals in the
CAG for further
evaluation
(Section 2.2.5)
en
       6. Characterize
         dose-response,
         point of departure
         and relative
         potency as a
         function of tissue
         dose
         (Section 2.3.1)
                              Dosimetry-Based Cumulative Risk Assessment
7. Develop      8. Quantify
  detailed  	    parameters for
  exposure        exposures to
  scenarios        be evaluated
  (Section 2.3.2)   (Section 2.3.3)
                   9. Conduct dosimetry-     10. Characterize the
                     based CRA
                      • Develop
                        pharmacokinetic
                        analysis
                      • Predict responses
                        as a function of
                        tissue dose
                        (Section 2.3.4)
                          cumulative risk
                          estimated via
                          pharmacokinetic
                          analysis
                          (Section 2.3.5)
                                                    FIGURE 1
                            Data Evaluation for Dosimetry-Based Cumulative Risk Assessment

-------
like to maintain consistency with the published U.S. EPA approach (U.S. EPA, 2002a);
and (3) for any chemical mixture which is to undergo dosimetry-based cumulative risk
assessment, it would have gone through the first five initial screening steps. Our
framework/approach simply organizes the 10 steps into two phases to underscore the
rationale for introducing PBPK modeling at the 6th step.

2.2.   PHASE I: INITIAL ANALYSIS
      In this phase, the principal purpose is to determine, from all available information
on the chemical mixture and its exposure scenarios, whether a dosimetry-based
cumulative risk assessment is warranted.  Central to this determination or decision are
the answers to the following questions:

   •  Is there the likelihood of toxicological interactions among the chemicals in the
      mixture? At what concentration levels?
   •  Are these toxicological interactions likely to cause higher toxicity (i.e., synergism
      or potentiation) or lower toxicity (i.e., antagonism) than the simple addition of the
      toxicity from all components of the mixture?
   •  Under the exposure conditions and levels, are these toxicological interactions
      possible?

2.2.1. Stepl. Identify Common Mechanism Group (CMG). A cumulative risk
assessment begins with the identification of a group of chemicals, a Common
Mechanism Group (CMG), that induce a common toxic effect by the same mechanism
of toxicity.  When chemicals are identified, data on their interactions should be sought.
Two sources for such data are the general toxicological literature and the Interactions
Profiles series developed by ATSDR (httgV/wwv^atsdrcdc.                  les/).  An
approach developed to guide OPP through conducting a mandated cumulative risk
assessment of organophosphate pesticides (U.S. EPA, 2002a) refers to the definition of
mechanism of toxicity from the Food Quality Protection Act (FQPA): "the major steps
leading to an adverse health effect following interaction of a pesticide with biological
targets." However, this definition is somewhat ambiguous regarding whether the
interactions between the chemical and the biological target are PK or PD in nature.
Existing U.S. EPA guidance (U.S. EPA, 2002a) considers this issue in Step 3 of the
CRA process.  However,  if a chemical is not included in the CMG group because it does
not share relevant elements of PD processes with other chemicals in the group,  but the
chemical in fact can interact with other chemicals in the group through PK processes,
                                      16

-------
an underestimation (or overestimation) of the actual cumulative risk may occur.  For
example, suppose four chemicals (Chemicals A, B, C, Q) are being considered for
inclusion in a CMC based on Effect X. Chemical Q does not cause Effect X, but Q is an
inhibitor of metabolism of Chemical A, and this inhibition causes an increase in the
Effect X for the same exposure to Chemical A (see also Section 1.6).  If Chemical Q is
omitted from the CMC due to lack of Effect X, the CRA will not properly characterize the
cumulative  risk for Chemicals A, B, C when Chemical Q  is also present. For this
reason, all chemicals that share common types of PK or PD interactions with biological
targets should be considered as members of the CMC. This represents a specific
extension of the process, as recommended by the available (OPP) guidance to address
commonalities beyond those related to toxic manifestation: those  pertaining to tissue
doses of toxicologically active chemical species (metabolites). While OPP's guidance
defines the CMC based only on toxicodynamic events in the mechanism of action itself,
The CMC is here defined differently.  Our definition is broadened to include
commonalities or interactions in toxicokinetics that may alter tissue dosimetry and
modify the toxic response. This has been done to optimize the application of PBPK
modeling to assess cumulative risk.
      For this project, the first mixture consists of 6 OP pesticides: methyl-parathion,
parathion, chlorpyrifos, fenthion, diazinon and fenitrothion.  For one effect of concern,
neurotoxicity, the common mechanism of these pesticides is inhibition of AChE. Thus,
the mechanism of toxicity  includes a common pharmacodynamic process.  However,
many of these OP pesticides are also metabolized by common enzymes, primarily in
the liver, including cytochrome P450s and esterases in various tissues. Therefore, this
CMC includes chemicals that have similar pharmacokinetics, pharmacodynamics and
target organs. During the CRA process for OP pesticides, the risk assessor will
obviously also consider other toxicity endpoints, such as developmental neurotoxicity,
before deciding which endpoint is most relevant. For at least some of the OP pesticides
being considered here, there is some evidence that exposures during pregnancy or
during childhood development  may be important, because development is  occurring
and chemical may be transferred from the mother (e.g., placental  transfer or via breast
milk), because the child may have increased  sensitivity, or because the child may
exhibit pharmacokinetic differences from adults that lead to greater toxicity (Chanda and
Pope, 1996; Qiao et al., 2001).  Considerations of potentially susceptible
subpopulations add to the complexity of PBPK modeling, when subpopulations exhibit
anatomic, physiologic and biochemical differences compared to the general population.
                                      17

-------
      The second mixture consists of four chlorinated hydrocarbons or volatile organic
chemicals that are often present in drinking water as disinfection byproducts or
groundwater contaminants: TCE, PERC, MC and chloroform (CHF).  The mechanisms
of toxicity for these chemicals are somewhat different depending on the type of toxic
endpoints.  For instance, at higher doses, particularly through inhalation exposure, all of
these chemicals cause acute central nervous system (CMS) depression through the
common mechanism of interference in ion channel function (Bruckner and Warren,
2001). In general, narcotic effects by solvents are caused by the following mechanisms:
(1) membrane expansion and associated conformational changes of proteins (Shibata
et al., 1991); (2) interaction with macromolecules (e.g., membrane lipids and proteins)
and associated changes of membrane fluidity (Ueda and Kamaya, 1984) and (3)
interaction with ligand gated ion channels (Franks and Lieb, 1994).  Specific
mechanisms for such effects by the TCE, PERC, MC and CHF have not been revealed
yet.  Since we are specifically interested in the exposure to contaminants found in
drinking water, the likely exposure levels are low. Thus, acute neurotoxicity is highly
unlikely and does not need to be considered as an endpoint. However, a hazard
quotient for such an effect could be developed. For volatile compounds, exposure
should also consider volatilization and inhalation of compounds during water use.
      Long-term, low-level exposure to this mixture in drinking water is a highly likely
scenario. Therefore, in addition to other chronic effects, the possible carcinogenic
potential of this mixture as a critical effect should also be considered.  In that sense,
damage to cellular targets from reactive metabolites from this group of chemicals may
be considered as a common mode of action. However, the specific targets of PD
mechanisms that are the critical determinants of toxicity at low dose may vary from  one
chemical to another in this group.  All of the four chemicals to be considered here share
common elements of PK processes, namely shared metabolic pathways by CYP2E1,
glutathione S-transferase (GST), and possibly other enzymes.  These enzymes detoxify
some of the volatile organic chemicals and their metabolites, while bioactivating others.
Based on the discussion above, all four chemicals should be considered to be part of
the CMC based on the potential for PK and/or PD interactions.

2.2.2. Step 2. Identify Potential Exposures. For each CMC member, evaluate
proposed and registered uses and use patterns to identify  potential exposure pathways
(i.e.,  food, drinking water, residential) and routes (oral, inhalation, dermal). For this
project, the primary exposure is through consumption of drinking water although the
potential exists for inhalation and dermal exposure (i.e., during showers).  Additionally,
                                      18

-------
pesticide (and other chemical) exposures from food intake and dust or residue through
home application remains a potential.  In this latter case, for instance, Whyatt et al.
(2002) reported that, of the 72 pregnant women residing in northern Manhattan and
South Bronx who underwent personal air monitoring for 48 hrs during their third
trimester, all (100%) had detectable levels of three insecticides: chlorpyrifos, diazinon,
and propoxur.  Fenske et al. (2002) reported that chlorpyrifos was measurable in the
house dust of all homes they monitored in a central Washington State agricultural
community; children in these homes had detectable urinary metabolite of chlorpyrifos.
Here, if the DBCRA is to be conducted, the PBPK modeling has the advantage of being
able to carry out computer simulations with a variety of exposure scenarios and
assumptions with relative ease. This translates external concentrations to internal
(target tissue) doses, representing a valuable advance in dose-response  evaluation for
animal studies, for species extrapolation  of dose-response, and for the purpose of
providing a more useful characterization  of human exposure.
      In Step 2, members of the CMC should be assessed for the potential for
significant and/or overlapping exposures. Various sources of  information on the range
of levels of each chemical in public and private drinking water supplies should be
consulted.  When analyzing these data, the risk assessor will confront one of the
conundrums to be encountered in this CRA: data on contaminants are often expressed
only as ranges observed, with  no information indicating which co-occur and which do
not.  A CRA can be based on these data, assuming that exposure to all of the chemicals
coincide. However, it is likely that there is some correlation between the exposure data,
such that some exposures do not coincide.  Suppose that the  concentration of Chemical
A ranges from 0.01 to 1.2 mg/L in various drinking water supplies and that the
concentration of Chemical B ranges from 0.005 to  1.3 mg/L, with no data available to
demonstrate a correlation between the two.  Should a worst-case CRA be performed at
the upper limit of exposure data, or at some confidence interval based on those data?
Such an approach would likely overstate the cumulative risk if, as is likely the case, the
geographical coincidence of maximal exposures is not complete. An alternate approach
would consider using a geographical scenario-based approach.  Such an approach
could perform a CRA based on regional exposure data, such as was previously
performed for OP pesticides (U.S. EPA, 2002b). A further alternative could take a point
sampling approach.  In this approach, specific risk assessment calculations could be
performed based on exposure data at discrete locations or at discrete points in the
concentration distributions (see US EPA, 2006). Statistical analysis of the results could
then be performed on the CRA output to  determine the percentages (and locations) of
                                      19

-------
the population that are above various hazard indices.  This approach would involve an
iterative approach; the information addressed in step 10 would communicate which
subpopulations and to what extent their risk may be increased above a given hazard
index.  We note that these latter approaches are more calculation-intensive, and the
ability of the risk assessor to perform this more detailed analysis would depend on
resources.  We also note that, to the extent that interactions between chemicals are
found to occur, a non-PBPK modeling approach would have difficulty in accurately
determining the cumulative risks.  This would be due, primarily, to a lack of mechanism
to address the interactions of chemicals within the body, which may occur between the
portal of entry and the target tissue.  However, once validated interaction-based PBPK
models have been developed, it is a relatively simple matter to implement them for any
number of exposure-concentration scenarios (including the point sampling approach);
this could even be automated in such a way that a large exposure dataset could be
evaluated.
      The potential exposures to each component in the mixture under as many
exposure scenarios as possible (primarily use of drinking water and showering) should
be evaluated. This analysis should include exposure information that may vary by
season and geographic location.  A formalized approach, similar to the five-step
process developed by ATSDR (2005) may be undertaken. This systematic approach
includes: (1) identification of contaminant source(s), (2) evaluation of environmental fate
and transport, (3) identification of point or area of exposure, (4) identification of
exposure route(s) and (5) characterization  of the exposed population.

2.2.3.  Step 3. Characterize and Select Health  Endpoint(s) for Evaluation. For
each CMC  member, evaluate common health effects and affected tissues and affected
organs attributable to the common mechanism of toxicity across all exposure routes and
durations of interest, determine the time-frames of expression for the common toxicity,
and evaluate the quality of the dose-response data for each CMC member.
Recommend  endpoints/species/sex that can serve as a uniform basis for determining
relative potency. In the case of the OP pesticide mixture, the  most reasonable endpoint
to use should be the inhibition of brain AChE (because they are neurotoxic through this
mechanism).  This endpoint should be correlated with clinical  signs which are usually
available in the toxicology literature. During this step, attention must be given to
subpopulations or life stages that may be more susceptible to the common toxic effect
and mechanism (here, inhibition of erythrocyte AChE activity). For example, infants and
children may not have fully developed metabolic pathways for detoxifying or
                                      20

-------
bioactivating chemicals in a common mechanism grouping (U.S. EPA, 2002a).  The
importance of describing the potential increased sensitivity of infants and children is
described in Executive Order 13045, and U.S. EPA's guidance is provided in EPA's
Rule Writer's Guide to Executive Order 13045: Guidance for Considering Risks to
Children During the Establishment of Public Health-Based and Risk-Based Standards
(U.S. EPA, 1998).
      The selection of endpoints requires a consideration, at least semi-quantitatively,
of the time-course of biological effects so that overlapping pharmacokinetics can be
considered.  Thus, it is necessary to consider "biologically effective dose."  This would
represent the concentration of toxicant in the target tissue; and in the case of the
drinking water disinfection byproducts and contaminant mixture,  could be taken as the
area-under the concentration vs. time curve (AUC) of the reactive species or related
dose metrics.
      Another issue to be determined is the choice of sex and age. Traditionally, risk
assessments and PBPK models are most commonly developed for adult male humans.
This is probably because of the preference for performing initial studies in male rodents
(because the male may be a simpler model from an endocrine standpoint). However,
U.S. EPA has been clear in its position that work should not stop with the male model
and should be extended to females as well. Thus, the logical process may be as
follows: determine the sex for which the most data are available to support the DBCRA
(PBPK models).  Initially develop PBPK models (and DBCRAs) for this sex. This would
likely be the male model. Subsequently, extend this to female using as much female-
specific data as possible and extrapolating from pooled or male data as required.
Discuss uncertainty in this context, i.e., there  may be more certainty with the  initial sex
selected and more uncertainty with the latter sex.  Specific anatomic, physiologic and
biochemical factors thought to differ among sexes should be identified. These issues
apply to both the PBPK modeling and the DBCRA itself.
      The same thoughts apply to the issue of age. At this stage, the PBPK models
and the DBCRA would probably be applied to adults, and possibly to children, although
it would certainly be  preferable if other age groups could be considered as well. Again,
both PBPK models and DBCRA analysis will probably be supported by more data for
adults than for children. Therefore, the models and DBCRAs should be  initially
conducted for adults and subsequently extended to children. These issues (issues
relating to potential differences in sensitivity among subpopulations) should be kept in
consideration while determining the common  mechanism endpoint.
                                      21

-------
2.2.4. Step 4. Determine the Need for a Dosimetry-Based Cumulative Risk
Assessment.  The relationship between exposure and acceptable exposure limits
should serve as the trigger in the decision whether to proceed with conducting a more
technically based dosimetry-based cumulative risk assessment (DBCRA). The
guidelines established by the U.S. EPA (1986) describe a Hazard Index (HI) screening
approach that represents the first step in this evaluation (U.S. EPA, 1989).  In that
approach, the exposure (E) and the acceptable level (AL) are determined for routes of
exposure and expressed in the same units. That approach favors AL measurements
(reference dose [RfD] and reference concentration [RfC] values) available on the U.S.
EPA's Integrated Risk Information System  (IRIS). However,  in the absence of such
values, other reliable values (e.g., ATSDR  Minimal Risk Level values) may be relied
upon, or values may be developed from original toxicity findings in the absence of more
reliable information. A Hazard Quotient (HQ; E/AL) is determined for each component
of the mixture and a Hazard Index for the mixture is determined as the sum of the HQ
values. When a group of chemicals is targeted for a CRA, a  HI should be developed for
the entire group, and HI values should also be developed for each CMC, independently.
A hypothetical example of a four chemical mixture (chloroform, trichloroethylene,
1,1,1-trichloroethane and tetrachloroethylene) is presented below in Table 1.  The
exposures are hypothetical, for demonstration purposes only.
      This is  a conservative approach, based on the assumption of dose additivity for
the mixtures components, and unconstrained in that the AL values are not segregated
by target organ or tissue.  Although there is no "bright line," as HI values increase above
1.0, risk from the mixture is considered to increase. Given the conservative nature of
the screening approach, if the HI is less than 1.0, additional methods to  calculate HI
should be considered. Subsequent to the development of this approach (U.S. EPA,
1986), additional works have suggested the employment of a weight of evidence (WOE)
determination to evaluate the potential for toxicological  interactions2 (ATSDR, 2004;
Mumtaz and Durkin, 1992; Mumtaz et al., 1998; U.S. EPA, 2000a); a WOE exercise  is
recommended by the U.S. EPA in its Supplementary Guidance for Conducting Health
Risk Assessment of Chemical Mixtures (U.S. EPA, 2000a). The WOE serves as the
basis for the decision whether component chemicals interact to alter (increase or
decrease) the toxicity of other components of the mixture.  When HI values are below
1.0, the WOE will serve as the basis for further consideration of the mixture, when the
2 A Weight of Evidence Determination is a judgment reflecting the quality of the available information that
categorizes the most plausible nature of any potential influence of one compound on the toxicity of
another compound for a given exposure scenario (U.S. EPA, 2000a).

                                       22

-------
TABLE 1
Screening Hazard Index Approach
Chemical
Chloroform
Tetrachloroethylene
Trichloroethylenea'b
1,1,1-Trichloroethaneb'c
Exposure (mg/kg-d)
0.0025
0.005
0.0025
0.05
RfD (mg/kg-d)
0.01
0.01
0.04
0.5
Hazard Index:
Hazard
Quotient
0.25
0.5
0.06
0.1
0.91
 IRIS RfD values were not available.
b RfD value was not available on IRIS; it was developed from a chronic no-observed-
adverse-effect level (NOAEL) value for renal effects (ATSDR, 1997a) and an
uncertainty factor of 1000.
c RfD value was not available on IRIS at the time of this report's completion; it was
developed from a chronic lowest-observed-adverse-effect level (LOAEL) value for body
weight reduction (ATSDR, 1995) and an uncertainty factor of 1000.
                                      23

-------
HI value alone would suggest that further consideration is not warranted. Thus,
developing a weight of evidence increases the basis for decision making. Some points
that should be considered in developing the WOE include:

   •  Information on the chemical mixture or combinations of components that can
      inform the magnitude and direction of toxicologic interactions,
   •  Information on mode or mechanism of action of component chemicals, and
   •  Information on mode or mechanism of action of related chemicals.

      While the U.S. EPA (2000a) has codified a technical approach to refining the HI
by incorporating mathematical values for WOE and additional factors, incorporating
such technical guidance here is beyond the scope of the intended exercise.
      When the HI for the mixture exceeds 1.0 or when the HI is considered in light of
the WOE and cannot without uncertainty be reduced to a value below 1.0, then the
mixture should be subjected to a more technical estimation of HI, based on the Target
Organ Toxicity Dose (TTD) approach (ATSDR, 2004; Mumtaz et al., 1997; U.S. EPA,
2000a).  In this approach, effect levels for multiple organs are compiled and combined
with uncertainty factors to develop TTD values for each organ system affected whether
the effect is the critical effect or a secondary effect.  These TTDs are analogous to  RfD
values.  HQ values are derived, but the AL values can be either TTD values or RfD
values for the most sensitive tissue or organ. HQ is determined as, E/TTD; HI values
are developed for each organ, by summing the HQ values for each chemical, within the
organ. This approach is more technical in that it incorporates organ-specific measures
of toxicity, expanding considerations of toxicity from only the critical target to additional
target systems, and it ensures that major systems affected are taken into account.  In
the tables that follow, chronic oral NOAEL values for effects were taken from ATSDR
Toxicological Profiles for the four respective chemicals (notable  exceptions are that the
IRIS RfD values for chloroform and tetrachloroethylene, where liver was the critical
target, were incorporated). Information on chronic NOAEL values were duration-
adjusted (days/week) when necessary, and uncertainty factors (10, 100 or 1000) were
employed to adjust for animal to human extrapolation, human interindividual  variability
and,  in some cases, subchronic to chronic duration). Tables 2, 3, 4 and 5 present the
derivation of TTD values for chloroform, tetrachloroethylene, trichloroethylene and
1,1,1 -trichloroethane, respectively.
                                      24

-------
TABLE 2
Derivation of TTD Values for Chloroform (ATSDR, 1997b)
Target
Species
NOAEL (mg/kg-d)
N GAEL Adjusted
Uncertainty Factors
TTD (mg/kg-d)
Resp
mouse
60
51.4
100
0.51
cv
dog
30
25.7
100
0.26
Gl
rat
200
142
100
1.4
Hemat
human
21
21
10
2.1
Muse
rat
200
143
100
1.4
Liver
human
0.96
0.96
10
0.10
Kidney
human
0.96
0.96
10
0.10
Repro
dog
30
25.7
100
0.26
b.w.
rat, dog
30
30
100
0.30
cn
    Resp = respiratory system
    CV = cardiovascular
    Gl = gastrointestinal
    Hemat = hematological
    Muse = musculoskeletal
    Repro = reproductive system
    b.w. = body weight reduction

-------
TABLE 3
Derivation of TTD Values for Tetrachloroethylene (ATSDR, 1997c)
Target
Species
NOAEL (mg/kg-d)
NOAEL Adjusted
Uncertainty Factors
TTD (mg/kg-d)
Resp
rat
941
672
100
6.7
CV
rat
941
672
100
6.7
Gl
rat
941
672
100
6.7
Hemat
human
NR



Muse

NR



Liver
rat
20
14.3
1000**
0.01
Kidney
mouse*
386
275
1000
0.28
Repro

NR



b.w.
rat
941
672
100
6.7
O)
    * LOAEL
    ** IRIS UF = 1,000.  10 each for UFS, UFA , UFH
    NR = not reported
    Additional abbreviations are listed with Table 2.

-------
TABLE 4
Derivation of TTD Values for Trichloroethylene (ATSDR, 1997a)
Target
Species
NOAEL (mg/kg-d)
N GAEL Adjusted
Uncertainty
Factors
TTD (mg/kg-d)
Resp
rat
250
179
1000
0.18
cv
rat
250
179
1000
0.18
Gl
rat
250
179
1000
0.18
Hemat

NR



Muse
rat
250
179
1000
0.18
Liver
rat
250
179
1000
0.18
Kidney
rat
50
35.7
1000
0.036
Repro
rat
NR

1000
0
b.w.
rat
250
179
1000
0.18
NR = not reported
Additional abbreviations are listed with Table 2.

-------
TABLE 5
Derivation of TTD Values for 1,1,1-Trichloroethane (ATSDR, 1995)
Target
Species
NOAEL
NOAEL
Adjusted
Uncertainty
Factors
TTD
Resp
rat
1500
1071
100
10.7
CV
rat
1500
1071
100
10.7
Gl
rat
1500
1071
100
10.7
Hemat
rat
1500
1071
100
10.7
Muse
rat
1500
1071
100
10.7
Liver
rat
1500
1071
100
10.7
Kidney
rat
1500
1071
100
10.7
Repro
rat
1500
1071
100
10.7
b.w.
rat
750*
536
1000
0.54
Immuno/
Lymphat
rat
1500
1071
100
10.7
00
     * LOAEL
     Additional abbreviations are listed with Table 2.

-------
      Employing TTD values, rather than simply relying on RfD values gives a more
complete treatment. For example, liver was the critical target for chloroform and
trichloroethylene, but kidney and body weight gain were the critical endpoints for
trichloroethylene and 1,1,1-trichloroethane, respectively.  When the HI is calculated
based on segregated values (i.e., the TTD approach), tissue dose-response information
is more optimally employed. While liver served as the RfD values for two of the four
compounds, the HI value changed from 0.91  under the screening approach (Table 1) to
0.77 for liver under the TTD approach (Table 6), reducing concern for toxicological
interaction. This change was due to  liver not being the most sensitive target for
trichloroethylene or 1,1,1-trichloroethane. Further, the TTD based HI appears quite low
for the other potentially affected organs/tissues,  serving to focus attention on the liver as
the site most likely affected. Again, a weight of evidence should be developed, but
specifically for liver effects, to further refine the decision whether to proceed to a full
DBCRA for this mixture.  This should include the number and types of possible
exposure scenarios in conjunction with the associated information on the concentration
of contaminant in food and environmental media available.  The NOAEL and LOAEL
values for the common health endpoints and target tissues should be compiled. This
evaluation may suggest that an initial analysis for the CMC will indicate that there is no
risk concern for this group of chemicals and no further detailed assessment will be
necessary. The available information may also suggest that expending resources on a
dosimetry-based cumulative risk assessment may not be warranted.
      One possible scenario is that,  for some or all of the component chemicals in the
mixture of interest, there is insufficient data or a  lack of data. In this  case, depending on
the importance of the mixture of interest to public health,  the U.S. EPA may decide: (1)
there are insufficient data to conduct a cumulative  risk assessment, and the component
chemicals in the mixture would be assessed by single chemical risk assessment
methods; or (2) new research initiatives are to be launched to fill the data gaps for the
specific purpose of conducting dosimetry-based cumulative risk assessment for the
mixture of interest.
      A thorough evaluation of the toxicological literature for each chemical in the CMC
must be performed. This may be performed in general accordance with risk
assessment guidance for single chemicals because interactions between chemicals are
considered separately.  The type and quality of toxicological data should  be assessed.
Moreover, the data must be evaluated in relation to the species, gender and age of the
animal as surrogates for humans. Ultimately, studies must be identified in which data
describing doses and quantified responses are presented in sufficient detail to serve as
                                       29

-------
TABLE 6
Target Organ Toxicity Doses, Hazard Quotients and Hazard Indices

Chloroform

TTD
Hazard Quotient
Tetrachloroethylene

TTD
Hazard Quotient
Trichloroethylene

TTD
Hazard Quotient
1,1,1-Trichloroethane

TTD
Hazard Quotient
Hazard Index
Exposure
0.0025


0.005


0.0025


0.050



Effect
Resp

0.51
0.0049

6.7
0.00074

0.18
0.014

11
0.0047
0.024
CV

0.26
0.0097

6.7
0.00074

0.18
0.014

11
0.0047
0.029
Gl

1.4
0.0018

6.7
0.00074

0.18
0.014

11
0.0047
0.021
Hemat

2.1
0.0012







11
0.0047
0.0059
Muse

1.4
0.0018

0


0.18
0.014

11
0.0047
0.020
Liver

0.01*
0.25

0.01*
0.5

0.18
0.014

11
0.0047
0.77
Kidney

0.096
0.026

0.28
0.018

0.04
0.063

11
0.0047
0.11
Repro

0.26
0.0097







11
0.0047
0.014
b.w.

0.3
0.0083

6.7
0.00074

0.18
0.014

0.54
0.093
0.12
CO
o
     * Liver is the critical organ for chloroform and fortetrachloroethylene. RfD values were taken from IRIS (U.S. EPA, 2007b). All other chronic TTD

     values (NOAEL values) were taken from ATSDR toxicological profiles for the respective chemicals.

-------
the basis for the dose-response relationship.  If the PBPK modeling approach will
ultimately be used, PK or PD studies will have to be identified that provide information
on the dose-response of the animal where the response is expressed relative to
appropriate markers of exposure.  These types of data can be then transformed to
determine the appropriate NOAEL or benchmark dose (BMD) for each chemical.
      For the two mixtures considered in this project, as an illustration below, the three
questions are answered and related discussions are provided.

   •  Is there the likelihood of toxicological interactions among the chemicals in the
      mixture? At what concentration levels?

      As shown in Figures 2 and 3, respectively for the metabolic pathways of OP
compounds (Mixture 1) and the drinking water disinfection byproducts and contaminants
(Mixture 2), there are several possible pharmacokinetic interactions.  All  OP compounds
share virtually the same metabolic pathways  involving: (1) cytochrome P450 mediated
oxidative desulfuration, dearylation, or dealkylation;  (2)  calcium-dependent hydrolysis of
phosphoester bond; and (3) GST conjugation reactions. Similarly, there are interwoven
metabolic pathways in the biotransformation of TCE, PERC, MC and CHF (see
Figure 3). At the pharmacodynamic level, various oxon derivatives would certainly
compete for the inhibition of brain AChE; thus, toxicological interactions are highly likely.
Cancer and noncancer effects have been observed following exposure to the
component chemicals of Mixture 2.  During the CRA screening process,  both cancer
and noncancer effects, including reproductive effects, should be reviewed for potential
(non-additive) interaction and significance. Three (TCE, PERC,  CHF) of the four
chemicals are thought to be carcinogenic at some doses; there is limited evidence that
MC could be carcinogenic at this time. Thus, if cancer is the endpoint and MC is
regarded as non-carcinogenic, the other three chemicals will be contributors to the
overall risk, but the potential for MC to affect the carcinogenic potency of the other
chemicals through PK interaction needs to be evaluated. The appropriate dose metrics
for cancer risk assessment for TCE, PERC and CHF are not clearly established, but
generally involve metabolites such as trichloroacetic acid or phosgene. There is limited
information  regarding potential pharmacodynamic interactions between these chemicals
that can be addressed  within a CRA. On the other hand, there have been several
studies of pharmacokinetic interactions that will be useful (Dobrev et al.,  2001, 2002;
Haddad et al., 2000; Thrall and  Poet, 2000).  Indeed, for some of the mixtures of these
four chemicals, PBPK interaction models exist (see below).  A more likely scenario
                                      31

-------
                Y
                   O
                    OH
YOx\p*S*

    OZ
                           1
  OZ
 I

V
 OZ
                                    o
                                     OH
0
Y
 OZ
                                   \l
                                   0


                                    OZ
                                                  T"
                                                   OH

                                                  Binding
                                                                AChE
                                                     OZ
                                 FIGURE 2
A Generalized Metabolic Schematic for OP Pesticides. Step 1 is activation and all other
steps are detoxication. Pharmacokinetic interactions are most likely during Step 1.
Interactions in other steps can be evaluated using PBPK models.  Pharmacodynamic
interactions may also occur during binding of the active moiety to AChE (Step 2).
                                    32

-------
C'H-
 H    SG
  GSH
a    a

Cl    H
           CYP450
  GSH
   Cl
 CYP450
   I
   Fe

   6" Cl

H—^C+
                                    CHO
                   ;'HH
      SG
 ci.
   ^
 Cf    S  NH3+


S-1,1-DCVC  COO-
 OH

-f'
 OH




   Cl

   Cl
                                                   TCA
                                                      .0


                                                       OH
                                                TCOH
HH
    S   NHCOCHj


        COOH

     S
        ^^M

     H
                                                                  Cytotoxicity
                                                                  Mutagenicity
                                                               Excretion
                                                               in Urine
                                                                 .0


                                                                  OH
                     DCA

                         /
                    glue—O

                     TCOG
                                                               Excretion
                                                               in Urine
                                                                      cc\
                                                          Cytotoxicity
                                                          Mutagenicity
       Potential interaction between PERC and TCE

   2   Potential interaction between MC and TCE

    )  Potential interaction between CHF and TCE
                                   FIGURES
Metabolic Pathway for Solvents Indicating Potential Sites for Interactions. This is
superimposed on a metabolic schematic for TCE (Dobrev et al., 2001), metabolic steps
where pharmacokinetic interactions could occur are denoted by numbers as shown.
Identification of additional potential interactions should occur during CRA Steps 3 and 6.
                                      33

-------
would be that some of the components could be excluded from the DBCRA (not from
the cumulative risk assessment; for these chemicals, only a screening level assessment
may be justified). It is important to avoid the expenditure of extensive resources to
develop PBPK models for all components, only to find that the models indicate that a
given component could have been omitted from the analysis based on a careful
evaluation of data presented in the Initial Analysis.
      As suggested by Figure 3, there are several common metabolic processes that
overlap forTCE, PERC, MC and CHF. All are initially oxidized by Cytochrome P450,
mostly by CYP2E1. Several studies have shown competitive inhibition at this step
(Koizumi et al., 1982;  Dobrev et al., 2001). Other steps are common to two or more of
the chemicals, including metabolism by GST and (3-lyase family enzymes. Potential
interactions at these stages should be considered.
      These interactions are likely to happen in vivo at the tissue levels of
approximately nmolar to umolar range which generally correspond to a mmole/kg
administered dosing range (i.e., see Dobrev et al., 2002).
      At this point, we believe that it is appropriate to address the toxicological
interaction issue using a specific example reported in the earlier attempt of cumulative
risk assessment of OP pesticides. In the Guidance on Cumulative Risk Assessment of
Pesticide Chemicals That Have a Common Mechanism of Toxicity (U.S. EPA, 2002a),  it
was assumed that at lower levels of exposure typically encountered environmentally, no
chemical interactions are expected (i.e., simple dose/concentration additivity would
apply). For additivity to hold true, a further assumption must be that all the common
mechanism chemicals exhibit similar pharmacokinetic and pharmacodynamic
characteristics (i.e., distribute to the same target tissue and have  similar elimination
half-lives; exert toxicity via similar biochemical mechanisms,  without specific regard to
which organ is the critical target). In reality though, a case study  of cumulative risk
assessment of 33 organophosphorus pesticides provided BMDL (lower bound
benchmark dose at EDi0) with a range of 3977-fold (female)  to 5528-fold (male)
difference between the highest BMDL (for malathion) to the lowest BMDL for
(dicrotophos) (U.S. EPA, 2002b; Table I.B.-4).  These 3-4 order of magnitude
differences among "common mechanism chemicals" suggest that the PK and PD may
each be quantitatively different.  For example, if each compound  is active through its
oxon form, and each has a similar potency, then one may expect a 3-4 order of
magnitude difference in PK, which seems unlikely. If the PK of these two compounds
are similar,  then potency would have to differ some 3-4 orders of  magnitude.  Inasmuch
as neither potency nor PK may be envisioned to differ to that degree between these
                                      34

-------
chemicals, it seems logical to assume that both PK and PD differ between these
chemicals.  Differences in the extent of tissue exposure to toxicologically active
chemical moieties with markedly difference potencies can be anticipated.  These
toxicological interactions are discussed a bit further in a hypothetical example in the
following paragraph. Thus, the probability of toxicological interactions at the level of PK
and PD exists.
      Are these toxicological interactions likely to cause higher toxicity (i.e., synergism
      or potentiation) or lower toxicity (i.e., antagonism or inhibition) than the simple
      addition of the toxicity (dose addition or response addition) from all components
      of the mixture?
      With the OP-pesticide mixture, synergism/potentiation or antagonism/inhibition
are likely depending on the rates of pharmacokinetic or pharmacodynamic interactions.
For instance, all six OP compounds are phosphorothioates (i.e., P=S compounds); they
require activation to their respective oxon (i.e., P=0) compounds to become potent
AChE inhibitors.  This requires oxidative desulfuration mediated by cytochrome P450.
Depending on the substrates (e.g., the parent OP insecticides) that can compete among
themselves for P450-mediated oxidation (bioactivation) and the involvement of other
metabolic processes (e.g., competing enzymes that may conjugate the parent chemical
with glutathione, a detoxication reaction), more or less potent oxon compounds may
emerge (the oxon of one insecticide may be more potent than the oxon of another
pesticide in the mixture) as the predominant (most prevalent) metabolite.  If available,
information on this competition between parent chemicals for oxidation and competition
between different enzymes for bioactivation/detoxication will determine if
synergism/potentiation or antagonism may result.
      In addition to compound differences in activation and detoxication, differences in
the biological fate of the inhibition complex (Oxon-AChE) exist. Regardless of the OP
examined, once the oxon form becomes bound to the AChE enzyme, that enzyme
molecule is no longer capable of metabolizing acetylcholine. Once bound, other events
occur, leading to either destabilization and ultimate dissociation of the  inhibitory
complex or stabilization and aging of the inhibitory complex.  Each of the OP
compounds identified is an effective inhibitor, but they differ in the half-life of the
inhibitory complex with  some dissociating within hours whereas others may take days.
Once bound, the inhibitory complex may "age," and become permanent.  Once aging
occurs, the enzyme cannot be reactivated and AChE activity can only be restored with
                                       35

-------
the generation of new enzyme.  Each of these OP compounds have different inhibition
half lives and they differ in the extent to which they age, particularly so for dimethyl
versus diethyl substituted compounds.  When encountered in a mixture, compounds
which bind readily, but which have a short inhibition half-life and relatively reduced
potential to age, may actually be protective against AChE inhibition produced by other
OP compounds.
      Similarly, depending on the competitive inhibition of cytochrome P450 2E1 in the
initial metabolism of TCE by the other chlorinated hydrocarbon solvents/volatile
organics, the GST conjugation pathway may be favored (Dobrev et al., 2002); this
would lead to more production of reactive thiols for possible renal carcinogenesis—an
example of interactive toxicity.
      One of the greatest challenges in dealing empirically with chemical mixture
toxicology is the exponential rise of experimental groups as the number of component
chemicals and doses increase.  Frequently, systematic experimental studies of
chemical mixtures become untenable because of the limitation of resources. Here, the
possible application of reaction network modeling3 (Klein et al., 2002; Liao et al., 2002;
Reisfeld and Yang, 2004; Yang et al., 2004a,b) can offer a workable solution to the
complexity of chemical mixture toxicology, particularly relating to metabolism.

   •  Under the exposure conditions and levels, are these toxicological interactions
      possible?
      In the cases of pesticide applicators, total exposure comprises both occupational
exposures and  additional exposures from drinking water, foods and household residues.
For these individuals, total exposures may approximate the range where toxicological
interactions are possible (for Mixture 1). Similarly, occupational  exposures in chemical
industry or manufacturing plants  where TCE, PERC, MC and CHF are used, plus
additional exposures from drinking water, foods, showers, etc. could also render the
cumulative exposure concentrations to where toxicological interactions are  likely.

2.2.5. Step 5.  Identify Chemicals in the Candidate Cumulative Assessment Group
(CAG) for Further Evaluation.  Select chemicals, chemical uses, routes and pathways
3 Reaction Network Modeling: A chemical/petroleum engineering computer simulation technology to
model the entire oil refinery based on initial chemical analyses of the feedstocks, as well as using graph
theory, linear free energy relationship (LFER), computational quantum chemical calculations, quantitative
structural activity correlation, Monte Carlo and quardrature modeling techniques. It is capable of
simulating the fate of thousands of chemicals and tens of thousands of reactions simultaneously and
predicting the outcomes of these chemical reactions.

                                        36

-------
from the CMC that have an exposure and hazard potential to result in cumulative effects
(joint toxicity) for inclusion in the quantitative estimates of cumulative risk.
      This determination flows from the information acquired during Steps 1-4 and
compiled in Step 4. For the purpose of an illustration for the present project, we will
assume that all the component chemicals in the respective Mixtures 1 and 2 are the
respective CAGs.  In the event that (1) exposures are sufficiently small or (2) exposure
or toxicity data are too limited to support including a chemical in the CAG, it will be
omitted at this stage.  Once again, depending on the importance of the mixture of
interest to public health, this (the omission) may not be acceptable. In that case, there
are two alternatives: (1) There are insufficient data  to conduct cumulative risk
assessment and the component chemicals in the mixture should revert back to the
traditional single chemical risk assessment; or (2) New research initiatives are to be
launched to fill the data gaps for the specific purpose of conducting dosimetry-based
cumulative risk assessment for the mixture of interest.

2.3.   PHASE II: DOSIMETRY-BASED CUMULATIVE RISK ASSESSMENT
      For those chemical mixtures selected to undergo DBCRA, the following
additional five steps should be carried out. It is proposed that starting from Step 6 (i.e.,
the first step in Phase II), PBPK modeling should be incorporated.  Common belief is
that PBPK modeling is resource-intensive and it is difficult, particularly for models
involving chemical interactions at the  level of either pharmacokinetics or
pharmacodynamics.  This warrants some special discussion: First, all of the component
chemicals in the mixtures used as examples in this document are industrially or
environmentally important chemicals with uses and/or exposures of concern. When
many chemicals reach this stage of commercialization, a substantial number of studies
may have been  conducted on them. Those chemicals have already become resource-
intensive during the developmental stage to become successful chemicals in
commerce.  However, this is not the case for all commercially important chemicals.  The
important point is that quantitative, time-course data useful for PBPK model
development may have already been generated during the product developmental
phase. If the incentive (i.e., risk assessment-driven scientific studies) exists, such
quantitative,  time-course data would be generated  during the product developmental
phase automatically.  In fact, PBPK modeling,  being a hypothesis-testing tool in
toxicology, may  be utilized to conduct many different kinds of experiments on computers
(i.e., in silico toxicology). Development of in silico toxicology such as PBPK modeling
and other biologically based computer modeling will improve the "attrition rate" of drug
                                      37

-------
 or chemical product development. Thus, it may save precious resources by avoiding
 unnecessary experimental studies or minimizing animal experimentation. Second, while
 PBPK modeling is by no means a very easy technology, it is not any more difficult than
 some of the statistical modeling (e.g.,  linearized multistage carcinogenesis model)
 carried out in the routine risk assessment process.  Furthermore, excellent training
 opportunities are available and the development of software is such that more and more
 user-friendly tools are going to be available.
       The advantages of incorporating PBPK modeling in CRA are many: First, PBPK
 modeling  may bring the tissue dose, instead of the applied dose, into dose-response
 assessment. Thus, it  is much more accurate in that pharmacokinetics has been
 employed to refine "exposure" in terms of tissue doses, reducing uncertainty.  Second,
 PBPK modeling has the capability of incorporating toxicological interactions of multiple
 chemicals, representing a valuable feature for CRA. Third, PBPK modeling has the
 capability of extrapolation,  be it dose-, species-, route-, age-, gender-dependent
 extrapolations.  In many cases, such extrapolation may reach the region (e.g., very low
 doses) where experimental studies are impossible to conduct.  Finally,  PBPK modeling,
 in conjunction with Monte Carlo simulation, may estimate the true means (i.e., of tissue
 AUC values) by carrying out  numerous repeated exposure simulations via computer-
 based programs.  However, confidence in mean values will be tempered by the degree
 to which the distributions of the variable  input parameters (i.e., chemical specific values
 such as partition coefficients  and  species specific values such as organ blood flows) are
 characterized.

2.3.1.  Step 6.  Characterize Dose-Response, Point of Departure  and Determine
Relative Potency from Tissue Doses. The risk assessor should select and apply an
appropriate method to characterize the dose-response relationship for effects and
determine  the relative toxic potencies of the CAG chemicals by each exposure route and
duration of interest. Subsequently, the point(s) of departure for extrapolating the risk of
the CAG should be chosen. As indicated in the guidance document (U.S. EPA, 2002a),
the preferred point of departure is  one derived from modeling the dose-response curve of
the index chemical to derive a BMD that estimates a pre-specified level  of response
(e.g., BMDio, BMD5 or  BMDi). The utility of PBPK modeling in the dose-response
analyses is to provide tissue dose  metrics such that in the BMD estimation, the tissue
doses rather than the applied  doses will be used in the Benchmark Dose Software
(BMDS). Obviously, this requires  identification of the target tissue. If a  PBPK model
capable of addressing chemical interactions (be they additive or not) cannot be
                                       38

-------
constructed because of data gaps, PBPK models for individual chemicals (e.g., the OP
pesticides) should be constructed to determine tissue dose metrics (e.g., concentration
of active metabolite in brain tissue) based on PBPK modeling of given applied doses.
That is, given the exposure level (applied doses), what would be level of the dose metric
in the affected tissue (e.g., brain) relative to the level of inhibition of brain AChE (e.g., a
10% reduction in AChE activity) and the related clinical signs? Pharmacokinetic models
can simulate concentrations of toxicants  in tissues for which biological monitoring cannot
be conducted in humans and for which concentration data may not be available in test
animals.  Certainty in PBPK-based risk assessments is increased when there are tissue
toxicant concentration data against which to compare PBPK model predictions. The
availability of tissue toxicant concentration data for  comparison should be considered
when choosing a pharmacokinetic outcome (tissue  concentration, dose  metric) for risk
assessment application. Of course,  preferably, an  interactive PBPK model is available at
this point such that pharmacokinetic and  pharmacodynamic interactions may be taken
into consideration  and "interaction thresholds" (EI-Masri et al., 1996b; Dobrev et al.,
2001, 2002) may be estimated to obtain interactive NOAELs, LOAELs (which may be
designated as NOAELint and LOAELint) and higher dose-response levels.
       When dose additivity is selected as the cumulative risk assessment model and
 once the models have been constructed, the index chemical should be selected and the
 potencies of other chemicals in the  CMC should be determined based  on dose
 response relationships characterized at  the level of the dose metric most associated
 with the response. This might, for example, be the level of the dose metric
 corresponding to  the animal ED10. Specifically, the dose-response relationships should
 be developed on the basis of internal, rather than external dose.  In this manner,
 pharmacokinetic interactions that alter the dose metrics of interest can  be made
 valuable in estimating the response from the mixture.

       2.3.1.1.  Recommendations on  Modeling  — The U.S. EPA guidance document
 included an informative discussion on "Modeling the Data" (Section 6.2.1.6, U.S. EPA,
 2002a).  Although it is not necessarily for PBPK modeling, the discussion is relevant to
 the present effort.  Some passages are quoted below and the CRA risk assessors using
 PBPK modeling should abide by these recommendations as much as possible:

       ...The selection of a mathematical model structure to fit the data being
       analyzed should be guided by the biology of the common mechanism of
       toxicity, the toxicokinetics of the chemicals,  and the observed shapes of
       their dose-response curves and the experimental designs used to

                                       39

-------
      generate the data.  If available, pharmacodynamic and pharmacokinetic
      data should be considered in order to account for tissue concentrations
      and to aid in defining dose-response relationships across different
      species, routes and time-frames of exposure...
      ...Although it is not possible to recommend the use of specific models, a
      few points that should be considered in modeling the data follow:

         •  Modeling of  individual animal data is desirable; however, if this is
            not practical, then use of summary data such as means and
            standard deviations can be alternatives

         •  Care should be taken with modeling high-dose data (particularly
            extreme doses) because the model shape in the low-dose region
            can be influenced by high-dose data

         •  Log transformation of data should be justified because such a
            transformation may distort the dose-response curve

         •  Data variability should be described by appropriate statistical
            techniques and reflected in the potency estimate (e.g., by weighting
            the data in the fitting procedure)

         •  Confidence intervals or limits should be included in the analysis
            because they can be valuable for evaluating the influence of
            variability on the potency estimates

         •  An estimate  for the uncertainty of the model used in the analysis
            should be included

         •  The statistical fitting method used must be clearly described.


      2.3.1.2.  Interactive PBPK Model: General Information — As this is the first
step of the dosimetry-based cumulative risk assessment process and the step where
PBPK modeling initiates, it is appropriate to discuss the issues related to the
incorporation of PBPK modeling.  In addition to the discussion below,  a recent paper on
PBPK/PD modeling is attached as Appendix A.
      From the perspective of interactive PBPK modeling, two aspects need to be
addressed: pharmacokinetic interactions and pharmacodynamic interactions. As
indicated earlier under Step 4, pharmacokinetic and pharmacodynamic interactions are
likely in both Mixtures 1  and 2. Thus, it is necessary to construct interactive PBPK
models after a thorough review of the existing literature. The most ideal and
scientifically defensible data requirement for establishing an interactive PBPK model is
that each component chemical in the mixture already has its respectively established
PBPK model and that there are many pharmacokinetic datasets in laboratory animals
                                      40

-------
as well as in humans available for each of these component chemicals.  In some cases,
even more specific and stringent data requirements are needed; an example may be
the ultralow dose pharmacokinetic data in perinatal developmental stages of laboratory
animals for extrapolation to human fetuses, neonates and infants using PBPK modeling.
However, until such time that quantitative time-course data useful for PBPK modeling
are automatically part of the product developmental process, the most likely scenario is
imperfect datasets. In Mixture 1, no PBPK models for fenthion, fenitrothion, diazinon,
and methyl parathion could be located at this time. However, PBPK models do exist for
parathion (Sultatos, 1990; Gearhart et al., 1994; Abbas and Hayton, 1997; Gentry et al.,
2002) and chlorpyrifos (Timchalk et al., 2002a,b; Kousba et al., 2003).  In Mixture 2,
because all the chemicals are important solvents with huge production volumes, PBPK
models are available for TCE, PERC, MC and CHF. TCE has obtained perhaps the
most attention of these chlorinated hydrocarbons/volatile organics in terms of the
development of PBPK models, many of which have also been used to support risk
assessment (Sato etal.,  1977, 1991; Bogen,  1988; Koizumi, 1989; Fisher, 1993; Fisher
and Allen, 1993; Fisher etal., 1989,  1990, 1991; Allen and Fisher, 1993; Clewell et al.,
1995; Croninetal., 1995; Poet etal., 2000; Dobrevetal.,  2001, 2002).  Several PBPK
models for the disposition of tetrachloroethylene were presented for animals and/or
humans (Ward et al., 1988; Koizumi, 1989; Bois et al., 1990; Gearhart et al.,  1993; Rao
and Brown, 1993;  Byczkowski et al., 1994; Dallas et al., 1994, 1995; Wilson and Knack,
1994; Byczkowski and Fisher, 1995; Reitz et  al., 1996; Loizou, 2001; Poet et al., 2002).4
PBPK models for MC alone or for mixtures of MC and TCE or other chlorinated solvents
have been published (Reitz et al., 1988; Koizumi, 1989; Tardif and Charest-Tardif,
1999; Dobrev et al., 2001).  Several  PBPK models that included progressively more
sophisticated levels of biochemical complexity (e.g., relating to  metabolite formation,
cellular regeneration, enzyme inhibition) have been developed for chloroform (Corley et
al., 1990, 2000; Gearhart et al., 1993; Chinery and Gleason, 1993; McKone,  1993; Roy
et al., 1996;  Levesque et al., 2000).

      2.3.1.3. PBPK/PD Models: Data Needs — What are the specific data needed
for building PBPK models? And what happens when such data are missing because no
4 PBPK models developed for a single chemical may differ for many reasons, including the species
addressed, route of exposure simulated, effect evaluated, choice for dose metric developed, differential
reliance on in vivo datasets and/or in vitro datasets, choices for the values of physiological and chemical-
specific parameter values employed, etc.

                                      41

-------
PBPK modeling has been attempted on some of the components in the mixtures of
interest? First, the data necessary for establishing a PBPK model for a single chemical
will be presented.  Obviously, well conducted in vivo pharmacokinetic experiments are
essential and usually the more varied the datasets (e.g., different doses, routes,
species), the better.  Other PBPK modeling-specific information such as the chemical-
specific parameters (e.g., tissue partition coefficients, Vmaxs and Kms) is needed.
Enzyme kinetic data, particularly human data, of the bioactivation (e.g., P=S to P=0
conversion) and detoxication (e.g.,  oxidative cleavage, hydrolysis, oxidative
dealkylation, etc.)  processes will be important for the interactive PBPK model.  In vitro
determination of tissue partition coefficients and enzyme kinetic data are relatively
straightforward. A later section is devoted totally to high quality organ donor liver
enzyme studies. Alternatively, with modern genetic engineering technologies, many
human enzymes are available commercially,  though their expression, singly, in non-
mammalian species may complicate the extrapolation of results obtained from such
preparations.  For instance, several recombinant human CYP enzymes are now
commercially available, including CYP 1A1, CYP 1A2, CYP 2B6,  CYP 2C8, CYP 2C9,
CYP 2C18, CYP 2C19, CYP 2D6, CYP 3A4,  CYP 3A5, and CYP 3A7 (e.g., from
GenTest [Woburn, MA], Cypex Ltd. [Dundee, Scotland], and  Sigma-Aldrich [St. Louis,
MO]). Thus, heretofore unavailable human enzyme kinetic information for many of the
environmentally important chemicals is within easy reach for  many laboratories. These
experiments should be performed.
      Pharmacodynamically, for instance,  once the oxon (i.e., P=0) derivatives are
formed, a number of questions arise: What is the competition between the irreversible
binding  of the oxon derivative with red cell AChE, non-specific esterases, and brain
AChE for any given single pesticide?  The same question, although much more
complex, should be asked about the mixture of 6 oxon derivatives.  Here the ultra-low
dose pharmacokinetic/pharmacodynamic studies in the binding of brain AChE with  any
given OP pesticide in the presence of others  similar to those  reported by Vogel et al.
(2002) will be critical for the interactive PBPD modeling in the cumulative risk
assessment.  Both chemical specific and species specific data are required.  Chemical
specific data include metabolic rate constants and measures of solubility (e.g.,  partition
coefficients).  Species specific data include information on body weight, organ or
compartment sizes, region/organ blood flow rates and media specific intake rates (i.e.,
alveolar ventilation rates).  Body weight and relative contributions of specific organs
(i.e.,  liver) to fractional  body mass are fairly constant; they are usually modeled as point
values.  Regional or organ  specific blood flows may vary, obviously with exercise;
                                      42

-------
recent evidence suggests that hepatic blood flow may vary three-fold at rest among
adults.  Virtually all species specific data may vary, to some extent, among individuals
and with age, as may metabolic rate constants.  Chemical-specific partition coefficients
may vary approximately two-fold among individuals, and metabolic parameter values
may vary markedly, depending on the chemical and enzyme involved.

      2.3.1.4.  PBPK/PD Modeling: Model Structures — While the structure of the
PBPK/PD model will have to be determined in detail based on the analysis conducted in
Steps 1-5, a general suggestion on what the model might contain can be provided at
this time. Figure 4 presents a graphical representation of the PBPK model that could be
initially considered.  The model consists of several familiar compartments. The
CNS/PNS (central nervous  system/peripheral nervous system) compartment should be
added and indeed, can be split into two compartments if kinetic processes are
significantly different in the two and if endpoints or validation datasets suggest doing so.
The liver compartment is where metabolism is usually regarded as occurring, but
extrahepatic metabolism  should also be considered.  For the drinking water scenario (or
other ingestion scenarios) chemical exposure of the liver should be primarily via the
portal flow from the  Gl compartment in addition to systemic flow.  Portal blood flow may
play a less prominent role (compared to arterial flow) for substances encountered via
inhalation. Usually, the chemical is dosed directly to the Gl compartment and first order
uptake is often assumed, but more complex formulations can be used if modeling
parsimony is not overly compromised.
      In the diagram, the liver compartment includes a schematic of simplified
metabolic pathways.  For metabolism, parsimony is important: only include discrete
pathways if the data will support developing kinetic parameter estimates. The overall
metabolism can be simplified in  a number of ways: sequential metabolic steps can be
modeled as one step that represents the rate-limiting step  in  the sequence. Parallel
pathways can be lumped together.  Insignificant pathways can be ignored.  The extent
to which a given pathway may be deemed insignificant should be judged on the basis of
all information known for each component of the chemical  mixture; a pathway deemed
minor or insignificant for one chemical may be critical for another.  Because the same
model structure should be employed for each chemical (within reason) and for the
mixture, the selection of pathways for inclusion should not be taken lightly.  Some
pathways can be multiple steps  (e.g., oxidative desulfuration and hydrolysis) while
others can be single staged (e.g., clearance pathways to non-toxic non-interacting
metabolites). The importance of parsimony is emphasized by the limited amount of
                                      43

-------
[
Venous blood Binding to esterases






•"
Lung

CNS/PNS

Fat

Slowly perfused

Rapidly perfused

XO
YOOp«S
f
voH°ks_ yo'3 ,s __Yox3p,o J^o
i i I °»
HOX3,S
Liver ?
t
Gl Tract

	
Skin









o
•*=
w
Arterial blood Binding to esterases f

* Dermal
•* 	
i

               FIGURE 4
A Preliminary PBPK Model Structure for OPs
                  44

-------
data (particularly in vivo) and the difficulty of extrapolating in vitro kinetic constants to in
vivo or determining the constants from model optimization. Automated model
optimization of parameters should be minimized because, for the single chemical
models, each pathway can require 2 metabolic constants, there can be 2, 3, 4...
different metabolic steps, some of these steps could actually inhibit each other if the
same enzymes are used at high enough doses, and inhibition would require additional
parameters. So for metabolism alone, it would be easy to get over 10 unknown
parameters, which would be difficult to optimize with any certainty. This is before
consideration of inter-component PK interactions, which require yet more optimization.
However, more than one metabolic step will nevertheless be required.
      In the schematic (Figure 4), lung and skin compartments are dotted boxes to
represent the fact that they may or may not be required for the drinking water
assessment.  If dermal exposure to drinking water sources does not contribute
significantly (a judgment call) to the aggregate exposures, the dermal compartment may
be unnecessary for the  initial model, but may be desirable for future modeling work
(especially for occupational exposure PBPK models).  Depending on the chemical and
its volatile properties, only occupational exposure models may require the inhalation
route of exposure. As with the discussion of age and gender, the model is usually built
for the route where the most complete datasets are available and then extended to
other routes.  However, for this study, the ingestion route should be given decided
preference  unless data  are severely lacking.  If inhalation models are developed,
attention must be paid to the form of the chemical (i.e., aerosols) and non-equilibrium
processes in uptake kinetics.
      The CNS/PNS compartment will need to include a submodel for binding to AChE.
A simple conceptualization of this is proposed in Figure 5.  As OP binding is considered
largely irreversible and causes complete inhibition of the enzyme (individual enzymes
that are bound), the degree of  inhibition would reflect the number of bound enzymes.
However, the activity of AChE  is not necessarily proportional to unbound (free) enzyme;
there may be a "reserve capacity" in that not all enzyme is required to clear
acetylcholine  in synapses. Thus, the  response may be a non-linear function of the
molecular events.  PD models  will be  a useful way to describe these processes once
developed.
      The PD submodel should include a description of the process of synthesis of
AChE, an on-going process. The kinetics of this process may be zero order (as is often
assumed for glutathione synthesis (D'Souza and Andersen, 1988), or could be
inducible.  Binding of free AChE by activated OPs may be a saturable process
                                      45

-------
AChE resynthesis
 Ko?, Inducible?
CNS/PNS
Native AChE Pool
Binding by OPs
                        Competitive inhibition?
                            B max* [S]
                            -Kd*(\-
                                                   Ki
                       FIGURES
      A Pharmacodynamic Submodel for CNS/PNS Compartment
                          46

-------
describable by a Bmax or Hill's equation [Bmax + S/(S+Kd)] with or without inhibition by
competing OPs. However, since it is usually regarded as irreversible, a linear equation
may do equally well and will involve fewer parameters. An empirical relationship
between the concentration of free AChE and neurological effects (e.g., tremors) will
complete the PD portion of the model.  Similar submodels will need to be included for
binding to other esterases, primarily  in the blood.
      In order to attain reasonable level of confidence, the interactive PBPK models for
mixtures should be developed only after the component single chemical models are
developed and validated. Likewise, the PD submodels should not be incorporated until
the PBPK models are as validated as possible.
      The models can be exercised with existing animal data to determine practical
thresholds for toxicokinetic interactions. In the event that human exposure is at lower
ranges than the model is validated for,  particular attention should be paid to the
presence or absence of interactions  at the lowest validated level. It would be useful to
perform low-dose PK studies in animals (or humans) to extend the range of the model
downwards, especially if it can be extended down to the range of the exposures under
consideration.  Moreover, it would be even more important to do the low dose
experiments if (non-additive) interactions are still occurring at levels below the model
validation range, to determine the shape of the dose response curve below the point of
departure.

      2.3.1.5. Human PBPK Modeling:  Incorporation of In Vitro Enzyme
Studies — The ultimate goal of PBPK  modeling is to provide scientifically defensible
computer simulations of the fate of the  chemical or chemical mixtures in humans.  There
are ethical and other problems with in vivo human toxicology studies, particularly with
highly toxic pesticidal chemicals such as OPs.  Thus, the building of human PBPK
models must rely on allometric extrapolation of animal  data and/or in vitro studies using
human tissues. While many of the physiological parameters on humans are readily
available in the physiology literature, some parameters which are chemical specific such
as tissue/blood partition coefficients and metabolic parameters (Km and Vmax) are best
empirically derived. Recent advances in PBPK modeling have demonstrated the ability
of the technique to include extrapolations made from data on chemical metabolism and
enzyme contents derived from in vitro human and research animal tissue preparations
(Kedderis and Held, 1996; Kedderis, 1997; Lipscomb et al., 1997, 1998, 2003a,b;
Lipscomb and Garrett, 1998; Snawder and Lipscomb, 2000; Kedderis and Lipscomb,
2001;  Lipscomb and Kedderis, 2002).  In particular, careful experimental measurements
                                      47

-------
and statistical evaluations on the content of microsomal protein, cytochrome P450
enzymes, in human liver samples derived under organ transplant conditions had been
carried out (Lipscomb et al., 1997, 1998, 2003a,b; Lipscomb and Garrett, 1998;
Snawder and Lipscomb, 2000). This is important because some past human studies
employed liver samples from cadavers, which were metabolically compromised.
Lipscomb et al. (2003b) recently integrated their various human in vitro studies from
organ donors and demonstrated the application of such information via PBPK modeling
in a risk assessment framework. This paper (Lipscomb et al., 2003b) provided a
stepwise illustration of how to incorporate three different datasets (the microsomal
protein content of human liver, the CYP2E1 [the principal metabolic enzyme for TCE]
content of human liver microsomal protein, and the in vitro Vmax for TCE oxidation by
humans) into a PBPK model for risk-relevant pharmacokinetic outcome in humans.
Using a variety of statistical  analyses, the 5th  and 95th percentiles of the resulting
distribution on Vmax (TCE oxidized per minute per gram liver) differed by approximately
6-fold.  These values were converted to the in vivo Vmax (mg TCE oxidized/hr/kg body
weight) and incorporated into a human PBPK model for TCE. Model simulation under
the conditions of 8-hr inhalation exposure to 50 ppm or drinking water exposure at 5 |j,g
TCE in 2 L/day revealed that 6-fold variation in Vmax (i.e., Vmax at 5th or 95th percentiles
levels) resulted in only 2% or less differences in metabolism of a key intoxicating step,
the formation of chloral  hydrate (Lipscomb et al., 2003b). On the surface, this finding,
which is suggestive of Vmax insensitivity under the model simulation conditions, has
strong  implications in risk assessment. In essence, it suggests that, at low
environmental or occupational exposure conditions, individual variability of metabolic
capacity as large as 600% (6-fold) has little or no impact on the toxic outcome of TCE.
However, upon closer examination, valid scientific explanation is available.  Kedderis
(1997), in studying the effect of enzyme induction on the bioactivation of TCE and other
volatile organic compounds, indicated that the hepatic blood flow limitation plays an
important role in the kinetics of bioactivation.  What happens in the above situation
(Lipscomb et al., 2003b) is that, even though the individual metabolic capacity varies
greatly (i.e., 600%), the rate of hepatic blood  flow delivery of TCE and its related
metabolites (formed from earlier passages through liver cells) to the liver is much slower
than the rate of bioactivation in the liver. Thus, this "flow-limited process" (that delivers
concentrations below those  that are in the linear range of the metabolic rate versus
substrate concentration curve) is a more important factor than maximum metabolic rate,
thus limiting the impact  of the large variability in metabolic capacity in the population.
These  results underscore the importance of considering the overall dynamic equilibrium
                                      48

-------
of all the relevant biological processes in the body—in some sense, a broader
application of the systems biology approach. The Lipscomb et al. (2003b) paper is
attached as Appendix B as instructional material for the incorporation of human in vitro
data into the PBPK modeling process.
      Appendix C provides summary information for each of the component chemicals
in the two  mixtures identified for this project.  It includes information on the toxicity,
metabolism and pharmacokinetics for each chemical.

2.3.2. Step 7. Develop Detailed Exposure Scenarios for All Routes and Durations.
The contribution to isolated or individual chemical exposures to the potential magnitude
of exposure should be characterized. Once these are characterized, a decision should
be reached as to whether these exposure scenarios should be included in a qualitative
exposure assessment. Consideration should be given to the identification of
subpopulations and their locations, as well as to co-exposures to multiple chemicals.
As drinking water is the primary concern for this project, emphasis will be placed on the
oral consumption of finished drinking water.  Review of pertinent literature on the
component chemicals in  the mixture is the first step (see Appendix C). In the guidance
document (U.S. EPA, 2002a), it was stressed that cumulative risk assessments should
reflect use patterns and practices on  a scale sufficient to capture the variability in
pesticide use, but not so large as to inappropriately dilute real and significant
differences.  A specific example was  given on fenthion, one of the six OP pesticides
selected by the  U.S. EPA for the first mixture. Apparently, fenthion was used for
localized mosquito control in parts of southern Florida; therefore, it was stated that this
pesticide should have only limited consideration in an assessment of other OP
pesticides, including those used for mosquito control (U.S. EPA, 2002a). Exposure
assessments should take into account those factors that impact the targeted group of
humans or the geographic region of interest.

2.3.3. Step 8. Quantify Parameters for Exposure.  Exposure should be presented in
as quantitative a manner as is possible. The magnitude, frequency and duration for all
pertinent exposure pathway/route combinations should be determined. Appropriate
sources of use/usage information, chemical concentrations in all appropriate media, and
any modifying factors necessary should be included in the assessment.  Where
necessary, surrogate datasets developed for similar chemicals, published literature or
generic datasets should be identified and justified for inclusion.  Here, emphasis will be
placed primarily on drinking water.  The guidance document (U.S. EPA, 2002a)
                                       49

-------
indicated that evidence of the co-occurrence of pesticides within a drinking water source
for a CAG is a critical piece of information needed prior to making a decision to include
more than one pesticide in a cumulative drinking water exposure assessment.  Direct
measurements of combinations of pesticides in finished drinking water are rarely
available.  However, U.S. Geological Survey's National Water Quality Assessment
Program databases do contain information on the co-occurrence of a wide variety of
pesticides in ambient surface water, and some registrant-sponsored studies provide
co-occurrence data for some compounds in drinking water. The exposure scenarios
may take on one of several forms, as previously described.  In accordance with U.S.
EPA guidance (2002a), the exposure data will be formatted for input into the
quantitative DBCRA taking  into account,  as appropriate, seasonal variations,
geographical variations, and other variations in exposure. The range of exposures
determined as such must be part of the exposure input. Yet more sophisticated
approaches, such as the one suggested  by Figure 6, can also be considered.  In
Figure 6, the distributions of exposures are determined and used as input to a PBPK
model. As interactions between component chemicals in the CAG will depend on the
magnitude and duration of the exposure to each chemical, non-PBPK models would
necessarily have to attempt to simplify the basis for the interactions in some empirical
manner.  Alternately, PBPK models (should validated ones be developed) can be used
to estimate the actual extent of toxicity (AChE inhibition or as otherwise desired). The
model can be run iteratively for as many  exposure scenarios as necessary. Indeed, it is
quite feasible to use a Monte Carlo sampling approach to sample from the distributions.
For example, within each of 500 communities, the simulation of 1000 exposures (i.e.,
what 1000 persons would be exposed to) would not be an unreasonable task.  This
would allow the  determination of cumulative risks in various locations, times of year,
etc., as well  as for the nation as a whole.

2.3.4. Step  9. Conduct Dosimetry-Based Cumulative Risk Assessment. This is
the point in the process where the pharmacokinetic data and  model are combined with
the defined exposure scenario(s) to estimate internal dosimetry.  The resulting internal
doses are combined with data describing dose-response and potency defined  in step 6.
To accomplish this, route/duration-specific dose metrics associated with specific risks
should be identified and internal doses should be expressed in these terms. A trial  run
should be conducted initially and its results evaluated. The model should be subjected
to a sensitivity analysis, which will identify model assumptions and parameters that  most
influence the production of the risk related dose metric. An assessment of the
                                      50

-------
1. Population-weighted,
geographically-defined
exposure data for each
mixture component
Monte Carlo sampling of
exposure data from estab-
lished distributions
  2. Interaction PBPK
     Model-Based
Determination of Tissue
  Doses and Toxicity
                                                         CMS AChE
                                                         inhibition resulting
                                                         from all mixture
                                                         components
                                                        _Q
                                                         ro
                                                         o
                                                         ol
                                    Statistical analysis of
                                    model outputs for all
                                    exposure scenarios

                                        Region XYZ
                                                                      Hazard Index

                                                             3. Geographically defined
                                                             distribution of risk estimates
                                       FIGURES
A Schematic for Integrating Data for Cumulative Risk Assessment.  This is a possible
method for integrating exposure distributions with PBPK models for several interacting
chemicals and analyzing model output to determine distribution of cumulative
risk estimates.
                                          51

-------
subpopulations of concern should be undertaken, and uncertainty and FQPA safety
factors should  be recommended.  The documents Cumulative Risk: A Case Study of the
Estimation of Risk from 24 Organophosphate Pesticides (U.S. EPA, 2000b) and
Organophosphate Pesticides: Revised OP Cumulative Risk Assessment (U.S.  EPA,
2002b), as well as the documents addressing the cumulative risk of N-methyl
carbamates, triazines and the chloroacetanilides should be consulted for carrying out
cumulative risk assessment. The results of the DBCRA should be presented so that the
reader will understand which dose metrics for which chemicals and effects were
selected and simulated, and the impact that a cumulative exposure will have on the
production of these dose metrics and, ultimately, on risk.
      Assigning FQPA Safety Factors is restricted to children and restricted to
pesticide chemicals only (refer to OPP guidance and to Step 10, below).

2.3.5.  Step 10. Characterize Cumulative Risk via Pharmacokinetic Analysis.
Describe the results and  conclusions of the cumulative risk analysis, including the
relative confidence in toxicity and  exposure data sources  and model inputs.  Discuss
major areas of uncertainty, the magnitude and direction of likely bias, and the impact on
the final assessment. Evaluate the risk contributions from each pathway and route
individually, as well as in combination. Identify risk contributors with regard to
chemical(s),  pathway, source, time of year, and impacted subpopulation (with particular
attention to children).  Conduct sensitivity analyses to determine those factors most
likely to impact the risk. Determine need for additional uncertainty and safety factors.
      A summary of the risk characterization should include a  restatement of the scope
of the issue being  addressed, explaining the chemicals under evaluation, data  available
and their strengths and weaknesses including uncertainties, the assumptions made
during the analysis, and how the results are interpreted for each of the demographic
groups represented. These  data include those pertaining to the temporal and
geographic nature of exposure, including food and water ingestion rates, contamination
levels (including non-detects) and the methods used to develop distributions for variable
data.  Special emphasis should be placed on how groups differ, and the likely bases for
those differences,  including anatomic, biochemical and physiologic differences, as well
as differences  related to exposures. With respect to PBPK modeling, these differences
may be due to age and sex-dependent differences in body composition and/or
metabolic capacity. The risk characterization section should report biases in datasets
employed, and a general evaluation of the level of confidence placed in the analysis.
Significant sources of uncertainty  should be communicated, and when possible, the
                                      52

-------
results of a sensitivity analysis should be used to demonstrate the level of impact that
these uncertainties may have on the overall outcome. Outcome should be
communicated as the level of the dose metric(s) most closely associated with the health
endpoints of concern.  Regarding sensitivity within the human population, the
application of FQPA safety factors is restricted to children and to pesticide chemicals.
The safety factor and the bases for its derivation (a value of up to a factor of ten) have
been described in other documents developed by OPP.  When data and  circumstances
warrant, the additional FQPA safety factor may not be applied. When applied, however,
the factor is applied to the results for the chemical  mixture, rather than to the individual
chemicals.
                                      53

-------
                              3. CONCLUSIONS

      In this report, we outline some of the considerations that are involved in
performing CRA using a PBPK-modeling based approach.  While this report makes a
number of specific suggestions and observations, additional details of a PBPK-based
approach to CRA will  depend on the specific chemicals in a CAG, which are not known
until the initial analysis is complete; and issues that arise during the risk assessment
process itself, which are largely unknown until the work is undertaken.  For
methodological issues that are decided on the basis of scientific judgment, decisions
should be made with the assistance of appropriate advisory committees dedicated to
the subject at hand.
      PBPK modeling-based approaches offer several advantages in CRA. First, U.S.
EPA has recently advocated using biologically-based approaches in risk assessment.
Indeed, these biologically based approaches offer the best means of performing many
of the extrapolations that are necessary in the risk assessment process.  Biologically
based approaches were further recommended by U.S. EPA for use in CRA,  when they
are available.  Because of their ability to extrapolate dosimetry across dose, species,
sex, route and age, PBPK modeling is often the favored biologically-based method for
determining tissue dosimetry. Second, PBPK modeling is the simplest method for
characterizing PK interactions in the body that is directly based on the biology of the
process. Other methods are empirical and not only take a significant effort to develop,
but have uncertainties when extrapolations are made to scenarios that are untested
(e.g., empirical methods developed for A+B and B+C do not translate easily into a
method for A+C). Third, if significant PK or PD interactions occur, DBCRA cannot be
reasonably performed without addressing the interactions at the level of target tissue
concentration. In other words, why go from single chemical risk assessments to
multiple chemical risk assessments when one of the ramifications of multiple chemical
exposures (toxicokinetic interactions resulting in altered tissue dosimetry)  is  omitted?
      Implementing PBPK modeling-based approaches to CRA would require the
development of PBPK models for the individual  chemicals in the CAG as well as a
quantitative definition  of the nature of the interactions. While PBPK models exist for
some of the chemicals considered here, some data are available to support  the
development of PBPK models for the rest.  Most likely, the biggest data gap consists of
data regarding interactions.  For the chlorinated solvents, sufficient data may be
available at this time;  indeed some  interaction models are already in the literature
(Dobrev et al., 2001, 2002; Haddad et al., 2000; Thrall and Poet, 2000). For the OP
                                      54

-------
pesticides, data gaps for some combinations of chemicals are likely.  Nevertheless, the
burden of developing PBPK models for this group is greatly reduced by the fact that the
same model structure can likely be used for all six chemicals.  With sufficient data,
building the interactive PBPK models is likely to be feasible.
      For drinking water scenarios, multiple route (drinking, inhalation and dermal
through bathing) exposure is a possibility. Even if the CRA is restricted to direct
consumption of drinking water itself, the pharmacokinetic and pharmacodynamic
processes involved may be non-linear, including differing rates of metabolism for
different chemicals, inhibition and replenishment of AChE, etc. These types of non-
linear processes are well characterized by PBPK models. Moreover, the drinking water
scenario involves a complex set of exposures.  For each chemical, a different
distribution of exposures, varying by season of the year, spatial location, and other
factors must be incorporated into the assessment.  When this level of complexity is
added to the variety of issues mentioned above, simple algebraic approaches to
calculating risks become strained. Other modeling approaches may be better suited to
address the exposure assessment part of the CRA, and these approaches are easily
tied into the PBPK model.
      The approach recommended in this report is completely consistent with existing
U.S. EPA guidance. The PBPK model itself largely affects the dose-response analysis
of the CRA.  In essence, a PBPK modeling approach is one alternative  method
suggested by U.S. EPA in performing dose-response analysis.  However, if a PBPK
model is to be used, other aspects of the CRA are also affected.  The approach used for
exposure assessment, for example, must be tailored in such a way that the structure of
the dataset is appropriate as an input to the PBPK model.  As a second example, the
criteria used to determine whether exposure to a chemical  is sufficient to warrant its
inclusion in the CAG is also modified by the fact that chemical interactions are
addressed within the CRA. Specifically, this report points out that a chemical could
conceivably be retained in the CAG based on an ability to interact with another
chemical, regardless of whether the retained chemical can directly cause the toxicity
that is the critical endpoint for the analysis.  In this way, a PBPK modeling-based CRA
may more accurately determine the actual cumulative risk from a  set of chemicals than
an approach that does not quantify the impact of interactions on toxicity. Last but not
least, the ultimate goal for risk assessment is the protection of public health.  Therefore,
PBPK modeling of chemical and chemical mixtures in humans is absolutely essential.
Accordingly, the incorporation of quality human  tissue studies into the PBPK  modeling
process is of critical importance.  As a final concluding remark, as U.S. EPA advances
                                      55

-------
the science of cumulative risk assessment, PBPK modeling should be incorporated
where appropriate.
                                     56

-------
                              4. REFERENCES
Abbas, R. and W.L. Hayton.  1997.  A physiologically based pharmacokinetic and
pharmacodynamic model for paraoxon in rainbow trout. Toxicol. Appl. Pharmacol.
145(1): 192-201.

Allen, B.C. and J.W. Fisher.  1993.  Pharmacokinetic modeling of trichloroethylene and
trichloroacetic acid in humans.  Risk Anal. 13(1):71-86.

ATSDR (Agency for Toxic Substances and Disease Registry). 1995. Toxicological
Profile for 1,1,1 -Trichloroethane (CAS# 71-55-6).  U.S. Department of Health and
Human Services, Public Health Service, Atlanta, GA.

ATSDR (Agency for Toxic Substances and Disease Registry). 1997a. Toxicological
Profile for Trichloroethylene (TCE) (CAS# 79-01-6). U.S.  Department of Health and
Human Services, Public Health Service, Atlanta, GA.

ATSDR (Agency for Toxic Substances and Disease Registry). 1997b. Toxicological
Profile for Chloroform (CAS# 67-66-3). U.S. Department of Health and Human
Services, Public Health Service, Atlanta, GA.

ATSDR (Agency for Toxic Substances and Disease Registry). 1997c. Toxicological
Profile for Tetrachloroethylene (PERC) (CAS# 127-18-4).  U.S. Department of Health
and Human Services, Public Health Service, Atlanta, GA.

ATSDR (Agency for Toxic Substances and Disease Registry). 2004. Guidance Manual
for the Assessment of Joint Toxic Action of Chemical Mixtures.  U.S. Department of
Health and Human Services, Public Health Service, Atlanta, GA. Available at


ATSDR (Agency for Toxic Substances and Disease Registry). 2005. ATSDR Public
Health Assessment Guidance Manual. U.S. Department of Health and Human
Services, Atlanta, GA. Available at

Bogen, K.T.  1988.  Pharmacokinetics for regulatory risk analysis: The case of
trichloroethylene. Regul. Toxicol. Pharmacol.  8(4):447-466.

Bois, F.Y., L. Zeise and T.N.  Tozer.  1990. Precision and sensitivity of pharmacokinetic
models for cancer risk assessment - tetrachloroethylene in mice, rats, and humans.
Toxicol. Appl. Pharmacol.  102(2):300-315.

Bruckner, J.V. and D.A. Warren. 2001.  Toxic effects of solvents and vapors.  In:
Casarett and Doull's Toxicology: The Basic Science of Poisons, C.D. Klaassen, Ed.
McGraw-Hill, New York, NY.  p. 892.
                                     57

-------
Byczkowski, J.Z. and J.W. Fisher. 1995. A computer-program linking physiologically-
based pharmacokinetic model with cancer risk assessment for breast-fed infants.
Comput. Meth. Programs Biomed. 46(2): 155-163.

Byczkowski, J.Z., E.R.  Kinkead, H.F. Leahy, G.M. Randall and J.W. Fisher. 1994.
Computer-simulation of the lactational transfer of tetrachloroethylene in rats using a
physiologically-based model. Toxicol. Appl. Pharmacol.  125(2):228-236.

Chambers, J.P., S.L. Hartgraves, M.R. Murphy, M.J. Wayner,  N. Kumar and J.J.
Valdes.  1991.  Effects of three reputed carboxylesterase inhibitors upon rat serum
esterase activity. Neurosci.  Biobehav. Rev. 15:85-88.

Chanda, S.M. and C.N. Pope. 1996.  Neurochemical and neurobehavioral effects of
repeated gestational exposure to chlorpyrifos in maternal and  developing rats.
Pharmacol. Biochem. Behav. 53(4):771-776.

Chinery, R.L. and A.K.  Gleason.  1993.  A compartmental model for the prediction of
breath concentration and absorbed dose of chloroform after exposure while showering.
Risk Anal.  13:51-62.

Clewell, H.J., P.R. Gentry, J.M. Gearhart, B.C. Allen and M.E.  Andersen.  1995.
Considering pharmacokinetic and mechanistic information in cancer risk assessments
for environmental contaminants: Examples with vinyl chloride and trichloroethylene.
Chemosphere. 31(9):2561-2578.

Cohen, S.D. 1984.  Mechanisms of toxicological interactions involving
organophosphate insecticides. Fundam. Appl. Toxicol. 4:315-324.

Corley, R.A., A.L. Mendrala, F.A. Smith et al. 1990. Development of a physiologically
based pharmacokinetic model for chloroform.  Toxicol. Appl. Pharmacol.  103:512-527.

Corley, R.A., S.M. Gordon and L.A. Wallace.  2000. Physiologically based
pharmacokinetic modeling of the temperature-dependent dermal absorption of
chloroform by humans  following bath water exposures.  Toxicol. Sci.  53:13-23.

Cronin, W.J., E.J. Oswald, M.L. Shelley, J.W. Fisher and C.D.  Flemming. 1995.  A
trichloroethylene risk assessment using a Monte Carlo analysis of parameter
uncertainty in conjunction with physiologically-based pharmacokinetic modeling.  Risk
Anal.  15(5):555-565.

D'Souza, R.W. and M.E. Andersen.  1988.  Physiologically based pharmacokinetic
model for vinylidene chloride. Toxicol. Appl. Pharmacol. 95:230-240.

Dallas, C.E., X.M. Chen, S. Muralidhara, P. Varkonyi, R.L. Tackett and J.V. Bruckner.
1994. Use of tissue disposition data from rats and dogs to determine species-
differences in input parameters for a physiological model for perchloroethylene.
Environ. Res. 67:54-67.
                                      58

-------
Dallas, C.E., X.M. Chen, S. Muralidhara, P. Varkonyi, R.L. Tackett and J.V. Bruckner.
1995. Physiologically-based pharmacokinetic model useful in prediction of the influence
of species, dose, and exposure route on perchloroethylene pharmacokinetics.  J.
Toxicol. Environ. Health. 44:301-317.

Dobrev, I., M.E. Andersen and R.S.H. Yang. 2001. Assessing  interaction thresholds for
trichloroethylene, tetrachloroethylene, and 1,1,1-trichloroethane using gas uptake
studies and PBPK modeling. Arch. Toxicol.  75:134-144.

Dobrev, I., M.E. Andersen and R.S.H. Yang. 2002. In silico toxicology: Simulating
interaction thresholds for human exposure to mixtures of trichloroethylene,
tetrachloroethylene, and 1,1,1-trichloroethane.  Environ. Health Perspect.
110(10): 1031-1039.

el-Masri, H.A., R.S. Thomas, G.R. Sabados etal.  1996a. Physiologically based
pharmacokinetic/pharmacodynamic modeling of the toxicologic interaction between
carbon tetrachloride and Kepone.  Arch Toxicol. 70(11):704-713.

el-Masri, H.A., A.A. Constan, H.S. Ramsdell and R.S.H. Yang.  1996b. Physiologically
based pharmacodynamic modeling of an interaction threshold between trichloroethylene
and 1,1-dichloroethylene in Fischer 344 rats. Toxicol. Appl. Pharmacol.  141:124-132.

Fenske, R.A., C. Lu, D. Barr and L. Needham.  2002.  Children's exposure to
chlorpyrifos and parathion in an agricultural community in central Washington State.
Environ. Health Perspect. 110(5):549-553.

Fisher, J.W.  1993.  Estimating the risks of liver and lung cancer in humans exposed to
trichloroethylene using a physiological model. Toxicol. Lett.  68(1-2):127-129.

Fisher, J.W.  and B.C. Allen. 1993. Evaluating the risk of liver cancer in humans
exposed to trichloroethylene using physiological models. Risk Anal.  13(1 ):87-95.

Fisher, J.W., T.A. Whittaker, D.H. Taylor, H.J. Clewell III and M.E. Andersen.  1989.
Physiologically based pharmacokinetic modeling of the pregnant rat:  A multiroute
exposure model for trichloroethylene and its metabolite, trichloroacetic acid. Toxicol.
Appl. Pharmacol.  99:395-414.

Fisher, J.W., T.A. Whittaker, D.H. Taylor, H.J. Clewell III and M.E. Andersen.  1990.
Physiologically based pharmacokinetic modeling of the lactating rat and nursing pup: A
multiroute exposure model for trichloroethylene and its metabolite, trichloroacetic acid.
Toxicol. Appl. Pharmacol. 102:497-513.

Fisher, J.W., M.L. Gargas, B.C. Allen and M.E. Andersen. 1991.  Physiologically based
pharmacokinetic modeling with trichloroethylene and its metabolite, trichloroacetic acid,
in the rat and mouse. Toxicol. Appl. Pharmacol.  109:183-195.

Franks,  N.P. and W.R. Lieb.  1994. Molecular and cellular mechanisms of general
anesthesia.  Nature. 367:607-614.
                                       59

-------
Gearhart, J.M., C. Seckel and A. Vinegar.  1993.  In vivo metabolism of chloroform in
B6C3F1 mice determined by the method of gas uptake: The effects of body temperature
on tissue partition coefficients and metabolism. Toxicol. Appl Pharmacol. 119:258-266.

Gearheart, J.M., G.W. Jepson, H.J.  Clewell, M.E. Andersen and R.B. Conolly. 1994.
Physiologically based pharmacokinetic model for the inhibition of acetylcholinesterase
by organophosphate esters.  Environ. Health Perspect.  102(Suppl. 11):51-60.

Gentry, P.R., D.E. Hack, L. Haber, A. Maierand H.J. Clewell III. 2002. An approach for
the quantitative consideration of genetic polymorphism data in chemical risk
assessment: Examples with warfarin and parathion. Toxicol. Sci.  70:120-139.

Haddad, S.,  G. Charest-Tardif and K. Krishnan. 2000. Physiologically based modeling
of the maximal effect of metabolic interactions on the kinetics of components of complex
chemical mixtures. J. Toxicol. Environ. Health A. 61(3):209-223.

Kedderis, G.L. 1997.  Extrapolation of in vitro enzyme induction data to humans in vivo.
Chem. Biol. Interact.  107:109-121.

Kedderis, G.L. and S.D. Held. 1996. Prediction of furan pharmacokinetics from
hepatocyte studies: Comparison of bioactivation and hepatic dosimetry in rats, mice and
humans.  Toxicol. Appl. Pharmacol.  140:124-130.

Kedderis, G.L. and J.C. Lipscomb. 2001.  Application of in vitro biotransformation data
and pharmacokinetic modeling to risk assessment.  Toxicol. Ind. Health. 17:315-321.

Klein, M.T., G. Hou, R. Quann et al. 2002.  BioMOL: A computer-assisted biological
modeling tool for complex chemical  mixtures and biological processes at the  molecular
level. Environ. Health Perspect.  110(Suppl. 6): 1025-1029.

Koizumi, A.  1989.  Potential of physiologically based pharmacokinetics to amalgamate
kinetic data of trichloroethylene and tetrachloroethylene obtained in rats and man.  Br. J.
Ind. Med. 46:239-249.

Koizumi, A.,  M. Kumai and M. Ikeda. 1982.  In vivo suppression of 1,1,1-trichloroethane
metabolism by co-administered tetrachloroethylene: an inhalation study. Bull. Environ.
Contam. Toxicol.  29:196-199.

Kousba, A.A., T.S. Poet and C. Timchalk.  2003.  Characterization of the in vitro kinetic
interaction of chlorpyrifos-oxon with  rat salivary cholinesterase: A potential
biomonitoring matrix. Toxicology. 188:219-232.

Levesque, B., P. Ayotte, R. Tardif et al. 2000.  Evaluation of the health risk associated
with exposure to chloroform in indoor swimming pools. J.  Toxicol. Environ. Health A.
61:225-243.
                                       60

-------
Liao, K.H., I. Dobrev, J.E. Dennison Jr. et al. 2002. Application of biologically based
computer modeling to simple or complex mixtures.  Environ. Health  Perspect.
110(Suppl. 6):957-963.

Lambert, J.C. and J.C. Lipscomb.  2007. Mode of action as a determining factor in
human health risk assessment of chemical mixtures.  Regul. Toxicol. Pharmacol.
49(3): 183-194.

Lipscomb, J.C. and C.M. Garrett.  1998.  Effect of organ procurement conditions on
cytochrome P-450 activity in rat liver microsomes. In  Vitro Mol. Toxicol.  11(3):265-270.

Lipscomb, J.C. and G.L. Kedderis. 2002. Incorporating human interindividual
biotransformation variance in health risk assessment. Sci. Total Environ.  288:13-21.

Lipscomb, J.C., C.M. Garrett and J.E. Snawder.  1997. Cytochrome P450-dependent
metabolism of trichloroethylene: Interindividual differences in humans.  Toxicol. Appl.
Pharmacol.  142:311-318.

Lipscomb, J.C., J.W. Fisher, P.O. Confer and J.Z. Byczkowski. 1998.  In vitroio in vivo
extrapolation for trichloroethylene metabolism in humans. Toxicol. Appl. Pharmacol.
152:376-387.

Lipscomb, J.C., L.K. Teuschler, J.C.  Swartout,  C.A.F. Striley and J.F. Snawder. 2003a.
Variance of microsomal protein and cytochrome P450 2E1 and 3A forms in adult human
liver. Toxicol. Mech. Methods. 13:45-51.

Lipscomb, J.C., L.K. Teuschler, J.C.  Swartout,  D. Popken, T. Cox and  G.L. Kedderis.
2003b. The impact of cytochrome P450 2E1-dependent  metabolic variance on a risk-
relevant pharmacokinetic outcome in humans.  Risk Anal. 23:1221-1238.

Loizou, G.D.  2001. The application  of physiologically based pharmacokinetic modelling
in the analysis of occupational  exposure to perchloroethylene.  Toxicol. Lett.  124:59-69.

McKone, T.E.  1993. Linking a PBPK model for chloroform with measured breath
concentrations in showers: Implications for dermal exposure models. J. Exposure Anal.
Environ. Epi.  3(3):339-365.

Mehendale, H.M.  2005. Tissue repair: An important determinant of final outcome of
toxicant-induced injury.  Toxicol. Pathol.  33:41-51.

Mumtaz, M.M. and P.R. Durkin. 1992. A weight of evidence approach for assessing
interactions in chemical mixtures.  Toxicol. Ind. Health. 8:377-406.

Mumtaz, M.M., K.A. Poirier and J.T. Colman. 1997. Risk assessment for chemical
mixtures: fine tuning the hazard index approach.  J. Clean Technol. Environ. Toxicol.
Occup. Med.  6:189-204.
                                      61

-------
Mumtaz, M.M., C.T. DeRosa, J. Groten, V.J. Feron, H. Hansen and P.R. Durkin. 1998.
Estimation of toxicity of chemical mixtures through modeling of chemical interactions.
Environ Health Perspect.  106(Suppl. 6): 1353-1360.

Poet, T.S., R.A. Corley, K.D. Thrall et al. 2000. Assessment of the percutaneous
absorption of trichloroethylene in rats and humans using MS/MS real-time breath
analysis and physiologically based pharmacokinetic modeling.  Toxicol. Sci.  56:61-72.

Poet, T.S., K.K. Weitz, R.A. Gies et al. 2002.  PBPK modeling of the percutaneous
absorption of perchloroethylene from a soil matrix in rats and humans.  Toxicol. Sci.
67:17-31.

Qiao, D., F.J. Seidler and T.A. Slotkin. 2001.  Developmental neurotoxicity of
chlorpyrifos modeled in vitro'.  Comparative effects of metabolites and other
cholinesterase inhibitors on DMA synthesis in PC12 and C6 cells.  Environ. Health
Perspect.  109(9):909-913.

Rao, H.V. and D.R. Brown. 1993.  A physiologically based pharmacokinetic
assessment of tetrachloroethylene in ground water for a bathing and showering
determination.  Risk Anal. 13(1):37-49.

Reisfeld, B. and R.S.H. Yang. 2004. A reaction network model for CYP2E1-mediated
metabolism of toxicant mixtures. Environ. Toxicol. Pharmacol. 18:173-179.

Reitz, R.H., J.N. McDougal, R.J. Nolan and A.M. Schumann.  1988.  Physiologically
based pharmacokinetic modeling with  methylchloroform: Implications for interspecies,
high dose/low dose, and dose route extrapolations. Toxicol. Appl. Pharmacol.
95:185-199.

Reitz, R.H., M.L. Gargas, A.L. Mendrala and A.M. Schumann.  1996. In vivo and in vitro
studies of perchloroethylene metabolism for physiologically based pharmacokinetic
modeling in rats, mice, and humans. Toxicol. Appl. Pharmacol. 136:289-306.

Roy, A., C.P. Weisel, P.J. Lioy and P.G. Georgopoulos.  1996. A distributed parameter
physiologically-based pharmacokinetic model for dermal and inhalation exposure to
volatile organic compounds.  RiskAnal.  16(2):147-160.

Sato, A., T. Nakajima, Y. Fujiwara and N. Murayama.  1977. A pharmacokinetic model
to study the excretion of trichloroethylene and  its metabolites after an inhalation
exposure.  Br. J. Ind. Med. 34:55-63.

Sato, A., K. Endoh, T. Kaneko and G. Johanson.  1991.  A simulation study of
physiological factors affecting pharmacokinetic behavior of organic solvent vapours.  Br.
J. Ind. Med.  48(5):342-347.

Shibata, A., K.  Morita, T. Yamashita, H. Kayama and I. Ueda.  1991. Anesthetic-protein
interaction:  Effects of volatile  anesthetics on the secondary structure of poly-(l-lysine).
J. Pharm. Sci.  80(11): 1037-1041.
                                      62

-------
Snawder, J.E. and J.C. Lipscomb. 2000. Interindividual variance of cytochrome P450
forms in human hepatic microsomes: Correlation of individual forms with xenobiotic
metabolism and implications in risk assessment. Regul. Toxicol. Pharmacol.
32:200-209.

Sultatos, L.G.  1990. A physiologically based pharmacokinetic model of parathion
based on chemical-specific parameters determined in vitro.  J. Amer. Coll. Toxicol.
9(6):611-619.

Tardif, R. and G. Charest-Tardif.  1999.  The importance of measured end-points in
demonstrating the occurrence of interactions: A case study with methylchloroform and
m-xylene.  Toxicol Sci. 49:312-317.

Thrall, K.D. and T.S. Poet.  2000.  Determination of biokinetic interactions in chemical
mixtures using real-time breath analysis and physiologically based pharmacokinetic
modeling.  J. Toxicol. Environ. Health A. 59(8):653-670.

Timchalk, C., A. Kousba and T.S. Poet.  2002a.  Monte Carlo analysis of the human
chlorpyrifos-oxonase (PON1) polymorphism using a physiologically based
pharmacokinetic and pharmacodynamic (PBPK/PD) model.  Toxicol Lett. 135:51-59.

Timchalk, C., R.J.  Nolan, A.L. Mendrala, D.A. Dittenber, K.A. Brzakand J.L. Mattsson.
2002b. A physiologically based pharmacokinetic and pharmacodynamic (PBPK/PD)
model for the organophosphate insecticide chlorpyrifos in rats and humans. Toxicol Sci.
66:34-53.

Ueda, I. and H. Kamaya.  1984. Molecular mechanisms of anesthesia. Anesth. Analg.
63:929-945.

U.S. EPA. 1986. Guidelines for Health Risk Assessment of Chemical Mixtures. Fed.
Reg. 51 (185):34014-34025.

U.S. EPA. 1989. Risk Assessment Guidance for Superfund, Volume I (RAGS), Human
Health Evaluation Manual, Part A, Interim Final.  December. OSWER 9285.7-02B,
EPA/540/1-89/002. Office of Emergency and Remedial Response, U.S. Environmental
Protection Agency, Washington, DC. Available at
http://www.epa.gov/superfund/programs/risk/ragsa/index.htm.

U.S. EPA. 1998. Rule Writer's Guide to Executive Order 13045:  Guidance for
Considering Risks to Children During the Establishment of Public Health-Based and
Risk-Based Standards. U.S. Environmental Protection Agency, Office of Children's
Health Protection.  Washington, DC.  November.

U.S. EPA. 2000a. Supplementary Guidance for Conducting Health Risk Assessment of
Chemical Mixtures.  U. S. Environmental Protection Agency, Risk Assessment Forum,
Washington, DC. EPA/630/R-00/002. Available at
                                   m       	rnjx08_20p_1.pdf.
                                      63

-------
U.S. EPA. 2000b. Cumulative Risk: A Case Study of the Estimation of Risk from 24
Organophosphate Pesticides.  U.S. Environmental Protection Agency,  Office of
Pesticide Programs, Health Effects Division, Washington, DC. Novembers. Available


U.S. EPA. 2002a. Guidance on Cumulative Risk Assessment of Pesticide Chemicals
that Have a Common Mechanism of Toxicity. U.S. Environmental Protection Agency,
Office of Pesticide Programs, Washington, DC. January 14.  Available at


U.S. EPA. 2002b. Organophosphate  Pesticides: Revised OP Cumulative Risk
Assessment. U.S. Environmental Protection Agency, Office of Pesticide Programs,
Washington, DC. Available at

U.S. EPA. 2003a. Physiologically-Based Pharmacokinetic/Pharmacodynamic
Modeling: Preliminary Evaluation and Case Study for the n-Methyl Carbamate
Pesticides.  U.S. Environmental Protection Agency, Office of Prevention, Pesticides,
and Toxic Substances, Washington, DC.  November 10.  Available at


U.S. EPA. 2003b. A Set of Scientific Issues Being Considered by the Environmental
Protection Agency Regarding: Physiologically-based Pharmacokinetic/
Pharmacodynamic Modeling: Preliminary valuation and Case Study for the N-Methyl
Carbamate  Pesticides: A consultation. Presented at FIFRA Scientific Advisory Panel
Meeting,  Sheraton Crystal City Hotel, Arlington, VA.  December 11-12. SAP Minutes
No. 2003-06. Available at


U.S. EPA. 2003c. Framework for Cumulative Risk Assessment.  U.S. Environmental
Protection Agency, Risk Assessment Forum, Washington, DC. EPA/630/P-02/001F.
Available at

U.S. EPA. 2006. Exposures and Internal Doses of Trihalomethanes in Humans: Multi-
Route Contributions from Drinking Water. Office of Research and Development,
National Center for Environmental Assessment, Cincinnati, OH. EPA/R-06/087.

U.S. EPA. 2007a. Revised A/-Methyl Carbamate Cumulative Risk Assessment U.S.
Environmental Protection Agency Office of Pesticide Programs. Docket # EPA-HQ-
OPP-2007-0935



U.S. EPA. 2007b. Integrated Risk Information System (IRIS). Online. Office of
Research and Development, National Center for Environmental Assessment,
Washington, DC. Available at
                                     64

-------
Vogel, J.S., G.A. Keating II and B.A. Buchholz. 2002. Protein binding of isofluorophate
in vivo after coexposure of multiple chemicals.  Environ. Health Perspect.  110(Suppl.
6): 1031-1036.

Ward, R.C., C.C. Travis, D.M. Hetrick, M.E. Andersen and M.L. Gargas. 1988.
Pharmacokinetics of tetrachloroethylene.  Toxicol. Appl. Pharmacol.  93:108-117.

Whyatt,  R.M., D.E. Camann, P.L. Kinney et al.  2002.  Residential pesticide use during
pregnancy among a cohort of urban minority women.  Environ.  Health Perspect.
110(5):507-514.

Wilson, J.D. and J. Knaak. 1994.  Health risk assessment of tetrachloroethylene using
a physiologically-based pharmacokinetic model. Abstr. Pap. Am. Chem. Soc.
207:62-ENVR.

Yang, R.S.H., K.H. Liao and B.  Reisfeld. 2004a. The integration of computer modeling
and experimental toxicology for the study of chemical mixtures and multiple stressors.
Arch. Complex Environ. Factors. In press.

Yang, R.S.H., H.A. EI-Masri, R.S. Thomas et al. 2004b. Chemical mixture toxicology:
From descriptive to mechanistic, and going on to in silico toxicology.  Environ. Toxicol.
Pharmacol. 18:65-81.
                                      65

-------
                              APPENDIX A
          A GENERAL REFERENCE ON PHYSIOLOGICALLY-BASED
       PHARMACOKINETIC/PHARMACODYNAMIC (PBPK/PD) MODELING

Yang, R.S.H., M.E. Andersen, J.E. Dennison, Y.C. Ou, K.H. Liao and B. Reisfeld.  2004.
Physiologically based pharmacokinetic and pharmacodynamic modeling. Chapter 23.
In: Mouse Models of Cancer, E.G. Holland, Ed.  Wiley Inc., New York, NY. p. 391-405.
                                  A-1

-------
23
PHYSIOLOGICALLY BASED PHARMACOKINETIC
AND PHARMACODYNAMIC  MODELING


RAYMOND S. H. YANG AND JAMES E. DENNISON
Quantitative and Computational Toxicology Group, Center for Environmental Toxicology and Technology, Department of
Environmental and Radiological Health Sciences, Colorado State University, Fort Collins, Colorado
MELVIN E. ANDERSEN
Department of Biomathematics and Physical Sciences, CUT Centers for Health Research, Research Triangle Park, North Carolina
YING C. Ou
Human Genome Sciences, Inc., Rockville, Maryland
KAI H. LIAO AND BRAD REISFELD
Quantitative and Computational Toxicology Group, Center for Environmental Toxicology and Technology, Department of Chemical
Engineering, Colorado State University, Fort Collins, Colorado
INTRODUCTION

At  first  glance,  the  topic  of physiologically  based
pharmacokinetic/pharmacodynamic (PBPK/PD) modeling
appears to be outside the central  theme of this  book.
However, with the advances of modern  biology and
computational technology, it is just a matter  of time
for  any area  of  biomedical sciences to be integrated
with computer science.  Two  aspects  are  particularly
important for  the application of PBPK/PD modeling  in
cancer research, thus underscoring the relevance for the
inclusion of this chapter in the book: (1) incorporation  of
PBPK/PD modeling in any biomedical research can avofd32
unnecessary experiments, thereby  conserving precious
resources, and (2) computer  simulations (i.e., in  silico
experimentation) using validated PBPK/PD models will
minimize animal usage.
  The  intent of this chapter is to (1) introduce the gen-
eral concept and background knowledge of PBPK/PD
modeling;  (2) provide  some  examples  of application
of  PBPK/PD modeling;  (3) illustrate  the  utility  of
PBPK/PD modeling, particularly in the  pharmaceutical
Mouse Models of Human Cancer, edited by Eric C. Holland
ISBN 0-471-44460-X Copyright © 2004 John Wiley & Sons, Inc.
drug development process; and (4) project future devel-
opment on "second-generation" PBPK/PD modeling and
In silico toxicology. An emphasis was given to  intro-
ducing the concepts of PBPK/PD modeling rather than
the details of its techniques and processes. For more
detailed background and conceptual information on PBPK
modeling, the readers are referred to  our earlier  intro-
ductory discussion (Yang and  Andersen, 1994) and the
two  "timeless papers"  by Bischoff and Brown (1966)
and Dedrick (1973). For more extensive information, par-
ticularly regarding modeling techniques, the reader should
consult the two future books from our laboratories (Reddy
et al.,» 2004; Yang et al, 2004).
WHAT IS PHYSIOLOGICALLY BASED
PHARMACOKINETICS? WHAT ARE THE
DIFFERENCES BETWEEN PBPK AND
CLASSICAL PHARMACOKINETICS?

Physiologically based pharmacokinetics (PBPK), as the
name  implies, is  a special branch of  pharmacokinetics
                                                                                                   391
                                                A-2

-------
392
       PHYSIOLOGICALLY BASED PHARMACOKINETIC AND PHARMACODYNAMIC MODELING
where physiology and anatomy of the animal or human
body as well as the biochemistry of the chemical or chem-
icals of interest are incorporated into the conceptual model
for computer simulation. Classicalpharmacokinetics refer
to those empirical noncompartmental or compartmental
pharmacokinetic studies routinely practiced in the pharma-
ceutical industry (van de Waterbeemd and Gifford, 2003).
As will  be illustrated later, the compartments of a PBPK
model have anatomic  and physiologic significance. This
is a major difference  from empirical noncompartmental
or compartmental pharmacokinetic modeling approaches.
PBPK models can be used to describe concentration-time
profiles  in individual  tissue/organ and in the plasma or
blood. When the concentration of a certain target tissue,
rather than the plasma  concentration, is highly related to a
compound's efficacy or toxicity, PBPK modeling will be
a more useful tool than classical pharmacokinetic models
for describing  PBPK/PD relationships and thus make a
better prediction of the time course of drug effects result-
ing from  a certain  dose regimen for the compound of
interest. Furthermore, PBPK models in combination with
absorption simulation  and quantitative structure-activity
relationship (QSAR) approaches will bring us closer to
a full prediction of drug disposition for pharmaceutical
new  entities and  help streamline the selection of lead
drug candidates in  the  drug discovery  process (van de
Waterbeemd and Gifford, 2003). Lastly,  unlike empirical
noncompartmental and compartmental pharmacokinetics,
PBPK modeling is a powerful tool for  extrapolation, be
it for interspecies, interroutes, interdoses, interlife stages,
and so on.
   The concept of PBPK had its embryonic development
in the  1920s  and  1940s;  for  a more detailed  early
history,  readers are referred to two books which are in
preparation (Reddy et al, 2004; Yang et al, 2004). PBPK
modeling blossomed and flourished in the late 1960s and
early 1970s in the chemotherapeutic area mainly due to
the efforts of investigators  with expertise in chemical
engineering  process design and  control. Two  notable
pioneers in this development are Kenneth B. Bischoff,
then at the University of Texas, Austin, Texas, and Robert
Dedrick of Biomedical Engineering and Instrumentation
Branch, Division of Research Services, National Institutes
of Health, Bethesda, Mariland. Two timeless publications
by these investigators are, respectively, "Drug Distribution
in Mammals" (Bischoff and Brown, 1966) and  "Animal
Scale-Up" (Dedrick,  1973);  these  articles  are  highly
recommended  to  those  who  are interested in  PBPK
modeling. In  the mid-1980s,  two  articles  on  PBPK
modeling  of styrene and methylene chloride (Andersen
et al., 1987; Ramsey  and Andersen,  1984) started yet
another "revolution" in the toxicology and risk assessment
arena. Today, there are more than 700 publications directly
related  to PBPK  modeling  on industrial  chemicals,
drugs, environmental pollutants, and simple and complex
chemical mixtures (Reddy et al., 2004).
   A PBPK model, graphically illustrated in Figure 23.1,
reflects  the  incorporation  of  basic  physiology  and
anatomy.  The  compartments  actually  correspond to
anatomic entities such as  liver, lung, and  so on, and
the blood circulation conforms to the basic  mammalian
physiology. In this specific model, a published example on
methylene chloride, it is quite obvious that the exposure
route of interest is inhalation because the lung and gas
exchange compartments are prominently  displayed with
intake (CI) and exhalation (CX) vapor  concentrations
indicated.  Oral  and/or  dermal exposures  may  be added
easily to the gastrointestinal tract compartment or general
venous circulation, respectively. Some tissues (e.g., richly
or slowly perfused tissues in  Fig.  23.1)  are  "lumped"
together because there is insufficient evidence to conclude
that each of these tissues is kinetically distinct enough for
the specific chemical to warrant a separate compartment.
   If one  draws an  analogy of the "scale-up"  from a
laboratory chemical  engineering process to  a chemical
plant to the scale-up of a mouse to a human,  one finds
that both  situations  are  governed  by a  great number
of physical and chemical  processes. In  mammals, the
physical processes (i.e., mass balance, thermodynamics,
transport,  and flow) often vary in  a  predictable way.
However, chemical processes such as metabolic reactions
may vary greatly and are  less predictable among species.
These physical and chemical processes interact in the body
such that  the pharmacokinetics of  any given chemical
between one  species and  another may be more (or less)
predictable depending  on  the amount of  background
information available.
HOW DOES A PHYSIOLOGICALLY BASED
PHARMACOKINETIC MODEL WORK?

A PBPK  model applies fundamental physiologic, bio-
chemical,  and engineering principles to describe the dis-
tribution and disposition of chemicals in the body at any
given time. The process and approach may be summarized
in a flowchart (Fig. 23.2). Once the chemical of interest
and the  problems needing to be addressed  are identified,
a thorough literature evaluation is  conducted.
   The fundamentals of PBPK modeling are to identify
the principal organs or tissues involved in the disposition
of the chemical of interest and to correlate the chemical
absorption, distribution, metabolism, and excretion within
and among these organs and tissues in an integrated and
biologically plausible manner. A scheme is usually formed
where the  normal  physiology  is  followed in a  graphi-
cal manner (i.e., a conceptual  model  as  in Fig. 23.1).
Within the boundary of the identified compartment (e.g.,
                                                    A-3

-------
                                     HOW DOES A PHYSIOLOGICALLY BASED PHARMACOKINETIC MODEL WORK?    393
                         Cl
                    cv
                                  QP
                                             CX
                                                         AM1LU
                                                             if
                                                              ""
                                  Gas
                                exchange
                                                  QC
                                                  CAI
               AM2LU

                 A
               KF
      Lung
   metabolism
                           QR
                           CVR
                           QF
                                              Richly perfused
                                                                         QR
                           CVF
                           QS
                                                    Fat
                                                                         CA
                                                                         QF
                                                                         CA
                           cvs
                           QL
                                             Slowly perfused
                           CVL
                                             .
                                           /
                                                   Liver
.
\
                                                                         CA
                                                                         QL
                                                                         CA
                                    AM2L
                                                                AM1L
                                                                                 c/
                                                                                     KZER
                                                                              G.I. TRACT
                Figure 23.1. Graphical representation of a PBPK  model for methylene chloride. (Andersen
                et al., 1987).
an organ or tissue or a group of organs or tissues), what-
ever "comes" in  must be  accounted for via whatever
"goes out"  or  whatever is transformed into something
else. This "mass balance" is expressed as a mathemati-
cal equation with appropriate parameters carrying biologic
significance. A series of such equations representing all of
the interlinked compartments are formulated to express a
mathematical representation, or model, of the biologic sys-
tem. This model can then be used for computer simulation
to predict the time course behavior of any given parame-
ter. Three sets of parameters are needed for PBPK model
building: physiologic parameters (e.g., ventilation  rates,
cardiac output, organs as percent body weight), thermody-
namic parameters  (e.g., tissue partition coefficients, flow
rates), and biochemical parameters (e.g.,  Km  and Vmax).
Most, if not all, of the parameters for laboratory animals
are available in relevant literature, such as the Biologi-
cal Data Book and the special report by the International
Life Sciences Institute (ILSI) on the compilation of physi-
ologic parameters for PBPK models (Brown et al., 1997).
When information  gaps exist, the solution is either  an
empirical one via experimentation  or through allometric
extrapolation, usually based on a power function of the
body weight (Lindstedt, 1987).
   The  U.S.  Environmental  Protection  Agency (EPA)
guidance document (2002) includes a very nice discussion
on "modeling the data." Although it is not necessarily for
PBPK modeling, the discussion reflects some "dos" and
"don'ts" on computer  modeling. We quote some of the
passages below:
   The* selection of a mathematical model structure to fit
   the data being analyzed should be guided by the biology
   of the common mechanism of toxicity,  the toxicokinet-
   ics of the chemicals,  and the  observed shapes  of their
   dose-response curves  and the experimental designs used
   to generate the data. If available, pharmacodynamic and
   pharmacokinetic  data should be  considered in order to
   account for tissue concentrations and to aid in defin-
   ing dose-response relationships across different  species,
   routes and time-frames of exposure ....

     ... Although it is not possible to recommend the use
   of specific models, a few points that should be considered
   in modeling the data follow:


     •  Modeling  of individual animal data is desir-
        able; however, if this is not practical, then use
• Q3
                                                      A-4

-------
394
       PHYSIOLOGICALLY BASED PHARMACOKINETIC AND PHARMACODYNAMIC MODELING
                         Mechanisms
                          of toxicity
                          Refine
                          model
                                                      Problem
                                                    identification
                                                     Literature
                                                     evaluation
Biochemical
 constants
                                                      Model
                                                    formulation
                                                     Simulation
Compare to
kinetic data
                                                   Design/conduct
                                                 critical experiments
Physiological
 constants
  Validate
   model
                            Extrapolate
                            to humans
                Figure 23.2. Flowchart illustrating processes  involved in PBPK.  (From M.  E.  Andersen,
                Pharmacokinetics in Risk Assessment, Drinking Water and Health, National Academy Press,
                Washington, D.C., 1987, pp. 8-23.)
       of summary data such as means and standard
       deviations can be alternatives
       Care should be taken with modeling high-dose
       data (particularly extreme doses) because the
       model shape in the  low-dose region can  be
       influenced by high-dose data
       Log transformation of data  should be justified
       because  such a transformation may distort the
       dose-response curve
       Data variability should be described by appro-
       priate statistical techniques and reflected in the
       potency  estimate (e.g., by weighting the data
       in the fitting procedure)
       Confidence  intervals or  limits should  be
       included  in  the  analysis  because  they can
       be  valuable for  evaluating the influence  of
       variability on the potency estimates
       An estimate for the uncertainty  of the model
       used in the analysis should  be included
       The statistical fitting method used  must  be
       clearly described.
       For the most well-studied chemicals or drugs, it is likely
       that the biochemical constants  such as Km's and Vmax's
       are known and readily retrievable from  the  information
       data base.  However, it  must be  made  clear  here  that
       the Km,  Vmax,  and  KF  (first-order rate  constant) in a
       PBPK  model (known as in vivo  Km,  Vmax, KF for a
       given chemical) such as the ones given in Figure 23.1 are
       hybrid  constants of all the saturable or linear metabolic
       pathways, respectively, for  the chemical of interest  in
       the organ  and/or  body.  They  are different  from  the
       in  vitro  Km, Vmax,  or  KF  of a  given  pure  enzyme.
       While they are not directly  interchangeable,  the in vitro
       constants  in the  literature  may  be used  to  estimate
       in  vivo constants for modeling purposes (Kedderis  and
       Lipscomb, 2001; Lipscomb  et al.,  1998).  Also,  for most
       well-known  chemicals, it is  likely  that enough is known
       about the  mechanism of toxicity  to  be incorporated
       into  the  model for  computer  simulation.  Physiologic
       constants  such  as  organ  volumes  and blood flow rates
       are usually  available in  the literature for the  common
       laboratory animals  as well as humans.  Therefore,  at
       least in those instances  of "well-known" chemicals, a
                                                     A-5

-------
                                                  PHYSIOLOGICALLY BASED PHARMACODYNAMIC MODELING
                                                                                                          395
model may be conceptually illustrated as in Figure 23.1
and mathematically represented by a number of mass
balance  differential equations. Computer simulations may
be made for any number of desired time course endpoints
such as  the blood levels of the parent compound, liver
level  of a  reactive metabolite, and similar information
on  different  species,  at lower or  higher  dose  levels,
and/or via a different route of exposure. The experimental
pharmacokinetic data may then be compared with PBPK
model simulation  to see if they are superimposable upon
each  other. If this is indeed  the  case, the model is
consistent with actual results.  Validation of the PBPK
model with data sets other than the working set (or training
set) to develop the model is necessary. Once validated, the
PBPK model is ready for extrapolation  to  other animal
species,  including humans. However, if the  experimental
data and PBPK  model  simulation are not  consistent,
the model  might  be deficient  because critical scientific
information might  be missing or  certain  assumptions
are incorrect. The investigator, with knowledge of the
chemical and a general understanding of the physiology
and biochemistry  of the animal species, can design and
conduct critical experiments  for refining the model to
reach consistency with experimentation (Fig. 23.2). This
refinement  process may be repeated again  and again
when necessary;  such an iterative  process  is critically
important for the  development of a  PBPK model. There
is  always the possibility  that  a  good model may not
be  obtained at the time  because of the limitation of
our knowledge on the  chemical.  An  emphasis must
be made here that the investigator and the knowledge
possessed are the single  most important determinant of
the outcome of the results; mathematical modeling  and
advanced computational  capabilities are nothing more
than good tools.
MYTHS ABOUT PHYSIOLOGICALLY BASED
PHARMACOKINETICS; ARE THEY REAL?

Common belief is that PBPK modeling is highly resource
intensive and very difficult to do, particularly  involving
interactive pharmacokinetics or pharmacodynamics. This
warrants some special  discussion:  First,  by  the  time
chemicals such as drugs or pesticides reach the stage of
commercialism, many types of studies have already been
conducted, including pharmacokinetic studies. Thus, they
are already resource intensive during the developmental
stage to become successful chemicals in commerce. The
important point is  how may quantitative,  time course
data useful for PBPK model development be  generated
during the product development phase? The data required
for PBPK modeling are really not too much different
from those required for the  classical pharmacokinetics
in the  present  IND/NDA  (investigational new  drug
application/new drug application) process. A slight new
orientation to the  existing battery of  studies would
generate adequate quantitative time course data for PBPK
modeling. Thus, it  is definitely  not any more  resource
intensive from the existing requirements. Furthermore, if
the incentive (e.g., regulatory guideline-driven scientific
studies) is there, such quantitative, time course data would
have been automatically generated during the product
development  phase.  In  fact,  PBPK modeling,  being a
hypothesis testing tool  in  toxicology,  may be utilized
to conduct many different kinds of experiments on a
computer  (i.e., in  silico toxicology).
   A great deal of  research effort has been devoted to
the development of  simple,  high-throughput, and in vitro
predictive tools in the pharmaceutical industry (van de
Waterbeemd and Gifford, 2003).  While these short-term,
rapid assays certainly offer some utility, particularly in
the early drug development process, an inherent danger
of such tools is the "by-pass" of integrated mammalian
physiology and architecture of the body. In that regard,
development  of in  silico toxicology such as  PBPK/PD
modeling  and other biologically  based  computer model-
ing has the advantage of integrating  whole-body phar-
macokinetics and  pharmacodynamics with computational
technology. The resulting predictive tools will minimize
unnecessary experiments and improve the attrition rate of
drug or chemical  product development with much more
scientific validity and confidence. In that sense, PBPK/PD
modeling, once integrated as a part of the product devel-
opment plan, will  actually save expenses and resources as
well as minimize unnecessary animal experiments.
   Second, while PBPK modeling  is  by  no  means a
very  easy technology, it is not  any more difficult than
some of the classical pharmacokinetic and statistical mod-
eling  carried out in  the routine product  development
process.  Furthermore,  excellent  training   opportunities
(e.g., www.cvnibs.colostate.edu/enhealth/cett/) are avail-
able, and the development of software is such that more
and more  user-friendly tools are going to be available.
PHYSIOLOGICALLY BASED
PHARMACODYNAMIC MODELING

Using plain English, pharmacokinetics can be considered
as "What the body does to the chemicals," and pharma-
codynamics can thus be considered as "What the chemi-
cals do to the body." Physiologically based pharmacody-
namic (PBPD) modeling is therefore computer simulation
of pharmacologic or toxicologic effects  of chemicals or
drugs based on  the biology  of  chemical/drug-receptor
interactions. From the point  at which a chemical or a
drug enters into the body to the point of pharmacologic or
                                                    A-6

-------
396
       PHYSIOLOGICALLY BASED PHARMACOKINETIC AND PHARMACODYNAMIC MODELING
toxicologic effect, it is a continuum of pharmacokinetics
and pharmacodynamics. It is difficult to distinguish where
pharmacokinetics end and pharmacodynamics start. When
we consider the pharmacology or toxicology of a chem-
ical or a drug, we must consider both pharmacokinetics
and pharmacodynamics to  have a full understanding. To
consider either area alone  is to understand only  part of
the picture.
   PBPK modeling preceded PBPD modeling  by many
years because of the slower and later advances of  mecha-
nistic understanding of modes of action of chemicals and
drugs. It is a natural course of evolution that PBPK blos-
somed and flourished first because it was developed based
on the fundamentals of mammalian physiology, analytical
chemistry, engineering principles of mass  transport and
mass balance, and desk-top computer hardware and soft-
ware development. These were all  well-developed areas
in the earlier days. PBPD modeling is dependent upon the
mechanistic basis of chemical/drug-receptor interactions,
and the biology associated with it took time to mature. The
recent trend is such that more and more PBPD  modeling
work is  evident. PBPK modeling has matured to the point
that research endeavors are centered around more com-
plex systems (e.g., multichemical interactions, subcom-
partmentalization of  organs) or  special  problem-driven
models  (e.g., dermal uptake of chemicals from showers,
exposure dose reconstruction). As will be discussed later,
the "delumping" of organ  and tissue compartments and
the linkage  of PBPK models  with other  types  of bio-
logically based models may lead  us to second-generation
PBPK/PD modeling.
DATA REQUIREMENTS FOR PBPK OR PBPD
MODELING

What are the specific data needed for building PBPK mod-
els?  Obviously, well-conducted in vivo pharmacokinetic
data  are essential, and usually the more the data sets (e.g.,
different doses, routes, species), the better. In each study,
time course blood and tissue concentration data are essen-
tial.  These time  course  data should include  at least the
following tissues and organs: blood (or plasma if blood
cell binding is not an issue), liver (organ of metabolism),
kidney (representing well-perfused organs/tissues), muscle
(representing  slowly  perfused organs/tissues), and target
organ(s)/tissue(s). We also need other PBPK-modeling
specific information,  such  as (1) physiologic constants,
including body size, organ and tissue volumes, blood flow,
and  ventilation rates; (2) biochemical constants,  includ-
ing the chemical-specific metabolic rate constants  such
as Vmax  and  Km, partition coefficients for  tissues;  and
(3) mechanistic factors  such as target tissues, metabolic
pathways, and receptor interactions. Enzyme kinetic data,
particularly human data, of at least the key metabolic pro-
cesses  will be important for  the PBPK model. In vitro
determination of tissue partition coefficients and enzyme
kinetic data is relatively  straightforward  and inexpen-
sive. With modern genetic engineering technologies, many
human enzymes are available commercially. Thus, hereto-
fore unavailable human enzyme kinetic information for
many of the environmentally important chemicals  are
within  easy reach for many laboratories. These experi-
ments should be performed.
   For  PBPD modeling,  the  data requirement is  much
more variable because of the many different types of
chemical/drug-receptor interactions.  Thus, it is  much
more of a compound-specific nature. However, as  a rule
of thumb, quality time course data on  the key biologic
processes to be modeled are essential.
PBPK OR PBPK/PD MODELS FOR CHEMICAL
INTERACTIONS (INTERACTIVE PBPK
OR PBPK/PD MODELS)

Human exposure to chemicals is rarely, if ever, to single
chemicals. In the area of clinical pharmacology, adverse
drug interactions are undoubtedly serious concerns. For
instance, Lazarou et al. (1998) estimated that there  were
over 2.2 million cases of serious adverse drug reactions
(ADRs) in hospital patients in  1994 in the United States
and among these cases 106,000 were fatal.  During  their
hospital stay, the patients in the  survey statistics  were
given an average of eight drugs. Comparing with other
statistics of causes  of death,  these investigators (Lazarou
et al., 1998) indicated that ADRs became the fourth to the
sixth leading  cause of death for that year in the United
States. Thus,  it  is important  to discuss the issues related
to the development  of  interactive PBPK/PD modeling
(see  particularly the example  given later on the work
of Kanamitsu et al., 2000).
  From the perspective of interactive PBPK/PD model-
ing, two aspects need to be addressed: pharmacokinetic
interactions and pharmacodynamic interactions.  Because
PBPD  modeling is  a relatively  recent effort and few
such models are available for even single chemicals, we
will concentrate all our  discussion on interactive PBPK
models. The most ideal and scientifically defensible data
requirement for establishing  an interactive PBPK model
is that each component chemical in the mixture already
have its respective established PBPK model and that there
are many pharmacokinetic data sets in laboratory animals
as well as in  humans available for each of these compo-
nent chemicals.  The interactive PBPK model is then  built
on the basis of  known pharmacokinetic interactions. For
instance, the component chemicals may inhibit each oth-
ers' biotransformation. The individual PBPK models may
                                                    A-7

-------
                                                                       APPLICATION OF PBPK/PD MODELING
                                                                                                            397
then be linked together at the liver compartment by intro-
ducing competitive inhibition (or other types of inhibition)
terms in the mass balance differential equation.
   In some cases, even more specific and stringent data
requirements are needed; an example may be the ultralow-
dose  pharmacokinetic data in perinatal developmental
stages of laboratory  animals for extrapolation to human
fetuses and babies using  PBPK modeling. This  area is
still  in its infancy  and further development is antici-
pated.
APPLICATION OF PBPK/PD MODELING

Dosing Schedule of a Chemotherapeutic Agent

Methotrexate  is a  folate analogue  and  a well-known
cancer  Chemotherapeutic agent.  This  is one of  the
Chemotherapeutic agents which  was studied  extensively
with  PBPK  modeling.  The  mechanism  of  toxicity is
due  to the  binding  of methotrexate to  dihydrofolate
reductase, a  key enzyme in  DNA synthesis, leading to
cessation  of DNA  synthesis and cell  death.  In  some
organs (liver, kidney, intestine,  and marrow) and  many
tumors, methotrexate undergoes metabolism to  active
polyglutamate derivatives (even more potent inhibitors for
dihydrofolate  reductase).  Because these polyglutamates
are retained  in some  tissues far longer than the parent
compound,  it has  been  suggested  that  this  effect is
of great importance for the antineoplastic property of
methotrexate.
   Table 23.1  shows the toxicity of methotrexate to mice
under  a variety  of dosing  schedules (Morrison  et al.,
1987). Obviously,  toxicity  does not directly  correlate
with  total  dose. Decreasing  total dose by a factor of
117  (350/3) led to an increase, rather than a substantial
decrease,  in  toxicity. In addition, the area under  the
plasma concentration-time  curve (AUC), a frequently
used pharmacokinetic parameter for bioavailability,  did
not correlate with toxicity. For instance, the  AUC for a
bolus dose of 350 mg/kg is about two orders of magnitude
greater than that of the 96-h infusion at 0.8 |xg/h», yet
toxicity is higher with infusion.
   The reason for the above phenomenon turned out to
be intimately  associated with  the  pharmacokinetics of
methotrexate (and its polyglutamate metabolites) and the
threshold and length of time that dihydrofolate reductase
is inhibited (thus inhibition of DNA synthesis). At a bolus
dose of 350 mg/kg, even though there is a short period of
very high blood and tissue concentration, the inhibition
of DNA synthesis did not persist long enough to cause
lethality  in  at least some of the mice. At an  infusion
rate of 0.8 |xg/h» for 96 h,  even though the blood and
tissue levels were low, they were nevertheless high enough
to  cause sustained inhibition of DNA  synthesis, which
ultimately translated into higher lethality in  the animals.
These scientific discoveries eventually led to the  revision
of the methotrexate PBPK model by incorporating the
inhibition of DNA synthesis into the model.
   From  the modeling perspective,  the iterative process
of making new scientific discoveries and then refining the
PBPK model by incorporating such new information into
the model is a wonderful illustration of what we discussed
in  relation  to  Figure 23.2.  An even  more significant
illustration is the fact that a validated methotrexate PBPK
model can be used  to  conduct all  the experiments in
Table 23.1 on a computer. With such complicated dosing
schedules, it is apparent how the lives of hundreds of mice
may be saved and how much time and resources may be
diverted to other more efficient  usages.
"Electronic Rats"

One of our earlier examples was the PBPK/PD modeling
of a toxicologic interaction between kepone (also known
as chlordecone) and carbon tetrachloride (CCLt) based on
mechanisms of interactive toxicity and  the application
of computer technology in acute toxicity studies.  This
was a collaboration among three research groups: Melvin
E. Andersen, CUT  (presently CUT  Centers for  Health
Research);  Harihara M. Mehendale, Northeast Louisiana
>Q4
                                                                                                                  >Q5
Table 23.1.  Dosing Schedule Dependence of Methotrexate Toxicity in Mice
Dose
(mg/kg)
350
25
3

0.5
0.8 [ig/h
Schedule
Single dose
Twice daily
Every 3 h, 5 times, rest 8 h,
then every 3 h, 3 times
Every 3 h, 20 times
Infusion 96 h
Total Dose
(mg/kg)
350
50
24

10
3
Peak Plasma
Concentration (M)
io-3
io-4
io-5

io-6
io-8
Effect
LD50
LD50
>LD50a

>LD50a
>LD50a
"Higher toxicity than LDso.
Source:  Morrison et al., 1987.
                                                     A-8

-------
398
       PHYSIOLOGICALLY BASED PHARMACOKINETIC AND PHARMACODYNAMIC MODELING
                CCI4 + kepone
                                                 KINJ
                                           KREP
                Figure 23.3.  A conceptual PBPD  model for CCU and  Kepone interaction.  KMIT = rate
                constant for mitosis; KBIR = rate constant for cell birth; KINJ = rate constant for cell injury;
                KDIEI = rate constant for general cell death; KDIE1 = rate constant for cell death due to injury;
                KPH = rate constant for phagocytosis (El-masri et  al., 1996).
University (presently University of Louisiana at Monroe);
and our Quantitative and Computational Toxicology group
at Colorado State University. The details of this study
may be found elsewhere (El-masri et al., 1996). Briefly,
CC14 is a well-known hepatotoxin. Following free-radical
formation through the P450 enzyme system, the toxicity
of CCU  can  be an  accumulation  of  lipids (steatosis,
fatty liver)  and degenerative processes leading  to  cell
death (necrosis). Kepone is found in  the  environment
as a result of photolytic oxidation  of Mirex, a pesticide
used for  the control of fire ants, or as  a pollutant from
careless and irresponsible discharge. At relatively  low
levels  (e.g., 10 ppm  in the diet), even  repeated dosing
of kepone in the diet up to 15 days caused no apparent
toxicity to the liver. The toxicologic interaction between
kepone and CC14 was reported by Curtis et al. (1979).
They illustrated that a 15-day dietary exposure of male
rats to kepone at 10 ppm,  an environmentally realistic
level,  markedly  enhanced  liver  toxicity  produced  by
an i.p. injection of  a  marginally toxic dose of CCU
(100 (iL/kg). This  toxicologic interaction is unique in
that (1) unlike many other toxicologic interaction studies
which  were usually dealing  with acute  toxicity at very
high doses,  kepone in this instance was administered at
a very  low environmental level; (2) CCU was also dosed
at a  low,  marginally  toxic level; and (3) the magnitude
of toxicologic  interaction, 67-fold, is  very large.  The
mechanism of this toxicologic interaction was elucidated
by Mehendale' s group through a series of studies to be the
impairment, by kepone, of the liver's regeneration process.
These mechanistic studies were summarized in a number
of publications (Mehendale, 1984, 1991, 1994).
   As shown in Figure 23.3,  a PBPD  model  was con-
structed by El-masri et al. (1996) based on the mechanism
of toxicologic interaction between kepone and CCU. This
PBPD model was  verified by  literature  information, and
it was capable of providing time course computer simu-
lations of mitotic, injured, and pyknotic  (dead) cells after
treatment with CCU alone or  with kepone pretreatment.
This PBPD model was further linked with Monte Carlo
simulation  to provide predictability of the acute lethality
of CC14 alone and in combination with kepone. As shown
in Table  23.2, the a priori predictions of lethality were in
very good  agreement with experimentally derived values
except at very high CCU levels. In this  latter case,  the
underprediction of lethality was due to toxicity other than
in the liver, that is, the neurotoxic effects of CCU on  the
central nervous  system.  When this  study was first pre-
sented at the International Congress of Toxicology Sym-
posium in  1995, a reporter for Food and Chemical News
wrote a  section  titled "Colorado Researchers Use Elec-
tronic Rats." Although it was somewhat  amusing at  the
time, the term "electronic rats" nevertheless reflects our
ultimate  goal of in silico toxicology.
Clonal Growth in Relation to Carcinogenesis

The  U.S. National Toxicology Program (NTP)  and its
predecessor, the National Cancer Institute's Carcinogen-
esis  Bioassay Program,  collectively, form the  world's
largest toxicology program (NTP, 2002). In its nearly 40
years of operation, fewer than  600 chemicals have been
studied  for carcinogenicity and other chronic toxicities
                                                     A-9

-------
                                                                              APPLICATION OF PBPK/PD MODELING
                                                                                                                   399
       Table 23.2. PBPK/PD Modeling and Monte Carlo
       Simulation of Kepone/CCLi Toxicologic Interaction
        DEN+corn oil
>Q6
• Q7
Dose Observed**
Kepone %
(ppm)
Dead
0
0.0
0
11.1
0
44.4
0
88.8
10
0.0
10
44.4
10
88.8

CC14
dxL/kg)
100

1000

3000

6000

10

50

100

Model

Dead
Rats
0

1-2

3

4-5

0

4-5

8-9

Prediction0

Percent
Dead
0.0

13.2

32.8

47.8

0.0

47.5

84.0



Dead*
Rats
0

1

4

8

0

4

8

                                                                        L
               I
       "Mortalities in 48 h, given a hypothetical condition of n = 9; Monte
       Carlo simulation, n = 1000.
       * Actual lethality studies (n = 9).
         rce: Modified after Yang et al., Toxicol. Lett., 82/82:497-504, 1995.
       (NTP, 2002, 2003). These "gold standard" chronic tox-
       icity/carcinogenicity studies are extremely expensive (i.e.,
       up to several million dollars per chemical), require large
       number of animals (i.e., about 2000 animals per chemical),
       and are lengthy (i.e., 5-12 years per chemical). Thus, con-
       sidering the approximately 80,000 chemicals in commerce
       (NTP, 2002), the number of compounds  for which we
       have adequate toxicology information for risk assessment
       so far is miniscule. With the mode and  rate of study-
       ing these chemicals as indicated above, it is unlikely that
       our society will ever have thorough toxicology informa-
       tion on the majority of the chemicals that we  use now or
       may use in the future. Considering further the "real-world"
       issue of the health  effects of chemical mixtures, it would
       be impossible to obtain adequate information  on most of
       the chemicals or chemical mixtures that humans might be
       exposed to using the conventional approach (Yang,  1994,
       1997). Thus, the PBPK/PD modeling approach described
       below represents a possible alternative  to the expensive
       and time-consuming cancer bioassays.
          We have used  a modification of the medium-term
       bioassay of Ito and co-workers (1989a,b) to study the car-
       cinogenicity of chlorobenzenes.  The Ito assay involves
       the sequential administration of a potent initiator, diethyl-
       nitrosamine (DEN), followed by chemical treatment and
       mitogenic stimulation of hepatocyte growth  via  partial
       hepatectomy. As shown  in  Figure 23.4, this  protocol
       allows  the  evaluation  of carcinogenic  potential  within
                     ZWTA          A^
        DEN + HCB, PECB, TECB, or DCB
1
1

(Days)1
I
J3L
v^x

|*'§x;six;isx;six;;«.
S S S
14 23 2628
I III
DEN fciitf&ssi
A
PH
— *y—A
s s
47 56
I I
HCB, PECB, TECB or DCB

Sacrifice
Figure 23.4.  Experimental design  for  initiation-promotion
study  and estimated parameters for clonal growth model. The
initiation agent,  DEN, was administered  i.p. (200 mg/kg)  at
week 0. HCB, PECB, TECB, or DCB was delivered by gavage
starting week 2 at dose of 0.1 mmol/kg per day, 7 days/week. A
two-thirds partial hepatectomy was performed on all animals on
week 3. Liver tissues were collected 23, 26, 28, 47, and 56 days
following initial DEN dosing. Concurrent controls without DEN
initiation were also performed for all groups (Ou et al., 2003).
eight weeks by identification of glutathione S-transferase
placental form (GST-P) positive preneoplastic foci as end-
point marker lesions. A large number of chemicals have
been tested using this protocol. When compared with the
two-year chronic bioassay, results from the Ito medium-
term bioassay have correctly identified 97% of genotoxic
hepatocarcinogens and 86% of the known nongenotoxic
hepatococarcinogens (Ogiso et al.,  1990). The principal
modification of this protocol in our laboratory  is the incor-
poration of pharmacokinetics and pharmacodynamics by
conducting time course studies on tissue dosimetry, cell
division rates, cell death rates, and GST-P foci formation.
   To collect experimental data, briefly, as shown in
Figure 23.4,  male  Fisher  344  rats, 8  weeks  of age,
were  initiated   with a  single  dose  (200 mg/kg i.p.)
of diethylnitrosamine. Two  weeks  later, animals  were
exposed to daily gavage consisting of 0.1 mmol/kg  1,4-
dichlorobenzene (DCB),       1,2,4,5-tetrachlorobenzene
(TECB), pentachlorobenzene (PECB), or hexachloroben-
zene (HCB) in  corn oil vehicle for 6 weeks.  Partial hep-
atectomy was performed 3  weeks after initiation. Liver
weight,  5-bromo-2'-deoxyuridine labeling index for analy-
sis of cell division rate, and number and volume of GST-P
positive foci were measured at 23, 26, 28, 47, or 56 days
after initiation*.
   We then  used a clonal  growth stochastic model (Ou
et al., 2001, 2003;  Thomas et al., 2000) to  describe the
dynamic growth of  preneoplastic foci  during the  Ito
medium-term bioassay. The clonal growth model is based
              >Q8
                                                           A-10

-------
400
       PHYSIOLOGICALLY BASED PHARMACOKINETIC AND PHARMACODYNAMIC MODELING
    Portion of two-stage model applicable
             to GST-P data
Figure 23.5. Simple two-stage model of carcinogenesis. The
analysis  presented here focuses  on  normal  and initiated
states (Ou et al., 2003).
on the two-stage model  of carcinogenesis (Moolgavkar
and  Luebeck, 1990;  Moolgavkar and  Venzon, 2000),
where  the  carcinogenesis process is  described by the
two  critical rate-limiting steps: (a) from normal to ini-
tiated cells  and (2) from initiated cells to malignant states
(Fig. 23.5). The  model allows the incorporation of rele-
vant biologic information such as the kinetics of tissue
growth and differentiation and mutation rates. The clonal
growth stochastic model adopts a discrete-time numerical
approach, where  the time  axis is decomposed  into a series
of time intervals, where parameters are allowed to change
between  but not with segments. To represent the multi-
plicity of the cellular states and the time-varying nature of
the numerous cell behavior variables, the numerical model
resorts to a recursive  simulation.  The  growth of normal
liver is described deterministically, whereas other cellular
events use stochastic simulation. This approach facilitates
description of the complex biologic process with time-
dependent values.
  A combination of experimental toxicology and com-
puter simulation  described above was used to analyze
clonal  growth  of GST-P enzyme-altered foci during
liver carcinogenesis in an initiation-promotion  regimen
for  DCB, TECB, PECB,  and HCB. The  clonal growth
stochastic model (Ou  et al., 2001, 2003; Thomas et al.,
2000) incorporating the hypothesis of two initiated cell
populations (referred  to  as  A and  B cells) (Fig. 23.5)
was  successfully used to describe the foci development
data for  four chlorobenzenes (Fig. 23.6  presented as an
example  for HCB and PECB). The B cells are  initiated
cells that display selective growth advantage under condi-
tions that inhibit the growth of initiated A cells and normal
hepatocytes. A sensitivity analysis of  model parameters
indicated that the net growth rate of B cells during the two-
week regenerative period following partial hepatectomy
is among the most sensitive parameters for  determining
the final foci  volume. Furthermore, the estimated  val-
ues of this model parameter among four chlorobenzenes
appear to  be positively correlated with the induction of
CYP2B1/2, CYP1A2, c-fos, enlarged liver, and final foci
volume, indicating that  examining effects of chemicals
on regenerative responses following partial hepatectomy
may be a way  to understand the carcinogenicity potential
of chlorobenzene compounds. While TECB, PECB, and
HCB all increased significant foci volume, only HCB had
effects on normal hepatocyte proliferation. The use  of a
two-cell hypothesis for the description  of DEN  control
data (with partial hepatectomy) also indicated the pres-
ence of multiple phenotypes of initiated clones following
DEN treatment, with resistant phenotypes arising during
early carcinogenesis.
   As initiation-promotion protocols are widely used in
the study of carcinogenesis, the clonal simulation of foci
growth, in combination with PBPK modeling, will be a
useful quantitative tool for examining the time course of
a dose-concentration relationship at critical target tissues
and  concentration-dependent pharmacodynamic  changes
at cellular levels during carcinogenesis.
HOW ARE PBPK/PD MODELS VALIDATED?
HOW CAN THEY BE USEFUL PREDICTIVE
TOOLS?

To develop a PBPK/PD model, quantitative time course
experimental data sets are essential for comparisons with
computer simulations. As indicated earlier, the goal and
the hope are that computer simulations are superimposable
with experimental data.  The data sets used  for model
building are working sets (or training  sets). For model
validation, it is very important that  data sets  other than
the working sets  are used to compare with  the model
simulation  results. For  instance,  assume* a  PBPK/PD
model was developed using PK and PD data sets  from
intravenous dosing of two dose levels of drug A reported
by laboratory X. The validation of this  PBPK/PD model
for drug A should be carried out using different data sets
such as PK and PD following oral and dermal dosing
of drug A reported,  respectively,  by laboratories Y and
Z. Generally,  the more different data  sets used in the
validation process, the more robust the PBPK/PD model.
   Once a PBPK/PD model is validated, it may be used
for predictive purposes. However, the present state-of-the-
art is such that the predictive capability is limited to the
very compound for which the PBPK/PD model is devel-
oped, or at the very least  to  close analogues with the
same mechanism of action. For example, Kanamitsu  et al.
(2000), using a PBPK model, predicted in vivo drug-drug
interaction between triazolam  and erythromycin based on
in vitro  enzyme kinetic studies using human liver micro-
somes and recombinant  human cytochrome P450  3A4
>Q9
                                                   A-11

-------
                   FUTURE PERSPECTIVES: IN SILICO TOXICOLOGY AND SECOND-GENERATION PBPK/PD MODELING
                                                                                                            401
                   9000.
         • Q1
                   7000 -
                   5000-
                £ 3000 -


                   1000 •
                     1.4 .

                     1.2 -
                                          DEN _
            PH
             I   I   I    I   I   I
                                                            DEN+PECB
                                    PH
                                      n   i    i   i   r
                             PH
                                                                    DEN+HCB
                            ~rf-i—i—i—i—i—T '  KI\  i—i—i—i—i—r  ir  i—i—i—i—r
                         0  2Q 40 60  80 100120 0  20 40  60  80 100120 0  2ffi 40  60 80 100120
                                                    PH
                                                                           PH
                                                        Time (days)
Figure 23.6.  Comparison of the clonal model outputs with experimental measurements of foci
growth for PECB and HCB congeners. Time-dependent changes in foci growth were measured
in animals subjected to an initiation/promotion protocol using DEN as an initiator and PECB or
HCB as a promoting agent. Using the standard stereologic methods, two-dimensional transection
data of GST-P foci were converted to three-dimensional foci number (foci number/cm3) and foci
volume (volume percentage of foci) in the liver. To illustrate the stochastic nature of foci growth,
the figures here show results from five runs of simulation. For comparison, experimental data
of Jang et al. (1993)* (triangle symbols) and simulation results for the DEN controls without
partial hepatectomy (lower solid gray lines) are shown along with those of the DEN controls (Ou
et al., 2003).
(CYP3A4).  The  mechanistic  basis  for this drug-drug
interaction which involved 15 fatalities in Japan in  1993
is  a  mechanism-based inhibition  (or  "suicide  inhibi-
tion," meaning the enzyme  metabolizes  a chemical to
a reactive metabolite which,  in  turn,  irreversibly  inac-
tivates the enzyme) for macrolide  antibiotics,  such as
erythromycin. Once the enzyme is  permanently inacti-
vated, its metabolism of the other coadministered  drug
is  impaired,  leading to a scenario of "drug overdose."
With the present state-of-the-art, Kanamitsu et al. (2000)
should be able to use their PBPK model to predict poten-
tial serious dose-limiting toxicities for combination thera-
pies involving macrolide antibiotics and coadministered
drugs metabolized  by CYP3A4  provided certain  basic
PK parameters  are known for  the drugs  involved as
well as in vitro enzyme kinetic information using human
CYP3A4. When more and more  such mechanistic infor-
mation is available for classes of chemicals, it is likely that
                                          reaction rules may be established for certain molecular
                                          attributes, and reaction network modeling (see the fol-
                                          lowing section on second-generation PBPK/PD model-
                                          ing) may be linked with PBPIC/PD modeling  such that
                                          more generalized  prediction of toxicities might become
                                          a reality.
                                          FUTURE PERSPECTIVES: IN SILICO
                                          TOXICOLOGY AND SECOND-GENERATION
                                          PBPK/PD MODELING

                                          In essence,  in silico toxicology means integrating com-
                                          puter modeling with focused, mechanistic animal exper-
                                          imentation such that experiments  which are impractical
                                          (e.g., too large, too expensive) or impossible (e.g., human
                                          experiments with carcinogens) to perform are conducted
                                          on  a computer.  We believe that utilization  of computer
                                                    A-12

-------
402
       PHYSIOLOGICALLY BASED PHARMACOKINETIC AND PHARMACODYNAMIC MODELING
modeling is essential in the studies of toxicology of chem-
icals/drugs, chemical/drug mixtures, and their risk/safety
assessment. The area of biology, in general, will be well
served by the application of computer technology  as an
alternative research method to conserve  resources and
minimize the killing of laboratory animals. Looking into
the future, we believe that the linkage of PBPK modeling
and "reaction network modeling," described briefly  in the
following two subsections, has the  potential of providing
a computer simulation platform for complex biologic sys-
tems involving chemicals/drugs, chemical/drug mixtures,
and/or multiple stressors. For a more detailed discussion,
the readers are referred to three publications (Liao  et al.,
2002; Yang et al., 2003a,b) from our laboratory.


BUILDING A SECOND-GENERATION PBPK/PD
MODEL

In classical pharmacokinetics and  physiologically  based
pharmacokinetics,  human or animal bodies were  often
described by  a  few compartments.  By second-generation
PBPK/PD modeling, we  refer to integrating PBPK with
reaction network modeling, thus including  many  more
compartments (i.e., a  delumping process). Our thoughts
may best be explained in  this way: If one draws a parallel
between an oil refinery, where application of the reaction
network modeling approach has been very successful,
and a human body, the individual processing units  in the
oil refinery may be considered as equivalent to the vital
organs of the human body. Even though the cell or  organ
may be much more complicated, the complex biochemical
reaction networks  in  each organ may be modeled and
linked much the same way as the modeling of the entire oil
refinery through linkage of the individual processing  units.


REACTION NETWORK MODELING

From  the  perspective  of its  original  application  in
petroleum engineering, a Reaction Network Model is a
tool for predicting the amounts of reactants, intermediates,
and  products as a function of  time for a  series  of
coupled chemical reactions (potentially numbering  in the
tens  of thousands  of reactions  for some systems). It is
usually a mathematical or symbolic formulation  suitable
for solution on the computer. A reaction network model
builder is a tool for the computer generation of a reaction
network model. The model builder can thus be used not
only to solve the  kinetic equations of interest but also
to generate the reaction mechanisms, rate  constants, and
reaction equations themselves.
   Essentially, the model builder works as  follows:

   1.  The  concentrations of the species to be reacted or
      metabolized are  input to the model builder.
   2. For each species in turn, the model builder performs
     a test against each of a set of "reaction rules" to
     determine whether or not the species is susceptible
     to a particular chemical reaction.
   3. If it is not susceptible to any reactions, no further
     action is taken on this species.
   4. If it is susceptible, a transformation of the species
     into one or more product species is performed based
     on  the particular chemical reaction.
   5. Each of  these product species then undergoes the
     same susceptibility tests and a similar transforma-
     tion sequence. This leads to a linking of all reactants
     with intermediates and, ultimately, with final prod-
     ucts. This linking forms the structure of the chemical
     "reaction network."
   6. After  the  reaction  network  is  established,  the
     rate constants  for the reactions are retrieved  or
     are computed.
   7. The coupled differential equations governing the
     reaction kinetics for the network are then formulated
     by  the model builder.
   8. Finally, the kinetic  equations, that is, the  model
     equations,  are  solved numerically,  leading  to the
     concentrations of all species as a  function of time.

More details on  reaction network modeling, particularly
the initial application to biomedical research, are avail-
able in a  number  of  recent  publications (Klein et al.,
2002; Liao et al., 2002; Reisfeld and Yang, 2003; Yang
et al., 2003a,b).
CONCLUSION

It is fitting to conclude with some recent testimonials from
reputable scientists for the importance of the integration
of different fields and the central role computer modeling
will play in  biology:

   In  an AAAS Plenary Lecture on  February 13, 1998,
     Dr. Harold  Varmus,  then Director of the National
     Institutes  of Health, emphasized,  among others,
     two specific themes: "Discoveries in biology and
     medicine  depend on progress  in many  fields  of
     science" and "Methods  that dramatically expand
     biological data  also demand new modes of analysis
     and new ways to ask scientific questions." He said,
     "In short, biology is  not only for biologists."*
   Craig Venter of the human genome fame stated, "If we
     hope to understand  biology,  instead of looking at
     one little protein at a time, which is not how biology
     works, we will need to understand the integration
     of thousands of proteins in a dynamically  changing
>Q10
                                                    A-13

-------
                                                                                                    REFERENCES
                                                                                                                    403
            environment. A computer will be  the  biologist's
, cm         number one tool."* (Bulter, 1999).
         Tyson and colleagues (2001) indicated, "Many promi-
            nent molecular biologists have pointed out the press-
            ing need for theoretical and computational tools to
            show the spatial and temporal organization implicit
            in the way the macromolecules are 'wired together'
, Q12         to create a living cell."*

         PBPK models have proven useful in uncovering deter-
       minants of disposition  of carcinogens and other  com-
       pounds in the body; PBPD cancer models have shown the
       role of mutation and cell proliferation in the time course of
       tumor development. Both of these types of models initially
       lacked considerable biologic  detail  due to limitations of
       our knowledge of fundamentals of cell and molecular biol-
       ogy.  The revolution in genomic technologies in the past
       decade has revolutionized the database to support mech-
       anistic modeling of chemical disposition and of biologic
       responses and these technologies now provide  a basis for
       expansion of these models with increasing biologic detail.
       Reaction network models of cell constituents and genetic
       regulatory networks of  cellular  controls appear particu-
       larly attractive candidates  for  approaches  to uncover the
       interactions and perturbations controlling neoplastic trans-
       formations and growth.
         Another contemporary extension of PBPK/PD models
       is closely tied to current attempts to unravel the circuitry
       of living cells and to discover the manner in which  these
       circuits  lead to biologic function and health. Mathemat-
       ical models of gene networks and  their perturbation by
       disease  or by chemical  exposures  are  now being devel-
       oped (Andersen and Barton, 1999). In some cases, simple
       prokaryotic cells with specific circuit elements (e.g., bio-
       logic oscillators,  switches, amplifiers) have been produced
       and examined by laboratory experiments and by compu-
       tation (Guet et al., 2002;  Hasty et al,  2002;  McMillen
       et al., 2002).  These computational models  evaluate the
       protein networks within cells, the genetic control of these
       networks, and the  logic  of cellular responses affected by
       these  networks (Aim and  Arkin, 2003; Davidson et al.,
       2002; Ferrell, 2002).
         Gene  network modeling  in   intact  animals   will
       inevitably draw  on PBPK  and reaction network  models
       to  track concentrations of endogenous  and  exogenous
       signaling compounds and on PBPD models to simulate
       consequences  of the  interactions   of  these compounds
       with signaling pathways within cells. Reverse engineering
       approaches attempt to uncover the circuitry of working
       cells (Csete  and  Doyle, 2002). Large-scale  simulation
       modeling that has formed the major core of PBPK/PD
       models  remains important in examining these  genetic
       regulatory networks. However, these types of models are
       being augmented by Boolean approaches using  on-off
logic to increase the scope of genomic coverage (Bolouri
and  Davidson, 2002). Efforts  in  understanding genetic
networks  may  be  especially  important  in  providing
insights into neoplastic diseases in which cell signaling
networks  become   impaired  or  deranged (Hahn  and
Weinberg, 2002; Hanahan and Weinberg, 2000).
ACKNOWLEDGMENTS

The concepts and work discussed in this presentation  were
  partially contributed by many colleagues associated or col-
  laborating with the Quantitative and Computational Toxicol-
  ogy group at Colorado State University; we are grateful to
  their contribution and intellectual stimulation. Any advances
  in science require funding support from many  agencies. We
  thank NIEHS (Superfund Basic Research Program Project P42
  ES05949; research grants RO1 ES09655 and RO3 ES10116
  ZES1; training grant T32 ES 07321;  and Career Develop-
  ment Award K25  ESI 1146), ATSDR (Cooperative Agree-
  ment U61/ATU881475), and U.S. Air Force (research grants
  F33615-91-C-0538; F49620-94-1-0304). Without the gener-
  ous support of these agencies, the  development of research
  described herein could have never been possible.
REFERENCES

Aim E, Arkin AP (2003): Biological networks. Curr Opinion
  Struct Biol 13:193-202.
Andersen ME,  Barton HA (1999):  Biological regulation  of
  receptor-hormone complex concentrations in relation to dose
  response assessments for endocrine active compounds. Toxicol
  Sci 48:38-50.
Andersen ME,  Clewell HJ, Gargas ML,  Smith FA,  Reitz RH
  (1987): Physiologically-based pharmacokinetics and the risk
  assessment for methylene chloride. Toxicol Appl Pharmacol
  87:185-205.
Bischoff KB, Brown RG (1966): Drug distribution in mammals.
  Chem Eng Prog Symp Series 62:33-45.
Bolouri H, Davidson EH (2002): Modeling transcriptional reg-
  ulatory networks, BioEssays 24:1118-1129.
Brown RP, Delp MD, Lindstedt SL, Rhomberg LR, Beliles RP
  (1997): Physiological parameter values for physiologically
  based  pharmacokinetic  models.  Toxicol Ind Health  13:
  407-484.
Butler D (1999): Computing 2010 from black holes to biology.
  Nature 402:C67-C70.
Csete ME, Doyle JC (2002): Reverse engineering of  biological
  complexity. Science 295:1664-1669.
Curtis LR, Williams WL, Mehendale HM (1979): Potentiation
  of hepatotoxicity of carbon tetrachloride following preexpo-
  sure to chlordecone (Kepone)  in the male rat. Toxicol Appl
  Pharmacol 51:283-293.
Davidson EH, Rast JP, Oliver! P, Ransick A, Calestani C, Yuh
  CH,  Minokawa T, Amore G,  Hinman V,  Arenas-Mena C,
  Otim O, Brown CT, Livi CB,  Lee PY, Revilla R,  Rust AG,
                                                           A-14

-------
404
       PHYSIOLOGICALLY BASED PHARMACOKINETIC AND PHARMACODYNAMIC MODELING
, Q13
  Pan Z, Schilstra MJ, Clarke PJ, Arnone MI, Rowen L, Cam-
  eron RA, McClay DR, Hood L, Bolouri H (2002): A genomic
  regulatory  network for development. Science 295:1669-
  1678.
Dedrick RL (1973): Animal scale-up. J Pharmacok Biopharm
  1:435-461.
El-Masri HA,  Thomas RS,   Sabados GR,  Phillips JK,  Con-
  stan AA, Benjamin  SA, Andersen ME, Mehendale HM, Yang
  RSH (1996): Physiologically based pharmacokinetic/pharma-
  codynamic  modeling of the toxicologic interaction between
  carbon tetrachloride and Kepone. Arch Toxicol 70:704-713.
Ferrell  JE (2002): Self -perpetuating states in signal transduction:
  Positive feedback,  double-negative feedback and bistability.
  Curr Opin Chem Biol 6:140-148.
Guet CC, Elowitz MB, Hsing W, Liebler S (2002): Combinato-
  rial synthesis of genetic networks. Science 296:1466-1470.
Hahn WC,  Weinberg RA (2002): Modelling the molecular
  circuitry of cancer.  Nature Rev Cancer 2:331-341.
Hanahan D, Weinberg RA (2000): The hallmarks of cancer. Cell
  100:57-70.
Hasty J, McMillen D, Collins JJ (2002): Engineered gene cir-
  cuits. Nature 420:224-230.
Ito N, Imaida  K, Hasegawa R, Tsuda H (1989a): Rapid bioassay
  methods for carcinogens and modifiers of hepatocarcinogen-
  esis.  CritRev Toxicol 19:385-415.
Ito N, Tatematsu M, Hasegawa R, Tsuda H (1989b): Medium-
  term bioassay system for detection of carcinogens  and mod-
  ifiers  of  hepatocarcinogenesis utilizing the GST-P positive
  liver  cell focus as  an endpoint marker.  Toxicol Pathol
  17:630-641.
Jang et al. (1993): Cancer Lett 71:89-95. •
Kanamitsu S,  Ito K,  Green CE,  Tyson AA, Shimada M, Su-
  giyama Y (2000): Prediction of in vivo interaction between
  triazolam and erythromycin based  on in vitro  studies using
  human liver microsomes and recombinant human  CYP3A4.
  Pharmaceut Res 17:419-426.
Kedderis GL,  Lipscomb JC   (2001):  Application  of in vitro
  biotransformation data and  pharmacokinetic modeling to risk
  assessment. Toxicol Ind Health 17:315-321.
Klein MT,  Hou G, Quann R, Wei W, Liao KH, Yang RSH,
  Campain JA,  Mazurek M,  Broadbelt LJ  (2002): BioMOL:
  A computer-assisted biological modeling tool for  complex
  chemical mixtures  and biological processes at the molecular
  level. Environ Health Perspect 110(Suppl. 6):1025-1029.
Lazarou J,  Pomeranz BH, Corey PN  (1998):  Incidence of
  adverse drug reactions in  hospitalized patients.  A meta-
  analysis of prospective studies. JAMA 279:1200-1205.
Liao KH, Dobrev I, Dennison Jr,  JE, Andersen ME, Reisfeld B,
  Reardon KF,  Campain JA,  Wei W,  Klein MT, Quann RJ,
  Yang RSH  (2002):  Application of biologically based  com-
  puter modeling to  simple  or complex  mixtures. Environ
  Health Perspect 110(Suppl. 6):957-963.
Lindstedt SL  (1987):  Allometry:  Body size constraints in
  animal  design, In  Pharmacokinetics  in  Risk  Assessment.
  Drinking Water and Health, Vol. 8, National Academy Press,
  Washington, DC, pp 65-79.
Lipscomb  JC, Fisher JW, Confer PD, Byczkowski JZ (1998):
  In vitro  to  in vivo extrapolation for trichloroethylene meta-
  bolism in humans. Toxicol Appl Pharmacol 152:376-387.
McMillen  D, Kopell N, Hasty J, Collins JJ (2002):  Synchroniz-
  ing genetic relaxation oscillators by intercell signaling. Proc
  NatAcadSci 99:679-684.
Mehendale HM (1984): Potentiation of halomethane hepatotox-
  icity:  Chlordecone and carbon tetrachloride. Fundam Appl
  Toxicol 4:295-308.
Mehendale HM (1991): Role of hepatocellular regeneration and
  hepatolobular* healing in the final outcome of  liver injury.
  Biochem Pharmacol 42:1155-1162.
Mehendale HM (1994): Mechanism of the interactive amplifi-
  cation of halomethane hepatotoxicity  and  lethality  by  other
  chemicals. In Yang RSH (Ed), Toxicology  of Chemical Mix-
  tures: Case Studies, Mechanisms, and Novel Approaches, Aca-
  demic, San Diego, pp 299-334.
Moolgavkar SH,  Luebeck G (1990): Two-event model for car-
  cinogenesis: Biological, mathematical, and statistical consid-
  erations. Risk Anal 10:323-341.
Moolgavkar SH,  Venzon DJ (2000): Two-event  model for
  carcinogenesis. Math Biosci 47:55-77.
Morrison PF, Dedrick RL, Lutz RJ (1987): Methotrexate: Phar-
  macokinetics and Assessment of Toxicity, In Pharmacoki-
  netics in Risk Assessment, Drinking Water and Health,  Vol.
  8, National Academy  Press,  Washington,  DC, pp 410-
  427.
National Toxicology Program (NTP) (2002):  National Toxicol-
  ogy Program:  Annual Plan  for Fiscal Year 2001.  National
  Toxicology Program, U.S. Department of Health and Human
  Services, Washington, DC.
National Toxicology Program (NTP) (2003):  National Toxicol-
  ogy Program,  Management Status Report, U.S.  Department
  of Health and Human Services, Washington,  DC, April 7.
Ogiso T, Tatematsu  M, Tamano S, Hasegawa R, Ito N (1990):
  Correlation between  medium-term liver bioassay system data
  and results  of long-term testing in  rats.  Carcinogenesis
  11:561-566.
Ou  YC,  Connolly RB, Thomas RS, Gustafson DL, Long ME,
  Dobrev  I, Chubb L,  Xu Y, Lapidot SA, Andersen ME,  Yang
  RSH (2003): Stochastic simulations of liver foci development
  in a medium-term bioassay for four chlorobenzene congeners.
  Toxicol Sci  73:301-314.
Ou  YC,  Connolly RB,  Thomas RS,  Xu Y,  Andersen ME,
  Chubb L,  Pitot HC, Yang RSH  (2001):  A clonal growth
  model: Time-course simulations of liver foci growth following
  penta- or hexa-chlorobenzene  treatment in  a medium-term
  bioassay. Cancer Res 61:1879-1889.
                                                                                                                                >Q14
                                                       A-15

-------
                                                                                                         REFERENCES
                                                                                                                         405
       Ramsey JC, Andersen ME (1984): A physiological model  for
         the inhalation pharmacokinetics of inhaled styrene monomer
         in rats and humans. Toxicol Appl Pharmacol 73:159-175.
       Reddy M,  Yang RSH, Clewell IE HJ, Andersen ME  (2004):
         Physiologically Based Pharmacokinetics: Science and Appli-
> Q15     cations.  Wiley, Hoboken, NJ.*
       Reisfeld B, Yang RSH (2003):  A reaction network model  for
         CYP2E1-mediated metabolism of toxicant mixtures. Environ
>Q16     Toxicol Pharmacol*
> 017   Thomas RS, Conolly* RB, Gustafson DL, Long ME, Benjamin
         SA, Yang RSH  (2000):  A physiologically-based pharmaco-
         dynamic analysis of hepatic foci within a medium-term liver
         foci bioassay using pentachlorobenzene as a promoter and
         diethylnitrosamine as an initiator. Toxicol Appl Pharmacol
         166:128-137.
       Tyson JJ, Chen K, Novak B (2001): Network dynamics and cell
         physiology. Nature Rev/Mol Cell Biol 2:908-916.
       U.S. Environmental Protection Agency (EPA) (2002): Guidance
         on Cumulative Risk Assessment of Pesticide Chemicals That
         Have  a  Common Mechanism of Toxicity.  Office of Pesti-
         cide Programs, U.S. Environmental Protection Agency, Wash-
         ington,  DC 20460. (Online  at www.epa.gov/oppfeadl/trac/
         science/cumulative_guidance.pdf.)
       van  de  Waterbeemd H,  Gifford E (2003):  ADMET   in sil-
         ico modeling:  Towards prediction  paradise. Nature  Rev
         2:192-204.
Yang RSH  (1994): Introduction to  the  toxicology of chemi-
  cal  mixtures. In Yang, RSH (Ed),  Toxicology of Chemical
  Mixtures: Case Studies, Mechanisms, and Novel Approaches.
  Academic, San Diego, pp 1-10.
Yang RSH (1997): Toxicologic interactions of chemical mix-
  tures. In  Bond J  (Ed),  Comprehensive  Toxicology. Vol.
  1, General Principles, Toxicokinetics, and Mechanisms of
  Toxicity.  Elsevier  Science,  Oxford,  England,  pp   189-
  203.
Yang RSH, Andersen ME (1994): Pharmacokinetics. In  Hodg-
  son E, Levi P (Eds), Introduction to Biochemical Toxicology,
  2nd ed., Elsevier, pp 49-73. •
Yang RSH, Andersen ME,  Clewell  III HJ  (2004):  Physio-
  logically Based Pharmacokinetic/Pharmaco dynamic Model-
  ing:  Principles,  Methods, and Applications. Academic, San
  Diego.*
Yang RSH, El-Masri HA, Thomas RS,  Dobrev I,  Dennison
  Jr JE, Bae DS,  Campain JA, Liao KH,  Reisfeld B, Ander-
  sen ME,  Mumtaz MM (2003a): Chemical mixture toxicol-
  ogy: From  descriptive to mechanistic, and going on to in
  silica toxicology. Environ Toxicol Pharmacol Submitted for
  publication.*
Yang RSH, Liao KH, Reisfeld  B  (2003b): The integration of
  computer modeling and experimental toxicology for the study
  of chemical  mixtures  and multiple  stressors. Arch  Complex
  Environ Factors In press.*
Q19
020
                                                                                                                                 021
                                                              A-16

-------
                             APPENDIX B
    A REFERENCE ON HOW TO INCORPORATE CREDIBLE HUMAN ENZYME
  STUDIES INTO THE PHYSIOLOGICALLY-BASED PHARMACOKINETIC (PBPK)
              MODELING AND RISK ASSESSMENT PROCESS

Lipscomb, J.C., L.K. Teuschler, J. Swartout, D. Popken, T. Cox and G.L. Kedderis.
2003. The impact of cytochrome P450 2E1-dependent metabolic variance on a risk-
relevant pharmacokinetic outcome in humans.  Risk. Anal. 23(6): 1221-1238.
                                 B-1

-------
Risk Analysis, Vol. 23, No. 6, 2003
The Impact of Cytochrome P450 211-Dependent Metabolic
Variance on a Risk-Relevant Pharmacokinetic Outcome
in H umans
John C. Lipscomb,1* Linda K. Teuschler,1 Jeff Swartout,1 Doug Popken,2
Tony Cox,2 and Gregory L. Kedderi§3
                         Risk assessments include assumptions about sensitive subpopulations, such as the fraction of
                         the general population that is sensitive and the extent that biochemical or physiological at-
                         tributes influence sensitivity. Uncertainty factors (UF) account for both phartnacokinetic (PK)
                         and pharmacodynamic (PD) components, allowing the inclusion of risk-relevant information
                         to replace default assumptions about PK and PD variance (uncertainly).  Large numbers of
                         human organ donor samples and recent advances in methods to extrapolate in vitro enzyme
                         expression and activity data to the intact human enable the investigation of the impact of PK
                         variability on human susceptibility. The hepatotoxicity of trichloroethylcne (TCE) is mediated
                         by acid metabolites formed by cytochrome P450 2E1 (CYP2E1) oxidation, and differences in
                         the CYP2E1 expression are hypothesized to affect susceptibility to TCE's liver injury, This
                         study was designed specifically to examine the contribution of statistically quantified variance
                         in enzyme content and activity on the risk of hepatotoxic injury among adult humans. We com-
                         bined data sets describing (1) the microsomal protein content of human liver, (2) the CYP2E1
                         content of human liver microsomal protein, and (3) the in vitro Vnmx for TCE oxidation by
                         humans. The 5th and 95th percentiles of the resulting distribution (TCE oxidized per minute
                         per gram liver) differed by approximately sixfold. These values were converted to mg TCE
                         oxidized/h/kg body mass and incorporated in a human PBPK model. Simulations of 8-hour
                         inhalation exposure to 50 ppm and oral exposure to 5 g TCE/L in 2 L drinking water showed
                         that the amount of TCE oxidized in the liver differs by 2% or less under  extreme values of
                         CYP2E1 expression and activity (here, selected as the 5th and 95th percentiles of the resulting
                         distribution). This indicates that differences in enzyme expression and TCE oxidation among
                         the central 90% of the adult human population account for approximately 2% of the difference
                         in production of the risk-relevant PK outcome for TCE-mediated liver injury. Integration of
                         in vitro metabolism information into physiological models may reduce the uncertainties asso-
                         ciated with risk contributions of differences in enzyme expression and the  UF that represent
                         PK variabilitv.
                         KEY WORDS; Cytochrome P450 2E1 (CYP2E1); interindividual differences: physiologically-based phar-
                         macokinetic modeling; risk assessment; triehtoroethylene bioaclivation; uncertainty factors
 i US. EPA, ORD, NCEA. Cincinnati, OH, USA.
 2 Cox Associates.Denver, CO, USA.                          '"
 -i Independent Consultant, Chapel Hill, NC. USA.                    ~  ....    .           .  ,               ,  TT_
 • Address correspondence 10 John C. Lipscomb, PhD, DABT, U.S.         Traditional noncancer risk assessments in the U.S.
  EPA/ORD/NCEA. 26 w. M.L.King Drive, MD-190, Cincinnati,      EPA apply uncertainty  factors  to  extrapolate the
  OH 45268, USA; Lipscomb.John@EPA.GOV                   measures of effects between animals and  humans,

                                                1221       0272-4332/(>3/l200~I221$22,0()/l € 2003 Society for Risk Analysis


                                                   B-2

-------
1222
                                  Lipscomb et at.
and  among humans. These two factors (UFA and
UFH, respectively), may be further subdivided into
their respective pharmacodynamic (PD) and phar-
macokinetic (PK) components/ '"3) WHO(4) and the
International  Programme  on Chemical Safety^5'6)
have provided guidance and application of the sep-
arate consideration of PD and PK, and the U.S. EPA
has also separately quantified PD and PK variability
in the UFA applied to reference concentration (RfC)
values and reference dose (RfD) values, where each
has been ascribed a default value of one-half log (10" ~\
or 3.16),(7"9) In addition, PD and PK components of
UFH for RfC values have also been separately con-
sidered for some substances such as methyl rnethacry-
late/10' Studies with humans can be conducted to as-
sess  the pharmacodynamics or pharmacokinetics of
environmental or occupational chemicals. While hu-
man clinical trials can assess the PK and PD of poten-
tial therapeutic substances, human studies  with po-
tentially toxic environmental or occupational chem-
icals are not usually conducted over concentration
ranges known or predicted to result in adverse effects.
The limited information available from human studies
with environmental chemicals provides critical (but
often limited) information, which can be extended by
in vitro  studies using preparations from human tis-
sues. Care must be taken so that the in vitro inves-
tigations are focused on risk-relevant endpoints, and
are conducted with the relevant tissues, tissue prepa-
rations, and chemical concentrations. It is critical that
the concentrations used for in vitro studies are within
the range of tissue concentrations observed or pre-
dicted in vivo in humans following chemical exposure.
Studies with human subjects or human tissue prepara-
tions in  vitro can identify variability in PK outcomes
such as the blood concentrations of parent chemical
and  metabolites or the rates of metabolite  produc-
tion  or elimination. When these PK outcomes over-
lap with the PK outcomes most linked to risk (verified
by results from mode of action and PK studies with
research animals), then additional information on the
variability of these PK outcomes will advance our un-
derstanding of susceptibility, and provide information
with which to replace default values for uncertainty
in extrapolations of risk. Although data from multiple
human subjects may seem preferable as the basis from
which to determine human PK variability, those data
seldom exist, and when they do the data usually offer
little information on risk-relevant PK outcomes such
as target tissue dosimetry. Physiologically-based PK
(PBPK) models allow the application of physiologic
 Toxieity and PK Testing
Defined
Mode of
Action
Exposure

Internal Dose

Distribution

Metabolism

Concentration in
Target Tissue
               Response -
 Risk Assessment

  Exposure
     t
  Internal Dose
     t
  Distribution
     t
  MetaboHsm
     t
  Concentration in
  Target Tissue
     t
». Response
               (1 ,(MF.l, NOAKI,.
Fig, 1. Application of PBPK modeling to link external dose with
concentration of toxicant in target organs. The approach builds
upon information on mode of action, which demonstrates the rela-
tionship between tissue, response, a PD phenomenon, and the PK
outcome directly related to that response, PBPK models are then
developed and employed to define the relationship between the ex-
ternal exposure and the target organ toxicant concentration, usu-
ally pet • orrncd in test animal species. Once completed, and based
on assumptions about the similarity in the qualitative and quanti-
tative nature of the PD effect (mode of action) between the lest
species and humans, parallel PBPK models are developed for the
human, and are exercised to "back -track" the toxicant from a pre-
determined concentration in the target tissue to the corresponding
exposure concentration (external dose),
and anatomic constraints to clarify the linkage be-
tween external concentrations and target tissue con-
centrations (Fig. 1) and offer a mechanism through
which information obtained ex vivo or in vitro may be
evaluated in the proper context. Results from PBPK
model simulations of relevant exposure scenarios pro-
vide a useful approach for estimation of PK variabil-
ity between research animals and humans and among
humans when other data are limiting. This technique
offers the opportunity to extrapolate concentrations
of bioactive chemical moieties in target tissues across
species, doses, and routes of exposure. The inclusion of
data derived in vitro through the exposure of human
tissue preparations offers an advance over exposing
humans to noxious agents, and several studies  have
demonstrated the applicability of in vitro findings in
refining PBPK models.
    While in vitro measurements of specific biochem-
ical reactions from multiple human samples can  yield
qualitatively valuable data on human variance, they
must  be tied to human anatomy and physiology, and
the impact of their variance evaluated under real ex-
posure scenarios, to be  of  quantitative  value.  This
                                               B-3

-------
Tridiloroethylene and Human CYP2E1
                                                                                               1223
study was constructed on the framework for extrap-
olation of in vitro metabolic rate information and
PBPK model incorporation previously suggested.^11'
    Enzymes are protein  molecules that catalyze
chemical reactions.^12^ Over 100 years ago, the study
of enzymes and their properties demonstrated that
the rates of enzyme-catalyzed reactions are directly
proportional to the total enzyme present in the sys-
temJ12-13) This property of enzymes provides the basis
for extrapolation of in vitro biotransformation data to
whole animals and humans.114' Therefore, data gener-
ated with subeellular fractions such as microsomes or
cytosols can  be extrapolated to in vivo based on pro-
tein content.(l5) Human liver is approximately 2.6%
of body weight,f16J
     In addition to enzyme content or activity and or-
gan weight, the kinetic mechanism of the enzyme (the
comings and goings of substrates and products) needs
to be taken into account to extrapolate in vitro data to
whole animals or humans/14' The CYP2El-catalyzed
oxidation of TCE follows Michaelis-Menten satura-
tion kinetics:^17-'

          v - (Vmax * [S})/(KM + [S])        (I)

where v is the initial velocity of the reaction, Vmax is
the maximal rate of the reaction at infinite substrate
concentration, [S] is the substrate concentration, and
KM  is the Michaelis  constant for the reaction. The
Michaelis-Menten (Equation  (1)) indicates  that the
initial velocity of the reaction will increase hyperbol-
ically as  a function of substrate concentration. The
Vmax is a horizontal  tangent to the top (saturated)
part of the curve,  while the tangent to the initial lin-
ear portion of the hyperbolic curve is the initial rate
of the reaction, V/K. The V/K is  the pseudo-flrst-
order rate constant for the reaction at low substrate
concentrations. The point where these two tangents
intersect corresponds to the KM-^ The  KM is de-
fined as the  substrate concentration that gives  one-
half the Vmax. The KM for each substrate is an inherent
property of the enzyme/12' A lower KM for one sub-
strate compared to a second substrate indicates that
the first substrate has a more rapid initial rate (V/K)
of metabolism. The value of KM can vary with the
structure of the enzyme; for example, in the polymor-
phism of the CYP2D6-mediated oxidation of debriso-
quine and related drugs/19) Experimentally, KM can
vary with pH, temperature, and ionic strength in vitro.
Therefore, in vitro kinetic measurements intended for
extrapolation to intact animals and humans should be
done under experimental conditions mirroring the in
vivo situation as closely as possible/14*
    The in vitro kinetic data can be incorporated into
PBPK models after rearrangement of the Vmax to the
appropriate units. Values of KM have units of con-
centration and can be used directly if the solubility of
the chemical in the in vitro system is known.  Incor-
poration of the extrapolated kinetic parameters into
a PBPK model for humans allows prediction  of tar-
get tissue dosimetry following a variety of exposure
scenarios.
    We have adapted an existing PBPK model to pre-
dict the difference among humans in the risk-relevant
PK outcome for the hepatotoxicity of trichloroethy-
lene (TCE) under  conditions of human variance in
the rates of TCE oxidation. We focused on the hep-
atotoxicity of TCE because: (1) the PK of TCE have
been characterized and modeled in research animals
and humans (reviewed in Reference (20); (2) more
than 95% of an absorbed dose of TCE is oxidized
in research animals and humans/21' (3) CYP2E1 has
been demonstrated to be the enzyme responsible for
the oxidation of TCE in research animals and humans
and in vitro preparations at low concentrations;*'22' (4)
the hepatotoxicity of TCE has been demonstrated to
depend on acid metabolite(s) derived from oxidative
metabolites of TCE/23-24} (5) the CYP2E1-mediated
oxidation  of TCE is rate limiting in the further for-
mation of acid  metabolite(s)/2i) (6) the expression of
CYP2E1 is modulated by genetic, environmental, and
lifestyle factors; and (7) large numbers of human liver
tissue samples and human liver tissue preparations
are currently available in contrast with preparations
and tissues from other human organs. The  results on
the variance of the distribution of CYP2E1 in adult
human liver will be especially applicable to other en-
vironmental contaminants that are also substrates for
this enzyme.
    The investigation was accomplished by first char-
acterizing the variance about the CYP2E'l-mediated
oxidation  of TCE  among human samples in vitro,
second by quantifying  the variance of human hep-
atic CYP2E1 content, and third by extrapolating the
bounds of variance  of TCE oxidation among the adult
human population to a human PBPK model. Two sep-
arate statistical analyses were conducted—one based
on convenience and one based on  technical accu-
racy. The amount of TCE metabolized (oxidized) in
the liver was simulated as a  dose surrogate for the
hepatotoxicity of TCE. The goal of the present in-
vestigation was to  quantify the variability in  a risk-
relevant PK outcome for the hepatotoxicity of TCE,
and to demonstrate the usefulness of advanced data
on human biochemical individuality in quantifying the
                                                  B-4

-------
1224
                                 Lipscomh ct al.
variability of risk-relevant PK outcomes  for inclu-
sion in risk assessments. Data on the distribution of
CYP2E1 in the intact liver has not been used to esti-
mate the degree of susceptibility to risk for metabo-
lized chemicals. We hypothesized that the degree of
natural variance in human hepatic levels of CYP2E1
would result in similar differences in the oxidation of
TCE in the intact human.
2. METHODS
    Several sets of information describing or based
on microsomal protein (MSP) were collected for as-
similation, extrapolation, and  incorporation into a
PBPK model. The objective of the extrapolation was
to transition expression of apparent Vmax from units of
"pmoles TCE oxidized/min/mg MSP" to units of con-
ventional PBPK modeling, mg/h/kg body mass. Nec-
essary data were compiled from multiple sources, and
used to describe the various parameters, whose distri-
butions were analyzed and combined. Table I demon-
strates the relationship between those data sets and
parameters. TCE  is oxidized by CYP2E1, and that
metabolic rate had been previously measured and
presented in units of MSP (nmol/min/mg MSP). Thus,
the need to express apparent VmaK as pmol TCE oxi-
dized/min/pmol CYP2EL CYP2E1 is isolated in MSP,
thus the need to quantify CYP2E1 in MSP  (pmoles
CYP2El/mg MSP). Because MSP is one constituent
of liver, the amount of MSP per gram liver tissue (mg
MSP/gram liver) needed computing. These data facil-
itate the extrapolation of in vitro metabolic capacity
(comprising enzyme activity and enzyme content) to
the intact liver. Units of calculation cancel (pmol TCE
oxidizedAnin/pmol CYP2E1) x (pmol CYP2El/mg
MSP) x (mg MSP/gram liver), leaving units of pmol
TCE oxidized/min/gram liver. Correction for molec-
ular weight of TCE, 60 min/h, and assumptions about
the fractional composition of body mass attributed to
the liver compartment (liver = BW x 0.026) results
in units of mg TCE oxidized/h/kg.


2,1. Human Samples and Quantification
    of CYP Proteins

    Both prepared MSP and intact liver tissue were
obtained for this investigation from various sources
(Human Cell Culture Center,  Laurel, MD; Interna-
tional Institute for the Advancement of Medicine, Ex-
ton, PA; Vitron, Tucson, AZ; Tissue Transformation
Technologies, Inc., Edison, NJ). All tissues and prepa-
rations were derived from adult human organ donors
that were devoid of antibodies directed against infec-
tious diseases. The MSP content of CYP2E1 and other
CYP forms was previously investigated and reported
for 40 donors.^15' In the present analysis, 20 samples of
intact tissue were obtained, and MSP prepared via the
method of Guengerich(26> (Fig. 2).  CYP2E1 content
of aliquots of (post 100 x g) homogenate protein and
MSP were determined by enzyme-linked immunosor-
bent assay (ELISA) following the method of Snawder
and
2.2. Distribution of CYP2E1 to Human
    Hepatic MSP
    Data on the CYP2E1 content from 40 samples
of human hepatic MSP were available from Snawder
and  Lipscomb/15' and data derived  from an  ad-
ditional 20 samples of human hepatic MSP were
                      Table I. Identification of Data Sets and Parameters for Statistical Evaluation
Information Data Sel
CYP2E1
content of
intact liver
CYP2E1 con- Data Set 1 (n = 60)
tent of MSP
MSP content of Data Set 2 (n = 20)
intact liver
TCE metaboli- Data Set 3 (« = 1 5)
?.cd per unit
CYP2E1
TCE metaboli-
zed per unit
of intact liver
Units
pmol CYP2El/gram liver


prnolCYP2El/mgMSP

mg MSP/gram liver

pmol TCE/min/ pmol CYP2E1


pmol TCE oxidized/minute/gram
liver

Parameter
A


B

C

D


E


Notes
Directly measured via ELISA (n — 20);
and separately predicted statistically

Directly measured via ELISA

Derived: C = A/B

Original measurements from
Reference 11 corrected by CYP2E1
content from Reference 18
Statistically estimated: E= B x C x D;
extrapolated to Vmax

                                             B-5

-------
Triehloroethylene and Human CYP2E1
                                                                  1225
        Isolation of Microsomal Protein
    Enzyme Activity
    Metabolite per:
    • mg Microsomal Protein,
    *pmolCYP2El
Enzyme Content
pmolesCYPZEl /mg
Microsomal Protein
Fig, 2. Relationship between intact liver, microsomal protein, and
some CYP forms. The isolation of microsomal protein from intact
liver via homogenation of tissue and differential cenlrifugation re-
sults in a 10CMMX) x g pellet, which is enriched for endoplasmic retie-
ulum content. The enrichment results in an artificial increase in the
concentration of biological components associated with the endo-
plasmic reticulum. This isolation produces a fraction (microsomes;
MSP), which is subjected to in vitro investigations of metabolic ac-
tivity and enzyme content. However, a quantitative relationship to
the intact liver is not possible without further information on the
distribution of rnicrosomai protein to the intact liver.

combined to yield a total sample of 60 adult human or-
gan donors for which data on the CYP2E1 content of
MSP (CYP2MSP, parameter B, data set 1) were avail-
able^ (Table II, described in the following section).
Of this set of 60 samples of MSP (representing 60
organ donors), 15 were used to estimate the in vitro
metabolic parameters for TCE and CYP2E1 content
of MSP; 45 were subjected only to the determination
of CYP2E1 in MSP (and 20 of that 45 were  paired
with liver homogenate to determine the MSP content
of liver).

2.3.  Estimation of Proteins in Intact Liver

    In this analysis, 20 samples of intact liver tissue
were assayed (Table II). The total amount of protein
(CYP and non-CYP; microsomal and cytosolic pro-
teins) in intact liver (PROUv) was empirically deter-
mined based on the protein content of the post 100 x g
liver supernatant, after correcting for volume accord-
ing to Equation (2), It was assumed that no protein
was lost during the sedimentation of nuclei and debris
at 100 x g.

    (mg protein/ml homogenate)
      x (ml homogenate/gram tissue)
     = (mg homogenate protein/gram tissue)    (2)
The  content  of CYP2E1  in  total hepatic  protein
(CYP2Pro)  and in MSP (CYP2MSP;  parameter B)
was determined so that a measure of the liver con-
tent  of MSP (MSPi.jV; parameter C) could be de-
rived. The content of CYP2E1  in  liver (pmoles
CYP2El/gram liver; CYP2Liv; parameter A) was de-
rived empirically by combining two data sets (PROyv
CYPpRo). described in. Equation (3), and was esti-
mated via the statistical method of moments (Sec-
tion  2.5,1)  and by computational  statistics (Section
2,5.2). For the 20 individual organ donors, the sep-
arate amounts of CYP2E1  per  gram  liver were
empirically determined according to the following
equation;
                          (pmol CYP2El/mg homogenate protein)
                            x (mg homogenate protein/gram tissue)
                            — (pmol CYP2E1/grain tissue).
                                             (3)
                       The amount of MSP per gram liver was estimated ac-
                       cording to Equation (4). This is the data set (MSPLiv;
                       parameter C, data set 2) that will be combined with
                       information on the distribution of CYP2E1 to MSP
                       (CYP2Msp; parameter B, data set 1) to determine the
                       distribution of CYP2E1 to the intact liver (CYPUv;
                       parameter A),

                                 (pmol CYP2El/gram tissue)/
                                   (pmol CYP2El/mg MSP)
                                   = (mg MSP/gram tissue)          (4)


                       2.4. CYP2El-Dependent Oxidation of TCE

                          The Michaelis-Menten  kinetic constants  were
                       available for 23 samples of MSP from Lipscomb
                       et  al.{[1) The metabolism  of TCE  to chloral hy-
                       drate, representing oxidation by CYP2E1, was quan-
                       tified  by measuring  the formation  of chloral hy-
                       drate. Apparent l/max was  expressed as pmol TCE
                       oxidized/min/mg MSP, From  this set of 23 origi-
                       nal samples, 15 remained, and CYP2E1 content of
                       those microsomal protein samples was quantified by
                       ELISA.(15)  We  sought to develop a more  techni-
                       cal description of Fmax (the theoretical maximal ini-
                       tial rate of the reaction in  the presence of unlim-
                       iting  substrate concentration), and one that would
                       be more readily extrapolable to the in vivo setting
                       through incorporation of the information on the hep-
                       atic content of CYP2E1, To accomplish this, the
                       Vmm values (pmoles/min/mg MSP)  available from
                       the previously published study'17' were divided by the
                       CYP2E1 content of MSP (pmoles CYP2El/mg MSP)
                                                 B-6

-------
1226                                                                                        Lipscomb et al.






                                         Table II. Liver Enzyme Data*27*
Samples
i
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
1?
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Homog. Protein
(rag/Gram Liver)
134
301
137
100
113
151
154
148
115
137
181
180
126
124
122
137
152
130
69
126








































CYP2Elpcrmg
Homog. Protein
16.1
25.6
17.9
15.2
10.9
12.2
25.0
12.4
21.5
20.9
24.6
27.8
21.4
21.9
22.3
24.4
16.3
24.1
25.2
14.0








































Parameter A = pmol
CYP2E1 per Gram Liver
2157.4
2585.6
2452.3
1520.0
1231.7
1842.2
3850.0
1835.2
2472.5
2863.3
4452.6
5004.0
26%.4
2715.6
2720.6
3342.8
2477.6
3133.0
1738.8
1764.0








































Parameter B = pmol
CYP2E1 per mg MSP
85.5
99.8
83.4
23.0
34.0
36.3
76.8
46.0
69.0
58.5
54.0
64.0
68.0
53.0
46.0
66.0
41.0
24.0
42.0
41.0
52.5
94,0
46,5
90.0
11.0
64.0
41.0
64.0
30.0
57.5
53.5
55.0
52.0
29.0
39.0
39.0
73.0
70.0
130.0
34.0
31.0
48.0
29.5
19.0
77.0
91.0
37.0
74.0
44.0
50.0
75.0
29.5
69.0
91.0
48.0
36.0
41.0
26.0
26.0
53.0
Parameter C = mg MSP
per Gram Liver = A/B
25.2
25.9
29.4
66.1
36.2
50.7
50.1
39,9
35.8
48.9
82.5
78,2
39.7
51.2
59.1
50.6
60.4
130.5
41.4
43.0








































                                                   B-7

-------
Trichloroethylene and Human CYP2E1
                                                                                  1227
          Table III. CYP2E1 Content and TCE Metabolic Activity Used to Produce Data Set 3, Describing Parameter D
Samples
     Sample
     Number
(from Reference 17)
     pmol TCE
   Oxidize d/min/rng
MSP (from Reference 17)
  pmol CYP2E1/
     mg MSP
(from Reference 15}a
pmol TCE Oxidized/
 min/pniol CYP2E1
  (ParatneterD)
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75


HHM67
HHM84
HHM 86
HHM88
HHM 55
HHM 60
HHM 77
HHM 78
HHM 81
HHM 82
HHM 89
HUM 144
HHM 58
HHM 61
HHM 79


1113
1724
1039
1432
1422
1746
943
1627
1416
2353
890
1584
2078
2623
3455
Geometric Mean
Geometric Standard Deviation
11
64
44
50
52.5
91
19
77
37
74
30
30
94
90
91


101.2
26.9
23.6
2H.6
27.1
19.2
49.6
21.1
38.3
31.8
29.7
53.7
22.1
29,1
38.0
32.5
1 .538
"Data not used in estimation of parameter B.
to yield Vmax values expressed as pmolcs TCE ox-
idized/minute/pmol CYP2E1 (Table 111), This mea-
sure (parameter D) and its distribution are referred
to as data set 3 and are described as TCECyp2-

2,5.  Statistical Analysis
    Tables II and III summarize the data employed in
the statistical analyses.
    Probability  distributions were fitted  by  the
SAS8 8.0 Analyst routine to data describing the fol-
lowing variables  (with mnemonic  variable name):
A = pmol CYP2El/gram liver (CYP2Uv): B = pmol
CYP2EI/mg MSP (CYP2MSp): C = mg MSP/gram
liver (MSPtiv) =  A/B. StatFit software was used to
determine an optimal distribution fit to the 15 ob-
servations for parameter D,  pmoles TCE oxidized,
min/pmol CYP2E1 (TCEcypa)- The log-normal dis-
tribution was selected with parameters ju = 3.4812
and  a  — 0.4156 for the imbedded  normal distribu-
tion, implying a geometric mean and standard devi-
ation of 32.5 and  1.515, and an arithmetic mean and
standard deviation of 35.4 and 15,4.  This distribution
was accepted via chi-squared, Kolmogorov-Smirnov,
and Anderson-Darling statistical tests at the a = 0.05
(95%) confidence level.
    Three sets  of data were  available: a set of « =
60 samples, for which laboratory measurements were
available on B  - (CYP2Msp),  an n = 20 subset of
the 60 samples, for which several additional labora-
                                       tory measurements were available (PROLiv, CYP2Pro,
                                       CYP2MSP), and a set of n = 15 samples, for which one
                                       laboratory measurement was  available  (TCEcypz).
                                       These three sets of available data were first analysed
                                       separately. The additional variables (CYP2uv  and
                                       C = MSPuv) were calculated from the measurement
                                       data.
                                          For all variables,  normal,  log-normal, exponen-
                                       tial, and Weibull  distributions were fit using stan-
                                       dard statistical tests ol goodness-of-fit (Kolmogorov-
                                       Smirnoff, Cramer-von Mises, and Anderson-Darling)
                                       and a visual  examination of quantile-quantile plots.
                                       The null hypothesis  was  that  the distribution fit
                                       the  data  well, with  a  rejection of the null at p
                                       < 0.10. All  these analyses  were performed  us-
                                       ing  SAS°. Each  of  the  distributions was  ade-
                                       quately approximated by a log-normal distribution,
                                       the  parameters of which  are the mean (/x)  and
                                       standard  deviation (s) of the logarithms  of  the
                                       observations.
                                       2.5.7. Analysis via Method of Momenta

                                          For convenience, ignoring the dependence be-
                                       tween data set 2 and  the n = 20  (matched sub-
                                       set) of data set 1, and because the consistency of
                                       goodness-of-fit of the data to  the  log-normal  dis-
                                       tributions  (excluding data set 3: TCECYP2; pmol
                                       TCE oxidized/min/pmol CYP2E1), we applied the
                                                  B-8

-------
1228
                                                                                   Lipscomb ef al.
statistical method of moments  (addition of errors,
(Equation (5)) to combine data sets 1-3 to estimate
parameter E, pmol TCE oxidized/min/gram liver. All
goodness-of-fit p-values were greater than 0.15. As a
convenience, the log-normal parameter will be rep-
resented by the geometric mean (GM = e'1) and
geometric standard deviation (GSD =  es), respec-
tively, in this article. Equation (5) demonstrates the
method used to estimate the distribution of Vm!ix val-
ues, where the distributions for parameters B (pmol
CYP2El/mg MSP), C (mg MSP/gram liver), and D
(pmol TCE oxidized/min/pmol  CYP2E1) are com-
bined mathematically. The  values at the 5th (Xos)
and 95th (A^s) percentiles for the resulting distribu-
tion (parameter E. pmoles TCE oxidized/min/gram
liver)  were  calculated by Equations (6) and  (7),
respectively.
Lnorm[M=

                                             (5)
where M; is mean of logs of observations, s, is stan-
dard deviation of logs of observations, 1 is data set
l^(CYP2Msp), 2 is data set 2— (MSPuv), 3 is data set
                 = e["-L645x$1
             X95 =
                                             (6)
                                             (7)
2.5.2. Analysis via Computational Statistics
    We next sought to model the distribution of A =
pmol CYP2El/gram  liver with greater precision by
using all of the available data, including the corre-
lation (Fig. 3) on variables B = pmol CYP2E1/mg
MSP and C = mg MSP/gram liver. Since A = B x C,
these three variables are not statistically independent.
Moreover, it is perhaps not obvious how or whether
the 40 measurements of B that are not matched to
measurements on A and C  (observations 21-60 in
Table II) can be used to  improve estimation of the
distribution of A. However, we were able to  synthe-
size  and apply two techniques from computational
statistics—mixture distribution modeling*285 and clas-
sification trees^29-30^—to use all of the B and  C data,
including the 40 unmatched  measurements on B, to
model the distribution of A.
    The methodology for estimating the frequency
distribution of A using all available measurements
(i.e., using the joint distribution of A and B, as well as
the derived variable C) was as follows.
                                                      140
                                                      120
                                                               20     40     60     80    100
                                                                  pmoles CYP2E1/mg MSP
                                                                                               120
                                                 Fig. 3. Correlation between mg MSP/gram and pmol CYP2El/mg
                                                 MSP. The slight, but statistically significant, correlation between
                                                 the two parameters dictated the choice of statistical methods, Data
                                                 from Reference 27.
1. The frequency distribution for A can be ex-
  pressed using marginal and conditional prob-
  abilities as follows:

     Pr(A = a) = E(b,c5Pr(A|B = b & C - c)
                 Pr(B = b & C = c)
               = E(b,c)Pr(A|B = b & C = c)
                 Pr(B = b)Pr(C = c|B = b)
  where the sum (or integral) is taken over all
  (b, c) pairs  of values. Thus, A is interpreted
  as having a distribution that depends on the
  (perhaps unobserved) values of B and C.
2. The terms Pr(A [ B = b & C = c), Pr(B =
  b), and Pr(C = c | B = b) are estimated em-
  pirically from all of the available data using a
  classification tree estimator. Fig. 4 shows the
  classification tree fit to the first 20 cases in
  Table II, i.e., those with data on both A and B
  (and hence C). This tree provides an estimate
  of the distribution of A conditioned on the
  values of B  and C. The fit was performed us-
  ing the KnowledgeSeeker™ package (http://
  www.angoss.com/ProdServ/indexH.html). In-
  terpretively, the distribution  of A is mod-
  eled as a finite mixture distribution^28^ with
  a  number of components to be  estimated
  from the data. These components correspond
  to leaves  in the classification tree in Fig. 4.
                                                B-9

-------
Trichloroethylene and Human CYP2E1
                                                                                 1229
 Average A
 # of records
2642.7?
20
           [23, 53]
         [53, 99.8]
Avg=2029.23
sUl=614.41
n=9
Cluster 1



Avg -3144.75
n=ll
C



  [25.233, 36.226]
 [36.226,59.143]
[59.143, 1,30.542]
avg=24 16.88
std=182.68
n=4


avg=3093.62
std=496.97
n=5


avg=4728.3
std=389.9
«-2
    Cluster I
                      Cluster 3
                                       Cluster 4
Fig. 4. Classification tree model for the distribution of A = pmol
CYP2El/gram liver. The distribution of A is modeled as a finite
mixture distribution with components corresponding to the leaves
in the depicted classification tree. Trie conditional distribution of A
depends on which component distribution a case belongs to. The
components arc bounded by breakpoints in the observed values for
B and C, The sample means and sample standard deviations for the
four component distributions arc estimated from the first 2(3 cases
in Table 11.
       The conditional distribution of A depends on
       which component distribution a case belongs
       to.  Fig.  4 shows the sample means and stan-
       dard deviations for four component distribu-
       tions (leaves), estimated from the first 20 eases
       in Table II.  Using only these data, the es-
       timated distribution of A would correspond
       to the following finite (4-eomponent) mixture
       distribution:
       Cluster 1: weight  =  9/20, sample  mean =
                 2029.23, sample standard deviation
                 = 614.41
       Cluster 2: weight  =  4/20, sample  mean =
                 2416.88, sample standard deviation
                 = 182.68
       Cluster 3: weight  =  5/20, sample  mean =
                 3093.62, sample standard deviation
                 = 496.97
       Cluster 4: weight  =  2/20, sample  mean =
                 4728.3, sample standard deviation
                 = 389.9
       Here, the four components are termed "clus-
       ters" since they correspond to sets of cases for
       which the distribution of A values is approx-
       imately the same (i.e., the classification tree
       algorithm is unable to find any statistically sig-
       nificant difference among them).
    3. This initial tree based on the first 20 (full-
       data) cases in Table II was  refined by using
       the remaining 40 observations of B values in
       Table II (i.e, cases 21-60) to better estimate the
       fraction of all cases for which B < 53 (the defin-
       ing characteristic of Cluster 1). The pooled es-
       timate from all 60 eases is that 32/60 (= 0.53,
       95% CI = 0.40 to 0.66) of A values are drawn
       from Cluster 1, The revised cluster weights us-
       ing all 60 observations on B are: 0.53 for Clus-
       ter 1; 0.17 for Cluster 2; 0.21 for Cluster 3; and
       0.09 for Cluster 4, While the cluster-specific
       sample sizes are very small  (n = 2 for Clus-
       ter 4), this decomposition of the  distribution
       of A into a  weighted mixture of component
       distributions actually decreases the variance in
       estimates of the true mean (and other statis-
       tics) of A compared to using a single estimated
       distribution J31)

The methodology summarized in Steps  1-3 can be
further refined, e.g., by using resampling to establish
robust boundaries for the classification tree splits, or
by using a Bayesian posterior distribution for the frac-
tion of cases belonging to different clusters. However,
given the small number of cases (« = 20) with full data.
additional refinements of the tree estimator in Fig. 4
with the cluster weights obtained from all 60 measure-
ments for B are not expected to greatly improve the
estimation of the distribution of A.
                                     2,6. Combination of Data Sets

                                        A program was developed in the MATLAB soft-
                                     ware (see Appendix A) to produce A, D, and A x D
                                     random variates in accordance with the distributions
                                     for A and D derived above. The distribution of A was
                                     taken from that determined by computational statis-
                                     tics. This program identified 100,000 random variates.
                                     Eight thousand of the generated A  x  D values were
                                     selected at random and were subjected to  a further
                                     analysis to find an optimal distributional fit (StatFit
                                     has a limit of 8,000 values).
                                     2.7. PBPK Model

                                         Human metabolism of TCE was simulated us-
                                     ing the PBPK  model of  Allen and Fisher(32) and
                                                    B-10

-------
1230
                                  Lipscomb et al.
SimuSolv  software  (Dow Chemical  Co., Midland,
MI), The  model structure consisted  of four  tissue
compartments (liver, rapidly perfused tissues, slowly
perfused tissues, and fat) and a gas exchange compart-
ment (lung) connected by blood flows. AH TCE bio-
transformation was assumed to take place in the liver
and follow Michaelis-Menten kinetics. The liver was
described  as  a well-stirred homogeneous compart-
ment. Previous studies have demonstrated that more
complex heterogeneous models of the liver, such as
the parallel tube and dispersion models, were not bet-
ter than the simple well-stirred model at predicting the
in vivo clearance of 28 drugs from in vitro data.'33^ The
model was set to simulate two extreme exposure sce-
narios: (1) simulating a higher, but permitted, occupa-
tional exposure at the Threshold Limit Value (TLV)
for TCE, which is 50 ppm in air for an 8-hour working
day,*:34' antj (2) simulating a low-dose environmen-
tal exposure via drinking water  containing the max-
imally allowable concentration of TCE (5 ug/L).(33)
With knowledge that the hepatic metabolism of TCE
in vivo is  limited by blood flow, the model was set
to simulate a "worst case" scenario of an oral bolus
dose, because this (rather than  a slower oral  inges-
tion rate)  would be the more likely scenario to pro-
duce differences in the amount of TCE metabolized
(AML). AML is presented in units of mg TCE me-
tabolized over the course of the simulation per liter of
liver tissue. Simulations of AML (amount of TCE me-
tabolized in the liver compartment) were evaluated
because the CYP2El-dependent oxidation of TCE is
a required step in the formation of the hepalotoxic
metabolite, trichloroacetic acid (TCA), in vivo. Chlo-
ral hydrate, the oxidative metabolite of TCE, does
not have a measurable half-life  in vivo following its
formation from TCE, but is immediately converted
to TCA and trichloroethanol. For these simulations.
the model incorporated both extremes of the distri-
bution of the Vmax for TCE oxidation (5th and 95th
percentilcs).


3. RESULTS

3.1. Distribution of CYP2E1 to Human
    Hepatic MSP

    Analysis of 60 samples of MSP derived from in-
dividual adult human organ donors for the content
of CYP2E1 (pmolcs CYP2El/mg MSP, parameter B,
data set 1) indicated that the log-normal distribution
adequately represented the set of observations. The
geometric mean and geometric standard  deviations
required to reconstruct the overall distribution  and
simulate the value for a percentile of interest are pre-
sented in Table IV. These values agree well with those
reported by Shimada et alS^ Variance between the
values at the 5th and the 95lh percentiles  of the dis-
tribution was approximated fourfold.
3.2. Distribution of CYP2E1 to Intact Human Liver

    Three types of analytical procedures were used to
determine the distribution of CYP2E1 to intact liver
tissue (pmoles CYP2El/gram liver, parameter A) de-
rived from adult human organ donors. First, the most
direct measure, but one for which only 20 observa-
tions are available, is depicted in Equations (1) and
(2) and involved the application of the ELISA tech-
nique to liver homogenale (post 100 x g) protein. The
empirical distribution of the 20 observations indicates
that the magnitude of variance between  the obser-
vations representing approximately the 5th and 95th
    Table IV. Distributions of TCE Metabolism Rate Constant. Microsomal Protein, and CYP2H1 Content of Adull Human Liver
Parameters
Description
Distribution
GM
OSD
Range
5lh Percentile
95th Percentile
A
CYP2?iv (Derived)
Discrete
2562
930
1232-5004
1232
4453
B
CYP2£ISP (Data Sell)
Log Norma!
48.9
1.6
11-130
22.5
106
C
MSPyjv (Data Set 2)
Log Normal
52.9
1 .476
27-108
27.9
100
D
TCEjtyp, (Data Set 3)
Log Normal
~32.5
1.515
19.2-101.2
16.4
64.4
E
TCE£jv (Derived)
Log Normal
78.810
1.7274
32,069
193.679
apmoles CYP2El/grani liver; data are presented as the arithmetic mean, arithmetic standard deviation, and values at the 5.89th and 95.5th
percentiles, n = 20. GM and GSD derived by computational statistics.
bpmoles CYP2El/mg MSP, n = 60.
cmg MSP/gram liver, n = 20.
d Vmax of CYP2E1 in human liver MSP toward TCE, pmoles/min/pmol CYP2H1. n - 15.
e^mai as pmoles TCE oxidized/rnin/gram liver, derived via computational statistics.
                                              B-11

-------
Trichloroethylene and Human CYP2E1
                                                                                              1231
percentiles of the empirical distribution is approxi-
mately threefold. These raw data indicate a mean
value of 2,643 and a standard deviation of 962 pmoles
CYP2El/gram liver.
    Second, the  application of the statistically lim-
ited method of moments required the characteriza-
tion of the two underlying distributions: (1) param-
eter B, pmoles CYP2El/mg MSP, data set 1  and (2)
parameter C, mg MSP/gram liver, data set 2. Because
the observations in these two sets of data were ad-
equately fit by a log-normal distribution, the values
for the geometric mean and geometric standard devi-
ations  for each data set (Table IV) were combined
(Equation (4)). The log-normal distribution of the
liver content of MSP (parameter C) was characterized
by a geometric mean of 52.9 mg MSP/gram liver and
a geometric standard deviation of 1.476 (arithmetic
mean and standard deviation of 57 ± 23 mg MSP/gram
liver). The MSP  content of CYP2E1 (parameter B)
was determined  in a  sample set of 60, and  demon-
strated a log-normal  distribution, with a geometric
mean of 48.9 pmoles CYP2El/mg MSP and a geomet-
ric standard deviation of 1.6. The arithmetic mean ±
standard deviation was 54 ± 23 pmoles CYP2El/mg
MSP. When combined, the results indicated a geo-
metric mean of  2,587 pmoles CYP2E I/gram intact
adult human liver, with a  geometric standard devi-
ation of  1.48.  From analysis via Equations  (5)  and
(6), these parameter values indicate values at the 5th
and 95th percentile of the distribution to be  949 and
7,053 pmoles CYP2El/gram. These values are simi-
lar  to those indicated by the direct measurement  of
CYP2E1 in homogenate protein, above. These data
indicate that the central 90% of the population rep-
resented by these 60 adult  organ donors expresses a
CYP2E1 content that varies 7.4-fold,
    Finally, a specific probability distribution for the
parameter A was developed, based upon the clusters
derived in Section 2.5.2. Recall that clusters of values
for A were identified, into which values for param-
eter B  (pmoles CYP2El/mg MSP) were segregated.
In this manner, the influence of parameter B, or its
determinant qualities, on parameter A was character-
ized. A continuous probability distribution was not
fit to the individual clusters due to the small num-
ber of  observations within each  cluster; instead, the
distribution of A was assumed to be discrete and con-
sist only  of the observed values (4th column. Param-
eter A, of Table  II). Clusters were assumed to occur
with proportional frequencies equal to the  weights
{0.53, 0.17, 0.21, 0,09} and to have counts of {9, 4,
5, 2} as described previously. Within a cluster, each
value belonging to that cluster is assumed to occur
with equal frequency. This approach provides prob-
abilities of 0.53/9 (= 0.0589) for values in Cluster
1, 0.17/4 (= 0.0425) for values in Cluster 2, 0.21/5 (=
0.0420) for values in Cluster 3, and 0.09/2 (= 0.0450)
for values in Cluster 4. Cluster statistics are presented
in Fig. 4. If we use x/ to represent the j'th value of pa-
rameter A, and p,- to represent the probability of the
j'th value, then the mean of the resulting distribution is
                20
                    iXt -2561.77
                                            (8)
        20
                   20
while the variance of the distribution is
                         v 2
                      pixA  =865,563.40    (9)
       1=1       \i=i    /
providing a standard deviation of 930.36.
    Note that the mean of the raw data for parameter
A (Table II) is 2,642.8 while the standard deviation
(using Equation (9)) is 937.21 (the sample standard
deviation is 962). We see that accounting for the in-
fluence of parameter B has shifted our estimates of
parameter A slightly downward, and has slightly de-
creased the standard deviation. This is a result of the
individual probabilities being shifted slightly upward
or downward from 0.05 in accordance with the distri-
bution shown in empirical distribution. This distribu-
tion was used in the recombination of data describing
parameter A with data describing parameter D, as
discussed below.


3.3. In vitro Metabolic Rate Constant (Vmax)
    The Fmax for the oxidation of TCE by CYP2E1
(pmoles  TCE oxidized/min/pmol CYP2E1, param-
eter D,  data set 3, Table III) was evaluated in
a data  set of 15  samples. The  apparent Frnax ob-
served in vitro (pmoles TCE oxidized/min/mg MSP)
was converted to the more specific units of pmol
TCE oxidized/min/pmol CYP2E1 by dividing the ob-
served Vmax value by the content of CYP2E1  in the
MSP (pmoles CYP2El/nig MSP). The resulting set
of 15 observations  (pmol  TCE oxidized/min/pmol
CYP2E1) were fit optimally with the log-normal dis-
tribution; its parameters were fj. = 3.4812 and a —
0.4156 for the embedded  normal distribution, im-
plying a geometric mean and standard deviation of
32.5 and 1.515, and an arithmetic mean  and  stan-
dard deviation of 35.4 and 15.4. This distribution was
accepted via chi-squared, Kolmogorov-Smirnov, and
Anderson-Darling statistical  tests  at  the  a =  0.05
                                                 B-12

-------
1232
                                  Lipscomb el al.
(95%) confidence level. The difference between val-
ues at the 5th and 95th percentiles of the distribution
approximated fourfold.


3,4, Determining the Metabolic Capacity of Intact
    Tissue and Extrapolation of Units

    MATLAB results of the simulations of 100,000
random variates of A x D revealed a plot (not shown)
suggestive of a log-normal distribution, and a nor-
mal distribution (not shown) of the logs of values
of A  x  D. StatFit analyzed 8,000 of these variates,
and indicated  that the most likely distribution was
the log-normal, with parameters n — 11.2748 and 
-------
Trichloroethylene and Human CYP2E1
                                                                             1233
Magnitude of Variance in PK Output
  LojKHpmill TCE/min(
  puwlCVTZEl), n- IS
Fig, S. Extrapolation  and  incorporation of  in vitro derived
metabolic rates in PBPK modeling. This figure  depicts the frame-
work for deriving appropriate in vitro measures and their extrap-
olation into a PBPK model. The model was exercised to simulate
environmentally and occupationally relevant exposures.
Genetic polymorphisms and enzyme induction due to
environmental and lifestyle factors can affect the level
of expression of xenobiotic metabolizing  enzymes.
Thus, genetic polymorphisms become critical to risk
only when they alter PK outcomes. The refinement of
human health risk assessments for chemicals metab-
olized by the liver to reflect data on human interindi-
vidual PK variability can be accomplished through
(1) the characterization of enzyme expression in large
banks of human liver samples, (2) the employment of
appropriate techniques for the quantification and ex-
trapolation of metabolic rates derived in vitro, and (3)
the judicious application of PBPK modeling.
    Numerous PK outcomes may be simulated by
PBPK modeling; the identification of the risk-relevant
PK outcome(s) from toxicity studies allows the study
of their variability through adequately constructed
PBPK models. When PK models are  constructed to
include metabolic rates (and rate constants) derived
in vitro, several extrapolations are necessary, not the
least  of which is the extrapolation of enzyme con-
tent  (Fig. 5).  PBPK models include the  apparent
Vmax  expressed as mg/h/kg  body mass, while typi-
cal in vitro studies express Kmax  in terms of nmoles
product formed/minute/mg microsornal protein. Ac-
curate extrapolation requires initially that enzyme
content be expressed per unit intact liver (i.e., pmoles
CYP2El/gram liver), and the extrapolation has usu-
ally included a numerical estimation of the MSP con-
tent of liver (i.e., 50 mg MSP/gram liver). The MSP
content  of intact liver has  been measured and used
to extrapolate in vifro-derived metabolism kinetic
constants for use in PBPK modeling efforts in hu-
mans^37i38) and to infer measures of intrinsic clear-
ance (Clint) in traditional rat-based PK models/39^
In the previous PBPK-based approach for TCE,(37}
samples expressing extreme values for kinetic con-
stants (KM and Vmax) were chosen for extrapolation
to a PBPK model. Those extrapolations were based
on the hepatocelMarity of intact liver tissue, and on
microsornal protein content of liver, rather than the
content of CYP2E1 of liver. The Fmax value was not
previously extrapolated on the basis of CYP2E1 con-
tent as no data  existed at the time through which
to quantify  the  distribution of  the key  metabolic
enzyme within  human  liver. With respect to the
distribution of the cytochromes P450 in one prepa-
ration of human liver (microsomes), several inves-
tigations^17'36-40'41^ have revealed quantitative infor-
mation about the content of these multiple enzyme
forms in this preparation, but reveal no direct infor-
mation on the type of distribution (i.e., log-normal)
of the enzymes within MSP, their content or distribu-
tion to the intact liver in situ. In the present study, we
developed measures of the liver content of microso-
rnal protein, of which the CYP  enzymes (and other
important xenobiotic-metabolizing enzymes, i.e., glu-
curonyl transferases) are a constituent. This key piece
of information is necessary to estimate the content of
the enzyme(s) in the intact liver. By combining the
two data sets on (1) the MSP content of CYP (pmoles
CYP/mg MSP), and (2) the liver  content of MSP (mg
MSP/gram liver), we derived the liver content of CYP
(pinoles CYP/gram liver), and developed measures of
that variance, employing a total number of 60 sam-
ples derived from adult human organ donors. The ap-
proach also included the determination and inclusion
of the human interindividual  variance in metabolic
activity toward TCE derived in  an additional set of
15 samples. This analysis allowed the variance of that
critical enzyme kinetic parameter (Vmax) to be exam-
ined among humans, and expressed as pmoles TCE
oxidized per minute per pmol CYP2E1. This parame-
ter (pmol TCE oxidized/min/pmol CYP2E1) did vary
among the human samples evaluated, not surprisingly.
This may be explained, in part,  by potential under-
lying genetic differences impacting CYP2E1 activity,
differences in the presence of other CYP forms that
also metabolize TCE at higher  concentrations, and
human-to-human interindividual differences in the
lipid composition (both qualitative and quantitative)
of isolated MSP preparations. The activity of isolated
enzymes represents the functional status of their re-
spective donors. The stability of these enzymes upon
                                                 B-14

-------
1234
                                  Lipscomb el al.
isolation and storage seems not to be a major contrib-
utor to this variance. The level of detail in this expres-
sion of Vmax allowed for a direct combination with in-
formation on the variance of CYP2E1 in intact human
liver, which enabled the resulting PBPK analysis of
the impact of that variance on the risk-relevant phar-
macokinetic (PK) outcome, amount of TCE metabo-
lized in the liver, among adult humans. Together, these
data sets separately describing enzyme activity and
enzyme  content combine to  describe the metabolic
capacity of the liver. The resulting combined distribu-
tion for the Vmm value demonstrated that this param-
eter (mg/h/kg) differed more than six-fold between
the values at the 5th and 95th percentiles of the dis-
tribution.  When these values were separately inte-
grated into the PBPK model, resulting  estimates of
the amount of TCE oxidized over the exposure pe-
riod differed by only 2%. Thus, widely divergent val-
ues for apparent Vm&yi, resulting from both variance
in enzyme content and activity, had little effect on the
m vivo metabolism of TCE, and will have little effect
on the hepatotoxic injury following TCE exposure in
humans.
    The present work demonstrates a significant ad-
vantage over earlier studies in that statistically valid
and robust measures of enzyme content and enzyme
activity  have been developed and incorporated into
the PBPK-based approach. This advance allows the
application of the approach  to estimate population
distributions of risk, when chemical dose-response
parameters (e.g., slope factors) are available. With
the availability of large banks of well-characterized
subcellular fractions  (mainly hepatic MSP) derived
from the livers of human organ donors comes the
opportunity to  determine several measures  of hu-
man biochemical individuality, which will be appli-
cable to many environmental, occupational, arid ther-
apeutic  compounds. Although several investigations
have failed to identify an inverse relationship between
post mortem cold-clamp time (the time interval be-
tween the  perfusion,  removal, and refrigeration  of
liver tissue and the freezing  of the tissue or micro-
somal protein isolation) and  microsomal enzyme ac-
tivity, the assumption that the activity  of these en-
zymes in vitro represents their activity  in vivo must
be recognized as such. From these samples, we can
measure interindividual differences in enzyme ac-
tivity  and differences in enzyme content in isolated
MSP. The in vitro metabolism of several CYP2E1 sub-
strates, such as furan,(41<43) perchloroethylene/37) and
trichloroethylene/38'  have been successfully extrap-
olated to  the in  vivo  setting through application of
adequately developed and validated PBPK models.
The additional validation of the extrapolation proce-
dure for metabolic activity based on enzyme recovery
data is important. This demonstrates the applicabil-
ity of the methodology to determine the interindi-
vidual differences of risk-relevant PK outcomes (i.e..
the amount of metabolite  formed in the liver for a
bioactivated hepatotoxicant) for xenobiotics to which
humans cannot be  safely  exposed for the genera-
tion of experimental data.  It is anticipated that tox-
icological data can  be generated  in test species in
vivo and in vitro to  determine the metabolic species
responsible  for toxicity, the PK of the xenobiotic
and metabolite(s), and the identity of the enzyme
responsible for metabolism. With this information,
an adequate test animal-based PBPK model can be
extrapolated to humans, using human tissue partition
coefficients and the appropriate physiological param-
eters. Data on human enzyme recovery could be used
to develop appropriate bounds on the distribution
of metabolic activity for evaluation  with the PBPK
model  to represent predefined proportions of the
population.
    The successful application of this approach re-
quires  the avoidance  of several pitfalls. It requires
(1) the metabolic process  under investigation must
be as directly linked to the risk-relevant PK outcome
as possible:  the correct identification of the critical
toxic effect, against which protection  is warranted,
or toward which  susceptibility requires evaluation.
In the absence of a defined link between this effect
and its most closely related and measurable or pre-
dictable PK outcome (e.g., AML), then further effort
will not advance the goals of the approach. (2) The
tissues/preparations included in the experiments must
be viable. The reliance on human tissues of research
grade can be troublesome; the comparison of in vitro-
derived metabolic rates and rate constants, especially
in humans, requires some justification that these ex
vivo or in vitro systems maintain the metabolic ca-
pacity they possess  in vivo. The isolated hepatocyte
model is more closely related to the in vivo situation
than the isolated microsomal protein preparation, but
metabolic rates from both systems require extrapola-
tion based on recovery information to the in vivo situ-
ation. Reliance upon data derived from compromised
in vitro systems can lead to underpredictions of in vivo
metabolism. The inclusion of data from compromised
systems (i.e., lengthy 37e incubations of microsomal
protein,  the application of immortalized cell lines,
etc.) must be avoided. The evaluation of metabolic
activity toward recognized marker substrates  and
                                               B-15

-------
Trichloroethylene and Human CYP2E1
                                            1235
assessment of cellular viability provide some evidence
of in vitro system stability. (3) There must be suffi-
cient data to enable extrapolation based on protein
recovery. Lack of data or uncertainty in the available
data quantifying the relationship between the in vitro
system and the in  vivo situation greatly complicate
the extrapolation procedure. Values for hepatocellu-
larity in rats and humans, and values for microsomal
protein content of rats and humans, are available for
use in extrapolation procedures. (4) The derivation of
metabolic rate constants must be accomplished under
valid experimental conditions. Rate constants must be
derived under conditions where rate is proportionate
to an increase in protein content, over time and with
increasing substrate concentrations (for first-order re-
actions). The value of such data is enhanced when rate
constants are lied specifically to  the enzyme, rather
than the subcellular fraction (e.g., pmoles/min/pmol
CYP2E1 vs. pmoles/min/mg MSP). (5) Data should
be used to identify the pertinent enzyme;  the contri-
bution of more than one enzyme complicates enzyme
kinetic evaluations. Additional uncertainty is encoun-
tered in the metabolic evaluation of substrates, toward
which multiple enzymes are active. Given the human
interindividual variability on enzyme expression as a
result of genetic, dietary, and lifestyle choices, differ-
ent ratios of two potentially active enzymes may be
observed. In this instance, the approach to  in vitro en-
zyme kinetic investigations must be robust enough to
separately identify the kinetic constants applicable to
each of the enzymes. Kinetic constants derived for the
preparation, without regard to the pertinent enzymes,
can falsely indicate that the apparent Vmax value  is
shifted upward due to the contribution of a low affin-
ity form, when in vivo substrate concentrations would
not be sufficient to drive an appreciable contribution
of this enzyme to the reaction. (6) This approach re-
lies on the availability of a "validated" PBPK model.
While generalization of model structure and physio-
logical components across chemicals is often the case,
the models must include parameters demonstrated or
judged to be relevant  to the study chemical. In addi-
tion to metabolic rate constants, tissue partition coeffi-
cients (PC) are highly chemical-specific, and differ for
the same tissue type among species. The application of
PC values derived in other species or adapted from PC
values of related chemicals requires justification. (7)
Finally, the approach  is aimed specifically at quanti-
fying human interindividual differences in metabolic
capacity. This approach is not specifically aimed at
quantifying human interindividual PK difference for
TCE oxidation; it was developed to test the hypothesis
that variability in metabolic capacity alters suscepti-
bility to hepatotoxic injury from TCE exposure. The
approach here demonstrates the applicability of the
statistical bounds established for the population un-
der investigation. In that regard, the application of a
representative sample set is required. The data must
support the identification of distribution type and in-
clude enough observations so that confidence can be
placed in the identification of values at predetermined
points of the distribution. While these are some of the
general pitfalls, investigators trained in the disciplines
of the individual investigatory steps will  be quite fa-
miliar with many of the more technical pitfalls.
    To members of the risk assessment community
who are advocating the development of approaches
that provide more  information than just "safe expo-
sure limits" (e.g., RfC and RfD values), the present
approach  may be  useful. The  approach is centered
on the idenlification of the risk-relevant PK outcome
through evaluation of toxicity and PK investigations,
not necessarily through PBPK investigations. Under
optimal conditions, the linking of PBPK modeling
approaches with data  describing  human biochemi-
cal individuality (enzyme content  and enzyme activ-
ity) will allow the  quantification of the  PK compo-
nent of UFH. The collection of advanced measures
of human biochemical individuality (e.g., differences
in the liver's content of critical xenobiotic metabo-
lizing enzymes) will broaden the applicability of this
approach to other chemicals whose PK are modulated
by the same enzyme. It is conceivable that when this
parameter (enzyme content and activity) modulates
the production of the risk-relevant PK outcome, in-
formation about the population distribution of the pa-
rameter (i.e., hepatic content of CYP2E1) will lead to
applications demonstrating the fraction of the popu-
lation that will be protected by regulations that spec-
ify a given level of  chemical exposure. Similarly, with
carcinogenicity slope factors, risk-relevant PK out-
comes can be converted directly to measures of risk,
indicating the level of risk corresponding to a given
level of enzyme content and activity. By converting
exposure to tissue dose, and having information to
link tissue dose to risk, the PBPK modeling approach
may be usefully employed to develop distributions of
risk, rather than simply assessing or demonstrating
the health protective nature of a given exposure.
    The purpose of the present study was to explore
the potential advantage of including additional, spe-
cific information on human biochemical individuality
as a process to refine the human health risk assessment
process. Because an ever-increasing amount of data is
                                                  B-16

-------
1236
                                 Lipscomh et al.
being developed through the analysis of tissues de-
rived not only from surgical resections, but also from
human organ donors, these sources of tissues offer
a unique potential to increase our knowledge about
human biochemical individuality.

5. SUMMARY AND CONCLUSIONS
    Human interindividual PK variability is impor-
tant both  for chemicals with adequate human PK
data and for those chemicals to which humans cannot
be experimentally exposed. Because CYP2E1 activity
limits (in vitro) the production of oxidative, hepato-
toxic metabolites of TCE, we evaluated the distribu-
tion of that enzyme in liver from up to 75 adult human
organ donors by applying published and accepted bio-
chemical and statistical methods. The extrapolation
of in vitro data captured both the variance in enzyme
content and enzyme activity among adult humans.
    CYP2E1 content and metabolic activity toward
TCE are described by log-normal distributions. The
central 90% of the human population represented by
these adult organ donors differs by less than fourfold
in the hepatic content of this critical xenobiotic me-
tabolizing enzyme; that same fraction of the popula-
tion differed by approximately sixfold with respect to
the oxidation of TCE. The finding and additional in-
formation to be gained from the now-characterized
distribution of CYP2E1 to intact human liver will be
useful not only to the assessment of risk from TCE
exposure,  but also to the assessment of risks from
other environmental chemicals that are also metab-
olized by this enzyme, including chloroform, carbon
tetrachloride, benzene, toluene, and styrene. Because
the metabolism of TCE is limited by blood flow to
the liver, divergent  values of Fmax do not result in
appreciable differences in the risk-relevant PK out-
come, the amount of TCE metabolized in the liver.
Therefore, factors that increase the hepatic expres-
sion of CYP2E1 and/or its metabolic activity will not
always result in proportionate changes in key PK out-
comes. This is because of the relatively low solubil-
ity of TCE in blood, and the  relatively high capacity
of the liver to metabolize TCE  (due to a relatively
high level of expression of the enzyme and the rela-
tively high metabolic activity of the enzyme toward
TCE), the limiting factor, in vivo, for TCE oxidation
becomes the rate at which TCE is delivered to liver tis-
sue by hepatic blood flow. In this situation, increases
in TCE metabolic capacity, even from the 5th to the
95th percentiles of the distribution, result in only a
2% increase in the amount of TCE metabolized. With
respect to the hepatotoxicity of TCE resulting from
exposure scenarios similar to those employed in this
analysis, these data indicate that the amount of PK
variability attributed to enzymic variance among hu-
mans is approximately 2%. The approach described
here is especially applicable to chemicals to which hu-
mans cannot be experimentally exposed for ethical
reasons. The application of actual, not hypothesized,
bounds of variance and the definition of the distribu-
tion of enzyme content and activity among humans
can allow the calculation of finite levels of risk (when
dependent on  the PK outcome) at different chosen
percentiles of the distribution of enzyme content and
activity.
    Several conditions must first be met for this strat-
egy to be successful:

     1. The target organ, mode, or mechanism of ac-
      tion, and metabolic species responsible for
      toxicity must be known.
     2. The target tissue-toxic chemical species dose-
      response relationship must be known.
     3. The biotransforming enzyme must be known
      and information on the variance and type of its
      distribution among humans must be known.
     4. The kinetic mechanism of metabolism must be
      known and expressed per unit of enzyme.
     5. An adequately characterized  PBPK model
      must be available for adaptation.

We  have  quantified  the  extent of variance in en-
zyme content  of  a critical xenobiotic metabolizing
enzyme, CYP2E1, and the variance in the  hepatic
biotransformation of a key environmental contami-
nant, trichloroethylene. The parameters of the result-
ing log-normal distributions can be used to identify
the bounds of biochemical and pharmacokinetic vari-
ance (e.g., 90% of the population), within which sus-
ceptibility can be determined and allows the replace-
ment of hypothesized magnitudes of difference with
actual measurement  of such when determining the
impact of enzyme variance on risk.
    This article identifies  the conditions and types of
data required, communicates and applies a logical ap-
proach, and describes the limitations of the approach
in estimating the  human  interindividual variance of
risk-relevant PK outcomes that may signify suscepti-
bility to chemical injury.  While data set 3 is unique
to TCE, data set 2 will  be useful  in estimating the
hepatic content of all enzymes contained in the mi-
crosomal fraction, when their distribution character-
istics are known,  and the information derived from
the  combination  of data sets 1 and  2 are  directly
                                              B-17

-------
Trichloroethylene and Human CYP2E1
                                              1237
applicable to other environmental contaminants that
are also substrates for CYP2E1.

ACKNOWLEDGMENTS

    The  views expressed  in  this article (NCEA-C-
0956)  are those of the individual authors  and do
not necessarily reflect the views and policies of the
U.S. Environmental Protection Agency (EPA). This
research  was supported by  an  interagency agree-
ment between U.S. EPA/NCEA and CDC/N1OSH
(No. DW75851501; John C, Lipscomb, Project Offi-
cer) and through a cooperative agreement between
U.S. EPA/NCEA and Dr. Gregory  Kedderis  (No.
CR828047-01-0; John C. Lipscomb, Project Officer).
The authors wish to express their gratitude to Glenn
Suter and Bob Bruce (EPA, NCEA, Cincinnati), and
Ulrike Bernauer (BgVV, Berlin, Germany) for in-
sightful comments during preparation of the article.
Preliminary results from this investigation have been
presented at the  annual  meeting of  the Society for
Risk Analysis, December 2000,  Arlington, VA and
December 2001, Seattle, WA; at the annual meeting
of the Society of Toxicology, March 2001, San Fran-
cisco, CA, and March 2002, Nashville, TN; and at the
Spring Toxicology Conference,  April 2001, Wright-
Patterson Air Force Base, OH. We are sincerely grate-
ful to organ donors and their families; these and other
studies would not be possible without their generous
contributions.

APPENDIX A: MATLAB CODE TO
GENERATE A x D

numReps  = 250000;
cluster{i} =  [1520 1231.7 1842.2 1835.2
   2720,6 2477.6 3133 1738.8 1764];
cluster{2} =  [2157.4  2585.6 2452.3 2472.2];
cluster{3} =  [3850 2863.3 2696.4 2715.6
   3342.8];
cluster{4} =  [4452.6  5004];
clusterCDF =  [.53  .70 .91 1.0];
A  = zeros(1,numReps);
result = zeros(1, munReps);
for i =  1:numReps
   D = exp(3.4812  +.4156*randn);
   clusterNum = min(find(rand  <
     clusterCDF));
   clusterSize = length(cluster{clusterNum}) ;
   clusterIndex =  ceil(rand  *  clusterSize);
   A(i)  =  cluster{clusterNum}(clusterIndex);
   result(i)  = A(i) * D;
end
writeArray = result *;
save Dvalues.txt writeArray  -ASCII
dispCA parameters')
disp([mean(A) std(A)])
dispCAxD  parameters')
dispC[mean(result) std(result)])
hist(result,100)
title('Empirical Distribution of  A x D1)
figure(2)
hist(log(result),100)
title('Empirical Distribution of  ln(A x
   D)3)

REFERENCES

 1.  Bogdanffy, M. S., & Jarabek, A. M. (1995). Understanding
    mechanisms of inhaled toxicants: Implications for replacing
    default factors with chemical-specific data. Toxicology Letters,
    82/83,919-932.
 2.  Jarabek, A. M. (1995). Interspecies extrapolation based on
    mechanistic determinants of chemical disposition. Human and
    Ecological Risk Assessment, 1, 641-662,
 3.  Bogdanffy, M, S., Sarangapani, R., Plowchalk, D, R.. Jarabek,
    A. M., & Andersen. M. E. (1999). A biologically-based risk
    assessment for vinyl acetate-induced cancer and noncancer
    inhalation toxicity. Toxicological Sciences, 51,19—35,
 4.  WHO (World Health Organization). (1998). Guidelines for
    Drinking Water Quality. Addendum to Volume 2, Health Cri-
    teria and Other Supporting Information, 2nd ed. (pp. 15-30).
    Geneva, Switzerland. ISBN 92-4-1545143.
 5.  IPCS (International Programme on Chemical Safety). (1994).
    Environmental Health Criteria 170. Assessing Human Health
    Risks of Chemicals: Derivation of Guidance. Values for Health-
    Based Exposure Limits. Geneva, World Health Organization.
 6.  IPCS (International Programme on Chemical Safety). (1998).
    Environmental Health Criteria 204: Boron. International Pro-
    gramme on Chemical  Safety. Geneva, Switzerland, World
    Health Organization, ISBN 92-41572043.
 7.  U.S. EPA. (1999). Toxicological Review of Ethylene Glycol
    Monohutyl Ether (111-76-2) in Support of Summary Informa-
    tion on the Integrated Risk Information System (IRIS), Avail-
    able at http://www.epa.gov/iris.
 8.  US.   EPA.  (2000).   Toxicological  Review  of  1,3-
    Dichloropropene  (542-75-6)  in  Support of  Summary
    Information on  the Integrated Risk  Information  System
    (IRIS),  National  Center for  Environmental Assessment,
    Washington, DC. Available at http;//www,epa,gov/iris,
 9.  U.S. EPA.  (2001). Toxicological Review of Hexachlomcy-
    clopentadiene (77-47-4) in Support of Summary Information
    on the Integrated Risk Information System (IRIS). Available
    at http://www.epa.gov/iris.
10.  U.S. EPA. (1998). Toxicological Review of Methyl Methacry-
    late (80-62-6) in Support of Summary Information on the
    Integrated Risk  Information System (IRIS). Available  at
    http://www.epa.gov/iris.
11.  Lipscomb. I C, & Kedderis, G. L. (2002). Incorporating hu-
    man interindividual biotransformation. Variance in health risk
    assessment. Science of the Total Environment, 288,13-21.
12.  Lehninger, A. L. (1975). Biochemistry, 2nd ed, (pp. 183-216),
    New York: Worth.
13.  Segel, I. H. (1975). Enzyme Kinetics (pp. 1-17). New York:
    Wiley.
14.  Kedderis, G. L. (1997). Extrapolation of in vitro enzyme induc-
    tion data to humans In vivo. Chemico-Biological Interactions,
    107,109-121,
                                                      B-18

-------
1238
                                     Lipscomb et al.
15.   Snawder, J. E., & Lipscomb, J. C. (2000), Interindividual vari-      30,
     ance of cytochrome P450 forms in human hepatic microsomes:
     Correlation of individual forms with xenobiotic metabolism
     and implications in risk assessment. Regulatory Toxicology      31.
     and Pharmacology, 32,200-209.
16.   Arias, I, M., Popper, H., Schachter, D., & Shafritz,  D.  A.
     (1982). The Liver:  Biology  and Pathohiology. New  York:      32.
     Raven Press.
17.   Lipscomb, J. C., Garrett, C. M., & Snawder, J, E. (1997). Cy-
     tochrome P450-dcpendent metabolism of trichloroethylene:      33.
     Interindividual differences in humans.  Toxicology and Ap-
     plied  Pharmacology, 142, 311-318.
18.   Northrop, D. B. (1983), Fitting enzyme-kinetic data to V7K,      34.
     Analytical Biochemistry, 132,457-461.
19.   Gut, J., Catin, T., Dayer, P., Kronbach, X, Zanger, U., & Meyer.
     U. A, (1986). Debrisoquine/sparteine-type  polymorphism of
     drug oxidation: Purification and characterization of two func-      35.
     tionally different human liver cytochrome P-450 isozymes in-
     volved in impaired hydroxylation of the prototype substrate
     bufuralol. Journal of Biological Chemistry, 261,11734-11743.      36,
20.   Fisher, J. W. (2000). Physiologically based pharmacokinetic
     models for  trichloroethylene and its oxidative metabolites.
     Environmental Health Perspectives, 108,265-273,
21.   Bloemen. L. I, Monster, A. C., Kezic, S., Commandeur, X
     N.. Veulemans, H., Vermeulen, N, P., & Wilmer, J. W, (2001).      37.
     Study on the cytochrome P-450- and glutathione-dependent
     biotransformation of trichloroethylene in  humans.  Interna-
     tional Archives of Occupational and Environmental Health.
     74, 102-108.
22.   Nakajima, T., Okino, X., Okuyama, S,, Kaneko, X, Yonekura,      38.
     I.,  & Sato, A. (1988).  Ethanol-induccd  enhancement  of
     trichloroethylene metabolism and hepatotoxicity: Difference
     from  the effect of  phenobarbital.  Toxicology and  Applied
     Pharmacology, 94,227-237,                '                    39.
23.   Barton, H.  A., & Clewell, H. 1, HI. (2000), Evaluating the
     noneancer effects of triehloroethylene: Dosimctry, mode of
     action, and  risk assessment. Environmental Health Perspec-
     tives, 108,323-334.
24.   Bull,  R. I (2000). Mode of action of liver tumor induction      40.
     by trichloroethylene and its metabolites, trichioroacetate and
     dichloroacetate. Environmental health Perspectives, 108,241-
     259.
25.   Ikeda, M.,  Miyake, Y,, Ogata,  M., & Ohmori. S.  (1980),      41.
     Metabolism of trichloroethylene. Biochemical Pharmacology,
     29.2983-2992.
26.   Guengcrich, F. P, (1989). Analysis and characterization of en-
     zymes. In A.W. Hayes, (Ed.), Principles and Methods of Tox-
     icology (pp. 777-814). New York: Raven Press.
27.   Lipscomb, J, C,, Teuschler, L. K., Swartout, J. C, Strilcy. C. A.
     K, & Snawder, J. E. (2003). Variance of microsomal protein      42.
     and cytochrome P450 2E1 and 3A forms in adult human liver.
     Toxicology Mechanisms and Methods, 13, 45-51,
28.   Titterington, D. M., Smith, A. F, M., & Makov, U, E, (1985).
     Statistical Analysis of Finite Mixture Distributions. New York:      43.
     Wiley.
29.   Zhang, H., & Singer, B. (1999). Recursive Partitioning in  the
     Health Sciences. New York: Springer.
Breiman,  I,.,  Friedman.  J.,  Olshcn,  R,,  &  Stone, C.
(1984). Classification and Regression Trees.  Wadswoith, CA:
Wadsworlh Publishing.
Feller, W. (1968). An Introduction to Prohability Theory and
Its Applications Volume  1.3rd ed.,'(pp. 229-232). New York:
Wiley.
Allen, B. C..& Fisher, J, W. (1993). Pharraacokincticmodeling
of trichloroethylene and  trichloroacetic acid in humans. Risk
Analysis. 13, 71-86.
Houston, J. B., & Carlile, D. J. (1997). Prediction of hepatic
clearance from microsomes, hepaiocytcs, and liversliccs. Drug
Metabolism Review;;, 29,  891 -922.
American Conference of Governmental Industrial Hygienisls
(ACGIH). (2001). Threshold Limit Values for Chemical Sub-
stances and Physical Agents and Biological Exposure Indices,
Cincinnati, OH: ACGIH.
U.S.  EPA. (1987). National primary drinking water regula-
tions; synthetic organic chemicals: monitoring for unregulated
contaminants. Federal Register, 52(130). 25690.
Shimada, X, Tsumura, K, & Yamazaki, H. (1999). Prediction
of human liver microsomal oxidations of 7-elhoyxcoumarin
and chlorzoxazone with  kinetic parameters of reeombinant
cylochrome  P-450 enzymes.  Drug Metabolism and Disposi-
tion, 27. 1274-1280.
Reitz. R. H., Gargas, M. L.,  Mcndrala, A. L.. & Schumann,
A. M. (1996). In vivo and in vitro studies of  perehloroethy-
lene metabolism for physiologically based  pharmacokinetic
modeling in rats, mice and humans. Toxicology and Applied
Pharmacology. 136. 289-306.
Lipscomb, J. C.. Fisher,X W., Confer. P. D..&  Byc/kowski. J. Z.
(1998). In vitro to in vivo extrapolation for trichloroethylene
metabolism  in humans. Toxicologv and Applied Pharmacol-
ogy. 152, 376-387.
Carlile. D. I, Zomorodi, K., &  Houston, J. B.  (1997). Scaling
factors to relate drug metabolic clearance in  hepatic micro-
somes. isolated hepatocytes, and the intact liver: Studies with
induced livers involving dia/.epam. Drug Metabolism and Dis-
position. 25, 903-911.
Iyer.  K. R.. & Smz. M.W. (1999). Characterization of phase
I and phase II hepatic drug metabolism activities in a panel
of human liver preparations. Chemico-Biological Interactions.
118, 151-169.
Shimada,  T.,  Yamazaki,  H..  Mimura. M, Inui,  Y..  &
Gucngerich. F. P. (1994).  Interindividual variations in hu-
man liver cytochrome P450 enzymes involved in the oxida-
tion of drugs, carcinogens and toxic chemicals:  Studies with
liver microsomes of 30 Japanese and 30 Caucasians. Journal
of Pharmacology and Experimental Therapeutics, 270, 414—
423.
Kedderis, G. I... Carfagna, M. A., Held, S. D.. Batra, R.,  Mur-
phy, J, E., & Gargas, M, L.  (1993).  Kinetic  analysis of furan
biotransfortnation by F-344 rats in vivo and in vitro. Toxicol-
ogy and Applied Pharmacology. 123. 274-282.
Kedderis, G. L., & Held, S. D. (1996). Prediction of furan phar-
macokinelics from  hepatocytc studies: Comparison of bioac-
tivation and hepatic dosimelry in rats, mice, and humans. Tox-
icology and Applied Pharmacology, 14U. 124-130.
                                                          B-19

-------
USEPA Contract No. 3C-R102-NTEX                            Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
                                              APPENDIX C
     INFORMATION ON CHEMICAL MIXTURES AND THEIR COMPONENT CHEMICALS
                                                    C-l

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

                                            Introduction

   This appendix summarizes data on chemicals considered for examples in the present analysis.  These
   data can be used in a component-based approach to cumulative risk, particularly in developing an
   understanding of toxicokinetics relevant to determining the likelihood of an interaction. For the data
   contained in this appendix, several points bear emphasis.

    1.  No data on either of the two example chemical mixtures was located.  Thus, this report contains
       information only on each of 10 components chemicals (i.e., 6 component chemicals for Mixture 1
       and 4 component chemicals for Mixture 2).
    2.  This summary of data is by no means an exhaustive review for each of these 10 component
       chemicals.  Only brief information is provided for each of the sections for each chemical. We
       placed a great deal of emphasis on not repeating the effort of existing reviews; rather, we provide
       an update on each chemical with current information.
    3.  We focused more on pharmacokinetics/pharmacodynamics with special emphasis on
       physiologically-based pharmacokinetic  (PBPK) modeling.

   In this report, the toxicological and pharmacokinetic characteristics often chemicals are discussed. The
   ten chemicals consist of two groups that can  potentially form mixtures in drinking water. The first
   mixture consists of the organophosphorus pesticides parathion, methyl parathion, chlorpyrifos, diazinon,
   fenthion, and fenitrothion. The second mixture consists of the chlorinated hydrocarbons chloroform,
   trichloroethylene, tetrachloroethylene, and 1,1,1 trichloroethane. Each chemical is discussed separately
   in one section of the report. In each section,  a brief description of the toxicology of the chemical is
   provided as a background to the selection of  appropriate dose metrics for risk assessments that can be
   quantified using PBPK modeling.  Subsequently, available data describing the pharmacokinetics (PKs)
   of each chemical in laboratory animals and humans is provided.  Finally, available studies regarding
   potential PK interactions between the chemicals are provided.

   The studies incorporated in this review are necessarily limited.  The review is based on a detailed search
   of the open literature. However, inevitably there are additional studies to be considered, especially those
   that are not published.

   The principal purpose of this review is to compile data that may be useful in performing a PBPK model-
   based cumulative risk assessment (CRA)  for  the two groups of chemicals in drinking water.

   Literature Cited

   Yang, R. S. H. (2004). Final Report and Final Approach. A Report to USEPA under USEPA Contract
   No. 3C-R102-NTEX. May 24, 2004.
                                                C-2

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

                                            Parathion
   1.0    Introduction

   Parathion (O, O-diethyl 4-nitrophenyl phosphothioate) is a phosphorothionate insecticide that has no
   registered uses in the U.S. but is widely used elsewhere in agriculture and is present in food and
   environments (Brack etal. 1999; Fenske et al. 2002; Leblanc etal. 2000; Lifshitz etal. 1999; Ripley et
   al. 2000; Simcoxetal. 1995).

   1.1    Toxic effects

   Parathion exerts its toxicological effects via inhibition of acetylcholinesterase (Nigg and Knaak 2000;
   Thiermann etal. 1997). Metabolism of parathion exploits CYP450 as a metabolizing enzyme
   (Atterberry et al.  1997; Attia 2000; Attia etal. 1995; Besser etal. 1993; Halpert etal. 1980; Halpert and
   Neal 198la, b; Howard and Pope 2002; Jett etal. 1994; Katz etal. 1997). Other toxicities such as
   reproductive toxicities, immunotoxicity, cytotoxicity, carcinogenicity, and other effects have also been
   shown (Bustos-Obregon and Diaz 1999; Bustos-Obregon  etal. 2001; Cabello etal. 2001;  Cao etal.
   1999; Carlson and Ehrich 2001; Carlson etal. 2000; Galloway and Handy 2003; Grellner  and
   Glenewinkel 1997; Ivens etal. 1998; Levario-Carrillo etal. 2001; Li and Zhang 2001; Liu etal. 1999;
   Melendez Camargo and Lopez Hernandez 1998; Olivier et al.  200la; Olivier et al. 200Ib; Padungtod et
   al. 1999; Padungtod etal. 1998; Padungtod etal. 2000; Rojas  etal.  1998; Sal eh etal. 2003; Segura etal.
   1999; Selgrade etal. 1984; Senel etal. 2001; long etal. 1988; Undeger etal. 2000; VanDenBeukel et
   al. 1997; van den Beukel etal. 1998; Wagner et al. 2003;  Zaidi etal. 2000).

   1.2    Pharmacokinetics

   There are a number of pharmacokinetic studies of parathion and its toxic metabolite, paraoxon,
   conducted both in non-mammalian and mammalian species such as mice, rat, pig and dog via many
   routes of exposure including intravenous, oral and dermal exposure (Braeckman et al.  1983; Brimer et
   al. 1994; Chang etal.  1997; Chang etal. 1994a; Chang and Riviere 1991, 1993; Chang etal. 1994b;
   Denga et al. 1995; Eigenberg et al. 1983; Hurh et al. 2000a; Hurh et al. 2000b, c; Lessire et al. 1996;
   Oneto etal. 1995; Pena-Egido etal. 1988a; Pena-Egido etal.  1988b).

   Parathion at the dose of 3 mg/kg was intravenously administered to a rat. From the pharmacokinetic
   analysis, the terminal half-life and clearance of parathion were 3.4 hr and 93 ml/min/kg respectively
   (Eigenberg et al. 1983). Similar results were obtained from another study in rats, where the terminal
   half-life, AUC and clearance of parathion were 321 min.,  52.5 |j,g-min/mL and 57.1 ml/min/kg
   respectively (Hurh et al. 2000a; Hurh et al. 2000b, c). In these studies, paraoxon levels were lower than
   their detection limits.

   Parathion pharmacokinetics in dogs were somewhat different from that in rats. After 30 mg intravenous
   dosing, plasma clearance and terminal half-life were 21 ml/min and 8.5-11.2 hr respectively. The plasma
   clearance in dogs appeared to be less than one-third of the plasma clearance for the rats.

   1.2.1  Absorption

   In a pharmacokinetic study in dogs (Braeckman etal. 1983), parathion was administered at 5 mg/kg
   intravenously and 10 mg/kg orally to determine its absolute bioavialability (F). The fraction absorbed
   was high (57-98%). However, the bioavailability of parathion  appeared to have a comparatively large
   variation because of its first pass metabolism and intersubject  variation in parathion hepatic extraction

                                                C-3

-------
USEPA Contract No. 3C-R102-NTEX                         Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   ratio (range = 63-97%) (Braeckman et al. 1983).  Oral absorption of parathion was also studied in rats
   (Beubler et al. 1985).

   Numerous dermal exposure studies have been performed (Antunes-Madeira and Madeira 1984; Bucks et
   al. 1990; Campbell et al. 2000; Carver and Riviere 1989; Carver ef al. 1989; Fisher ef al. 1985; Gyrd-
   Hansen et al. 1993; Hawkins and Reifenrath 1986; Knaak et al. 1984; Murphy 1980; Qiao et al. 1996;
   Qiao etal. 1994; Reifenrath et al. 1984; Reifenrath et al. 1991; Riley and Kemppainen 1985; Shah and
   Guthrie 1983; Skinner and Kilgore 1982; Wester et al. 2000; Williams et al. 1990; Williams et al.
   1996). After dermal application of 50 mg/kg parathion was performed along the midline of the entire
   back of pigs, the dermal bioavailability (F) was 0.0993.  Tissue distribution of parathion in back skin,
   back fat, liver, kidney, muscle, adipose tissue was also determined. It appeared that 25.0-60.8% of the
   administered dose remained at the application site (Brimer et al.  1994).

   1.2.2  Distribution

   Protein binding in dog serum and in human serum  were 99% and 98% respectively (Braeckman et al.
   1983). Tissue distributions were also reported in some species (Brimer et al. 1994).

   1.2.3  Metabolism

   Parathion is metabolized into paraoxon and 4-nitrophenol by desulfuration  and dearylation (Fig. 1),
   respectively. 4-Nitrophenol formation is considered as the inactivation pathway, whereas paraoxon
   formation is considered as an activation pathway (Benke GM 1975; Bulusu and Chakravarty 1986,
   1988; Butler and Murray 1993, 1997; Chambers and Forsyth 1989; Chambers etal. 1994; Chaturvedi et
   al. 1991; Contreras et al. 1999; Halpert et al. 1980; Halpert and Neal 198la, b; Hou et al. 1996;
   Kulkarni and Hodgson 1982; Kuo and Perera 2000; Lapadula et al. 1984; Levi and Hodgson 1985;
   Martinez-Zedillo et al.  1979; Monnet-Tschudi  et al. 2000; Morgan et al. 1994; Murray and Butler 1994,
   1995; Mutch etal. 1999; Mutch etal. 2003; Nadin and Murray 1999; Pond etal. 1995; Pond etal.  1998;
   Purshottam and Srivastava 1987; Ramos and Sultatos 1998; Rowland et al. 1991; Soranno and Sultatos
   1992; Sultatos 1986;  Sultatos etal. 1984; Sultatos and Gagliardi  1990; Sultatos and Minor 1986;
   Sultatos etal. 1985; Sultatos and Murphy 1983; Tang and Chambers  1999; Vargas Loza etal.  1997;
   Venera et al. 1978; Vitarius et al. 1995; Wallace  and Dargan 1987; Watson et al.  1994; Zhang and
   Sultatos 1991; Zhu and Liu 1994). The primary metabolizing organ is the liver by the enzyme
   cytochrome P450 3A4.

   In mouse liver microsomes, the apparent Km's for the formation of paraoxon and p-nitrophenol were
   29.6 and 26.5 |jM respectively, and the apparent Vmaxs were 5.8 and 6.7 nmols/100 mg liver/min
   respectively (Sultatos 1986; Sultatos etal. 1984;  Sultatos and Gagliardi 1990; Sultatos and Minor 1986;
   Sultatos et al. 1985; Sultatos and Murphy 1983).

   In rat liver microsomes, the kinetic curve for desulfuration of parathion is baphasic with apparent Km's
   of 0.23 and 71.3 |jM and Vmaxs of 3.62 and 4.56 nM/min/mg protein. For the dearylation reaction,
   parathion has an apparent Km and Vmax of 56 |jM and  1.49 nM/min/mg protein,  respectively (Ma and
   Chambers 1994, 1995).
                                                C-4

-------
USEPA Contract No. 3C-R102-NTEX
                                                  Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
        C2H50S
                          No2
        C2H50X||
-N02           p_0
        C2H50X
                              GS-C2H5
                                         Oxidation   C2H5<\ II
                                     NO2 	*~       P-O-
                                           P450
                                                       Paraoxon
      C2H5Cf
     Phosphorothioic Acid
                                       Para-Nitrophenol
C2H5cf
 Phosphoric Acid
                  Figure 1. Metabolic pathway of parathion (adapted from Benke 1975).
   In a human microsomal study, parathion demonstrated biphasic behaviors in both individual microsomal
   and pooled samples. Its apparent Knii and Km2 in individual microsomes were 0.30 and 165.5 |jM and
   Vmaxi and Vmax2 were 290 and 821 pmol oxon/mg protein/min respectively. In pooled liver
   microsomes, parathion's apparent Knii and Km2 were 9.0 and 69.6 |jM and Vmaxi and Vmax2 were
   106.6 and 2,478 pmol oxon/mg protein/min, respectively (Buratti etal. 2003; Ma and Chambers 1994,
   1995). Another study in human liver microsomes indicated that CYP3A4 is the major enzyme
   responsible for catalyzing parathion oxidation to paraoxon (Butler and Murray 1993, 1997).

   In addition to the liver, the brain is also capable of metabolizing parathion in various regions such as
   cortex, olfactory bulb/hypothalamus, striatum, cerebellum, midbrain, medulla and pons, and
   hippocampus. However, the total activity appears to be highest in the cortex (Soranno and Sultatos
   1992).

   1.2.4  Pharmacokinetic studies in special population

   There have been a number of pharmacokinetic studies in specific populations (Benjaminov etal. 1992;
   Jaramillo and Reyes 1990). Neilsen et al. conducted a pharmacokinetic study in neonatal and young
   pigs. Intravenous parathion (0.5 mg/kg) was administered to newborn, 1 week and 8 weeks old piglets.
   The total body clearance was 7, 35 and 121 ml/min/kg, respectively. Tissue distribution in all groups
   was also presented. Interestingly, the newborn piglets seemed to retain parathion in significant amounts
   in many organs such as the liver, lung, brain, heart and muscle indicating that reduced total body
   clearance in the newborn markedly influenced tissue distribution (Nielsen et al. 1991).

   Pregnancy also affects parathion disposition and its toxicity. Concentrations of parathion were
   significantly higher in blood and brain of pregnant mice at most times after administration (5 mg of
   parathion/kg) of parathion when compared to non-pregnant mice (Weitman etal. 1983, 1986a, b).
                                                 C-5

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   1.3    Pharmacokinetic interaction between parathion and other compounds

   Due to its metabolic pathway via CYP450, there are many possibilities that parathion pharmacokinetics
   may be affected by certain compounds particularly drugs, CYP450 inducers, inhibitors, other
   environmental pollutants and foods (Agyeman and Sultatos 1998;  Carr etal. 2002; Chakravarty and
   Sreedhar 1982; Costa and Murphy 1984; Delaunois etal. 1999; Gelal etal. 2001; Graziano etal. 1985;
   Guilhermino etal. 1998a, b; Hurh etal. 2000a; Hurh etal. 2000b,  c; Joshi and Thornburg 1986;  Karanth
   etal. 2001; Miranda etal. 1998; Mourelle etal. 1986; Murphy 1980; O'Shaughnessy and Sultatos 1995;
   Purshottam and Kaveeshwar 1982; Purshottam and Srivastava 1984; Ramos and Sultatos 1998;
   Sawahata and Neal 1982; Siller et al. 1997; Wester et al. 2000).

   Cimetidine, a non-specific inhibitor of CYP450, was capable of antagonizing methyl parathion toxicity
   but failed to decrease parathion-induced toxicity in mice and rats (Weitman et al. 1983). Rats pretreated
   with dexamethasone, a specific inducer of CYP3A23, showed faster clearance of parathion than control
   rats (Hurh et al. 2000a).

   1.4    PBPK modeling of parathion and Monte Carlo simulation

   A PBPK model of parathion was developed by Gearhart et al. In brief, the model describes the
   metabolism of parathion to paraoxon by the liver, the inhibition of acetylcholinesterase,
   butyrylcholinesterase, and carboxylesterase by paraoxon in the brain, liver, kidneys, rapidly perfused
   tissues and the  arterial and venous blood (Gearhart 1994). Physiologic parameters are available in the
   literature. (Jepson etal.  1994; Kousba and  Sultatos 2002). Due to the existence of paraoxonase
   polymorphisms (Costa et al. 2003; Diepgen and Geldmacher-von Mallinckrodt 1986; Eaton 2000;
   Furlong et al. 2000; Haber et al. 2002; Laplaud et al. 1998; Lee et al. 2003; Shih et al. 1998), a
   subsequent study was conducted by using this existing PBPK model with Monte Carlo simulation to
   elaborate the effect of polymorphic paraoxonase (PON1) on its toxicity (Gentry et al. 2002).

   1.5    Literature Cited

   Agyeman, A. A., and  Sultatos, L. G. (1998). The actions of the H2-blocker cimetidine on the toxicity
      and biotransformation of the phosphorothioate insecticide parathion. Toxicology 128, 207-18.
   Antunes-Madeira, M.  C., and Madeira, V. M. (1984). Partition of parathion in synthetic and native
      membranes. Biochim Biophys Acta 778, 49-56.
   Atterberry, T. T., Burnett, W. T., and Chambers, J. E. (1997). Age-related differences in parathion and
      chlorpyrifos toxicity in male rats: target and nontarget esterase sensitivity and cytochrome P450-
      mediated metabolism. Toxicol ApplPharmacol 147, 411-8.
   Attia, A. M. (2000). Possible involvement of beta-adrenergic receptors in the enhancement of nocturnal
      pineal N-acetyltransferase activity due to parathion administration. Toxicology  142, 79-86.
   Attia, A. M., Mostafa, M. H., Richardson, B. A., and Reiter, R. J. (1995). Changes  in nocturnal pineal
      indoleamine metabolism in rats treated with parathion are prevented by beta-adrenergic antagonist
      administration. Toxicology 97', 183-9.
   Benjaminov, O., Hoffer, E., Taitelman, U., Urbach, J., and Brandes, J. M.  (1992). Parathion transfer and
      acetylcholinesterase activity in an in-vitro perfused term human placenta. Vet Hum Toxicol 34, 10-2.
   Benke GM, M. S. (1975). The influence of age on the toxicity and metabolism of methyl parathion and
      parathion in male and female rats. Toxicol Appl Pharmacol. 31, 254-69.
   Besser, R., Gutmann, L., and Weilmann, L. S. (1993). Polyneuropathy following parathion poisoning. J
      NeurolNeurosurgPsychiatry 56, 1135-6.
                                                C-6

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

  Beubler, E., Dirnhofer, R., and Ranner, G. (1985). Parathion and gastrointestinal transit in the rat. Arch
      Toxicol 57, 72-3.
  Brack, W., Altenburger, R., Ensenbach, U., Moder, M., Segner, H., and Schuurmann, G. (1999).
      Bioassay-directed identification of organic toxicants in river sediment in the industrial region of
      bitterfeld (Germany)-A contribution to hazard assessment. Arch Environ Contam Toxicol 37, 164-
      74.
  Braeckman, R. A., Audenaert, F., Willems, J. L., Belpaire, F. M., and Bogaert, M. G. (1983).
      Toxicokinetics of methyl parathion and parathion in the dog after intravenous and oral
      administration. Arch Toxicol 54, 71-82.
  Brimer, L., Gyrd-Hansen, N., and Rasmussen, F. (1994). Disposition of parathion after dermal
      application in pigs. J Vet Pharmacol Ther 17, 304-8.
  Bucks, D. A., Hinz, R. S., Sarason, R., Maibach, H. I, and Guy, R. H. (1990). In vivo percutaneous
      absorption of chemicals: a multiple dose study in rhesus monkeys. FoodChem Toxicol28, 129-32.
  Bulusu, S., and Chakravarty, I. (1986). Subacute administration of organophosphorus pesticides and
      hepatic drug metabolizing enzyme activity in normal and malnourished rats. Bull Environ Contam
      Toxicol 36, 73-80.
  Bulusu, S., and Chakravarty, I. (1988). Profile of drug metabolizing enzymes in rats treated with
      parathion, malathion, and phosalone under various conditions of protein energy malnutrition. Bull
      Environ Contam Toxicol 40, 110-8.
  Buratti, F. M., Volpe, M. T., Meneguz, A., Vittozzi, L., and Testai, E. (2003). CYP-specific
      bioactivation of four organophosphorothioate pesticides by human liver microsomes. Toxicol Appl
      Pharmacol 186, 143-54.
  Bustos-Obregon, E., and Diaz, O. (1999). Ultrastructure of mouse teratozoospermia induced by
      parathion. Asian JAndrol 1, 79-80.
  Bustos-Obregon, E., Diaz, O., and Sobarzo, C. (2001). Parathion induces mouse germ cells apoptosis.
      Ml JAnat Embryol 106, 199-204.
  Butler, A. M., and Murray, M. (1993). Inhibition and inactivation of constitutive cytochromes P450 in
      rat liver by parathion. Mol Pharmacol 43, 902-8.
  Butler, A. M., and Murray, M. (1997). Biotransformation of parathion in human liver: participation of
      CYP3A4 and its inactivation during microsomal parathion oxidation. JPharmacol Exp Ther 280,
      966-73.
  Cabello, G., Valenzuela, M., Vilaxa, A., Duran, V., Rudolph, I, Hrepic, N., and Calaf, G. (2001). A rat
      mammary tumor model induced by the organophosphorous pesticides parathion  and malathion,
      possibly through acetylcholinesterase inhibition. Environ Health Perspect 109, 471-9.
  Campbell, J.  L., Smith, M. A., Eiteman, M. A., Williams, P. L., and Boeniger, M. F. (2000).
      Comparison of solvents for removing pesticides from skin using an in vitro porcine model. Aihaj 61,
      82-8.
  Cao, C. J., Mioduszewski, R. J., Menking, D. E., Valdes, J. J., Katz, E. J., Eldefrawi, M. E., and
      Eldefrawi, A. T. (1999). Cytotoxicity of organophosphate anticholinesterases. In Vitro Cell Dev Biol
      Anim 35, 493-500.
  Carlson, K., and Ehrich, M. (2001). Organophosphorus compounds alter intracellular F-actin content in
      SH-SY5Y human neuroblastoma cells. Neurotoxicology 22, 819-27.
  Carlson, K., Jortner, B. S., and Ehrich, M. (2000). Organophosphorus compound-induced apoptosis in
      SH-SY5Y human neuroblastoma cells. Toxicol Appl Pharmacol 168, 102-13.
  Carr, R. L., Richardson, J. R., Guarisco, J. A., Kachroo, A., Chambers, J. E., Couch, T. A., Durunna, G.
      C., and Meek, E. C. (2002). Effects of PCB exposure on the toxic impact of organophosphorus
      insecticides. Toxicol Sci 67, 311-21.
  Carver, M. P., and Riviere, J. E. (1989). Percutaneous absorption and excretion of xenobiotics after
      topical  and intravenous administration to pigs. Fundam Appl Toxicol 13, 714-22.

                                                C-7

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Carver, M. P., Williams, P. L., and Riviere, J. E. (1989). The isolated perfused porcine skin flap. III.
      Percutaneous absorption pharmacokinetics of organophosphates, steroids, benzoic acid, and caffeine.
      Toxicol Appl Pharmacol 97', 324-37.
   Chakravarty, L, and Sreedhar, R. (1982). Interaction between parathion toxicity and protein
      malnutrition. Environ Res 27, 179-84.
   Chambers, J. E., and Forsyth, C. S. (1989). Lack of inducibility of brain monooxygenase activities
      including parathion desulfuration. JBiochem Toxicol 4, 65-70.
   Chambers, J. E., Ma, T., Boone, J. S., and Chambers, H. W. (1994). Role of detoxication pathways in
      acute toxicity levels of phosphorothionate insecticides in the rat. Life Sci 54, 1357-64.
   Chang, M. J., Chen, Y. C., and Yang, H. J. (1997). Comparative evaluation on the biological monitoring
      of exposure to parathion and its methyl analog. Arch Environ Contain Toxicol 32, 422-5.
   Chang, S. K., Brownie, C., and Riviere, J. E. (1994a). Percutaneous absorption of topical parathion
      through porcine skin: in vitro studies on the effect of environmental perturbations. J Vet Pharmacol
      Ther 17, 434-9.
   Chang, S. K., and Riviere, J. E. (1991). Percutaneous absorption of parathion in vitro in porcine skin:
      effects of dose, temperature, humidity, and perfusate composition on absorptive flux. Fundam Appl
      Toxicol 17, 494-504.
   Chang, S. K., and Riviere, J. E. (1993). Effect of humidity and occlusion on the percutaneous absorption
      of parathion in vitro. Pharm Res 10, 152-5.
   Chang, S. K., Williams, P. L., Dauterman, W. C., and Riviere, J. E. (1994b). Percutaneous absorption,
      dermatopharmacokinetics and related bio-transformation studies of carbaryl, lindane, malathion, and
      parathion in isolated perfused porcine skin. Toxicology 91, 269-80.
   Chaturvedi, A. K., Kuntz, D. J., and Rao, N. G. (1991). Metabolic aspects of the toxicology of mixtures
      of parathion, toxaphene and/or 2,4-D in mice. J Appl Toxicol 11, 245-51.
   Contreras, H. R., Badilla, J., and Bustos-Obregon, E. (1999). Morphofunctional disturbances of human
      sperm after incubation with organophosphorate pesticides. Biocell23, 135-41.
   Costa, L. G., and Murphy, S. D. (1984). Interaction between acetaminophen and organophosphates in
      mice. Res Commun Chem PatholPharmacol 44, 389-400.
   Costa, L. G., Richter, R. J., Li, W. F., Cole, T., Guizzetti, M., and Furlong, C. E. (2003). Paraoxonase
      (PON 1) as a biomarker of susceptibility for organophosphate toxicity. Biomarkers 8,  1-12.
   Delaunois, A., Florquin, S., Segura, P., Montano, L.  M., Vargas, M. H., and Gustin, P.  (1999).
      Interactions between cytochrome P-450 activities and ozone-induced modulatory effects on
      endothelial permeability in rabbit lungs: influence of gender. Inhal Toxicol 11, 999-1014.
   Denga, N., Moldeus, P., Kasilo, O. M., and Nhachi, C. F. (1995). Use of urinary p-nitrophenol as an
      index of exposure to parathion. Bull Environ Contain Toxicol 55, 296-302.
   Diepgen, T. L., and Geldmacher-von Mallinckrodt, M. (1986). Interethnic differences in the
      detoxification of organophosphates: the human serum paraoxonase polymorphism.  Arch Toxicol
      Suppl9, 154-8.
   Eaton, D. L. (2000). Biotransformation enzyme polymorphism and pesticide susceptibility.
      Neurotoxicology 21, 101-11.
   Eigenberg, D. A., Pazdernik, T. L., and Doull, J. (1983). Hemoperfusion and pharmacokinetic studies
      with parathion and paraoxon in the rat and dog. DrugMetab Dispos 11, 366-70.
   Fenske, R. A., Lu, C., Barr, D., and Needham, L. (2002). Children's exposure to chlorpyrifos and
      parathion in an agricultural community in central Washington State. Environ Health Perspect 110,
      549-53.
   Fisher, H. L., Most, B., and Hall, L. L. (1985). Dermal absorption of pesticides calculated by
      deconvolution. J Appl Toxicol 5,  163-77.
   Furlong, C. E., Li, W. F., Brophy, V. H., Jarvik, G. P., Richter, R. J., Shih, D. M., Lusis, A. J., and
      Costa, L. G. (2000). The PON1 gene and detoxication. Neurotoxicology 21, 581-7.

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Galloway, T., and Handy, R. (2003). Immunotoxicity of organophosphorous pesticides. Ecotoxicology
      12, 345-63.
   Gearhart JM, J. G., Clewell HJ, Andersen ME, Conolly RB (1994). Physiologically based
      pharmacokinetic model for the inhibition of acetylcholinesterase by organophosphate esters. Environ
      Health Per sped. 102, 51-60.
   Gelal, A., Gumustekin, M., Kalkan, S., Guven, H., and Eminoglu, O. (2001). Effects of subchronic
      parathion exposure on cyclosporine pharmacokinetics in rats. J Toxicol Environ Health A 62,  289-
      94.
   Gentry, P. R., Hack, C. E., Haber, L., Maier, A., and Clewell, H. J., 3rd (2002). An approach for the
      quantitative consideration of genetic polymorphism data in chemical risk assessment: examples with
      warfarin and parathion.  Toxicol Sci 70, 120-39.
   Graziano, M. J., Gairola, C., and Borough, H. W. (1985). Effects of cigarette smoke and dietary vitamin
      E levels on selected lung and hepatic biotransformation enzymes in mice. DrugNutr Interact 3, 213-
      22.
   Grellner, W., and Glenewinkel, F. (1997). Exhumations: synopsis of morphological and toxicological
      findings in relation to the postmortem interval. Survey on a 20-year period and review  of the
      literature. Forensic Sci Int 90,  139-59.
   Guilhermino, L., Scares, A. M., Carvalho, A. P., and Lopes, M. C. (1998a). Correlation between whole
      blood cholinesterase activity and cerebral cortex cholinesterase activity in rats treated with
      parathion. Chemosphere 37, 1385-93.
   Guilhermino, L., Scares, A. M., Carvalho, A. P., and Lopes, M. C. (1998b). Effects of cadmium and
      parathion exposure on hematology and blood biochemistry of adult male rats. Bull Environ Contain
      Toxicol 60, 52-9.
   Gyrd-Hansen, N., Brimer, L., and Rasmussen, F. (1993). Percutaneous absorption of organophosphorus
      insecticides in pigs—the influence of different vehicles. J Vet Pharmacol Ther 16, 174-80.
   Haber, L. T., Maier, A., Gentry, P. R., Clewell, H. J., and Dourson, M. L. (2002). Genetic
      polymorphisms in assessing interindividual variability in delivered dose. Regul Toxicol Pharmacol
      35, 177-97.
   Halpert, J., Hammond, D., and Neal, R. A. (1980). Inactivation of purified rat liver cytochrome P-450
      during the metabolism of parathion (diethyl p-nitrophenyl phosphorothionate). J Biol Chem 255,
      1080-9.
   Halpert, J., and Neal, R. A.  (1981a). Formation and fate of reactive intermediates of parathion.  Adv Exp
      MedBiol 136 PtB, 1037-50.
   Halpert, J., and Neal, R. A.  (1981b). Inactivation of rat liver cytochrome P-450 by the suicide substrates
      parathion and chloramphenicol. DrugMetab Rev 12, 239-59.
   Hawkins, G. S., and Reifenrath, W. G. (1986). Influence of skin source, penetration cell fluid, and
      partition coefficient on in vitro skin penetration. JPharm Sci 75, 378-81.
   Hou, X., Maser, R. L., Magenheimer, B.  S., and Calvet, J. P. (1996). A mouse kidney- and liver-
      expressed cDNA having homology with a prokaryotic parathion hydrolase (phosphotriesterase)-
      encoding gene: abnormal expression in injured and polycystic kidneys. Gene 168, 157-63.
   Howard, M. D., and Pope, C. N. (2002). In vitro effects of chlorpyrifos, parathion, methyl  parathion and
      their oxons on cardiac muscarinic receptor binding in neonatal and adult rats. Toxicology 170, 1-10.
   Hurh, E., Lee, E., Lee, A., Kim, Y., Kim,  S., and Lee, M. (2000a). Effects of enzyme inducers  or
      inhibitors on the pharmacokinetics of intravenous parathion in rats. Biopharm Drug Dispos 21, 193-
      204.
   Hurh, E., Lee, E. J., Kim, Y. G, Kim, S. Y., Kim, S. H., Kim, Y. C., and Lee, M. G.  (2000b). Effects of
      neostigmine on the pharmacokinetics of intravenous parathion in rats. Res CommunMolPathol
      Pharmacol 108, 261-73.
                                                C-9

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Hurh, E., Lee, E. J., Kim, Y. G., Kim, S. Y., Kim, S. H., Kim, Y. C., and Lee, M. G. (2000c). Effects of
      physostigmine on the pharmacokinetics of intravenous parathion in rats. Biopharm DrugDispos 21,
      331-8.
   Ivens, I. A., Schmuck, G., and Machemer, L. (1998). Learning and memory of rats after long-term
      administration of low doses of parathion. Toxicol Sci 46, 101-11.
   Jaramillo, F., and Reyes, J. L. (1990). Intrauterine exposure to parathion increases its disposition rate in
      postnatal life. Biol Neonate 57, 200-6.
   Jepson, G. W., Hoover, D. K., Black, R. K., McCafferty, J. D., Mahle, D.  A., and Gearhart, J. M. (1994).
      A partition coefficient determination method for nonvolatile chemicals in biological tissues. Fundam
      Appl Toxicol 22, 519-24.
   Jett, D. A., Fernando, J. C., Eldefrawi, M. E., and Eldefrawi, A. T. (1994). Differential regulation of
      muscarinic receptor subtypes in rat brain regions by repeated injections of parathion. Toxicol Lett 73,
      33-41.
   Joshi, U. M., and Thornburg, J. E. (1986). Interactions between cimetidine, methylparathion, and
      parathion. J Toxicol Environ Health 19, 337-44.
   Karanth, S., Olivier, K., Jr., Liu, J., and Pope, C. (2001).  In vivo interaction between chlorpyrifos and
      parathion in adult rats: sequence of administration can markedly influence toxic outcome. Toxicol
      ApplPharmacol 111, 247-55.
   Katz, E. J., Cortes, V. L, Eldefrawi, M. E., and Eldefrawi, A. T. (1997). Chlorpyrifos,  parathion, and
      their oxons bind to and desensitize a nicotinic acetylcholine receptor: relevance to their toxicities.
      Toxicol Appl Pharmacol 146, 227-36.
   Knaak, J. B., Yee, K., Ackerman,  C. R., Zweig, G., Fry, D. M., and Wilson, B. W. (1984). Percutaneous
      absorption and dermal dose-cholinesterase response studies with parathion and carbaryl in the rat.
      Toxicol Appl Pharmacol 76, 252-63.
   Kousba, A., and Sultatos, L. G. (2002). Continuous system modeling of equilibrium dialysis for
      determinations of tissue partitioning of parathion and paraoxon. Toxicol Lett 133, 153-9.
   Kulkarni, A. P., and Hodgson, E. (1982). Mouse liver microsomal hexose-6-phosphate dehydrogenase.
      NADPH generation and utilization in monooxygenation reactions. Biochem Pharmacol 31,  1131-7.
   Kuo, L. Y., and Perera, N. M. (2000). Paraoxon and parathion hydrolysis by aqueous molybdenocene
      dichloride (Cp2MoC12): first reported pesticide hydrolysis by an organometallic complex. Inorg
      Chem 39, 2103-6.
   Lapadula, D. M., Carrington, C. D., and Abou-Donia, M. B. (1984). Induction of hepatic microsomal
      cytochrome P-450 and inhibition of brain, liver, and plasma esterases by an acute dose of S,S,S-tri-
      n-butyl phosphorotrithioate (DEF) in the adult hen. Toxicol Appl Pharmacol 73, 300-10.
   Laplaud, P. M., Dantoine, T., and Chapman, M. J. (1998). Paraoxonase as a risk marker for
      cardiovascular disease: facts and hypotheses. Clin Chem LabMedl>6, 431-41.
   Leblanc, J. C., Malmauret, L., Guerin, T., Bordet, F., Boursier, B., and Verger, P. (2000). Estimation of
      the dietary intake of pesticide residues, lead, cadmium, arsenic and radionuclides in France. Food
      AdditContam 17, 925-32.
   Lee, B. W., London, L., Paulauskis, J., Myers, J., and Christiani, D. C. (2003). Association between
      human paraoxonase gene polymorphism and chronic  symptoms in pesticide-exposed workers. J
      Occup Environ Med 45, 118-22.
   Lessire, F., Gustin, P., Delaunois,  A., Bloden, S., Nemmar, A., Vargas, M., and Ansay, M. (1996).
      Relationship between parathion and paraoxon toxicokinetics, lung metabolic activity, and
      cholinesterase inhibition in guinea pig and rabbit lungs. Toxicol Appl Pharmacol 138, 201-10.
   Levario-Carrillo, M., Feria-Velasco, A., De Celis, R., Ramos-Martinez, E., Cordova-Fierro, L., and
      Solis, F. J. (2001). Parathion, a cholinesterase-inhibiting plaguicide induces changes in tertiary villi
      of placenta of women exposed: a scanning electron microscopy study. Gynecol Obstet Invest 52,
      269-75.

                                                C-10

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Levi, P. E., and Hodgson, E. (1985). Oxidation of pesticides by purified cytochrome P-450 isozymes
      from mouse liver. Toxicol Lett 24, 221-8.
   Li, H., and Zhang, S. (2001). In vitro cytotoxicity of the organophosphorus pesticide parathion to FG-
      9307 cells. Toxicol In Vitro 15, 643-7.
   Lifshitz, M., Shahak, E.,  and Sofer, S. (1999). Carbamate and organophosphate poisoning in young
      children. Pediatr Emerg Care 15, 102-3.
   Liu, J., Olivier, K., and Pope, C. N. (1999). Comparative neurochemical effects of repeated methyl
      parathion or chlorpyrifos exposures in neonatal and adult rats. Toxicol ApplPharmacol 158, 186-96.
   Ma, T., and Chambers, J. E. (1994). Kinetic parameters of desulfuration and dearylation of parathion
      and chlorpyrifos by rat liver microsomes. Food Chem Toxicol 32, 763-7.
   Ma, T., and Chambers, J. E. (1995). A kinetic analysis of hepatic microsomal activation of parathion and
      chlorpyrifos in control and phenobarbital-treated rats. JBiochem Toxicol 10, 63-8.
   Martinez-Zedillo, G., Castilho-Alonso,  C., Magdaleno, V.  M., and Gonzalez-Angulo, A. (1979).
      Cytochrome P-450 and parathion metabolism in the fetal and adult gonads of the horse. Life Sci 25,
      327-32.
   Melendez Camargo, M. A., and Lopez Hernandez, S. (1998). Effect of cadmium and parathion on renal
      function in rat. Proc West Pharmacol Soc 41, 65-7.
   Miranda, C. L., Henderson, M. C., and Buhler, D. R. (1998). Evaluation of chemicals as inhibitors of
      trout cytochrome P450s. Toxicol Appl Pharmacol 148, 237-44.
   Monnet-Tschudi, F., Zurich, M. G., Schilter, B., Costa, L.  G., and Honegger, P. (2000). Maturation-
      dependent effects of chlorpyrifos and parathion and their oxygen analogs on acetylcholinesterase and
      neuronal and glial markers in aggregating brain cell cultures. Toxicol Appl Pharmacol 165, 175-83.
   Morgan, E. W., Yan, B.,  Greenway, D., Petersen, D. R., and Parkinson, A. (1994). Purification and
      characterization of two  rat liver microsomal carboxylesterases (hydrolase A and B). Arch Biochem
      Biophys 315, 495-512.
   Mourelle, M., Giron, E., Amezcua, J. L., and Martinez-Tabche, L. (1986). Cimetidine enhances and
      phenobarbital decreases parathion toxicity. J Appl Toxicol 6, 401-4.
   Murphy, S. D. (1980). Toxic interactions with dermal exposure to organophosphate insecticides. Dev
      Toxicol Environ Sci 8, 615-21.
   Murray, M., and Butler, A.  M. (1994). Hepatic biotransformation of parathion: role of cytochrome P450
      in NADPH- and NADH-mediated microsomal oxidation in vitro. Chem Res Toxicol 7, 792-9.
   Murray, M., and Butler, A.  M. (1995). Identification of a reversible component in the in vitro inhibition
      of rat hepatic cytochrome P450 2B1 by parathion. JPharmacol Exp Ther 272, 639-44.
   Mutch, E., Blain, P. G., and Williams, F. M. (1999). The role of metabolism in determining
      susceptibility to parathion toxicity in man. Toxicol Lett 107, 177-87.
   Mutch, E., Daly, A. K., Leathart, J. B., Blain, P. G., and Williams, F. M. (2003). Do multiple
      cytochrome P450 isoforms contribute to parathion metabolism in man? Arch Toxicol 77, 313-20.
   Nadin, L., and Murray, M. (1999). Participation of CYP2C8 in retinoic acid 4-hydroxylation in human
      hepatic microsomes. Biochem Pharmacol 58, 1201-8.
   Nielsen, P., Friis, C., Gyrd-Hansen, N.,  and Kraul, I. (1991). Disposition of parathion in neonatal and
      young pigs. Pharmacol Toxicol 69,  233-7.
   Nigg, H. N., and Knaak, J. B. (2000). Blood cholinesterases as human biomarkers of organophosphorus
      pesticide exposure. Rev Environ Contam Toxicol 163, 29-111.
   Olivier, K., Jr., Liu, J., and Pope, C. (2001a). Inhibition of forskolin-stimulated cAMP formation in vitro
      by paraoxon and chlorpyrifos oxon in cortical slices from  neonatal, juvenile, and adult rats. J
      Biochem Mol Toxicol 15, 263-9.
   Olivier, K., Liu, J., Karanth, S., Zhang,  H., Roane, D. S., and Pope, C. N. (2001b). Glucose feeding
      exacerbates parathion-induced neurotoxicity. J Toxicol Environ Health A 63, 253-71.
                                                C-ll

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Oneto, M. L., Basack, S. B., and Kesten, E. M. (1995). Total and conjugated urinary paranitrophenol
      after an acute parathion ingestion. Sci Justice 35, 207-11.
   O'Shaughnessy, J. A., and Sultatos, L. G. (1995). Interaction of ethanol and the organophosphorus
      insecticide parathion. Biochem Pharmacol 50, 1925-32.
   Padungtod, C., Hassold, T. J., Millie, E., Ryan, L. M., Savitz, D. A., Christian!, D. C., and Xu, X.
      (1999). Sperm aneuploidy among Chinese pesticide factory workers: scoring by the FISH method.
      AmJIndMed 36, 230-8.
   Padungtod, C., Lasley, B. L., Christiani, D. C., Ryan, L. M., and Xu, X. (1998). Reproductive hormone
      profile among pesticide factory workers. J Occup Environ Med 40, 1038-47.
   Padungtod, C., Savitz, D. A., Overstreet, J. W., Christiani, D. C., Ryan, L. M., and Xu, X. (2000).
      Occupational pesticide exposure and semen quality among Chinese workers. J Occup Environ Med
      42, 982-92.
   Pena-Egido, M. J., Marino-Hernandez, E. L., Santos-Buelga, C., and Rivas-Gonzalo, J. C. (1988a).
      Urinary excretion kinetics of p-nitrophenol following oral administration of parathion in the rabbit.
      ArchToxicol 62, 351-4.
   Pena-Egido, M. J., Rivas-Gonzalo, J. C., and Marino-Hernandez, E. L. (1988b). Toxicokinetics of
      parathion in the rabbit. Arch Toxicol 61, 196-200.
   Pond, A. L., Chambers, H. W., and Chambers, J. E. (1995). Organophosphate detoxication potential of
      various rat tissues via A-esterase and aliesterase activities. Toxicol Lett 78, 245-52.
   Pond, A. L., Chambers, H. W., Coyne, C. P., and Chambers, J. E. (1998). Purification of two rat hepatic
      proteins with A-esterase activity toward chlorpyrifos-oxon and paraoxon. JPharmacol Exp Ther
      286, 1404-11.
   Purshottam, T., and Kaveeshwar, U. (1982). Effect of phenobarbitol pretreatment on regeneration of
      plasma cholinesterase activity inhibited by parathion or dichlorovos. Arch Environ Health 37, 53-8.
   Purshottam, T., and Srivastava, R. K. (1984). Effect of high-fat and high-protein diets on toxicity of
      parathion and dichlorvos. Arch Environ Health 39, 425-30.
   Purshottam, T., and Srivastava, R. K. (1987). Parathion toxicity in relation to liver microsomal oxidases,
      lipid composition and fluidity. Pharmacology 35, 227-33.
   Qiao, G. L., Brooks, J. D., Baynes, R. E., Monteiro-Riviere, N. A., Williams, P. L., and Riviere, J. E.
      (1996). The use of mechanistically defined chemical mixtures (MDCM) to assess component effects
      on the percutaneous absorption and cutaneous disposition of topically exposed chemicals. I.  Studies
      with parathion mixtures in isolated perfused porcine skin. Toxicol ApplPharmacol 141, 473-86.
   Qiao, G. L., Williams, P. L., and Riviere, J. E. (1994). Percutaneous absorption, biotransformation, and
      systemic disposition of parathion in vivo in swine. I. Comprehensive pharmacokinetic model. Drug
      Metab Dispos 22, 459-71.
   Ramos, S., and Sultatos, L. (1998). Flavonoid-induced alterations in cytochrome P450-dependent
      biotransformation of the organophosphorus insecticide parathion in the mouse. Toxicology 131, 155-
      67.
   Reifenrath, W.  G., Chellquist, E. M., Shipwash, E. A., Jederberg, W. W., and Krueger, G. G. (1984).
      Percutaneous penetration in the hairless dog, weanling pig and grafted athymic nude mouse:
      evaluation of models for predicting skin penetration in man. Br JDermatol 111 Suppl  27, 123-35.
   Reifenrath, W.  G., Hawkins, G. S., and Kurtz, M. S.  (1991). Percutaneous penetration and skin retention
      of topically applied compounds: an in vitro-in vivo study. JPharm Sci 80,  526-32.
   Riley, R. T., and Kemppainen, B. W. (1985). Effect of serum-parathion interactions on cutaneous
      penetration of parathion in vitro. Food  Chem Toxicol 23, 67-71.
   Ripley, B. D., Lissemore, L. L, Leishman, P. D., Denomme, M. A., and Ritter, L. (2000). Pesticide
      residues on fruits and vegetables from Ontario, Canada, 1991-1995. JAOACInt 83,  196-213.
                                                C-12

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Rojas, M., Bustos-Obregon, E., Martinez-Garcia, F., Contreras, H., and Regadera, J. (1998). The effect
      of parathion on mouse testicular and epididymal development cultured in chicken allantochorion.
      Adv Exp Med Biol 444, 201 -6.
   Rowland, S. S., Speedie, M. K., and Pogell, B. M. (1991). Purification and characterization of a secreted
      recombinant phosphotriesterase (parathion hydrolase) from Streptomyces lividans. ApplEnviron
      Microbiol 57, 440-4.
   Saleh, A. M., Vijayasarathy, C., Fernandez-Cabezudo, M., Taleb, M., and Petroianu, G. (2003).
      Influence of paraoxon (POX) and parathion (PAT) on apoptosis: a possible mechanism for toxicity
      in low-dose exposure. J Appl Toxicol 23, 23-9.
   Sawahata,  T.,  and Neal, R. A. (1982). Inhibition of rat liver cytochrome P-450 by benzyl hydrodisulfide.
      Mol Pharmacol 21, 464-7.
   Segura, P., Chavez, J., Montano, L. M., Vargas, M. H., Delaunois, A., Carbajal, V., and Gustin, P.
      (1999). Identification of mechanisms involved in the acute airway toxicity induced by parathion.
      Naunyn Schmiedebergs Arch Pharmacol 360, 699-710.
   Belgrade, M. K., Daniels, M. J., Illing, J. W., Ralston, A. L., Grady, M.  A., Charlet, E., and Graham, J.
      A. (1984). Increased susceptibility to parathion poisoning following murine cytomegalovirus
      infection. ToxicolApplPharmacol76, 356-64.
   Senel, A. C., Ulusoy, H., and Erciyes, N. (2001). An intermediate syndrome after parathion poisoning.
      Intensive Care Med 27, 333.
   Shah, P. V., and Guthrie, F. E. (1983). Percutaneous penetration of three insecticides in rats: a
      comparison of two methods for in vivo determination. J Invest Dermatol 80, 291-3.
   Shih, D. M., Gu, L.,  Xia, Y. R., Navab, M.,  Li, W. F., Kama, S., Castellani, L. W., Furlong, C. E., Costa,
      L. G., Fogelman, A. M., and Lusis, A. J. (1998). Mice lacking serum paraoxonase are susceptible to
      organophosphate toxicity and atherosclerosis. Nature 394, 284-7.
   Siller, F. R., Quintanilla-Vega, B.,  Cebrian,  M. E., and Albores, A. (1997). Effects of arsenite
      pretreatment on the acute toxicity of parathion. Toxicology 116, 59-65.
   Simcox, N. J., Fenske, R. A., Wolz, S. A., Lee, I. C., and Kalman, D. A. (1995). Pesticides in household
      dust and soil: exposure pathways for children of agricultural families. Environ Health Perspect 103,
      1126-34.
   Skinner, C. S., and Kilgore, W. W. (1982). Percutaneous penetration of [14C]parathion in the mouse:
      effect of anatomic region. J ToxicolEnviron Health  9, 483-90.
   Soranno, T. M., and  Sultatos, L. G. (1992). Biotransformation of the insecticide parathion by mouse
      brain. Toxicol Lett 60,  27-37.
   Sultatos, L. G. (1986). The effects  of phenobarbital pretreatment on the metabolism and acute toxicity of
      the pesticide parathion in the mouse. Toxicol Appl Pharmacol 86, 105-11.
   Sultatos, L. G., Basker, K. M., Shao, M., and Murphy, S. D. (1984). The interaction of the
      phosphorothioate insecticides chlorpyrifos and parathion and their oxygen analogues with bovine
      serum albumin. Mol Pharmacol 26, 99-104.
   Sultatos, L. G., and Gagliardi, C. L. (1990). Desulfuration of the insecticide parathion by human
      placenta in vitro. Biochem Pharmacol 39, 799-801.
   Sultatos, L. G., and Minor, L. D. (1986). Factors affecting the biotransformation of the pesticide
      parathion by the  isolated perfused mouse liver. DrugMetab Dispos  14, 214-20.
   Sultatos, L. G., Minor, L. D., and Murphy, S. D. (1985). Metabolic activation of phosphorothioate
      pesticides: role of the liver. J Pharmacol Exp Ther 232, 624-8.
   Sultatos, L. G., and Murphy, S. D.  (1983). Kinetic analyses of the microsomal biotransformation of the
      phosphorothioate insecticides chlorpyrifos and parathion. FundamAppl Toxicol 3, 16-21.
   Tang, J., and Chambers, J. E. (1999). Detoxication of paraoxon by rat liver homogenate and serum
      carboxylesterases and A-esterases. J Biochem Mol Toxicol 13, 261-8.
                                                C-13

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Thiermann, H., Mast, U., Klimmek, R., Eyer, P., Hibler, A., Pfab, R., Felgenhauer, N., and Zilker, T.
      (1997). Cholinesterase status, pharmacokinetics and laboratory findings during obidoxime therapy in
      organophosphate poisoned patients. Hum Exp Toxicol 16, 473-80.
   Tong, J., Feng, Z. Y., and Yang, Y. N. (1988). Chronotoxicological studies on toxicity of parathion.
      ChinMedJ(Engl) 101, 715-8.
   Undeger, U., Institoris, L., Siroki, O., Nehez, M., and Desi, I. (2000). Simultaneous geno- and
      immunotoxicological investigations for early detection of organophosphate toxicity in rats.
      EcotoxicolEnviron Sa/45., 43-8.
   Van Den Beukel, L, Dijcks, F. A., Vanderheyden, P., Vauquelin, G., and Oortgiesen, M.  (1997).
      Differential muscarinic receptor binding of acetylcholinesterase inhibitors in rat brain, human brain
      and Chinese hamster ovary cells expressing human receptors. JPharmacolExp Ther 281, 1113-9.
   van den Beukel, L, van Kleef, R. G., and Oortgiesen, M. (1998). Differential effects of physostigmine
      and organophosphates on nicotinic receptors in neuronal cells of different species. Neurotoxicology
      19, 777-87.
   Vargas Loza, A. M., Montes de Oca, E. L, and Posadas del Rio, F. A. (1997). Sex differences in the
      enzymatic hydrolysis of acetylsalicylic acid by microsomes from various rat tissues. J Appl Toxicol
      17,347-51.
   Venera, G. D., Morisoli, L. S., and Rodriguez Garay, E. A. (1978). Degradation of parathion to p-
      nitrophenol by livers of rats treated with phenobarbital. Farmaco [Prat] 33, 549-53.
   Vitarius, J. A., O'Shaughnessy, J. A., and Sultatos, L. G. (1995). The effects of phenobarbital
      pretreatment on the metabolism  and toxicity of paraoxon in the mouse. Pharmacol Toxicol 77, 16-
      22.
   Wagner, E. D., Marengo, M. S., and Plewa, M. J. (2003). Modulation of the mutagenicity of
      heterocyclic amines by organophosphate insecticides and their metabolites. Mutat Res 536,  103-15.
   Wallace, K. B., and Dargan, J. E. (1987).  Intrinsic metabolic clearance of parathion and paraoxon by
      livers from fish and rodents. Toxicol Appl Pharmacol90, 235-42.
   Watson, A. M., Chambers, H., and Chambers, J. E. (1994). An investigation of activities and paraoxon
      sensitivities of hepatic aliesterases in beta-naphthoflavone-treated rats. Toxicol Lett 71, 217-25.
   Weitman, S. D., Vodicnik, M. J., and Lech, J. J. (1983). Influence of pregnancy on parathion toxicity
      and disposition. Toxicol Appl Pharmacoll'I, 215-24.
   Weitman, S. D., Vodicnik, M. J., and Lech, J. J. (1986a). Influence of pregnancy on the hepatic
      metabolism of parathion. Dev Pharmacol Ther 9, 23-31.
   Weitman, S. D., Vodicnik, M. J., and Lech, J. J. (1986b). Mechanism of enhanced parathion/paraoxon
      toxicity during pregnancy in the mouse. Fundam Appl Toxicol 6, 155-61.
   Wester, R. M., Tanojo, H., Maibach, H. L, and Wester, R. C. (2000). Predicted chemical  warfare agent
      VX toxicity to uniformed soldier using parathion in vitro human skin exposure and absorption.
      Toxicol Appl Pharmacol 168, 149-52.
   Williams, P. L., Carver, M. P., and Riviere, J. E. (1990). A physiologically relevant pharmacokinetic
      model of xenobiotic percutaneous absorption utilizing the isolated perfused porcine skin flap. J
      PharmSci 79, 305-11.
   Williams, P. L., Thompson, D., Qiao, G. L., Monteiro-Riviere, N., and Riviere, J. E. (1996). The use of
      mechanistically defined chemical mixtures (MDCM) to assess mixture component effects on the
      percutaneous absorption and cutaneous disposition of topically exposed chemicals. II. Development
      of a general dermatopharmacokinetic  model for use in risk assessment. Toxicol Appl Pharmacol 141,
      487-96.
   Zaidi, S. S., Bhatnagar, V. K., Gandhi, S.  J., Shah, M. P., Kulkarni, P. K., and Saiyed, H. N. (2000).
      Assessment of thyroid function in pesticide formulators. Hum Exp Toxicol 19, 497-501.
                                                C-14

-------
USEPA Contract No. 3C-R102-NTEX                           Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Zhang, H. X., and Sultatos, L. G. (1991). Biotransformation of the organophosphorus insecticides
      parathion and methyl parathion in male and female rat livers perfused in situ. DrugMetab Dispos
      19, 473-7.
   Zhu, S., and Liu, Y. G. (1994). Toxicity and toxicokinetics of MeISP in isolated rat hepatocytes. Biomed
      Environ Sci 7, 169-79.
                                                   C-15

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

                                       Methyl Parathion

   2.0    Introduction

   Methyl parathion ((9,0-dimethyl (9-4-nitrophenylphosphorothioate) is a highly toxic organophosphorus
   insecticide approved for specific agricultural crops. Its use is restricted by appropriately trained certified
   pesticide applicators (Garcia etal. 2003). However, it has been used illegally indoors in certain areas of
   the Southern and Midwestern parts of the United States due to its effectiveness and low cost (ATSDR
   2001; Rubin et al. 2002), leading to an increased health risk  in non-workers, children and pregnant
   women.

   2.1    Toxic effects

   Neurotoxicity is the major toxic effect of methyl parathion (MP) or its metabolite, methyl paraoxon, in
   various species caused by inhibition of acetylcholinesterase (AChE) enzymes, resulting in acetylcholine
   accumulation at postsynaptic receptors and overstimulation of cholinergic systems (Chambers and Carr
   1993; Gupta et al. 2000; Hahn et al. 1991; Ma et al. 2003).  The median lethal dose (LD50) of MP in
   mice applied orally and dermally was 14.5 and 1200 mg/kg body weight, respectively, while the dermal
   median effective dose (ED50) that caused 50% reduction in AChE was 550 mg/kg at 24 hours after
   dosing (Haley et al.  1975; Skinner and Kilgore 1982).  The developing animals are more sensitive to
   acute toxicity of MP than adults, indicating the age-related differences in sensitivity to MP exposure
   (Liu etal. 1999; Pope and Chakraborti 1992; Pope etal. 1991).

   In humans, manifestations of exposure to MP such  as shortness of breath, nose bleeding, vomiting,
   diarrhea, abdominal cramps, headache, eye pain, blurred vision, sweating, confusion, muscle
   contraction, contact burns and erythema multiforme eruption (following dermal exposure) were
   reported. The severe neurotoxic effects include loss of coordination, slurred speech, fatigue and death
   caused by respiratory or cardiac arrest (Azaroff and Neas 1999; Fisher 1986; Karki et al. 2001; Rehner
   et al. 2000). Cranial nerve palsies and intermediate syndromes have also been reported  in certain
   patients (Karki et al. 2001; Narendra et al. 1989).

   Other effects reported include genotoxic and mutagenic effects (Bartoli et al.  1991; Breau et al. 1985;
   Chen et al. 1981; de Cassia Stocco et al. 1982; Degraeve and Moutschen 1984; Dolara  et al. 1993;
   Griffin and Hill 1978; Grover and Malhi 1985; Lodovici et al.  1994; Lodovici et al. 1997; Mathew et al.
   1990; Mathew etal. 1992; Nehez etal. 1994; Rashid and Mumma 1984; Rupa etal. 1990; Rupa etal.
   1991; Singh etal. 1984; Tripathy etal. 1987; Undeger etal.  2000; Velazquez etal. 1990;
   Vijayaraghavan and Nagarajan 1994; Wiaderkiewicz et al. 1986), effects on calmodulin (Pala et al.
   1991), effects  on liver and muscle enzymes (Delia Morte et al. 1994; Gupta et al. 1994; Jabbar et al.
   1990), hematoxicity (Parent-Massin and Thouvenot 1993), immunotoxic effects (Crittenden et al. 1998;
   Institorise^a/. 1995; Institoris etal. 1992; Lee et al. 1979; Sunil Kumar et al.  1993; Undeger et al.
   2000), hormonal effects (Asmathbanu and Kaliwal  1997; Fatranska et al. 1978; Lukaszewicz-Hussain et
   al.  1985; Sortur and Kaliwal 1999), reproductive and developmental effects (Basha and Nayeemunnisa
   1993; Desi etal. 1998; Dhondup and Kaliwal 1997; Garcia etal. 2003; Gupta etal.  1985; Gupta etal.
   1984; Kumar and Desiraju 1992; Mahaboob Basha et al. 2001; Mahaboob Basha and Nayeemunnisa
   1993; Nagymajtenyi et al. 1995; Nayeemunnisa and Begum  1992; Sortur and Kaliwal 1999),
   embryotoxicity (Tanimura et al. 1967; Uzokwu 1974), cardiac toxicity (Howard and Pope 2002), and
   behavioral effects (George etal. 1992;  Liu etal.  1994; Schulz  etal. 1990; Zhu etal. 2001).
                                               C-16

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   2.2    Pharmacokinetics

   2.2.1  Absorption

   Because of its lipid solubility, MP can be absorbed through skin; therefore, the most likely route of
   human exposure is dermal, particularly from agricultural field reentry (Abu-Qare et al. 2000).  Oral
   exposure can also occur via contaminated food or water consumption and suicidal attempt (Garcia et al.
   2003) while exposure to MP via inhalation during spraying is questionable (Kummer and van Sittert
   1986).

   Oral absorption

   MP is well and rapidly absorbed through the gastrointestinal tract following oral gavage in mice
   (Hollingworth 1967), rats (Garcia-Repetto et al. 1997; Kramer and Ho 2002; Kramer et al. 2002;
   Miyamoto 1963), guinea pigs (Miyamoto 1963), dogs (Braeckman et al. 1983), and humans (Morgan et
   al. 1977). However, the oral bioavailability is very low (5-20%), which can be explained by a significant
   hepatic first-pass effect. The oral absorption rate constant after 1.5-2.5 mg/kg administration of MP in
   rats was  1.2 h"1 (Kramer and Ho 2002; Kramer et al. 2002).

   Dermal absorption

   An in vitro model using human skin in a static diffusion cell system demonstrated that 5.2% of the
   applied dose of MP from a commercial formulation was present after 24 h (Sartorelli et al. 1997). In
   adult female rats and pregnant rats, 20-50% of administered dose was absorbed following a single
   dermal dose of 10-50 mg/kg MP with the absorption rate constant of 0.41 h"1 (Abu-Qare et al. 2000;
   Kramer and Ho 2002; Kramer et al. 2002).

   2.2.2  Distribution

   Following oral and dermal administration, MP is extensively bound to plasma protein and rapidly
   distributed to tissues including placenta and fetus. Then it is slowly redistributed to the central
   compartment (Abu-Qare et al. 2000; Garcia-Repetto et al.  1997). The highest level of MP is found in
   adipose tissue.  Distribution coefficients of adipose tissue, liver and brain in rats and mice have been
   published (Garcia-Repetto et al. 1995; Sultatos et al.  1990). The terminal half-life varies from 7.2 h to
   15 days,  depending on species and gender (Abu-Qare etal. 2000; Braeckman et al. 1980; Garcia-
   Repetto et al. 1997; Kramer and Ho 2002; Kramer et al. 2002). The volume of distribution is relatively
   high (9.6 I/kg in dogs and 10.1 I/kg in female rats) (Braeckman et al. 1980; Kramer and Ho 2002).

   2.2.3  Metabolism

   MP is metabolized by hepatic and extrahepatic phase I and phase II enzymes (Figure 2.1) (Abu-Qare et
   al. 2000; Garcia et al. 2003).  Phase I metabolism include dearylation of MP, leading to the formation of
   />-nitrophenol and dimethyl thiophosphoric acid, which promotes detoxification. On the other hand,
   desulfuration by cytochrome P450 can activate MP to methyl paraoxon, the neurotoxic metabolite
   (Yamamoto et al. 1983; Zhang and Sultatos 1991).  This oxidation process is the major metabolic
   pathway of MP in the liver (Anderson et al. 1992; Sultatos 1987).  Methyl paraoxon  is also formed in
   the brain. CYP2B has been demonstrated to be responsible for MP activation in rat brain extracts
   (Albores et al. 2001). The activities of dearylation and desulfuration of MP were reduced when a low
   dose of MP was given repeatedly in rats (Yamamoto et al.  1982).

                                                C-17

-------
USEPA Contract No. 3C-R102-NTEX
                                                  Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
                                           GS-CH3
                                            P450
                                           Oxidation   CH3O
                                                                     N02
                                                      Methyl Paraoxon
                                                  Hydrolysis
                                                          3.
                   CH30X
              Dimethyl Phosphorothioic Acid
HO-f   V N02

 Para-Nitrophenol
      O
CH30S||
      P-OH
CH30X
Dimethyl Phosphoric Acid
           Figure 2.1. Metabolic pathways of MP (Benke and Murphy 1975; Garcia etal. 2003).

   Methyl paraoxon is then hydrolyzed by liver and plasma paraoxonase to form/>-nitrophenol and
   dimethyl phosphoric acid (Garcia et al. 2003).  The correlation between LDso of MP in rats of several
   ages and reaction rates of metabolism of methyl paraoxon, both hydrolysis and GSH-dependent
   pathways, was reported, indicating that these pathways contributed to age-related differences in MP
   toxicity (Benke and Murphy 1975). /7-Nitrophenol further undergoes glucuronidation and sulfuric
   conjugation.

   MP is also conjugated by glutathione S-aryl transferase to form/»-nitrophenyl mercapturic acid (Di Ilio
   et al. 1995; Huang and Sultatos 1993; Sultatos and Woods 1988) and by glutathione ^-alkyl transferase
   to yield ^-methyl glutathione (Radulovic et al.  1987; Radulovic et al. 1986).  Furthermore, a study has
   reported another non mixed-function oxidative pathway of MP in brain tissue subfractions that
   transformed MP to its isomer (de Lima et al. 1996).

   2.2.4   Excretion

   MP is rapidly eliminated after oral and dermal  administration. Renal excretion is the major route of MP
   elimination. In rats, 75-90% of administered dose was recovered in urine and less than 10% was found
   in feces (Abu-Qare etal. 2000; Abu-Qare and Abou-Donia 2000; Hollingworth 1967; Miyamoto 1963).
   In humans,  the urinary metabolites of MP are/>-nitrophenol, dimethylphosphate, and unidentified
   metabolites (Morgan et al. 1977).  Therefore, />-nitrophenol has been used as a biomarker of MP
   exposure in humans (Barr et al. 2002; Chang et al. 1997; Esteban et al. 1996; Hryhorczuk et al. 2002;
   Rubin etal. 2002).

   2.3     Interactions

   MP has been reported to produce behavioral alterations when given in combination with endosulfan
   (Castillo et al. 2002) or toxaphene (Crowder et al. 1980) and more likely to induce intermediate
   syndrome when combined with parathion (De Bleecker et al. 1992).  Conversely, the inhibition of
   cholinesterase enzyme activity was significantly lowered when MP was administered with either
   chlorpyrifos or diazinon, which could be due to competition for cytochrome P-450 enzymes, resulting in
                                                C-18

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   inhibition of oxon formation (Abu-Qare etal. 2001; Abu-Qare and Abou-Donia 2001).  Moreover,
   cimetidine, chlordecone, mirex and linuron, gentamicin and rifamycin, polychlorinated biphenyls
   (PCBs), and permethrin have also been demonstrated to change the toxicity of MP (Carr et al. 2002;
   Joshi and Thornburg 1986; Ortiz etal. 1995; Tvede etal. 1989; Youssef etal. 1987). No interactions
   with acetaminophen and hexachlorocyclohexane (HCH) have been reported (Costa and Murphy 1984;
   Dikshith et al. 1991).
   2.4    PBPK models

   One- to three-compartment classical models has been used to fit the blood concentration data following
   intravenous, oral, and dermal administration (Abu-Qare et al. 2000; Braeckman et al. 1983; Braeckman
   et al. 1980; Kramer and Ho 2002; Kramer et al. 2002). No PBPK models for MP have been published.
   2.5    Literature Cited

   Abu-Qare, A. W., Abdel-Rahman, A., Brownie, C., Kishk, A. M., and Abou-Donia, M. B. (2001).
      Inhibition of cholinesterase enzymes following a single dermal dose of chlorpyrifos and methyl
      parathion, alone and in combination, in pregnant rats. J ToxicolEnviron Health A 63, 173-89.
   Abu-Qare, A. W., Abdel-Rahman, A. A., Kishk, A. M., and Abou-Donia, M. B. (2000). Placental
      transfer and pharmacokinetics of a single dermal dose of [14C]methyl parathion in rats.  Toxicol Sci
      53,5-12.
   Abu-Qare, A. W., and Abou-Donia, M. B. (2000). Urinary excretion of metabolites following a single
      dermal dose of [14C]methyl parathion in pregnant rats. Toxicology 150, 119-27.
   Abu-Qare, A. W., and Abou-Donia, M. B. (2001). Inhibition and recovery of maternal and fetal
      cholinesterase enzyme activity following a single cutaneous dose of methyl parathion and diazinon,
      alone and in combination, in pregnant rats. J Appl Toxicol 21, 307-16.
   Albores, A., Ortega-Mantilla, G., Sierra-Santoyo, A., Cebrian, M. E., Munoz-Sanchez, J. L., Calderon-
      Salinas, J. V., and Manno, M. (2001). Cytochrome P450 2B (CYP2B)-mediated activation of
      methyl-parathion in rat brain extracts. Toxicol Lett 124, 1-10.
   Anderson, P. N., Eaton, D. L., and Murphy, S. D. (1992). Comparative metabolism of methyl parathion
      in intact and  subcellular fractions of isolated rat hepatocytes. FundamAppl Toxicol 18, 221-6.
   Asmathbanu, I, and Kaliwal, B. B. (1997). Temporal effect of methyl  parathion on ovarian
      compensatory hypertrophy, follicular dynamics and estrous cycle in hemicastrated albino rats. J
      Basic Clin PhysiolPharmacol 8, 237-54.
   ATSDR (2001).  Toxicological profile for methyl parathion.
   Azaroff, L. S., andNeas, L. M. (1999). Acute health effects associated with nonoccupational pesticide
      exposure in rural El Salvador. Environ Res 80, 158-64.
   Barr, D. B., Turner, W. E., DiPietro, E., McClure, P. C., Baker, S. E., Barr, J. R., Gehle, K., Grissom, R.
      E., Jr., Bravo, R., Driskell, W. J., Patterson, D. G., Jr., Hill, R. H., Jr., Needham, L. L., Pirkle, J. L.,
      and Sampson, E.  J. (2002). Measurement of p-nitrophenol in the urine of residents whose homes
      were contaminated with methyl parathion. Environ Health Perspect 110 Suppl 6, 1085-91.
   Bartoli, S., Bonora, B., Colacci, A., Niero, A.,  and Grilli, S. (1991). DNA damaging activity of methyl
      parathion. Res Commun Chem PatholPharmacol 71, 209-18.
   Basha, P. M., and Nayeemunnisa (1993). Effect of methyl parathion on Na(+)-K+ and Mg2+ adenosine
      triphosphatase activity in developing central nervous system in rats. Indian JExp Biol 31, 785-7.
   Benke,  G. M., and Murphy, S. D. (1975). The influence of age on the toxicity and metabolism of methyl
      parathion and parathion in male and female rats. Toxicol Appl Pharmacol 31, 254-69.

                                               C-19

-------
USEPA Contract No. 3C-R102-NTEX                           Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Braeckman, R. A., Audenaert, F., Willems, J. L., Belpaire, F. M., and Bogaert, M. G. (1983).
      Toxicokinetics of methyl parathion and parathion in the dog after intravenous and oral
      administration. Arch Toxicol 54, 71-82.
   Braeckman, R. A., Godefroot, M. G., Blondeel, G. M., Belpaire, F. M., and Willems, J. L. (1980).
      Kinetic analysis of the fate of methyl parathion in the dog. Arch Toxicol 43, 263-71.
   Breau, A. P., Mitchell, W. M., Swinson, J., and Field, L. (1985). Mutagenic and cell transformation
      activities of representative phosphorothioate esters in vitro. J Toxicol Environ Health 16, 403-13.
   Carr, R. L., Richardson, J. R.,  Guarisco, J. A., Kachroo, A., Chambers, J. E., Couch, T. A., Durunna, G.
      C., and Meek, E. C.  (2002). Effects of PCB exposure on the toxic impact of organophosphorus
      insecticides. Toxicol Sci 67, 311-21.
   Castillo, C. G., Montante, M.,  Dufour, L., Martinez, M. L., and Jimenez-Capdeville, M. E. (2002).
      Behavioral effects of exposure to endosulfan and methyl parathion in adult rats. Neurotoxicol
      Teratol 24, 797-804.
   Chambers, J. E., and Carr, R. L. (1993). Inhibition patterns of brain acetylcholinesterase and hepatic and
      plasma aliesterases following exposures to  three phosphorothionate insecticides and their oxons in
      rats. Fundam Appl Toxicol 21, 111-9.
   Chang, M. J., Chen, Y. C., and Yang, H. J. (1997).  Comparative evaluation on the biological monitoring
      of exposure to parathion and its methyl analog. Arch Environ Contain Toxicol 32, 422-5.
   Chen, H. H., Hsueh, J. L., Sirianni, S. R., and Huang, C. C. (1981). Induction of sister-chromatid
      exchanges and cell cycle delay in cultured mammalian cells treated with eight organophosphorus
      pesticides. MutatRes 88, 307-16.
   Costa, L. G., and Murphy, S. D. (1984). Interaction between acetaminophen and organophosphates in
      mice. Res Commun Chem PatholPharmacol 44, 389-400.
   Crittenden, P. L., Carr, R., andPruett, S. B. (1998). Immunotoxicological assessment of methyl
      parathion in female B6C3F1 mice. J Toxicol Environ Health A 54, 1-20.
   Crowder, L. A., Lanzaro, G. C., and Whitson, R. S. (1980). Behavioral effects of methyl parathion and
      toxaphene exposure in rats. J Environ Sci Health B 15, 365-78.
   De Bleecker, J., Willems, J., Van Den Neucker, K., De Reuck, J., and Vogelaers, D. (1992). Prolonged
      toxicity with intermediate syndrome after combined parathion and methyl parathion poisoning. J
      Toxicol Clin Toxicol 30, 333-45; discussion 347-9.
   de Cassia Stocco, R., Becak, W., Gaeta, R., and Rabello-Gay, M. N. (1982). Cytogenetic study of
      workers exposed to methyl-parathion. MutatRes 103, 71-6.
   de Lima, J. S., Bastos Neto Ida, D., Bastos, V.  L., da Cunha, J. C., Moraes, F. F., Ferreira Mde, F.,
      Moreira Ida, D., and Faria, M. V. (1996). Methyl parathion activation by a partially purified rat brain
      fraction. ToxicolLett 81', 53-60.
   Degraeve,  N., and Moutschen, J. (1984). Absence of genetic and cytogenetic effects in mice treated by
      the organophosphorus insecticide parathion, its methyl analogue, and paraoxon. Toxicology 32, 177-
      83.
   Delia Morte, R., Villani, G. R., Di Martino, E., Squillacioti, C., De Marco, L., Vuotto, P., Belisario, M.
      A., and Staiano, N. (1994). Glutathione depletion induced in rat liver fractions by seven pesticides.
      Boll Soc Ital Biol Sper 70,  185-92.
   Desi, I, Nagymajtenyi, L., Papp, A., and Schulz, H. (1998). Experimental  model studies of pesticide
      exposure. Neurotoxicology 19, 611-6.
   Dhondup, P.,  and Kaliwal, B. B. (1997). Inhibition of ovarian compensatory hypertrophy by the
      administration of methyl parathion in hemicastrated albino rats. Reprod Toxicol 11, 77-84.
   Di Ilio, C., Sacchetta, P., lannarelli, V., and Aceto,  A. (1995). Binding of pesticides to alpha, mu and pi
      class glutathione transferase. Toxicol Lett 76, 173-7.
                                                C-20

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Dikshith, T. S., Raizada, R. B., Singh, V., Pandey, M., and Srivastava, M. K. (1991). Repeated dermal
      toxicity of technical HCH and methyl parathion (50EC) to female rats (Rattus norvigicus). Indian J
      Exp Biol 29, 149-55.
   Dolara, P., Vezzani, A., Caderni, G., Coppi, C., and Torricelli, F. (1993). Genetic toxicity of a mixture
      of fifteen pesticides commonly found in the Italian diet. Cell Biol Toxicol 9, 333-43.
   Esteban, E., Rubin, C., Hill, R., Olson, D., and Pearce, K. (1996). Association between indoor
      residential contamination with methyl parathion and urinary para-nitrophenol. J Expo Anal Environ
      Epidemiol6,375-87.
   Fatranska, M., Vargova, M., Rosival, L., Batora, V., Nemeth, S., and Janekova, D. (1978). Circadian
      susceptibility rhythms to some organophosphate compounds in the rat. Chronobiologia 5, 39-44.
   Fisher, A. A. (1986). Erythema multiforme-like eruptions due to topical miscellaneous compounds: Part
      III. Cutis 37, 262-4.
   Garcia, S. J., Abu-Qare, A. W., Meeker-O'Connell, W. A., Borton, A. J., and Abou-Donia, M. B. (2003).
      Methyl parathion: a review of health effects. J Toxicol Environ Health B Crit Rev 6, 185-210.
   Garcia-Repetto, R., Martinez, D., and Repetto, M. (1995). Coefficient of distribution of some
      organophosphorous pesticides in rat tissue. Vet Hum Toxicol 37, 226-9.
   Garcia-Repetto, R., Martinez, D., and Repetto, M. (1997). Biodisposition study of the organophosphorus
      pesticide, methyl-parathion. Bull Environ Contain Toxicol 59, 901-8.
   George, J., Andrade, C., and Joseph, T. (1992). Delayed effects of acute oral and chronic inhalational
      exposure to methylparathion on learning and memory in rats. Indian JExp Biol 30, 819-22.
   Griffin, D. E., 3rd, and Hill, W. E. (1978). In vitro breakage of plasmid DNA by mutagens and
      pesticides. Mutat Res 52, 161-9.
   Grover, I. S., and Malhi, P. K. (1985). Genotoxic effects of some organophosphorous pesticides. I.
      Induction of micronuclei in bone marrow cells in rat. MutatRes 155, 131-4.
   Gupta, R. C., Goad, J. T., and Kadel, W. L. (1994). Cholinergic and noncholinergic changes in skeletal
      muscles by carbofuran and methyl parathion. J Toxicol Environ Health 43, 291-304.
   Gupta, R. C., Goad, J. T., Milatovic, D.,  and Dettbarn, W. D.  (2000). Cholinergic and noncholinergic
      brain biomarkers of insecticide exposure and effects. Hum Exp Toxicol 19, 297-308.
   Gupta, R. C., Rech, R. H., Lovell, K. L., Welsch, F., and Thornburg, J. E. (1985). Brain cholinergic,
      behavioral, and morphological development in rats exposed in utero to methylparathion. Toxicol
      ApplPharmacol 77', 405-13.
   Gupta, R. C., Thornburg, J. E., Stedman, D. B., and Welsch, F. (1984). Effect of subchronic
      administration of methyl parathion on in vivo protein synthesis in pregnant rats and their
      conceptuses. Toxicol Appl Pharmacoll'2, 457-68.
   Hahn, T., Ruhnke, M., and Luppa, H. (1991). Inhibition of acetylcholinesterase and
      butyrylcholinesterase by the organophosphorus insecticide methylparathion in the central  nervous
      system of the golden hamster (Mesocricetus auratus). Acta Histochem 91, 13-9.
   Haley, T. J., Farmer, J. H., Harmon, J. R., and Dooley, K. L. (1975). Estimation of the LD1 and
      extrapolation of the LDO. 1  for five organothiophosphate pesticides. Eur J Toxicol Environ Hyg 8,
      229-35.
   Hollingworth, R.  M., Metcalf,  R.L., Fukuto, T.R. (1967). The selectivity of sumithion compared with
      methyl parathion:  Metabolism in the white mouse. J Agric FoodChem 15, 242-9.
   Howard, M. D., and Pope, C. N. (2002).  In vitro effects of chlorpyrifos, parathion, methyl parathion and
      their oxons on cardiac muscarinic receptor binding in neonatal and adult rats.  Toxicology  170, 1-10.
   Hryhorczuk, D. O., Moomey, M., Burton, A., Runkle, K., Chen, E., Saxer, T., Slightom, J., Dimos, J.,
      McCann, K., and Barr, D. (2002). Urinary p-nitrophenol as a biomarker of household exposure to
      methyl parathion. Environ Health Perspect 110 Suppl 6, 1041-6.
   Huang, Y. S., and Sultatos, L. G. (1993). Glutathione-dependentbiotransformation of methyl parathion
      by mouse liver in vitro. Toxicol Lett 68, 275-84.

                                                C-21

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Institoris, L., Siroki, O., and Desi, I. (1995). Immunotoxicity study of repeated small doses of
      dimethoate and methylparathion administered to rats over three generations. Hum Exp Toxicol 14,
      879-83.
   Institoris, L., Siroki, O., Toth, S., and Desi, I. (1992). Immunotoxic effects of MPT-IP containing 60%
      methylparathion in mice. Hum Exp Toxicol 11, 11-6.
   Jabbar, A., Khawaja, S. A., Iqbal, A., and Malik, S. A. (1990). Effect of malathion and methyl-parathion
      on rat liver enzymes. JPakMedAssoc 40, 266-70.
   Joshi, U. M., and Thornburg, J. E. (1986). Interactions between cimetidine, methylparathion, and
      parathion. J Toxicol Environ Health 19, 337-44.
   Karki, P., Hansdak, S. G., Bhandari, S., Shukla, A., and Koirala, S. (2001). A clinico-epidemiological
      study of organophosphorus poisoning at a rural-based teaching hospital in eastern Nepal. Trop Doct
      31,32-4.
   Kramer, R. E., and Ho, I. K. (2002). Pharmacokinetics and pharmacodynamics of methyl parathion.
      Zhonghua Yi Xue Za Zhi (Taipei) 65, 187-99.
   Kramer, R. E., Wellman, S. E., Rockhold, R. W., and Baker, R. C. (2002). Pharmacokinetics of methyl
      parathion: a comparison following single intravenous, oral or dermal administration. JBiomedSci 9,
      311-20.
   Kumar, M. V., and Desiraju, T. (1992). Effect of chronic consumption of methylparathion on rat brain
      regional acetylcholinesterase activity and on levels of biogenic amines. Toxicology 75,  13-20.
   Kummer, R., and van Sittert, N. J. (1986). Field studies on health effects from the application of two
      organophosphorus insecticide formulations by hand-held ULV to cotton. Toxicol Lett 33, 7-24.
   Lee, T. P., Moscati, R., and Park, B. H. (1979). Effects of pesticides  on human leukocyte functions. Res
      Commun Chem Pathol Pharmacol 23, 597-609.
   Liu, J., Olivier, K., and Pope, C. N. (1999). Comparative neurochemical effects of repeated methyl
      parathion or chlorpyrifos exposures in neonatal and adult rats. Toxicol ApplPharmacol 158, 186-96.
   Liu, P.  S., Kao, L. S., and Lin, M. K. (1994). Organophosphates inhibit catecholamine secretion and
      calcium influx in bovine adrenal chromaffin cells. Toxicology 90, 81-91.
   Lodovici, M., Aiolli, S., Monserrat, C., Dolara, P., Medica, A., and Di Simplicio, P. (1994). Effect of a
      mixture of 15 commonly used pesticides on DNA levels of 8-hydroxy-2-deoxyguanosine and
      xenobiotic metabolizing enzymes in rat liver. J Environ Pathol Toxicol Oncol 13, 163-8.
   Lodovici, M., Casalini, C., Briani, C., and Dolara, P. (1997). Oxidative liver DNA damage in rats treated
      with pesticide mixtures. Toxicology 117, 55-60.
   Lukaszewicz-Hussain, A., Moniuszko-Jakoniuk, J., and Pawlowska,  D. (1985). Blood glucose and
      insulin concentration in rats subjected to physical exercise in acute poisoning with parathion-methyl.
      PolJPharmacolPharm 37, 647-51.
   Ma, T., Kramer, R. E., Baker, R. C., Fan, L. W., and Ho, I. K. (2003). Effects of chronic dermal
      exposure to nonlethal doses of methyl parathion on brain regional acetylcholinesterase and
      muscarinic cholinergic receptors in female rats. JNeurosci Res 71, 138-45.
   Mahaboob Basha, P., Begum, S., and Nayeemunnisa (2001). Methyl parathion induced alterations in
      monoaminergic system of developing rat pups. Indian J Exp Biol 39, 276-9.
   Mahaboob Basha, P., and Nayeemunnisa (1993). Methyl parathion induced alterations in GABAergic
      system during critical stage of central nervous system development in albino rat pups. Indian J Exp
      5/0/31,369-72.
   Mathew, G., Rahiman, M. A., and Vijayalaxmi, K. K. (1990). In vivo genotoxic effects in mice of
      Metacid 50, an organophosphorus insecticide. Mutagenesis 5, 147-9.
   Mathew, G., Vijayalaxmi, K. K., and Abdul Rahiman, M. (1992). Methyl parathion-induced sperm
      shape abnormalities in mouse. MutatRes 280, 169-73.
                                                C-22

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Miyamoto, J., Sato, Y., Kadota, T., Fujinami, A., Endo, M. (1963). Studies on the mode of action of
      organophosphorus compounds: Part I. metabolic fate  of P32 labeled sumithion and methylparathion
      in guinea pig and white rat. Agr Biol Chem 27, 381-9.
   Morgan, D. P., Hetzler, H. L., Slach, E. F., and Lin, L. I.  (1977). Urinary excretion of paranitrophenol
      and alkyl phosphates following ingestion of methyl or ethyl parathion by human subjects. Arch
      Environ Contain Toxicol 6, 159-73.
   Nagymajtenyi, L., Schulz, H., and Desi, I. (1995). Changes in EEG of freely-moving rats caused by
      three-generation organophosphate treatment. Arch Toxicol Suppl 17, 288-94.
   Narendra, J., Chethankumar, J. G., and Rao, B. B. (1989). Cranial nerve palsies in organophosphorus
      poisoning. JAssoc Physicians India 37, 732-3.
   Nayeemunnisa, and Begum, S. (1992). Methyl parathion  induced regional alterations in the regulatory
      proteins during critical stage of central nervous system development in albino rat pups. Indian J
      Physiol Pharmacone, 77-82.
   Nehez, M., Toth, C., and Desi, I. (1994). The effect of dimethoate, dichlorvos, and parathion-methyl on
      bone marrow cell chromosomes of rats in subchronic  experiments in vivo. Ecotoxicol Environ Saf
      29,365-71.
   Ortiz, D., Yanez, L., Gomez, H., Martinez-Salazar, J. A.,  and Diaz-Barriga, F. (1995). Acute
      toxicological effects in rats treated with a mixture of commercially formulated products containing
      methyl parathion and permethrin. Ecotoxicol Environ Saf '32, 154-8.
   Pala, L, Vig, P. J., Desaiah, D., and Srinivasan, A. (1991). In vitro effects of organophosphorus
      compounds on calmodulin activity. JAppl Toxicol 11, 391-5.
   Parent-Massin, D., and Thouvenot, D. (1993). In vitro study of pesticide hematotoxicity in human and
      rat progenitors. JPharmacol Toxicol Methods 30, 203-7.
   Pope, C. N., and Chakraborti, T. K. (1992). Dose-related  inhibition of brain and plasma cholinesterase  in
      neonatal and adult rats following sublethal organophosphate exposures. Toxicology 73, 35-43.
   Pope, C. N., Chakraborti, T. K., Chapman, M. L., Farrar,  J. D., and Arthun, D. (1991). Comparison of in
      vivo cholinesterase inhibition in neonatal and adult rats by three organophosphorothioate
      insecticides. Toxicology 68, 51-61.
   Radulovic, L. L., Kulkarni, A. P., and Dauterman, W. C.  (1987). Biotransformation of methyl parathion
      by human foetal liver glutathione S-transferases: an in vitro study. Xenobiotica 17, 105-14.
   Radulovic, L. L., LaFerla, J. J., and Kulkarni, A. P. (1986). Human placental glutathione S-transferase-
      mediated metabolism of methyl parathion. Biochem Pharmacol 35, 3473-80.
   Rashid, K. A., and Mumma, R. O. (1984). Genotoxicity of methyl parathion in short-term bacterial test
      systems. J Environ Sci Health B 19, 565-77.
   Rehner, T. A., Kolbo, J. R., Trump, R., Smith, C., and Reid, D. (2000). Depression among victims of
      south Mississippi's methyl parathion disaster. Health Soc Work 25, 33-40.
   Rubin, C., Esteban, E., Hill, R. H., Jr., and Pearce, K. (2002). Introduction—the methyl parathion story: a
      chronicle of misuse and preventable human exposure. Environ Health Perspect 110 Suppl 6, 1037-
      40.
   Rupa, D. S., Reddy, P. P., and Reddi, O. S. (1990). Cytogeneticity of quinalphos and methyl  parathion
      in human peripheral lymphocytes. Hum Exp  Toxicol 9, 385-7.
   Rupa, D. S., Reddy, P. P.,  Sreemannarayana, K., and Reddi, O. S.  (1991). Frequency of sister chromatid
      exchange in peripheral lymphocytes of male  pesticide applicators. Environ MolMutagen 18, 136-8.
   Sartorelli, P., Aprea, C., Bussani, R., Novelli, M. T., Orsi, D., and Sciarra, G. (1997). In  vitro
      percutaneous penetration of methyl-parathion from a  commercial formulation through the human
      skin. Occup Environ Med 54, 524-5.
   Schulz, H., Desi, I, and Nagymajtenyi, L. (1990). Behavioral effects of subchronic intoxication with
      parathion-methyl in male Wistar rats. Neurotoxicol Teratol 12, 125-7.
                                                C-23

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Singh, S., Lehmann-Grube, B., and Goedde, H. W. (1984). Cytogenetic effects of paraoxon and methyl-
      parathion on cultured human lymphocytes: SCE, clastogenic activity and cell cycle delay. Int Arch
      Occup Environ Health 54, 195-200.
   Skinner, C. S., and Kilgore, W. W. (1982). Acute dermal toxicities of various organophosphate
      insecticides in mice. J Toxicol Environ Health 9, 491-7.
   Sortur, S. M., and Kaliwal, B. B. (1999). Effect of methyl parathion formulation on estrous cycle and
      reproductive performance in albino rats. Indian JExp Biol 37, 176-8.
   Sultatos, L. G. (1987). The role of the liver in mediating the acute toxicity of the pesticide methyl
      parathion in the mouse. DrugMetab Dispos 15,  613-7.
   Sultatos, L. G., Kim, B., and Woods, L. (1990). Evaluation of estimations in vitro of tissue/blood
      distribution coefficients for organothiophosphate insecticides.  ToxicolApplPharmacol 103, 52-5.
   Sultatos, L. G., and Woods, L. (1988). The role of glutathione in the detoxification of the insecticides
      methyl parathion and azinphos-methyl in the mouse. Toxicol Appl Pharmacol 96, 168-74.
   Sunil Kumar, K. B., Ankathil, R., and Devi, K. S. (1993). Chromosomal aberrations induced by methyl
      parathion in human peripheral lymphocytes of alcoholics and smokers. Hum Exp Toxicol 12, 285-8.
   Tanimura, T., Katsuya, T., and Nishimura, H. (1967). Embryotoxicity of acute exposure to methyl
      parathion in rats and mice. Arch Environ Health 15, 609-13.
   Tripathy, N. K., Dey, L., Majhi, B., and Das, C. C. (1987). Genotoxicity of metacid established through
      the somatic and germ line mosaic assays and the sex-linked recessive lethal test in Drosophila. Arch
      Toxicol 61, 53-7.
   Tvede, K. G., Loft, S., Poulsen, H. E., and Schou, J. S.  (1989). Methyl parathion toxicity in rats is
      changed by pretreatment with the pesticides chlordecone,  mirex and linuron. Arch Toxicol Suppl 13,
      446-7.
   Undeger, U., Institoris, L., Siroki, O., Nehez, M., and Desi, I.  (2000). Simultaneous geno- and
      immunotoxicological investigations for early detection of organophosphate toxicity in rats.
      EcotoxicolEnviron Saf45, 43-8.
   Uzokwu, M. (1974). Comparative feto-toxicity of organophosphate insecticide in mice. Bull Epizoot Dis
      Afr22, 161-6.
   Velazquez, A., Xamena, N., Creus, A., and Marcos, R.  (1990). Mutagenic evaluation of the
      organophosphorus insecticides methyl parathion and triazophos in Drosophila melanogaster. J
      Toxicol Environ Health 31, 313-25.
   Vijayaraghavan, M., and Nagarajan, B. (1994). Mutagenic potential of acute exposure to
      organophosphorus and organochlorine compounds. Mutat Res 321, 103-11.
   Wiaderkiewicz, R., Walter, Z., and Reimschussel, W. (1986).  Sites of methylation of DNA bases by the
      action of organophosphorus insecticides in vitro. Acta Biochim Pol 33, 73-85.
   Yamamoto, T., Egashira, T., Yoshida, T., and Kuroiwa, Y. (1982). Comparison of the effect of an
      equimolar and low dose of fenitrothion and methylparathion on their own metabolism in rat liver. J
      Toxicol Sci 7, 3 5 -41.
   Yamamoto, T., Egashira, T., Yoshida, T., and Kuroiwa, Y. (1983). Comparative metabolism of
      fenitrothion and methylparathion in male rats. Acta Pharmacol Toxicol (Copenh) 53, 96-102.
   Youssef, S. H., el-Sayed, M. G., and Atef, M. (1987). Influence of gentamicin and rifamycin on toxicity
      and biotransformation of methyl parathione in rats.  Dtsch  Tierarztl Wochenschr 94, 203-5.
   Zhang, H. X., and Sultatos, L. G. (1991). Biotransformation of the organophosphorus insecticides
      parathion and methyl parathion in male and female rat livers perfused in situ. DrugMetab Dispos
      19, 473-7.
   Zhu, H., Rockhold, R. W., Baker, R. C., Kramer, R. E., and Ho, I. K. (2001). Effects of single or
      repeated dermal exposure to methyl parathion on behavior and blood cholinesterase activity in rats. J
      BiomedSci 8, 467-74.
                                                C-24

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.


                                           Chlorpyrifos

   3.0    Introduction

   Chlorpyrifos ((9,(9-diethyl 0-(3,5,6-trichloro-2-pyridyl) phosphorothiolate) is an OP pesticide with
   restricted uses in the U.S.  It is produced in the U.S. and marketed under various trade names, including
   Dursban®,  Lorsban®, and other names (ATSDR 1997). Chlorpyrifos (CP) has been used against pests
   in turfgrass, commercial agriculture, and in residential settings, although indoor uses have been
   restricted.  CP has been found in drinking water supplies, although often at low levels (ATSDR 1997).

   3.1    Toxic Effects

   Numerous  systemic effects of CP have been reported, but the critical effect for risk assessment is
   inhibition of acetylcholinesterase by CP or CP-oxon (ATSDR 1997).  This can result in headache,
   diaphoresis, nausea, vomiting,  diarrhea, epigastric cramping, bradycardia, blurred vision, miosis,
   bronchoconstriction and excess mucous secretions, pulmonary edema, dyspnea, muscle fasciculations,
   salivation,  lacrimation, urination, tremors, anxiety, drowsiness, confusion, ataxia, abnormal gait,
   hypotension, and memory impairment (Ballantyne and Marrs  1992).

   Neurodevelopmental effects have also been described (Auman et al. 2000; Campbell etal.  1997; Carr et
   al. 2001; Chakraborti et al. 1993; Chanda and Pope 1996; Crumpton et al. 2000; Dam et al. 1999; Dam
   et al. 1998, 2000, 2003; Das and Barone 1999; Garcia et al. 2002; Garcia et al. 2003; Gore 2001, 2002;
   Howard and Pope 2002; Jett et al. 2001; Lassiter et al. 1998; Levin et al.  2002; Levin et al. 2001;
   Olivier et al.  2001; Qiao et al. 2002; Qiao et al. 2001; Qiao et al. 2003; Raines et al. 2001; Richardson
   and  Chambers 2003; Roy et al. 1998; Sachana et al. 2001; Slotkin et al. 2001; Slotkin et al. 2002; Song
   etal. 1998; Tang etal.  1999; Whitney etal.  1995; Won etal. 2001). Some of these studies suggested
   that developmental neurotoxicity occurred at lower exposure levels than did acetylcholinesterase
   inhibition in adult animals, and developmental neurotoxicity may be worthy of consideration in future
   risk assessments (Abdel-Rahman et al. 2002).

   3.2    Pharmacokinetics

   3.2.1  Absorption

   CP is well  absorbed through the gut after oral exposure to CP in drinking water or food.  In humans or
   rodents, 70-90% of the oral dose was absorbed (Ahdaya et al.  1981; Bakke et al. 1976; Nolan et al.
   1984; Smith et al. 1967). Only 3% of the dermal dose was absorbed; (Nolan et al. 1984) however, this
   is dependant on the vehicle.  When acetone was used as vehicle in another study, 46-99% of the dose
   was absorbed (Shah et al.  1987). Absorption rates through the gut (Cook and Shenoy 2003) and skin
   (Griffin et  al. 2000; Sartorelli et al 1998) have been quantified.

   3.2.2  Distribution

   CP rapidly distributes to tissues after absorption (Shah et al. 1987; Shah et al. 1981; Smith et al. 1967).

   3.2.3  Metabolism
                                               C-25

-------
USEPA Contract No. 3C-R102-NTEX
                                               Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
   CP is primarily metabolized in the liver due to the high concentration of cytochrome (CYP) P450s in
   that organ (Ma and Chambers 1994; Sultatos and Murphy 1983b). CP is rapidly bioactivated to CP-
   oxon by multiple isoforms of CYP P450. Both CP and CP-oxon are hydrolyzed by acetylcholinesterase
   to 3,5,6-trichloro-2-pyridinol (TCP), diethyl thiophosphate, and diethyl phosphate (Bakke et al. 1976;
   Nolan etal. 1984; Smithed al.  1967; Sultatos et al. 1985; Sultatos and Murphy 1983a, b). The principal
   metabolic pathways are shown in Figure 3.1 below.

                                           Figure 3.1
                          Principal Metabolic Pathways for Chlorpyrifos
                                 CYP450
                            Cl
                  Cl
           Chlorpyrifos
 Chlorpyrifos oxon
                 CYP450
                                                             CYP450
                     /
        (CH3CH20)2P

   Diethylthiophosphate (DETP)  + TCP
              \\^0
   (CH3CH20)2P"
Diethylphosphate (DEP)  +  TCP
                                   Modified from ATSDR (1997)

  For risk assessment or PBPK model development, studies of metabolism are often used, but further
  incorporation of pharmacodynamic responses such as acetylcholinesterase inhibition would improve the
  risk assessment or PBPK model development. Therefore, either type of study is included in this review.
  Several human pharmacokinetic (PK) or biomonitoring studies have been conducted (Cocker et al.
  2002; Drevenkar et al. 1993; Fenske etal. 2002; Sams and Mason 1999).

  A number of animal studies have been conducted addressing PKs, binding, cholinestase inhibition, or
  other endpoints with CP (Abdel-Rahman etal. 2002; Atterberry etal.  1997; Bushnell etal. 1994;
  Bushnell et al. 1993; Carr and Chambers 1996; Carr et al. 2002; Carr et al. 1995;  Chanda et al. 1997;
  Chiappa et al. 1995; Cowan et al. 2001; Hunter et al. 1999; Karanth and Pope 2000; Lassiter et al. 1999;
  Li etal. 2000; Li etal. 1995; Liu etal. 1999; Mattsson etal.  2000; Mortensen etal. 1996; Mortensen et
  al. 1998; Moser et al. 1998; Moser and Padilla 1998; Padilla et al. 2000; Padilla et al. 1994; Pond et al.
  1995; Pond etal. 1998; Stantonetal. 1994).

  Fetal transfer of CP or metabolites was assessed in several studies, including two  recent ones (Abdel-
  Rahman etal. 2002; Ashry etal. 2002).
                                              C-26

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
  In vitro PK studies with CP have been conducted. These include studies using cell lines (Barber and
  Ehrich 2001; Ehrich et al.  1997; Monnet-Tschudi etal. 2000), microsomes (Buratti et al. 2003; Katz et
  al.  1997; Ma and Chambers 1994; Poet et al. 2003; Sams et al. 2000; Tang et al. 2001; Usmani et al.
  2003), antibodies (Buratti et al. 2003), tissue slices (Liu et al. 2002), or other systems (Amitai et al.
  1998).

  Genetic polymorphisms relevant to CP metabolism have been described (Brophy et al. 2000; Costa et al.
  1999; Costa etal. 2003; Dai etal.  2001; Furlong etal. 2000a; Furlong etal. 1998; Furlong etal. 2000b).

  3.2.4  Excretion

  Most metabolites are found as conjugated metabolites of TCP in the urine. The half-life of elimination
  in rats was 10-16 hours in most tissues and 62 hours for fat (Smith et al. 1967). The half life for
  elimination in humans was estimated at 27 hours (Nolan et al. 1984).

  3.3    Interactions with other OP pesticides

  Interactions of CP with other OP pesticides have been studied (Axelrad et al. 2002; Karanth et al. 2001;
  Richardson etal. 2001; Tang etal. 2002; Usmani etal. 2002).

  3.4    PBPK models for  chlorpyrifos

  One group has developed PBPK models for CP (Timchalk et al.  2002a; Timchalk et al. 2002b). These
  models were based on adult rat and human exposures in gavage, dietary, or dermal studies in a seven-
  compartment model structure. They included saturable metabolism of CP by CYP and "a-esterases,"
  and binding of CP-oxon to "b-esterases" as a second order process. Regeneration of b-esterases was
  also included.

  A classical PK model for CP was also described (Rigas et al. 2001).

  The Timchalk et al. CP PBPK model could be adapted for use in a PBPK-model based risk assessment.
  Adaptation should include alteration for drinking water scenarios.  Also, during the process of
  determining the  common mechanism of toxicity for a mixture of OP pesticides, if developmental
  neurotoxicity is an important part of the risk assessment, PBPK models for relevant endpoints in the
  fetus or neonates should be considered.

  3.5    Literature Cited

  Abdel-Rahman,  A. A., Blumenthal, G. M., Abou-Donia, S. A., Ali, F. A., Abdel-Monem, A. E., and
      Abou-Donia, M. B. (2002). Pharmacokinetic profile and placental transfer of a single intravenous
      injection of [(14)C] chlorpyrifos in pregnant rats. Arch Toxicol 76, 452-9.
  Ahdaya, S. M., Monroe, R. J., and Guthrie, F. E. (1981). Absorption and distribution of intubated
      insecticides in fasted mice. Pestic Biochem Physiol 16, 38-46.
  Amitai, G., Moorad, D., Adani, R., and Doctor, B. P. (1998). Inhibition of acetylcholinesterase and
      butyrylcholinesterase by chlorpyrifos-oxon. Biochem Pharmacol 56, 293-9.
  Ashry, K. M., Abu-Qare, A. W., Saleem, F. R.,  Hussein, Y. A., Hamza, S. M., Kishk, A.  M., and Abou-
      Donia, M. B. (2002). Inhibition and recovery of maternal and fetal cholinesterase enzymes following
      a single oral  dose of chlorpyrifos in rats. Arch Toxicol 76, 30-9.

                                               C-27

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   ATSDR (1997). Toxicological profile for chlorpyrifos. Agency for Toxic Substances and Disease
      Registry, Atlanta, GA.
   Atterberry, T. T., Burnett, W. T., and Chambers, J. E. (1997). Age-related differences in parathion and
      chlorpyrifos toxicity in male rats: target and nontarget esterase sensitivity and cytochrome P450-
      mediated metabolism. Toxicol ApplPharmacol 147, 411-8.
   Auman, J. T., Seidler, F. J., and Slotkin, T. A. (2000). Neonatal chlorpyrifos exposure targets multiple
      proteins governing the hepatic adenylyl cyclase signaling cascade: implications for neurotoxicity.
      Brain Res Dev Brain Res 121, 19-27.
   Axelrad, J. C., Howard, C. V., and McLean, W. G. (2002). Interactions between pesticides and
      components of pesticide formulations in an in vitro neurotoxicity test. Toxicology 173, 259-68.
   Bakke, J. E., Feil, V. J., and Price, C. E. (1976). Rat urinary metabolites from O,O-diethyl-O-(3,5,6-
      trichloro-2-pyridyl) phosphorothioate. J Environ Sci Health B 11, 225-30.
   Ballantyne, B., andMarrs, T. C. (1992). Clinical and experimental toxicology of organophosphates and
      carbamates. Buterworth-Heinemann, Ltd., Lindacre House, Jordan Hill Oxford.
   Barber, D. S., and Ehrich, M. (2001). Esterase inhibition in SH-SY5Y human neuroblastoma cells
      following exposure to organophosphorus compounds for 28 days. In Vitr Mol Toxicol 14, 129-35.
   Brophy, V. H., Jarvik, G. P., Richter, R. J., Rozek, L.  S., Schellenberg, G. D., and Furlong, C. E. (2000).
      Analysis of paraoxonase (PON1) L55M status requires both genotype and phenotype.
      Pharmacogenetics 10, 453-60.
   Buratti, F. M., Volpe, M. T., Meneguz, A., Vittozzi, L., and Testai, E. (2003). CYP-specific
      bioactivation of four organophosphorothioate pesticides by human liver microsomes. Toxicol Appl
      Pharmacol 186, 143-54.
   Bushnell, P. J., Kelly, K. L., and Ward, T. R. (1994). Repeated inhibition of cholinesterase by
      chlorpyrifos in rats: behavioral, neurochemical and pharmacological indices of tolerance. J
      Pharmacol Exp Ther 270, 15-25.
   Bushnell, P. J., Pope, C. N., and Padilla, S. (1993). Behavioral and neurochemical effects of acute
      chlorpyrifos in rats: tolerance to prolonged inhibition of cholinesterase. J Pharmacol Exp Ther 266,
      1007-17.
   Campbell, C.  G., Seidler, F. J., and Slotkin, T. A. (1997). Chlorpyrifos interferes with cell development
      in rat brain regions. Brain Res Bull 43, 179-89.
   Carr, R. L., Chambers, H. W., Guarisco, J. A., Richardson, J. R., Tang, J., and Chambers, J. E. (2001).
      Effects of repeated oral postnatal exposure to chlorpyrifos on open- field behavior in juvenile rats.
      Toxicol Sci 59, 260-7.
   Carr, R. L., and Chambers, J. E. (1996). Kinetic analysis of the in vitro inhibition, aging, and
      reactivation of brain acetylcholinesterase from rat and channel catfish by paraoxon and chlorpyrifos-
      oxon. Toxicol Appl Pharmacol 139, 365-73.
   Carr, R. L., Richardson, J. R., Guarisco, J. A., Kachroo, A.,  Chambers, J. E., Couch, T. A., Durunna, G.
      C., and Meek, E. C. (2002). Effects of PCB exposure on the toxic impact of organophosphorus
      insecticides. Toxicol Sci 67, 311-21.
   Carr, R. L., Straus, D. L., and Chambers, J. E. (1995). Inhibition and aging of channel catfish brain
      acetylcholinesterase following exposure to two phosphorothionate insecticides and their active
      metabolites. J Toxicol Environ Health 45, 325-36.
   Chakraborti, T. K., Farrar, J. D., and Pope, C. N. (1993). Comparative neurochemical and
      neurobehavioral effects of repeated chlorpyrifos exposures in young and adult rats. Pharmacol
      Biochem Behav 46, 219-24.
   Chanda, S. M., Mortensen, S. R., Moser, V. C., and Padilla, S. (1997). Tissue-specific effects of
      chlorpyrifos on carboxylesterase and cholinesterase activity in adult rats: an in vitro and in vivo
      comparison. Fundam Appl Toxicol 38, 148-57.
                                                C-28

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Chanda, S. M., and Pope, C. N. (1996). Neurochemical and neurobehavioral effects of repeated
      gestational exposure to chlorpyrifos in maternal and developing rats. PharmacolBiochem Behav 53,
      771-6.
   Chiappa,  S., Padilla, S., Koenigsberger, C., Moser, V., and Brimijoin, S. (1995). Slow accumulation of
      acetylcholinesterase in rat brain during enzyme inhibition by repeated dosing with chlorpyrifos.
      Biochem Pharmacol 49, 955-63.
   Cocker, J., Mason, H. J., Garfitt, S. J., and Jones, K. (2002). Biological monitoring of exposure to
      organophosphate pesticides. Toxicol Lett 134, 97-103.
   Cook, T. J., and Shenoy, S. S. (2003). Intestinal permeability of chlorpyrifos using the single-pass
      intestinal perfusion method in the rat. Toxicology 184, 125-33.
   Costa, L.  G., Li, W. F., Richter, R. J., Shih, D. M., Lusis, A., and Furlong, C. E. (1999). The role of
      paraoxonase (PON1) in the detoxication of organophosphates and its human polymorphism. Chem
      Biol Interact 119-120, 429-38.
   Costa, L.  G., Richter, R. J., Li, W. F., Cole, T., Guizzetti, M., and Furlong,  C. E. (2003). Paraoxonase
      (PON 1) as a biomarker of susceptibility for organophosphate toxicity. Biomarkers 8, 1-12.
   Cowan, J., Sinton, C. M., Varley, A. W., Wians, F. H., Haley, R.  W., and Munford, R. S.  (2001). Gene
      therapy to prevent organophosphate intoxication. Toxicol ApplPharmacol 173,  1-6.
   Crumpton, T. L., Seidler, F. J., and Slotkin, T. A. (2000). Developmental neurotoxicity of chlorpyrifos
      in vivo and in vitro: effects on nuclear transcription factors involved in  cell  replication and
      differentiation. Brain Res 857, 87-98.
   Dai, D., Tang, J., Rose, R., Hodgson, E., Bienstock, R. J., Mohrenweiser, H. W., and Goldstein, J. A.
      (2001). Identification of variants of CYP3A4 and characterization of their abilities to  metabolize
      testosterone and chlorpyrifos. JPharmacol Exp Ther 299, 825-31.
   Dam, K., Garcia, S. J., Seidler, F. J., and Slotkin, T. A.  (1999). Neonatal chlorpyrifos exposure alters
      synaptic development and neuronal activity in cholinergic and catecholaminergic pathways. Brain
      Res Dev Brain Res 116, 9-20.
   Dam, K., Seidler, F. J., and Slotkin, T. A. (1998). Developmental neurotoxicity of chlorpyrifos: delayed
      targeting of DNA synthesis after repeated administration. Brain Res Dev Brain Res 108, 39-45.
   Dam, K., Seidler, F. J., and Slotkin, T. A. (2000). Chlorpyrifos exposure during a critical  neonatal
      period elicits gender- selective deficits in the development of coordination skills and locomotor
      activity. Brain Res Dev Brain Res 121, 179-87.
   Dam, K., Seidler, F. J., and Slotkin, T. A. (2003). Transcriptional biomarkers distinguish  between
      vulnerable periods for developmental neurotoxicity of chlorpyrifos:  Implications for
      toxicogenomics. Brain Res Bull 59, 261-5.
   Das,  K. P., and Barone, S., Jr. (1999). Neuronal differentiation in PC12  cells is  inhibited by chlorpyrifos
      and its metabolites: is acetylcholinesterase inhibition the site of action?  Toxicol Appl Pharmacol
      160,217-30.
   Drevenkar, V., Vasilic, Z., Stengl, B., Frobe, Z., and Rumenjak, V.  (1993).  Chlorpyrifos metabolites in
      serum and urine of poisoned persons. Chem Biol Interact 87, 315-22.
   Ehrich, M., Correll, L., and Veronesi, B. (1997). Acetylcholinesterase and neuropathy target esterase
      inhibitions in neuroblastoma cells to distinguish organophosphorus compounds causing acute and
      delayed neurotoxicity. Fundam Appl Toxicol 38, 55-63.
   Fenske, R. A., Lu, C., Barr, D., and Needham, L. (2002). Children's exposure to chlorpyrifos and
      parathion in an agricultural community in central Washington State. Environ Health Perspect 110,
      549-53.
   Furlong, C. E., Li, W. F., Brophy,  V. H., Jarvik, G. P., Richter, R. J., Shih,  D. M., Lusis, A. J., and
      Costa, L. G. (2000a). The PON1 gene and detoxication. Neurotoxicology 21, 581-7.
                                                C-29

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Furlong, C. E., Li, W. F., Costa, L. G., Richter, R. I, Shih, D. M., and Lusis, A. J. (1998). Genetically
      determined susceptibility to organophosphorus insecticides and nerve agents: developing a mouse
      model for the human PON1 polymorphism. Neurotoxicology 19, 645-50.
   Furlong, C. E., Li, W. F., Richter, R. J., Shih, D. M., Lusis, A. J., Alleva, E., and Costa, L. G. (2000b).
      Genetic and temporal determinants of pesticide sensitivity: role of paraoxonase (PON1).
      Neurotoxicology 21, 91-100.
   Garcia, S. J., Seidler, F. J., Qiao, D., and Slotkin, T. A. (2002). Chlorpyrifos targets developing glia:
      effects on glial fibrillary acidic protein. Brain Res Dev Brain Res 133, 151-61.
   Garcia, S. J., Seidler, F. J., and Slotkin, T. A. (2003). Developmental neurotoxicity elicited by prenatal
      or postnatal chlorpyrifos exposure: effects on neurospecific proteins indicate changing
      vulnerabilities. Environ Health Perspect 111, 297-304.
   Gore, A. C. (2001). Environmental toxicant effects on neuroendocrine function. Endocrine 14, 235-46.
   Gore, A. C. (2002). Organochlorine pesticides directly regulate gonadotropin-releasing hormone gene
      expression and biosynthesis in the GT1-7 hypothalamic cell line. Mol Cell Endocrinol 192, 157-70.
   Griffin, P., Payne, M., Mason, H., Freedlander, E., Curran, A. D., and Cocker, J. (2000). The in vitro
      percutaneous penetration of chlorpyrifos. Hum Exp Toxicol 19, 104-7.
   Howard, M. D., and Pope,  C. N. (2002). In vitro effects of chlorpyrifos, parathion, methyl parathion and
      their oxons on cardiac muscarinic receptor binding in neonatal and adult rats. Toxicology 170, 1-10.
   Hunter, D. L., Lassiter, T. L., and Padilla, S. (1999). Gestational exposure to chlorpyrifos: comparative
      distribution  of trichloropyridinol in the fetus and dam. ToxicolApplPharmacol 158,  16-23.
   Jett, D. A., Navoa, R. V., Beckles, R. A., and McLemore, G. L. (2001). Cognitive function and
      cholinergic neurochemistry in weanling rats exposed to chlorpyrifos. Toxicol Appl Pharmacol 174,
      89-98.
   Karanth, S., Olivier, K., Jr., Liu, J., and Pope, C. (2001). In vivo interaction between chlorpyrifos and
      parathion in adult rats:  sequence  of administration can markedly influence toxic outcome. Toxicol
      Appl Pharmacol 111, 247-55.
   Karanth, S., and Pope, C. (2000). Carboxylesterase and A-esterase activities during maturation and
      aging: relationship to the toxicity of chlorpyrifos and parathion in rats. Toxicol Sci 58, 282-9.
   Katz, E. J.,  Cortes, V. L, Eldefrawi, M. E.,  and Eldefrawi, A. T. (1997). Chlorpyrifos, parathion, and
      their oxons bind to and desensitize a nicotinic acetylcholine receptor: relevance to their toxicities.
      Toxicol Appl Pharmacol 146, 227-36.
   Lassiter, T. L., Barone, S.,  Jr., Moser, V. C., and Padilla,  S. (1999). Gestational  exposure to
      chlorpyrifos: dose response profiles for cholinesterase and Carboxylesterase  activity.  Toxicol Sci 52,
      92-100.
   Lassiter, T. L., Padilla, S., Mortensen, S. R., Chanda, S. M., Moser, V.  C., and Barone, S., Jr. (1998).
      Gestational exposure to chlorpyrifos: apparent protection of the fetus? Toxicol Appl Pharmacol 152,
      56-65.
   Levin, E. D., Addy, N., Baruah, A., Elias, A., Christopher, N. C., Seidler, F. J., and Slotkin, T. A.
      (2002).  Prenatal  chlorpyrifos exposure  in rats causes persistent behavioral alterations. Neurotoxicol
      Teratol 24, 733-41.
   Levin, E. D., Addy, N., Nakajima, A., Christopher, N. C., Seidler, F. J., and Slotkin, T. A. (2001).
      Persistent behavioral consequences of neonatal chlorpyrifos exposure in rats. Brain Res Dev Brain
      Res 130, 83-9.
   Li, W. F., Costa, L. G., Richter, R. J., Hagen, T., Shih, D. M., Tward, A., Lusis, A. J., and Furlong,  C. E.
      (2000).  Catalytic efficiency determines the in-vivo efficacy of PON1 for detoxifying
      organophosphorus compounds. Pharmacogenetics  10, 767-79.
   Li, W. F., Furlong, C. E., and Costa, L.  G. (1995). Paraoxonase protects against chlorpyrifos toxicity in
      mice. Toxicol Lett 76, 219-26.
                                                 C-30

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Liu, J., Chakraborti, T., and Pope, C. (2002). In vitro effects of organophosphorus anticholinesterases on
      muscarinic receptor-mediated inhibition of acetylcholine release in rat striatum. ToxicolAppl
      Pharmacol 178, 102-8.
   Liu, J., Olivier, K., and Pope, C. N. (1999). Comparative neurochemical effects of repeated methyl
      parathion or chlorpyrifos exposures in neonatal and adult rats. Toxicol ApplPharmacol 158, 186-96.
   Ma, T., and Chambers, J. E. (1994). Kinetic parameters of desulfuration and dearylation of parathion
      and chlorpyrifos by rat liver microsomes. Food Chem Toxicol 32, 763-7.
   Mattsson, J. L., Maurissen, J. P., Nolan, R. J., and Brzak, K. A. (2000). Lack of differential sensitivity to
      cholinesterase inhibition in fetuses and neonates compared to dams treated perinatally with
      chlorpyrifos. ToxicolSci 53, 438-46.
   Monnet-Tschudi, F., Zurich, M. G., Schilter, B., Costa, L. G., and Honegger, P. (2000). Maturation-
      dependent effects of chlorpyrifos and parathion and their oxygen analogs on acetylcholinesterase and
      neuronal and glial markers in aggregating brain cell cultures. Toxicol Appl Pharmacol 165, 175-83.
   Mortensen, S. R., Chanda,  S. M., Hooper, M. J., and Padilla, S. (1996). Maturational differences in
      chlorpyrifos-oxonase activity may contribute to age-related sensitivity to chlorpyrifos. JBiochem
      Toxicol 11,279-87.
   Mortensen, S. R., Hooper, M. J., and Padilla, S. (1998). Rat brain acetylcholinesterase activity:
      developmental profile and maturational sensitivity to carbamate and organophosphorus inhibitors.
      Toxicology 125, 13-9.
   Moser, V. C., Chanda, S. M., Mortensen, S. R., and Padilla, S. (1998). Age- and gender-related
      differences in sensitivity to chlorpyrifos in the rat reflect developmental profiles of esterase
      activities. Toxicol Sci 46, 211-22.
   Moser, V. C., and Padilla, S. (1998). Age- and gender-related differences in the time course of
      behavioral and biochemical effects produced by oral chlorpyrifos in rats. Toxicol Appl Pharmacol
      149, 107-19.
   Nolan, R. J., Rick, D. L., Freshour, N. L., and Saunders, J. H. (1984). Chlorpyrifos: pharmacokinetics in
      human volunteers. Toxicol Appl Pharmacol73, 8-15.
   Olivier, K., Jr., Liu, J., and Pope, C. (2001). Inhibition of forskolin-stimulated cAMP formation in vitro
      by paraoxon and chlorpyrifos oxon in cortical slices from neonatal, juvenile, and adult rats. J
      Biochem Mol Toxicol 15, 263-9.
   Padilla, S., Buzzard, J., and Moser, V. C. (2000). Comparison of the role of esterases in the differential
      age-related sensitivity to chlorpyrifos and methamidophos. Neurotoxicology 21, 49-56.
   Padilla, S., Wilson, V. Z., and Bushnell, P. J. (1994). Studies on the correlation between blood
      cholinesterase inhibition and 'target tissue' inhibition in pesticide-treated rats. Toxicology 92, 11-25.
   Poet, T. S.,  Wu, H., Kousba, A. A., and Timchalk, C. (2003). In vitro rat hepatic and intestinal
      metabolism of the organophosphate pesticides chlorpyrifos and diazinon. Toxicol Sci 72, 193-200.
   Pond, A. L., Chambers, H. W., and Chambers, J. E. (1995). Organophosphate detoxication potential of
      various  rat tissues via A- esterase and aliesterase activities. Toxicol Lett 78, 245-52.
   Pond, A. L., Chambers, H. W., Coyne, C. P., and Chambers, J. E. (1998). Purification of two rat hepatic
      proteins with A-esterase activity toward chlorpyrifos-oxon and paraoxon. JPharmacol Exp Ther
      286, 1404-11.
   Qiao, D., Seidler, F. J., Padilla, S., and Slotkin, T. A. (2002). Developmental neurotoxicity of
      chlorpyrifos: what is the vulnerable period? Environ Health Perspect  110, 1097-103.
   Qiao, D., Seidler, F. J., and Slotkin, T. A. (2001). Developmental neurotoxicity of chlorpyrifos modeled
      in vitro: comparative effects of metabolites and other cholinesterase inhibitors on DNA synthesis in
      PC12 and C6 cells. Environ Health Perspect 109, 909-13.
   Qiao, D., Seidler, F. J., Tate, C. A., Cousins, M. M., and Slotkin, T. A. (2003). Fetal chlorpyrifos
      exposure: adverse effects on brain cell development and cholinergic biomarkers emerge postnatally
      and continue into adolescence and adulthood. Environ Health Perspect  111, 536-44.

                                                C-31

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Raines, K. W., Seidler, F. J., and Slotkin, T. A. (2001). Alterations in serotonin transporter expression in
      brain regions of rats exposed neonatally to chlorpyrifos. Brain Res Dev Brain Res 130, 65-72.
   Richardson, J., and Chambers, J. (2003). Effects of gestational exposure to chlorpyrifos on postnatal
      central and peripheral cholinergic neurochemistry. J Toxicol Environ Health A 66, 275-89.
   Richardson, J. R., Chambers, H. W., and Chambers, J. E. (2001). Analysis of the additivity of in vitro
      inhibition of cholinesterase by mixtures of chlorpyrifos-oxon and azinphos-methyl-oxon.  Toxicol
      Appl Pharmacol 172, 128-39.
   Rigas, M. L., Okino, M. S., and Quackenboss, J. J. (2001). Use of a pharmacokinetic model to assess
      chlorpyrifos exposure and dose in children, based on urinary biomarker measurements. Toxicol Sci
      61,374-81.
   Roy, T. S., Andrews, J. E., Seidler, F. J., and Slotkin, T. A. (1998). Chlorpyrifos elicits mitotic
      abnormalities and apoptosis in neuroepithelium of cultured rat embryos. Teratology 58, 62-8.
   Sachana, M., Flaskos, J., Alexaki, E., Glynn, P., and Hargreaves, A. J. (2001). The toxicity of
      chlorpyrifos towards differentiating mouse N2a neuroblastoma cells. Toxicol In Vitro 15,  369-72.
   Sams, C., and Mason, H. J. (1999). Detoxification of organophosphates by A-esterases in human serum.
      Hum Exp Toxicol 18, 653-8.
   Sams, C., Mason, H. J.,  and Rawbone, R. (2000). Evidence for the activation of organophosphate
      pesticides by cytochromes P450 3A4 and 2D6 in human liver microsomes. Toxicol Lett 116, 217-21.
   Sartorelli, P., Aprea, C., Cenni, A., Novell!, M. T., Orsi, D., Palmi, S., and Matteucci, G. (1998).
      Prediction of percutaneous absorption from physicochemical data: a model based on data  of in vitro
      experiments. Ann Occup Hyg 42, 267-76.
   Shah, P. V., Fisher, H. L., Sumler, M. R., Monroe, R. J., Chernoff, N., and Hall, L. L. (1987).
      Comparison of the penetration of 14 pesticides through the skin of young and adult rats. J Toxicol
      Environ Health 21, 353-66.
   Shah, P. V., Monroe, R. J., and Guthrie, F. E. (1981). Comparative rates of dermal penetration of
      insecticides in mice. Toxicol Appl Pharmacol 59, 414-423.
   Slotkin, T. A., Cousins,  M.  M., Tate, C. A., and Seidler, F. J. (2001). Persistent cholinergic presynaptic
      deficits after neonatal chlorpyrifos exposure. Brain Res 902, 229-43.
   Slotkin, T. A., Tate, C. A., Cousins, M. M., and Seidler, F. J. (2002). Functional alterations in CNS
      catecholamine systems in adolescence and adulthood after neonatal chlorpyrifos exposure. Brain Res
      Dev Brain Res 133,  163-73.
   Smith, G. N., Watson, B. S., and Fischer, F. S. (1967). Investigations on dursban insecticide:
      Metabolism of [36C1] O,O-diethyl-O- 3,5,6-trichloro-2-pyridyl phosphorothioate in  rats. J Agric
      FoodChem 15, 132-8.
   Song, X., Violin, J. D., Seidler, F. J., and Slotkin, T.  A. (1998). Modeling the developmental
      neurotoxicity of chlorpyrifos in vitro: macromolecule synthesis in PC 12 cells. Toxicol Appl
      Pharmacol 151, 182-91.
   Stanton, M. E., Mundy,  W. R., Ward, T., Dulchinos, V., and Barry, C. C. (1994). Time-dependent
      effects of acute chlorpyrifos administration on spatial delayed alternation and cholinergic
      neurochemistry in weanling rats. Neurotoxicology 15, 201-8.
   Sultatos, L. G., Minor, L. D., and Murphy, S. D. (1985). Metabolic activation of phosphorothioate
      pesticides: role of the liver. J Pharmacol Exp Ther 232, 624-8.
   Sultatos, L. G., and Murphy, S. D. (1983a). Hepatic microsomal detoxification of the organophosphates
      paraoxon and chlorpyrifos oxon in the mouse. DrugMetab Dispos 11, 232-8.
   Sultatos, L. G., and Murphy, S. D. (1983b). Kinetic analyses of the microsomal biotransformation of the
      phosphorothioate insecticides chlorpyrifos and parathion. FundamAppl Toxicol 3, 16-21.
   Tang, J., Cao, Y., Rose, R. L., Brimfield, A. A., Dai, D., Goldstein, J. A., and Hodgson, E. (2001).
      Metabolism of chlorpyrifos by human cytochrome P450 isoforms and human, mouse, and rat liver
      microsomes. DrugMetab Dispos 29, 1201-4.

                                                C-32

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Tang, J., Cao, Y., Rose, R. L., and Hodgson, E. (2002). In vitro metabolism of carbaryl by human
      cytochrome P450 and its inhibition by chlorpyrifos. Chem BiolInteract 141, 229-41.
   Tang, J., Carr, R. L., and Chambers, J. E. (1999). Changes in rat brain cholinesterase activity and
      muscarinic receptor density during and after repeated oral exposure to chlorpyrifos in early postnatal
      development. Toxicol Sci 51, 265-72.
   Timchalk, C., Kousba, A., and Poet, T. S. (2002a). Monte Carlo analysis of the human chlorpyrifos-
      oxonase (PON1) polymorphism using a physiologically based pharmacokinetic and
      pharmacodynamic (PBPK/PD)  model. Toxicol Lett 135, 51-9.
   Timchalk, C., Nolan, R. J., Mendrala, A. L., Dittenber, D. A., Brzak, K. A., and Mattsson, J. L. (2002b).
      A physiologically based pharmacokinetic and pharmacodynamic (PBPK/PD) model for the
      organophosphate insecticide chlorpyrifos in rats and humans. Toxicol Sci 66, 34-53.
   Usmani, K. A., Rose, R. L., Goldstein, J. A., Taylor, W. G., Brimfield, A. A., and Hodgson, E. (2002).
      In vitro human metabolism and interactions of repellent N,N-diethyl-m- toluamide. DrugMetab
      Dispos 30, 289-94.
   Usmani, K. A., Rose, R. L., and Hodgson, E. (2003). Inhibition and activation of the human liver
      microsomal and human cytochrome P450 3A4 metabolism of testosterone by deployment-related
      chemicals. DrugMetab Dispos 31, 384-91.
   Whitney, K. D.,  Seidler, F. J., and Slotkin, T. A. (1995). Developmental neurotoxicity of chlorpyrifos:
      cellular mechanisms. Toxicol ApplPharmacol 134, 53-62.
   Won, Y. K., Liu, J., Olivier, K., Jr., Zheng, Q., and Pope, C. N. (2001). Age-related effects of
      chlorpyrifos  on acetylcholine release in rat brain. Neurotoxicology 22, 39-48.
                                                C-33

-------
USEPA Contract No. 3C-R102-NTEX
                                                 Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
                                             Diazinon
   4.0    Introduction
  Diazinon (0,0-diethyl O-2 isopropyl-6-methylpyrimidinyl phosphothiolate) is an organophosphate
  insecticide that is used in agriculture and as a topically applied pesticide in animal use (ATSDR 1996;
  Garfitt et al. 2002; USEPA 2003). It is a colorless liquid and was available as granules, emulsifiable
  concentrate, dust, and wettable powder.  It is soluble in most organic solvents and is stable in neutral
  media but is slowly hydrolyzed in alkaline media and more rapidly in acidic media (HSDB 2003).

  4.1    Toxic effects

  Many effects have been determined to occur in chronic bioassays in laboratory animals. Some of the
  target organs affected include the respiratory, kidney, cardiovascular, gastrointestinal, hematological,
  hepatic, endocrine, lymphatic, reproductive (ATSDR 1996), immune (Galloway and Handy 2003), and
  nervous systems (Gordon and Mack 2003).

  Acute oral LDSOs in rats between 76 and 408 mg/kg have been reported (ATSDR 1996).  Dermal
  LDSO's in rats ranged between 455  and 1100 mg/kg (ATSDR 1996).

  Diazinon has been reported to cause genotoxicity in a number of assays, including in the S.
  typhimurium, mouse lymphoma cell forward mutation assay, and Chinese hamster cell chromosomal
  aberration assay, but was negative in several other assays (ATSDR 1996; Hatjian et al. 2000).

  Risk assessment for exposure to diazinon, however, has generally been based on inhibition of brain
  acetylcholinesterase (AChE) as the  critical endpoint of toxicity (ATSDR 1996).  As with other OP
  pesticides, the mode of action of diazinon is inhibition of AChE in the central and peripheral nervous
  system. Diazinon is a weak inhibitor of AChE while the oxon analog is much more potent.  Therefore,
  activation by mixed function oxygenases, primarily in the liver, is an important bioactivating step.
  Other metabolic pathways (see Figure 4.1 below) are generally detoxifying.  Symptoms of acute toxic
  exposure include vomiting, unconsciousness, giddiness, sweating, diarrhea, tachycardia, muscle
  fasciculations, abdominal pain, and bronchospasm (ATSDR 1996).

  ATSDR has published an oral Minimal Risk Level for intermediate term exposure to diazinon of 0.0002
  mg/kg/day (ATSDR 1996). USEPA established a chronic reference dose (RfD) of 0.0002 mg/kg/day in
  the diet (USEPA 2000).
   4.2    Pharmacokinetics

   Diazinon pharmacokinetics are qualitatively similar to other organophosphate pesticides described in
   this report.

   4.2.1  Absorption

   Several oral absorption studies have been performed. 85% of the single oral dose of 4.0 mg/kg diazinon
   was absorbed by Beagle dogs in one study (Iverson et al. 1975).  Other oral absorption studies were
                                               C-34

-------
USEPA Contract No. 3C-R102-NTEX
                                                  Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
   conducted in rats, goats, sheep and cows (Abdelsalam and Ford 1986; Janes et al. 1973; Machin et al.
   1974; Machin et al. 1971; Mount 1984; Wu et al. 1996a). In a dermal study, human volunteers absorbed
   34% of the dose applied to the abdomen or forearm for 24 hours (Wester et al 1993).

   4.2.2  Distribution

   Diazinon is found widely distributed in all tissues examined after oral absorption (Abdelsalam and Ford
   1986; de Blaquiere et al 2000; Janes et al 1973; Machin et al 1974; Machin et al 1971; Mucke et al
   1970; Tomokuni and Hasegawa  1985; Tomokuni et al 1985). No studies of distribution after inhalation
   or dermal exposures are available. After an i.v. dose of 0.2 mg/kg in ethanol, the terminal halflife was
   1.5 hours (Iverson et al. 1975).  Distribution coefficients for diazinon were reported (Garcia-Repetto et
   al  1995).

   4.2.3  Metabolism

   The principal metabolic pathways of diazinon are shown in Figure 4.1.

                   Diazinon

                     CH,
      CH3CH2
                                                            CHCH,
                      2-isopropyl-4-methyl-6-hydroxypyrimidine
                     plus diethylphosphate or diethylthiophosphate

             Figure 4.1  Principal metabolic pathways for diazinon. Adapted from Poet, 2003.

   Diazinon is subject to oxidative desulfurization and hydrolysis of the ester. Hydrolysis of the ester can
   occur before or after desulfurization, i.e., either diazinon or diazoxon can be hydrolyzed to yield 2-
   isopropyl-4-methyl-6-hydroxypyrimidine (IMHP) and either diethylphosphate or diethylthiophosphate
   (Iverson et al. 1975; Machin et al.  1975; Mucke et al. 1970); however, the P450 catalyzed oxidative
   cleavage of phosphorothioate (i.e., diazinon) triester bond is much more efficient (Yang et al 1971).
   Desulfurization is mediated by cytochrome P450 isoenzymes while oxidative cleavage or hydrolysis is
   mediated by cytochrome P450 or various esterases, respectively (Poet et al 2003b; Walker and
   Mackness 1987). Diazinon and its metabolites may also be oxidized at alkyl carbons (Aizawa 1989;
   Yang etal  1971).
                                                C-35

-------
USEPA Contract No. 3C-R102-NTEX                         Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   In a human self-poisoning case, diazinon was found in serum and monoethyl phosphate, diethyl
   phosphate, and diethyl phosphorothioate were detected in the urine (Klemmer et al.  1978). In rat liver
   microsomes, diazoxon hydrolysis occurred without NADPH (Yang et al. 1971).

   Rates of metabolism in rat liver and intestinal microsomes of diazinon and diazoxon were recently
   reported by Poet and coworkers. The authors measured CYP450 mediated oxidative desulfurization (to
   form diazoxon) or hydrolysis (to form IMHP and diethylthiophosphate) as well as hydrolysis by esterase
   (PON1) in microsomes from both tissues.  Based on the measured rates, the  authors conclude that
   intestinal metabolism may be important, especially for low level oral doses (Poet et al. 2003b). Also in
   microsomes, Vittozzi et al. measured the activity of expressed cytochrome P450s for desulfurization and
   hydrolysis (Vittozzi et al. 2001).  Significant activity was obtained for all nine cytochrome P450s tested
   that varied over about one order of magnitude (CYP2C19, 3A4, 2B6, 1A2, 1A1, 2C8, 2C9, 2D6, and
   2A6).   A separate study by the same group indicated that CYP2C11, 3A2, and 2B1/2 were involved but
   that 2E1 and 1A1 were not (Fabrizi et al. 1999). Kappers et al.  indicated that CYP2C19 was the major
   isoform involved in diazinon metabolism, but that others such as CYP1A2 and CYP3 A4 may also be
   showing some activity (Kappers et al. 2001).  However, using immunoinhibition and other techniques,
   Buratti  and coworkers found that the principal isoforms involved in diazinon metabolism were CYP3 A4,
   1A2, and 2B6 (Buratti et al. 2003), while Sams et al. felt that CYP 2D6 and 3 A4 were the most
   important isozymes (Sams et al. 2000).

   Toxicity and acetylcholinesterase inhibition was studied in PON1 knockout  mice (Li et al. 2000).

   Rates of inhibition of acetylcholinesterase were measured in some studies (Kamal and Al-Jafari 2000).

   4.2.4   Excretion

   Most diazinon is excreted as metabolites in urine, while smaller amounts are excreted in feces or, after
   extensive metabolism, as CO2 in expired air.  Approximately 75% of a 4.0 mg/kg oral dose to rats was
   excreted as urinary metabolites, 20% in the feces, and about 6% as CO2 (Mucke et al. 1970).
   Approximately 85% of total label was recovered in a 24 hour urine sample from dogs receiving a single
   oral dose of diazinon. After an i.v. dose, the dogs excreted approximately 58% of the label in urine
   (Iverson et al.  1975). In human volunteers, urinary excretion of diethyl phosphate and diethyl
   thiophosphate was reported after oral and dermal dosing (Cocker et al. 2002; Garfitt et al. 2002).
   Diazinon has been found in hair (of rabbits) as a potential biomarker of exposure (Tutudaki et al.  2003).
   Blood cholinesterase inhibition has also been used as a biomarker for diazinon exposure (Nigg and
   Knaak 2000).

   4.2.5   Special populations and variability

   Several studies have suggested high variability in human metabolism of diazinon (Buratti et al. 2003;
   Kappers et al.  2001). Polymorphisms in PON1 have been described (Brophy et al. 2000; Cherry et al.
   2002; Costa et al. 2003; Davies et al.  1996; Mackness  et al. 2003).

   4.3     Interactions with other chemicals

   The toxicity was increased and pharmacokinetics of diazinon were affected by cimetidine (Wu et al.
   1996b). Interactions between diazinon and methyl parathion were reported in the blood and brain of
   pregnant rats and the fetus after a single dermal dose (Abu-Qare and Abou-Donia 2001). Neurite
   outgrowth was assessed for mixtures of diazinon and chlorpyrifos (Axelrad et al. 2002).

                                               C-36

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
   4.4    Diazinon PBPK models

   One PBPK model for diazinon has been published in abstract form (Poet et al. 2003a).  The model
   incorporated an oral exposure route, desulfurization and "hydrolysis" (an oxidation reaction actually) by
   a CYP450 enzyme, hydrolysis of the oxon by PON1 in liver and blood, and second order binding and
   inhibition and regeneration of B-esterases in the liver, blood, diaphragm, and brain.

   4.5    Literature Cited

   Abdelsalam, E. B., and Ford, E. J. (1986). Effect of pretreatment with hepatic microsomal enzyme
      inducers on the toxicity of diazinon in calves. Res Vet Sci 41, 336-9.
   Abu-Qare, A. W., and Abou-Donia, M. B. (2001). Inhibition and recovery of maternal and fetal
      cholinesterase enzyme activity following a single cutaneous dose of methyl parathion and diazinon,
      alone and in combination, in pregnant rats. J Appl Toxicol 21, 307-16.
   Aizawa, H.  (1989). Metabolic maps of pesticides. Vol. 2.  Ecotoxicology and Environmental Quality
      Series. Academic Press.
   ATSDR (1996). Toxicological profile  for diazinon.  Agency for Toxic Substances and Disease Registry,
      Atlanta, GA.
   Axelrad, J. C., Howard, C. V., and McLean, W. G. (2002). Interactions between pesticides and
      components of pesticide formulations in an in vitro neurotoxicity test. Toxicology 173, 259-68.
   Brophy, V. H., Jarvik, G. P., Richter, R. J., Rozek, L.  S., Schellenberg, G. D., and Furlong, C. E. (2000).
      Analysis of paraoxonase (PON1) L55M status requires both genotype and phenotype.
      Pharmacogenetics 10, 453-60.
   Buratti, F. M., Volpe, M. T., Meneguz, A., Vittozzi, L., and Testai, E. (2003). CYP-specific
      bioactivation of four organophosphorothioate pesticides by human liver microsomes. Toxicol Appl
      Pharmacol 186, 143-54.
   Cherry, N., Mackness, M., Durrington, P., Povey, A.,  Dippnall, M., Smith,  T., and Mackness, B. (2002).
      Paraoxonase (PON1) polymorphisms in farmers attributing ill health to sheep dip. Lancet 359, 763-
      4.
   Cocker, J., Mason, H. J.,  Garfitt, S. J.,  and Jones, K. (2002). Biological monitoring of exposure to
      organophosphate pesticides. Toxicol Lett 134, 97-103.
   Costa, L. G., Richter, R. J., Li, W. F., Cole, T., Guizzetti, M., and Furlong,  C. E.  (2003). Paraoxonase
      (PON 1) as a biomarker of susceptibility for organophosphate toxicity. Biomarkers 8, 1-12.
   Davies, H. G., Richter, R. J., Keifer, M., Broomfield, C. A., Sowalla, J., and Furlong, C. E. (1996). The
      effect of the human serum paraoxonase polymorphism is reversed with  diazoxon, soman and sarin.
      Nat Genet 14, 334-6.
   de Blaquiere, G. E., Waters, L., Blain,  P. G., and Williams, F. M. (2000). Electrophysiological and
      biochemical effects of single and multiple doses of the organophosphate diazinon in the mouse.
      Toxicol Appl Pharmacol 166, 81-91.
   Fabrizi, L., Gemma,  S., Testai, E., and Vittozzi, L. (1999). Identification of the cytochrome P450
      isoenzymes involved in the metabolism of diazinon in the rat liver. JBiochem Mol Toxicol 13, 53-
      61.
   Galloway, T., and Handy, R. (2003). Immunotoxicity  of organophosphorous pesticides. Ecotoxicology
      12, 345-63.
   Garcia-Repetto, R., Martinez, D., and Repetto, M. (1995). Coefficient of distribution of some
      organophosphorous pesticides in rat tissue. Vet Hum Toxicol 37,  226-9.
   Garfitt, S. J., Jones, K., Mason, H. J., and Cocker, J. (2002). Exposure to the organophosphate diazinon:
      data from a human volunteer study with oral and dermal doses. Toxicol Lett 134, 105-13.

                                               C-37

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Gordon, C. J., and Mack, C. M. (2003). Influence of gender on thermoregulation and cholinesterase
      inhibition in the long-evans rat exposed to diazinon. J ToxicolEnviron Health A 66, 291-304.
   Hatjian, B. A., Mutch, E., Williams, F. M., Blain, P. G., and Edwards, J. W. (2000). Cytogenetic
      response without changes in peripheral cholinesterase enzymes following exposure to a sheep dip
      containing diazinon in vivo and in vitro. MutatRes 472, 85-92.
   HSDB (2003). Diazinon, Vol. 2003. Hazard Substances Data Bank.
   Iverson, F., Grant, D. L., and Lacroix, J. (1975). Diazinon metabolism in the dog. Bull Environ Contam
      Toxicol 13, 611-8.
   Janes, N. F.,  Machin, A. F., Quick, M. P., Rogers, H., Mundy, D. E., and Cross, A.  J. (1973).  Toxic
      metabolites of diazinon in sheep. J Agric Food Chem 21, 121-4.
   Kamal, M. A., and Al-Jafari, A. A. (2000). Dual substrate model for novel approach towards  a kinetic
      study of acetylcholinesterase inhibition by diazinon. J Enzyme Inhib 15, 201-13.
   Kappers, W.  A., Edwards, R. J., Murray, S., and Boobis, A. R. (2001). Diazinon is activated by
      CYP2C19 in human liver. ToxicolApplPharmacol 111, 68-76.
   Klemmer, H. W., Reichert,  E. R., and Yauger, W. L., Jr. (1978). Five cases of intentional ingestion of 25
      percent diazinon with treatment and recovery. Clin Toxicol 12, 435-44.
   Li, W. F., Costa, L. G., Richter, R. J., Hagen, T., Shih, D. M., Tward, A., Lusis, A.  J., and Furlong, C. E.
      (2000). Catalytic efficiency determines the in-vivo efficacy of PON1 for detoxifying
      organophosphorus compounds. Pharmacogenetics 10, 767-79.
   Machin, A. F., Anderson, P. H., and Hebert, C. N. (1974). Residue levels and cholinesterase activities in
      sheep poinsoned experimentally with diazinon. Pest Sci 5, 49-56.
   Machin, A. F., Quick, M. P., Rogers, H., and Anderson, P. H. (1971). The  conversion of diazinon to
      hydroxydiazinon in the  guinea-pig and sheep. Bull Environ Contam Toxicol 6, 26-7.
   Machin, A. F., Rogers, H., and Cross, A. J. (1975). Metabolic aspects of the toxicology of diazinon. I.
      Hepatic metabolism in the sheep,  cow, pig, guinea-pig, rat, turkey, chicken and  duck. Pest Sci 6,
      461-73.
   Mackness, B., Durrington, P., Povey, A., Thomson, S., Dippnall, M., Mackness, M., Smith, T., and
      Cherry, N. (2003). Paraoxonase and susceptibility to organophosphorus poisoning in farmers dipping
      sheep. Pharmacogenetics 13, 81-8.
   Mount, M. E. (1984). Diagnostic value of urinary dialkyl phosphate measurement in goats exposed to
      diazinon. Am J Vet Res  45, 817-24.
   Mucke, W., Alt, K.  O., and Esser,  H. O. (1970). Degradation of 14 C-labeled Diazinon in the rat. J Agric
      FoodChem 18,208-12.
   Nigg, H. N.,  and Knaak, J. B. (2000). Blood cholinesterases as human biomarkers of organophosphorus
      pesticide exposure. Rev Environ Contam Toxicol 163, 29-111.
   Poet, T. S., Kousba, A. A., Wu, H., Dennison, S. L., and Timchalk, C. (2003a). Development of a
      physiologically  based pharmacokinetic and pharmacodynamic (PBPK/PD) model for the
      organophosphate pesticide, diazinon. The Toxicologist. Abstract # 1485.
   Poet, T. S., Wu, H., Kousba, A. A., and Timchalk, C. (2003b). In vitro rat hepatic and intestinal
      metabolism of the organophosphate pesticides chlorpyrifos and diazinon. Toxicol Sci 72, 193-200.
   Sams, C., Mason, H. J., and Rawbone, R.  (2000). Evidence for the activation of organophosphate
      pesticides by cytochromes P450 3A4 and 2D6 in human liver microsomes. Toxicol Lett 116, 217-21.
   Tomokuni, K., and Hasegawa, T. (1985). Diazinon concentrations and blood cholinesterase activities in
      rats exposed to diazinon. Toxicol Lett 25,  7-10.
   Tomokuni, K., Hasegawa, T., Hirai, Y., and Koga, N. (1985). The tissue distribution of diazinon and the
      inhibition of blood cholinesterase activities in rats and mice receiving a single intraperitoneal dose of
      diazinon. Toxicology 37, 91-8.
   Tutudaki, M., Tsakalof, A. K., and Tsatsakis, A. M. (2003). Hair analysis used to assess chronic
      exposure to the  organophosphate diazinon: a model study with rabbits. Hum Exp Toxicol 22, 159-64.

                                                C-38

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   USEPA (2000). Refined Anticipated Residues/Acute and Chronic Dietary Risk Assessment (Including
      Beef Fat). U.S. Environmental Protection Agency. November 14, 2000.
   USEPA (2003). Restricted Use Pesticides Report. U.S. Environmental Protection Agency. Accessed
      9/28/03.  http://www.epa.gov/opprd001/rup/rupjun03.htm.
   Vittozzi, L., Fabrizi, L., Di Consiglio, E., and Testai, E. (2001). Mechanistic aspects of
      organophosphorothionate toxicity in fish and humans. Environ Int 26, 125-9.
   Walker, C. H., and Mackness, M. I. (1987). "A" esterases and their role in regulating the toxicity of
      organophosphates. Arch Toxicol 60, 30-3.
   Wester, R. C., Sedik, L., Melendres, J., Logan, F., Maibach, H. I, and Russell, I. (1993). Percutaneous
      absorption of diazinon in humans. FoodChem Toxicol 31, 569-72.
   Wu, H. X., Evreux-Gros, C., and Descotes, J. (1996a). Diazinon toxicokinetics, tissue distribution and
      anticholinesterase activity  in the rat. BiomedEnviron Sci 9, 359-69.
   Wu, H. X., Evreux-Gros, C., and Descotes, J. (1996b). Influence of cimetidine on the toxicity and
      toxicokinetics of diazinon  in the rat. Hum Exp Toxicol 15, 391-5.
   Yang, R. S., Hodgson, E., and Dauterman, W.  C. (1971). Metabolism in vitro of diazinon and diazoxon
      in rat liver. J Agric Food Chem  19, 10-3.
                                                C-39

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.


                                            Fenthion

   5.0    Introduction

   Fenthion  [0,0-dimethyl  O-(4-(methylthio)-m-tolyl)  phosphorothioate,  (DMTP),  Figure  1]  is  an
   organophosphorus insecticide used against mosquitoes, pests and bugs (EXTOXNET 1996).  It is
   available in dust, emulsifiable or liquid concentrate, and granular and wettable powder formulations.
   Fenthion is soluble in organic solvents such as DMSO, acetone, methanol  and ether, but not in water
   (NTP 2003).
                              Figure 5.1: Chemical structure of fenthion

   5.1    Toxic effects

   Fenthion is moderately toxic to laboratory animals (mice, rats, guinea pigs and rabbits). The acute oral
   and intraperitoneal LD50's varied from approximately 125 to >1000 mg/kg body weight (DuBois and
   Kinoshita  1964; IPCS 1971;  Ma 1995).  Mortality  of rats orally treated with fenthion in  subchronic
   studies (30 mg/kg for 13 weeks, or 5.0 mg/kg for 3 months) was reported (NIH 1979).

   Fenthion predominately causes  cholinergic toxicity in animals and humans.  Its  oxon inhibits plasma,
   erythrocyte and brain cholinesterase activity (Bai et al.  1990; De Bleecker et al. 1994; Bellinger and
   Mostrom  1988; Ma  1995; Misra et al.  1985; Misra  et al.  1994; Tsatsakis et al.  2002; Tsatsakis et al.
   1998). An acute no-observed-adverse -effect-level (NOAEL) of 0.07 mg/kg/day was determined in a 2-
   year oral  monkey study  (USEPA 1999a, 2001).  Other effects unrelated to  cholinergic mechanisms,
   however,  were also reported (Bagchi et al. 1995; Bagchi et al. 1996; Cova et al. 1995; Kojima et al.
   1992).

   Fenthion did not show mutagenic effect in the bacterial reverse mutation test or the in vitro chromosome
   aberration test in Chinese hamster ovary cells, but did in unscheduled DNA synthesis study and mouse
   micronucleus assays  (USEPA 1999a). In a 103-week chronic feeding study, no elevated incidence of
   tumor was observed in both sexes of F344 rats and female B6C3F1 mice; a slightly increased incidence
   of sarcomas, fibrosarcomas, and rhabdomyosarcomas of the integumentary system in male B6C3F1
   mice was  observed (NIH  1979). Fenthion is not considered a carcinogen (Ma 1995; USEPA 1999a).

   5.2    Pharmacokinetics

   5.2.1  Absorption

   Depending on application, fenthion may be absorbed from the gastrointestinal tract, skin and respiratory
   tract. The former two pathways, however, have been  more intensively studied. Generally,  fenthion is
   readily absorbed from GI tract. Blood levels peak a few hours after oral dosing in rats (Ma 1995), rabbits
   (Emteres et al 1985) and lactating goat (Ma 1995). Absorption was almost complete (96-100% at 72
                                               C-40

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   hours) and not dose-dependent (at 10 or 100 mg/kg) in Wistar rats fasted for 16-24  hour before gavage
   (Ma 1995).

   Dermal absorption of fenthion is slow and incomplete. Eighteen hours after a single dermal dose,
   prepared as an application formulation, in pigs or lactating cows, the tissue residue levels were generally
   low, whereas at the application site the levels were much higher (Ma 1995).  USEPA set a dermal
   absorption factor as 20% in 1996 and re-set it as 3% in 1999 based on the LOAEL's (lowest observed
   adverse effect level) of cholinesterase inhibition from an oral development toxicity study and a 21-day
   dermal toxicity study in rabbits (USEPA 1999b).

   5.2.2  Distribution

   From the limited information on its distribution in body, fenthion and its metabolites had relatively high
   concentration in fat, liver and kidney (EXTOXNET 1996). In milk from fenthion-treated dairy  cows,
   the fenthion level was 50 times higher in the "fat" fraction than that in the "non-fat" fraction (O'Keeffe
   etal. 1983).

   5.2.3  Metabolism

   Principal metabolic  pathways are shown in  Figure 5.2.   Fenthion has  several  possible oxidative
   metabolites in body such as sulfoxide (SO), sulfone  (802), oxygen analogue (P=O), oxygen analogue
   sulfoxide (P=O, SO) and sulfone (P=O, SO2) (Wright and Riner 1979). Of them, the oxygen analogues
   are bioactivated forms with higher anti-cholinesterase  activity (IPCS 1971, 1976).

   Incubated with rat liver microsomes, fenthion was oxidized to oxygen analogue and fenthion sulfoxide.
   Fenthion sulfone,  however, was not detected.  The  main enzymes involved were cytochrome P450s
   (especially CYP1A1) and flavin-containing monooxygenase (Kitamura  et al. 2003; Venkatesh et al.
   1991). In liver cytosol of rats, fenthion sulfoxide was  reduced to fenthion catalyzed by aldehyde oxidase
   (Kitamura  et al. 2003).

   14C-Fenthion was  extensively metabolized in  rats (Ma  1995).  No unchanged parent compound  was
   detected in the urine and very little (< 2%) in the feces.  The major group of metabolites (about 60% of
   the total label) was composed of the three  phenols (phenol  fenthion,  phenol sulfoxide and phenol
   sulfone) and  their  glucuronide and sulfate conjugates. Four demethyl metabolites accounted for about
   30% of the label, whereas the oxygen analogue sulfoxide constituted only 1-4%. The metabolite profiles
   were not affected by dosing route, dose, sex or pre-treatment with fenthion.

   In pigs,  fenthion was oxidized to fenthion sulfoxide, fenthion sulfone, oxygen analogue and oxygen
   analogue sulfoxide and sulfone.  These  metabolites were further hydrolyzed and excreted via  urine in
   conjugated forms (Ma 1995).

   No data from human studies is available.
                                               C-41

-------
USEPA Contract No. 3C-R102-NTEX
                                                     Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
                                                         SCH,
                                                                                         SCH,
                                           Phenol fentnion sulfoxide

                                                   O
Fenthion oxygen analogue
                              Fenthion oxygen analogue sulfoxide
                                                                 CH3O
                                                                 Fenthion oxygen analogue sulfone
                         Figure 5.2: Proposed metabolism pathways of fenthion in the rats.
                                             Adapted from Ma, 1995.
   5.2.4    Excretion

   Fenthion was rapidly  eliminated after  a  single  dose in Wistar rats, over  90% of the  administered
   radiolabel being excreted within 48 hours, and less than 1% retained 72 hours  after treatment (Ma 1995).
   In pigs and dairy cows rapid elimination and low bioaccumulation were also observed (Ma 1995).  In
   New Zealand white rabbits, the halflife of a single dose (20mg/kg) was about  11-12 hours regardless the
   route of administration (Emteres et al.  1985).
                                                   C-42

-------
USEPA Contract No. 3C-R102-NTEX                           Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   The polar metabolites of fenthion are mainly excreted via urine in rats, pigs and dairy cows (Ma 1995).
   Milk is a significant pathway for the elimination of the parent compound  from lactating dairy cows
   (IPCS 1971; O'Keeffe etal.  1983; Wright and Riner 1979).

   5.3    Interactions of fenthion with other OP pesticides

   Fenthion potentiated the acute intraperitoneal toxicity of malathion, dioxathion, and coumaphos in rats,
   but  intraperitoneal administration of 13 other organophosphate or carbamate insecticides to rats in
   combination with fenthion  did  not  result  in greater than  additive toxic effects  (Ma 1995).  Dietary
   combination of equitoxic doses (2 mg/kg) of fenthion with coumaphos, neither of which alone affected
   cholinesterase  activity when fed to dogs for six weeks, was found to potentiate the anticholinesterase
   activity in serum and erythrocytes by 75 and 30%, respectively.  The potentiation was less evident with
   malathion, and no potentiation was noted with dioxathion (Ma  1995).  Pretreatment with  fenthion
   significantly potentiated the acute toxicity of 2-sec-butylphenyl TV-methycarbamate (BPMC) in mice and
   dogs, which may be a result of the inhibited detoxification of the carbamate (Ma  1995; Miyaoka et al.
   1984; Miyaoka etal. 1987).

   5.4    PBPK models

   No PBPK models on fenthion have been reported yet.

   5.5    Literature Cited
   Bagchi, D., Bagchi, M., Hassoun, E. A., and Stohs, S. J. (1995). In vitro and in vivo generation of
      reactive oxygen species, DNA damage and lactate dehydrogenase leakage by selected pesticides.
      Toxicology 104, 129-40.
   Bagchi, D., Bhattacharya, G., and Stohs, S. J. (1996). In vitro and in vivo induction of heat shock
      (stress) protein (Hsp) gene expression by selected pesticides. Toxicology 112, 57-68.
   Bai, C. L., Qiao, C. B., Zhang, W. D., Chen, Y. L., and Qu, S. X. (1990). A study of the pesticide
      fenthion: toxicity, mutagenicity, and influence on tissue enzymes. BiomedEnviron Sci 3, 262-75.
   Cova, D., Perego, R., Nebuloni, C., Fontana, G., and Molinari, G. P. (1995). In vitro cytotoxicity of
      fenthion and related metabolites in human neuroblastoma cell lines. Chemosphere 30, 1709-15.
   De Bleecker, J., Lison, D., Van Den Abeele, K., Willems, J., and De Reuck, J. (1994). Acute and
      subacute organophosphate poisoning in the rat. Neurotoxicology 15, 341-8.
   Dellinger, J., and Mostrom, M. (1988). Effects of topical fenthion on blood cholinesterase and vagal
      tone in dogs. Vet Hum ToxicollO, 229-34.
   DuBois, K. P., and Kinoshita, F. (1964). Acute toxicity and anti-cholinesterase action of O,O-dimethyl
      O-4-(methylthio)-m-tolyl phosphorothioate (DMTP; Baytex) and related compounds. ToxicolAppl
      Pharmacol6, 86-95.
   Emteres, R., Abdelghani, A., and Anderson, A. C. (1985). Determination of the half life of fenthion in
      New Zealand White rabbits using three routes of administration. J Environ Sci Health B 20, 577-91.
   EXTOXNET (1996). Pesticide Information Profiles: Fenthion. Oregon State University.
      http: //ace. ace. orst.edu/info/extoxnet/pips/fenthi on. htm.
   IPCS (1971). WHO Pesticide Residues Series, No. 1. World Health Organization, Geneva.
   IPCS (1976). Data Sheets on Pesticides No. 23: Fenthion. World Health Organization, Geneva.
   Kitamura, S., Suzuki, T., Kadota, T., Yoshida, M., Ohashi, K., and Ohta, S. (2003). In vitro metabolism
      of fenthion and fenthion sulfoxide by liver preparations of sea bream, goldfish, and rats. DrugMetab
      Disposal, 179-86.

                                                C-43

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

  Kojima, T., Tsuda, S., and Shirasu, Y. (1992). Non-cholinergic mechanisms underlying the acute lethal
      effects of P = S type organophosphorus insecticides in rats. J VetMedSci 54, 529-33.
  Ma, S. (1995). Fenthion. Pesticide Residues in Food: 1995 evaluations Part II Toxicological &
      Environmental. Health Canada, Ottawa, Canada.
  Misra, U. K., Nag, D., Bhushan, V., and Ray, P. K. (1985). Clinical and biochemical changes in
      chronically exposed organophosphate workers. Toxicol Lett 24, 187-93.
  Misra, U. K., Prasad, M., and Pandey, C. M. (1994). A study of cognitive functions and event related
      potentials following organophosphate exposure. Electromyogr Clin Neurophysiol 34, 197-203.
  Miyaoka, T., Takahashi, H., Tsuda, S., and Shirasu, Y. (1984). Potentiation of acute toxicity of 2-sec-
      butylphenyl N-methylcarbamate (BPMC) by fenthion in mice. Fundam Appl Toxicol 4, 802-7.
  Miyaoka, T., Tsuda, S., and Shirasu, Y. (1987). Effect of O, O-dimethyl O-(3-methyl-4-
      methylthiophenyl) phosphorothioate (fenthion) pretreatment on acute toxicity of 2-sec-butylphenyl
      N-methylcarbamate (BPMC) in dogs. Nippon Juigaku Zasshi 49, 173-5.
  NIH (1979). Bioassay of fenthion for possible carcinogenicity. U.S. Department of Health, Education
      and Welfare, Bethesda, Maryland. DHEW Pub No. (NIH) 79-1353.
  NTP. NTP Chemical Repository: Baytex.  http://ntp-server.niehs.nih.gov/htdocs/
      CHEM_H&S/NTP_Chem5/Radian55-38-9.html
  O'Keeffe, M., Eades, J. F., and Strickland, K. L. (1983). Fenthion residues in milk and milk products
      following treatment of dairy cows for warble-fly. J Sci FoodAgric 34, 192-197.
  Tsatsakis, A. M., Bertsias, G. K., Liakou, V., Mammas, I. N., Stiakakis, I, and Tzanakakis, G. N.
      (2002). Severe fenthion intoxications due to ingestion and inhalation with survival  outcome. Hum
      Exp Toxicol 21, 49-54.
  Tsatsakis, A. M., Manousakis, A., Anastasaki, M., Tzatzarakis, M., Katsanoulas, K., Delaki, C., and
      Agouridakis, P. (1998). Clinical and toxicological data in fenthion and omethoate acute poisoning. J
      Environ Sci Health B 33, 657-70.
  USEPA (1999a).  Human Health Risk Assessment: Fenthion. U.S. Environmental Protection Agency,
      Washington, D.C., October 13, 1999.
  USEPA (1999b). Fenthion: Re-evaluation of the  dermal absorption factor. U.S. Environmental
      Protection Agency, Washington, D.C. HED DOC. No.013746.
  USEPA (2001). Interim Reregi strati on Eligibility Decision for Fenthion. U.S. Environmental Protection
      Agency, Washington, D.C. Report # EPA738-R-00-013.
  Venkatesh, K., Levi, P. E., and Hodgson, E. (1991). The effect of detergents on the purified flavin-
      containing monooxygenase of mouse liver, kidney and lungs. Gen Pharmacol 22, 549-52.
  Wright, F. C., and Riner, J. C. (1979). Biotransformation and deposition of residues of fenthion and
      oxidative metabolites in the fat of cattle. J Agric Food Chem 27, 516-1.
                                               C-44

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.


                                           Fenitrothion

   6.0    Introduction

   Fenitrothion ((9,(9-dimethyl O-(3-methyl-4-nitrophenyl) phosphorodithioate), also commonly called
   Sumithion™, is an organophosphorus insecticide and was registered for use in ant and roach bait. There
   are no approved domestic food or feed uses for fenitrothion, and exposure to fenitrothion in the U.S. is
   minimal. However, fenitrothion is used in countries to control pests on crops, stored grains, and cotton.
   Fenitrothion is also used elsewhere in forest spraying and public health campaigns. As a result, the
   human health effects associated with exposure to fenitrothion remain a concern, especially among
   pesticide workers and applicators whose acute exposure to organophosphorus pesticides can sometimes
   occur at levels high enough to inhibit blood acetylcholinesterase activity (Nigg and Knaak 2000; Ohayo-
   Mitoko et al. 2000; Satoh and Hosokawa 2000).

   Degradation rates reported for fenitrothion are as follows:
   •  Soil and groundwater -» T,/2 < one week (Meister 1994; U.S.EPA  1987)
   •  Surface water —» Ti/2 = 1.5-2 days (Novathion 1987)
   •  Surface water in dark -» Ti/2 = 21.6 - 49.5 days (Novathion 1987)
   •  Plants -» Ti/2 = 1-2  days (Mollhoff 1968)
   •  Plants (fenitrooxon) -»Ti/2 = few hours (Mollhoff 1968)

   6.1    Toxic effects

   Like other organophosphorus compounds, fenitrothion acts in the organism as a cholinesterase inhibitor,
   after conversion to fenitrooxon. Some evidence indicates that acetylcholinesterase inhibition in brain
   depends more on the rate of penetration than on the rate of oxidation and decomposition of fenitrothion
   (JMPR 1988; Miyamoto 1969). Fenitrothion appears to affect cytochrome P450 enzyme activity in the
   liver and testes of rats (Clos et al. 1994; Gradowska-Olszewska et al. 1984).

   Fenitrothion has anticholinesterase activity and moderate acute toxicity with oral LD50 values in rats and
   mice ranging from 330 to 1,416 mg/kg body weight (Miyamoto et al. 1963b). Acute dermal toxicity in
   rodents is reported to range from 890 to more than 2,500 mg/kg body weight (WHO 1992). The LCso
   value in rats exposed for 8 h is estimated to be more than 186 mg/m3 (WHO 1992). In short-term studies
   on rats and dogs, no-observed-adverse-effect levels (NOAELs) based on brain cholinesterase activity
   were 10 mg/kg diet and 50 mg/kg diet, respectively. Long-term studies on rats and mice indicated a
   NOAEL of 10 mg/kg diet (WHO 1992). An acceptable daily intake (ADI) of 0.003 mg/kg body weight
   was established in 1984 (WHO  1986), but no occupational exposure limits (OEL) have been published.
   No carcinogenic effects were found in any of the long-term fenitrothion studies (WHO 1992).
   Fenitrothion was not found to be mutagenic in in vitro and in vivo studies or teratogenic at doses of up to
   30 mg/kg body weight in rabbits and up to 25 mg/kg body weight in rats (Benes et al.  1975; WHO
   1992).  Other toxicity studies have been conducted (Chevalier et al. 1982; Groszek et al. 1995; Khan et
   al. 1990; Misu et al. 1966; Myatt et al. 1975; Trottier et al.  1980; Yoshida et al. 1987). Fenitrothion has
   also been shown to be neurotoxic, immunosuppressive, a pulmonary toxicant, and cause disturbances of
   prenatal development (Berlinska and Sitarek 1997; Kunimatsu et al.  1996; Khan et al. 1990; Lehotzky
   and Ungvary 1976).
                                                C-45

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   In a field spraying operation in Nigeria and Kenya, humans exposed to fenitrothion exhibited depressed
   plasma cholinesterase activity (Ohayo-Mitoko et al. 2000; Vandekar 1965; Wilford et al.  1965). A study
   of fenitrothion on 24 human subjects was also conducted and showed that both plasma and
   cholinesterase activity was not depressed in all but one case (Nosal and Hladka 1968). Fenitrothion has
   been reported as causing "intermediate syndrome" due to acute poisoning (Groszek et al.  1995).

   The presence of chemicals in the environment that have antiandrogenic activity and thus the ability to
   disrupt the endocrine system is a source of concern. In several studies fenitrothion has been shown to be
   a competitive androgen receptor antagonist both in vivo and in vitro (Curtis 2001; Sohoni  et al. 2001;
   Tamura et al. 2001; Turner et al. 2002). One study, however, exhibited fenitrothion not having
   significant androgenic or antiandrogenic activity in vivo (Sunami et al. 2000). Fenitrothion might also
   alter estradiol metabolism  by inhibition of certain P450 enzymes and produce changes in adrenal
   function (Berger and Sultatos 1997; Yamamoto et al.  1982b).

   6.2     Pharmacokinetics

   Various studies in mouse,  rat, guinea pig, and humans have dealt with the pharmacokinetic and
   biochemical aspects of fenitrothion and its metabolites (Aprea et al. 1999; Douch et al.  1968; Hladka
   and Nosal 1967; Hollingworth et al. 1967; Meaklim et al. 2003; Miyamoto 1964a; Miyamoto  1964b;
   Miyamoto and Sato 1969;  Miyamoto 1969; Miyamoto et al.  1963a; Nishizawa et al. 1961; Vandanis and
   Crawford 1964).

   6.2.1   Absorption

   Fenitrothion is presumably rapidly absorbed from the mammalian intestinal tract when given orally.
   Additionally, it can also be absorbed by the intact skin and by inhalation. (Kohli et al. 1974; Moody and
   Franklin 1987; Moody et al.  1987).

   6.2.2   Distribution

   The presence of the oxygen analogue was demonstrated in all tissues examined (brain, heart, lung, liver,
   kidney, spleen, and muscle), and it was detectable in blood one min after intravenous injection of
   fenitrothion (Muller 2000).

   6.2.3   Metabolism

   The oxygen analogue is the most important metabolite with respect to toxicity. It is formed in the
   microsomal fraction of the cell, the main organs responsible for the transformation being the liver and
   kidney. The major excretion product found is 3-methyl-4-nitrophenol which can be oxidized further to
   3-carboxyl-4-nitrophenol.  Other metabolites are the dimethyl derivatives, which, with increasing dose,
   are excreted in increasing  amounts. Nine metabolites have been isolated, most of which have also been
   identified. In vitro, formation of the oxygen analogue depended on the availability of reduced nicotine
   adenine dinucleotide phosphate (NADPFk) and oxygen (Miyamoto et al. 1963a; Miyamoto 1969). Liver
   slices incubated with fenitrothion did not produce measurable amounts of fenitrooxon, while liver
   homogenates and the supernatant fraction of such homogenates appreciably activated added fenitrothion
   (Miyamoto et al. 1963a; Miyamoto 1969). No correlation between the toxicity and rate  of formation  of
   fenitrooxon could, however, be demonstrated (JMPR 1988; Miyamoto et al. 1963a; Miyamoto 1969).
   No observations were made in these studies on the distribution into fatty tissues, but studies of residues
   in milk, meat, and fat from cattle indicated the presence of approximately 0.001 mg/kg in  these samples

                                                C-46

-------
USEPA Contract No. 3C-R102-NTEX
                                                Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
   (JMPR 1988; Miyamoto and Sato 1969). Other studies involving the metabolism of fenitrothion have
   been described (Anjum and Qadri 1986; Kasagami etal. 2002; Sultatos 1991; Yamamoto et al. 1983;
   Yamamoto et al. 1982; Yoshida etal. 1975).

                Figure 6.1: Metabolic pathway of fenitrothion in vivo (Kumar et al. 1993).
                                           CYP450
                                                         o-ci-u
                      Fenitrothion

                             CYP450

                        o

                     Fenitroxon
                                           Esterase
                                                        o
                                                  .        ^
                                                                  +
   6.2.4  Excretion

   Fenitrothion and its metabolites are excreted mainly in the urine (90-95%) (Aprea et al. 1999;
   Hollingworth et al. 1967). Up to 10% was recovered in feces (Hollingworth et al. 1967). Within three
   days nearly complete recovery of an orally administered dose (15 mg/kg) could be obtained
   (Hollingworth et al. 1967). The ratios between the amounts of metabolites was dependent upon the dose
   given (Hollingworth et al. 1967). Other urinary excretion studies have been described (Aprea et al.
   1999; Hladka and Nosal 1967; Kojima et al. 1989; Nosal and Hladka 1968).

   6.3    Interactions of fenitrothion with other chemicals

   Interactions of fenitrothion with other compounds such as malathion (Hladka et al. 1974), diethyl
   maleate (Sultatos etal. 1991), 2-sec-butylphenol methylcarbamate (BPMC) (Takahashi etal. 1984), and
   N,N-diethyl-m-toluamide (DEBT)  (Moody et al. 1987) have  also been studied.

   6.4    PBPK models

   To date there are no published PBPK models for fenitrothion; however, there are numerous
   pharmacokinetic data that could be used in model development (Aprea et al. 1999; Douch et al. 1968;
   Hladka and Nosal 1967; Hollingworth et al. 1967; Kojima et al. 1989; Meaklim et al.  2003; Meaklim
   and McNeil 1999; Miyamoto 1964a; Miyamoto 1964b; Miyamoto and Sato 1969; Miyamoto 1969;
   Miyamoto et al. 1963a; Muller 2000; Nishizawa et al. 1961; Nosal and Hladka 1968; Vandanis and
   Crawford 1964).

   6.5    Literature Cited
                                               C-47

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Anjum, F., and Qadri, S. S. (1986). In vivo metabolism of fenitrothion (0,0-dimethyl-0-(4-nitro-m-tolyl)
      phosphorothioate) in fresh water teleost (Tilapia mossambica). Bull Environ Contain Toxicol 36,
      140-145.
   Aprea, C., Sciarra, G., Sartorelli, P., Ceccarelli, F., and Centi, L. (1999). Multiroute exposure
      assessment and excretion of urinary metabolites of fenitrothion during manual operations on treated
      ornamental plants in greenhouses. Arch Environ Contain Toxicol 36, 490-497.
   Benes, V., Sram, R. J., and Tuscany, R. (1975). Fenitrothion. I. Study of mutagenic activity in rats. J
      Hyg Epidemiol Microbiol Immunol 19, 163-172.
   Berger, C. W., Jr., and Sultatos, L. G. (1997). The effects of the phosphorothioate insecticide
      fenitrothion on mammalian cytochrome P450-dependent metabolism of estradiol. Fundam Appl
      Toxicol 37, 150-157.
   Berlinska, B., and Sitarek, K. (1997). Disturbances of prenatal development in rats exposed to
      fenitrothion. Rocz Panstw Zakl Hig 48, 217-228.
   Chevalier, G., Bastie-Sigeac, L, and Cote, M. G. (1982). Morphological assessment of fenitrothion
      pulmonary toxicity in the rat. Toxicol Appl Pharmacol 63, 91-104.
   Clos, M. V., Ramoneda, M., and Garcia, G. (1994). Modification of testicular cytochrome P-450 after
      fenitrothion administration. Gen Pharmacol 25, 499-503.
   Curtis, L. R. (2001).  Organophosphate antagonism of the androgen receptor.  Toxicol Sci 60, 1-2.
   Douch, P. G. C., Hook, C. E. R., and Smith, J. N. (1968). Metabolism of Folithion (dimethyl-4-nitro-3-
      methylphenyl phosphorothionate). Australasian J Pharmacol 49, Nr.66, 2.S.
   Gradowska-Olszewska,  I, Brzezinski, J., and Rusiecki, W. (1984). Excretion and peripheral metabolism
      of 1, 2-3H-testosterone and androgens in rats following intoxication with organophosphorous
      insecticides. 1—Acute exposure. J Appl Toxicol 4, 261-264.
   Groszek, B., Pach, J., and Klys, M. (1995). Intermediate syndrome in acute fenitrothion poisoning.
      PrzeglLek 52, 271-274.
   Hladka, A., Krampl,  V., and Kovac, J. (1974). Effect of malathion on the content of fenitrothion and
      fenitrooxone in the rat. Bull Environ Contam Toxicol 12, 38-45.
   Hladka, A., and Nosal, M. (1967). The determination of the exposition to metathion (fenitrotion) on the
      basis of excreting its metabolite p-nitro-m-cresol through urine in rats. Int Arch Arbeitsmed  23, 209-
      214.
   Hollingworth, R. M., Fukuto, T. R., and Metcalf, R. L. (1967). Selectivity of Sumithion compared with
      methyl parathion. Influence of structure on anticholinesterase activity. Metabolism in the white
      mouse. Agric Food Chem 15,235 -249.
   JMPR (1988). Pesticide residues in food. Report of the joint  meeting of the FAO panel of experts on
      pesticide residues on food and the environment and a WHO expert group on pesticide residues.
   Kasagami, T., Miyamoto, T., and Yamamoto, I.  (2002). Activated transformations of organophosphorus
      insecticides in the case of non-AChE inhibitory oxons. PestManag Sci 58, 1107-1117.
   Khan, M. F., Abidi, P., Anwer, J., Ray, P. K., and Anand, M. (1990). Pulmonary biochemical
      assessment of fenitrothion toxicity in rats. Bull Environ Contam Toxicol  45, 598-603.
   Kohli, J. D., Hasan, M. Z., and Gupta, B. N. (1974). Dermal  absorption of fenitrothion in rat. Bull
      Environ Contam  Toxicol 11, 285-290.
   Kojima, T., Yashiki, M., Miyazaki, T., Chikasue, F., and  Ohtani, M. (1989). Detection of S-
      methylfenitrothion, aminofenitrothion, aminofenitroxon and acetylaminofenitroxon in the urine of a
      fenitrothion intoxication case. Forensic Sci Int 41, 245-253.
   Kumar, R., Roy, S., Rishi, R., and Sharma, C. B. (1993). Metabolic fate of fenitrothion in liver,  kidney
      and brain of rat. Biomed Chromatogr 7, 301-305.
   Kunimatsu, T., Kamita,  Y., Isobe, N., and Kawasaki, H. (1996). Immunotoxicological insignificance of
      fenitrothion in mice  and rats. Fundam Appl Toxicol 33, 246-253.
                                                C-48

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

  Lehotzky, K., and Ungvary, G. (1976). Experimental data on the neurotoxicity of fenitrothion. Ada
      Pharmacol Toxicol (Copenh) 39, 374-382.
  Meaklim, J., Yang, J., Drummer, O. H., Killalea, S., Staikos, V., Horomidis, S., Rutherford, D.,
      loannides-Demos, L. L., Lim, S., McLean, A. J., and McNeil, J. J. (2003). Fenitrothion:
      toxicokinetics and toxicologic evaluation in human volunteers. Environ Health Perspect  111, SOS-
      SOS.
  Meaklim, J. F., and McNeil, J. J. (1999). Fenitrothion ingestion in humans: Subacute effects.
      Unpublished report no. HT-0539 from the department of Epidemiology and Preventative Medicine.
  Meister, R. T. (1994). Farm Chemical Handbook. Meister Publishing Co., Willoughby, OH.
  Misu, Y., Segawa, T., Kuruma, L, Kojima, M., and Takagi, H. (1966). Subacute toxicity of O, o-
      dimethyl O-(3-methyl-4-nitrophenyl) phosphorothioate (Sumithion) in the rat. Toxicol Appl
      Pharmacol 9, 17-26.
  Miyamoto, J. (1964a). Studies on the mode of action of organophosphorus compounds. Part III.
      Activation and degradation of Sumithion and Methylparathion in vivo. Agric Biol Chem  28, 411-
      421.
  Miyamoto, J. (1964b). Studies on the mode of action of organophosphorus compounds. Part IV.
      Pentration of Sumition, Methylparathion and their oxygen analogs into guinea pig brain and
      inhibition of cholinesterase in vivo. Agric Biol Chem 28, 422-430.
  Miyamoto, J. (1969). Mechanism of low toxicity of Sumithion toward mammals. Residue Rev 25, 251-
      264.
  Miyamoto, J., and Sato, Y. (1969). Determination of insecticide residue in animal and plant tissues. VI.
      Determination of Sumithion residue in cattle tissues. Botyu Kagaku 34, 3-6.
  Miyamoto, J., Sato, Y., Kadota, T., Fujinami, A., and Endo, M. (1963a). Studies on the mode of action
      of organophosphorus compounds. Part I. Metabolic fate of P32 labelled Sumition and methyl
      parathion in the guinea-pig and white rat. Agric Biol Chem  27, 381-389.
  Miyamoto, J., Sato, Y., Kadota, T., Fujinami, A., and Endo, M. (1963b). Studies on the mode of action
      of organophosphorus compounds. Part II. Inhibition of mammalian cholinesterase in vivo following
      administration of Sumithion and methylparathion. Agric Biol Chem 27, 669-676.
  Mollhoff, E. (1968). Beit zur Frage der Riickstande und ihrer Bestimmung in Pflanzen nach Anwendung
      von Praparaten der E 605, pp. 331-58, Pflanzenschutz-Nachrichten Bayer.
  Moody, R. P., and Franklin, C. A. (1987). Percutaneous absorption of the insecticides fenitrothion and
      aminocarb in rats and monkeys. J Toxicol Environ Health 20, 209-218.
  Moody, R. P., Riedel, D., Ritter, L., and Franklin, C. A. (1987). The effect of DEBT (N,N-diethyl-m-
      toluamide) on dermal persistence and absorption of the insecticide fenitrothion in rats and monkeys.
      J Toxicol Environ Health 22, 471-479.
  Muller, U. (2000). Pesticide residues in food 2000: Fenitrothion. Chemicals and Non-Presciption
      Medicines Branch, Therapeutics Goods Administration, Canberra, ACT, Australia.
  Myatt, G. L., Ecobichon, D. J., and Greenhalgh, R. (1975). Fenitrooxon and S-methyl fenitrothion: acute
      toxicity and hydrolysis in mammals. Environ Res 10, 407-414.
  Nigg, H. N., and Knaak, J. B. (2000). Blood cholinesterases as human biomarkers of organophosphorus
      pesticide exposure. Rev Environ Contam Toxicol 163, 29-111.
  Nishizawa, Y., Fujii, K., Kadota, T., Miyamoto, J., and Sakamoto, H. (1961). Studies of the
      organophosphorus insecticides. Part VII. Chemical and biological preperties of new low toxic
      organophosphorus insecticide. O,O-Dimethyl-O-(3-methyl-4-nitrophenyl) thiophosphorothioate.
      Agric Biol Chem  25,605-610.
  Nosal, M., and Hladka, A. (1968). Determination of the exposure to fenitrothion (0,0-dimethyl-0-3-
      methyl-4-nitrophenyl-thiophosphate)  on the basis of the excretion of p-nitro-m-cresol by  the urine of
      the persons tested. Int Arch Arbeitsmed 25, 28-38.
  Novathion (1987). Data Manual. Cheminova Agro A/S, Lemvig, Denmark.

                                               C-49

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Ohayo-Mitoko, G. 1, Kromhout, H., Simwa, J. M., Boleij, J. S., and Heederik, D. (2000). Self reported
      symptoms and inhibition of acetylcholinesterase activity among Kenyan agricultural workers. Occup
      Environ Med 57, 195-200.
   Satoh, T., and Hosokawa, M. (2000). Organophosphates and their impact on the global environment.
      Neurotoxicology 21, 223-227.
   Sohoni, P., Lefevre, P. A., Ashby, J., and Sumpter, J. P. (2001). Possible androgenic/anti-androgenic
      activity of the insecticide fenitrothion. JAppl Toxicol 21, 173-178.
   Sultatos, L. G. (1991). Metabolic activation of the organophosphorus insecticides chlorpyrifos and
      fenitrothion by perfused rat liver. Toxicology 68,  1-9.
   Sultatos, L. G., Huang, G. J., Jackson, O., Reed, K., and Soranno, T. M. (1991). The effect of
      glutathione monoethyl ester on the potentiation of the acute toxicity of methyl parathion, methyl
      paraoxon or fenitrothion by diethyl maleate in the mouse.  Toxicol Lett 55, 77-83.
   Sunami, O., Kunimatsu, T., Yamada, T., Yabushita, S., Sukata, T., Miyata, K., Kamita, Y., Okuno,  Y.,
      Seki, T., Nakatsuka, L, and Matsuo, M. (2000). Evaluation of a 5-day Hershberger assay using
      young mature male rats: methyltestosterone and p,p'-DDE, but not fenitrothion, exhibited androgenic
      or antiandrogenic activity in vivo. J ToxicolSci  25, 403-415.
   Takahashi, H., Miyaoka, T., Tsuda, S., and Shirasu, Y.  (1984). Potentiated toxicity of 2-sec-butylphenyl
      methylcarbamate (BPMC) by O,O-dimethyl  O-(3-methyl-4-nitrophenyl)phosphorothioate
      (fenitrothion) in mice; relationship between acute toxicity and metabolism of BPMC. Fundam Appl
      Toxicol 4, 718-723.
   Tamura, H., Maness, S.  C., Reischmann, K., Dorman, D. C., Gray, L. E., and Gaido, K. W. (2001).
      Androgen receptor antagonism by the organophosphate insecticide fenitrothion. Toxicol Sci  60, 56-
      62.
   Trottier, B., Fraser, A. R., Planet, G., and Ecobichon, D. J. (1980). Subacute toxicity of technical
      fenitrothion in male rats. Toxicology 17, 29-38.
   Turner, K. J., Barlow, N. J., Struve, M. F., Wallace, D.  G., Gaido, K. W., Dorman, D. C., and Foster, P.
      M. (2002). Effects of in utero exposure to the organophosphate insecticide fenitrothion on androgen-
      dependent reproductive development in the Crl:CD(SD)BR rat. Toxicol Sci  68, 174-183.
   U.S.EPA (1987). Pesticide Fact Sheet Number 142. US EPA, Offices of Pesticide Programs,
      Registration Division, Washington, DC.
   Vandanis, A., and Crawford, L. G. (1964). Comparative metabolism of O,O-dimethyl-O-p-nitrophenyl
      phosphorothioate (methylparathion and O,O-dimethyl-O-(3-methyl-1-nitrophenyl) phosphorothioate
      (Sumithion). JEcon Entomol  57, 136-139.
   Vandekar, M. (1965). Observations on the toxicity of carbaryl, folithion and 3-isopropylphenyl n-
      methylcarbamate in  a village-scale trial in Southern Nigeria. Bull World Health Organ 33, 107-115.
   WHO (1986). Environmental health criteria 63, Geneva.
   WHO (1992). Environmental health criteria 133, Geneva.
   Wilford, K., Lietaert, P. E. A., and Foil, C. V. (1965). Toxicological observations during large scale
      field trial of OMS-43 in Northern Nigeria (preliminary report), Geneva, 18-24 February 1965.
   Yamamoto, T., Egashira, T., Yoshida, T., and Kuroiwa, Y. (1982a).  Comparison of the effect of an
      equimolar and low dose of fenitrothion and methylparathion on their own metabolism in rat liver. J
      Toxicol Sci 7,35-41.
   Yamamoto, T., Egashira, T., Yoshida, T., and Kuroiwa, Y. (1982b).  Increase of adrenal weight in rats by
      the repeated administration of fenitrothion. Toxicol Lett 11, 187-191.
   Yamamoto, T., Egashira, T., Yoshida, T., and Kuroiwa, Y. (1983). Comparative metabolism of
      fenitrothion and methylparathion in male rats. Ada Pharmacol Toxicol (Copenh) 53, 96-102.
   Yoshida, M., Shimada, E., Yamanaka, S., Aoyama, H., Yamamura, Y., and Owada, S. (1987). A case of
      acute poisoning with fenitrothion (Sumithion). Hum Toxicol  6, 403-406.
                                                C-50

-------
USEPA Contract No. 3C-R102-NTEX                           Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Yoshida, T., Homma, K., Suzuki, Y., and Uchiyama, M. (1975). Effect of fenitrothion on hepatic
      microsomal components of drug metabolizing system in mice. Chem Pharm Bull (Tokyo) 23, 2155-
      2157.
                                                  C-51

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.


                                           Chloroform

   7.0    Introduction

   Chloroform (trichloromethane, CHCb) is a dense liquid that is insoluble in water and volatile under
   environmental conditions (McCulloch 2003). The major domestic use for chloroform is in the
   manufacture of refrigerant HCFC-22 (Chemical Marketing Reporter 1995). Chloroform is also used as
   a laboratory reagent and extraction solvent for pharmaceuticals.  A significant amount of chloroform has
   been released to the environment as a by-product of the treatment of drinking and waste waters and
   through reactions of chlorine with organic chemicals (Meek et al. 2002).

   7.1    Mechanisms of toxicity

   Oral and inhalation exposures to chloroform cause toxicity to the liver, kidney, and nasal epithelium
   (USEPA 2001).  Chloroform can also cause reproductive or developmental toxicity, although most of
   the effects are secondary to maternal toxicity (USEPA 2001). Increased incidences of liver and kidney
   tumors have been observed in several animal species after exposures to chloroform via several routes,
   although there is no adequate human data for carcinogenicity (USEPA 2001). The mode of toxicity of
   chloroform is probably through oxidative metabolism to form phosgene (Pohl et al. 1977), which can
   react to form covalent bonds with microsomal proteins (Corley et al. 1990; Rosenthal 1987).

   7.2    Pharmacokinetics

   7.2.1  Absorption

   Chloroform is generally absorbed rapidly in humans and animals. It is easily absorbed into the blood
   from the lungs after inhalation exposures.  Human inhalation studies include exposures via surgical
   anesthesia (Smith et al. 1973), indoor swimming pools (Aggazzotti et al.  1993; Cammann and Hubner
   1995; Levesque et al.  1994; Levesque et al. 2000),  and shower air (Jo et al. 1990; Levesque et al. 2002).
   Chloroform can also be absorbed through the skin easily. Dermal exposures were considered in
   conjunction with inhalation exposures in some of the indoor swimming pool studies (Cammann and
   Hubner 1995; Levesque et al.  1994; Levesque et al. 2000)  and shower air studies (Jo et al. 1990;
   Levesque et al. 2002).  Other human studies of dermal-only exposures include showering with facemask
   (Corley et al. 2000; Gordon et al. 1998) and topical administration of chloroform to volunteers (Dick et
   al. 1995).  Dermal absorption in animals was studied in guinea pigs (Jakobson et al. 1982) and hairless
   rats (Islam etal. 1995). Gastrointestinal absorption of chloroform is also fast and extensive (USEPA
   2001). Oral exposure studies of humans were done in  volunteers using 13C-labeled chloroform (Fry et
   al. 1972) and additional information is available from an accidental chloroform poisoning case (Rao et
   al. 1993).  Animal studies via oral exposure were reported in mice and rats by Withey et al. (Withey et
   al. 1983) and Pereira (Pereira 1994), respectively.

   7.2.2  Distribution

   Chloroform is widely distributed throughout the body after being absorbed. Radiolabeled chloroform in
   mice was reported to distribute to the liver, kidney, lungs, spleen, body fat, muscle, and nervous tissue
   (Bergman 1979; Cohen and Hood 1969). The highest  levels of chloroform detected in human
   postmortem samples are in the body fat (5-68 ng/kg) and lower levels (1-10 ng/kg) were  detected in the
   kidney, liver, and brain (McConnell etal. 1975). In a study with 14C-chloroform injected in male mice

                                               C-52

-------
USEPA Contract No. 3C-R102-NTEX
                                                  Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
   intraperitoneally, the maximum radioactivity levels were observed in the liver, kidney, and blood within
   10 minutes of dosing (Gemma et al. 1996). It was also found that the presence of testosterone affected
   chloroform accumulations in mouse kidney (Ilett et al. 1973; Pohl et al. 1984; Smith et al. 1973) and
   resulted in higher nephrotoxicity in male mice.

   7.2.3   Metabolism

   The major metabolic pathway of chloroform in humans and animals (Figure 7.1) is oxidative
   metabolism that produces reactive phosgene and the minor pathway is reductive metabolism that forms
   dichloromethyl free radical (USEPA 2001).  In the presence of oxygen, chloroform is converted to
   trichloromethanol, which spontaneously dehydrochlorinates to produce phosgene (Pohl et al. 1981;
   Stevens and Anders 1981). These reactions are catalyzed by cytochrome P450 in liver and kidneys (Ade
   et al. 1994; Branchflower et al. 1984;  Smith and Hook 1984). Phosgene can in turn react with
   nucleophilic groups in cellular macromolecules and form covalent adducts (Noort et al. 2000; Pereira
   and Chang 1981; Pereira et al. 1984; Pohl  et al.  1977;  Pohl et al.  1981; Pohl et al. 1980).  Phosgene can
   also undergo hydrolysis to form  carbon dioxide and hydrochloric acid, or react with glutathione to form
   diglutathionyl dithiocarbonate, gluathione  disulfide,  and carbon monoxide (Pohl et al. 1981; USEPA
   2001).  In the absence of oxygen, chloroform is converted to dichloromethyl free radical,  which can
   form covalent adducts with microsomal  enzymes and can also cause lipid peroxidation (USEPA 2001).
   Metabolic pathways of chloroform overlapping  with the other three volatile organics in Mixture 2 are
   shown in Figure 8.1 under trichloroethylene.
                P450
      CHC13
               O2   H2O
HOCC13
                           O = CC12
                           Phosgene
                                         H2O
                                          V
                   ->  2HC1 + CO2
               Cysteine
          Adducts
                         GSH
  \
                         Her
                               O
                               II
                          GS—C—Cl
                       S-chloro-carbonyl glutathione
     CCLH
 H2  /   2
 C-CH
VNH
                                                2-oxothiazolidine-
                                                4-carboxylic acid
              fl
                   CO + GSSG
         GS—C—SG
    diglutathionyl dithiocarbonate
  Figure 7.1 Major metabolic pathways of chloroform (adapted from USEPA (2001))
                                                C-53

-------
USEPA Contract No. 3C-R102-NTEX
                                                Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
   7.2.4  Excretion

   Chloroform is excreted through the lungs either unchanged or as carbon dioxide, with small amounts
   detected in urine after inhalation (Corley etal.  1990; Fry etal.  1972; Gordon etal. 1988), oral (Fry etal.
   1972), and dermal (Dick et al. 1995) exposures.

   7.3    Physiologically based pharmacokinetic (PBPK) models

   The first PBPK model for chloroform was developed by Corley and colleagues (Corley et al. 1990) to
   describe the fate of chloroform in several species via numerous exposure routes.  Several subsequent
   PBPK models (Chinery and Gleason 1993; Corley et al. 2000;  Gearhart et al. 1993; Levesque et al.
   2000; McKone 1993; Roy et al. 1996) were developed based on the Corley model to include a variety of
   exposure scenarios.  A schematic representation of a general PBPK model for chloroform is shown in
   Figure 7.2.
             Inhalation
             Exposure

          Intravenous _
          Infusion
           Metabolites
Alveolar Air
Lung Blood
                                Richly Perfused Tissues
                                Slowly Perfused Tissues
                                        Kidney
                                         Liver
                                      f
                                Dermal Exposure
Exhaled Air
                            Intraperioneal
                            Injection
                                                                 Oral Exposure
  Figure 7.2 PBPK model for chloroform in rats, mice, and human (adapted from (Corley et al. 2000))

  In the Corley model (Corley et al. 1990), the exposure routes include oral, inhalation, and
  intraperitoneal. Liver and kidney were both sites of metabolism for chloroform. The amount of
                                               C-54

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   metabolite binding to cellular macromolecules was used as the indicator for chloroform toxicity. Due
   to the lower rates of metabolism, ventilation, and cardiac output in larger species than in smaller species,
   the relative potency of chloroform was predicted as mice > rats > humans in the Corley model (Corley et
   al. 1990).

   Reitz and colleagues (Reitz et al. 1990) extended the Corley model to include pharmacodynamic
   endpoints for cancer risk assessment.  Two dose metrics were used for the liver compartment, while the
   kidney compartment was not considered. The first type of dose metric used was the average daily
   macromolecular binding. The other type of dose surrogate was cytotoxicity due to the formation of
   reactive chloroform metabolite, phosgene.  It was concluded that cytotoxicity is the dose metric best
   reflecting carcinogenicity (Reitz et al. 1990).  These two dose metrics were later analyzed for
   interindividual variability and parameter uncertainty by Allen and colleagues (Allen et al. 1996). The
   cytotoxicity dose metric was much more sensitive to interindividual variability than the average daily
   macromolecular binding was.

   Gearhart and colleagues (Gearhart et al. 1993) adjusted partition coefficients, rate of metabolism,
   cardiac output, and minute ventilation according to body temperature. These adjustments strengthened
   the Corley model (Corley et al. 1990) according to the fitting of gas uptake data of mice by loosening
   the assumption of enzyme loss and resysthesis.

   Chinery and Gleason (Chinery and Gleason 1993) further included the skin compartment to describe the
   fate of chloroform after adsorption through dermal exposure. The skin compartment was further divided
   into three sub compartments: the aqueous solution, stratum corneum, and viable skin.  The model was
   able to predict the concentration of chloroform in the exhaled air from humans exposed while showering
   through inhalation only and the combination of dermal and inhalation routes.

   In a PBPK model similar to Chinery and Gleason's, McKone (McKone 1993) assumed skin to be only
   one compartment. It was demonstrated that chloroform metabolism by the liver was not linear with
   respect to higher exposure concentrations (60-100 mg/L).

   Based on the Corley model  (Corley et al. 1990), Levesque and colleagues (Levesque et al. 2000) used a
   PBPK model to predict the fate of chloroform for individuals exposed while swimming through dermal
   and inhalation routes. Dermal exposure was described using an overall skin permeability constant. The
   levels of macromolecular binding in swimmers calculated from the model are much lower than the
   smallest no-observed-effect level for liver tumors in animals.

   Corley and colleagues (Corley et al. 2000) added a single skin compartment to the original  Corley
   model (Corley et al.  1990) and described the kinetics of human dermal exposure to chloroform while
   bathing.  With the adjustment of model parameters according to temperature, the model can predict the
   relationship between water temperature (30-40°C) and exhaled chloroform observed from experiments
   (Gordon et al. 1998).

   Constan and colleagues (Constan et al. 2002) used cytolethality and regerative cell proliferation as
   pharmacodynamic endpoints to perform a chloroform inhalation cancer risk assessment. The NOAEL
   for chloroform-induced hepatotoxicity in humans was estimated to be 110 ppm using experimental data
   from B6C3F1 mouse and PBPK-PD model calculations.

   Meek and colleagues (Meek et al.  2002) recently performed an assessment of exposure-response
   analyses and risk characterization using PBPK models.  Inhalation, oral, and dermal exposures were

                                               C-55

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   considered from ten-minute shower, discrete periods of water and food consumption, as well as
   inhalation of chloroform at a variety of concentrations. Dose metrics used for carcinogenicity were the
   maximum rate of metabolism per unit kidney cortex volume and mean rate of metabolism per unit
   kidney cortex volume during each dose interval. For non-neoplastic effects, the dose metrics used were
   the mean rate of metabolism per unit centrilobular region of the liver and the average concentration of
   chloroform in the non-metabolizing centrilobular region of the liver.

   7.4    Literature Cited

   Ade, P., Guastadisegni, C., Testai, E., and Vittozzi, L. (1994). Multiple activation of chloroform in
      kidney microsomes from male and female DBA/2J mice. Journal ofBiochem Pharmacol 9, 289-295.
   Aggazzotti, G., Fantuzzi, G., Righi, E., Tartoni, P., Cassinadri,  T., andPredieri, G. (1993). Chloroform
      in alveolar air of individuals attending indoor swimming pools. Arch Environ Health 48, 250-254.
   Allen, B. C.,  Covington, T. R., and Clewell, H.  J. (1996). Investigation of the impact of pharmacokinetic
      variability and uncertainty on risks predicted with a pharmacokinetic model for chloroform.
      Toxicology  111,289-303.
   Bergman, K.  (1979). Whole-body autoradiography and allied tracer techniques in distribution and
      elimination studies of some organic solvents: benzene, toluene, xylene, styrene, methylene chloride,
      chloroform, carbon tetrachloride and trichloroethylene. Scand J Work Environ Health 5 Suppl 1,1-
      263.
   Branchflower, R. V., Nunn, D. S., Highet, R. J., Smith, J. H., Hook, J. B., and Pohl, L. R. (1984).
      Nephrotoxicity of chloroform: metabolism to phosgene by the mouse kidney. ToxicolAppl
      Pharmacoll2, 159-168.
   Cammann, K., and Hubner, K. (1995). Trihalomethane concentrations in swimmers' and bath attendants'
      blood and urine after swimming or working in indoor swimming pools. Arch Environ Health 50, 61-
      65.
   Chemical Marketing Reporter. (1995). Chemical profile: Chloroform. Schnell Publishing, New York,
      NY.
   Chinery, R. L.,  and Gleason, A. K. (1993). A compartmental model for the prediction of breath
      concentration and absorbed dose of chloroform after exposure while showering. Risk Anal 13, 51-62.
   Cohen, E. N., and Hood, N. (1969). Application of low-temperature autoradiography to studies of the
      uptake and metabolism of volatile anesthetics in the mouse. I. Chloroform. Anesthesiology 30, 306-
      314.
   Constan, A. A., Wong, B. A., Everitt, J. I, and Butterworth, B. E. (2002). Chloroform  inhalation
      exposure conditions necessary to initiate liver toxicity in female B6C3F1 mice. Toxicol Sci 66, 201-
      208.
   Corley, R.  A., Gordon, S. M., and Wallace, L. A. (2000). Physiologically based pharmacokinetic
      modeling of the temperature-dependent dermal absorption of chloroform by humans following bath
      water exposures. Toxicol Sci 53, 13-23.
   Corley, R.  A., Mendrala,  A. L., Smith, F. A., Staats, D. A., Gargas, M. L., Conolly, R.  B., Andersen, M.
      E., and Reitz, R. H. (1990). Development of a physiologically based pharmacokinetic model for
      chloroform. ToxicolApplPharmacol 103, 512-527.
   Dick, D., Ng, K. M., Sauder, D. N., and Chu, I.  (1995). In vitro and in vivo percutaneous absorption of
      14C-chloroform in humans. Hum Exp Toxicol 14, 260-265.
   Fry, B. J., Taylor, T., and Hathway, D. E.  (1972). Pulmonary elimination of chloroform and its
      metabolite in man. Arch Int Pharmacodyn Ther 196, 98-111.
   Gearhart, J. M., Seckel, C., and Vinegar, A.  (1993). In vivo metabolism of chloroform in B6C3F1 mice
      determined by the method of gas uptake: the effects of body temperature on tissue partition
      coefficients and metabolism. Toxicol Appl Pharmacol 119,  258-266.

                                               C-56

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
   Gemma, S., Faccioli, S., Chieco, P., Sbraccia, M., Testai, E., and Vittozzi, L. (1996). In vivo
      bioactivation, toxicokinetics, toxicity, and induced compensatory cell proliferation in B6C3F1 male
      mice. Toxicol Appl Pharmacol 141, 394-402.
   Gordon, S. M., Wallace, L. A., Callahan, P. 1, Kenny, D. V., and Brinkman, M. C. (1998). Effect of
      water temperature on dermal exposure to chloroform. Environ Health Perspect 106, 337-345.
   Gordon, S. M., Wallace, L. A., Pellizzari, E. D., and Oneill, H. J. (1988). Human breath measurements
      in a clean-air chamber to determine half-lives for volativle organic-compounds. Atmos Environ 22,
      2165-2170.
   Ilett, K. F., Reid, W. D., Sipes, I. G., and Krishna, G. (1973). Chloroform toxicity in mice: correlation of
      renal and hepatic necrosis with covalent binding of metabolites to tissue macromolecules. ExpMol
      Pathol 19, 215-229.
   Islam, M. S., Zhao, L., McDougal, J. N.,  and Flynn, G. L. (1995). Uptake of chloroform by skin during
      short exposures to contaminated water. Risk Anal 15, 343-352.
   Jakobson, L, Wahlberg, J. E., Holmberg, B., and Johansson, G. (1982). Uptake via the blood and
      elimination of 10 organic solvents following epicutaneous exposure of anesthetized guinea pigs.
      Toxicol Appl Pharmacol 63, 181-187.
   Jo, W. K., Weisel, C. P., and Lioy, P. J. (1990). Routes of chloroform exposure and body burden from
      showering with chlorinated tap water. Risk Anal 10, 575-580.
   Levesque, B., Ayotte, P., LeBlanc, A., Dewailly, E., Prud'Homme, D., Lavoie, R., Allaire, S., and
      Levallois, P. (1994). Evaluation of dermal and respiratory chloroform exposure in humans. Environ
      Health Perspect 102, 1082-1087.
   Levesque, B., Ayotte, P., Tardif, R., Charest-Tardif, G., Dewailly, E., Prud'Homme, D., Gingras, G.,
      Allaire,  S., and Lavoie, R. (2000). Evaluation of the health risk associated with exposure to
      chloroform in indoor swimming pools. J Toxicol Environ Health A  61, 225-243.
   Levesque, B., Ayotte, P., Tardif, R., Ferron, L., Gingras, S., Schlouch, E.,  Gingras, G., Levallois, P., and
      Dewailly, E. (2002). Cancer risk associated with household exposure to chloroform. J Toxicol
      Environ Health A 65, 489-502.
   McConnell, G., Ferguson, D. M., and Pearson, C. R. (1975). Chlorinated hydrocarbons and the
      environment. Endeavour 34, 13-18.
   McCulloch, A. (2003). Chloroform in the environment: occurrence, sources, sinks and effects.
      Chemosphere 50, 1291-1308.
   McKone, T. E. (1993). Linking a PBPK model for chloroform with measured breath concentrations in
      showers: implications for dermal exposure models. J Expo Anal Environ Epidemiol 3 , 339-365.
   Meek, M. E., Beauchamp, R., Long, G., Moir, D., Turner, L., and Walker, M. (2002). Chloroform:
      exposure estimation, hazard characterization, and exposure-response analysis. J Toxicol Environ
      Health B CritRev 5, 283-334.
   Noort, D., Hulst, A. G., Fidder, A., van Gurp, R. A., de Jong, L. P.,  and Benschop, H. P. (2000). In vitro
      adduct formation of phosgene with albumin and hemoglobin in human blood. Chem Res Toxicol 13,
      719-726.
   Pereira, M. A. (1994). Route of administration determines whether chloroform enhances or inhibits cell
      proliferation in the liver of B6C3F1 mice. Fundam Appl Toxicoly 23, 87-92.
   Pereira, M. A., and Chang, L. W. (1981). Binding of chemical carcinogens and mutagens to rat
      hemoglobin. Chem Biol Interact 33, 301-305.
   Pereira, M. A., Chang, L. W., Ferguson, J. L., and Couri, D. (1984). Binding of chloroform to the
      cysteine of hemoglobin. Chem Biol Interact 51, 115-124.
   Pohl, L. R., Bhooshan, B., Whittaker, N.  F., and Krishna, G. (1977). Phosgene: a metabolite of
      chloroform. Biochem Biophys Res Commun 79, 684-691.
                                               C-57

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Pohl, L. R., Branchflower, R. V., Highet, R. 1, Martin, J. L., Nunn, D. S., Monks, T. 1, George, J. W.,
      and Hinson, J. A. (1981). The formation of diglutathionyl dithiocarbonate as a metabolite of
      chloroform, bromotrichloromethane, and carbon tetrachloride. DrugMetab Dispos 9, 334-339.
   Pohl, L. R., George, J. W., and Satoh, H. (1984). Strain and sex differences in chloroform-induced
      nephrotoxicity. Different rates of metabolism of chloroform to phosgene by the mouse kidney. Drug
      Metab Dispos 12, 304-308.
   Pohl, L. R., Martin, J. L., and George, J. W. (1980). Mechanism of metabolic activation of chloroform
      by rat liver microsomes. Biochem Pharmacol 29, 3271-3276.
   Rao, K. N., Virji, M. A., Moraca, M. A., Diven, W. F., Martin, T. G., and Schneider, S. M. (1993). Role
      of serum markers for liver function and liver regeneration in the management of chloroform
      poisoning. J Anal Toxicol 17, 99-102.
   Reitz, R. H., Mendrala, A. L., Corley, R. A., Quast, J. F., Gargas, M. L., Andersen, M. E., Staats, D. A.,
      and Conolly, R. B. (1990). Estimating the risk of liver cancer associated with human exposures to
      chloroform using physiologically based pharmacokinetic modeling. Toxicol Appl Pharmacol 105,
      443-459.
   Rosenthal, S. L. (1987). A review of the mutagenicity of chloroform. Environ MolMutagen 10, 211-
      226.
   Roy, A., Weisel, C. P., Lioy, P. J., and Georgopoulos, P. G. (1996). A distributed parameter
      physiologically-based pharmacokinetic model for dermal and inhalation exposure to volatile organic
      compounds. Risk Anal 16, 147-160.
   Smith, A. A., Volpitto, P. P., Gramling, Z. W., DeVore, M. B., and Glassman, A. B. (1973).
      Chloroform, halothane, and regional anesthesia: a comparative study. AnesthAnalg 52, 1-11.
   Smith, J. H., and Hook, J. B. (1984). Mechanism of chloroform nephrotoxicity. III. Renal and hepatic
      microsomal metabolism of chloroform in mice. Toxicol Appl Pharmacol 73, 511-524.
   Stevens, J. L., and Anders, M. W. (1981). Effect of cysteine, diethyl maleate, and phenobarbital
      treatments on the hepatotoxicity of [1H]chloroform. Chem BiolInteract 37, 207-217.
   USEPA (2001). Toxicological Review of Chloroform. U.S. Environmental Protection Agency,
      Washington, DC.
   Withey, J. R.,  Collins, B. T., and Collins, P. G. (1983). Effect of vehicle on the pharmacokinetics and
      uptake of four halogenated hydrocarbons from the gastrointestinal tract of the rat. J Appl Toxicol 3,
      249-253.
                                                C-58

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.


                                       Trichloroethylene

   8.0    Introduction

   Trichloroethylene (TCE) is one of the most important industrial chemicals of our time. It is an organic
   solvent that has been used widely in dry cleaning, metal degreasing, and as a solvent for oils and resins.
   Because of the large production volume and its wide applications, TCE is one of the top, if not the top,
   environmental pollutants in ground water. Recently, an entire supplemental volume (Volume 108,
   Supplement 2, May 2000) of Environmental Health Perspective was devoted to Trichloroethylene
   Health Risks.  This volume contains many excellent review articles which cover the areas of toxicology
   and risk assessment extensively. The USEPA, in its re-assessment of TCE health risks, devoted a great
   deal of effort to publishing a document on Trichloroethylene Health Risk Assessment: Synthesis and
   Characterization (EPA/600/P-01/002A; quoted in this write-up as USEPA, 2001). This document and
   its related Science Advisory Board review (www.epa.gov/sciencel/pdf/ehc03002.pdf) also provided an
   excellent source of information on TCE.  Thus, the summary below  represents a brief update of the
   current information.

   8.1    Toxic effects

   8.1.1  Noncancer effects

   Neuro- or neuro-behavioral (Boyes et al., 2000; Ohta et al., 2001; USEPA, 2001; Waseem et al., 2001;
   Kilburn, 2002; Moser et al., 2003), male reproductive (Kumar et al., 2000; 2001; Forkert et al., 2002;
   2003), developmental (Boyer et al., 2000; Rodenbeck et al., 2000; USEPA, 2001; Johnson et al., 2003),
   liver (USEPA, 2001), renal (USEPA, 2001; Mensing et al., 2002), immuno- (Griffin et al., 2000a,b,c;
   Kaneko et al., 2000; USEPA, 2001) toxicities in experimental  animals and/or humans are reported or
   implicated. In vitro studies using isolated cell cultures have demonstrated and reconfirmed many of the
   species-, sex-, and tissue-dependent differences in hepato- and renal- toxicities observed in vivo
   (Cummings et al., 2000a,b; Lash et al., 2001).

   Halogenated hydrocarbons such as TCE are among the most common water supply contaminants in the
   U.S. and elsewhere. Epidemiological studies have found an association, but not a cause-and-effect
   relation, between halogenated hydrocarbon contamination and increased incidence of congenital cardiac
   malformations or other defective birth outcomes. However, some animal studies in birds and rats as
   well as in tissue cultures had demonstrated statistically significant increased incidence of congenital
   cardiac malformations or other defective birth outcomes (Boyer et al., 2000; Johnson et al., 2003) while
   others turned out negative (Fisher et al., 2001). The most recent  study (Johnson et al., 2003) reported
   that maternal rats exposed to more than 250 ppb TCE,  a very low dose study, showed an associated
   increased incidence of cardiac malformations in their developing fetuses.

   8.1.2  Cancer effects

   TCE causes liver, lung tumors and lymphomas in mice and kidney and testicular tumors in rats (Bull,
   2000; Green, 2000; Lash et al., 2000a; USEPA, 2001). In humans, TCE was implicated to be a
   carcinogen (Wartenberg et al., 2000; USEPA, 2001).  It is well established that two metabolites of TCE,
   dichloroacetic acid (DCA) and trichloroacetic acid (TCA) are important contributors to carcinogenicity
   of TCE (Bull, 2000; Tao et al., 2000; Bull et al., 2002).
                                               C-59

-------
USEPA Contract No. 3C-R102-NTEX                         Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

  Regarding renal cancer in humans, German epidemiological studies of prevalence of renal cancer
  following high exposure of TCE in workers have been the subject of considerable scientific debate, re-
  evaluation, and repeated studies (Brauch et al., 1999; Bruning et al., 1999; Green and Lash,  1999;
  Schraml et al., 1999; Brauch et al., 2000; Bruning and Bolt, 2000; Green, 2000; Lash et al., 2000a;
  Wartenberg et al., 2000; USEPA, 2001; Bruning et al., 2003).

  8.2    Pharmacokinetics

  The pharmacokinetics of TCE has been reviewed thoroughly (ATSDR, 1997; Fisher, 2000).  More
  recent updates are provided below.

  8.2.1  Absorption

  Dose-dependent gastrointestinal absorption of TCE and its kinetics in male Sprague-Dawley rats over a
  wide range of oral bolus doses were characterized by Lee et al. (2000b). Dietary incorporation of guar
  gum, a thickener and stabilizer in foods and pharmaceuticals, was found to decrease TCE accumulation
  in the body by reducing absorption and fat tissue mass (Nakashima and Ikegami, 2001).

  8.2.2  Distribution

  PBPK models for the systemic transport of TCE to various tissues and organs with a special emphasis to
  fat tissues were established by Albanese et al. (2002).

  8.2.3  Metabolism

  Human and animal studies

  The principal metabolic pathways for TCE and metabolic steps where interactions with chloroform
  (CHL),  tetrachloroethylene (PERC), and/or 1,1,1 trichloroethane (MC) may occur are denoted in Figure
  8.1.  The metabolism of TCE has been  reviewed thoroughly (Lash et al., 2000b;  USEPA, 2001); more
  recent updates are provided below.

  Lash et al.  (1999a) reported direct, in vivo, evidence of GSH conjugation of TCE in human volunteers
  exposed to 100 ppm TCE and demonstrated markedly higher amounts of S-(l,2-dichlorovinyl)
  glutathione (DCVG) in males than females. However, Bloemen et al. (2001) studied urinary
  concentrations of metabolites from GST-dependent pathway in human volunteers exposed to 50 and 100
  ppm TCE for 15 min or occupationally exposed (0.4 to 21 ppm TWA) workers.  They found little or no
  such metabolites and suggested the glutathione-mediated metabolism is of minor importance in humans.

  There were evidences suggesting that TCE is metabolized in the reproductive tract of the mouse and
  monkey; the fact that TCE and its metabolites accumulated in seminal fluid in human diagnosed with
  clinical  infertility also suggested associations between production of TCE metabolites, reproductive
  toxicity, and impaired fertility (Forkert et al., 2003).

  In vitro studies

  Extensive in vitro biotransformation studies have been published on a variety of enzyme preparations
  and  cell culture systems ranging from cell free tissue preparations (Lipscomb etal., 1997, 1998;
  Lipscomb and Garrett, 1998; Lash et al., 1999b; Cai and Guengerich, 2000; Snawder and Lipscomb,

                                               C-60

-------
USEPA Contract No. 3C-R102-NTEX
                                                   Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
   2000; Cummings et al., 2001; Lipscomb etal., 2003a) to highly purified human enzymes (Cai and
   Guengerich, 2001), to primary and other cell cultures (Lash et al., 1999b; Cummings and Lash, 2000;
   Cummings et al., 2000a,b; Walgren et al., 2000; Cummings et al., 2001) including collagen gel
   sandwich cultures of rat hepatocytes (De Smet et al., 2000).
re °-' ci ci fci^
\ / 	 ^. c=C=S ^_^_k Cytotoxicity
/ \ ' ^^^* Mutagenicity
/3-Lyase -^< H SH |_ H J
X^>C9 1,2-DVT Thioketene
Cl Cl Cl Cl Cl Cl
\ 	 / 	 »- 	 »- \ — / , - \ 	 / 	 •- Excretion
H SG H S NH3+ H NS NHCOCH3 '" U»ne
\ I, \ I,
._. S-f^-DCVC TH T
A (i) COO- COOH
GSH Tr^N
Cl\ /Cl CYP450
/~Y~^
GSH W
CYP450
1
Fe
6" Cl
Cl Cl
O O
CI3C— \ 	 *• CI2CH— \
_^_ ^ TCA °H
*• CCI3 | H x^^ y 	 CCI3
OH >V H0^CC| 	 »- gluc-0
CHO TCOH TCOG
                 CI
                      SG
                                                     ci
                                                             NHCOCH
                                                                          Excretion
                                                                          in Urine
                                 S-1,1-DCVC  COO-
                                                             COOH
                                                         /
                                                           H
                    Potential interaction between PCE and TCE
                    Potential interaction between MC and TCE
                    Potential interaction between CHL and TCE
^^-   Cytotoxicity
^^^   Mutagenicity
 Figure 8.1 Superimposed on a metabolic
 schematic for TCE (Dobrev, et al., 2001),
 metabolic steps where pharmacokinetic
 interactions could occur are denoted by
 symbols. Identification of additional
 potential interactions should occur during
 CRA Steps 3 and 6.
   Cai and Guengerich (2001) demonstrated that the direction reaction of TCE oxide with either human
   P450 2E1, P450 2B1, or NADPH-P450 reductase was shown to lead to enzyme inactivation, and no
   recovery of either enzyme occurred.

   8.2.4   Elimination and Excretion

   Presystemic elimination of TCE has been shown by Lee et al. (1996) to be inversely related to dose.
   When relatively high doses were administered to rats via the portal vein, first-pass hepatic extraction
   became negligible. This phenomenon could result not only from metabolic saturation, but from suicidal
   destruction of cytochrome P450 and hepatocellular injury as well (Lee et al., 2000a). Subsequent
   pharmacokinetic analysis by Lee et al. (2000b) indicated that TCE was eliminated by capacity-limited
   hepatic metabolism, no evidence for P450 2E1  destruction, with incursion into nonlinear kinetics with
   bolus doses greater or equal to 8 to 16 mg/kg.
                                                  C-61

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   8.3    Interactions of TCE with other chemicals

   Interaction studies reported in the recent literature covered diverse subject areas. Each of the relevant
   papers is discussed briefly below.

   Dobrev et al. (2001; 2002) studied "Interaction Thresholds" in rats and humans using interactive PBPK
   modeling of a ternary mixture of TCE, tetrachloroethylene, and 1,1,1-trichloroethane. Because of
   competitive inhibition of the primary metabolic system, P450 2E1, an alternative pathway, the GST
   conjugation system, becomes important. It was demonstrated that at or below the current threshold limit
   values (TLVs) for these three chemicals, the coexposure to these chemicals would result in significant
   interactions.

   Very high doses (2000 to 5000 mg/kg, ip) of TCE induced anticonvulsive effect of a number of drugs
   (Shih et al., 2001); it was suggested that this effect might be predominantly mediated by GABA
   receptors.

   A full-factorial design for neurobehavioral evaluations of mixtures of TCE, heptachlor,  and di (2-
   ethylhexyl) phthalate in F344 rats was carried out by Moser et al. (2003).  In general, significant overall
   interactions that deviated from response additivity were detected for most endpoints (11 of 14). Most of
   the interactions are antagonistic in nature.

   Pretreatment of TCE in Sprague-Dawley rats altered drug kinetics of theophylline, quinidine, and
   pentobarbital (Kukongviriyapan et al., 2001).

   Dietary incorporation of guar gum, a thickener and stabilizer in foods and pharmaceuticals, was found to
   decrease TCE accumulation in the body by reducing absorption and fat tissue mass (Nakashima and
   Ikegami, 2001).

   8.4    PBPK models

   TCE is undoubtedly one of the chemicals, which were most extensively studied using PBPK modeling
   technique.  The initial development of PBPK models was reported by Andersen et al. (1987). This
   initial PBPK model for TCE was followed by a number of variations by others for different goals
   (Fisher et al., 1989; 1990; Koizumi,  1989; Dallas et al., 1991). As the science advances, more and more
   sophistication were incorporated into the later PBPK models.  Thus, PBPK models with incorporation of
   TCE metabolites, as well as reproductive physiology and toxicology (Fisher et al., 1989; 1990; 1991;
   Abbas et al.,  1996; Abbas and Fisher, 1997; Fisher et al., 1998; Greenberg et al., 1999), and
   pharmacokinetic and pharmacodynamic interactions (Elmasri et al., 1996; Byczkowski et al., 1999) were
   seen in the literature. Furthermore, the application of PBPK modeling in risk assessment received
   progressively more emphasis (Allen and Fisher, 1993; Fisher and Allen, 1993; Gearhart et al., 1993;
   Clewell et al., 1995; Cronin et al., 1995; Bogen and Gold,  1997; Simon, 1997).  The 2000 Monograph in
   EHP and more recent PBPK modeling efforts included its application in risk assessment (Clewell et al.,
   2000; Fisher, 2000),  statistical analyses for variability  and uncertainty (Bois, 2000a,b), further
   toxicological interaction studies to define "Interaction Thresholds" (Dobrev et al., 2001; 2002). PBPK
   modeling studies for other specific purposes or toxic endpoints are also seen.  Thus, Poet et al. (2000)
   utilized PBPK modeling for assessing percutaneous absorption of TCE in rats and humans.  PBPK
   models for the transport of TCE in adipose tissues were reported by Albanese et al. (2002) and PBPK
   modeling for male Long-Evans rats to aid in evaluation of neurotoxicity data was published by Simmons
   et al. (2002).

                                               C-62

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
   8.5    Risk assessment related

   Because TCE is a very important industrial chemical and a prevalent environmental pollutant, the risk
   assessment, particularly cancer risk assessment became an area of much scientific debate.
   Consequently, quite a number of publications, review articles, and documents are available specifically
   dealing with mechanisms of toxicity of TCE and PBPK modeling in relation to risk assessment, as well
   as the process of risk assessment of TCE (Allen and Fisher, 1993; Fisher and Allen, 1993; Gearhart et
   al., 1993; Clewell et al., 1995; Cronin et al., 1995; Bogen and Gold, 1997; Simon, 1997; Brauch et al.,
   1999; Bruning et al., 1999; Green and Lash, 1999; Motohashi et al., 1999a,b; Schraml et al., 1999;
   Barton and Clewell, 2000; Bois, 2000a,b; Bruning and Bolt, 2000; Bull, 2000; Chen,  2000; Clewell et
   al., 2000; Fisher, 2000; Green, 2000; Lash et al., 2000a,b; Moore and Harington-Brock, 2000; Pastino et
   al., 2000; Rhomberg, 2000; Wartenberg et al., 2000; Ruden, 2001a,b; Stewart, 2001; USEPA, 2001;
   Ruden 2001a,b; 2002a,b; Bruning et al., 2003; Lipscomb etal., 2002; 2003b; Ruden,  2003).

   8.6    Literature Cited

   Abbas, R., and Fisher, J. W. (1997). A physiologically based pharmacokinetic model  for
      trichloroethylene and its metabolites, chloral hydrate, trichloroacetate, dichloroacetate,
      trichloroethanol, and trichloroethanol glucuronide in B6C3F1 mice. Toxicol Appl Pharmacol 147,
      15-30.
   Abbas, R. R., Seckel, C. S., Kidney, J. K., and Fisher, J.  W. (1996). Pharmacokinetic  analysis of chloral
      hydrate and its metabolism in B6C3F1 mice.[erratum appears in Drug Metab Dispos 1997
      Dec;25(12):1449]. Drug Metab Dispos 24, 1340-6.
   Albanese, R. A., Banks, H. T., Evans, M. V., and Potter, L. K. (2002). Physiologically based
      pharmacokinetic models for the transport of trichloroethylene in adipose tissue. Bull Math Biol 64,
      97-131.
   Allen, B. C., and Fisher, J. W. (1993). Pharmacokinetic modeling of trichloroethylene and
      trichloroacetic acid in humans. Risk Anal 13, 71-86.
   Andersen, M. E., Gargas, M. L., Clewell, H. J., 3rd, and Severyn, K. M. (1987). Quantitative evaluation
      of the metabolic interactions between trichloroethylene and 1,1-dichloroethylene in vivo using gas
      uptake methods. Toxicol Appl Pharmacol 89, 149-57.
   ATSDR.  1997.  Toxicological Profile for Trichloroethylene. Agency for Toxic Substances and Disease
      Registry. 335 pp.
   Barton, H. A., and Clewell, H. J., 3rd (2000). Evaluating noncancer effects of trichloroethylene:
      dosimetry, mode of action, and risk assessment. Environ Health Perspect 108 Suppl 2, 323-34.
   Bloemen, L. J., Monster, A. C., Kezic, S., Commandeur, J. N., Veulemans, H., Vermeulen, N. P., and
      Wilmer, J. W. (2001).  Study on the cytochrome P-450- and glutathione-dependentbiotransformation
      of trichloroethylene in humans. Int Arch Occup Environ Health 74, 102-8.
   Bogen, K. T., and Gold, L. S. (1997). Trichloroethylene cancer risk: simplified calculation of PBPK-
      based MCLs for cytotoxic end points. Reg Toxicol Pharmacol 25, 26-42.
   Bois, F. Y. (2000a). Statistical analysis of Fisher et al. PBPK model of trichloroethylene kinetics.
      Environ Health Perspect 108 Suppl 2, 275-82.
   Bois, F. Y. (2000b). Statistical analysis of Clewell et al. PBPK model of trichloroethylene kinetics.
      Environ Health Perspect 108 Suppl 2, 307-16.
   Boyer, A. S., Finch, W. T., and Runyan, R. B. (2000). Trichloroethylene inhibits development of
      embryonic heart valve precursors in vitro.[comment]. ToxicolSci 53, 109-17.
                                                C-63

-------
USEPA Contract No. 3C-R102-NTEX                           Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Boyes, W. K., Bushnell, P. J., Crofton, K. M., Evans, M., and Simmons, J. E. (2000). Neurotoxic and
      pharmacokinetic responses to trichloroethylene as a function of exposure scenario. Environ Health
      Perspect 108 Suppl 2, 317-22.
   Brauch, H., Weirich, G., Brieger, J., Glavac, D., Rodl, H., Eichinger, M., Feurer, M., Weidt, E.,
      Puranakanitstha, C., Neuhaus, C., Pomer, S., Brenner, W., Schirmacher, P., Storkel, S., Rotter, M.,
      Masera, A., Gugeler, N., and Decker, H. J. (2000). VHL alterations in human clear cell renal cell
      carcinoma: association with advanced tumor stage and a novel hot spot mutation. Cancer Res 60,
      1942-8.
   Brauch, H., Weirich, G., Hornauer, M. A., Storkel, S., Wohl, T., and Bruning, T. (1999).
      Trichloroethylene exposure and specific somatic mutations in patients with renal cell carcinoma. J
      Natl Cancer Inst 91, 854-61.
   Bruning, T., and Bolt, H. M. (2000). Renal toxicity and carcinogenicity of trichloroethylene: key results,
      mechanisms, and controversies. CritRev Toxicol 30, 253-85.
   Bruning, T., Mann, H., Melzer, H., Sundberg, A. G., and Bolt, H. M. (1999). Pathological excretion
      patterns of urinary proteins in renal cell cancer patients exposed to trichloroethylene. OccupMed
      (Lond) 49, 299-305.
   Bruning, T., Pesch, B., Wiesenhutter, B., Rabstein, S., Lammert, M., Baumuller, A., and Bolt, H. M.
      (2003). Renal cell cancer risk and occupational exposure to trichloroethylene: results of a
      consecutive case-control study in Arnsberg, Germany. Am JIndMed43, 274-85.
   Bull, R. J.  (2000). Mode of action of liver tumor induction by trichloroethylene and its metabolites,
      trichloroacetate and dichloroacetate. Environ Health Perspect 108 Suppl 2, 241-59.
   Bull, R. J., Orner, G. A., Cheng, R. S., Stillwell, L., Stauber, A. J., Sasser,  L. B., Lingohr, M. K., and
      Thrall, B. D. (2002). Contribution of dichloroacetate and trichloroacetate to liver tumor induction in
      mice by trichloroethylene. ToxicolApplPharmacol 182, 55-65.
   Byczkowski, J. Z., Channel, S. R., and Miller, C. R. (1999). A biologically based pharmacodynamic
      model  for lipid peroxidation stimulated by trichloroethylene in vitro. JBiochemMol Toxicol 13,
      205-14.
   Cai, H., and Guengerich, F. P. (2000). Acylation of protein lysines by trichloroethylene oxide. Chem Res
      Toxicol 13, 327-35.
   Cai, H., and Guengerich, F. P. (2001). Reaction of trichloroethylene and trichloroethylene oxide with
      cytochrome P450 enzymes: inactivation and sites of modification. Chem Res Toxicol 14, 451-8.
   Chen, C. W. (2000). Biologically based dose-response model for liver tumors induced by
      trichloroethylene. Environ Health Perspect 108 Suppl 2, 335-42.
   Clewell, H. J., 3rd, Gentry, P. R., Covington, T. R., and Gearhart, J. M. (2000). Development of a
      physiologically based pharmacokinetic model of trichloroethylene and its metabolites for use in risk
      assessment. Environ Health Perspect 108 Suppl 2, 283-305.
   Clewell, H. J., Gentry, P. R., Gearhart, J. M., Allen, B. C., and Andersen, M. E.  (1995). Considering
      pharmacokinetic and mechanistic information in cancer risk assessments for environmental
      contaminants: examples with vinyl chloride and trichloroethylene. Chemosphere 31, 2561-78.
   Cronin, W. J. t, Oswald, E. J., Shelley, M. L., Fisher, J. W., and Flemming, C. D.  (1995). A
      trichloroethylene risk assessment using a Monte Carlo analysis of parameter uncertainty in
      conjunction with physiologically-based pharmacokinetic modeling. Risk Anal 15, 555-65.
   Cummings, B. S., and Lash, L. H. (2000). Metabolism and toxicity of trichloroethylene and S-(l,2-
      dichlorovinyl)-L-cysteine in freshly isolated human proximal tubular cells. Toxicol Sci 53, 458-66.
   Cummings, B. S., Parker, J. C., and Lash, L. H. (2000b). Role of cytochrome P450 and  glutathione S-
      transferase alpha in the metabolism and cytotoxicity of trichloroethylene in rat kidney. Biochem
      Pharmacol 59, 531-43.
   Cummings, B. S., Parker, J. C., and Lash, L. H. (2001).  Cytochrome p450-dependent metabolism of
      trichloroethylene in rat kidney. Toxicol Sci 60, 11-9.

                                                C-64

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Cummings, B. S., Zangar, R. C., Novak, R. F., and Lash, L. H. (2000a). Cytotoxicity of
      trichloroethylene and S-(l, 2-dichlorovinyl)-L-cysteine in primary cultures of rat renal proximal
      tubular and distal tubular cells. Toxicology 150, 83-98.
   Dallas, C. E., Gallo, J. M., Ramanathan, R., Muralidhara, S., and Bruckner, J. V. (1991). Physiological
      pharmacokinetic modeling of inhaled trichloroethylene in rats. Toxicol Appl Pharmacol 110, 303-14.
   De Smet, K., Bruning, T., Blaszkewicz, M., Bolt, H. M., Vercruysse, A., and Rogiers, V. (2000).
      Biotransformation of trichloroethylene in collagen gel sandwich cultures of rat hepatocytes. Arch
      Toxicol 74, 587-92.
   Dobrev, I. D., Andersen, M. E., and Yang, R. S. H. (2001). Assessing interaction thresholds for
      trichloroethylene in combination with tetrachloroethylene and  1,1,1-trichloroethane using gas uptake
      studies and PBPK modeling. Arch Toxicol 75, 134-44.
   Dobrev, I. D., Andersen, M. E., and Yang, R. S. H. (2002). In silico toxicology: simulating interaction
      thresholds for human exposure to mixtures of trichloroethylene, tetrachloroethylene, and 1,1,1-
      trichloroethane. Environ Health Perspect 110, 1031-9.
   El-Masri, H.  A., Constan,  A. A.,  Ramsdell, H. S., and Yang, R. S. H. (1996). Physiologically based
      pharmacodynamic  modeling  of an interaction threshold  between  trichloroethylene  and  1,1-
      dichloroethylene in Fischer 344 rats. Toxicol Appl Pharmacol 141, 124-132.
   Fisher, J. W. (2000). Physiologically based pharmacokinetic models for trichloroethylene and its
      oxidative metabolites. Environ Health Perspect 108 Suppl 2, 265-73.
   Fisher, J. W., and Allen, B. C. (1993). Evaluating the risk of liver cancer in humans exposed to
      trichloroethylene using physiological models. Risk Anal 13, 87-95.
   Fisher, J. W., Channel, S. R., Eggers, J. S., Johnson, P. D., MacMahon, K. L., Goodyear, C. D.,
      Sudberry, G. L., Warren, D. A., Latendresse, J. R., and Graeter, L. J.  (2001). Trichloroethylene,
      trichloroacetic acid, and dichloroacetic acid: do they affect fetal rat heart development? Int J Toxicol
      20, 257-67.
   Fisher, J. W., Gargas, M. L.,  Allen, B. C., and Andersen, M. E. (1991). Physiologically based
      pharmacokinetic modeling with trichloroethylene and its metabolite, trichloroacetic acid, in the rat
      and mouse. Toxicol Appl Pharmacol 109, 183-95.
   Fisher, J. W., Mahle, D., and Abbas, R. (1998). A human physiologically based pharmacokinetic model
      for trichloroethylene and its metabolites, trichloroacetic acid and free trichloroethanol. Toxicol Appl
      Pharmacol 152, 339-59.
   Fisher, J. W., Whittaker, T. A., Taylor, D. H., Clewell, H. J., 3rd, and Andersen, M. E. (1989).
      Physiologically based pharmacokinetic modeling of the pregnant rat: a multiroute exposure model
      for trichloroethylene and its metabolite, trichloroacetic acid. Toxicol Appl Pharmacol99, 395-414.
   Fisher, J. W., Whittaker, T. A., Taylor, D. H., Clewell, H. J., 3rd, and Andersen, M. E. (1990).
      Physiologically based pharmacokinetic modeling of the lactating rat and nursing pup: a multiroute
      exposure model for trichloroethylene and its metabolite, trichloroacetic acid. Toxicol Appl
      Pharmacol 102, 497-513.
   Forkert, P. G., Lash, L. H., Nadeau, V., Tardif, R.,  and Simmonds, A. (2002). Metabolism and toxicity
      of trichloroethylene in epididymis and testis. Toxicol Appl Pharmacol 182, 244-254.
   Forkert, P. G., Lash, L., Tardif, R., Tanphaichitr, N., Vandevoort, C., and Moussa, M. (2003).
      Identification of trichloroethylene and its metabolites in human seminal fluid of workers exposed to
      trichloroethylene. DrugMetab Dispos 31, 306-11.
   Gearhart, J. M., Mahle, D. A., Greene, R. J., Seckel, C. S., Flemming, C. D., Fisher, J. W., and Clewell,
      H. J., 3rd (1993). Variability of physiologically based pharmacokinetic (PBPK) model parameters
      and their effects on PBPK model predictions in a risk assessment for perchloroethylene (PCE).
      Toxicol Lett 68, 131-44.
   Green, L. C., and Lash, T. L. (1999). Re: "Renal cell cancer correlated with occupational exposure to
      trichloroethylene". J Cancer Res Clin Oncol 125, 430-2.

                                                C-65

-------
USEPA Contract No. 3C-R102-NTEX                           Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Green, T. (2000). Pulmonary toxicity and carcinogen!city of trichloroethylene: species differences and
      modes of action. Environ Health Perspect 108 Suppl 2, 261-4.
   Greenberg, M. S., Burton, G. A., and Fisher, J. W. (1999). Physiologically based pharmacokinetic
      modeling of inhaled trichloroethylene and its oxidative metabolites in B6C3F1 mice. ToxicolAppl
      Pharmacol 154, 264-78.
   Griffin, J. M., Blossom, S. J., Jackson, S. K., Gilbert, K. M., and Pumford, N. R. (2000a).
      Trichloroethylene accelerates an autoimmune response by Thl T cell activation in MRL +/+ mice.
      Immunopharmacology 46, 123-37.
   Griffin, J. M., Gilbert, K. M., Lamps, L. W., and Pumford,  N. R. (2000b). CD4(+) T-cell activation and
      induction of autoimmune hepatitis following trichloroethylene treatment in MRL+/+ mice. Toxicol
      Sci 57, 345-52.
   Griffin, J. M., Gilbert, K. M., and Pumford, N. R. (2000c).  Inhibition of CYP2E1 reverses CD4+ T-cell
      alterations in trichloroethylene-treated MRL+/+ mice. Toxicol Sci 54, 384-9.
   Johnson, P. D., Goldberg, S. J., Mays, M. Z., and Dawson,  B. V. (2003). Threshold of trichloroethylene
      contamination in maternal drinking waters affecting fetal heart development in the rat. Environ
      Health Perspect 111, 289-92.
   Kaneko, T., Saegusa, M., Tasaka, K., and Sato, A. (2000). Immunotoxicity of trichloroethylene: a study
      with MRL-lpr/lpr mice. J Appl Toxicol 20, 471-5.
   Kilburn, K. H. (2002). Is neurotoxicity associated with environmental trichloroethylene (TCE)? Arch
      Environ Health 57, 113-20.
   Koizumi, A.  (1989). Potential of physiologically based pharmacokinetics to amalgamate kinetic data of
      trichloroethylene and tetrachloroethylene obtained in rats and man. Br JIndMed 46, 239-49.
   Kukongviriyapan, V., Simajareuk,  S., Kukongviriyapan, U., Cha-on, U., and Airarat, W. (2001).
      Alteration of drug kinetics in rats following exposure to trichloroethylene. Pharmacology 63, 90-4.
   Kumar, P., Prasad, A. K., and Dutta, K. K. (2000). Steroidogenic alterations in testes and sera of rats
      exposed to trichloroethylene (TCE) by inhalation. Hum Exp Toxicol 19, 117-21.
   Kumar, P., Prasad, A. K., Mani, U., Maji, B. K.,  and Dutta, K. K. (2001). Trichloroethylene induced
      testicular toxicity in rats exposed by inhalation. Hum Exp Toxicol 20, 585-9.
   Lash, L. H., Fisher, J. W., Lipscomb, J. C., and Parker, J. C. (2000b). Metabolism of trichloroethylene.
      Environ Health Perspect 108 Suppl 2, 177-200.
   Lash, L. H., Lipscomb, J. C., Putt, D. A., and Parker, J.  C. (1999b). Glutathione conjugation of
      trichloroethylene in human liver and kidney:  kinetics and individual variation. DrugMetab Dispos
      27,351-9.
   Lash, L. H., Parker, J. C., and Scott, C. S. (2000a). Modes of action of trichloroethylene for kidney
      tumorigenesis. Environ Health Perspect 108  Suppl 2, 225-40.
   Lash, L. H., Putt, D. A., Brashear, W. T., Abbas, R., Parker, J. C., and Fisher, J. W. (1999a).
      Identification of S-(l,2-dichlorovinyl)glutathione in the blood of human volunteers exposed to
      trichloroethylene. J Toxicol Environ Health A 56, 1-21.
   Lash, L. H., Qian, W., Putt, D. A., Hueni, S. E., Elfarra, A. A., Krause, R. J., and Parker, J. C. (2001).
      Renal and hepatic toxicity of trichloroethylene and its glutathione-derived metabolites in rats and
      mice: sex-, species-, and tissue-dependent differences. J Pharmacol Exp Ther297, 155-64.
   Lee, K. M., Bruckner, J. V., Muralidhara, S., and Gallo, J. M. (1996). Characterization of presystemic
      elimination of trichloroethylene and its nonlinear kinetics in rats. Toxicol Appl Pharmacol 139, 262-
      71.
   Lee, K. M., Muralidhara, S., Schnellmann, R. G., and Bruckner, J. V. (2000a). Contribution of direct
      solvent injury to the dose-dependent kinetics of trichloroethylene: portal vein administration to rats.
      Toxicol Appl Pharmacol 164, 46-54.
                                                C-66

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Lee, K. M., Muralidhara, S., White, C. A., and Bruckner, J. V. (2000b). Mechanisms of the dose-
      dependent kinetics of trichloroethylene: oral bolus dosing of rats. Toxicol Appl Pharmacol 164, 55-
      64.
   Lipscomb, J. C., Fisher, J. W., Confer, P. D., and Byczkowski, J. Z. (1998). In vitro to in vivo
      extrapolation for trichloroethylene metabolism in humans. Toxicol Appl Pharmacol 152, 376-387.
   Lipscomb, J. C., and Garrett, C. M. (1998). Effect of organ procurement conditions on cytochrome P-
      450 activity in rat liver microsomes. In Vitro Mol Toxicol 11, 265-270.
   Lipscomb, J. C., Garrett, C. M., and Snawder, J. E. (1997). Cytochrome P450-dependent metabolism of
      trichloroethylene: Interindividual differences in humans. Toxicol Appl Pharmacol 142, 311-318.
   Lipscomb, J. C., and Kedderis, G. L. (2002). Incorporating human interindividual biotransformation
      variance in health risk assessment. Sci Total Environ 288, 12-21.
   Lipscomb, J. C., Teuschler, L. K., Swartout, J. C., Striley, C. A. F., and Snawder, J. E. (2003a).
      Variance of microsomal  protein and cytochrome P450 2E1 and 3 A forms in adult human liver.
      ToxicolMechMethods 13, 45-51.
   Lipscomb, J. C., Teuschler, L. K., Swartout, J., Popken, D., Cox, T.,  and Kedderis, G. L. (2003b). The
      impact of cytochrome  P450 2El-dependent metabolic variance on a risk-relevant pharmacokinetic
      outcome in humans. Risk Anal In press.
   Mensing, T., Welge, P., Voss, B., Fels, L. M., Fricke, H. H., Bruning, T., and Wilhelm, M. (2002).
      Renal toxicity  after chronic inhalation exposure of rats to trichloroethylene. Toxicol Lett 128, 243-7.
   Moore, M. M., and Harrington-Brock, K. (2000). Mutagenicity of trichloroethylene and its metabolites:
      implications for the risk  assessment of trichloroethylene. Environ Health Perspect 108 Suppl 2, 215-
      23.
   Moser, V. C., MacPhail, R. C., and Gennings, C. (2003). Neurobehavioral evaluations of mixtures of
      trichloroethylene, heptachlor, and di(2-ethylhexyl)phthlate in a full-factorial design. Toxicology 188,
      125-37.
   Motohashi, N., Nagashima, H., and Molnar, J. (1999a). Trichloroethylene. I. Carcinogenicity of
      trichloroethylene. In Vivo 13, 211-4.
   Motohashi, N., Nagashima, H., and Molnar, J. (1999b). Trichloroethylene. II. Mechanism of
      carcinogenicity of trichloroethylene. In Vivo 13, 215-9.
   Nakashima, Y., and Ikegami, S. (2001). Guar gum reduces trichloroethylene accumulation in the body
      by reducing TCE absorption and fat tissue mass. JAgFoodChem 49, 3499-505.
   Ohta, M., Saito, T., Saito,  K., Kurasaki, M., and Hosokawa, T. (2001). Effect of trichloroethylene on
      spatiotemporal pattern of LTP in mouse hippocampal slices. Int JNeurosci 111, 257-71.
   Pastino, G. M., Yap, W. Y.,  and Carroquino, M. (2000). Human variability and susceptibility to
      trichloroethylene. Environ Health Perspect 108 Suppl 2, 201-14.
   Poet, T. S., Corley, R. A.,  Thrall, K. D., Edwards, J. A., Tanojo, H., Weitz, K.  K., Hui, X., Maibach, H.
      L, and Wester, R. C. (2000).  Assessment of the percutaneous absorption of trichloroethylene in rats
      and humans using MS/MS real-time breath analysis and physiologically based pharmacokinetic
      modeling. Toxicol Sci  56, 61-72.
   Rhomberg, L. R. (2000). Dose-response analyses of the carcinogenic effects of trichloroethylene in
      experimental animals.  Environ Health Perspect 108 Suppl 2, 343-58.
   Rodenbeck, S. E.,  Sanderson, L. M., and Rene, A. (2000). Maternal exposure to trichloroethylene in
      drinking water and birth-weight outcomes. Arch Environ Health 55, 188-94.
   Ruden, C. (200la). Interpretations of primary carcinogenicity data in 29 trichloroethylene risk
      assessments. Toxicology 169, 209-25.
   Ruden, C. (200 Ib). The use  and evaluation of primary data in 29 trichloroethylene carcinogen risk
      assessments. Reg Toxicol Pharmacol 34,  3-16.
   Ruden, C. (2002a). Scrutinizing three trichloroethylene carcinogenicity classifications in the European
      Union—implications for the risk assessment process. Int J Toxicol 21, 441-50.

                                               C-67

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Ruden, C. (2002b). The use of mechanistic data and the handling of scientific uncertainty in carcinogen
      risk assessments. The trichloroethylene example. Reg ToxicolPharmacol 35, 80-94.
   Ruden, C. (2003). Science and transscience in carcinogen risk assessment—the European Union
      regulatory process for trichloroethylene. J Toxicol Environ Health B 6, 257-77'.
   Schraml, P., Zhaou, M., Richter, J., Bruning, T., Pommer, M., Sauter, G.,  Mihatsch, M. J., and Moch, H.
      (1999). Untersuchung von Nierentumoren bei Trichlorathylen-exponierten Arbeitern mittels
      Comparative Genomic Hybridization und DNA Sequenzierung. Verh Dtsch Ges Pathol 83, 218-24.
   Shih, C. L., Chen, H. H., and Chiu, T. H. (2001). Acute exposure to trichloroethylene differentially
      alters the susceptibility to chemoconvulsants in mice. Toxicology  162, 35-42.
   Simmons, J. E., Boyes, W. K., Bushnell, P. J., Raymer, J. H., Limsakun, T., McDonald, A., Sey, Y. M.,
      and Evans, M. V. (2002). A physiologically based pharmacokinetic model for trichloroethylene in
      the male long-evans rat. Toxicol Sci 69, 3-15.
   Simon, T. W. (1997). Combining physiologically based pharmacokinetic modeling with Monte Carlo
      simulation to derive an acute inhalation guidance value for trichloroethylene. Reg Toxicol
      Pharmacol 26, 257-70.
   Snawder, J. E., and Lipscomb, J. C. (2000). Interindividual variance of cytochrome P450 forms in
      human hepatic microsomes: correlation of individual  forms with xenobiotic metabolism and
      implications in risk assessment. Reg Toxicol Pharmacol 32, 200-9.
   Stewart, B. W. (2001). Trichloroethylene and cancer: a carcinogen on trial. MedJAust 174, 244-7.
   Tao, L., Yang,  S., Xie, M., Kramer, P. M., and Pereira, M. A. (2000). Effect of trichloroethylene and its
      metabolites, dichloroacetic acid and trichloroacetic acid,  on the methylation and expression of c-Jun
      and c-Myc  protooncogenes in mouse liver: prevention by methionine.  Toxicol Sci 54, 399-407.
   USEPA. (2001). Trichloroethylene Health Risk Assessment: Synthesis and Characterization.
      EPA/600/P-01/002A, August 2001, USEPA Office or Research and Development, Washington, DC
      20460.
   Walgren, J. E., Kurtz, D. T., and McMillan, J. M. (2000). The effect of the trichloroethylene metabolites
      trichloroacetate and dichloroacetate on peroxisome proliferation and DNA synthesis in cultured
      human hepatocytes. Cell Biol Toxicol 16, 257-73.
   Wartenberg, D., Reyner, D., and Scott, C. S.  (2000).  Trichloroethylene and cancer: epidemiologic
      evidence.[comment]. Environ Health Perspect 108 Suppl 2, 161-76.
   Waseem, M., Ali, M., Dogra, S., Dutta, K. K., and Kaw, J. L. (2001). Toxicity of trichloroethylene
      following inhalation and drinking contaminated water. JAppl Toxicol 21, 441-4.
                                                C-68

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
                                         Tetrachloroethylene
   9.0    Introduction

   Tetrachloroethylene  is  made  by  direct  chlorination or  oxychlorination  of certain  hydrocarbons.
   Tetrachloroethylene  is used as  a chemical intermediate,  as solvent for  metal cleaning and  vapor
   degreasing, and for dry-cleaning and textile processing. (Aggazzotti et al.  1994) It is found in many
   household products, including  paint removers,  water repellents,  silicone  lubricants,  spot  removers,
   adhesives, and wood cleaners (ATSDR 1997).

   9.1    Toxic effects

   Liver, kidney, blood, and the central nervous system are the target organs for systemic effects (Calabrese
   1983; Chen et al. 2002; Echeverria et al.  1995; Ferroni et  al. 1992; Umezu et al. 1997; Utzinger and
   Schlatter 1977; Zavon 1967). Exposure to  high concentrations of tetrachloroethylene induces dizziness,
   headache, sleepiness, confusion, nausea, unconsciousness, and death.  Irritation could occur when skin is
   exposed to tetrachloroethylene.  Breathing the vapor may  irritate  the lungs, causing coughing and/or
   shortness of breath (Stewart et al. 1961).   Animal studies  showed that tetrachloroethylene  can cause
   liver and kidney damage (Schimmelpfennig et al. 1987; Kylin et al.  1963; Kylin et al. 1965; Lash et al.
   2002).   The  developing fetus and children may  be particularly  susceptible  to the toxic  effects of
   tetrachloroethylene (Ahlborg 1990; Fredriksson et al.  1993;  Motohashi et al.  1993; Spector et al. 1999).
   Exposure to pregnant rodents induces behavioral deficits in pups (Mattsson et al. 1998; Seeber 1989).

   The neurotoxicities of tetrachloroethylene  may result from  the alterations of fatty acid patterns  in the
   brain (ATSDR 1997; Burger et al.  1991).  In contrast to the nervous system, the effects on the liver
   including cancer are thought to be a result of the metabolite, trichloroacetic acid (ATSDR 1997). It is
   believed that trichloroacetic acid may play a role in  inducing hepatocellular peroxisomes, resulting in
   the production of hydrogen peroxide as a by-product (Bentley et al. 1993).  The increased hydrogen
   peroxide may increase DNA damage.  Kidney cancer may in part be a result of the formation  of the
   genotoxic metabolites from S-(l,2,2-trichlorovinyl) glutathione by p-lyase (Birner et al. 1997; Cooper et
   al. 2002; Green  et  al. 1990).  Tetrachloroethylene is classified as a group 2A carcinogen  (probably
   carcinogenic  to human) (Aschengrau et al. 1993; Aschengrau et al. 1998; Aschengrau et al.  2003;
   Wartenberg et al. 2000).

   9.2    Pharmacokinetics

   Tetrachloroethylene is readily absorbed through oral, skin, and inhalation exposure (Ward et al. 1988).
   Once  it  is absorbed, tetrachloroethylene  is distributed to  fatty  tissues because  of high lipophilicity
   (fat/blood partition coefficient is about 140) (Dallas et al. 1994c).  The half-life of tetrachloroethylene in
   fat tissues is 55  hours (ATSDR  1997).   One  to three percent  of absorbed tetrachloroethylene is
   metabolized  to trichloroacetic acid in the  liver (ACGIH  1991).  Unmetabolized tetrachloroethylene is
   exhaled  (ATSDR 1997).  This is the primary route of excretion.  Trichloroacetic acid is excreted in the
   urine (ATSDR 1997).

   9.2.1  Absorption
                                                C-69

-------
USEPA Contract No. 3C-R102-NTEX
                                                    Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
   Tetrachloroethylene is readily absorbed in the G.I. tract and lungs.  Pulmonary uptake is proportional to
   ventilation rate, duration of exposure, and the concentration in the inspired air (ATSDR 1997).  In rats,
   the proportion absorbed was approximately 55-70% after 1 minute, gradually declining to 40-50% after
   2 hours (Dallas et al. 1994b).  Dermal absorption has been studied in guinea pigs (Bogen et al.  1992).

   9.2.2   Distribution

   Tetrachloroethylene is preferentially stored in fat tissues.  In rats, distribution to brain, liver, and kidneys
   has also been demonstrated (Frantz and Watanabe 1983; Dallas et al.  1994a; Dallas et al.  1994b).  In
   animal  studies, transplacental and  lactational  transport  of unchanged  tetrachloroethylene  has been
   reported (Byczkowski et al. 1994; Hamada and Tanaka 1995).

   9.2.3   Metabolism

   The  metabolic  pathways  of  tetrachloroethylene are  summarized in Figure  9.1.   The overlapping
   pathways with  the other  three  volatile organics  in Mixture 2  can be seen  in Figure  8.1   under
   trichl oroethy 1 ene.
  CI     ^Cl

  ci'    NCI
  Tetrachloroethylene
                                               -+- "y-v01
                           GSH, GSH-transferase
                            a
                                     NADPH, O,
                                      P4SO          CI ^       X Ci
                                                Teiraohloroethylene oxide
                                                             w'cfe hydrase
                 _CI    SG

                 •ff'Xj
                 CI
  y/.

5=<

   I
                            llH2
                            i1"?
                _CI     S -CH2-9H
                            COOH
                       N-acetyl
                       transferase
                            _c
                            H3
 pi
              ,
             Ci    S-CHjTpH
                       COOH _J
I
               Urine
              Interactions with
              proteins, DNA
           CI   ,01
           cf   ^si-
           ^
            H-C-C+=S
                       NH3   Ammonia

                        OOH
                       CH,
                            Acetic acid
                                        I  epoxid
                                                     ci    CI
        Tetrachloroethylene glycol

               J^  +2HC/
   Oxalic acid  O^
   dichloride
            CI   Cl
                            o   p
                                        Oxalic acid
                                            CI     OH
                                                NOH
Glyoxylic
acid
chloride
                                                  DeG®rboxyl&tion
                                          CO,  +  HC
                                                     OH
                                       Carbon dioxide   Formic acid
                                                              Fig 9.1 Metabolic pathways
                                                                 of tetrachloroethylene
                                                                    (ATSDR, 1997)
   Human pharmacokinetic studies have been performed in volunteers and workers. The pharmacokinetics
   of tetrachloroethylene by inhalation  exposure has been described (Ikeda  1977; Monster et al.  1979;
                                                   C-70

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Ohtsuki  et  al.  1983;  Imbriani  et al.   1988).  One  study  described  the  pharmacokinetics   of
   tetrachloroethylene in a boy who ingested the chemical (Koppel et al. 1985).

   The pharmacokinetics of tetrachloroethylene following inhalation exposure  have been described  for
   rodents (Pegg et al.  1978;  Schumann  and  Watanabe 1979).  The dermal  pharmacokinetics   of
   tetrachloroethylene in hairless guinea pigs was also studied (Bogen et al. 1992).  Pharmacokinetics of
   tetrachloroethylene following oral exposure were  reported in several  studies  including rats, mice, and
   dogs (Frantz and Watanabe 1983; Dallas etal. 1994c).

   In vitro metabolic studies of tetrachloroethylene have been conducted using rat hepatic microsome and
   other subcellular systems (Huang et al. 2001; Costa and Ivanetich 1980; Reitz et al. 1996; Dekant et al.
   1998). Some studies focused on the interaction of tetrachloroethylene with rat hepatic microsomal P450
   enzymes (Hani oka et al. 1995a; Hani oka et al. 1995b; Hani oka et al. 1997).

   9.2.4  Excretion

   In humans and animals, the major part of the absorbed  amount is exhaled unchanged.  In humans,  80-
   100% of  the  amount was exhaled as  parent compound.  In rats, about 70% was exhaled  in same
   conditions (ATSDR 1997). Excretion of metabolites in urine is 2% of exposed dose with a half-life of
   75-80 hours (Ikeda et al.  1972; Imbriani et al.  1988; ATSDR 1997). In rats, elimination via maternal
   milk was high (Byczkowski et al. 1994; Byczkowski and Fisher 1995).

   9.3    Interactions with other chemicals

   The hepatic monooxygenase  system  is mainly responsible for oxidation of tetrachloroethylene.  Thus,
   chemicals  that  affect the monooxygenase  system  could  affect  the metabolism  and  toxicity   of
   tetrachloroethylene.   Two papers were  published dealing with  pharmacokinetic interactions  between
   tetrachloroethylene and other chlorinated contaminants (Dobrev et al.  2001,  2002).  Toxicological
   interactions between tetrachloroethylene and ethanol or  other chemicals were also reported  (Koizumi et
   al.  1982; Dobrov andPoluekto 1971;  Kobayashi etal.  1982; Seiji etal. 1989; Giovannini etal. 1992).

   9.4    PBPK models

   Several PBPK models for the disposition of tetrachloroethylene were presented in animals  and humans
   (Gelman etal. 1996; Haddad etal. 2000; Ward etal. 1988; Koizumi 1989; Bois etal. 1990; Gearhart et
   al.  1993; Dallas et al. 1994b; Dallas  et al.  1994c;  Wilson  and Knaak 1994;  Dallas et al.  1995; Reitz et
   al.  1996; Poet et al. 2000; Loizou 2001; Poet et al. 2002).  The majority of the available PBPK models
   are concerned with the carcinogenesis of tetrachloroethylene.  One model has been developed to predict
   brain  concentrations  following exposure  to tetrachloroethylene during  showering (Rao  and Brown
   1993).  PBPK models for the lactational transfer of tetrachloroethylene  through breast milk were
   developed to  estimate the risk of tetrachloroethylene  exposure to infants (Byczkowski et al. 1994;
   Byczkowski and Fisher 1995).

   9.5    Literature Cited

   ACGIH (1991). Documentation of the threshold limit values and biological exposure indices. American
      Conference of Governmental Industrial Hygienists, Cincinnati.
                                               C-71

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Aggazzotti, G., Fantuzzi, G., Righi, E., Predieri, G., Gobba, P.M., Paltrinieri, M., and Cavalleri, A.
      (1994). Occupational and environmental exposure to perchloroethylene (Pee) in dry cleaners and
      their family members. Arch Environ Health 49, 487-493.
   Ahlborg, G. (1990). Pregnancy outcomes among women working in laundries and dry- cleaning shops
      using tetrachloroethylene. Am JIndMed 17, 567-575.
   Aschengrau, A., Ozonoff, D., Paulu, C., Coogan, P., Vezina, R., Heeren, T., and Zhang, Y. Q. (1993).
      Cancer risk and tetrachloroethylene-contaminated drinking-water in Massachusetts. Arch Environ
      Health 48, 284-292.
   Aschengrau, A., Paulu, C., and Ozonoff, D. (1998). Tetrachloroethylene-contaminated drinking water
      and the risk of breast cancer. Environ Health Perspect 106, 947-953.
   Aschengrau, A., Rogers, S., and Ozonoff, D. (2003). Perchloroethylene-contaminated drinking water
      and the risk of breast cancer: Additional results from Cape Cod, Massachusetts, USA. Environ
      Health Perspect 111,167-173.
   ATSDR (1997). Toxicological profile for tetrachloroethylene. US Department of Health and Human
      Services.
   Bentley P, Calder I, and C, E. (1993). Hepatic peroxisome proliferation in rodents and its significance
      for humans. Food Chem Toxicol 31, 857-907.
   Birner, G., Bernauer, U., Werner, M., and Dekant, W. (1997). Biotransformation,  excretion and
      nephrotoxicity of haloalkene- derived cysteine S-conjugates. Arch Toxicol72, 1-8.
   Bogen, K. T., Colston, B. W., and Machicao, L.  K. (1992). Dermal absorption of dilute aqueous
      chloroform, trichloroethylene, and tetrachloroethylene in hairless guinea- pigs. Fundam Appl Toxicol
      18, 30-39.
   Bois, F. Y., Zeise, L.,  and Tozer, T. N. (1990). Precision and sensitivity of pharmacokinetic models for
      cancer risk assessment - tetrachloroethylene  in mice, rats, and humans. Toxicol Appl Pharmacol 102,
      300-315.
   Burger, A., Gehrisch,  S., Jaross, W., Dietz, E., Sucker, M. L., and Gutewort, T.  (1991). Investigations on
      the effect of tetrachloroethylene on the  lipoprotein metabolism of man by taking into account the
      alcohol uptake. Zeitschrift Fur Klinische Medizin-Zkm 46, 671-674.
   Byczkowski, J. Z., and Fisher,  J. W. (1995). A computer-program linking physiologically-based
      pharmacokinetic model with cancer risk assessment for breast- fed infants. ComputMeth Programs
      Biomed46, 155-163.
   Byczkowski, J. Z., Kinkead, E. R., Leahy, H. F., Randall, G. M.,  and Fisher, J. W. (1994). Computer-
      simulation of the lactational transfer of tetrachloroethylene in rats using a physiologically-based
      model. Toxicol Appl Pharmacol 125, 228-236.
   Calabrese, E. J. (1983).  Tetrachloroethylene in drinking water - a toxicologists perspective - discussion.
      J Am Water WorkAssoc 75, 190-190.
   Chen, H. H., Chan, M. H., and Fu, S. H. (2002).  Behavioural effects of tetrachloroethylene exposure in
      rats: acute and subchronic studies. Toxicology 170, 201-209.
   Cooper, A. J. L., Bruschi, S. A., and Anders, M. W. (2002). Toxic, halogenated cysteine S-conjugates
      and targeting of mitochondrial enzymes of energy metabolism. Biochem Pharmacol 64, 553-564.
   Costa, A. K., and Ivanetich, K. M. (1980). Tetrachloroethylene metabolism by the hepatic-microsomal
      cytochrome-P-450 system. Biochem Pharmacol 29, 2863-2869.
   Dallas, C. E., Chen, X. M., Muralidhara, S., Varkonyi, P.,  Tackett, R. L., and Bruckner, J. V. (1994a).
      Use of tissue disposition data from rats and dogs to determine species-differences in input
      parameters for a physiological model for perchloroethylene. Environ Res 67, 54-67.
   Dallas, C. E., Chen, X. M., Muralidhara, S., Varkonyi, P.,  Tackett, R. L., and Bruckner, J. V. (1995).
      Physiologically-based pharmacokinetic model useful in prediction of the influence of species, dose,
      and exposure route on perchloroethylene pharmacokinetics. J Toxicol Environ Health 44, 301-317.
                                                C-72

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Dallas, C. E., Chen, X. M., Obarr, K., Muralidhara, S., Varkonyi, P., and Bruckner, J. V. (1994b).
      Development of a physiologically-based pharmacokinetic model for perchloroethylene using tissue
      concentration-time data. Toxicol Appl Pharmacol 128, 50-59.
   Dallas, C. E., Muralidhara, S., Chen, X. M., Ramanathan, R., Varkonyi, P., Gallo, J. M., and Bruckner,
      J. V. (1994c). Use of a physiologically-based model to predict systemic uptake and respiratory
      elimination of perchloroethylene. Toxicol Appl Pharmacol 128,  60-68.
   Dekant, W., Birner, G., Werner, M., and Parker, J. (1998). Glutathione conjugation of perchloroethene
      in subcellular fractions from rodent and human liver and kidney. Chem Biol Interact 116, 31-43.
   Dobrev, I. D., Andersen, M. E., and Yang, R. S. H. (2001). Assessing interaction thresholds for
      trichloroethylene in combination with tetrachloroethylene and 1,1,1-trichloroethane using gas uptake
      studies and PBPK modeling. Arch Toxicol 75, 134-144.
   Dobrev, I. D., Andersen, M. E., and Yang, R. S. H. (2002). In silico toxicology: simulating interaction
      thresholds for human exposure to mixtures of trichloroethylene,  tetrachloroethylene, and 1,1,1-
      trichloroethane. Environ Health Perspect 110, 1031-1039.
   Dobrov, I. V., and Poluekto, V.A. (1971). Inhibition of tetrachloroethylene oxidation by benzene.
      Doklady Akademii Nauk Sssr 200, 367-&.
   Echeverria, D., White, R. F., and Sampaio, C. (1995). A behavioral-evaluation of PCE exposure in
      patients and dry cleaners - a possible relationship between clinical  and preclinical effects. J Occup
      EnvironMed37, 667-680.
   Ferroni, C., Selis, L., Mutti, A., Folli, D., Bergamaschi, E., and Franchini, I. (1992). Neurobehavioral
      and neuroendocrine effects of occupational exposure to perchloroethylene. Neurotoxicology 13, 243-
      247.
   Frantz, S. W., and Watanabe, P. G. (1983). Tetrachloroethylene - balance and tissue distribution in male
      sprague-dawley rats by drinking-water administration. Toxicol Appl Pharmacol 69,  66-72.
   Fredriksson, A., Danielsson, B. R. G., and Eriksson, P. (1993). Altered behavior in adult mice orally
      exposed to trichloroethylene and tetrachloroethylene  as neonates. Toxicol Lett 66, 13-19.
   Gearhart, J. M., Mahle, D. A., Greene, R. J., Seckel, C. S., Flemming,  C. D., Fisher, J. W., and Clewell,
      H. J. (1993). Variability of physiologically-based pharmacokinetic (PBPK) model parameters and
      their effects on PBPK model predictions in a risk assessment for perchloroethylene (PCE). Toxicol
      Lett 68, 131-144.
   Gelman, A., Bois, F., and Jiang, J. M. (1996). Physiological pharmacokinetic analysis using population
      modeling and informative prior distributions. J Am Stat Assoc 91, 1400-1412.
   Giovannini, L., Guglielmi, G., Casini, T., Bertelli, A., Galmozzi, E., and Bertelli, A. A.  E. (1992). Effect
      of ethanol chronic use on hepatotoxicity in  rats exposed to tetrachloroethylene. IntJ Tissue React
      Exp ClinAsp 14, 281-285.
   Green T, Odum J, and Nash J (1990). Perchloroehtylene-induced rat kidney tumors: an investigation of
      the mechanisms involved and their relevance to humans. Toxicol Appl Pharmacol 103, 77-89.
   Haddad, S., Charest-Tardif, G., and Krishnan, K. (2000). Physiologically based modeling of the
      maximal effect of metabolic interactions on the kinetics of components of complex chemical
      mixtures. J Toxicol Environ Health A 61, 209-223.
   Hamada, T., and Tanaka, H. (1995). Transfer of methyl chloroform, trichloroethylene and
      tetrachloroethylene to milk, tissues and expired air following intraruminal or oral-administration in
      lactating goats and milk-fed kids. Environ Pollut 87,  313-318.
   Hanioka, N., Jinno, H., Takahashi, A., Nakano, K., Yoda, R., Nishimura,  T., and Ando, M. (1995a).
      Interaction of tetrachloroethylene with rat hepatic-microsomal P450-dependent monooxygenases.
      Xenobiotica 25,151-165.
   Hanioka, N., Jinno, H., Toyooka, T., Nishimura, T.,  and Ando, M. (1995b). Induction of rat-liver drug-
      metabolizing-enzymes by tetrachloroethylene. Arch Environ Contain Toxicol28, 273-280.
                                                C-73

-------
USEPA Contract No. 3C-R102-NTEX                           Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Hanioka, N., Omae, E., Yoda, R., Jinno, H., Nishimura, T., and Ando, M. (1997). Effect of
      trichloroethylene on cytochrome P450 enzymes in the rat liver. Bull Environ Contain Toxicol 58,
      628-635.
   Huang, R. N., Wang, J. L., Chen, W. L., Tsai, S. Y., and Sung, P. Y. (2001). Toxicokinetics of
      trichloroethylene and tetrachloroethylene in cultural medium and their toxicity to CHO-K1 cells.
      Toxicology 164, 161-161.
   Ikeda, M. (1977). Metabolism of trichloroethylene  and tetrachloroethylene in human subjects. Environ
      Health Perspect21, 239-245.
   Ikeda, M., Imamura, T., Ohtsuji, H., and Komoike, Y. (1972). Urinary-excretion of total trichloro-
      compounds, trichloroethanol, and trichloroacetic acid as a measure of exposure to trichloroethylene
      and tetrachloroethylene. BritJIndMed29, 32S-&.
   Imbriani, M., Ghittori, S., Pezzagno, G., and Capodaglio, E. (1988). Urinary excretion of
      tetrachloroethylene (perchloroethylene) in experimental and occupational exposure. Arch Environ
      Health 43, 292-298.
   Kobayashi, S., Hutcheon, D. E.,  and Regan, J. (1982). Cardiopulmonary toxicity of tetrachloroethylene.
      J Toxicol Environ Health 10, 23-30.
   Koizumi, A.  (1989). Potential of physiologically based pharmacokinetics to amalgamate kinetic data of
      trichloroethylene and tetrachloroethylene obtained in rats and man. BritJIndMed46., 239-249.
   Koizumi, A., Kumai, M., and Ikeda, M. (1982). In vivo suppression of 1,1,1-trichloroethane metabolism
      by co- administered tetrachloroethylene - an inhalation  study. Bull Environ Contam Toxicol 29, 196-
      199.
   Koppel, C., Arndt, I, Arendt, U., and Koeppe, P. (1985). Acute tetrachloroethylene poisoning - blood
      elimination kinetics during hyperventilation therapy. J Toxicol Clin Toxicol 23, 103-115.
   Kylin, B., Sumegi, L, Reichard, H., and Yllner, S. (1963). Hepatotoxicity of inhaled trichloroethylene,
      tetrachloroethylene and chloroform - single exposure. Act Pharmacol Toxicol 20, 16-&.
   Kylin, B., Sumegi, L, and Yllner, S. (1965). Hepatotoxicity of inhaled trichloroethylene and
      tetrachloroethylene . Long-term exposure. Act Pharmacol Toxicol22, 379-&.
   Lash, L. H., Qian, W., Putt, D. A., Hueni, S. E., Elfarra, A. A., Sicuri, A. R., and Parker, J. C. (2002).
      Renal toxicity of perchloroethylene and S-(l,2,2- trichlorovinyl) glutathione in rats and mice: sex-
      and species- dependent differences. Toxicol ApplPharmacol 179,  163-171.
   Loizou, G. D. (2001). The application of physiologically based pharmacokinetic modelling in the
      analysis of occupational exposure to perchloroethylene. Toxicol Lett 124,  59-69.
   Mattsson, J. L., Albee, R. R., Yano, B. L., Bradley, G. J., and Spencer, P. J. (1998). Neurotoxicologic
      examination of rats exposed to 1,1,2,2- tetrachloroethylene (Perchloroethylene) vapor for 13  weeks.
      Neurotoxicol Teratol 20, 83-98.
   Monster, A. C., Boersma, G., and Steenweg, H. (1979). Kinetics of tetrachloroethylene in volunteers -
      influence of exposure concentration and work load. Int Arch Occup Environ Health 42, 303-309.
   Motohashi, Y., Miyazaki, Y., and Takano, T. (1993). Assessment of behavioral effects of
      tetrachloroethylene using a set of time series analyses. Neurotoxicol Teratol 15, 3-10.
   Ohtsuki, T., Sato, K., Koizumi, A., Kumai, M., and Ikeda, M. (1983). Limited capacity of humans  to
      metabolize tetrachloroethylene. Int Arch Occup Environ Health 51, 381-390.
   Pegg, D. G.,  Zempel,  J. A., Braun, W. H., and Gehring, P. J. (1978). Disposition of tetrachloroethylene-
      C-14 following oral and inhalation exposure in  rats. Toxicol Appl Pharmacol 45, 276-277'.
   Poet, T. S., Corley, R. A., Thrall, K. D., Edwards, J. A., Tanojo, H., Weitz, K. K., Hui, X. Y., Maibach,
      H. L, and Wester, R. C. (2000).  Assessment of the percutaneous absorption  of trichloroethylene in
      rats and humans using MS/MS real-time breath analysis and physiologically based pharmacokinetic
      modeling. Toxicol Sci 56, 61-72.
                                                C-74

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Poet, T. S., Weitz, K. K., Gies, R. A., Edwards, J. A., Thrall, K. D., Corley, R. A., Tanojo, H., Hui, X.
      Y., Maibach, H. I, and Wester, R. C. (2002). PBPK modeling of the percutaneous absorption of
      perchloroethylene from a soil matrix in rats and humans. Toxicol Sci 67, 17-31.
   Rao H, and Brown DR (1993). A physiologically based pharmacokinetic assessment of
      tetrachloroethylene in ground water for a bathing and showering determination. Risk Anal 13, 37-49.
   Reitz, R. H., Gargas, M. L., Mendrala, A. L., and Schumann, A. M. (1996). In vivo and in vitro studies
      of perchloroethylene metabolism for physiologically based pharmacokinetic modeling in rats, mice,
      and humans. Toxicol ApplPharmacol 136, 289-306.
   Schimmelpfennig, W., Lun, A., Gutewort, T., Dietz, E., and Pannier, R. (1987).  Investigations of
      hepatotoxicity by tetrachloroethylene. Zeitschrift Fur Klinische Medizin-Zkm 42, 2241-2243.
   Schumann, A. M., and Watanabe, P. G. (1979). Species differences between rats and mice on the
      metabolism and hepatic macromolecular binding of tetrachloroethylene. Toxicol Appl Pharmacol 48,
      A89-A89.
   Seeber, A. (1989). Neuro-behavioral toxicity of long-term exposure to tetrachloroethylene. Neurotoxicol
      Teratoll 1,579-583.
   Seiji, K., Inoue,  O., Jin, C., Liu, Y. T., Cai, S. X., Ohashi, M., Watanabe, T., Nakatsuka, H., Kawai, T.,
      and Ikeda, M. (1989). Dose excretion relationship in tetrachloroethylene-exposed workers and the
      effect of tetrachloroethylene co-exposure on trichloroethylene metabolism. Am JIndMed 16, 675-
      684.
   Spector, J., Lewandowski, A. G., Mott, J. A., and Schreiber, J. S. (1999). Neuropsychological and
      behavioral functioning in tetrachloroethylene-exposed pre-school children and controls. Arch Clin
      Neuropsychol 14, 661-662.
   Stewart, R. D., Hake, C. L., Erley, D. S., Gay, H. H., and Schaffer, A. W. (1961). Human exposure to
      tetrachloroethylene vapor - relationship of expired air and blood-concentrations to exposure and
      toxicity. Arch Environ Health 2, 516-&.
   Umezu, T., Yonemoto, J., Soma, Y., and Miura, T. (1997). Behavioral effects of trichloroethylene and
      tetrachloroethylene in mice. PharmacolBiochem Behav 58, 665-671.
   Utzinger, R., and Schlatter, C. (1977). Review on toxicity of trace amounts of tetrachloroethylene in
      water. Chemosphere 6, 517-524.
   Ward, R. C., Travis, C. C., Hetrick, D. M., Andersen, M. E., and Gargas, M. L. (1988).
      Pharmacokinetics of tetrachloroethylene. Toxicol Appl Pharmacol 93, 108-117.
   Wartenberg, D., Reyner, D., and Scott, C. S. (2000). Trichloroethylene and cancer: Epidemiologic
      evidence. Environ Health Perspect 108, 161-176.
   Wilson, J. D., and Knaak, J. (1994). Health risk assessment of tetrachloroethylene using a
      physiologically-based pharmacokinetic model. Abstr Pap Am Chem Soc 207, 62-ENVR.
   Zavon, M. R. (1967). Liver disease from tetrachloroethylene. JAMA 199, 135-&.
                                                C-75

-------
USEPA Contract No. 3C-R102-NTEX                         Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.


                                     1, 1, 1-Trichloroethane

   10.0   Introduction

   1, 1, 1-Trichloroethane (TCA; Figure 10.1) is a common organic solvent, often used commercially for
   industrial degreasing as well as dry-cleaning.  The 2001 Comprehensive Environmental Response,
   Compensation, and Liability Act (CERCLA) Priority List of Hazardous Substances includes TCA in the
   top 100 hazardous substances based upon its environmental distribution, especially at hazardous waste
   sites (ATSDR, 200la). Furthermore, TCA ranks 13th in the CERCLA Completed Exposure Pathway;
   therefore, humans are frequently exposed to TCA (ATSDR, 200 Ib). Because of its ability to induce
   central nervous system depression, TCA has been abused, and thus purposeful human exposure also
   occurs.  TCA is considered to be a group III carcinogen due to lack of adequate evidence of
   carcinogenicity in rodents and humans (IARC, 1999).
  Figure 10.1. Structure of 1, 1, 1-trichloroethane.

  10.1   Toxic effects

  TCA  has various  systemic effects,  most notably  central  nervous  system  (CNS)  depression,
  hepatotoxicity, and cardiovascular complications. Central nervous system depression is the principal
  CNS effect observed in individuals and animals following exposure to TCA (Hall and Hine, 1966; Stahl
  et al.,  1969; Jones and Winter, 1983; Bowen and Balster,  1998; Bowen et al., 1998; Bruckner et al.,
  2001). Descriptions of plausible mode(s)  of action for CNS depression are given  in numerous reports
  (Rosengren  et al,  1985; Nilsson, 1986b;  Nilsson,  1986a; Nilsson,  1987; Fernicola et  al,  1991;
  Beckstead et al, 2000; Warren et al, 2000; You and Dallas, 2000; Beckstead et al, 2001; Okuda et al,
  2001; Beckstead etal, 2002; Wiley etal,  2002; Lopreato etal, 2003).

  Various reports cite changes in serum enzyme chemistry, which serve as indicators  of hepatotoxicity for
  both humans and animals (Halevy et al, 1980; Hodgson et al, 1989). Another marker for hepatotoxicity
  observed following exposure to TCA is accumulation of fat in the liver (Hall and Hine, 1966; Caplan et
  al, 1976; Hodgson et al, 1989). However,  many  contradictory studies on both humans and animals
  report failure of  serum enzymes  to change or  extremely  mild changes,  indicating  no apparent
  hepatotoxic effects (Domette and Jones, 1960; Carlson, 1973; Kramer etal, 1978; Kelafant etal, 1994;
  Wang et al, 1996).  Although observed hepatic  alterations are  reversible, they tend to indicate  mild
  hepatotoxicity induced by TCA and/or a metabolite (Halevy et al, 1980; Bruckner et al, 2001). Cardiac
  sensitization to epinephrine,  resulting in  arrhythmia,  has been linked with exposure to TCA in  both
  humans and animals (Clark and  Tinston,   1973;  Guberan et al, 1976; Macdougall et al, 1987).
  Additionally, cardiac depression, resulting in decreased blood pressure, is caused by exposure to TCA.
  Toraason  and  coworkers demonstrated decreases  in contractility  of cultured cardiac cells occurred in a
  dose-dependent manner following treatment with TCA (Toraason et al,  1990).  Some  reproductive
  effects have also been reported, ranging  from increased mammary  adenocarcinomas to  decreases in

                                               C-76

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   sperm motility,  however effects were usually slight and in some cases confounded by exposure to
   chemical mixtures (Rudolph and  Swan, 1986; Swan et al., 1989; Yang,  1993; Coleman et al.,  1999;
   Lemasters et al., 1999; NTP, 2000; Wang et al., 2002). Other studies found no association between TCA
   exposure and reproductive effects (George et al., 1989; Wrensch etal., 1990a; Wrensch et al., 1990b)

   10.2   Pharmacokinetics

   10.2.1 Absorption

   Exposure to TCA primarily  occurs through inhalation, and has been described  in both humans  and
   animals (Morgan et al., 1972a; Morgan et al.,  1972b; Monster et al., 1979; Hobara  et al.,  1982;
   Jakobson et al.,  1982; Hobara et al., 1983; Koizumi et al., 1983; Nolan et al., 1984; Dallas et al.,  1989;
   Boman et  al., 1995). Dermal and gastrointestinal exposures are plausible as TCA is a groundwater
   contaminant, though  due to the volatility  of TCA, the most common exposure route is inhalation
   (ATSDR, 1995). Alternative routes of TCA exposure have been explored by many researchers, as TCA
   is efficiently and rapidly absorbed via the lung, skin, and gastrointestinal tract of humans and animals
   (Stewart and Dodd,  1964; Riihimaki and Pfaffli,  1978; Mitoma et al,  1985; RTI,  1987; Reitz  et al,
   1988; Morgan et al,  1991; Yoshida et al,  1998; Giardino et al, 1999;  Kezic et al, 2000; Poet  et al,
   2000;  Kezic et  al,  2001).  Steady-state blood levels in rats exposed to 50 or 500  ppm TCA were
   approached at 2 hours  following  initiation of continuous exposure  (Dallas et al,  1989).  Reitz  and
   colleagues  noted  achievement of maximal  blood levels of TCA  at 10-15  minutes  following
   administration of a 14.2 mg/kg dose of TCA in water via gavage  (Reitz et al, 1988). Following the
   initial  phases, absorption rates plateau as steady-state levels are approached in blood  and tissues;
   generally, blood levels approach steady-state within a few hours following onset of exposure (Monster
   etal, 1979; Nolan etal, 1984).

   10.2.2 Distribution

   TCA is widely  distributed,  with preferential  distribution to fatty tissues due to  its lipophilic nature,
   regardless of exposure scenario (Takahara, 1986; RTI, 1987;  Shimada, 1988; Katagiri etal, 1997; You
   and Dallas, 1998). Detectable levels of TCA are found in the fat, liver, kidney, spleen, blood, lung,
   heart, brain, placenta, and fetus following inhalation exposure (Danielsson et al, 1986; Takahara,  1986;
   Shimada,  1988). In mice  exposed for 1  hour  to  1,000  ppm TCA, tissue concentrations of TCA
   immediately following exposure resulted in preferential accumulation of TCA (in descending order) in
   the fat, liver, kidney,  spleen and blood, followed by lung, heart and brain (Takahara, 1986). Schumann
   and coworkers supported these findings, as they  reported  significantly  higher TCA concentrations in
   fatty tissues than in the liver and kidneys following exposure of mice and rats to either 150 or 1,500 ppm
   TCA for  6 hours (Schumann  et al,  1982b). Distribution of TCA  is  regulated by various factors,
   including tissue  blood flow rate, tissue volume and tissue:blood partition coefficients, the latter  likely
   being most influential (ATSDR, 1995).

   10.2.3 Metabolism

   Metabolism of TCA has been studied extensively (Carlson, 1973; Ivanetich and Van den Honert,  1981;
   Casciola and Ivanetich,  1984; Takano et al, 1985; Takano et al, 1988; Kawai et al, 1991; Durk  et al,
   1992; Baker and Ronnenberg,  1993).  Regardless  of exposure route, TCA is metabolized at low rates
   (<10%), mainly  to four metabolites: trichloroethanol, trichloroethanol glucuronide, trichloroacetic acid,
   and carbon dioxide (Monster, 1979; Schumann et al, 1982a; Nolan et al, 1984;  Mitoma et al,  1985;
   Reitz etal, 1988; Dallas etal, 1989; Kawai etal, 1991); Figure 10.2). Oxidative metabolism of TCA

                                               C-77

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   by the cytochrome P-450 mixed-function oxidase system combined with other metabolic dehydrogenase
   enzymes forms trichloroethanol and trichloroacetic acid; trichloroethanol may be further metabolized
   via conjugation to form a glucuronide derivative. Cytochrome P450 2E1, specifically, is believed to play
   a  role in TCA metabolism (Nakajima and Sato, 1979; Guengerich et al., 1991; Kaneko et al.,  1994).
   Monster and coworkers found in humans exposed to 70 or 145 ppm of TCA for 4 hours, trichloroethanol
   and trichloroacetic acid excreted in the urine only accounted for 2 and 0.5%, respectively, of absorbed
   TCA (Monster et al., 1979). Another byproduct, acetylene, may also be formed from TCA in mammals
   via reductive dechlorination,  though only under hypoxic conditions (Durk et al., 1992). The metabolic
   pathway of TCA is shown in Figure 10.2, as well as in Figure 8.1 where interactive reaction network
   with the other three volatile organics is evident.
   1,1,1 trichloroethane     1,1,1 trichloroethanol   1,1,1 trichloroacetic acid
                                                               0
                                                                             •*•     CO,
                                                    glucuronide conjugates
             Figure 10.2. Metabolism of 1, 1, 1-Trichloroethane, reproduced from Agency for Toxic Substances
                           and Disease Registry's toxicological profile (ATSDR, 1995).
   10.2.4 Elimination

   The primary route of TCA elimination is exhalation of the parent compound, which occurs fairly rapidly
   following exposure due to TCA's highly volatile nature (Monster, 1979; Monster et al., 1979; Hobara et
   al., 1982; Schumann et al., 1982c; Schumann et al., 1982a; Schumann et al., 1982b; Nolan etal, 1984).
   In humans exposed to 35 or 350 ppm for 6 hours, more than 91% of TCA absorbed was eliminated,
   unchanged, in exhaled air (Nolan et al, 1984). Similarly, in animals given 20 daily doses of TCA by
   gavage in vegetable oil followed by a single 14C-labeled bolus,  85.1 and 92.3% of TCA was excreted as
   parent compound via exhalation, from rats and mice, respectively,  (Mitoma et al.,  1985).    Both
   acetylene and carbon dioxide are excreted in expired air (Durk et al., 1992; ATSDR, 1995). The other
   major metabolites, trichloroethanol, trichloroethanol glucuronide, and trichloroacetic  acid are mainly
   eliminated via urinary excretion, though fecal excretion has also been observed (Caperos et al.,  1982;
   Mitoma etal., 1985; Ghittori etal., 1987; Imbriani etal., 1988; Kawai etal., 1991).

   10.2.5 Species variations

   Because physiologically based  pharmacokinetic  (PBPK) modeling attempts to extrapolate  between
   various species for risk assessment purposes, variations among species can affect model precision and
   accuracy.  Most aspects of TCA pharmacokinetics are similar among species, including absorption and
   elimination route.  However,  quantitative  differences  in blood:air partition coefficients as well  as
   metabolism rates have been noted (Schumann et al, 1982b). Specifically, mice tend to have higher rates
   of TCA metabolism compared to rats and humans. Furthermore, blood:air partition coefficients, which

                                               C-78

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   dramatically effect inhalation  absorption differ: 2.53,  5.76,  and 10.8 for humans,  rats,  and mice
   respectively (Reitz etal., 1988).

   10.3   PBPK modeling

   Attempts to construct PBPK models appropriate for TCA's disposition have commonly used approaches
   similar to those developed in 1984 by Ramsey and Andersen (Ramsey and Andersen,  1984; Reitz et a/.,
   1988).  Based upon the Ramsey and Andersen model (RAM), a modified model was used to estimate
   metabolic kinetic constants using a closed, recirculated atmosphere representative of those used for gas
   uptake studies (Gargas et a/., 1986). This study found that to  adequately describe TCA disposition, its
   metabolism required  only  a first-order pathway,  which was abolished  when oxidative microsomal
   metabolism is inhibited. Reitz and colleagues utilized a model  similar to the RAM to  simulate exposure
   to  TCA  via  inhalation, intravenous  administration, bolus  gavage,  and  in  drinking water,  and
   demonstrated the plausibility of using  PBPK models in TCA risk assessment, based upon  successful
   interspecies extrapolation (Reitz etal., 1988).

   Bogen and Hall used a derivation of the RAM with  an additional compartment for skin  to assess risk
   associated with  TCA  in  drinking water,  and found  PBPK  modeling predicted  nontoxic  TCA
   concentrations  lower than the existing NOAELs (Bogen and Hall,  1989). Attempts  to  determine
   metabolic constants via PBPK  modeling concluded that, due to low metabolism of TCA, gas uptake
   study techniques were too insensitive to sufficiently form a TCA PBPK model (Gargas and  Andersen,
   1989). Absorption and elimination of TCA across time following an inhalation exposure was  measured,
   and a PBPK model was built to predict TCA levels in blood  and  expired air (Dallas et a/.,  1989). As
   TCA contaminates both water and soil, percutaneous absorption has been modeled in rats and humans,
   including simulations specific for exposure to children (Poet et a/., 2000). Notably, combination of
   quantitative structure-property relationships with traditional  PBPK modeling has successfully predicted
   inhalation pharmacokinetics for TCA, as well as other volatile organic chemicals (Beliveau etal., 2003).

   Within the context of  utilizing biological  monitoring to  assess  exposure, especially in a  work
   environment,  PBPK  models of TCA  have been  applied. Droz and coworkers first  developed  a
   population physiological model to  investigate  variability in  biological monitoring, then applied the
   model to assess how  alterations in components such as workload,  organ function, and body build
   affected the model's ability to accurately determine TCA exposure (Droz  et a/., 1989a,b). Comparison
   of various exposure scenarios on alterations in biological monitoring using PBPK modeling has been
   used  to determine which  biological  indices,  i.e.,  parent compound versus metabolite  in various
   biological  media,  are  best  suited to assess exposure to TCA (Lapare et a/., 1995). A linear four-
   compartment mass-balance model was used to not only assess  uptake and elimination of TCA in human
   subjects at  environmentally feasible levels,  but also predict  exhaled  TCA concentrations in another
   human  study  (Wallace et  a/.,  1997).  Analysis  of various PBPK models for a series  of chemicals,
   including TCA,  has allowed analysis of pharmacokinetic model output sensitivity to variability in both
   biochemical and metabolic input parameters (Hetrick et a/., 1991).

   Because exposure to a  single chemical compound in industrial or environmental exposure  settings is
   unlikely,  examination of  chemical in mixtures is necessary. Koizumi  and coworkers performed
   inhalation  studies  to investigate  co-exposure of TCA with perchloroethylene, and  found significant
   decreases in formation of TCA  metabolites due to co-exposure with perchloroethylene (Koizumi  et al.,
   1983). Further, Tardif and  Charest-Tardif noted decreases in  excretion of TCA metabolites following
   co-exposure with  w-xylene  (Tardif and Charest-Tardif, 1999).  Dobrev  and associates successfully
   modeled competitive  inhibition of trichloroethylene  by  TCA and tetrachloroethylene, likely due to a

                                                C-79

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   shared metabolic pathway with limited enzymatic capacity, specifically the cytochrome P450s (Dobrev
   et a/., 2001).  Further work by Dobrev and associates also used a combination of tetrachloroethylene,
   perchloroethylene,  and TCA to  assess possible interactions  which might change observed  toxicity
   (Dobrev et a/., 2002). The findings indicated that co-exposure to the three chlorinated hydrocarbons lead
   to a nonlinear increase in toxic conjugative metabolites of tetrachloroethylene (which are associated
   with  renal toxicity and/or  carcinogenicity), possibly indicating a greater than additive risk associated
   with  exposure to the chemical mixture. Although  metabolism of TCA is relatively  low (<10%), its
   ability to interact with essential metabolic enzymes may confer TCA the  ability to inhibit or decrease
   metabolism/detoxification of other chemicals, especially other organic solvents. Alternatively, because
   of comparatively low affinity for the cytochrome  P450s, formation of TCA metabolites (especially
   trichloroethanol and trichloroacetic acid)  may be reduced due to enzymatic inhibition caused  by co-
   exposure with other chemicals.
   10.4   Literature Cited

   ATSDR (1995). Toxicological profile for 1,1,1-trichloroethane. Agency for Toxic Substances and
         Disease Registry. U.S. Department of Health and Human Services, Public Health Service,
         Atlanta, GA.
   ATSDR (200la). 2001 CERCLA priority list of hazardous substances. Agency for Toxic Substances and
         Disease Registry. U.S. Department of Health and Human Services, Public Health Service,
         Atlanta, GA.
   ATSDR (200Ib). 2001 Substances most frequently found  in completed exposure pathways (CEPs) at
         hazardous waste sites. Agency for Toxic Substances and Disease Registry. U.S. Department of
         Health and Human Services, Public Health Service, Atlanta, GA.
   Baker, M. T., and Ronnenberg, W. C., Jr. (1993). Contrasting effects of 1,1,1-trichloroethane on
         [14C]vinyl chloride metabolism and activation in hepatic microsomes from phenobarbital- and
         isoniazid-treated rats. Toxicol ApplPharmacol 119, 17-22.
   Beckstead, M. J., Phelan, R., and Mihic, S. J. (2001). Antagonism of inhalant and volatile anesthetic
         enhancement of glycine receptor function. JBiol Chem 276, 24959-24964.
   Beckstead, M. J., Phelan, R., Trudell, J. R., Bianchini, M.  J., and Mihic, S. J. (2002). Anesthetic and
         ethanol effects on spontaneously opening glycine receptor channels. JNeurochem 82, 1343-
         1351.
   Beckstead, M. J., Weiner, J. L., Eger, E. I, 2nd, Gong, D.  H., and Mihic, S.  J. (2000). Glycine and
         gamma-aminobutyric acid(A) receptor function is enhanced by inhaled drugs of abuse. Mol
         Pharmacol 57, 1199-1205.
   Beliveau, M., Tardif, R., and Krishnan, K. (2003). Quantitative structure-property relationships for
         physiologically-based pharmacokinetic modeling of volatile organic chemicals in rats. Toxicol
         Appl Pharmacol 189, 221-232.
   Bogen, K. T., and Hall, L. C. (1989). Pharmacokinetics  for regulatory risk analysis: the case of 1,1,1-
         trichloroethane (methyl chloroform). Regul Toxicol Pharmacol  10, 26-50.
   Boman, A., Hagelthorn, G.,  and Magnusson, K. (1995).  Percutaneous absorption of organic solvents
         during intermittent exposure in guinea pigs. Ada Derm Venereal 75, 114-119.
   Bowen, S. E., and Balster, R. L. (1998).  A direct  comparison of inhalant effects on locomotor activity
         and  schedule-controlled behavior in mice. Exp Clin Psychopharmacol 6, 235-247.
   Bowen, S. E., Hamilton, J., and Balster,  R. L. (1998). A method for adjusting exposure levels of volatile
         solvents based on effects  on schedule-controlled behavior. Neurotoxicol Teratol 20, 169-180.
                                                C-80

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Bruckner, J. V., Kyle, G. M., Luthra, R., Acosta, D., Mehta, S. M., Sethuraman, S., and Muralidhara, S.
         (2001). Acute, short-term, and subchronic oral toxicity of 1,1,1- trichloroethane in rats. Toxicol
         Sci 60, 363-372.
   Caperos, J. R., Droz, P. O., Hake, C. L., Humbert, B. E., and Jacot-Guillarmod, A. (1982). 1,1,1-
         Trichloroethane exposure, biologic monitoring by breath and urine analyses. Int Arch Occup
         Environ Health 49, 293-303.
   Caplan, Y. H., Backer, R. C., and Whitaker, J. Q. (1976). 1,1,1-trichloroethane: report of a fatal
         intoxication. Clin Toxicol 9, 69-74.
   Carlson, G. P. (1973). Effect of phenobarbital and 3-methylcholanthrene pretreatment on the
         hepatotoxicity of 1,1,1-trichloroethane and 1,1,2-trichloroethane. Life Sci 13, 67-73.
   Casciola, L. A., and Ivanetich, K. M. (1984). Metabolism of chloroethanes by rat liver nuclear
         cytochrome P-450. Carcinogenesis 5,  543-548.
   Clark, D. G., and Tinston, D. J. (1973). Correlation of the cardiac sensitizing potential of halogenated
         hydrocarbons with their physicochemical properties. Br JPharmacol 49, 355-357.
   Coleman, C. N., Mason, T., Hooker, E. P., and Robinson, S. E. (1999). Developmental effects of
         intermittent prenatal exposure to 1,1,1- trichloroethane in the rat. Neurotoxicol Teratol2l, 699-
         708.
   Dallas, C. E., Ramanathan, R., Muralidhara, S., Gallo, J. M., and Bruckner, J. V. (1989). The uptake and
         elimination of 1,1,1-trichloroethane during and following inhalation exposures in rats. Toxicol
         ApplPharmacol 98, 385-397.
   Danielsson, B. R., Ghantous, H., and Dencker, L. (1986). Distribution of chloroform and methyl
         chloroform and their metabolites in pregnant mice. Biol Res Pregnancy Perinatol 7, 77-83.
   Dobrev, I. D., Andersen, M. E., and Yang, R.  S. (2001). Assessing interaction thresholds for
         trichloroethylene in combination with  tetrachloroethylene and 1,1,1-trichloroethane using gas
         uptake studies and PBPK modeling. Arch Toxicol 75, 134-44.
   Dobrev, I. D., Andersen, M. E., and Yang, R.  S. (2002). In silico toxicology: simulating interaction
         thresholds for human exposure to mixtures of trichloroethylene, tetrachloroethylene, and 1,1,1-
         trichloroethane. Environ Health Perspect 110, 1031-1039.
   Domette, W. H. L., and Jones, J. P. (1960).  Clinical experiences with 1,1,1-trichloroethane: A
         preliminary report of 50 anesthetic administrations. AnesthAnalg 39, 249-252.
   Droz, P.  O., Wu, M. M., and Cumberland, W. G. (1989a). Variability in biological monitoring of
         organic solvent exposure.  II. Application of a population physiological model. Br JIndMed 46,
         547-558.
   Droz, P.  O., Wu, M. M., Cumberland, W. G.,  and Berode, M. (1989b). Variability in biological
         monitoring of solvent exposure. I. Development of a population physiological model. Br JInd
         Med 46, 447-460.
   Durk, H., Poyer, J. L., Klessen, C., and Frank, H. (1992). Acetylene, a mammalian metabolite of 1,1,1-
         trichloroethane. BiochemJ286, 353-356.
   Fernicola, C., Govoni,  S., Coniglio, L., Daniele, E., and Trabucchi, M.  (1991). The use of 1,1,1-
         trichloroethane (methylchloroform) in industrial operations: the neurotoxicity risk. Med Lav 82,
         38-49.
   Gargas, M. L., and Andersen, M.  E. (1989). Determining kinetic constants of chlorinated ethane
         metabolism in the rat from rates of exhalation. Toxicol Appl Pharmacol 99, 344-353.
   Gargas, M. L., Andersen, M. E., and Clewell, H. J., 3rd (1986). A physiologically based simulation
         approach for determining metabolic constants from gas uptake data. Toxicol Appl Pharmacol 86,
         341-352.
   George, J. D., Price, C. J., Marr, M. C., Sadler, B. M., Schwetz, B. A., Birnbaum,  L. S., and Morrissey,
         R. E. (1989). Developmental toxicity of 1,1,1-trichloroethane in CD rats. FundamAppl Toxicol
         13,641-651.

                                                C-81

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Ghittori, S., Imbriani, M., Pezzagno, G., and Capodaglio, E. (1987). The urinary concentration of
         solvents as a biological indicator of exposure: proposal for the biological equivalent exposure
         limit for nine solvents. Am IndHygAssoc J 48, 786-790.
   Giardino, N. 1, Gordon, S. M., Brinkman, M.  C., Callahan, P. I, and Kenny, D. V. (1999). Real-time
         breath analysis of vapor phase uptake of 1,1,1 trichloroethane through the forearm: implications
         for daily absorbed dose of volatile organic compounds at work. Appl Occup Environ Hyg 14,
         719-727.
   Guberan, E., Fryc, O., and Robert, M. (1976).  Sudden death from ventricular fibrillation after voluntary
         inhalation of chlorothene in a mechanics apprentice. SchweizMed Wochemchr 106, 119-121.
   Guengerich, F. P., Kim, D. H., and Iwasaki, M. (1991). Role of human cytochrome P-450 IIE1 in the
         oxidation of many low molecular weight cancer suspects. Chem Res Toxicol 4, 168-179.
   Halevy, J., Pitlik, S., Rosenfeld, J., and Eitan,  B. D. (1980). 1,1,1,-Trichloroethane intoxication: a case
         report with transient liver and renal damage. Review of the literature.  Clin Toxicol 16, 467-472.
   Hall, F. B., and Hine, C. H. (1966). Trichloroethane intoxication: a report of two cases. J Forensic Sci
         11,404-413.
   Hetrick, D. M., Jarabek, A. M., and Travis, C. C. (1991). Sensitivity analysis for physiologically based
         pharmacokinetic models. JPharmacokinet Biopharm 19, 1-20.
   Hobara, T., Kobayashi, H., Higashihara, E., Iwamoto, S., Kawamoto, T., Sakai, T., and Tsubota, N.
         (1982). Experimental examinations and toxicokinetic analysis of the absorption and excretion of
         1,1,1-trichloroethane by the lung. Sangyo Igaku 24, 599-607.
   Hobara, T., Kobayashi, H., Higashihara, E., Kawamoto, T., and Sakai, T. (1983). Factors affecting
         1,1,1-trichloroethane absorption and excretion by the lung. Nippon Eiseigaku Zasshi 38, 642-
         648.
   Hodgson, M. J., Heyl, A. E., and Van Thiel, D. H. (1989). Liver disease associated with exposure to
         1,1,1-trichloroethane. Arch Intern Med 149, 1793-1798.
   IARC (1999). 1,1,1-Trichloroethane. IARCMonogr Eval CarcinogRisks Hum 71, 881-903.
   Imbriani, M., Ghittori, S., Pezzagno, G., Huang, J., and Capodaglio, E. (1988). 1,1,1-Trichloroethane
         (methyl chloroform) in urine as biological index of exposure. Am JIndMed 13, 211-222.
   Ivanetich, K. M., and Van den Honert, L. H. (1981).  Chloroethanes : their metabolism by hepatic
         cytochrome P-450 in vitro. Carcinogenesis 2, 697-702.
   Jakobson, L, Wahlberg, J. E., Holmberg, B., and Johansson, G. (1982). Uptake via the blood and
         elimination of 10 organic solvents following epicutaneous exposure of anesthetized guinea pigs.
         Toxicol Appl Pharmacol 63, 181-187.
   Jones, R.  D., and Winter, D. P. (1983).  Two case reports of deaths on industrial premises attributed to
         1,1,1- trichloroethane. Arch Environ Health 38, 59-61.
   Kaneko, T., Wang, P. Y., and Sato, A. (1994). Enzymes induced by ethanol differently affect the
         pharmacokinetics of trichloroethylene  and 1,1,1-trichloroethane. Occup Environ Med 51, 113-
         119.
   Katagiri, H., Aoki, N., Soma, K., Karube, H., Aizawa, Y., Kadowaki, T., and Inoue, Y. (1997).
         Concentration in blood and organs of dogs after high dose 1,1,1- trichloroethane inhalation. Ind
         Health 35,  461-466.
   Kawai, T., Yamaoka, K., Uchida, Y., and Ikeda, M. (1991). Exposure of 1,1,1-trichloroethane and dose-
         related excretion of metabolites in urine of printing workers. Toxicol Lett 55, 39-45.
   Kelafant, G. A., Berg, R. A., and Schleenbaker, R. (1994). Toxic encephalopathy due to 1,1,1-
         trichloroethane exposure. Am JInd Med'25, 439-446.
   Kezic, S., Monster, A. C., Kruse, J., and Verberk, M. M. (2000).  Skin absorption of some vaporous
         solvents in volunteers. IntArch  Occup Environ Health 73, 415-422.
   Kezic, S., Monster, A. C., van de Gevel, I. A., Kruse, J., Opdam, J. J., and Verberk, M. M. (2001).
         Dermal absorption of neat liquid solvents on brief exposures in volunteers. Aihaj 62,  12-18.

                                                C-82

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Koizumi, A., Kumai, M., and Ikeda, M. (1983). Dose-dependent induction and suppression of liver
         mixed-function oxidase system in chlorinated hydrocarbon solvent metabolism. Journal of
         Applied Toxicology 3, 208-217.
   Kramer, C. G., Imbus, H. R., Ott, M. G., Fulkerson, J. E., and Hicks, N. (1978). Health of workers
         exposed to 1, 1, 1,-trichloroethane: a matched-pair study. Arch Environ Health 33, 331-342.
   Lapare, S., Tardif, R., and Brodeur, J. (1995). Effect of various exposure scenarios on the biological
         monitoring of organic solvents in alveolar air. II. 1,1,1-Trichloroethane and trichloroethylene. Int
         Arch Occup Environ Health 67, 375-394.
   Lemasters, G. K., Olsen, D. M., Yiin, J. H., Lockey,  J. E., Shukla, R., Selevan, S. G., Schrader, S. M.,
         Toth, G. P., Evenson, D. P., and Huszar, G. B. (1999). Male reproductive effects of solvent and
         fuel exposure during aircraft maintenance. Reprod Toxicol 13, 155-166.
   Lopreato, G. F., Phelan, R., Borghese, C. M., Beckstead, M. J., and Mihic, S. J. (2003). Inhaled drugs of
         abuse enhance serotonin-3 receptor function.  Drug Alcohol Depend 70, 11-15.
   Macdougall, I. C., Isles, C., Oliver, J. S., Clark, J. C., and Spilg, W. G. (1987). Fatal outcome following
         inhalation of Tipp-Ex. ScottMedJ'32, 55.
   Mitoma, C., Steeger, T., Jackson, S. E., Wheeler, K.  P., Rogers, J. H., and Milman, H. A. (1985).
         Metabolic disposition study of chlorinated hydrocarbons in rats and mice. Drug Chem Toxicol 8,
         183-194.
   Monster, A. C. (1979). Difference in uptake,  elimination, and metabolism in exposure to
         trichloroethylene, 1,1,1-trichloroethane and tetrachloroethylene. Int Arch Occup Environ Health
         42,311-317.
   Monster, A. C., Boersma, G., and Steenweg, H. (1979). Kinetics of 1,1,1-trichloroethane in volunteers;
         influence of exposure concentration and workload. Int Arch Occup Environ Health 42, 293-301.
   Morgan, A., Black, A., and Belcher, D.  R. (1972a). Studies on the absorption of halogenated
         hydrocarbons and their excretion in breath using 38C1 tracer techniques. Ann Occup Hyg 1.5,
         273-282.
   Morgan, A., Black, A., Walsh, M., and Belcher, D. R. (1972b). The absorption and retention of inhaled
         fluorinated hydrocarbon vapours. Int J ApplRadiat hot 23, 285-291.
   Morgan, D. L., Cooper, S. W., Carlock, D. L., Sykora, J. J., Sutton, B., Mattie, D. R., and McDougal, J.
         N. (1991). Dermal absorption of neat and aqueous volatile organic chemicals in the Fischer 344
         rat. Environ Res 55, 51-63.
   Nakajima, T., and Sato, A.  (1979). Enhanced activity of liver drug-metabolizing enzymes for aromatic
         and chlorinated hydrocarbons following food deprivation. Toxicol Appl Pharmacol 50, 549-556.
   Nilsson, K. B. (1986a). Actions of 1,1,1-trichloroethane on the cAMP metabolism in mouse brain. Acta
         Pharmacol Toxicol (Copenh) 59, 362-369.
   Nilsson, K. B. (1986b). Effects of 1,1,1-trichloroethane on the cGMP metabolism in mouse brain. Acta
         Pharmacol Toxicol (Copenh) 58, 318-326.
   Nilsson, K. B. (1987). Effects of 1,1,1-trichloroethane on synaptosomal calcium accumulation in mouse
         brain. Pharmacol Toxicol 61, 215-219.
   Nolan, R. J., Freshour, N. L., Rick, D. L., McCarty, L. P.,  and Saunders, J. H. (1984).  Kinetics and
         metabolism of inhaled methyl chloroform (1,1,1- trichloroethane) in male volunteers. Fundam
         Appl Toxicol 4, 654-662.
   NTP (2000). NTP Technical Report on the Toxicity Studies of 1,1,1-Trichloroethane (CAS No.  71-55-6)
         Administered in Microcapsules in Feed to F344/N Rats and B6C3F1 Mice. Toxic Rep Ser 41,1-
         E20.
   Okuda, M., Kunitsugu, I, Kobayakawa, S., and Hobara, T. (2001). Inhibitory effect of 1,1,1-
         trichloroethane on calcium channels of neurons. J Toxicol Sci 26, 169-176.
   Poet, T. S., Thrall, K. D., Corley, R. A., Hui, X., Edwards, J. A., Weitz, K. K., Maibach, H. I, and
         Wester, R. C. (2000). Utility of real time breath analysis and physiologically based

                                               C-83

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

         pharmacokinetic modeling to determine the percutaneous absorption of methyl chloroform in
         rats and humans. Toxicol Sci 54, 42-51.
   Ramsey, J. C., and Andersen, M. E. (1984). A physiologically based description of the inhalation
         pharmacokinetics of styrene in rats and humans. Toxicology and Applied Pharmacology 73, 159-
         175.
   Reitz, R. H., McDougal, J. N., Himmelstein, M. W., Nolan, R. J., and Schumann, A. M. (1988).
         Physiologically based pharmacokinetic modeling with methylchloroform: implications for
         interspecies, high dose/low dose, and dose route extrapolations.  Toxicol ApplPharmacol 95,
         185-199.
   Riihimaki, V., and Pfaffli, P. (1978). Percutaneous absorption of solvent vapors in man. ScandJ Work
         Environ Health 4, 73-85.
   Rosengren, L. E., Aurell, A., Kjellstrand, P., and Haglid, K. G. (1985). Astrogliosis in the cerebral
         cortex of gerbils after long-term exposure to 1,1,1-trichloroethane. ScandJ Work Environ Health
         11,447.455.
   RTI (1987). Absorption, disposition, metabolites, and excretion of 1,1, 1 -trichloroethane (TCEN).
         RTI-213/311T-3662.
   Rudolph, L., and Swan, S. H. (1986). Reproductive hazards in the microelectronics industry. OccupMed
         1, 135-143.
   Schumann, A. M., Fox, T. R., and Watanabe, P. G. (1982a). [14C]Methyl chloroform (1,1,1-
         trichloroethane): pharmacokinetics in rats and mice following inhalation  exposure. Toxicol Appl
         Pharmacol 62, 390-401.
   Schumann, A. M., Fox, T. R., and Watanabe, P. G. (1982b). Carbon-16 labeled methyl chloroform (1 ,1
         ,1 -trichloroethane): Pharmacokinetics in rats and mice following inhalation exposure. Toxicol
         Appl Pharmacol 62, 3 90-401.
   Schumann, A. M., Fox, T. R., and Watanabe, P. G. (1982c). A comparison of the fate of inhaled methyl
         chloroform (1,1,1- trichloroethane) following single or repeated exposure in rats and mice.
         Fundam Appl Toxicol 2, 27-32.
   Shimada, Y. (1988).  Studies on monochlorobenzene poisoning.  III. Distribution  of monochlorobenzene
         in the organs of pregnant mice and its transfer to the fetus through the placenta: Comparison with
         trichloroethylene and 1,1, 1-trichloroethane. OkayamaIgakkai Zasshi 100, 147- 153.
   Stahl, C. J., Fatteh, A. V., and Dominguez, A. M. (1969). Trichloroethane poisoning: observations on
         the pathology and toxicology in six fatal cases. J Forensic Sci 14, 393-397.
   Stewart, R. D., and Dodd, H. C. (1964). Absorption of carbon tetrachloride, trichloroethylene,
         tetrachloroethylene, methylene chloride, and 1,1,1-trichloroethane through the human skin. Ind
         HygJ, 439-446.
   Swan, S. H., Shaw, G., Harris, J. A., and Neutra, R. R. (1989). Congenital cardiac anomalies in  relation
         to water contamination, Santa Clara County, California,  1981-1983. Am JEpidemiol 129, 885-
         893.
   Takahara, K. (1986). Experimental study on toxicity of trichloroethane. I. Organ distribution of 1,1,1-
         and 1,1,2-trichloroethanes in exposed mice. Okayama Igakkai Zasshi 98, 1079-1089.
   Takano,  T., Miyazaki, Y., and Motohashi,  Y. (1985). Interaction of trichloroethane isomers with
         cytochrome P-450 in the perfused rat liver. Fundam Appl Toxicol 5, 353-360.
   Takano,  T., Miyzaki, Y., and Araki, R. (1988). Interaction of 1,1,1-trichloroethane with the mixed-
         function  oxidation system in rat liver microsomes. Xenobiotica 18, 1457-1464.
   Tardif, R., and Charest-Tardif, G. (1999). The importance of measured end-points in demonstrating the
         occurrence of interactions: a case study with methylchloroform and m-xylene. Toxicol Sci 49,
         312-317.
                                                C-84

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.

   Toraason, M., Krueger, J. A., and Breitenstein, M. J. (1990). Depression of contractility in cultured
         cardiac myocytes from neonatal rat by carbon tetrachloride and 1,1,1-trichloroethane. Toxicology
         In Vitro 4, 363-368.
   Wallace, L. A., Nelson, W. C., Pellizzari, E. D., and Raymer, J. H. (1997). Uptake and decay of volatile
         organic compounds at environmental concentrations: application of a four-compartment model to
         a chamber study of five human subjects. J Expo Anal Environ Epidemiol 7, 141-163.
   Wang, F. L, Kuo, M. L., Shun, C. T., Ma, Y. C., Wang, J. D., and Ueng, T. H. (2002). Chronic toxicity
         of a mixture of chlorinated alkanes and alkenes in ICR mice. J Toxicol Environ Health A 65,
         279-291.
   Wang, R. S., Nakajima,  T., Tsuruta, H., and Honma, T. (1996). Effect of exposure to four organic
         solvents on hepatic cytochrome P450 isozymes in rat. Chem BiolInteract 99, 239-252.
   Warren, D. A., Bowen, S. E., Jennings, W. B., Dallas, C. E., and Balster, R. L. (2000). Biphasic effects
         of 1,1,1-trichloroethane on the locomotor  activity of mice: relationship to blood and brain
         solvent concentrations. Toxicol Sci 56, 365-373.
   Wiley, J. L., Fagalde, R. E., Buhler, K. G., LaVecchia, K. L., and Balster, R. L. (2002). Evaluation of
         1,1,1-trichloroethane and flurothyl locomotor effects following diazepam treatment in mice.
         PharmacolBiochem Behav 71, 163-169.
   Wrensch, M., Swan, S., Lipscomb, J., Epstein, D., Fenster, L., Claxton, K., Murphy, P. J.,  Shusterman,
         D., and Neutra, R. (1990a).  Pregnancy outcomes in women potentially exposed to solvent-
         contaminated drinking water in San Jose, California. Am J Epidemiol 131, 283-300.
   Wrensch, M., Swan, S., Murphy, P. J., Lipscomb, J., Claxton, K., Epstein, D.,  and Neutra, R. (1990b).
         Hydrogeologic assessment of exposure to  solvent-contaminated drinking water: pregnancy
         outcomes in relation to exposure. Arch Environ Health 45, 210-216.
   Yang, R. (1993). NTP technical report on the toxicity studies of a Chemical Mixture of 25 Groundwater
         Contaminants Administered in Drinking Water to F344/N Rats and B6C3F(1) Mice. Toxic Rep
         Ser35, 1-112.
   Yoshida, T., Andoh, K., and Fukuhara, M. (1998). Estimation of absorption of environmental
         contaminants in low-level exposure by pharmacokinetic analysis. J Toxicol Environ Health A 54,
         145-158.
   You, L., and Dallas, C. E.  (1998). Regional brain  dosimetry of trichloroethane in mice and rats
         following inhalation exposures. J Toxicol Environ Health A 54, 285-299.
   You, L., and Dallas, C. E.  (2000). Effects of inhaled 1,1,1-trichloroethane on the regional brain cyclic
         GMP levels in mice and rats. J Toxicol Environ Health A 60, 331-341.
                                                C-85

-------
USEPA Contract No. 3C-R102-NTEX                          Principal Investigator/Program Director (Last, first, middle): Yang. Raymond S. H.
                                       Acknowledgments

   This work was funded by the USEPA under Contract No. 3C-R102-NTEX, "Method of Analysis to
   Perform a Tissue-Based Cumulative Risk Assessment for Mixtures of Chemicals" Numerous
   individuals at Colorado State University, Center for Toxicology and Technology participated in the
   production of the work.  Jim Dennison, Manupat Lohitnavy, Ornrat Lohitnavy, Yasong Lu, and Damon
   Perez, assisted with the review of organophosphorus pesticides. Amanda Ashley, Sun Ku Lee, and Ken
   Liao reviewed the literature for chlorinated hydrocarbons. Jim Dennison served as Co-Principal
   Investigator and worked throughout the project on every phase. Linda Monum provided administrative
   support and Christina Barinque provided valuable support with the bibliography.
                                               C-86

-------
                           APPENDIX D
                     EXTERNAL PANEL REVIEW:
Method of Analysis to Perform a Tissue-Based Cumulative Risk Assessment for
                 Mixtures of Chemicals (NCEA-C-1602)
                                 D-1

-------
The following panelists were identified and retained by Versar, Inc.

Richard J. Bull, Ph.D.
MoBull Consulting
Richland, WA

Harvey Clewell (Chair)
CUT Centers for Health Research
Research Triangle Park, NC

Gary L. Ginsberg, Ph.D.
Connecticut Department of Public Health
Hartford, CT

Margaret MacDonell, Ph.D.
Argonne National Laboratory
Argonne, IL

MoizM. Mumtaz, Ph.D.
Agency for Toxic Substances and Disease Registry (ATSDR)
Chamblee, GA

Clifford P. Weisel, Ph.D.
Environmental & Occupational Health Sciences Institute (EOHSI)/UMDNJ
Piscataway, NJ
I. PRE-MEETING COMMENTS

Following dissemination of the draft document and before the convening of the panel
review, comments were solicited from panelists. Two sets of comments were received
(Clewell and Ginsberg); those pre-meeting comments were discussed by the review
panel.  Where concerns remained, those concerns were treated in the final review
comments.  As such, there are no EPA responses to pre-meeting comments.

1.A.  Harvey Clewell

(1) Method of Analysis to Perform a Tissue-Based Cumulative Risk Assessment
for Mixtures of Chemicals (NCEA-C-1602)

1.  Does the document clearly distinguish interactions at the level of cumulative risk
assessment and pharmacokinetics? Does the document present the benefits of
employing tissue dose, rather than environmental concentration as the basis for
cumulative risk assessment?
                                      D-2

-------
      Reviewer Response: The document does provide a reasonably clear distinction
      between interactions as defined in default cumulative risk assessment and
      interactions in pharmacokinetics, although there are some awkward sentences,
      e.g.:

      "Form [stet] an "interactions" perspective, it is possible that pharmacokinetic
      interactions may explain chemical mixture (cumulative risk) interactions, that may
      not be "interactions" as defined by cumulative risk; such departures from
      additivity may only be departures when considered at the level of the external or
      applied dose, and may be strictly additive when "dose" is expressed as "dose
      metric" - the concentration of the toxicologically active chemical species in the
      target tissue."

      I think this sentence should be expanded into a paragraph that takes the
      necessary time to lay out the idea it is being attempted to convey.

      The document does a good job of describing the benefits of using tissue dose,
      but it would be strengthened if examples were presented where considering
      tissue dose made a difference.

2.  The decision point for continuing to develop  a comprehensive cumulative risk
assessment (a PBPK-based approach and assessment) is presented in Step 4 of the
analysis plan.  It is based on the assumption of chemical additivity, and includes a
Hazard Index type analysis coupled with an uncertainty factor intended to address
interactions (in the broad definitional sense). This point has drawn appreciable
comment previously.  The intent of the document is to communicate considerations
undertaken in the series of choices relating to resource expenditure for CRA.  Please
comment on the clarity and strength of the rationale presented and propose other
considerations you feel may be useful to incorporate.  Please comment specifically on
the proposed 1000-fold uncertainty factor and on conditions when it may be warranted.
Below is an example comment and response relating to this point:

Proposed "Rules of Thumb" for Estimating Threshold of Toxicological
Interactions
INTERNAL REVIEWER COMMENT.  "Thus, for practical purpose of conducting CRA"
Interesting rule, but the stated rationale  (stated at the end on page 19) is pretty weak.
The screening calculation  might make more sense if stated in terms of an average
hazard index (i.e., sum of each chemical's exposure potential /RfD, and that sum then
divided by N) or relative potency or some other normalizing method, then applying an
additional screening factor (like an uncertainty factor or margin of safety factor) to
account for potential synergistic interactions. One could also evaluate the potential for
antagonism and a lower level of hazard for a mixture.  A fixed screening factor for
synergistic interactions (100 or 1000) could be consistently applied for all CAGs, but I
think a CAG specific factor would be better.  The case by case factor could include
some weighting for the number in the group, although I think better based upon what we
know about the chemicals' metabolism or other binding, what we know about
                                       D-3

-------
competitive inhibition/binding at different dose levels for the CAG chemicals, or about
chemical reactivity for the chemicals in the CAG. A similar method could apply for a
cancer index using potential exposure over the 10*6 risk level (from slope factors or
linear extrapolations from the NOAELs as the POD 10%).
RESPONSE: This is a very good comment; likewise it was echoed in those from the
other reviewer.  This issue (how best to develop a method to cull out potential scenarios
for which a Comprehensive Cumulative Risk Assessment should be performed) will be
raised for discussion at the level of external review. It may be that such a point (when
to not do such an assessment) should not be specifically "codified." but be left to be
addressed on a case-by-case basis.  This issue will be forwarded to the external review
panel for comment.

      Reviewer Response: I agree with the reviewer above.  The rule of thumb should
      be characterized in terms of the hazard index for the mixture, and the numerical
      value of the "screening factor" for a given mixture should be left to scientific
      judgement.

3. Are you aware of any  additional information which can be used to improve the
present draft report?

      Reviewer Response: No.

4. Are there assumptions or uncertainties being made in this exercise that are not
articulated?

      Reviewer Response: No.

5. Are the figures and tables informative? Would revision of the tables or figures
improve the clarity of the  report  or its conclusions? If so, what are they?

      Reviewer Response: The figures and tables are adequate.

6. Are there important publications missing from the reference section?

      Reviewer Response: I think the report would be improved by doing a literature
      search and adding examples of interaction modeling that demonstrate its value.

7. What additional information would you like to see presented?

      Reviewer Response: Examples of modeling of interactions and how the
      modeling provided a quantitative understanding of the impact of the interaction
      on cumulative risk.

8. Please comment on the overall quality of the report and analysis.  What is your
overall evaluation of the scientific content, readability  and utility of the draft report? Do
                                       D-4

-------
you have any suggestions relative to structure or content that would improve the quality
of this draft report?

      Reviewer Response: For the most part, the report is well written and
      scientifically sound.  It should be useful as a an initial source of discussion on the
      merits of tissue dosimetry in cumulative risk assessment. It sometimes feels
      rather shallow, however, because there are no concrete examples of the
      successful use of tissue dosimetry to improve a cumulative risk assessment. I
      believe such examples exist for chemicals other than those discussed in the
      report, although they admittedly are for high exposure interactions (e.g.,
      TCEA/DC, BTEX).

      There is one scientific inaccuracy I noticed: the reaction of OPs with AChE is not,
      in general, irreversible, as suggested on p. 43.  In fact, characterization of the
      relative rates of regeneration (reversible binding) and aging (irreversible binding)
      is an important aspect of the identification of the more dangerous OPs (e.g.,
      soman).
1.B. Gary L. Ginsberg, Ph.D.

General Comments

This document presents some interesting arguments on how to screen multiple
chemical exposures and determine whether an extensive PBPK modeling assessment
would be needed to evaluate cumulative risk.  In this regard it is useful.  However, it is
written more like a white paper than a framework or general guidance document and so
would likely to be difficult to follow for the average risk assessor. The ultimate user of
this document is unclear (researchers/modelers? headquarters risk assessors involved
in policy? Risk assessors in the field?).  It would  be helpful at the outset to describe the
intended audience and how they might consider  using the approach.  Regardless of the
audience, a summary of the  proposed approach  which highlights the key analytical
phases and steps is needed. The document is wordy and somewhat redundant with the
purpose of individual sections relative to the overall flow not always clear. Starting with
a flow diagram or outline of the proposed approach (e.g.,  the 10 steps divided into 2
phases with brief outline of purpose of each step) would help keep the various sections
in perspective. Then at the end, a section is needed which summarizes the overall
approach to tie it together for the intended user.  For example, it might state that the
screening level approach can be conducted by most risk assessors to quickly evaluate
whether interactions are possible and whether a  more quantitative (modeling) approach
is needed. This then would require recruitment of analysts which can gather the PK
information, build interaction models, etc. etc. Right now, the document makes a
number of good suggestions but does not do a good job of presenting them as an
integrated and practical approach.
                                       D-5

-------
Charge Questions and Reviewer Responses

1.  Does the document clearly distinguish interactions at the level of cumulative risk
assessment and pharmacokinetics?  Does the document present the benefits of
employing tissue dose,  rather than environmental concentration as the basis for
cumulative risk assessment?

      Reviewer Response: The document spends considerable effort distinguishing
      between  interactions than can be judged based upon additivity of effect vs. those
      that occur on a biochemical or toxicokinetic basis.  While this generally comes
      across, it can be more clearly described in Section 1.  For example, the 2nd
      paragraph of page 7 thru page 8 is the place where this distinction is  made for
      the first time.  However, the paragraph starts out with other concepts
      (interactions that are passive/partitioning vs. active - metabolic) that are an aside
      to the key point that the interaction can be on a PD basis (additivity of effect such
      as with the OPs) or on a PK basis (interference with metabolic activation,
      detoxification, clearance, binding, etc.).  From such an  introduction, one could
      then go on to state how these  types of interactions are different and need to be
      analyzed differently (e.g., one  with RfD approach, one with  PBPK analysis) but
      that both types of interactions  need to be considered jointly as part of the overall
      CRA. Setting up the differences while also  explaining the need for a  common
      analytical framework is critical  to the success of the rest of the document. Right
      now the reader has to do too much work to  figure out the implications of the
      distinction that is being made on pages 7-8.  I don't think this is clarified to a
      satisfactory extent at later points where the difference is brought up again.
      However, I must add that the 2 types of mixtures examples as discussed on
      pages 13-14, the OPs vs. the chlorinated solvents, does help clarify the
      distinction in that they exemplify the two types of interactions quite well.  Some
      framing verbiage would help solidify that understanding (e.g., "The two case
      study mixtures were selected to exemplify	, OPs on the one hand represent
      	, while the chlorinated  water contaminants represent the other type of
      interaction at the PK level.  This document uses these different examples to
      show that different types of interactions can be analyzed jointly within a unified
      framework").

      The second part of this charge question, whether the case is adequately made
      for the benefits of analyzing at the tissue dose rather than environmental
      concentration level.  The case for this is argued fairly strongly with at  times
      overzealous language (e.g., - page 4 - "wonderful", "the best"). However, it is still
      somewhat unclear whether the tissue (PBPK) level of analysis is needed for
      CRAs involving interaction  at the PD level.  If this is the case (and it may well be),
      it is not argued well in  this document.  In contrast, the need for internal dose
      resolution for PK interactions is clear.

2.  The decision point for continuing to develop a comprehensive cumulative risk
assessment (a PBPK-based approach and assessment) is presented in Step 4 of the
                                       D-6

-------
analysis plan.  It is based on the assumption of chemical additivity, and includes a
Hazard Index type analysis coupled with an uncertainty factor intended to address
interactions (in the broad definitional sense).  This point has drawn appreciable
comment previously.  The intent of the document is to communicate considerations
undertaken in the series of choices relating to resource expenditure for CRA. Please
comment on the clarity and strength of the rationale presented and propose other
considerations you feel may be useful to incorporate. Please comment specifically on
the proposed 1000-fold uncertainty factor and on conditions when it may be warranted.
Below is an example comment and response relating to this point:

Proposed "Rules of Thumb" for Estimating Threshold of Toxicological
Interactions
INTERNAL REVIEWER COMMENT  "Thus, for practical purpose of conducting CRA"
Interesting rule, but the stated rationale (stated at the end on page 19) is pretty weak.
The screening calculation might make more sense if stated in terms of an average
hazard index (i.e., sum of each chemical's exposure potential/RfD, and that sum then
divided by N) or relative potency or some other normalizing method, then applying an
additional screening factor (like an uncertainty factor or margin of safety factor) to
account for potential synergistic interactions.  One could also evaluate the potential for
antagonism and a lower level of hazard for a mixture. A fixed screening factor for
synergistic interactions (100 or 1000) could be consistently applied for all CAGs, but I
think a CAG specific factor would be better. The case by case factor could include
some weighting for the number in  the group, although I think better based upon what we
know about the chemicals' metabolism or other binding, what we know about
competitive inhibition/binding at different dose levels for the CAG chemicals, or about
chemical reactivity for the chemicals in the  CAG. A similar method could apply for a
cancer index using potential exposure over the 10*6 risk level (from slope factors or
linear extrapolations from the NOAELs as the POD 10%).
RESPONSE: This is  a very good  comment; likewise it was echoed in those from the
other reviewer. This  issue (how best to develop a method to cull out potential
scenarios for which a  Comprehensive Cumulative Risk Assessment should be
performed) will be raised for discussion at the level of external review.  It may be that
such a point (when to not do such an assessment) should not be specifically "codified",
but be left to be addressed on a case-by-case basis.  This issue will be forwarded to the
external review panel for comment.

      Reviewer Response: This screening approach is both intriguing and troubling.
      The attraction to it is that it  offers a unified way to screen different types of
      interaction mechanisms for potential likelihood of occurrence.  The problem is
      that one approach may not fit different interaction mechanisms.  The  RfD/N
      additivity approach comes out  of the common PD MOA - additivity mechanism in
      which each ingredient in the mixture has a common  PD target/endpoint (e.g.,
      brain AChE inhibition). In this case,  a simplistic hazard equation (exposure
      dose/modified  RfD) makes  some sense in that the RfD for single compounds is
      set to be below an effects level for even sensitive individuals and lowering the
      RfD to take into account interacting chemicals  that have the same MOA has
                                       D-7

-------
some rationale. It may be a conservative default if the various chemicals have
similar potency. One gets nervous about this approach for similarly targeted
chemicals when one considers the malathion/dicrotophos comparison presented
on pages 28-29.  Here are similarly acting OPs which have vastly different dose
response for the same endpoint. If we are evaluating an interaction between
these 2 OPs, does it make sense to only lower the malathion RfD (the less potent
ingredient) by a factor of 2?  One might guess that the high potency of
dicrotophos might cause it to be more influential in an interaction scenario and so
should be more heavily weighted.  Perhaps whats needed is an approach which
weights each ingredient's contribution to the interaction based upon its potency
and its dose level  in the system. If dicrotophos is present at levels far  below  its
RfD, then  it likely doesn't matter that its so potent.  Thus, rather than adding
malathion and  dicrotophos together as two equally important constituents
(1+1=2) and then dividing each RfD by 2, perhaps one should consider the
dicrotophos contribution relative to the malathion contribution as follows:

(Malathion RfD/Dicrotophos RfD) * (Dicrotophos dose level/ Dicrotophos  RfD)

If Malathion RfD is 1000x greater than the Dicrotophos RfD and the dicrotophos
dose is 1000x  lower than its RfD, then the above equation yields 1 and the
dichrotophos contribution to the cumulative interaction would equal that
contributed by  malathion and you could divide the malathion RfD by 2 (N=2) as in
the proposed screening method. However, if the dicrotophos exposure level  is
0.1 of the  RfD, then the malathion RfD would be divided by 1 (for malathion) +
100 (from  above equation for dicrotophos contribution) (N = 101).  Obviously  this
can become more complicated and unwieldy for ternary or higher order
interactions and so one needs a simplifying assumption that the potency
adjustment is only needed for the more potent ingredients  and only if they
exceed a given ingredient's  potency (lower RfD) by 3x or more.  Thus,  if you
have 5 OPs in  the mixture with 3 having similar potency and 2 being considerably
more potent, you would only need to run the above equation twice to calculate
how much RfD lowering is contributed by these 2 more potent ingredients.

The above discusses issues with the screening approach for chemicals acting via
similar PD MOAs. However, this approach may not be so  relevant for chemicals
which  interact strictly via PK mechanisms (e.g., the chlorinated solvents
interacting at the level of CYP2E1). In this case, there may be no relationship
between a chemical's RfD and its ability to compete for throughput via 2E1. For
example, a chemical may have a high RfD but also a high  affinity for 2E1 which
may make it a  stronger metabolic interactor than a chemical with a low RfD but
weak affinity for 2E1.  Thus, for PK-interacting chemicals, screening on the basis
of RfD/N may have little relevance  to what is  going on biochemically, especially
since these chemicals may have different PD targets and MOAs (e.g., one
solvent affecting CMS while  other affects liver). The example of fenitrothion and
parathion  interaction is a good example of the fact that a chemical does not have
to exert a  measureable toxic effect (only a biochemical effect) to synergize
                                 D-8

-------
another chemical. Fenitrothion ties up peripheral carboxylesterases so that at
doses well below those required for fenitrothion AChE inhibition, a synergistic
interaction with parathion is set up.

It may well be that in most cases, the RfD is set low enough so that not only are
risks for toxicity averted but also there is no biochemical perturbation relevant to
interaction with other chemicals at the RfD/N dose.  However, the case for this
needs to be explored and made stronger. For example, dioxin is a potent
inducer of the CYP1A family which may cause interactions with other CYP1A
family substrates.  Do we know that RfD/N for dioxin (or PCBs) will make this
type of biochemical/interaction mechanism moot?

I believe that evidence supporting this screening approach may be obtainable via
interactive case studies which involve PBPK analyses exploring what goes on at
high dose vs. what occurs at  RfD-type doses.  My guess would be that for simple
competition interactions and protein binding interactions, the RfD/N approach
would be a reasonably conservative screen.  However, this may be less clear for
enzyme induction interactive  mechanisms, or perhaps for non-competitive
inhibition mechanisms.  The fenitrothion/parathion interaction mentioned above
would be interesting to test in this regard.

This charge question also raises the issue of whether a default approach
(NOAEL /1000 fold UF) is a useful default in the case where an RfD is
unavailable for a particular constituent.   The idea is that a 1000x cumulative UF
will create an RfD that is low  enough  to protect against the possibility of
synergistic interactions. This approach does not appear to be well thought out.
The RfD should be set based upon the relevance and strength of the underlying
database.  The greater the uncertainty, the larger the cumulative  UF. RfDs are
set based upon the dose-response assessment for individual chemicals and are
not based upon whether they have an interaction with other chemicals.  It would
be preferable to stick with the standard risk assessment approach to RfD setting
when starting from a NOAEL  and not use a default factor that incorporates
concerns over synergistic interaction. The analysis of interaction, whether
screening or detailed, should  be a separate step from dose-response
assessment. Since IRIS RfDs do not incorporate a group interaction UF, to use
one in the case of a missing RfD would create an apples and oranges situation in
which the different types of RfDs would not be comparable.

It is important to note that the focus should not be on how much potentiation is
possible in a synergistic interaction but whether the concentrations of the
reactants are high enough for a meaningful interaction  to take place. If this is so,
then one moves beyond the screening level to Phase II in the 10  step process.
Hopefully during this phase one can determine whether synergy is possible and
to what extent.
                                 D-9

-------
      If there is a particular concern over the possibility of synergy during the screening
      phase of the analysis (e.g., OPs that can interact the way that fenitrothion
      interacts with malathion) one could incorporate it into the RfD/N approach by
      increasing the value of N for synergistic chemicals in a manner similar to the
      dichrotophos example above. Such a method could be tested with known
      synergistic interactions on a pharmacokinetic level of analysis to document that
      the screening level approach is protective against synergists. Such analyses
      may show what additional factor (if any) is needed to modify N in the RfD/N
      paradigm to ensure protection against synergistic interactions.

      In summary, I would recommend against adjustment of the way in which one
      goes from a NOAEL to an RfD but would rather adjust for synergistic interactions
      on a case specific basis  by adjusting the value of N in RfD/N.

3.  Are you aware of any additional information which can be used to improve the
present draft report?

      Reviewer Response: As stated above, the fenitrothion/parathion synergistic
      interaction is important to mention. The document should also describe other
      types of interactions (non-competitive inhibition, enzyme induction, etc) and how
      they can be brought within the CRA analytical framework.

4.  Are there assumptions or uncertainties being made in this exercise that are not
articulated?

      Reviewer Response: No, nothing not already talked about.

5.  Are the figures and tables informative? Would revision of the tables or figures
improve the clarity of the report or its conclusions?  If so, what are they?

      Reviewer Response: As described above, a framework overview figure would
      be very helpful at the outset and a summary version at the end showing key
      decision points would also be helpful.

      Figure 1 does not capture the interaction between OPs at the level of peripheral
      carboxylestases (CE). A prime example of this is the fenitrothion synergism of
      parathion. This type of interaction should be included in the chart.

      Figure 2 might be improved by indicating the direction of the interaction at each
      step: e.g., if 1 inhibits 3 at level of 2E1 then there might be more of 3 available for
      GSH conjugation and renal toxicity. Without this context the figure can seem
      overwhelming and may not relate as clear a  message regarding  the importance
      of the interactions., vis-a-vis the text.
                                        D-10

-------
6.  What additional information would you like to see presented?

      Reviewer Response: The paper lacks citations of PBPK modeling studies in
      which interactions were evaluated. I believe several studies have been
      published along these lines. Additional background information on the various
      types of PD and PK interactions possible would be helpful (e.g., interaction at
      level of competitive or non-competitive enzyme inhibition; at level of GSH
      depletion predisposing to second toxicant that needs GSH; at level of enzyme
      induction, etc.

7.  Please comment on the overall quality of the report and analysis.  What is your
overall evaluation of the scientific content, readability,  and utility of the draft report?  Do
you have any suggestions relative to structure or content that would improve the  quality
of this draft report?

      Reviewer Response: For overall quality, readability and utility,  see general
      comments above. To reiterate, the intended end user(s) should be stated up
      front, how and when they might use the document should also be stated, and
      then this should be kept in  mind throughout the writing.  This may enable it to
      become less of a white paper and more of an analytical  framework document.

      Some specific comments are as follows:

      Section 1.1 -the lengthy quotes from the Hansen and Oilman  memos do not
      appear to be necessary.  The main points can be highlighted much more briefly,
      using selective quotes.

      Page 16, large para - This paragraph addresses the problem  of aggregating
      exposures across chemicals, each of which has its own  data distribution.  The
      discussion here seems too complex and misses what I think is the main option -
      a screening level approach where an upper bound value is used for each
      contaminant.  If there are not signficiant interactions with these across the board
      upper bound concentrations, then there is little  need to worry about the full data
      distributions for each  analyte. There are a number of options for more refined
      analyses. One in  particular that may be attractive is to run the interaction
      scenario 4 times (if there are 4 chemicals interacting in the mixture). Each run
      uses the 95th percentile for one chemical and the average concentration for the
      other 3 chemicals when the first chemical is at its 95th percentile. The  point
      sampling approach on pages 16 to 17 is worth mentioning but appears to  be
      impractical.

      Page 19, bottom para - The issue of age-related variability is briefly discussed.
      Some contextual discussion is needed of the types of immaturities that may
      predispose to PK or PD interactions in early life as opposed to adults.  One can
      cite the early life vs. adult OP studies showing greater sensitivity in young  rats
      with mechanistic work demonstrating both PK and PD bases for the sensitivity
                                       D-11

-------
      differences. This has obvious bearing on the potential for OP to OP interactions
      in young vs. old rats.

      Page 27 - middle of page - "some number of components would be at a
      significant fraction of the NOAEL. These components would survive the initial
      analysis."
            What is meant by a significant fraction?  Several lines later it states that
      others may be present at levels below or near RfD/N and so could be excluded.
      This language is too vague. The screening level RfD/N approach needs to be
      clarified with regards to what exposure level (relative to RfD/N) would constitute a
      significant interaction potential.

      Pages 36-38 and beyond - discussion of PBPK modeling of chemical
      interactions - this section mentions some key parameters that will be needed but
      does not describe how these parameters can come together to predict the
      outcome of the interaction - e.g., that the ratio of tissue concentration to Km  is a
      key factor for competitive interactions involving M-M kinetics. It would be useful
      to have a table of the key interactive equations (and parameter definitions) for the
      major types of interactions: M-M enzyme competive, non-competitive;
      competition for protein binding sites (e.g., albumin, CEs), enzyme induction, etc.

      Page 38,  bottom: should describe some of the issues with recombinant CYP
      systems (e.g., expression systems may not also express electron transport
      enzymes or they may be  in wrong proportions).
II.  FINAL REVIEW COMMENTS

II.A.  Document Overview

Dr. Lipscomb provided an overview of the document and reviewed the charge
questions. He explained that pharmacokinetics is important in evaluating human
exposure and refining the doses used in the dose-response relationship.  It is important
in cumulative risk and mixtures assessment because the internal doses of some
components of the mixture can alter the tissue dosimetry of other components. Dr.
Lipscomb provided as background information the fact that Congress mandated that
EPA conduct a cumulative risk assessment for organophosphorous (OP) pesticides that
share a common mode of action. Because of that mandate, the EPA's guidance
determines that when chemicals have the same mode of action, one can use the dose-
addition or response-addition approach to evaluate cumulative risk. One of EPA's
challenges is to develop methods to estimate  risks from  environmental mixtures
containing a variety of chemicals, often acting through different modes of action. PBPK
modeling is a valuable tool for developing the  approaches  to conduct cumulative risk
assessments, with internal tissue doses being a critical step in the process.  Dr.
Lipscomb suggested that the group should keep this important issue in mind when
reviewing the first document.
                                       D-12

-------
Method of Analysis to Perform a Tissue-Based Cumulative Risk Assessment for
Mixtures of Chemicals is a document that attempts to take a 10-step process,
developed by the Office of Pesticide Programs under a Congressional mandate, and
inform a cumulative risk assessor of the benefits and considerations that need to be
recognized when the assessor applies the 10-step process. The document presents
two examples of chemical mixtures,  OP pesticides and volatile organic chemicals.  Dr.
Lipscomb stated that the key step is the decision point which is after Step 5 in  the 10-
step process. The document discusses comprehensive cumulative risk assessment
which means a physiologically- or tissue-based  approach.  Dr. Lipscomb believes that it
will be difficult to get to that stage of risk assessment due to the lack of reliable data.
He stated that the analysis still needs to be done but there might be uncertainty in the
outcome.

Dr. Lipscomb indicated that confusion may arise about the term "interaction," which in
cumulative  risk assessment means that the observed toxic response cannot be
predicted by the addition of the doses of the individual components or responses from
the addition of individual components.  Pharmacokinetics can assist in considering
metabolic interactions and estimating chemical concentrations in the target tissue.

Dr. Lipscomb explained that the document is intended for NCEA internal use.  He
reviewed the charge questions and highlighted that he would like comments regarding
the decision point step of the process.
II.B. General Comments

Mr. Clewell requested general comments from the reviewers on the document, Method
of Analysis to Perform a Tissue-Based Cumulative Risk Assessment for Mixtures of
Chemicals. Dr. Richard Bull stated that the reaction network modeling references
interrupt the flow of the document and could be made into footnotes.  Dr. Gary Ginsberg
recommended toning down the "overzealous" language (terms like "wonderful").  Dr.
Moiz Mumtaz felt that the document did not reflect the EPA 2000 "Supplementary
Guidance for Mixtures" document. Also, he felt that the approach for assessing 3-4
chemicals is not adequate for more realistic environmental mixtures containing
hundreds of chemicals.  Dr. Clifford Weisel felt that the section on pharmacodynamics
was rushed and Dr. Bull agreed; they decided to address this issue in the responses to
charge questions.  Mr. Clewell agreed with the previous comments and added that there
should be discussion of all the different  levels of interactions that can occur, not just
metabolic enzyme interactions.  He also added that these metabolic interactions
typically do not occur at low doses, which  is needed for the assessment to be
interesting in terms of evaluating interactive effects.  Dr. Ginsberg agreed that it was
important to point out that other pharmacokinetic interactions exist, but to state that the
document focuses on the 10-step process. Dr. Mumtaz agreed that other types of
interactions should be mentioned and acknowledged, but that it should  be pointed out
that they wouldn't be completely explained or solved in the document.  Dr. Ginsberg did
                                       D-13

-------
not feel that the document was user-friendly and would like a flowchart of the 10-step
process included up front in the document to help the reader understand where they are
in the process as they move through the steps.  Mr. Clewell agreed and would also like
a diagram of the levels of interaction included.
II.C.  Response to Charge Questions

1.  Does the document clearly distinguish interactions at the level of cumulative risk
assessment and pharmacokinetics?

      Reviewer Response: Yes, it does, for the most part, but more discussion in the
      early part of the document would be helpful (section 1.6). Examples of published
      interactions and their impact would also help. E.g., mention of interaction
      threshold dose studies.  Case studies could be used to demonstrate the different
      levels of interactions.

      There needs to be some discussion of other levels of interactions (e.g.,
      pharmacodynamic) between the two extremes currently identified in the
      document.

      EPA Response: Section 1.6 has been substantially expanded.  Examples of
      some published interactions have been added. The second paragraph contains
      a description of a work by Dobrev and colleagues that demonstrates interactions
      thresholds.  This constitutes one case study where an interaction threshold has
      been determined.

      Some discussion of other levels (types) of interactions has also been added to
      section 1.6. The fifth paragraph contains the example of piperonyl butoxide
      (PBO) and a Detoxicated insecticide.  Here, PBO exerts an effect which is
      unlikely associated with  the toxicity of PBO, but which markedly decreases the
      ability of the organism to metabolize the active insecticide.  Also, the example of
      thioacetamide and carbon tetrachloride has been included in paragraph 6.  This
      example demonstrates how pretreatment with one toxic compound can be
      protective against subsequent and otherwise lethal doses of another compound.
      Finally, paragraph 8 contains an example indicated in the reviewers' comments
      to charge question 3. The OP inhibition data from the Chambers et al. and
      Cohen et al. publications have been included in this paragraph.

Does the document present the benefits of employing tissue dose,  rather than
environmental concentration as the basis for cumulative risk assessment?

      Reviewer Response: A better case is made for the value of tissue dose for
      pharmacokinetic interactions than for its value for cumulative risk assessment.
      Need to include a hypothetical case study of receptor interaction in section
      2.3.1.4.
                                       D-14

-------
      EPA Response: The modifications made in response to Charge question 1
      largely address this comment.  The additions specifically include measures of
      internal dose (especially the example by Dobrev and colleagues) for use in
      interactions as well as in cumulative risk assessment.  The report has been
      challenged to include characterizations of receptor interactions in multiple
      locations have been suggested by one particular reviewer.  We feel that the
      document adequately conveys its general points without the inclusion of a
      receptor interaction example.

2.  The decision point for continuing to develop a comprehensive cumulative risk
assessment (a PBPK-based approach and assessment) is presented in Step 4 of the
analysis plan. It is based on the assumption of chemical additivity, and includes a
Hazard Index type analysis coupled with an uncertainty factor intended to address
interactions (in the broad definitional sense). This point has drawn appreciable
comment previously.  The intent of the document is to communicate considerations
undertaken in the series of choices relating to resource expenditure for CRA. Please
comment on the clarity and strength of the rationale presented and propose other
considerations you feel may be useful to incorporate.  Please comment specifically on
the proposed  1000-fold uncertainty factor and on conditions when it may be warranted.
Below is an example comment and response from the  Internal Review relating to this
point:

Internal Review Comment, Proposed "Rules of Thumb" for Estimating Threshold
of Toxicological Interactions
INTERNAL REVIEWER COMMENT  "Thus,  for practical purpose of conducting CRA"
Interesting rule, but the stated rationale  (stated at the end on page 19) is pretty weak.
The screening calculation might make more  sense if stated in terms of an average
hazard index (i.e., sum of each chemical's exposure potential /RfD, and that sum then
divided by N)  or relative potency or some other normalizing method, then applying an
additional screening factor (like an uncertainty factor or margin of safety factor) to
account for potential synergistic interactions.  One could also evaluate the potential for
antagonism and a lower level of hazard for a mixture. A  fixed screening factor for
synergistic interactions (100 or 1000) could be consistently applied for all CAGs, but I
think a CAG specific factor would be better.  The case  by case factor could include
some weighting for the number in  the group, although I think better based upon what we
know about the chemicals' metabolism or other binding, what we know about
competitive inhibition/binding at different dose levels for the CAG chemicals, or about
chemical reactivity for the chemicals in the CAG. A similar method could apply for a
cancer index using potential exposure over the 10*6 risk level (from slope factors or
linear extrapolations from the NOAELs as the POD 10%).
RESPONSE:  This is a very good  comment; likewise it was echoed in those from the
other reviewer. This issue (how best to develop a method to cull out potential
scenarios for which a Comprehensive Cumulative Risk Assessment should be
performed) will be raised for discussion  at the level of external review.  It may be that
such a point (when to not do such an assessment) should not be specifically "codified",
                                       D-15

-------
but be left to be addressed on a case-by-case basis.  This Issue will be forwarded to the
external review panel for comment.

      Reviewer Response:
      A. Conceptually, such a screening rule of thumb is a good idea. Practically, the
         difficulty is knowing the relationship of the RfD and the interaction threshold.
         Should use target organ toxicity data, not most sensitive NOAEL, since the
         latter may not be for same site.
      B. A default factor of 1000 should not be applied; instead an RfD equivalent
         should be derived for the appropriate effect, as explained in the TTD
         document.  Secondary endpoints (e.g., developmental) should also be
         considered.
      C. The document should make the point that the interaction threshold may be
         determined by pharmacokinetic factors (for example, saturation of enzyme
         systems and enzyme inhibition or induction), and not a toxicity threshold.  The
         interaction threshold may determined by data on the dose/response for the
         induction of an enzyme, for example, but this should not be confused with a
         threshold for a toxic effect.
      D. Case by case informed scientific judgment based on knowledge of the
         mechanism of action is necessary. The proposed rule of thumb may be
         valuable for the organophosphate common mechanism group, but it is
         questionable to apply it for metabolic interactions, without data on the
         relationship of the dose-responses for metabolic interactions and toxicity.

      EPA Response:
      A. The proposed rule of thumb has been removed and the issue has been
         characterized as one that should be undertaken on a case-by-case basis.
      B. The default factor of 1,000 has been removed. The entire Rule of Thumb text
         has been removed. Section 2.2.4 has been substantially expanded  and now
         includes a worked example of a Target Organ Toxicity Dose example as
         suggested by the reviewers.
      C. That toxicokinetics (e.g., metabolic interactions) can be the determinant of
         mixtures interactions has been clearly included, see response to charge
         question 1 and section 1.6 of  the document.
      D. The approach has been restructured and communicated in a manner to
         clearly indicate that the decision to proceed with developing a physiologically
         based pharmacokinetic analysis to mixtures risk should be undertaken on a
         case bycase basis.

3.  Are you aware of any additional information which can be used to improve the
present draft report?

      Reviewer Response:
      A. Examples from the literature,  demonstrating the effect of interactions on
      toxicity.  Presentation of the literature on receptor interactions.  Mention of other
                                       D-16

-------
      types of interactions (enzyme induction, non-competitive inhibition,
      Equations describing interactions.
etc.).
      B. A good example of a mechanism of interaction that would have very little to
      do with the proposed RfD-based cumulative analysis is the potentiation of OP
      toxicity by tieing up serum carboxylesterases. Pretreatment  with low toxicity
      pesticides like fenitrothion and other low toxicity chemicals such as bis-p-
      nitrophenyl phosphate (BNPP) can potentiate malathion, soman, etc.  The
      LOAEL, NOAEL and RfD for the potentiating agents would be based upon other
      endpoints that would not reflect their ability to interact with OPs.

      References:   Chambers, et al., Effects of 3 reputed carboxylesterase  inhibitors
      upon rat serum esterase activity.  Neurosci Biobehav Rev 15: 85-88, 1991.

      Cohen, Mechanisms of toxicological interactions involving organophosphate
      insecticides.   FAT 4: 315-324, 1984.

      EPA Response:
      A Other types of interactions have been included, specifically in section 1.6. An
      example of enzyme induction has not, however, been included, neither has one
      on receptor binding been included. The report is constructed to guide the risk
      analyst through the process of justifying and undertaking physiologically based
      analysis of mixtures and or cumulative risk.  As such, more than a few examples
      are not warranted.

      B The information contained in the references suggested by the reviewers has
      been included in passages inserted into section 1.6.  This passage describes OP
      and AChE interactions in more detail.

4.  Are there assumptions or uncertainties being made in this exercise that are not
articulated?

      Reviewer Response: Use of Hazard Index approach for estimating threshold for
      metabolic interactions,  (although effect threshold is ultimate issue.)

      EPA Response: The application of the Hazard Index approach does  not and is
      not intended to address metabolic interactions.  Perhaps the reviewer
      misunderstood something. No change made.

5.  Are the figures and tables informative? Would revision of the tables or figures
improve the clarity of the report or its conclusions?  If so, what are they?

      Reviewer Response: Need flow chart of 10 steps and diagram of multiple levels
      of interactions (PK, PD).
      Need table of definitions.
                                       D-17

-------
      EPA Response: A flow chart of operations has been developed and inserted as
      Figure 1. It divides the activity into the Initial Assessment, comprising steps 1-5,
      and the  Dosimetry-Based Cumulative Risk Assessment, comprising steps 6-10.

6.  Are there important publications missing from the reference section?

      Reviewer Response: Mehendale et al. (effects on repair of injury). Swartout et
      al. (combining HQs). Target organ toxicity doses (EPA/ATSDR publication).
      Krishnan et al. (BTEX). Andersen et al. (TCE/DCE). Bull et al. (TCA/DCA).
      Chambers et al., Pope et al.  (OP mixture studies). ATSDR interaction profiles.
      Rider et al. (human OP studies).  Paul Price (P3M). Herzberg (MIXTOX
      database).

      EPA Response:
      •  The reviewers point out several areas of work which could enhance the depth
         of the report and/or broaden  its treatment of uncertainty. The report has been
         revised to include an example from Mehendale (Tissue repair: an important
         determinant of final outcome of toxicant-induced injury. Toxicol Pathol. 33:41-
         51, 2005) on pre-exposure and its effect on stimulating tissue repair near the
         end of section 1.6.
      •  Swartout's reference on probabilistic approaches to developing reference
         doses (A probabilistic framework for the reference dose (probabilistic RfD)
         Risk Anal. 18:271-82, 1998) was considered but not included due to its
         indirect relationship to pharmacokinetics and the assessment of exposure.
      •  A complete and substantial treatment of Target Organ Toxicity Doses has
         been incorporated in section 2.2.4. References from the US EPa and from
         Mumtaz have been cited.
      •  The reviewer suggests inclusion of a publication by Krishnan and colleagues
         on BTEX.  Instead, the report has  been revised to include a reference from
         Dobrev and colleagues that addresses the same issue (thresholds for
         metabolic interactions)  as does the Krishnan paper.  This example can be
         found a couple of pages into section 1.6.  The Bull publication on TCA/DCA
         could be one of several that,  en masse, conclude that dichloroacetic acid is a
         metabolite of trichloroacetic acid and that the liver tumors produced by DCA
         and TCA are phenotypically different, when cell membrane proteins are
         evaluated. Bull and colleagues concluded that, when administered
         separately, TCA and DCA promoted the outgrowth of different subpopulations
         of spontaneously-arising tumors in rodents. A complicating factor for
         estimating internal doses of DCA is that administration of DCA at high doses
         tends to inhibit its own metabolism, prolonging biological residence time.  It  is
         not immediately clear how the inclusion of these points would improve the
         manner in which the document informs the risk assessor in choosing whether
         and how to implement PBPK modeling to refine mixtures or cumulative risk
         assessment.
      •  A PubMed search revealed no returns for organophosphate mixture studies
         published by Jan Chambers or by Carey Pope.
                                       D-18

-------
•  The reviewer suggests that the reference list should include some
   Interactions Profiles developed by the ATSDR.  A reference to that database
   has been added to section 2.2.1.
•  Extensive searching for Rider and human organophosphate returns was
   fruitless.  However, a chapter in the PhD dissertation (2005) for Cynthia V
   Rider demonstrates the development and application of a mathematical
   model for inhibition of a detoxicating enzyme in a mixtures toxicity context:

       An  Integrated  Addition and  Interaction (IAI)  model of mixture toxicity  was
       constructed and  validated  using  a  ternary mixture  of organophosphates
       (malathion and parathion) and the P450 inhibitor piperonyl butoxide. Individual
       chemical  concentration-response parameters and  binary interaction data were
       used in the model. Modeled data was compared to experimentally derived  data
       from Daphnia magna acute toxicity assays. The IAI model provided a good  fit to
       the data. Results indicated that toxicokinetic interactions could be quantified and
       incorporated into mixture toxicity models.
       http://www.lib.ncsu.edu/theses/available/etd-01022006-
       223335/unrestricted/etd.pdf

   However, this is not a human data set, and the example is the one for
   piperonyl butoxide, already incorporated into Section 1.6. No change was
   made.

•  The reviewer has suggested  that the document could be improved by
   including reference to the work done by Paul Price and collaborators,
   especially that from the P3M  database. This database contains measured
   variability in physiological parameters in humans.  Because variation in these
   parameters may influence internal dosimetry, they can influence risk.
   Application of the data  in the  database would improve the estimates of
   variability in internal dosimetry for a given chemical, but because  variability in
   a given direction for one parameter can increase dosimetry for one compound
   while decreasing dosimetry for  another compound, it is not clear that inclusion
   of this work would increase the document's treatment of internal dosimetry as
   a modulator of risk, except when a PBPK model has already been developed
   for the chemical(s) of interest. As such, this work (abstract from the 2003
   meeting of the Society  for Risk  Analysis below) has  not been included in the
   revisions to the  report.

       Modeling Inter-individual Variation in Physiological Factors Used in PBPK
       Models of Humans, by PS Price et al. SRA, 2003,
       Modeling interindividual variation in internal dose in humans using PBPK  models
       requires data  on the   variation  in the  physiological parameters across the
       population of interest. These data should also capture the correlations between
       the values in each person. In  this project, we developed a tool to provide such
       data and  its  correlations.  The  tool provides a  source  of  data for  human
       physiological parameters where  1) the parameter values for an individual are
       correlated  with one  another,  and 2) values of parameters vary according  to
       interindividual variation in the general population, by gender, race, and age. The
       parameters investigated in this project include: 1) volumes of selected  organs
       and tissues; 2) blood flows for the organs and tissues; and 3) the total  cardiac
                                    D-19

-------
             output  under resting  conditions  and  average daily inhalation  rates. These
             parameters are expressed as records of correlated values for the approximately
             30,000 individuals evaluated in the NHANES III survey. Software was developed
             that allows records to be retrieved randomly from the database with specification
             of constraints  on,  age,  sex,  and  ethnicity.  The  [P3M] database  and
             accompanying software together provide a convenient tool for parameterization
             of human PBPK models for the study of interindividual variation. In addition, the
             data provides a useful information on the variation in physiological parameters in
             adults and children. This work was funded by the American Chemistry Council.

      •  The reviewer also suggested the MixTox database developed by Hertzberg
         and collaborators. While this database was  useful, especially in assessing
         the likelihood of interactions in binary mixtures, it has not been maintained
         and is no longer publicly available.

7.  What additional information would you like to see presented?

      Reviewer Response:
      A. Receptor theory (e.g., from textbook  by Pratt and Taylor). Mechanism vs.
      mode of action and implications for interactions  (AChE reversible inhibition vs.
      aging vs. ion channel disruption. The document over-simplifies the nature of OP
      inhibition of AChE, which can be either reversible or irreversible (aging).
      Complex mixture issues (DBPs, gasoline).

      B. Some discussion of time element (variation in exposure) and linkage to
      exposure modeling.

      C. How to  identify potential exposures is not covered. Need indication of
      completed exposure pathway for mixture to define need for CAG.

      D. Need clearer criteria for step 5 cutpoint. (availability of data?, evidence of
      interaction?).  Just common metabolism enzymes is not the only criteria.

      EPA Response:
      A.
      •  The document will not be revised to include receptor theory.  The reviewer
         seems to indicate that the document  would benefit from inclusion of the
         distinction between mechanism and mode of action.  The basis for deciding
         the grouping of chemicals for CRA is the Common mechanism Group (CMG)
         per US EPA guidance. Note that this distinction is made at the level of
         "Mechanism", not Mode.  A key point underlying this comment is perhaps
         sentiment that critical toxicologic interactions can be based on effects that are
         not directly involved in the mechanism of action, or are  based on events
         included in the mode, but not the mechanism.  Such an interaction based on
         the former is exemplified by piperonyl butoxide, as can  now be found in
         Section 1.6. The opening passages  in section 1.6 have also been revised to
         include a distinction between Mechanism and Mode and a caveat about how
                                        D-20

-------
         focusing only on mechanism can result in an overlooking of important bases
         for chemical interactions.

      •  The document has been revised to include a more technical treatment of OP
         interactions, specifically to include binding half-life and aging of enzyme
         complexes.  This can be found in a separate paragraph near the end of
         section 2.2.4.

      •  The reviewer indicates that complex mixtures such as gasoline and drinking
         water disinfection byproducts should be addressed. It is felt that the
         considerations presented in this document apply equally to simple and to
         complex mixtures.  No revision made.

      B. The reviewer indicates that complex mixtures such as gasoline and drinking
      water disinfection byproducts should be addressed.  It is felt that the
      considerations presented in this document apply equally to simple and to
      complex mixtures.  No revision made.

      C. Per the reviewer's suggestion, a passage has been added to the end of
      section 2.2.2 which communicates the 5 steps identified by ATSDR and
      references their 2005 guidance.

      D. There is not going to be a more clearly made cut-point. The document
      instead espouses a case by case approach to the decision.

8.  Please comment on the overall quality of the report and analysis.  What is your
overall evaluation of the scientific content, readability and utility of the draft report?  Do
you have any suggestions relative to structure or content that would improve the quality
of this draft report?

      Reviewer Response: Excellent explication of mixtures issues within the context
      of a standard paradigm.  The document makes its main point (the importance of
      considering metabolic interactions) well.  The main requirement to improve it is
      some discussion of other levels of interaction (PD) and examples from the
      literature of the impact of PK and PD interactions on  cumulative risk assessment.

      Put reaction network discussion in footnotes or delete.
      Tone down language ("wonderful").

      "Comprehensive" implies more than just target tissue based. Refer to as
      "dosimetry-based" instead.

      EPA Response:
      A. Several additional interaction types and results have been included in the
      report including metabolic induction, metabolic  inhibition, and alterations of
      biological response (dynamics, induction of tissue repair).
                                       D-21

-------
B. Reaction network modeling was mentioned in two places in the draft report.  It
was retained in the first instance, but was there relegated to a footnote.  It was
deleted in the second instance.

C. There were several subjective and biased terms included in the external
review draft. These sporadic instances of overzealousness have been replaces
with more objective terminology.

D. CCRA has been globally changed to Dosimetry-Based Cumulative Risk
Assessment (DBCRA).
                                 D-22

-------