United States      Science Advisory       EPA-SAB-EC-99-013
Environmental      Board (1400)          July 1999
Protection Agency     Washington DC       iviviv.epa.gov/sab

REVIEW OF THE ERA'S
PROPOSED
ENVIRONMENTAL
ENDOCRINE DISRUPTOR
SCREENING PROGRAM
REVIEW OF THE ENDOCRINE
DISRUPTOR SCREENING
PROGRAM BY A JOINT
SUBCOMMITTEE OF THE
SCIENCE ADVISORY BOARD AND
SCIENTIFIC ADVISORY PANEL

-------
                                     July 28, 1999

EPA-SAB-EC-99-013

Honorable Carol M. Browner
Administrator
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, DC 20460

             Subject:      Review of EPA's Proposed Environmental Endocrine Disrupter
                          Screening Program.

Dear Ms. Browner:

       In 1996, the passage of the Food Quality Protection Act (FQPA) and amendments to the
Safe Drinking Water Act (SOWA) required EPA to develop a screening and testing strategy for
environmental endocrine disrupters. The Agency established the Endocrine Disrupter Screening
and Testing Advisory Committee (EDSTAC) to provide advice on the screening and testing of
pesticides and other chemicals for their potential to disrupt the endocrine system.  The EPA
subsequently asked the Science Advisory Board (SAB) and the FIFRA Scientific Advisory Panel
(SAP) to form a Joint Subcommittee to review a set of scientific issues being considered by the
Agency concerning the development of the Agency's endocrine disrupter screening and testing
program as required by the legislation noted above. A Joint Subcommittee (the Joint
Environmental Disrupter Screening Program (EDSP) review Subcommittee) met on March 30-
April  1, 1999, in Arlington VA, and produced this report.

       The Charge was broad and complex,  posing 18 major questions within four broad areas:
a) scope of the program; b) priority-setting; c) the high throughput pre-screening approach; and
d) the proposed endocrine disrupter screening program (the complete Charge is provided in
section 2.2 of this report).

       At the outset, we wish to note that although our review identified several areas of
concern, and the EDSP has provided recommendations to improve EPA's planned program, we
wish to congratulate the Agency for dealing  effectively with an extraordinarily complex  set of
issues, many of which are on the cutting edge of the relevant science. The EDSP's detailed

-------
response to each element of the Charge can be found in section 3 of the report, and our major
issues and recommendations are summarized below:

       a)     Evaluating the Program: We find no provision for mid-course evaluation or
             optimization of the process. Although an approach may look fine on paper or in a
             small research setting, translating it into a volume-screening mode may be quite
             another thing.  There was broad support among the Subcommittee for the concept
             that the Agency should convene a panel of independent scientists to review all the
             screening data for 50-100 compounds, with an eye towards revising the process
             and eliminating those methods that don't work.

       b)     Mixture Issues: The Subcommittee agreed that the initial focus of the methods
             development effort must focus necessarily on single compounds and leave the
             question of testing of mixtures until accepted single-compound methods have
             been completed.

       c)     Case Studies: The Subcommittee strongly encourages the Agency to include
             more and better-detailed case studies in the evolution of the priority-setting
             scheme. This will facilitate a realistic test of the plan, checking the  sensitivity of
             the system  and its practicality to prioritize properly chemicals for further testing.

       d)     Sub-population Compartment:  The question of the need for a separate
             compartment to address sub-populations (e.g., developing children)  was
             addressed to the EDPS. Our conclusions supported the use of sub-populations as
             a criterion within the existing compartments already identified,  but not as a
             separate stand-alone compartment.

       e)     Use of the  Integrated Risk Information System (IRIS): The priority testing
             scheme relies on the use of several databases summarizing the environmental fate
             and effects of chemicals.  Several Members of the Committee expressed concern
             about problems with the validation of IRIS and other databases. Before placing
             heavy reliance on these computerized systems, users need to be aware of these
             validation problems and proceed with caution before incorporating these values
             unilaterally.

       f)     Exposure: The EDPS believes that consideration of the toxicological
             implications of exposure should include both dose and timing of exposure,
             particularly with respect to developmental or reproductive events. The current

-------
       scheme does not adequately cover the time aspect of exposure and this needs to
       be remedied before broad-scale application of the approach.

g)     Use of Animals and Routes of Exposure: We are concerned about the large
       number of animals that would be needed by the EDSTAC program.  The
       Subcommittee is cognizant of the essential role animals play in tests to detect
       endocrine disruption, and aware that there are no substitutes for tests currently
       available for the Tier 2 tests. This fact notwithstanding, the Agency has an
       obligation to conserve all resources in developing new testing protocols, and the
       use of animals in such tests poses both ethical and practical problems. In
       addition,  in this role of hazard assessment (as opposed to hazard definition)
       biologically relevant routes of exposure are indicated.. The current  EPA
       synthesis of the EDSTAC recommendations is inconsistent on the matter of route
       of exposure, but animal testing should be restricted to biologically relevant routes
       of exposure.

h)     Need for an Introductory Statement: The previous EDSTAC meeting
       suggested that the final document needed, as a introductory section, a description
       of the problem or the scientific or health-based reason for the EDSTAC program.
       The EDPS urges the EPA's EDSTAC team to include a description  of both the
       health and ecological problems associated with exposure to the endocrine
       disrupters and to show how the these findings relate to the program.

i)      Support  for Decisions: Decisions about which assays are selected,  and which
       protocols are adopted for those assays, should be supported with data that are
       generally available.

j)      Exceptions: Testing strategies will always have exceptions. Care should be taken
       to be aware of the imperfect nature of any future agreed strategy.

k)     Negative Control Agents:  There is a need to define and agree on some negative
       control agents for ED assay validation. Assay specificity will not be capable of
       assessment unless such agents can be made available for general study.

1)      Expanding the Universe of Agents: Developing massive amounts of screening
       information on a large universe of chemicals does not necessarily expedite the
       development of the appropriate scientific underpinning that the Agency needs to
       broaden this effort.  Consequently, the Subcommittee recommends that EPA

-------
             should not expand the set of agents until the Agency develops or adopts validated
             systems and can provide clear decision criteria.

      We appreciate the opportunity to review these proposed revisions, and look forward to
receiving your response to the issues raised.
                                       /Signed/
                                 Dr. Joan Daisey, Chair
                                 Science Advisory Board, and
                                 Co-chair, Endocrine Disrupter Screening
                                 Program Review Subcommittee
                                       /Signed/
                                 Dr. Gene McConnell, Co-chair
                                 Endocrine Disrupter Screening
                                 Program Review Subcommittee

-------
                                      NOTICE
  This report has been written as part of the activities of the Science Advisory Board, a public
 advisory group providing extramural scientific information and advice to the Administrator and
  other officials of the Environmental Protection Agency.  The Board is structured to provide
  balanced, expert assessment of scientific matters related to problems facing the Agency. This
 report has not been reviewed for approval by the Agency and, hence, the contents of this report
 do not necessarily represent the views and policies of the Environmental Protection Agency, nor
 of other agencies in the Executive Branch of the Federal government, nor does mention of trade
             names or commercial products constitute a recommendation for use.
Distribution and Availability: This Science Advisory Board report is provided to the EPA
Administrator, senior Agency management, appropriate program staff, interested members of the
public, and is posted on the SAB website (www.epa.gov/sab). Information on its availability is
also provided in the SAB's monthly newsletter {Happenings at the Science Advisory Board).
Additional copies and further information are available from the SAB Staff.

-------
                                     ABSTRACT
       The 1996 passage of the Food Quality Protection Act and amendments to the Safe
Drinking Water Act (SDWA) required EPA to develop a screening and testing strategy for
environmental endocrine disrupters.  The EPA subsequently asked the Science Advisory Board
(SAB) and the FIFRA Scientific Advisory Panel (SAP) to form a Joint Subcommittee to review a
set of scientific issues concerning the development of the Agency's endocrine disrupter
screening and testing program.  The review Subcommittee met on March 30-April 1, 1999, in
Arlington VA.

       The Charge was broad and complex, posing 18 major questions within four broad areas:
a) scope of the program; b) priority-setting; c) the high throughput pre-screening approach; and
d) the proposed endocrine disrupter screening program

       The Subcommittee recommended: a mid-course evaluation or optimization of the
screening; an initial focus on the methods development effort; the inclusion of more and better-
detailed case studies; the use of sub-populations as a criterion within the existing compartments
already identified, but not as a separate stand-alone compartment; making users aware of
validation problems in systems like IRIS; the inclusion of both dose and timing of exposure,
particularly with respect to developmental or reproductive events; minimizing the number of
animals needed for testing; inclusion of an introductory statement; support with data decisions
about which assays are selected, and which protocols are adopted for those assays, should be
with data; be aware of the imperfect nature of any future agreed strategy; define and agree on
some negative control agents for environmental disruption assay validation; do not expand the
set of agents until the Agency develops or adopts validated systems and can provide clear
decision criteria.

       Although the review identified several areas of concern, we wish to congratulate the
Agency for dealing effectively with an extraordinarily complex set of issues, many of which are
on the cutting edge of the relevant science.

KEYWORDS: endocrine; hormone; environmental endocrine disrupters; screening; assays;
environmental mixtures.
                                           11

-------
                                     ROSTER
            Joint SAB/SAP Subcommittee On Endocrine Disrupter Screening
                               March 30-April 1,1999

CO-CHAIRS
Dr. Joan Daisey, Lawrence Berkeley Laboratory, Indoor Environmental Program, Berkeley CA
Dr. Ernest McConnell, ToxPath, Raleigh NC

MEMBERS AND CONSULTANTS
Dr. John Ashby, Zeneca Corporation, Cheshire, United Kingdom
Dr. Richard Bull, Molecular Biosciences, Battelle Pacific Northwest Laboratories, Richland,
      WA
Dr. Charles Capen, Department of Veterinary Biosciences, The Ohio State University,
      Columbus, OH
Dr. Kenneth Davis, Ecological Research Center, University of Memphis, Memphis, TN
Dr. John Doull, Department of Pharmacology, Toxicology and Therapeutics, The University of
      Kansas Medical Center, Kansas City, KS
Dr. Paul M.D. Foster, Chemical Industry Institute of Toxicology, Research Triangle Park, NC
Dr. James Gibson, Dow AgroSciences, Indianapolis, IN
Dr. Philippe Grandjean, Institute of Community Health, Odense University, Denmark
Dr. Diane Henshel, School of Public and Environmental Affairs, Indiana University,
      Bloomington, IN
Dr. Alan Maki, Exxon Company, USA, Houston, TX
Dr. Genevieve Matanoski, School of Hygiene and Public Health, Johns Hopkins University,
      Baltimore, MD
Dr. Margaret McCarthy, Department of Physiology, School of Medicine, University of Maryland
      at Baltimore, Baltimore, MD
Dr. Michael McClain, University of Medicine and Dentistry of New Jersey, R.W. Johnson
      Medical School, Randolph, NJ
Dr. P.M. Anne McNabb, Department of Biology, Virginia Polytech. Institute & State University,
      Blacksburg, VA
Dr. Mary Anna Thrall, Department of Pathology, College of Veterinary Medicine & Biomedical
      Sciences, Colorado State University, Fort Collins, CO
Dr. John G. Vandenbergh, Department of Zoology, College of Agriculture and Life Sciences,
      North Carolina State University, Raleigh, NC
Dr. Tim Zacharewski, Department of Biochemistry, Michigan State University, East Lansing,
      MI

FEDERAL EXPERTS
Dr. Robert Chapin, National Institute of Environmental Health Sciences, Research Triangle
      Park, NC
Dr. James Hanson, National Cancer Center, Bethesda, MD
                                         in

-------
CO-DESIGNATED FEDERAL OFFICIAL (SAP)
Mr. Larry Dorsey, FIFRA Scientific Advisory Panel, Office of Prevention, Pesticides and Toxic
      Substances, Environmental Protection Agency, Washington, DC

CO-DESIGNATED FEDERAL OFFICIAL (SAB)
Mr. Samuel Rondberg, Science Advisory Board (1400), Environmental Protection Agency,
      Washington, DC
                                        IV

-------
                            TABLE OF CONTENTS

1 EXECUTIVE SUMMARY	1

2.  INTRODUCTION 	4
      2.1 Background	4
      2.2 Charge	4

3.  DETAILED RESPONSE TO THE CHARGE	11
      3.1 Scope of the Program  	11
             3.1.1 The proper Scope for the Endocrine Disrupter Screening Program  	11
             3.1.2 Use of the expanded set of Agents in the EDSP  	13
             3.1.3 Exemptions from the EDSP	14
             3.1.4 Mixtures	15
      3.2 Priority Setting  	16
             3.2.1 The Component-based Approach to Priority-setting	16
             3.2.2  The Relational Database and Priority Setting	17
      3.3 High Throughput Prescreening Approach  	19
             3.3.1 High Throughput Technology As A Tool for Priority Setting 	19
             3.3.2 High Throughput Technology As A Prescreening Device  	20
      3.4 The Proposed Endocrine Disrupter Screening Program	21
             3.4.1 The Two-phase Sorting Strategy	21
             3.4.2 Adequacy of the Screening Battery	22
             3.4.3 Adequacy of Thyroid Coverage	24
             3.4.4 In utero and In Ovo Screens and Single Dose Screening  	25
             3.4.5 Rigor of The Five Compartment Test Protocol Design	26
             3.4.6 Adequacy of Detection of Critical Endpoints in The EAT Systems	27
             3.4.7 Additional Screening for Agents Initially Found to Be Negative  	29
             3.4.8 Endocrine Disrupters and Hazard Assessment 	29
             3.4.9 Validation of The Proposed  Screens and Tests	29
             3.4.10 Subcommittee Recommendations to Help EPA Meet its Charge	31

4.  MAJOR FINDINGS AND RECOMMENDATIONS 	32

-------
                           1 EXECUTIVE SUMMARY

       In 1996, the passage of the Food Quality Protection Act (FQPA) and amendments to the
Safe Drinking Water Act (SDWA) required EPA to develop a screening and testing strategy for
endocrine disrupters within two years and implement the plan by August, 1999. EPA
established the Endocrine Disrupter Screening and Testing Advisory Committee (EDSTAC)
under the Federal Advisory Committee Act to advise the Agency on the screening and testing of
pesticides and other chemicals for their potential to disrupt the endocrine system.  Consequently,
the EPA asked the Science Advisory Board (SAB) and the FIFRA Scientific Advisory Panel
(SAP) to form a Joint Subcommittee to review a set of scientific issues being considered by the
Agency concerning the development of the Agency's endocrine disrupter screening and testing
program as required by the legislation noted above. This Joint Subcommittee (the Joint
Environmental Disrupter Screening Program (EDSP) review Subcommittee met on March 30-
April  1, 1999, and produced this report.

       The Charge was broad and complex, posing 18  major questions within four broad areas:
a) scope of the program; b)  priority-setting; c) the high throughput prescreening approach; and d)
the proposed endocrine disrupter screening program (the complete Charge is provided in section
2.2 of this report).

       The EDSP's detailed response to each element of the Charge is found in section 3 of the
report. The major issues and recommendations are:

       a)     Evaluating the Program: We find no provision for mid-course evaluation or
             optimization of the  process. The Agency is mandated to assemble and evaluate
             this proposed panel of tests and then to implement them, but a correlate
             responsibility is to make sure that what's being done is the best that can be.
             Although something looks fine on paper or in a small research setting, translating
             it into volume-screening mode may be quite another thing.  There was broad
             support among the Subcommittee for the concept that the Agency should convene
             a panel of independent scientists to review all the screening data for 50-100
             compounds, with an eye towards revising the process and eliminating those
             methods that don't work.

       b)     Mixtures Issues: The Subcommittee agreed that the initial focus of the methods
             development effort must focus necessarily on single compounds  and leave the
             question of testing of mixtures until accepted single-compound methods have
             been completed. The Subcommittee concluded that very promising methods
             already exist in the  field of ecotoxicology.  These include the Whole Effluent
             Testing (WET) and Toxicity Identification Evaluation (TIE) procedures
             developed by the Agency in concert with the Society for Environmental
             Toxicology and Chemistry.  Those methods have been developed to test effects of
             effluents and should have direct application to the Endocrine Disrupter Screening
             Program.


                                           1

-------
c)     Case Studies: The Subcommittee strongly encourages the Agency to include
       more and better-detailed case studies in the evolution of the priority-setting
       scheme. Case studies will enable a realistic test of the  scheme, checking
       sensitivity of the system and its practicality to prioritize chemicals for further
       testing.

d)     Sub-population Compartment: The question of the need for a separate
       compartment to address sub-populations (e.g., developing children) was
       addressed to the EDPS. Our conclusions supported the use of sub-populations as
       a criterion within the existing compartments already identified, but not as a
       separate stand-alone compartment.

e)     Use of the Integrated Risk Information System  (IRIS): The priority testing
       scheme relies on the use of several databases summarizing the environmental fate
       and effects of chemicals. Several Members of the Committee expressed concern
       that there are numerous problems with the validation of IRIS  and other databases.
       Before placing heavy reliance on these computerized systems, users need to be
       aware of these validation problems and proceed with caution before incorporating
       these values unilaterally.

f)     Exposure: The EDPS expressed concern that consideration of the toxicological
       implications of exposure should include both dose and timing of exposure,
       particularly with respect to developmental or reproductive events. The current
       scheme does not adequately cover the time aspect  of exposure and this needs to
       be remedied before broad-scale application of the approach.

g)     Animal Tests and Routes of Exposure: We are concerned about the large
       number of animals that would be needed by the EDSTAC program, and there is
       significant international concern on the proposed use of animals for such
       screening. The Subcommittee is cognizant of the essential role animals play in
       tests to detect endocrine disruption.  There are no substitutes for tests currently
       available for the Tier 1 and Tier 2 tests.  This notwithstanding, the Agency has an
       obligation to conserve all resources in developing  new testing protocols, and the
       use of animals in such tests poses both ethical and practical problems.  In this role
       of hazard assessment (as opposed to hazard definition) biologically relevant
       routes of exposure are indicated (oral gavage, diet, water, inhalation, skin
       painting).  At present, use of the subcutaneous injection or intraperitoneal
       injection routes are recommended in thequest of increasing assay sensitivity. In
       fact, irrespective of the outcome of this suggestion it should be noted that the
       current EPA synthesis of the EDSTAC recommendations is inconsistent on the
       matter of route of exposure, and that al animal testing should use only
       biologically relevant routes..

-------
       h)     Need for an Introductory Statement: The previous EDSTAC meeting
             suggested that the final document needed, as a introductory section, a description
             of the problem or the scientific or health-based reason for the EDSTAC program.
             The EDPS urges the EPA's EDSTAC team to include a description of both the
             health and ecological problems associated with exposure to the endocrine
             disrupters and to show how the program relates to these findings.

       i)     Support for Decisions: Decisions about which assays are selected, and which
             protocols are adopted for those assays, should be supported with data that are
             generally available.

       j)     Exceptions: Testing strategies will always have exceptions. Care should be taken
             to be aware of the imperfect nature of any future agreed strategy.

       k)     Negative Control Agents: There is a need to define and agree on some negative
             control agents for ED assay validation.  Assay specificity will not be capable of
             assessment unless reliable agents can be made available for general study.

       1)     Expanding the Universe of Agents: Developing massive amounts of screening
             information on a large universe of chemicals does not necessarily expedite the
             development of the appropriate scientific underpinning that the Agency needs to
             broaden this effort. Consequently, the Subcommittee recommends that EPA
             should not expand the set of agents until the Agency develops or adopts validated
             systems and can provide clear decision criteria.

       Although the review identified several areas of concern, and the EDSP has provided
recommendations to improve EPA's planned program, we wish to  congratulate the Agency for
dealing effectively with an extraordinarily complex set of issues, many of which are on the
cutting edge of the relevant science.

-------
                               2. INTRODUCTION

2.1 Background

       Chemicals which may interfere with endocrine system functioning (endocrine disrupters)
have concerned the U.S. Environmental Protection Agency (EPA) for some time. Such
chemicals have the potential to impact human and wildlife populations. A variety of human
health and ecological effects have been attributed to endocrine disrupters.

       In 1996, the passage of the Food Quality Protection Act (FQPA) and amendments to the
Safe Drinking Water Act (SDWA) required EPA to develop a screening and testing strategy for
endocrine disrupters within two years and implement the plan by August, 1999. The legislation
cites the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) and the Toxic Substances
Control Act (TSCA) as the two statutes under which EPA should implement an endocrine
screening and testing strategy. EPA established the Endocrine Disrupter Screening and Testing
Advisory Committee (EDSTAC) under the Federal  Advisory Committee Act to advise the
Agency on the screening and testing of pesticides and chemicals for their potential to disrupt the
endocrine system.

       Consequently, the Science Advisory Board (SAB) and the FIFRA Scientific Advisory
Panel (SAP) were asked to form a Joint Subcommittee to review a set of scientific issues being
considered by the Agency  concerning the development of the Agency's endocrine disrupter
screening and testing program as required by the legislation noted above.

2.2 Charge

       The specific issues to be addressed by the Joint Subcommittee are:

a) Scope of the Program

       1)     The amendments to the Food Quality Protection Act (FQPA) and the Safe
              Drinking Water Act (SDWA) mandate or support the development of a screening
              program that will determine whether pesticides and certain drinking water source
              contaminants "may have an effect in humans that is similar to an effect produced
              by a naturally-occurring estrogen, or other such endocrine effect as the
              Administrator may designate." Very early in its deliberations,  EPA's Endocrine
              Disrupter Screening and Testing Advisory Committee (EDSTAC) determined
              that there was both a strong scientific basis and feasibility, considering time and
              resource constraints, to expand the scope of the screening program to include the
              androgen- and  thyroid-hormone systems, and to include evaluation of the
              potential impact on wildlife as well as on human health. EPA agrees and is
              developing  a screening program which incorporates these modifications. Does
              the Joint Subcommittee agree that this expanded scope is appropriate to
              serve as the starting point for the Endocrine Disrupter Screening Program

-------
       (EDSP), given the understanding that the framework for the Program can
       support for further expansion at a later date?

2)     The FQPA and SDWA identify a universe of substances that should be evaluated
       in a an EDSP. EDSTAC noted that there exist many other substances in addition
       to pesticides and certain drinking water source contaminants that may exhibit
       endocrine-disrupting potential. They recommended that the "candidate pool" for
       the EDSP include substances on the Toxic Substances Control Act (TSCA)
       Inventory, certain complex environmental mixtures as well  as non-pesticide food
       additives, cosmetics and nutritional supplements. EPA agrees that there are
       substances in addition to pesticides and certain drinking water source
       contaminants that warrant consideration for inclusion in the EDSP..  Does the
       Joint Committee agree that this expanded universe of substances should be
       included in the EDSP process, at a minimum in the priority-setting phase,
       and continuing on if a potential for concern is identified?

3)     FQPA contains a provision which would exempt from the EDSP "any biological
       substance or other substance if the Administrator determines that the substance is
       anticipated not to produce any effect in humans similar to an effect produced by a
       naturally-occurring estrogen" or,  presumably, "such other endocrine effect as the
       Administrator may designate." EPA has identified some chemical categories that
       may be candidates for exemption. Examples include certain polymers with a
       number average molecular weight (NAMW) greater than 1000 daltons, certain
       List 4 pesticide inerts such as cookie crumbs, strong mineral acids and bases,
       which are most likely to interact with tissue at the portal of entry giving rise to
       localized lesions rather than systemic effects, certain biopesticides such as plant
       pesticides or microbials or non-chemical pesticides such as  parasitic wasps.  Does
       the Joint Committee  agree that there are categories of pesticides and other
       substances that should be exempt from the EDSP? In addition to the
       examples noted here, are there additional categories  that should be
       considered for exemption?

4)     EDSTAC concluded, and EPA agrees, that there are important complex
       environmental mixtures that deserve inclusion in the EDSP. EDSTAC
       recommended that EPA include in the EDSP representative mixtures to which
       large or identifiable key segments of the population (e.g. children) are exposed.
       They suggested that high-priority mixture categories include: Chemicals in breast
       milk, phytoestrogens in soy-based infant formulas, mixtures commonly found at
       Superfund sites, common pesticide/fertilizer mixtures found in ground and
       surface water, disinfection byproducts, and gasoline. EPA proposes to screen and
       test (if appropriate) one representative mixture from each category, after it
       confirms that the screening and testing components of the EDSP are satisfactory
       for the handling of single substances.

-------
             A)     Is the proposal a reasonable way to address the practicality of
                    screening and testing mixtures?
             B)     Are the six categories of mixtures the most appropriate to address
                    first?
             C)     Are there other mixture categories that should be included in addition
                    to, or instead of those identified (Note:  During the May Consultation,
                    it was suggested that mixtures found in fish tissue, benthic sites and
                    eggs of Great Lakes birds should replace gasoline as a priority
                    mixture).
             D)     Can/should standardized representative mixtures be developed? If so,
                    how should the chemical combinations, ratios, and doses be selected
                    for mixtures?
             E)     If a mixture is positive in the screening tier, should the whole mixture
                    be tested in the testing tier or should only the active component(s) in
                    the screen(s) be tested in the second tier?

b) Priority-setting

       1)     EDSTAC recommended a component-based approach to priority-setting. EPA
             agrees that this is the appropriate framework. Under this approach, EPA will
             group chemicals into sets, based on the existence of factual information in a given
             area.  Thus, priority ranking can be made fairly among substances, i.e., chemicals
             will compete for priority with others on the basis of comparable data and will not
             be assigned lower priority for lack of information.  Are these principles and the
             component-based approach to priority setting reasonable? Are there other
             approaches that would be more useful?

       2)     EPA is developing a relational database to assist in developing priorities for
             screening. The relational database is intended to import existing data and allow
             its synthesis, as well as the estimation of certain parameters through modeling.
             The relational database was considered to have great value in helping to identify
             the specific compartments under the EDSTAC's  component-based priority setting
             approach.  The database will also be helpful in selecting chemicals for the first
             and subsequent rounds of screening  Would the Joint Subcommittee comment
             on the approach and provide additional insights to improve the content of
             the relational database or its implementation?

c) High Throughput Prescreening Approach

       1)     EDSTAC recommended, and EPA proposes to implement a priority setting
             strategy that includes initial sorting based on an examination of existing
             information. This initial sorting strategy leads to four possible  outcomes: i)
             polymers;  ii) chemicals with sufficient data to proceed to testing; iii)chemicals
             with sufficient data to proceed to hazard assessment; and iv) chemicals with

-------
             insufficient data, which presumably, would go into the screening phase. EPA
             anticipates that a large number of substances will end up in category iv-chemicals
             with insufficient data. To provide at least a minimum number of biological data to
             assist in the sorting process, EPA proposes to conduct High Throughput
             Prescreening on a significant number of substances (perhaps, as many as 15,000),
             using in vitro assay systems incorporating transcriptional activation or reporter
             gene systems for the estrogen-, androgen- and thyroid-hormone systems.

                    A)    On the assumption that the technology can be shown to be
                          applicable to the large number and wide range of chemical
                          substances under consideration, and the limited relevant test data
                          which are available for many industrial chemicals, is this a
                          reasonable approach and sorting strategy to support priority
                          setting?

                    B)    EPA has been funding a pilot study, using about 80 chemicals, to
                          determine the applicability of the high throughput technology in a
                          prescreening component of the EDSP. Based upon your review of
                          the data developed to date, does the Joint Subcommittee
                          believe that this technique can be used as a prescreening
                          device? If not, what modifications/improvements must be
                          made in order to assure its usefulness?

d) The Proposed Endocrine Disrupter Screening Program

       1)     EPA is proposing to  develop and implement an Endocrine Disruption Screening
             Program that consists of two phases. The first phase is screening, currently
             consisting of eight components. The second phase is testing, currently envisioned
             to have one to five components, depending upon the  need for identifying effects
             in various sectors of the animal kingdom. Is it reasonable and appropriate to
             develop and implement a two-phase program, the first phase focused on
             identifying a substance's potential to interact with one or more of the three
             hormone systems, the second phase to characterize the effects of concern that
             interaction with these hormone systems might elicit?

       2)     EDSTAC recommended, and EPA proposes to implement, a screening battery
             consisting of eight assays, three in vitro and five in vivo, to address estrogen-,
             androgen- and thyroid-hormone system effects. At the time (a year ago or so) and
             continuing to today, based upon our knowledge  of the state-of-the-science, the
             Agency believed that these eight assays, once validated and standardized, would
             detect all substances currently known to interact with the three hormone systems
             to be covered in the Program. Does the Joint Subcommittee agree with this
             assessment? If not, what changes should be made  in the battery to assure the
             identification of substances of potential concern?

-------
3)     Interaction with a receptor is the principal or key mechanism by which substances
       exert their effects on the estrogen- and androgen-hormone systems. This appears
       NOT to be the case for the thyroid-hormone system. In light of this, does the
       Joint Subcommittee believe that there is adequate coverage of the thyroid
       provided in the proposed screening battery? If not, what modifications
       should/could be made?

4)     EPA would prefer to have a screening battery which included assays containing
       an in utero or in ovo exposure component, given its great and continuing concern
       about the potential for effects on the developing vertebrate organism. At the time
       the proposed screening battery was being assembled, EPA was not aware of the
       existence of any such screens. Is the Joint Subcommittee aware of any such
       assays that may exist or are under development that could supplant or
       complement one or more components of the proposed screening battery?
5)     EDSTAC recommended, and EPA would prefer, for efficiency and cost reasons
       given the numbers of substances that may be involved in the EDSP, to conduct
       each in vivo screening assay using only one dose, with the appropriate use of
       range finding studies and other information to inform dose selection. Does the
       Joint Subcommittee agree with this approach, and if not, what suggestions
       would it have to modify the approach, keeping pace, volume, cost and
       efficiency in mind? What would be the public health consequences of these
       false negatives? (Note: At the May consultation, some members raised concern
       about relying on a single dose and suggested that a minimum of two doses, and
       perhaps even three, be used to ensure that the screens do not yield false negative
       results.lt has also been suggested, elsewhere, that this issue could/should be
       solved during validation/standardization.).

6)     EDSTAC recommended, and EPA is proposing, a testing phase in the EDSP
       which could have as many as five components (i.e. covering mammals, birds,
       fish, invertebrates and amphibians). Each test would be designed to delineate the
       dose-response relationships of effects of concern for chemicals which yielded
       positive results during the screening phase. The testing protocols to be used are
       either upgrades or modifications of existing guidelines, except for the amphibian.
       In this case, a protocol is being developed de novo. Does the Joint
       Subcommittee believe that these test protocol designs will provide sufficient
       rigor to identify effects of concern and establish their dose responses for
       disruption in the estrogen-, androgen- and/or thyroid-hormone systems?

7)     There could be circumstances in which substances bypass the screening phase,
       and go directly into the testing phase. EPA is proposing for those cases that the
       chemical under evaluation be tested in all five tests. Does the Joint

-------
       Subcommittee believe that the tests in the testing phase will be adequate to
       detect all known critical endpoints in the estrogen-, androgen-, and thyroid-
       hormone systems? If not, what modifications should be made?

8)     If the results of any of the testing phase tests are negative, what, if any,
       additional screening or testing should be conducted to assure that the
       chemical is not an endocrine disrupter in the estrogen-, androgen- or thyroid
       hormone systems of that sector of the  animal kingdom?

9)     Testing phase tests will identify effects of concern that are the consequence of
       endocrine disruption. They may also identify effects of concern that are not the
       consequence of endocrine disruption. Thus, it may not be possible to determine if
       a substance is an endocrine disrupter if it has not been subjected to some or all
       components of the screening battery. Is it important to  be able to identify
       substances as endocrine disruptors from the standpoint of conducting a
       hazard assessment. If so, why? If not,  why not?

10)    EPA is proposing a validation program in which the maximum validation effort
       will consist of conducting each assay in  three laboratories. EPA believes that
       there currently is a wide variation in  the state  of validation of each of the
       proposed screens and tests, and that the validation efforts should be tailored for
       each assay/test accordingly. EPA plans to focus first on the validation of the
       mammalian assays as they are both better developed than the non-mammalian
       assays and are more directly relevant to  meeting the FQPA and SDWA mandates
       for a screening program for potential human health impacts.  EPA's preliminary
       assessment of the work needed is as follows:

             The uterotrophic assay requires the development of a standardized
             protocol but may need little or no additional laboratory/protocol
             development effort since the assay has been in extensive use for many
             years.

             The Hershberger assay may require some, but not much, additional
             lab oratory/protocol development in addition to standardization.

             The pubertal male and pubertal female assays need some additional
             developmental work and will require the full regime  of interlaboratory
             validation.
             The mammalian two-generation reproduction test will require limited
             testing in one laboratory to validate the new endpoints since the basic
             protocol is already considered to be valid.

-------
             Both of the non-mammalian screens and some of the non-mammalian tests
             will require the full validation regime; some will require further pre-
             validation development (e.g. amphibian test).

      The mammalian two-generation test will require limited testing in one laboratory
      to validate the new endpoints since the basic protocol is already considered to be
      valid. All of the non-mammalian assays will require the full validation regime
      and some will require further pre-validation development.  Does the Joint
      Committee agree with the Agency's assessment of the current status of the
      screens and tests? If not, what is the Joint Committee's own assessment of
      any screen or test which differs from EPA's, and what is the basis for your
      opinion?

11)   Does the Joint Subcommittee have any other suggestions or
      recommendations that would help EPA meet its charge?
                                   10

-------
                3.  DETAILED RESPONSE TO THE CHARGE

3.1 Scope of the Program

       The amendments (1996) to the Food Quality Protection Act (FQPA) and the Safe
Drinking Water Act (SDWA) mandate or support the development of a screening program that
will determine whether pesticides and certain drinking water source contaminants "may have an
effect in humans that is similar to an effect produced by a naturally-occurring estrogen, or other
such endocrine effect as the Administrator may designate."  Very early in its deliberations,
EPA's Endocrine Disrupter Screening and Testing Advisory Committee (EDSTAC) determined
that there was scientific basis for expanding the scope of the screening program to include the
androgen- and thyroid-hormone systems, and to include evaluation of the potential impact on
wildlife as well as on  human health (EDSTAC, 1999). EPA agrees and is developing a
screening program which incorporates these modifications.  Sections 3.4.1 through 3.4.4 address
significant issues in designing the screening program

  3.1.1 The proper Scope for the Endocrine Disrupter Screening Program (EDSP)

       The initial element of the Charge (a) (1) for this review asks if the Joint Subcommittee
agrees that this expanded scope is appropriate to  serve as the starting point for the Endocrine
Disrupter Screening Program (EDSP), given the understanding that the framework for the
Program can support further expansion at a later date.

       Expansion of the scope of the program under the FQPA and the SDWA from simply
estrogen-like effects on human health to androgen and thyroid active compounds is reasonable.
This expansion raises the sights above that demanded by the authorization language, but is
clearly within the guidance provided.

       The EDSTAC review of endocrine effects recognized that issues related to endocrine
disruption are even broader than the three categories identified (EDSTAC, 1998). However,
there will be significant technical difficulties in addressing estrogen, androgen and thyroid active
compounds. Further expansion at this stage in the development of the program would have
created an unmanageable task. Nevertheless, it should be recognized that modification of the
activity of other hormonal systems can be as important, or perhaps more important, than the
systems identified.  Moreover, the technology to begin integrating these systems is now
becoming available. Overall, the proposed framework should enable the agency to integrate
knowledge  of these systems and assay techniques as they mature.

       The expansion of the concerns to the broader environmental concerns over endocrine
disruption is not only  appropriate, but crucial.  Modifications in reproductive and developmental
processes in the environment have been related to endocrine disruption. Experience has shown
that effects  on wildlife and ecosystems are seen before there is any significant impact in humans,
particularly for chemicals that bioaccumulate.  It  is unfortunate that effects on populations in
actual ecosystems cannot be practically included  in the program because in some cases these

                                          11

-------
have been most sensitive measures for agents of this type. Consequently, there are substantive
reasons for expanding the scope to non-human species.  As with the expansion to the androgen
and thyroid hormone systems, however, it is important to recognize that this further complicates
implementation.

       The Joint Subcommittee expressed considerable concern with respect to whether the
program is focused on the critical endpoints.  The focus on the endocrine system is very much a
focus on mechanism(s) that do not necessarily relate in a holistic manner to the adverse health
and environmental effects that are of ultimate concern. In general, the major concern for
endocrine disrupters are effects on normal reproduction and development that also extend to the
induction of certain kinds of cancer.  Deviations in endocrine levels in either direction can have
biological consequences, although the consequences can be beneficial as well as detrimental in
some cases.  Each alteration needs to viewed in terms of the totality of the data."  The endocrine
disruptions, as characterized by interactions with these three hormone systems, are not the only
way in which such effects can be  produced.  As a consequence, there is some inversion of the
normal decision logic that makes  it crucial that a concise working definition of endocrine
disruption be developed. The EDSP is being developed on the apparent assumption that most
compounds that affect the three endocrine systems identified are likely to be of toxicological
concern.  Secondary testing will be triggered based on broad-based screening. However,
significantly more thought must be given to identifying quantitative "trigger points"  as well as
the question of whether the response is positive or not. These might include an exclusion  based
on excessive doses required to induce the effect or a requirement for a minimum level of
response in the test system. The most important issue is to identify the magnitude and perhaps
the duration of a response. Most  of this difficulty could be taken care of by crafting a more
concise definition of what constitutes endocrine disruption.  This could possibly be done by
adding a phrase to the endocrine disruption definition that states "... reproducible effects on any
endocrine sensitive system impairing successful reproduction and development..." " Broader
definitions could also raise issues relating to carcinogenesis or the  ability to maintain
homoeostatic balance in response to biotic, chemical and physical stresses.  Extension of such
definitions inevitably leads to complications in the application of the test scheme.

       The Joint Subcommittee also identified a need to establish an on-going review of
progress of the EDSP. There are  both methodological and interpretative issues involved.  The
methodological issues are more straightforward, involving a process for reviewing new
screening and testing methods for incorporation or substitution for current methods.  The
interpretative problems have more to do with how the data are going to be applied to improving
environmental protection.  EPA put in a heroic effort in reviewing  the available methodology
and putting together a framework based on that methodology.  They have addressed many of the
interpretative problems that evolve from the current structure in the EDSTAC report (EDSTAC,
1998; Federal Register,  1998). It is not as clear how these processes are to be moved forward in
a rational way.  A regular plan to  revisit both the methodological and interpretative issues  can be
used as a vehicle to stimulate that progress.
                                           12

-------
  3.1.2 Use of the expanded set of Agents in the EDSP

       The FQPA and SDWA identify a universe of substances that should be evaluated in an
EDSP. The EDSTAC report (1998) noted that there are many other substances in addition to
pesticides and certain drinking water source contaminants that may exhibit endocrine-disrupting
potential. They recommended that the "candidate pool" for the EDSP include substances on the
Toxic Substances Control Act (TSCA) Inventory, certain complex environmental mixtures as
well as non-pesticide food additives,  cosmetics and nutritional supplements. EPA agrees that
there are substances in addition to pesticides and certain drinking water source contaminants that
warrant consideration for inclusion in the EDSP. Charge element (a) (2) asked the Joint
Committee to comment on the use of an expanded universe of substances in the EDSP process,
both in the priority-setting phase, and continuing on to later phases if a potential for concern is
identified.

       Expansion of the scope of the mandated efforts under the FQPA and  the SDWA
makes sense only if EPA has developed or adopted validated systems and can apply clear
decision criteria for expanding (or  not expanding) the effort. Under such circumstances
there would little reason to exclude additional chemicals from consideration.  At  the present
time, however, this particular expansion seems to add a level of complexity that may be
counterproductive. The Subcommittee's concerns  arise  from considering what the underlying
objective(s) of a screening and testing program are in the environmental programs administered
by the EPA. The ultimate goal  is to protect health  and the environment from adverse effects. In
one sense, the Agency is to be congratulated for attempting to focus on the endocrine disrupter
issue, because it does move them in the direction of identifying and perhaps understanding more
subtle environmental hazards.  However, if the activity loses its  connectivity to recognized or
newly described forms of compromised health and ecological effects, it will difficult for the
program to establish a solid rational basis. Expanded consideration of diverse types of candidate
"endocrine disrupting" chemicals has the potential of getting ahead of our knowledge of the risks
actually represented by  screening and testing of large numbers of chemicals.  The interest in how
modified cellular function leads to adverse effects is a necessary step involved in improving
hazard identification and risk assessment. These advancements  must be built on  careful
development of the science that establishes clear causal  associations between new testing tools
and adverse impacts. The precedent that could be set by pursuing mechanisms without regard to
effect can not only greatly increase the expense of testing of products, it can compromise
confidence in the screening program  that will eventually have to include all aspects of endocrine
functions represented.  The application of these tests to a wide variety of chemicals (as many as
15,000) has the potential for building up a data base, but not necessarily increasing our
knowledge about the significance of any effects that are observed.  Consequently, we believe
that the most important  area on which to focus resources is that intended to improve our ability
to establish clear causal connections between exposure and effects in target organisms, rather
then expand the number of agents in the screening  program.

       This point may be illustrated by considering some multifaceted problems that could
evolve from the EDSP in the form of questions  some of which are actually discussed in the
                                           13

-------
report. Is the intent to focus entirely on direct effects mediated through recognized estrogen,
androgen and thyroid receptors? How will indirect effects on the endocrine receptor be
identified?  Are indirectly mediated effects on the estrogenic, androgenic and thyroid systems to
be handled differently from a risk assessment standpoint?  The High Throughput Prescreening
System (HTPS) is unlikely to detect indirect effects. However, this may provide the only data
that is available on most of those chemicals. Are the ones that are "negative" then neglected?
Their ability to harm health and the environment has not been evaluated.  Reporter systems can
only dependably detect those interactions that are mediated at the level of the receptor.  Many
results from higher level tests are likely to arise from these indirect effects. Does the inability of
seeing the effect on a cellular system containing a reporter system linked to the hormone
response element provide a rationale for dismissing endocrine disruption as a mechanism that is
likely to be active at low doses? What if the steroid hormone response is mimicked  by a
membrane hormone effect, as has been demonstrated with Epidermal Growth Factor? Where
does mechanism of action fit into this process and how does it modify perceived risks at low
dose?

       On the other hand, there is the clear long-term advantage of beginning to focus on the
association of adverse health impacts with modifications in hormonal control mechanisms. If an
adverse effect of a chemical can be clearly associated with an endocrine effect (or any other
biochemical/molecular response that can be clearly related to adverse effects), the dose response
relationships can be explored across test systems and into the impacted species. Thus, the
impacts of low dose exposure to environmental agents can be explored in greater detail and with
greater understanding.  This is true, however, only if the health impacts that are associated with
changes are understood in fairly explicit ways.  Developing massive amounts of screening
information on a large universe of chemicals does not necessarily expedite the development of
the appropriate scientific underpinning that the Agency needs to broaden this effort.
Consequently, the Subcommittee recommends that EPA should not expand the set of agents until
it has validated systems and can apply clear decision criteria.

  3.1.3 Exemptions from the EDSP

       The 1996 Food Quality Protection Act (FQPA) contains a provision which would exempt
from the  EDSP "...any biological substance or other substance if the Administrator determines
that the substance is anticipated not to produce any effect in humans similar to an effect
produced by a naturally occurring estrogen" or, presumably, "...such other endocrine effect as
the Administrator may designate." EPA has identified some chemical categories that may be
candidates for exemption.  Examples include certain polymers with a number average molecular
weight (NAMW) greater than  1000 daltons, certain List 4 pesticide inert  substances (such as
cookie crumbs, strong mineral acids and bases), which are most likely to interact with tissue  at
the portal of entry giving rise to localized lesions rather than systemic effects, certain bio-
pesticides such as plant pesticides or microbials or non-chemical pesticides such as parasitic
wasps. In Charge element (a) (3), EPA  asked if the Joint Subcommittee agreed with the
Agency's position that there are categories of pesticides and other substances that should be
                                           14

-------
exempt from the EDSP.  The Subcommittee was also asked to identify any additional categories
that should be considered for exemption.

       The Subcommittee believes that there are clearly categories of chemicals that should be
exempt. However, the boundaries between those compounds that would be exempted and those
that would not must be carefully considered.  The selection of 1000 daltons as a cutoff for
polymers based on nominal molecular weight appears to have precedent under TSCA, but the
scientific justification for this limit was not provided. Many compounds with molecular weight
approaching 500 are known to have biological activity. Presumably this precedent will allow
polymers with as much as 10% of their total mass to be 500 daltons or less. A more concise
statement of the scientific reason for taking the specific action on polymers would have been
useful. Clearly, there would be rationales for dismissing other types of chemicals (e.g. amino
acids, fatty acids, sugars that are part of normal diets) from the EDSP.

       The Joint Subcommittee did not respond to the second issue of this Charge element. We
did not think it appropriate for the Subcommittee to identify additional classes of chemicals for
exemption. The Subcommittee suggests that the Agency consider handling of exemptions
through a rule making process that is transparent and open to public comment.

  3.1.4 Mixtures

       EPA recognizes that there are important complex environmental mixtures that deserve
inclusion in the EDSP.  Consequently, EPA plans to include in the EDSP representative mixtures
to which large or identifiable key segments of the population are exposed. Initial choices for
these high-priority mixture categories include: chemicals in breast milk; phytoestrogens in soy-
based infant formulas; mixtures commonly found at Superfund sites; common pesticide/fertilizer
mixtures found in ground and surface water; disinfection byproducts; and  gasoline. EPA
proposes to screen and test (if appropriate) one representative mixture from each category, after
it confirms that the screening and testing components of the EDSP are satisfactory for the
handling of single substances.

       For this review, EPA  asked (in Charge elements (a) (4) (A-E)) if the proposal is a
reasonable way to address the practicality of screening and testing mixtures; if the chosen six
categories are the right ones to address first; if there other mixture categories that should be
included in addition to, or instead of those identified; and if standardized representative mixtures
be developed. The Agency also sought advice on dealing with those mixtures found to  be
positive in the screening tier, asking if the whole mixture should be tested in the testing tier or
only the active component(s) identified in the screens(s).

       The recommendations for handling mixtures are outlined in Section VII of Chapter 4 of
the EDSTAC Final Report (EDSTAC, 1998) and were discussed by EPA staff and were the
subject of several comments from the public during the Subcommittee's public meeting.
Although there was general recognition of the key importance of mixtures as a part of the overall
EDSP, there were concerns about the selection  process, the experimental design for testing the
                                           15

-------
mixtures, and the ability of the Agency to evaluate and interpret the results of the studies and to
effectively communicate this information to the risk managers and the public. There was a
consensus by the Subcommittee that the mixtures section of the EDSTAC document needed to
be re-worked and there were several public comments recommending that mixtures not be
included in the program.

       The Subcommittee suggests a compromise proposal: delay starting the mixtures testing
program until most of the single agent testing was completed.  This would have two advantages;
first it would provide a more extensive data base to use in selecting mixtures for testing; and
second it would enable the Agency to benefit from some of the current efforts underway with
pesticides (within EPA and by outside research groups) to improve our ability to define and test
mixtures.  The Subcommittee recognizes that the Agency is currently testing some mixtures
(wastewater, cholinesterase inhibiting pesticides etc.) and that relatively little effort would be
required to incorporate these into EDSP. We are also aware of studies that compare the effects
of mixtures having independent actions with those having identical actions as a way to
characterize the risk assessment of mixtures (Yang et a/., 1998; Feron et. al., 1995;
NAS/NRC/COT, 1989). Similar approaches could be used to standardize or characterize
mixtures for testing in the EDSTAC program and would provide more interpretable results than
those proposed in this report.

       The EPA's final question in the mixtures section addressed phase 2 testing of mixtures
and/or the components.  The Subcommittee believes that would be prudent to test both the
mixture and its components.

3.2 Priority Setting

  3.2.1  The Component-based Approach to Priority-setting

       The EDSTAC report (1998) recommended a component-based approach to priority-
setting. EPA agrees that this is the appropriate framework, and plans to group chemicals into
sets, based on the existence of factual information in a given area. Thus, comparisons can be
made between like substances (i.e., chemicals will  compete for priority with others on the basis
of comparable data and will not be assigned lower priority for lack of information).  In Charge
element (b) (1), EPA asked the Joint Subcommittee to comment on the principles of the
component-based approach to priority setting, and to suggest any other approaches that would be
more useful.

       The Subcommittee finds that this approach1 is supportable when  ranking is based on both
effect and exposure data following guidance in NRC and EPA risk assessment literature (NRC,
       lrThis approach first assigns environmental toxicants into four categories (based on available data): a)
Specially targeted priorities; b) Exposure-related information; c) Effects-related information; and d) Integrated
Effects and Exposure. Components (or sets) are defined within each category, into which agents are assigned on the
basis of exposure and/or effects information. The individual agents are then ranked within each component
from highest concern to lowest concern to set priorities for eventual Tier one screening.

                                            16

-------
 1994, 1983; EPA 1997, 1992, 1986). The greatest weight should be given to chemicals for
which we have data that indicates actual human or environmental exposure and effects. Lower
weight should be given to agents for which the data are indicative of probable exposure (in food
or drinking water) or probable effects (from well conducted animal studies). The lowest weight
and priority ranking should be given to chemicals for which the data are indicative of possible
exposure (based on Toxics Release Inventory data or known high production volume) or
possible effects from (in vitro research or from Structure Activity Research). The Subcommittee
supports the nomination concept (i.e., the process of identifying ("nominating")
probable/possible exposure or probable/possible effects as noted above by citizens who are
disproportionately exposed because of the group or community to which they belong, or because
an ecosystem is disproportionately exposed (EPA, 1999)) but advises the Agency that the
process needs further definition and that no unsubstantiated claims be allowed.

       The Subcommittee is concerned that the prioritization process is not as "transparent" as it
needs to be for public understanding. Also, concern was expressed at the public meeting that
health care professionals (both personal health clinicians and public health) may not understand
EPA's process, intent, and implementation methods, especially in the context of their own work.
Communication at this level needs to improve,  or the results will be valueless.  Additional
concern was expressed that the process appears to have undue emphasis on chemistry and
toxicology, with

less clear emphasis on health effects. In summary, prioritization should be based on a sound
scientific basis.

       Other than the comments provided above, the Subcommittee has no suggestions for
alternative approaches.

  3.2.2  The Relational Database and Priority  Setting

       EPA is developing a relational database2 to assist in developing priorities  for screening.
This database is intended to import and synthesize existing data, allowing EPA to estimate
certain parameters through modeling. It is expected to have great value in helping to identify the
specific components under the EDSTAC's component -based priority setting approach. The
database will also be helpful in selecting chemicals for the first and subsequent rounds of
screening.  The EPA asked the Joint Subcommittee to comment on this approach, and to provide
advice to improve the content of the relational database or its implementation (Charge  element
(b) (2)).

       The Subcommittee believes that the proposed relational database shows strong  promise
of being a useful tool, as long as it does prove to be truly relational This step is the very core of
        Such a database links (or relates) all data elements to each other, allowing a broad range of questions to be
answered. In this case, it would provide a means of relating environmental exposure data with lexicological effects
information.

                                            17

-------
the risk assessment process and will ultimately provide the most valuable guidance in the priority
setting approach.

       When designing a relational database, it is important to consider relationships that
accomplish more than simple indexing of component data, so that the database is not just a
resource to access information on a specific agent. Although this capability is valuable in itself,
one needs to consider biologic relationships in modeling the database, so that it can become a
more active investigational tool.  For instance, the data need to be collected and organized in a
way that can respond to our growing knowledge in gene sequence at specific loci and
implications for health and disease. Likewise, the data need to be organized so that
developmental gene networks and other biological hierarchies can be reflected appropriately in
the database.  We are moving beyond single major risk factors for particular outcomes and into
the complex gene-gene and gene-environment relationships which characterize common diseases
(cancer, heart disease, behavioral disorders, aging etc.) This is a very difficult challenge for
genetic epidemiologists and has enormous design and analytical ramifications. Since the type of
information provided by the database will affect greatly the interpretation of available data, its
design will inevitably affect prioritization.

       The database needs to reflect knowledge throughout the specified organism's life cycle,
and should be able both to examine longitudinal developmental changes within a system, and to
make cross-sectional comparisons across the organism. The goal is to facilitate creation of a
biologically plausible chain of causal inference.  The database also needs to be prepared to deal
with a rapidly growing genetic database on variation in endocrine system-related genes.

       The database should be designed so it can be readily interfaced with human health
surveillance data on disorders such as birth defects and cancer.  The National Institutes of
Health, the Center for Disease Control, and other agencies are working with states to strengthen
these surveillance systems, and in some states (e.g. Iowa) these  systems have been constructed
together with environmental quality databases.  These have been used for aggregate (ecologic)
studies of health outcome risk factors. It is expected that such capabilities will be substantially
expanded  in the next few years with corresponding implications for priority setting.

       Priority setting should also address those persons or organisms found to be "most
susceptible", but not be limited to this subpopulation alone. There needs also to be focus on
population disease burden.  Individual rare genes may be major risk factors for a few persons,
but may contribute less to the burden  of a disease in a population than do "minor risk factor
genes" which are common in the population.

       Finally, there are two important problem which must be considered in using the relational
database as proposed. The Subcommittee expressed concern  that the lack of effects data on the
universe of chemicals currently in commercial use will lead to a relational data base that only
identifies known problem chemicals that are already well studied. The Subcommittee
encouraged the development and use of new techniques including quantitative structural activity
relationships, molecular modeling, and androgen binding, in addition to solubility (Kow) and

                                           18

-------
other measures to help identify the bio-available, potentially active compounds for further testing
in the EDSP). In addition, we are concerned by potential problems caused by uninformed use of
data from EPA's Integrated Risk Information System (IRIS). The priority testing scheme relies
on the use of several databases (including the IRIS) summarizing the environmental fate and
effects of chemicals.  Caution was expressed by several members of the Committee that there are
numerous problems with the validation of IRIS (and other databases as well). Before placing
heavy reliance on these computerized systems, users need to be aware of these validation
problems and proceed with caution before incorporating these values on their face value.

3.3 High Throughput Prescreening Approach

       Based on recommendations from the EDSTAC (1998), EPA proposes to implement a
priority setting strategy that includes an initial sorting, based on an examination of existing
information. This initial sorting strategy leads to four possible classifications: i) polymers; ii)
chemicals with sufficient data to proceed to testing; iii) chemicals with sufficient data to proceed
to hazard assessment; and iv) chemicals with insufficient data, which presumably, would go into
the screening phase. EPA anticipates that a large number of substances will end up in category
iv.  To provide biological data to assist in the sorting process, EPA proposes to conduct High
Throughput Prescreening (HTPS) on a  significant number of substances (perhaps as many as
15,000), using in vitro assay systems incorporating transcript!onal activation or reporter gene
systems for the estrogen-, androgen- and thyroid-hormone systems.

       In Charge element (c) (1) EPA asks two questions about this approach: first, is this a
reasonable approach and sorting strategy to support priority setting?; and second, based upon the
data developed through the pilot study to date, can this technique can be used as  a prescreening
device, and what modifications/improvements must be made? (The latter two questions are
addressed in section 3.3.2, below.)

  3.3.1 High Throughput Technology As A Tool for Priority Setting

       EDSTAC has recommended the use of HTPS in order to address the problem that most
chemical substances on the TSCA Inventory have little or no data regarding their potential to
interact/modulate/disrupt the endocrine system.  HTPS is designed to: a) provide priority setting
information for chemicals to be examined in Tier one Screening (T1S); b) provide a prospective
on the effectiveness of HTPS relative to other methodologies such as QSARs; and c) satisfy the
receptor binding/in vitro gene expression T1S requirement for those chemicals that go through
HTPS.

       EPA does not intend to use HTPS data to establish the endocrine disrupting status of a
chemical. Nevertheless, there is considerable concern that results from HTPS will be the first
available data, and will thus be (inappropriately) used, resulting in a certain stigma or in product
de-selection.  This is a concern and appropriate measures should be spelled out and taken in
order to ensure that the data from HTPS is not misused.
                                           19

-------
       Unfortunately, the EPA-funded demonstration project with OSI Pharmaceuticals failed to
sufficiently demonstrate the utility of their HTPS system for the purpose of chemical sorting and
priority setting of estrogen, androgen and thyroid active chemicals. The Joint Subcommittee
raised several concerns regarding the responsiveness and selectivity of the assays developed to
date. We believe that the currently available data obtained from the OSI assays would not be of
assistance in chemical sorting and priority setting. However, it was acknowledged that this was
a work in progress and that, in general, the HTPS approach had merit but required further
development prior to implementation.  Therefore, the Joint Subcommittee agrees that in
conjunction with other priority setting data, results from estrogen and androgen receptor HTPS
assays could contribute to chemical priority setting provided the assays are validated and
standardized.  The Subcommittee also questioned the utility of the thyroid receptor HTPS assay,
since there are no known examples of endocrine disruption that occur as a result of chemical
interaction with this receptor.

       The Subcommittee had one additional concern. EPA's plan for increasing the quality of
the assay is appropriate, but there appears to be no contingency plan in the event that it is
eventually discovered that the assay is not working.  Also, the plan says nothing about a time
frame for making adjustments to the assay, nor at what point it would be prudent to discontinue
it and seek other approaches.

  3.3.2 High Throughput Technology As A Prescreening Device

       Eight transcriptional activation assays have been recommended by EPA. These assays
include the estrogen receptor (ER) alpha and beta, the androgen receptor (AR), and the thyroid
receptor (TR) in the absence and presence of metabolic activation/detoxification system.  The
OSI Pharmaceuticals Corporation (under an EPA contract as noted above) initiated a study to
determine the feasibility of using AR and TR transcriptional activation assays to pre-screen
chemicals in the presence and absence of a metabolic system.  Sixty-one chemicals were
examined including known ER or AR agonists and antagonists. The known ER and AR active
chemicals were selected in order to span a wide range of potencies. As of March 5, 1999, stably
transfected ER and TR transcriptional activation assays in the absence of metabolic systems have
been used to assess the 61 selected chemicals.  A stable AR cell line has been selected and was
used in an initial pilot screen of 16 chemicals.

       Following a review of the data and an up to date presentation by OSI at the public
meeting,  the Joint Subcommittee believes that the OSI HTPS assays were not ready to be used as
a pre-screen device. The following modifications/improvements are suggested in order to ensure
its usefulness:

       a)    improve responsiveness and selectivity of assays

       b)    conduct a thorough statistical analysis of the results to identify significant
             chemical effects on gene expression
                                           20

-------
       c)      validate and standardize all HTPS assays using a training set of agents known to
              be either positive or negative with regard to endocrine disruption. Use of this set

              should identify the error rate, i.e., the percentage of false positive and false
              negative findings.

       d)      verify the results by comparing to other bench gene expression assays

       e)      develop assays that would be capable of distinguishing interactions between
              estrogen receptors alpha and beta.

       f)      establish/define criteria for positive, negative and equivocal results

       The Joint Subcommittee  also made the following suggestions regarding the use of HTPS
assays for the purposes of priority setting:

       a)      re-open the bidding process to include other assays (e.g., receptor binding) and to
              identify additional analytical resources

       b)      consult with intramural EPA scientists and extramural scientists with expertise in
              receptor binding/gene expression assays to evaluate responses to any Agency
              request for proposals

       c)      investigate the development and utility of other HTPS assays such as gene
              chip/cDNA array assays and computer modeling of receptor ligand binding
              domain-chemical interactions

3.4 The Proposed Endocrine Disrupter Screening Program.

       EPA is proposing to develop and implement an Endocrine Disruption Screening Program
(EDSP) that consists of two phases: screening, currently consisting of eight components, and
testing, currently envisioned to have one to five components, depending upon the need for
identifying effects in various sectors of the animal kingdom.  EPA posed a number of questions
to the Subcommittee concerning the proposed EDSP, comprising Charge elements (d)(l-10).
The response to these questions are provided in the following report sections (3.4.1-3.4.10).

  3.4.1  The Two-phase Sorting Strategy

       In Charge element (d) (1), the Agency asked if it is reasonable and appropriate to develop
and implement a two-phase program,  the first phase focused on identifying a substance's
potential to interact with one or more  of the three hormone systems, the second phase to
characterize the effects of concern that interaction with these hormone systems might elicit.
                                           21

-------
       The Subcommittee supports the proposal to develop a two-phase program for endocrine
disrupter screening and testing (EDST). Further, a formal reevaluation of the screening and
testing process at regular intervals should be part of the program. The purposes of this
reevaluation process would be to evaluate the effectiveness of the protocols initially adopted for
screening and testing and to adopt new protocols in cases where none currently exist for
identifying endocrine alterations or the effects of those alterations.  Adoption of new screens and
tests should also mean the  elimination of previous, less useful ones.

       The suggestion was made that non-mammalian systems might serve as developmental
screening tests.  Amphibians, birds and fish have all been used for developmental screening to
provide an integrative assessment system.  The fish and the bird assays seem to be the most
sensitive. Of the three, the basic mechanisms underlying development are best understood in the
bird to date, but some fish  (especially zebra fish  and medaka species) are rapidly catching up. Of
the amphibians, the frog embryo teratogenesis assay Xenopus( FETAX) may be adaptable to a
fully integrative  screening  assay (Fort, 1995, 1996).

       As multiple laboratories are likely to be running the prescribed assays, it is important to
establish procedures for standardization among laboratories and for training of the technicians
and scientists who will run the screens and tests. Significant consideration and planning needs to
be conducted on how to ensure inter-laboratory standardization.

       The task  of testing  for endocrine-disrupting activity and related potential  adverse effects
is at the cutting edge of current science (and some feel that it is ahead of the state-of-the-art), and
that it is therefore particularly important to:

       a)    be as explicit as possible about the type and significance of the effects that the
             tests are attempting to assess.

       b)    incorporate updates of the screening and testing protocols early and often.

       c)    focus resources on the weakest  scientific links in the screening and testing process
             (such as the ability to clearly link the tests to risks of adverse effects).

       Finally, the absence of an anticipated report from the National Research Council (NRC)
study committee on endocrine disrupters made the task of this review Subcommittee more
difficult from a procedural perspective. We may need to reevaluate our recommendations on the
underlying science for screening and testing once conclusions drawn by the NRC study
committee are available.

 3.4.2 Adequacy of the Screening Battery

       The EDSTAC (1998) recommended, and EPA proposes to implement, a screening
battery consisting of eight  assays, three in vitro and five in vivo, to address estrogen, androgen,
and thyroid systems (EAT) effects. The Agency believes that these eight assays would detect all

                                           22

-------
substances currently known to interact with the EAT, and asked the Subcommittee (in Charge
element (d)(2)) if it agrees with this position. Also, EPA sought advice on making changes in
the battery to assure the identification of substances of potential concern.

       It is difficult to evaluate the proposals provided to the Subcommittee without a clear
determination of the scope and nature of the problem for both humans and wildlife. A concise
description of the scope and nature of the problem for both humans and wildlife would be useful
to demonstrate the relationship of the proposed screens to the effects of concern. The
Subcommittee agreed that, as a minimum, the Agency must develop an acceptable methods and
standardization and validation program for all proposed testing methods.  The program proposed
is clearly screening more for mechanism than adverse responses.  There are other potential
mechanisms in the EAT systems of which we are not fully aware.  Thus, the information on di-n-
butyl phthalate (Gray, 1999a) presented by EPA staff at the public meeting clearly show major
adverse effects produced by alterations in the androgen system during development. These
effects might not be found using the screens currently employed in Tier 1. This indicates that
there are potential critical events that would not be detected with the current screening battery.
An in utero (or in ovo) screen that is recognized as the most sensitive exposure window for
endocrine disrupter event (see section 3.4.) should be utilized.  The Joint Subcommittee firmly
agreed that an in utero assay should be developed by the EPA and that it should be considered as
a substitution, not an addition to the proposed battery. However, it is imperative that it be
validated before becoming a required assay.

       Thus, the screens in whole animals would provide access to more potential mechanisms
than receptor based screens  since these animals would have intact
hypothalamic/pituitary/gonadal or thyroid axes and also have multiple end points in the same
animals related to endocrine disturbances. Moreover, such screens would also provide positive
information on reproductive (and developmental) toxicants that act, for example, directly on the
gonads via non-endocrine primary mechanisms, such as methoxyethanol (Foster et a/., 1986),
but would also affect endocrine end points subsequent to gonadal damage over several days.

       The Subcommittee also suggests that developmental nervous system endpoints should be
incorporated into the screening assays. This could  be done at the level of a "to be developed"
integrative screening assay.  In addition, believe that  due regard should be given to the dose
route employed in the in vivo screens. Some flexibility should be employed but it was
considered that the most appropriate route of exposure (that which mimics the typical route of
exposure in humans and/or wildlife) be chosen with the oral route being the default exposure
route.  The use of the intra peritoneal route, especially for the uterotrophic assay, was considered
to be inappropriate.

       Further, addressing technique, it was not clear why the fish protocol presently being
refined and tested by EPA's Duluth laboratory specifically says not to use the organ weight
corrected for body weight (organ/somatic index). This practice is generically used and
recommended to account for any changes in overall body weight induced by the chemical
treatment. The protocols should be consistent with each other. Since exposure to the chemicals

                                           23

-------
may well induce changes in body weight as a separate phenomenon from any gonad or other
endocrine effects, it would seem that the protocol should be to calculate and report both the raw
organ weights and the body-weight-corrected somatic indices.

       The Subcommittee was also aware that there is a huge gulf in terms of effort, complexity
and cost between Tier 1 screening and Tier 2 testing.  The EPA may wish to consider if an
intermediate tier would be warranted that would provide valuable information without the
expense of multi-hundred thousand-dollar efforts.

  3.4.3 Adequacy of Thyroid Coverage

       Interaction with a receptor is the principal or key mechanism by which substances_exert
their effects on the EAT.  There is an exception, however, in that this appears not to be the case
for the thyroid-hormone  system. Consequently, the Agency asked the Joint Subcommittee to
comment on the adequacy of coverage of the thyroid provided in the proposed screening battery,
and suggest modifications if needed (Charge element (d)(3)).

       The Subcommittee believes that the proposed screening battery should detect alterations
in thyroid function. However, the screens proposed are more general and less robust than those
designed to detect alterations in estrogens and androgens. It would be prudent to have thyroid-
hormone-sensitive tests in the screen. Most known thyrotoxicants produce changes in thyroid-
related hormones and/or  clearance and/or thyroid histology. The proposed EDSTAC screening
process for thyroid hormone appears to address these requirements.  Measuring hormone levels
and thyroid histopathology in rats, and amphibian tail resorption, should effectively capture the
strongest thyrotoxicants.   The Subcommittee supports the inclusion of Thyroid Stimulating
Hormone (TSH) and T3 in addition to the measurement of T4, histopathology and amphibian tail
resorption.3  Only the proposed amphibian tail resorption test specifically evaluates an effect of
thyroid hormones on target tissues.

       No data were offered by EPA to support the inclusion of additional tests, other than the
fact that T3 is the biologically active form of the hormone, and that an elevation in TSH would
confirm a physiologically relevant reduction in T4 or T3 levels. The data from Cook and
O'Connor (in press) showed that for every thyrotoxicant that  reduced T3, there were also
changes in T4, which offers direct support for the EPA proposal.  If the EPA wishes to have the
extra confirmation of a thyroid effect (or lack thereof), the Subcommittee would support the
inclusion of TSH and T3 in addition to the measurement of T4, histopathology and amphibian
tail resorption.

       Because few chemicals that alter thyroid function do so by binding to thyroid receptors,
binding assays and gene  reporter screens for thyrotoxicants have been omitted from Tier 1.  This
       3 T3 and T4 are two forms of thyroid hormone, the digits "3" and "4" indicating the number of iodine
molecules in its atomic structure.

                                          24

-------
omission is appropriate because it would have generated false negatives for thyroid effects.
However, if these receptor assays are to play a large part in priority setting for testing in Tier 1, it
needs to be recognized that less information about potential thyroid alterations will be available
from the Tier 1 screens than is the case for E and A.

       There is broad agreement that most known thyrotoxicants produce changes in thyroid-
related hormones and/or clearance and/or thyroid histology.  And the Subcommittee agrees that
it is prudent to have some degree of overlap and complementarity in the screening tests.
Consequently, the same function should be evaluated by more than one test. Finally, because
hormone signals are both amplitude and frequency modulated signals, and a single time-point
measurement may not capture or identify an exposure-related change when one is primarily
measuring a hormone, it is desirable to also measure some downstream functional result of that
hormone.  The proposed EDSTAC screening process for thyroid appears to address these
requirements by measuring hormone levels and thyroid histopathology in rats, and tail resorption
in amphibians. These measures should effectively capture the strongest thyrotoxicants.

  3.4.4 In utero and In Ovo Screens  and Single Dose Screening

       EPA would prefer to have a screening battery which included assays containing an in
utero or in ovo exposure component, given the concern about the potential for effects on the
developing vertebrate organism. At the time the proposed screening battery was being
assembled, EPA was not aware of the  existence of any such screens.  The question posed to the
Subcommittee in Charge element (d) (4) asks for comment on any such assays that may exist or
are under development, and that could supplant or complement one or more components of the
proposed screening battery?  Charge element (d) (5) addressed EPA preference to conduct each
in vivo screening assay using only one dose, selected through the use of range finding studies
and other information. The Agency asked if the Subcommittee agreed with the single dose
approach, and what suggestions it had to modify it. The EPA also sought advice on the possible
public health consequences of these false negatives. The Subcommittee decided that, since the
issues
were inter-related, it would be best to address both elements of the Charge in a single response,
which follows below.

       The Subcommittee prefers those tests which bundle several endpoints into a single "test
unit." The Subcommittee consequently supports the use of gene reporter and binding assays as
part of Tier 1. Problems may be encountered because of differences in the specificity of
different cell systems and because of patent control of some assay components. A screen, using
animals exposed in utero and possibly during lactation, is appealing.  The Subcommittee
strongly encourages the continued development and evaluation of such a protocol. It could
replace several individual assays. No  protocol for such a test has been evaluated or validated as
a screen to date. However, such a test is easily developed by taking pieces of existing protocols
(see the discussion in section 3.4.1). The development of such a protocol would significantly
improve screening effectiveness, reduce the numbers of animals used, and could improve overall
efficiency of screening.

                                          25

-------
       The consensus of the Subcommittee regarding dose levels for in vivo screens was focused
around two issues.  First, for those relatively non-toxic agents, the employment of a single limit
dose (as specified in the Federal Register (1998) document, e.g., 1 g/kg/d oral) was considered
to be appropriate. Second, in other cases where non-specific toxicity could be possible, the
highest dose level tested should elicit some, but not overt, systemic toxicity in line with the
establishment of a maximum tolerated dose (MTD).  A second dose level should then be
employed at one quarter of the MTD.  The Subcommittee felt that the application of a multiple
of exposure as the highest dose level tested would not be appropriate since exposure data would
only be available infrequently. The Subcommittee also found that false negatives from high dose
exposure were unlikely, since even where multiple mechanisms may be operating at different
parts of the dose response curve, one would not expect to see effects only at low dose levels.  A
second lower dose level would also resolve some of these questions.

       The potential for identifying effects at low doses of putative endocrine disrupters was
discussed.  In view of the preliminary nature of these potentially important findings (see Nagel,
et a/., 1997), the Subcommittee recommends that EPA continue to remain alert to new
information on low dose effects.  The Subcommittee was pleased to learn that EPA is sponsoring
a workshop on this issue in May, 1999.

       It should be noted that the current EPA  synthesis of the EDSTAC recommendations is
inconsistent on the matter of route of exposure.  The uterotrophic assay uses subcutaneous or
intra peritoneal injection, the Hershberger assay oral gavage, the multi-generation assay uses
diet/oral/inhalation, and no route is identified for the pubertal male and female assays.
Consistency is preferred unless evidence requires otherwise.

       An integrative developmental assay using the chicken was proposed (Henshel, 1998,
1996; Henshel et a/.,  1997). The assay integrates both a rapid five-day screening component
with a more complete developmental assessment. Many chemicals that are developmental
toxicants interact with the embryo during organogenesis. Therefore, using modifications of
established procedures, and modeling the system after the mammalian embryo culture systems,
the avian embryo may provide a useful assay.

  3.4.5 Rigor of The Five Compartment Test Protocol Design

       EDSTAC recommended, and EPA is proposing, a testing phase in the EDSP which could
have as many as five  subject compartments(i.e., mammals, birds, fish,  invertebrates and
amphibians).  Tests for each compartment would be designed to delineate the dose response
relationships of effects of concern for chemicals which yielded positive results during the
screening phase. The testing protocols to be used are either upgrades or modifications of
existing guidelines, except for the amphibian.  In this case, a protocol is being developed de
novo. The  Subcommittee was asked if the planned test protocol designs would provide sufficient
rigor to identify the effects of concern and establish their dose response relationship for
disruption in the EAT systems (Charge element (d) (6)).
                                           26

-------
       The Subcommittee believes that the five specified compartments should be adequate for
detecting endocrine mediated events in most animal species.  However, the Agency needs to be
cognizant that there may be exceptions in some cases (e.g., reptiles, for which there are no EPA
guidelines).  The Subcommittee also concluded that many of the proposed tests were valid assays
of endocrine disrupters.  They also concluded that methods must be standardized and validated,
based on accepted criteria for validation and regulatory acceptance of toxicological test methods.
Other tests, however, met with considerable criticism:

       a)     A more comprehensive, in-utero test battery should be assembled to replace
             several tests in Tier 1  (see sections 3.4.1 and 3.4.4)

       b)    The Daphnia developmental assay should be considered as a replacement for the
             mysid assay because there is a better understanding of the endocrine mechanisms
             in Daphnia (Baldwin et a/., 1998;  Baldwin and LeBlanc,  1994).

       c)     The fish assay for endocrine disruption should include the measurement of
             vitellogenin in male fishes. Vitellogenin is a yolk precursor protein made by the
             liver in response to estrogen in female but normally not in male oviparous
             animals. Its detection in male fish is a highly  sensitive assay for estrogenic
             activity.  Many laboratories have the ability to detect vitellogenin by radio
             immunoassay in a variety of species. There are no known barriers to the
             development of such a vitellogenin test, although it would still have to  be
             standardized and validated.

       d)    The fish reproduction assay should include some measure of the reproductive
             fecundity of the selected compounds. Egg production and developmental success
             will detect effects which may not be obviously toxic to the organism but might
             have detrimental effects at the population level.  None of the proposed tests with
             wildlife detect breeding success. Further, exposure to a variety of compounds can
             alter the sex ratios to favor one sex or the other.  The effect of the test chemicals
             should include an evaluation of the sex ratios of eggs (or other stages of
             development) treated with the chemicals.

       e)     The Subcommittee recommends that the EPA  examine the use of the Japanese
             quail to substitute for the proposed avian tests on Bob-white quail and mallard
             ducks. Japanese quail have the advantage of short generation time and  provide a
             model with a great deal of background information.

       f)     Although the Tier 2 tests designed to indicate  thyroid alterations should "identify
             effects of concern" they will not effectively determine whether those hormone
             alterations have adverse effects on the development or function of the target
             tissues for thyroid hormones. Thus, the proposed tests may be adequate for
             detecting hormone perturbations but they don't give information about the effects
             of those perturbations (see additional information in section 3.4.6).

                                           27

-------
  3.4.6 Adequacy of Detection of Critical Endpoints in The EAT Systems

       There could be circumstances in which EPA decides to bypass the screening phase for a
particular agent and go directly into the testing phase. EPA is proposing for those cases that the
chemical under evaluation be tested in all five tests. The Agency is asking the Subcommittee if
the tests in the testing phase will be adequate to  detect all known critical endpoints in the EAT
hormone systems, and what modifications the Subcommittee might recommend (Charge
element (d) (7)).

       A central point in the Subcommittee's discussion was that the tests employed in Tier 2
will be the ones used in risk assessment. The number of tests employed would be dependent on
the use and potential exposure for each chemical.  Clearly no single test, or group of tests, has
the ability to cover all critical endpoints for the EAT systems; they should, however, cover most
endpoints. Many research techniques, especially with regard to the thyroid system, are not at a
stage where they can be ready for application in a regulatory testing scenario. The
Subcommittee recommends that the EPA should remain alert for new techniques and end points
to improve testing protocols when these become robust and applicable for routine testing.  The
Subcommittee also suggests that specific consideration be given for the use of Japanese quail in
the avian reproduction study and the use ofDaphnia spp. as a useful alternate species for
invertebrates.  It was also unclear during the discussions if the proposed Mysid species did
indeed have a functioning EAT system. Since these were the specific endocrine systems laid out
in the EDSP,  it would be inappropriate to propose a species in which estrogen, androgen and
thyroid hormones did not have a physiological role.

       The immediate focus of many of the proposed tests is on mechanisms. The ultimate goal,
however, is the capability to detect adverse effects on reproduction and development in a variety
of species.  Thus, all chemicals interfering with  reproduction and development should be
detected in these test systems,  including those whose primary mechanism is not via a disturbance
in the endocrine or EAT systems. Although the risk assessment for any adverse effect and the
dose response data for that adverse effect will be provided by these tests, it will be unfortunate if
all reproductive and developmental toxicants are labeled as "endocrine disrupters." This issue
further raises  the need for a clear definition of an endocrine disrupter - if it is to receive special
consideration — as opposed to being treated as any normal reproductive or developmental
toxicant.

       The advent of new test end points (especially for incorporation into the mammalian two-
generation reproduction  study) has raised questions about the adversity of specific responses and
the normal range for these end points (e.g. anogenital distance, preputial separation, vaginal
opening). Guidance from the EPA would be especially welcome in these specific areas of
testing.

       The Tier 2 tests include few endpoints that will detect critical target tissue effects of
thyroid hormone alterations. Such tests are needed to provide suitable information about
whether alterations in thyroid hormones (which  should be detected by the proposed screening

                                          28

-------
and testing) will affect other developmental, morphological or physiological endpoints in target
tissues.  Currently such tests are not available for quick adoption. However, the research
information about these effects is available and could be used as the basis for development of
such tests at later stages of this program. With respect to the evaluation of thyroid function, the
proposed Tier 2 tests seem marginally adequate for providing information to the final program
stages of hazard evaluation and risk assessment. Thus, although the addition of more tests to
Tier 2 should not be done lightly, there are serious questions about the adequacy of the thyroid
tests for assessing whether there are adverse effects of thyroid alterations.

       The proposed Tier 2 tests include some endpoints affected by alterations in thyroid
function ( e.g. growth). However, the proposed measurements are not very sensitive and most
are ones that involve the interactions of several hormone systems.

  3.4.7 Additional Screening for Agents Initially Found to Be Negative

       EPA wished to know what, if any, additional screening or testing would be required to
assure that  an agent is not an EAT disrupter, if the results of any of the testing phase tests are
negative (Charge element (d) (8)).

       The Subcommittee agreed that, if an agent is found to be inactive in the Tier 2 tests, it
would be regarded as being inactive as an endocrine disrupter. This is axiomatic, as the Tier 2
tests were selected to define the endocrine toxicity of agents found to be potentially active in the
Tier 1 tests. So the answer to the question posed is that no further testing would be required.

       The Members also noted that an agent found to be active in Tier 1 tests, but inactive in
Tier 2 tests, should be considered to be inactive as an endocrine disrupter. In particular, the
positive Tier 1 data should not assume" a  life of its own" after the Tier 2 tests are found  to be
negative.

  3.4.8 Endocrine Disrupters and Hazard Assessment

       Testing (as opposed to screening) phase tests will identify effects of concern that are the
consequence of endocrine disruption. They may also identify effects of concern that are  not the
consequence of endocrine disruption. Thus, it may not be possible to determine if a substance is
an endocrine disrupter if it has not been subjected to some or all components of the screening
battery. Because of this, EPA has asked the Subcommittee if it is important to be able to identify
substances  as endocrine disrupters from the standpoint of conducting a hazard assessment, and if
so, why (Charge element (d) (9).

       It is important to be able to identify substances as endocrine disrupters from the
standpoint  of conducting a hazard assessment. If a compound causes toxicity, it should be
treated like all other toxicants. On the other hand, knowing that a compound is more toxic to
developing hormonal systems means that the particularly vulnerable populations are more likely
                                           29

-------
to be protected. The perception among the Subcommittee Members is that hazard assessment
will not likely be impacted, but that risk assessment will be improved.

  3.4.9  Validation of The Proposed Screens and Tests

       EPA is proposing a validation program in which the maximum validation effort will
consist of conducting each assay in three laboratories.  EPA believes that there currently is a
wide variation in the state of validation of each of the proposed screens and tests, and that the
validation efforts should be tailored for each assay/test accordingly. EPA plans to focus first on
the validation of the mammalian assays, as they are both better developed than the non-
mammalian assays and are more directly relevant to meeting the FQPA and SDWA mandates for
a screening program for potential human health  impacts. EPA's preliminary determination of
the areas needing development are: a) the uterotrophic assay; b) the Hershberger assay; c) the
pubertal male and pubertal female assays;  d) the mammalian two-generation reproduction test;
and e) the non-mammalian screens and some of the non-mammalian tests.

        EPA asked if the Joint Subcommittee agrees with the Agency's assessment of the current
status of the screens and tests, and, if it reached  differing conclusions, to provide the background
and rationale for its findings (Charge element (d) (10)).

       It was agreed that the new mammalian multi-generation assay protocol would require
validation of its practicality.  It cannot be validated per se because it is an apical Tier 2 test. The
Subcommittee recommended that the validation should proceed sequentially. One laboratory
should establish practicality, and that result should then be confirmed in one or two additional
laboratories.  An objective appraisal  of the result of the first run could well indicate that the
protocol is practical, and the second phase of validation may be canceled. This point is
important, given the time taken to  conduct the assay, and the present need for the assay as the
most informative (Tier 2) test.

       The Subcommittee also agreed that the non mammalian Tier 2 tests, as well as the
mammalian tests, would require formal validation as to their practicality and
sensitivity/specificity.

       The purpose of the Hershberger assay is  to quantify the effects of potential
anti-androgenic and androgenic compounds on the hormone-dependant tissues in the immature
male rat (Hershberger, et al. 1953). Castrated immature male rats, reared under standardized
housing conditions, are treated with a potential xenobiotic or the vehicle via oral gavage.  The
animals are then euthanized and the relevant target tissues are fixed and stained and examined
for histopathology.  Serum thyroxin (T4) and TSH is measured,, as is serum Luteinizing
Hormone (androgen measurements are optional).  The data are then analyzed for statistical
significance of any differences found between the treated animals and the controls.

       The Subcommittee was concerned that the existing animal assays in Tier I may not be
sensitive to events occurring uniquely in the fetus  or in the developing neonate/weanling.  The

                                          30

-------
development of a limited in utero assay is currently under study, and several laboratories are
evaluating the effects of a range of endocrine active chemicals on sexual development of
perinatal rats and mice (Gray etal.,, 1999a; 1999b). Uterotrophic effects in the female weanling
rat are already incorporated as an alternative assay in the EDSTAC proposals, and work is being
done on the male weanling at present (Ashby, and Lefevre,  1997a; Gray et a/., 1997).

       We recommend that these assays be kept under close review, with attention focused on
the results obtained when testing the activity of the same agents with the different types of assay.
It may be that data will eventually indicate that one or other of these classes of assay can replace
the existing rodent assays in Tier I, but a well  constructed, robust, database will be needed before
such a decision can be made. The Subcommittee endorsed strongly the continuing evaluation of
endocrine disruption assays that cover the periods of gestation and sexual  development.

  3.4.10 Subcommittee Recommendations to Help EPA  Meet its Charge

       The final element of the Charge (d) (11) asked the Joint Subcommittee for any other
suggestions or recommendations that would help EPA meet its charge.

       The body of this report provides specific recommendations concerning the screening and
testing of endocrine disrupters, as posed by the Charge. The following section of this report
contains a summary of our major findings and recommendations including some issues not
included in the Charge which arose during the Subcommittee's public meeting and/or the
development of this report..

-------
             4. MAJOR FINDINGS AND RECOMMENDATIONS

       This section highlights a variety of recommendations and concerns discussed at the
public meeting, or generated during the preparation of this report. These findings are:

       a)     Evaluating the Program: We wish to reinforce the comments concerning the
             lack of on-going program evaluation noted in section 3.1.1 of this report.  We find
             no provision for mid-course evaluation or optimization of the process. The
             Agency is mandated to assemble and evaluate this proposed panel of tests and
             then to implement them, but a correlate responsibility is to make sure that what's
             being done is the best that can be. Edmund Burke's "You can never plan the
             future by the past," and Robert Burns' "The best-laid plans of mice and men oft
             gang agley." both apply here. For example, evaluation of minced testis and
             minced-ovary assays finds them to be only 50% effective in identifying
             compounds that inhibited steroid biosynthesis (Powlin et a/., 1998). Although
             something looks fine on paper or in  a small research setting, translating it into
             volume-screening mode may be quite another thing.  There was broad support
             among the Subcommittee for the concept that the Agency should convene a panel
             of independent scientists to review all the screening data for 50-100 compounds,
             with an eye towards revising the process and eliminating those methods that don't
             work. The dictum that, the more removed a screen is from a whole model, the
             more wrong the answers are likely to be, also supports the need for such a review.

             Finally, we believe that the regulated community and the public interest groups
             would be more willing to participate if they knew that the system was going to be
             optimized as it proceeded. The Agency should have one or more evaluations of
             the process as we proceed with this, and the Subcommittee strongly encourages
             this.

       b)     Mixture Issues: Discussions at the public meeting focused on whether to include
             mixtures in the listing of materials to be screened and tested. The Subcommittee
             agreed that the initial focus of the methods development effort must necessarily
             focus on single compounds and leave the question of testing of mixtures until
             accepted single-compound methods have been completed.  However, Agency
             representatives underscored the need to apply the methods to testing of effluents
             and source waters which are obviously complex mixtures.  The Subcommittee
             concluded that very promising methods already exist in the field of
             ecotoxicology. These include the Whole Effluent Testing (WET) and Toxicity
             Identification Evaluation (TIE) procedures developed by the Agency  in concert
             with the Society for Environmental Toxicology and Chemistry. Those methods
             have been developed to test effects of effluents and would have direct application
             to the Endocrine Disrupter Screening Program.
                                          32

-------
c)     Case Studies: The Subcommittee strongly encourages the Agency to include
       more and better-detailed case studies in the evolution of the priority-setting
       scheme. Case studies will enable a realistic test of the scheme, checking
       sensitivity of the system and its working practicality to actually prioritize
       chemicals for further testing.

d)     Sub-population Compartment: The question of the need for a separate defined
       compartment to address sub-populations (i.e., human infants) was addressed to
       the Subcommittee. Our conclusions supported the use of sub-populations as a
       criterion within the existing compartments already identified, but not as a  separate
       stand-alone compartment.

e)     Use of IRIS: The priority testing scheme relies on the use of several databases
       summarizing the environmental fate and effects of chemicals. Caution was
       expressed by several members of the Committee that there are numerous
       problems with the validation of IRIS and other databases  as well. Before placing
       heavy reliance on these computerized information systems, users need to be
       aware of these validation problems and proceed with caution before incorporating
       these data at their face value.

f)     Exposure: Although the issue was not explicitly stated in the Charge, the
       Subcommittee expressed concern that consideration of the toxicological
       implications of exposure should include both dose and timing of exposure,
       particularly with respect to developmental  or reproductive events. The current
       scheme does not adequately cover the time aspect of exposure and this needs to
       be remedied before broad-scale application of the approach,  e of exposure. The
       current scheme does not adequately cover the time aspect of exposure and this
       needs to be remedied before broad-scale application of the approach.

g)     Use of Animals: During the public meeting, concern was expressed about the
       large number of animals that would be needed in the EDSTAC program. The
       Subcommittee was asked whether alternatives and approaches to  minimize animal
       use had been appropriately considered in developing the protocols.  The
       Subcommittee pointed out the essential role animals play  in tests  to detect
       endocrine disruption to reveal adverse effects on humans.  There  are no
       substitutes for tests currently available for the Tier 1 or Tier 2 tests using animals.
       Because of the complexity of the biological systems involved in endocrine
       disrupter detection, animals will remain a necessary model for the foreseeable
       future. Additional comments by Subcommittee Members and others described
       protocol or method modifications which would be less expensive, faster and use
       fewer animals.  The Agency has an obligation to conserve all resources in
       developing new testing protocols, and the use of animals in such tests poses both
       ethical and practical problems
                                    33

-------
h)     Need for an Introductory Statement: The previous EDSTAC meeting
       suggested that the final document needed, as a introductory section, a description
       of the problem or the scientific or health-based reason for the EDSTAC program
       (1998).  Although the anticipated NAS/NRC report is expected to address this
       issue, the Joint Subcommittee now urges the EDSTAC team to include a
       description of both the health and ecological problems associated with exposure
       to the endocrine disrupters and to show how the program relates to these findings.
i)      Support for Decisions: Decisions about which assays are selected, and which
       protocols are adopted for those assays, should be supported with data that are
       generally available.

j)      Exceptions: Testing strategies will always have exceptions. Care should be taken
       to be aware of the imperfect nature of any future agreed strategy.  In particular,
       there is the present danger that the two chemicals dibutyl phthalate (DBF) and
       methoxychlor (MC) will have an undue influence on the future. The status of
       DBF as an anti-androgen (Gray et a/., 1999c; Bulger et a/., 1978) has yet to be
       fuly documented.. MC is sometimes reported as being inactive for in vitro
       estrogenicity assays, thereby providing a precedent for using animals in a
       screening mode.  However, this chemical has been published as being active as an
       estrogen in vitro (Ashby, 1997b).

k)     Negative Control Agents: There is a need to define and agree on some negative
       control agents for ED assay validation. It has been suggested that the only valid
       one at present is diethyl phthalate. (Foster, 1980). This position is supported by
       the fact that it gave negative results in a full and updated rodent multi- generation
       study at the National Institute of Environmental Health Sciences (Chapin, 1997).
       However, this view is questioned by work in progress by Ashby and by Gray
       indicating positive activity.  Assay specificity will not be capable of assessment
       unless additional reliable negative agents can be found and made available for
       general study.

1)      Animal Tests and Routes of Exposure: As noted above, MC stands alone as the
       only precedent for why animals should be used in the screening mode  (Tier 1).
       There is significant international concern on the proposed use of animals for
       screening. Ashby and Lefevre (1997a) proposed that short term animal studies
       should be recognized as an intermediate Tier, much as happens now with the
       anticipation of animal carcinogens [screen in vitro,  assess in short term in vivo
       assays, and then define in lifetime bioassays].  In this role of hazard assessment (as
       opposed to hazard definition) biologically relevant routes of exposure would be
       indicated (oral gavage, diet, water, inhalation, skin painting). At present, use of
       the subcutaneous injection or intra peritoneal injection routes are recommended in
       the frail quest of increasing assay sensitivity. In fact, irrespective of the outcome

                                    34

-------
       of this suggestion it should be noted that the current EPA synthesis of the
       EDSTAC recommendations is inconsistent on the matter of route of exposure —
       the uterotrophic assay uses subcutaneous or intraperitoneal injection, the
       Hershberger assay oral gavage, the multi generation assay uses
       diet/oral/inhalation, and no route is identified for the pubertal male and female
       assays.

       The main contributors to differences in test outcome between assays conducted in
       vitro and assays in rodents will be delivered dose, pharmacodynamics, and
       pharmacokinetics. Route of exposure will dominate these factors. The ultimate
       role to be adopted for animal studies, and the route of animal exposures, will have
       the greatest impact on the successful implementation of the EPA initiatives in the
       area of endocrine disruption.

m)     Expanding the Universe of Agents: Developing massive amounts of screening
       information on a large universe of chemicals does not necessarily expedite the
       development of the appropriate scientific underpinning that the Agency needs to
       broaden this effort.  Consequently, the Subcommittee recommends that EPA
       should not expand the set of agents until the Agency develops or adopts validated
       systems and can provide clear decision criteria.
                                    35

-------
                                   REFERENCES
Ashby, J., and P. A. Lefevre. 1997a. The weanling male rat as an assay for endocrine disruption:
       preliminary observations. Regul. Toxicol. Pharmacol 26(3): 330-337.
Ashby J., Harris, C.A., Lefevre, P.A., Odum, J., Routledge, E.J., and J. Sumpter.  1997b.
       Synergism between synthetic estrogens?  Nature 385: 494
Baldwin, W.S., Graham, S.E., Shea, D., and G.A, LeBlanc. 1998. Altered metabolic elimination
       of testosterone and associated toxicity following exposure of Daphnia magna to
       nonylphenol polyethoxylate. Ecotoxicol. Environ. Saf.39 ( 2): 104-111.
Baldwin, W.S., and G.A. LeBlanc.  1994.  Identification of multiple steroid hydroxylases in
       Daphnia magna and their modulation by xenobiotics. Environ. Toxicol. Chem.(l3) (7):
       1013-1021
Bulger, W. H., etal.  1978. Interactions of methoxychlor, methoxychlor base-soluble
       contaminant, and 2,2-bis(p-hydroxyphenyl)-l,l,l-trichloroethane with rat uterine
       estrogen receptor. J. Toxicol Environ Health. 1978 4(5-6):881-93..
Chapin, R. 1997. Reproductive assessment by continuous breeding: evolving study design and
       summaries of  88 studies. Env. Health Perspectives 105 (1): 199-395.
Cook. J.,  and J. C. O'Connor (In Press).
EDSTAC (Endocrine Disrupter Screening and Testing program Advisory Committee). 1998.
       EDSTAC Final report (Vols 1 and 2). U.S. Environmental Protection Agency,
       Washington DC.
EPA (Environmental  Protection Agency). 1999. "Strawman" priority setting process for
       endocrine disrupter screening and testing. Washington, DC.
EPA (Environmental  Protection Agency). 1997. Guiding principles for Monte Carlo analysis.
       EPA/63 O/R-97/001.
EPA (Environmental  Protection Agency). 1992. Guidelines for exposure assessment. Federal
       Register 57 (104), 22888-22938.
EPA (Environmental  Protection Agency). 1986. Guidelines for the health risk assessment of
       chemical mixtures.  Federal Register 51 (185), 34014-34025.
FR (Federal Register). 1998. Endocrine Disrupter Screening Program: Statement of Policy.
       United States  Government Printing Office, Volume 63, Number 248:71541-71568.
       Washington DC.
Feron, V.J., Groton, J.P., vanZorge, J.A., Cassee. F.R., Jonker, D., and PJ. vanBladerem. 1995.
       Toxicity studies in rats of simple mixtures of chemicals with the same or different target
       organs. Toxcology Lett.  82/83: 505-512.
Foster, P.M., Blackburn, D.M., Moore, R.B., and S.C. Lloyd.  1986. Testicular toxicity og 2-
       methoxyacetaldehyde, a possible metabolite  of glycol monomethyl ether, in  the rat.
       Toxicol Appl Pharmacol 3 2( 1 -2): 73 -80.
Foster, P.M., Thomas L., Cook, M.W., and S.D.  Gangoli.  1980. Study of the testicular effects
       and changes in zinc excretion produced by some n-alkyl phthalates in the rat. Toxicol.
       Appl. Pharmacol. 54(3):392-398.
                                          R-l

-------
Fort, D. J., Stover, E. L., and J.A. Bantle. 1996. Integrated ecological hazard assessment of
       waste site soil extracts using fetax and short-term fathead minnow teratogenesis assay.
       ASTM Spec Tech Publ. (1262): 93-109.
Fort, D,J., Stover, E.L., and D. Norton. 1995. Ecological hazard assessment of aqueous soil
       extracts using FETAX.  J. Appl. Toxicol. 15, (3):183-191.
Gray, L.E., Wolf, C., Lamhright, C., Mann, P., Price, M., Cooper, R.L.,and J. Ostby. 1999a.
       Administration of potentially antiandrogenic pesticides (procymidone, linuron, iprodione,
       chlozolinate, p,p'-DDE, and ketoconazole) and toxic substances (dibutyl- and
       diethylhexyl phthalate, PCB 169, and ethane dimethane sulphonate) during sexual
       differentiation produces diverse profiles of reproductive malformations in the male rat.
       Toxicology And Industrial Health(15)l/2: 94-118.
Gray, L.E., Ostby, J., Cooper, R.L., and W.R. Kelce.  1999c.  The estrogenic and antiandrogenic
       pesticide methoxychlor alters the reproductive tract and behavior without affecting
       pituitary size or LH and prolactin secretion in male rats. Toxicology and Industrial
       Health (15) (1/2): 37-47.
Gray, L.E., Ostby, J., Monosson, E., and W.R, Kelce.  1999b.  Environmental antiandrogens:
       Low doses of the fungicide vinclozolin alter sexual differentiation of the  male rat.
       Toxicology and Industrial Health 15 (1/2): 48-64.
Gray, L.E., Jr., Kelce, W.R., Wiese, T., Tyl, R., Gaido, K., Cook, J., Klinefelter,  G., Desaulniers,
       D., Wilson, E., Zacharewski, T., Waller, C., Foster, P., Laskey, J., Reel, J., Giesy, J.,
       Laws, S., McLachlan, J., Breslin, W., Cooper,  R., Di Giulio, R., Johnson, R., Purdy, R.,
       Mihaich, E., Safe, S., and T. Colborn.  1997. Endocrine screening methods workshop
       report: detection of estrogenic and androgenic hormonal and anti-hormonal activity for
       chemicals that act via receptor or steroidogenic enzyme mechanisms.  Reprod. Toxicol.
       11(5):719-750.
Henshel, D.S.  1998 Developmental and neurotoxic effects of dioxin and dioxin-like compounds
       on domestic and wild avian species. Environ.  Toxicol Chem. 17(1):88 - 98.
Henshel, D.S., Sparks, D.W., Allen,  C.A., Benson, K., Fox, C., Lam, Y., Sobiech,  S.A., Wagey,
       R.. 1997. Preliminary results using the early embryo teratogenesis assay: a comparison
       of early embryo abnormalities with late embryo and hatchling teratogenic changes.
       Environmental Toxicology and Risk Assessment: Modelling and  Risk Assessment (6th
       Volume),ASTM STP 1317, F. James Dwyer, Thomas R. Doane, and Mark L. Hinman,
       Eds., American Society for Testing and Materials, 391-401.
Henshel, D.S.  1996. An argument for the chicken embryo as a model for the developmental
       toxicological effects of the polyhalogenated aromatic hydrocarbons (PHAHs).
       Environmental Toxicology and Risk Assessment: Biomarkers and Risk Assessment (5th
       Volume),ATSM STP 1306, David A. Bengston and Diane S. Henshel, Eds., American
       Society for Testing and Materials, Philadelphia, pp. 219 - 229.
Hershberger, L.G.,  Shipley, E.G., and R.K.Meyer.  1953. Myotrophic activity of 19-nor
       testosterone and other steroids determined by modified levator ani muscle method. Proc.
       Expt. Biol. Med. 83:175-180.
NAS/NRC/COT (National  Academy of Science/National Research Council/Committee on
       Technology). 1989. Safe Drinking Water Committee, Vol. 9, Drinking Water and
       Health, Part II Mixtures. National Academy Press, Washington DC.
                                          R-2

-------
NRC, National Research Council. Risk Assessment in the Federal Government: Managing the
       Process. Washington, D.C., National Academy Press (1983).
NRC, National Research Council. Science and Judgment in Risk Assessment.  Washington,
       D.C., National Academy Press (1994).
Nagel, S.C., von Saal, F.S., Thayer,K.A., Dhar, M.G., Boecher, M., and Welshons, WV. 1997. R
       elative binding affinity-serum modified access (RBA-SMA) assay predicts the relative in
       vivo bioactivity of the xenoestrogens bisphenol A and octylphenol. Enviiron. Health
       Perspect. 105:70. (Nature, 385,494 1997).
Powlin, S.S., Cook, J.C., Novak, S., and J.C. O'Connor.  1998. Ex vivo and in vitro testis and
       ovary  explants: utility for identifying steroid bisosysthesis inhibitors and comparison to a
       tier 1 screening battery.  ToxicologicalSciences 46:51-74..
Yang, R.S.H., Thomas,  R.S., Gustafson, D. L., Campain, J. A., Benjamin, S.A., Verhaar,
       H.J.M., and M.M. Mumtaz. 1998.  Approaches to developing alternative and predictive
       toxicology based on PBPK/PD and QSAR modeling. Envir. Health Perspectives: 106
       (sup) 1385-1393.

-------