&EPA
                                                      EPA/635/R-13/171a
                                                Revised External Review Draft
                                                         www.epa.gov/iris
United States
Environmental Protection
Agency
            lexicological Review of Trimethylbenzenes

               (CASRN 25551-13-7, 95-63-6, 526-73-8, and 108-67-8}

                In Support of Summary Information on the
                Integrated Risk Information System (IRIS)
                                August 2013
                                  NOTICE

 This document is a Revised External Review draft. This information is distributed solely
 for the purpose of pre-dissemination peer review under applicable information quality
 guidelines. It has not been formally disseminated by EPA. It does not represent and should
 not be construed to represent any Agency determination or policy. It is being circulated for
 review of its technical accuracy and science policy implications.
                    National Center for Environmental Assessment
                        Office of Research and Development
                        U.S. Environmental Protection Agency
                               Washington, DC

-------
                                                 Toxicological Review of Trimethylbenzene

                                     DISCLAIMER

       This document is a preliminary draft for review purposes only. This information is
distributed solely for the purpose of pre-dissemination peer review under applicable information
quality guidelines. It has not been formally disseminated by EPA. It does not represent and should
not be construed to represent any Agency determination or policy. Mention of trade names or
commercial products does not constitute endorsement of recommendation for use.
        This document is a draft for review purposes only and does not constitute Agency policy.
                                             ii
                                                          DRAFT—DO NOT CITE OR QUOTE

-------
                                                Toxicological Review of Trimethylbenzene
CONTENTS
CONTENTS	iii
TABLES	v
FIGURES	vi
ABBREVIATIONS AND ACRONYMS	vii
AUTHORS | CONTRIBUTORS | REVIEWERS	viii
PREFACE	xi
PREAMBLE TO IRIS TOXICOLOGICAL REVIEWS	xv
EXECUTIVE SUMMARY	xxxiv
LITERATURE SEARCH STRATEGY | STUDY SELECTION AND EVALUATION	xlvi
1. HAZARD IDENTIFICATION	1-1
   1.1.  Synthesis of Evidence	1-1
      1.1.1.  Neurological Effects	1-1
      1.1.2.  Respiratory Effects	1-25
      1.1.3.  Reproductive and Developmental Effects	1-31
      1.1.4.  Hematological and Clinical Chemistry Effects	1-35
      1.1.5.  Carcinogenicity	1-45
      1.1.6.  Similarities Among TMB Isomers Regarding Observed Inhalation and Oral Toxicity	1-47
      1.1.7.  Similarities Among TMB Isomers Regarding Toxicokinetics	1-49
   1.2.  Summary and Evaluation	1-51
      1.2.1.  Weight of  Evidence for Effects Other Than Cancer	1-51
      1.2.2.  Weight of  Evidence for Carcinogenicity	1-54
      1.2.3.  Susceptible Populations and Lifestages	1-54
2. DOSE-RESPONSE ANALYSIS	2-1
   2.1.  Inhalation Reference Concentration for Effects Other Than Cancer for
      1,2,4-TMB	2-1
      2.1.1.  Identification of Studies and Effects for Dose-Response Analysis for 1,2,4-TMB	2-1
      2.1.2.  Methods of Analysis for 1,2,4-TMB	2-7

        This document is a draft for review purposes only and does not constitute Agency policy.
                                           iii
                                                        DRAFT—DO NOT CITE OR QUOTE

-------
                                                    Toxicological Review of Trimethylbenzene

    2.1.3.   Derivation of Candidate RfC Values for 1,2,4-TMB	2-12
    2.1.4.   Derivation of Organ/System Specific Reference Concentrations for 1,2,4-TMB	2-15
    2.1.5.   Selection of the Proposed Overall Reference Concentration for 1,2,4-TMB	2-16
    2.1.6.   Uncertainties in the Derivation of the Reference Concentration for 1,2,4-TMB	2-19
    2.1.7.   Confidence Statement for 1,2,4-TMB	2-19
2.2.  Inhalation Reference Concentration for Effects Other Than Cancer for
    1,2,3-TMB	2-20
    2.2.1.   Identification of Studies and Effects for Dose-Response Analysis for 1,2,3-TMB	2-20
    2.2.2.   Methods of Analysis for 1,2,3-TMB	2-24
    2.2.3.   Derivation of Candidate RfC Values for 1,2,3-TMB	2-26
    2.2.4.   Derivation of Organ/System Specific Reference Concentrations for 1,2,3-TMB	2-30
    2.2.5.   Selection of the Proposed Overall Reference Concentration for 1,2,3-TMB	2-31
    2.2.6.   Uncertainties in the Derivation of the Reference Concentration for 1,2,3-TMB	2-32
    2.2.7.   Confidence Statement for 1,2,3-TMB	2-33
2.3.  Inhalation Reference Concentration for Effects Other Than Cancer for
    1,3,5-TMB	2-34
    2.3.1.   Identification of Studies and Effects for Dose-Response Analysis for 1,3,5-TMB	2-34
    2.3.2.   Methods of Analysis for 1,3,5-TMB	2-36
    2.3.3.   Derivation of Candidate RfC Values for 1,3,5-TMB	2-38
    2.3.4.   Derivation of Organ/System Specific Reference Concentrations for 1,3,5-TMB	2-41
    2.3.5.   Selection of the Proposed Overall Reference Concentration for 1,3,5-TMB	2-42
    2.3.6.   Uncertainties in the Derivation of the Reference Concentration for 1,3,5-TMB	2-43
    2.3.7.   Confidence Statement for 1,3,5-TMB	2-43
2.4.  Oral Reference Dose for Effects Other Than Cancer for 1,2,4-TMB	2-44
    2.4.1.   Identification of Studies and Effects for Dose-Response Analysis for 1,2,4-TMB	2-44
    2.4.2.   Methods of Analysis for 1,2,4-TMB	2-44
    2.4.3.   Derivation of the Reference Dose for 1,2,4-TMB	2-45
    2.4.4.   Uncertainties in the Derivation of the Reference Dose for 1,2,4-TMB	2-46
    2.4.5.   Confidence Statement for 1,2,4-TMB	2-46
2.5.  Oral Reference Dose for Effects Other Than Cancer for 1,2,3-TMB	2-46
    2.5.1.   Identification of Studies and Effects for Dose-Response Analysis for 1,2,3-TMB	2-46
    2.5.2.   Methods of Analysis and Derivation of the Reference Dose for 1,2,3-TMB	2-46
    2.5.3.   Uncertainties in the Derivation of the Reference Dose for 1,2,3-TMB	2-47
    2.5.4.   Confidence Statement for 1,2,3-TMB	2-47

      This document is a draft for review purposes only and does not constitute Agency policy.
                                               iv
                                                              DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

    2.6.  Oral Reference Dose for Effects Other Than Cancer for 1,3,5-TMB	2-48
        2.6.1.   Identification of Studies and Effects for Dose-Response Analysis for 1,3,5-TMB	2-48
        2.6.2.   Methods of Analysis and Derivation of the Reference Dose for 1,3,5-TMB	2-48
        2.6.3.   Uncertainties in the Derivation of the Reference Dose for 1,3,5-TMB	2-49
        2.6.4.   Confidence Statement for 1,3,5-TMB	2-49
    2.7.  Cancer Risk Estimates for 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB	2-49
REFERENCES	R-l
TABLES
Table ES-1. Summary of inhalation reference concentrations (RfCs)	xxxv
Table ES-2. Summary of reference concentration (RfC) derivation for 1,2,4-TMB	xxxvi
Table ES-3. Summary of reference concentration (RfC) derivation for 1,2,3-TMB	xxxviii
Table ES-4. Summary of reference doses (RfDs) forTMB isomers	xli
Table ES-5. Summary of reference dose (RfD) derivation for 1,2,4-TMB	xli
Table LS-1: Details of the search strategy employed forTMBs	xlvii
Table 1-1. Evidence pertaining to neurological effects of TMBs in animals — inhalation exposures	1-10
Table 1-2. Evidence pertaining to neurological effects of 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB in animals —
           oral exposures	1-16
Table 1-3. Evidence pertaining to respiratory effects of TMBs in animals — inhalation exposures	1-27
Table 1-4. Evidence pertaining to reproductive and developmental effects of 1,2,4-TMB and 1,3,5-TMB in
           animals — inhalation exposures	1-33
Table 1-5. Evidence pertaining to hematological and clinical chemistry effects of 1,2,3-TMB, 1,2,4-TMB, or
           1,3,5-TMB in animals — inhalation exposures	1-38
Table 1-6. Evidence pertaining to hematological and clinical chemistry effects of 1,3,5-TMB in animals —
           oral exposures	1-41
Table 1-7. Similarities between 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB regarding observed inhalation and
           oral toxicity	1-49
Table 2-1. Target and actual inhalation concentrations, and internal blood dose metrics of 1,2,4-TMB
           calculated using the available rat PBPK model (Hissink et al., 2007)	2-3
Table 2-2. Endpoints considered for the derivation of the RfC for 1,2,4-TMB	2-5
Table 2-3. Summary of derivation  of points of departure for 1,2,4-TMB	2-11
Table 2-4. Effects and corresponding derivation  of candidate RfC values for 1,2,4-TMB	2-14
Table 2-5. Organ/system-specific RfCs and proposed overall RfC for 1,2,4-TMB	2-16

         This document is a draft for review purposes only and does not constitute Agency policy.
                                                   v
                                                                  DRAFT—DO NOT CITE OR QUOTE

-------
                                                         Toxicological Review of Trimethylbenzene

Table 2-6. Target and actual exposure concentrations used in BMD modeling of 1,2,3-TMB endpoints
           considered for the derivation of the RfC	2-22
Table 2-7. Endpoints considered for the derivation of the RfC for 1,2,3-TMB	2-23
Table 2-8. Summary of derivation of points of departure for 1,2,3-TMB	2-26
Table 2-9. Effects and corresponding derivation of candidate RfC values for 1,2,3-TMB	2-29
Table 2-10. Organ/system-specific RfCs and proposed overall RfC for 1,2,3-TMB	2-31
Table 2-11. Target and  actual exposure concentrations used  in BMD modeling of 1,3,5-TMB endpoints
           considered for the derivation of the RfC	2-35
Table 2-12. Endpoints considered for the derivation of the RfC for 1,3,5-TMB	2-36
Table 2-13. Summary of derivation of points of departure for 1,3,5-TMB	2-38
Table 2-14. Effects and corresponding derivation of candidate RfC values for 1,3,5-TMB	2-40
Table 2-15. Organ/system-specific RfCs and proposed overall RfC for 1,3,5-TMB	2-42
FIGURES
Figure LS-1. Literature search and study selection strategy for TMBs	xlviii
Figure 1-1. Exposure response array of neurological effects following inhalation exposure to 1,2,4-TMB	1-18
Figure 1-2. Exposure response array of neurological effects following inhalation exposure to 1,2,3-TMB	1-19
Figure 1-3. Exposure response array of neurological effects following inhalation exposure to 1,3,5-TMB	1-20
Figure 1-4. Exposure response array of neurological effects following oral exposure to 1,2,3-TMB,
             1,2,4-TMB, or 1,3,5-TMB	1-21
Figure 1-5. Exposure response array of respiratory effects following inhalation exposure to 1,2,3-TMB,
             1,2,4-TMB, or 1,3,5-TMB	1-29
Figure 1-6. Exposure response array of reproductive and developmental effects following inhalation
             exposure to 1,2,4-TMB or 1,3,5-TMB	1-34
Figure 1-7. Exposure response array of hematological and clinical chemistry effects following inhalation
             exposure to 1,2,4-TMB or 1,3,5-TMB	1-42
Figure 1-8. Exposure response array of hematological and clinical chemistry effects following inhalation
             exposure to 1,2,3-TMB	1-43
Figure 1-9. Exposure response array of hematological and clinical chemistry effects following oral
             exposure to 1,3,5-TMB. Mode of Action Analysis- Hematological and Clinical Chemistry
             Effects	1-44
Figure 2-1. Candidate RfC values with corresponding POD and composite UF for 1,2,4-TMB	2-15
Figure 2-2. Candidate RfC values with corresponding POD and composite UF for 1,2,3-TMB	2-30
Figure 2-3. Candidate RfC values with corresponding POD and composite UF for 1,3,5-TMB	2-41
          This document is a draft for review purposes only and does not constitute Agency policy.
                                                    vi
                                                                    DRAFT—DO NOT CITE OR QUOTE

-------
                                                    Toxicological Review of Trimethylbenzene

ABBREVIATIONS  AND ACRONYMS
AAQC      Ambient air quality criterion               OMOE
ACGIH      American Conference of                   OSHA
           Governmental Industrial Hygienists
ADME      absorption, distribution, metabolism,        p
           and excretion                            PBPK
AEGL      Acute Exposure Guideline Levels
AIC        Akaike Information Criterion               PEL
BAL       bronchoalveolar lavage                    POD
BMD       benchmark dose                          PODADj
BMDL      lower confidence limit on the               POI
           benchmark dose                          ppm
BMDS      benchmark dose software                  RBC
BMR       benchmark response                      RDso
BW        body weight                             REL
CAS        Chemical Abstracts Service                 RfC
CASRN     Chemical Abstracts Service Registry         RfD
           Number                                RGDR
CI         confidence interval                       ROS
CNS        central nervous system                    SCE
CYP450    cytochrome P450                         SD
DAF       dosimetric adjustment factor               SOA
DMBA      dimethylbenzoic acid                     TLV
DMHA      dimethylhippuric acid                     TMB
DNA       deoxyribonucleic acid                     TSCA
ECso       half maximal effective concentration         TWA
EEC        electroencephalogram                     UF
EPA       U.S. Environmental Protection              UFA
           Agency                                 UFH
GD         gestational day                           UFs
Hb/g-A    animal blood:gas partition coefficient
Hb/g-H    human blood:gas partition coefficient        UFi
HEC       human equivalent concentration            UFD
HER       human epidermal keratinocytes
HERO      Health and Environmental Research         UV
           Online                                  VOC
HEV       human epithelial keratinocytes             WBC
HSDB      Hazardous Substances Data Bank           WS
IL-8        interleukin-8                             %2
i.p.         intraperitoneal
IRIS        Integrated Risk Information System
JP-8        jet propulsion fuel 8
Km         Michaelis-Menten constant
LDH       lactate dehydrogenase
LOAEL     lowest-observed-adverse-effect level
NCEA      National Center for Environmental
           Assessment
NIOSH      National Institute for Occupational
           Safety and Health
NLM       National Library of Medicine
NO AEL     no-observed-adverse-effect level
              Ontario Ministry of the Environment
              Occupational Safety and Health
              Administration
              probability value
              physiologically based
              pharmacokinetic (model)
              permissible exposure limit
              point of departure
              duration adjusted POD
              point of impingement
              parts per million
              red blood cell
              50% respiratory rate decrease
              recommended exposure limit
              reference concentration
              reference dose
              regional gas dose ratio
              reactive oxygen species
              sister chromatid exchange
              standard deviation
              secondary organic aerosol
              threshold limit value
              trimethylbenzene
              Toxic Substances Control Act
              time-weighted average
              uncertainty factor
              interspecies uncertainty factor
              intraspecies uncertainty factor
              subchronic-to-chronic uncertainty
              factor
              LOAEL-to-NOAEL uncertainty factor
              database deficiency uncertainty
              factor
              ultraviolet
              volatile organic compound
              white blood cell
              white spirit
              chi-squared
         This document is a draft for review purposes only and does not constitute Agency policy.


                                                             DRAFT—DO NOT CITE OR QUOTE
vn

-------
                                                 Toxicological Review of Trimethylbenzene
AUTHORS  |  CONTRIBUTORS  |  REVIEWERS
Assessment Team
       J. Allen Davis, M.S.P.H. (Chemical Manager)
       Eva McLanahan, Ph.D. (LCDR, USPHS)
       Paul Schlosser, Ph.D.
       John Cowden1, Ph.D.
       Gary Foureman, Ph.D. (Currently ICF Int.)
       Andrew Kraft1, Ph.D.
       Ray Antonelli, B.S.
U.S. Environmental Protection Agency
Office of Research and Development
National Center for Environmental Assessment
Research Triangle Park, NC

U.S. Environmental Protection Agency
Office of Research and Development
National Center for Environmental Assessment
Washington, DC

Oak Ridge Institute for Scientific Education
Research Triangle Park, NC
Scientific Support Team
       Reeder Sams, Ph.D.
       John Stanek, Ph.D.
       Rob Dewoskin, Ph.D.
       George Woodall, Ph.D.
       Geniece Lehmann, Ph.D.
       Connie Meacham, M.S.
       Martin Gehlhaus1, MHS
       Nina Wang1, Ph.D.
U.S. Environmental Protection Agency
Office of Research and Development
National Center for Environmental Assessment
Research Triangle Park, NC
U.S. Environmental Protection Agency
Office of Research and Development
National Center for Environmental Assessment
Washington, DC
U.S. Environmental Protection Agency
Office of Research and Development
National Center for Environmental Assessment
Cincinnati, OH
1 Chemical Assessment Support Team (CAST) Member

        This document is a draft for review purposes only and does not constitute Agency policy.
                                            Vlll
         DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
Production Team
        Ellen Lorang, M.S.
        Deborah Wales
        Gerald Gurevich
        J. Sawyer Lucy
        Maureen Johnson
        Vicki Soto
U.S. Environmental Protection Agency
Office of Research and Development
National Center for Environmental Assessment
Research Triangle Park, NC
U.S. Environmental Protection Agency
Office of Research and Development
National Center for Environmental Assessment
Washington, DC
Contractor Support
        Battelle Memorial Institute, Pacific Northwest Division, Richmond, WA
        Karla D. Thrall, Ph.D.
        Battelle Memorial Institute, Columbus, OH
        Jessica D. Sanford, Ph.D.
        Maureen A. Wooton
        Robert A. Lordo, Ph.D.
        Anthony Fristachi

        Toxicology Excellence for Risk Assessment (TERA)
        Under Battelle Memorial Institute Contract EP-C-09-006
        Lisa M. Sweeney, Ph.D., DABT
        Melissa J. Kohrman-Vincent, B.A.
Executive Direction
        Kenneth Olden, Ph.D., Sc.D., L.H.D.
        (Center Director)
        Lynn Flowers, Ph.D., DABT
        (Associate Director for Health)
        Vincent Cogliano, Ph.D.2
        (IRIS Program Director—Acting)
        Samantha Jones, Ph.D.
        (IRIS Associate Director for Science)
        Jamie Strong, Ph.D.
        Ted Berner, MS
        Reeder Sams, Ph.D.
        Lyle Burgoon, Ph.D.
        John Vandenberg, Ph.D.
        Debra Walsh, M.S.
U.S. Environmental Protection Agency
Office of Research and Development
National Center for Environmental Assessment
Washington, DC
U.S. Environmental Protection Agency
Office of Research and Development
National Center for Environmental Assessment
Research Triangle Park, NC
2 Chemical Assessment Support Team (CAST) Lead

         This document is a draft for review purposes only and does not constitute Agency policy.

                                               ix             DRAFT—DO NOT CITE OR QUOTE

-------
                                                   Toxicological Review of Trimethylbenzene
Reviewers
This assessment was provided for review to scientists in EPA's Program and Regional Offices.
Comments were submitted by:
       Region/Office
       Region 8, Denver, CO
       Region 2, New York City, NY
       Office of Children's Health Protection, Washington, DC
       Office of Air and Radiation, Washington, DC
       Office of Environmental Information, Washington, DC
       Office of Solid Waste and Emergency Response, Washington, DC
       Office of Pesticides, Washington, DC
       Office of Water, Washington, DC
This assessment was provided for review to other federal agencies and the Executive Offices of the
President Comments were submitted by:
       Agency/ Office / Program
       Chemical Material Risk Management Program, Department of Defense
       Agency for Toxic Substances and Disease Registry, Centers for Disease Control and Prevention,
           Department of Health & Human Services
       The National Institute for Occupational Safety and Health, Centers for Disease Control and
           Prevention, Department of Health & Human Services
       National Toxicology Program, National Institute of Environmental Health Sciences, National
           Institutes of Health, Department of Health & Human Services
       Council on Environmental Quality, Executive Office of the President

This assessment was released for public  comment on June 26th, 2012 and comments were due on
August 28th, 2012. Comments were received from the following entities:
       Non-Government
       Leslie Berry                          .    .    „    . ,_  „    .,
        ....   „  J                         American Chemistry Council
       Allison Starmann

A public listening session was held by EPA on August 1st, 2012. Attendees external to the EPA are
listed below.
       Listening Session Attendees (Non-EPA)
       Leslia Berry                          .    .    „    . ,_  „    .,
       „.  ,  , „  ,                        American Chemistry Council
       Richard Becker
         This document is a draft for review purposes only and does not constitute Agency policy.

                                              x             DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
     PREFACE
 1          This Toxicological Review critically reviews the publicly available studies on the three
 2   isomers of trimethylbenzene (i.e., 1,2,3-trimethylbenzene [1,2,3-TMB], 1,2,4-trimethylbenzene
 3   [1,2,4-TMB], and 1,3,5-trimethylbenzene [1,3,5-TMB]) in order to identify their adverse health
 4   effects and to characterize exposure-response relationships. Because more types of studies are
 5   available for the 1,2,4-TMB isomer, it generally appears first when the individual isomers are listed.
 6   This assessment was prepared under the auspices of EPA's Integrated Risk Information System
 7   (IRIS) program.
 8          This assessment was prepared because of the presence of trimethylbenzenes (TMB) at
 9   Superfund sites. Of sites on EPA's National Priorities List that report TMB isomer contamination (38
10   sites), 93% report 1,3,5-TMB contamination, 85% report 1,2,4-TMB contamination, 12% report
11   1,2,3-TMB contamination, and 17% report contamination by unspecified TMB isomers.
12          The Toxicological Review of Trimethylbenzenes is a new assessment; there is no previous
13   entry on the IRIS Database for 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB. This assessment reviews
14   information on all health effects by all exposure routes.
15          This assessment was conducted in accordance with EPA guidance, which is cited and
16   summarized in the Preamble to IRIS Toxicological Reviews. The findings of this assessment and
17   related documents produced during its development are available on the IRIS website
18   (http://www.epa.gov/iris). Appendices for chemical and physical properties, toxicokinetic
19   information, summaries of toxicity studies, and other supporting materials are provided as
20   Supplemental Information (See Appendix A to C).

     Implementation of the 2011 National Research Council Recommendations
21          On December 23, 2011, The Consolidated Appropriations Act, 2012, was signed into law
22   (U.S. Congress, 2011). The report language included direction to EPA for the IRIS Program related
23   to recommendations provided by the National Research Council (NRC) in their review of EPA's
24   draft IRIS assessment of formaldehyde (NRC. 2011). The NRC's recommendations, provided in
25   Chapter 7 of the review report, offered suggestions to EPA for improving the development of IRIS
26   assessments. The report language included the following:
             This document is a draft for review purposes only and does not constitute Agency policy.

                                                 xi            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene

 1        The  Agency shall  incorporate, as appropriate, based  on chemical-specific datasets  and
 2        biological effects, the  recommendations of Chapter 7 of the National Research Council's
 3        Review of the Environmental Protection Agency's Draft IRIS Assessment of Formaldehyde
 4        into the IRIS process .... For draft assessments released in fiscal year 2012, the Agency shall
 5        include documentation describing how the Chapter 7 recommendations of the National
 6        Academy of Sciences (NAS) have been implemented or addressed, including an explanation
 7        for why certain recommendations were not incorporated.

 8          Consistent with the direction provided by Congress, documentation of how the
 9   recommendations from Chapter 7 of the NRC report have been implemented in this assessment is
10   provided in Appendix D. This documentation also includes an explanation for why certain
11   recommendations were not incorporated.
12          The IRIS Program's implementation of the NRC recommendations is following a phased
13   approach that is consistent with the NRC's "Roadmap for Revision" as described in Chapter 7 of the
14   formaldehyde review report. The NRC stated that, "the committee recognizes that the changes
15   suggested would involve a multi-year process and extensive effort by the staff at the National
16   Center for Environmental Assessment and input and review by the EPA Science Advisory Board and
17   others."
18          Phase 1 of implementation has focused on a subset of the short-term recommendations,
19   such as editing and streamlining documents, increasing transparency and clarity, and using more
20   tables, figures, and appendices to present information and data in assessments. Phase 1 also
21   focused on assessments near the end of the development process and close to final posting. The
22   IRIS TMBs assessment is one of the first assessments in Phase 2 of implementation, which
23   addresses all of the short-term NRC recommendations (see Appendix D, Table D-l). The IRIS
24   Program is implementing all  of these recommendations but recognizes that achieving full and
25   robust implementation of certain recommendations will be an evolving process with input and
26   feedback from the public, stakeholders, and external peer review committees. Phase 3 of
27   implementation will incorporate the longer-term recommendations made by the NRC as outlined in
28   Table D-2, including the development of a standardized approach to describe the strength of
29   evidence for noncancer effects. On May 16, 2012, EPA  announced [U.S. EPA. 2012c] that as a part of
30   a review of the IRIS Program's assessment development process, the NRC will also review current
31   methods for weight-of-evidence analyses and recommend approaches for weighing scientific
32   evidence for chemical hazard identification. This effort is included in Phase 3 of EPA's
33   implementation plan.

     Assessments by Other National and International Health Agencies
34          Toxicity information on 1,2,4-TMB, 1,2,3-TMB, and 1,3,5-TMB has been evaluated by the
35   National Institute for Occupational Safety and Health (NIOSH), the American Conference of

             This document is a draft for review purposes only and does not constitute Agency policy.

                                                xii           DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene

 1   Governmental Industrial Hygienists (ACGIH), the National Advisory Committee for Acute Exposure
 2   Guideline Levels for Hazardous Substances, and the Ontario Ministry of the Environment (MOE).
 3   The results of these assessments are summarized in Appendix A (Table A-l). It is important to
 4   recognize that these assessments may have been prepared for different purposes and may utilize
 5   different methods, and that newer studies may be included in the IRIS assessment.

     Chemical Properties and Uses
 6          TMBs are aromatic hydrocarbons with three methyl groups attached to a benzene ring and
 7   the chemical formula C9Hi2. The chemical and physical properties of the TMB isomers are similar to
 8   one another. TMBs are colorless, flammable liquids with a strong aromatic odor; an odor threshold
 9   of 0.4 parts per million (ppm) of air has been reported [U.S. EPA. 1994a). They are insoluble in
10   water but miscible with organic solvents such as ethyl alcohol, benzene, and ethyl ether [OSHA.
11   1996]. Production and use of TMBs may result in their release to the environment through various
12   waste streams. If released to the atmosphere, 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB will exist solely
13   in the vapor phase in the atmosphere under ambient conditions, based on measured vapor
14   pressures of 1.69, 2.10, and 2.48 mm Hg at 25°C, respectively [HSDB. 2011a. b, c). All three isomers
15   are expected to have limited mobility through soil based on their Log Koc values, but are expected to
16   volatilize from both moist and dry soil surfaces and surface waters based on their respective
17   Henry's law constants and vapor pressures (see Appendix B, Table B-l). Degradation of both
18   isomers in the atmosphere occurs by reaction with hydroxyl radicals, the half-life of which is 11-12
19   hours (HSDB. 2011a. b, c). Non-volatilized TMBs may be subject to biodegradation under aerobic
20   conditions (HSDB. 2011a.  b, c). The estimated bio-concentration factors (133-439) and high
21   volatility of TMBs suggest that bioaccumulation of these chemicals will not be significant (U.S. EPA,
22   1987). Additional information on the  chemical identities and physicochemical properties of TMBs
23   are listed in Table B-l in Appendix B.
24          The commercially  available substance known as trimethylbenzene, CAS No. 25551-13-7, is a
25   mixture of three isomers in various proportions, namely CAS No. 526-73-8 (1,2,3-TMB or
26   hemimellitene), CAS No. 95-63-6 (1,2,4-TMB or pseudocumene), and CAS No. 108-67-8 (1,3,5-TMB
27   or mesitylene). Production of TMB isomers occurs during petroleum refining, and 1,2,4-TMB
28   individually makes up approximately 40% of the C9 aromatic fraction (i.e., aromatic hydrocarbons
29   with nine carbons) (U.S. EPA. 1994a). The domestic production of the C9 fraction in 1991 was
30   estimated to be approximately 80 billion pounds (40 million tons) (U.S. EPA. 1994a). Vehicle
31   emissions are a major anthropogenic source of TMBs, due to the widespread use of the C9 fraction
32   as a component of gasoline (U.S. EPA.  1994a). Other uses of TMBs include solvents in research and
33   industry, dyestuff intermediate, paint thinner, and as a UV oxidation stabilizer for plastics (HSDB.
34   2011b. c).


             This document is  a draft for review purposes only and does not constitute Agency policy.

                                                 xiii           DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene

1          Occupational levels of exposure for TMBs have been measured between 20-8,540 |ig/m3
2   [HSDB, 2011a, b, c; Tiun-Horngetal., 2008], whereas residential exposures are generally much
3   lower: 0.29-7.8 [ig/m3 [Martins etal.. 2010: Choi etal.. 2009: Guo etal.. 2009]. Total atmospheric
4   releases of 1,2,4-TMB to the environment in 2008 equaled 5.8 million pounds (2,900 tons], 265,000
5   pounds (132.5 tons] were released to surface waters, underground injection sites, or land (TRI.
6   2008]. No information is currently available regarding 1,2,3-TMB or 1,3,5-TMB releases.
7          For additional information about this assessment or for general questions regarding IRIS,
8   please contact EPA's IRIS Hotline at 202-566-1676 (phone], 202-566-1749 (fax], or
9   hotline.iris@epa.gov.
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                xiv           DRAFT—DO NOT CITE OR QUOTE

-------
                                                    Toxicological Review of Trimethylbenzene
    PREAMBLE TO  IRIS TOXICOLOGICAL  REVIEWS
    1.  Scope of the IRIS Program

 1     Soon after the EPA was established in
 2  1970, it was at the forefront of developing
 3  risk assessment as a science and applying it
 4  in decisions to protect human health and the
 5  environment. The Clean Air Act, for example,
 6  mandates that the EPA provide "an ample
 7  margin of safety  to protect public health";
 8  the  Safe  Drinking Water  Act,  that  "no
 9  adverse effects on the health of persons may
10  reasonably be anticipated to occur, allowing
11  an adequate margin of safety." Accordingly,
12  the  EPA uses information  on  the  adverse
13  effects of chemicals and on exposure levels
14  below  which  these   effects  are   not
15  anticipated to occur.
16     IRIS  assessments  critically review  the
17  publicly available  studies to identify adverse
18  health effects from exposure to  chemicals
19  and  to  characterize   exposure-response
20  relationships.  In  terms set forth by  the
21  National Research Council [NRC. 1983). IRIS
22  assessments cover the hazard identification
23  and dose-response assessment steps of risk
24  assessment, not the exposure assessment or
25  risk   characterization   steps  that   are
26  conducted   by the  EPA's  program  and
27  regional offices and by other federal, state,
28  and local health agencies that evaluate risk
29  in   specific   populations   and  exposure
30  scenarios. IRIS assessments are distinct from
31  and do not  address political, economic, and
32  technical considerations that influence  the
33  design  and selection  of risk management
34  alternatives.
35     An IRIS assessment may cover  a single
36  chemical,   a  group   of   structurally   or
37  toxicologically  related  chemicals,   or  a
38  complex mixture. These agents may be found
39  in air, water, soil, or sediment Exceptions
40  are chemicals currently used exclusively as
41  pesticides,  ionizing  and   non-ionizing
42  radiation, and criteria air pollutants listed
43  under Section  108  of the  Clean Air  Act
44  (carbon  monoxide, lead,  nitrogen  oxides,
45  ozone, particulate matter, and sulfur oxides).
46     Periodically, the IRIS Program asks other
47  EPA  programs  and  regions, other  federal
48  agencies,  state  health agencies,  and  the
49  general public to  nominate chemicals  and
50  mixtures   for    future   assessment   or
51  reassessment Agents may be considered for
52  reassessment as significant new studies are
53  published. Selection  is based  on program
54  and  regional  office  priorities  and  on
55  availability  of  adequate   information  to
56  evaluate the potential for  adverse  effects.
57  Other agents  may  also  be  assessed  in
58  response to an urgent public health need.

    2.  Process for developing and peer-
       reviewing IRIS assessments

59     The   process   for   developing  IRIS
60  assessments  (revised  in  May 2009  and
61  enhanced  in July 2013)  involves  critical
62  analysis    of    the    pertinent   studies,
63  opportunities for public input,  and multiple
64  levels of scientific review. The EPA revises
65  draft assessments after each  review,  and
66  external drafts and comments  become part
67  of the public record (U.S. EPA. 2009).
68     Before beginning an assessment, the IRIS
69  Program discusses the scope with other EPA
70  programs and regions to ensure that the
71  assessment will  meet their needs. Then  a
72  public  meeting  on  problem  formulation
73  invites discussion of  the key issues and the
          This document is a draft for review purposes only and does not constitute Agency policy.

                                              xv            DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
 1  studies and analytical approaches that might
 2  contribute to their resolution.
 3  Step 1.    Development   of   a   draft
 4     Toxicological   Review.    The   draft
 5     assessment   considers  all  pertinent
 6     publicly available  studies  and  applies
 7     consistent  criteria  to  evaluate  study
 8     quality,  identify  health effects, identify
 9     mechanistic  events  and   pathways,
10     integrate the evidence of causation for
11     each effect, and derive toxicity values. A
12     public meeting prior to the integration of
13     evidence  and  derivation  of toxicity
14     values promotes public discussion of the
15     literature  search,  evidence,  and  key
16     issues.
17  Step 2.  Internal review by scientists in
18     EPA  programs and regions.  The draft
19     assessment  is revised  to   address the
20     comments from within the EPA.
21  Step 3.  Interagency science consultation
22     with other federal agencies and the
23     Executive Offices of the President. The
24     draft assessment is revised to address
25     the interagency comments.  The science
26     consultation     draft,      interagency
27     comments, and the EPA's  response to
28     major  comments become  part of the
29     public record.
30  Step 4.  Public  review  and   comment,
31     followed by external peer  review. The
32     EPA  releases the draft assessment for
33     public  review and  comment A  public
34     meeting  provides   an  opportunity to
35     discuss  the assessment prior to peer
36     review. Then the  EPA releases a draft for
37     external peer review. The peer review
38     meeting  is  open  to  the  public  and
39     includes time for oral public comments.
40     The peer reviewers assess whether the
41     evidence  has  been  assembled  and
42     evaluated  according to guidelines  and
43     whether the conclusions are justified by
44     the evidence.  The  peer  review draft,
45     written public  comments,  and  peer
46     review report become part of the public
47     record.
48  Step 5.  Revision of  draft  Toxicological
49     Review and development of draft IRIS
50     summary.  The  draft  assessment  is
51     revised to  reflect  the  peer  review
52     comments, public comments, and newly
53     published studies that are critical to the
54     conclusions  of  the  assessment  The
55     disposition of peer  review comments
56     and public comments becomes  part  of
57     the public record.
58  Step 6. Final EPA review and interagency
59     science discussion with other  federal
60     agencies and the Executive Offices  of
61     the President The draft assessment and
62     summary are revised  to address the EPA
63     and interagency comments. The  science
64     discussion  draft, written  interagency
65     comments, and EPA's response to major
66     comments become part  of the  public
67     record.
68  Step 7.  Completion  and  posting.  The
69     Toxicological Review  and IRIS summary
70     are  posted  on  the  IRIS  website
71     [http://www.epa.gov/iris/].
72  The remainder of this Preamble addresses
73     step 1, the  development  of a  draft
74     Toxicological Review. IRIS  assessments
75     follow  standard  practices  of evidence
76     evaluation and peer review, many  of
77     which  are discussed  in EPA guidelines
78     fU.S. EPA. 2005a. b,  2000.  1998. 1996.
79     1991. 1986a. b) and other methods [U.S.
80     EPA. 2012a. b, 2011.  2006a. b, 2002.
81     1994b).  Transparent  application   of
82     scientific   judgment  is  of paramount
83     importance.  To provide a  harmonized
84     approach across  IRIS assessments, this
85     Preamble  summarizes   concepts  from
86     these  guidelines    and   emphasizes
87     principles of general applicability.
          This document is a draft for review purposes only and does not constitute Agency policy.

                                              xvi             DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
    3.  Identifying and selecting
       pertinent studies

 1  3.1. Identifying studies

 2     Before beginning an assessment, the EPA
 3  conducts a  comprehensive  search  of the
 4  primary scientific literature. The literature
 5  search   follows  standard   practices  and
 6  includes the PubMed and ToxNet databases
 7  of the National Library  of Medicine, Web of
 8  Science, and other databases listed in the
 9  EPA's    HERO    system    (Health   and
10  Environmental      Research       Online,
11  http://hero.epa.gov/].     Searches     for
12  information  on  mechanisms of toxicity are
13  inherently  specialized   and  may  include
14  studies  on  other agents that act  through
15  related mechanisms.
16     Each assessment specifies  the  search
17  strategies, keywords, and cut-off dates  of its
18  literature  searches.  The  EPA posts  the
19  results of the literature search on the IRIS
20  web site and requests information from the
21  public  on additional studies and  ongoing
22  research.
23     The EPA also considers studies received
24  through  the IRIS  Submission Desk and
25  studies  (typically unpublished)  submitted
26  under the Toxic Substances Control Act or
27  the  Federal  Insecticide,   Fungicide,  and
28  Rodenticide  Act. Material  submitted  as
29  Confidential   Business   Information  is
30  considered  only if it includes  health and
31  safety data that can be publicly released. If a
32  study that may be critical to the conclusions
33  of  the  assessment  has not  been  peer-
34  reviewed,  the  EPA will  have  it  peer-
35  reviewed.
36     The EPA also examines the toxicokinetics
37  of the agent to identify  other chemicals (for
38  example, major  metabolites of the agent) to
39  include  in  the  assessment  if adequate
40  information  is available, in  order to more
41  fully explain the toxicity of the agent and to
42  suggest  dose   metrics   for   subsequent
43  modeling.
44     In assessments  of chemical  mixtures,
45  mixture  studies  are  preferred  for  their
46  ability   to  reflect   interactions  among
47  components.
48     The literature search seeks, in
49  decreasing order of preference (U.S. EPA.
50  2000. 32.2: 1986b. 32.1)]:
51     -  Studies  of   the  mixture   being
52        assessed.
53     -  Studies  of  a  sufficiently  similar
54        mixture. In evaluating similarity, the
55        assessment considers  the  alteration
56        of  mixtures  in  the  environment
57        through      partitioning       and
58        transformation.
59     -  Studies   of    individual   chemical
60        components of the mixture, if there
61        are  not   adequate   studies   of
62        sufficiently similar mixtures.

63  3.2.  Selecting pertinent epidemiologic
64       studies
65     Study design is the key consideration for
66  selecting  pertinent  epidemiologic  studies
67  from the results of the literature search.
68
69
70
71
72
73
74

75
76
77
78
79
80
81
82
83
Cohort studies, case-control studies,
and some population-based surveys
(for example, NHANES) provide the
strongest  epidemiologic  evidence,
especially if they collect information
about  individual   exposures  and
effects.
Ecological    studies    (geographic
correlation studies) relate exposures
and effects by geographic area. They
can provide strong evidence if there
are   large    exposure   contrasts
between geographic areas, relatively
little exposure variation within study
areas, and population migration is
limited.
          This document is a draft for review purposes only and does not constitute Agency policy.

                                               xvii            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1     -   Case reports  of high or accidental
 2         exposure  lack  definition   of  the
 3         population at risk and the expected
 4         number of cases. They can provide
 5         information about  a rare effect or
 6         about the relevance  of analogous
 7         results in animals.

 8     The   assessment  briefly   reviews
 9  ecological  studies  and case  reports  but
10  reports details  only if they suggest effects
11  not identified by other studies.

12  3.3.  Selecting pertinent experimental
13       studies

14     Exposure  route    is  a   key  design
15  consideration   for    selecting   pertinent
16  experimental  animal  studies  or  human
17  clinical studies.

18     -   Studies of oral, inhalation, or dermal
19         exposure involve passage through an
20         absorption    barrier   and   are
21         considered most pertinent to human
22         environmental exposure.
23     -   Injection or implantation studies are
24         often considered less pertinent but may
25         provide  valuable  toxicokinetic   or
26         mechanistic information. They also may
27         be  useful  for  identifying  effects in
28         animals if deposition or absorption  is
29         problematic (for example, for particles
30         and fibers).

31     Exposure duration is also a key design
32  consideration   for    selecting   pertinent
33  experimental animal studies.

34     -   Studies  of  effects  from   chronic
35         exposure  are  most  pertinent to
36         lifetime human exposure.

37     -   Studies  of  effects   from less-than-
38         chronic  exposure are pertinent but
39         less preferred  for identifying effects
40         from lifetime human exposure. Such
41         studies may be indicative of effects
42         from    less-than-lifetime    human
43         exposure.
44     Short-duration studies involving animals
45  or humans  may  provide  toxicokinetic or
46  mechanistic information.
47     For    developmental    toxicity    and
48  reproductive  toxicity,  irreversible effects
49  may result from a brief exposure  during a
50  critical period of development Accordingly,
51  specialized study designs are used for these
52  effects [U.S. EPA. 2006b. 1998. 1996. 19911.

    4. Evaluating the quality of
       individual studies

53     After   the    subsets    of   pertinent
54  epidemiologic   and  experimental  studies
55  have  been  selected  from  the  literature
56  searches,  the  assessment  evaluates  the
57  quality  of  each  individual  study.  This
58  evaluation considers the design,  methods,
59  conduct, and documentation of each study,
60  but not whether the results  are  positive,
61  negative, or null. The objective is to identify
62  the stronger, more informative studies based
63  on  a  uniform   evaluation   of  quality
64  characteristics  across  studies of similar
65  design.

66  4.1. Evaluating the quality of
67      epidemiologic studies
68     The  assessment  evaluates  design  and
69  methodological  aspects that can increase or
70  decrease  the   weight  given   to   each
71  epidemiologic study in the overall evaluation
72  [U.S. EPA. 2005a. 1998.1996.1994b. 19911:
73     -   Documentation   of   study  design,
74         methods, population  characteristics,
75         and results.
76     -   Definition and selection of the study
77         group and comparison group.
78     -   Ascertainment of exposure to the
79         chemical or mixture.
80     -   Ascertainment of disease  or health
81         effect.
          This document is a draft for review purposes only and does not constitute Agency policy.

                                               xviii             DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1     -   Duration of exposure and follow-up
 2         and  adequacy  for  assessing the
 3         occurrence of effects.
 4     -   Characterization of exposure during
 5         critical periods.
 6     -   Sample size and statistical power  to
 7         detect anticipated effects.
 8     -   Participation rates and potential for
 9         selection bias  as  a result of the
10         achieved participation rates.
11     -   Measurement  error (can  lead  to
12         misclassification of exposure, health
13         outcomes, and  other  factors) and
14         other types of information bias.
15     -   Potential  confounding  and   other
16         sources  of bias addressed in the
17         study  design or in the analysis  of
18         results. The  basis for consideration
19         of  confounding  is  a  reasonable
20         expectation that the confounder  is
21         related   to   both  exposure   and
22         outcome and is sufficiently prevalent
23         to result in bias.
24     For developmental toxicity, reproductive
25  toxicity, neurotoxicity,  and cancer there  is
26  further  guidance  on  the   nuances   of
27  evaluating epidemiologic studies  of  these
28  effects fU.S. EPA. 2005a. 1998. 1996. 19911

29  4.2.  Evaluating the quality of
30       experimental studies
31     The assessment  evaluates design and
32  methodological aspects that can increase  or
33  decrease  the   weight  given   to   each
34  experimental animal study, in-vitro study,  or
35  human clinical study [U.S. EPA. 2005a.  1998.
36  1996.  1991).  Research  involving human
37  subjects is considered only if  conducted
38  according to ethical principles.
39     -   Documentation  of   study  design,
40         animals    or   study   population,
41         methods, basic data,  and results.
42     -   Nature of the assay and validity for
43         its intended purpose.
44     -   Characterization of the nature and
45         extent     of    impurities     and
46         contaminants  of  the  administered
47         chemical or mixture.
48     -   Characterization of dose and dosing
49         regimen (including age at exposure)
50         and their adequacy to elicit  adverse
51         effects, including latent effects.
52     -   Sample sizes and statistical power to
53         detect  dose-related  differences  or
54         trends.
55     -   Ascertainment of survival, vital signs,
56         disease or effects, and cause of death.
57     -   Control of other variables that could
58         influence the occurrence of effects.
59     The assessment uses  statistical  tests to
60  evaluate whether the observations  may be
61  due to chance. The  standard for determining
62  statistical  significance  of a response  is a
63  trend test or comparison of outcomes in the
64  exposed groups against those of concurrent
65  controls. In some situations, examination of
66  historical  control   data   from  the  same
67  laboratory within a few years of the  study
68  may improve the analysis. For an uncommon
69  effect that is  not statistically  significant
70  compared   with    concurrent   controls,
71  historical controls may show that the  effect
72  is  unlikely to  be  due to  chance.  For a
73  response that appears significant against a
74  concurrent control response that is unusual,
75  historical  controls  may  offer a different
76  interpretation fU.S. EPA. 2005a. §2.2.2.1.3).
77     For developmental toxicity, reproductive
78  toxicity, neurotoxicity, and cancer there is
79  further  guidance   on  the  nuances   of
80  evaluating  experimental   studies of  these
81  effects  (U.S. EPA. 2005a. 1998. 1996. 1991).
82  In  multi-generation  studies,  agents  that
83  produce developmental effects at doses that
84  are not toxic  to the maternal animal are of
85  special concern. Effects that occur at  doses
86  associated with mild maternal toxicity  are
87  not assumed  to  result only from maternal
88  toxicity. Moreover, maternal effects  may be
89  reversible, while effects on the offspring may
          This document is a draft for review purposes only and does not constitute Agency policy.

                                               xix             DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1  be permanent [U.S.  EPA. 1998.33.1.2.4.5.4:
 2  1991 §3.1.1.4),.

 3  4.3. Reporting study results
 4     The assessment  uses evidence tables to
 5  present  the  design and key results  of
 6  pertinent studies. There may be separate
 7  tables for each site of toxicity or type of
 8  study.
 9     If a large number of studies observe the
10  same effect, the assessment  considers the
11  study quality characteristics  in this  section
12  to identify the strongest studies or types of
13  study. The tables present details from these
14  studies,  and  the  assessment  explains the
15  reasons  for  not reporting details of other
16  studies or groups of studies that do not add
17  new information. Supplemental information
18  provides    references   to    all    studies
19  considered, including those not summarized
20  in the tables.
21     The assessment discusses strengths and
22  limitations that affect the interpretation of
23  each study. If the interpretation of a study in
24  the assessment differs from that of the study
25  authors,  the assessment  discusses the  basis
26  for the difference.
27     As  a  check  on the   selection  and
28  evaluation of pertinent studies, the EPA asks
29  peer reviewers to identify studies that  were
30  not adequately considered.

    5.  Evaluating the overall evidence
       of each effect

31  5.1. Concepts of causal inference
32     For each  health effect, the assessment
33  evaluates  the  evidence as  a  whole  to
34  determine whether it is reasonable to infer a
35  causal association between exposure to the
36  agent and the  occurrence of the effect. This
37  inference is  based on  information  from
38  pertinent human studies, animal studies, and
39  mechanistic  studies  of  adequate  quality.
40  Positive, negative, and null results  are given
41  weight according to study quality.
42     Causal  inference  involves   scientific
43  judgment,  and  the   considerations  are
44  nuanced   and  complex.   Several  health
45  agencies  have developed frameworks for
46  causal  inference,  among  them  the  U.S.
47  Surgeon General fCDC. 2004: HEW. 19641
48  the  International Agency for Research on
49  Cancer  [IARC.   2006],  the  Institute  of
50  Medicine  [IOM.   2008],   and   the   EPA
51  f2010.31.6:     2005a.§2.51     Although
52  developed  for   different  purposes,  the
53  frameworks  are  similar  in nature  and
54  provide   an  established  structure   and
55  language  for   causal  inference.   Each
56  considers  aspects  of  an  association  that
57  suggest causation, discussed by  Hill [1965]
58  and elaborated by  Rothman and Greenland
59  (1998),   and  U.S.  EPA  (2005a, §2.2.1.7;
60  1994b, Appendix C).
61  Strength of association: The finding  of a
62     large  relative   risk  with   narrow
63     confidence intervals  strongly  suggests
64     that  an association is not due to chance,
65     bias,  or  other  factors. Modest relative
66     risks, however, may reflect a small range
67     of exposures, an agent of low potency, an
68     increase  in  an effect that is  common,
69     exposure  misclassification,   or  other
70     sources of bias.
71  Consistency of association: An inference of
72     causation is  strengthened  if  elevated
73     risks   are  observed  in independent
74     studies of  different populations  and
75     exposure  scenarios. Reproducibility of
76     findings constitutes one of the strongest
77     arguments  for  causation.  Discordant
78     results sometimes reflect differences in
79     study design, exposure, or confounding
80     factors.
81  Specificity of association:  As  originally
82     intended,  this   refers  to  one  cause
83     associated  with  one  effect.  Current
84     understanding that many agents cause
85     multiple  effects and many effects have
86     multiple   causes   make  this  a   less
87     informative  aspect  of causation, unless
          This document is a draft for review purposes only and does not constitute Agency policy.

                                               xx             DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1     the effect is rare or unlikely  to  have
 2     multiple causes.
 3  Temporal    relationship:    A   causal
 4     interpretation requires  that  exposure
 5     precede development of the effect.
 6  Biologic   gradient   (exposure-response
 7     relationship):       Exposure-response
 8     relationships strongly suggest causation.
 9     A  monotonic increase is not the only
10     pattern consistent with  causation.  The
11     presence   of  an   exposure-response
12     gradient also weighs against bias  and
13     confounding  as   the  source   of  an
14     association.
15  Biologic   plausibility:  An  inference  of
16     causation  is  strengthened   by  data
17     demonstrating     plausible    biologic
18     mechanisms, if  available.  Plausibility
19     may reflect subjective  prior beliefs  if
20     there is insufficient understanding of the
21     biologic process involved.
22  Coherence:  An  inference  of  causation  is
23     strengthened by supportive results from
24     animal    experiments,     toxicokinetic
25     studies, and short-term tests.  Coherence
26     may also be found  in  other  lines of
27     evidence,  such   as   changing   disease
28     patterns in the population.
29  "Natural   experiments":  A  change  in
30     exposure that brings about a change in
31     disease  frequency   provides   strong
32     evidence,  as  it tests  the hypothesis of
33     causation.  An  example  would be  an
34     intervention to reduce exposure in the
35     workplace   or   environment  that   is
36     followed by a reduction of an adverse
37     effect.
38  Analogy:    Information    on   structural
39     analogues or on  chemicals that induce
40     similar mechanistic events can provide
41     insight into causation.
42     These considerations are consistent with
43  guidelines   for   systematic  reviews  that
44  evaluate the quality and weight of evidence.
45  Confidence is increased if the magnitude of
46  effect is large,  if there is  evidence of an
47  exposure-response relationship, or  if  an
48  association was observed and the plausible
49  biases would tend to decrease the magnitude
50  of   the  reported effect.  Confidence   is
51  decreased     for    study    limitations,
52  inconsistency  of results,  indirectness   of
53  evidence,  imprecision,   or  reporting bias
54  [Guyattetal.. 2008b: Guyattetal.. 2008a).

55  5.2. Evaluating evidence in humans

56     For each effect, the assessment evaluates
57  the evidence from the epidemiologic studies
58  as a whole. The objective  is  to determine
59  whether a credible  association has  been
60  observed and, if so, whether that association
61  is consistent with causation. In  doing this,
62  the    assessment    explores    alternative
63  explanations  (such   as  chance,  bias,  and
64  confounding) and draws a conclusion about
65  whether these alternatives can satisfactorily
66  explain any observed  association.
67     To   make   clear   how    much   the
68  epidemiologic evidence  contributes to the
69  overall  weight  of   the  evidence,  the
70  assessment may select a standard descriptor
71  to characterize  the epidemiologic evidence
72  of association between exposure to the agent
73  and occurrence of a health effect.
74  Sufficient  epidemiologic evidence of  an
75     association consistent with  causation:
76     The  evidence  establishes  a  causal
77     association   for    which    alternative
78     explanations such as chance, bias, and
79     confounding  can  be ruled out  with
80     reasonable confidence.
81  Suggestive epidemiologic  evidence of an
82     association consistent with  causation:
83     The    evidence   suggests   a   causal
84     association   but   chance,   bias,   or
85     confounding  cannot  be ruled  out  as
86     explaining the association.
87  Inadequate epidemiologic  evidence  to
88     infer a causal association: The available
89     studies do  not  permit  a  conclusion
          This document is a draft for review purposes only and does not constitute Agency policy.

                                               xxi             DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1     regarding the presence or absence of an
 2     association.
 3  Epidemiologic evidence consistent with no
 4     causal  association:  Several  adequate
 5     studies covering the full range of human
 6     exposures  and considering susceptible
 7     populations, and for which alternative
 8     explanations   such   as   bias   and
 9     confounding can   be  ruled  out,  are
10     mutually  consistent in not finding an
11     association.

12  5.3.  Evaluating evidence in animals
13     For each effect, the assessment evaluates
14  the evidence from the animal experiments as
15  a whole to determine  the extent to which
16  they  indicate a  potential for  effects in
17  humans.  Consistent results across various
18  species and strains increase confidence that
19  similar  results would  occur in  humans.
20  Several  concepts discussed by Hill [1965]
21  are pertinent to the weight of experimental
22  results:   consistency  of  response,   dose-
23  response relationships,  strength of response,
24  biologic  plausibility, and coherence  [U.S.
25  EPA.  2005a. §2.2.1.7: 1994b. Appendix Cj.
26     In  weighing  evidence
27  experiments,    U.S.   EPA
28  distinguishes:
from  multiple
  [2005a. §2.5)
29  Conflicting evidence (that is, mixed positive
30     and negative results in the same sex and
31     strain  using a similar study protocol)
32     from
33  Differing  results (that is, positive results
34     and  negative  results  are  in  different
35     sexes or strains  or use different study
36     protocols).
37     Negative or null results do not invalidate
38  positive results in a different experimental
39  system.  The  EPA   regards  all  as  valid
40  observations and  looks to explain  differing
41  results using mechanistic information (for
42  example,     physiologic    or    metabolic
43  differences   across   test  systems)    or
44  methodological differences  (for example,
45  relative sensitivity of the tests, differences in
46  dose  levels,  insufficient  sample  size,  or
47  timing of dosing or data collection).
48     It is  well established that there  are
49  critical periods for some developmental and
50  reproductive  effects   (U.S.  EPA.  2006b.
51  2005a. b,  1998. 1996. 1991). Accordingly,
52  the assessment determines whether critical
53  periods have  been adequately investigated.
54  Similarly,   the   assessment   determines
55  whether  the  database   is  adequate  to
56  evaluate other critical sites and effects.
57     In  evaluating  evidence   of  genetic
58  toxicity:
59     -  Demonstration  of gene  mutations,
60        chromosome     aberrations,     or
61        aneuploidy    in    humans     or
62        experimental    mammals   [in vivo]
63        provides the strongest evidence.
64     -  This is followed by positive results in
65        lower  organisms or in cultured cells
66        [in vitro] or for other genetic events.
67     -  Negative  results carry less weight,
68        partly because  they cannot exclude
69        the  possibility  of effects  in  other
70        tissues [IARC. 2006).
71     For germ-cell mutagenicity, The EPA has
72  defined categories of evidence, ranging from
73  positive    results  of   human   germ-cell
74  mutagenicity  to  negative results for  all
75  effects of concern [U.S. EPA. 1986a. §2.3).

76  5.4.  Evaluating mechanistic data

77     Mechanistic  data  can  be  useful  in
78  answering several questions.
                   79
                   80

                   81
                   82

                   83
                   84
           The biologic plausibility of a causal
           interpretation of human studies.
       -   The   generalizability
           studies to humans.
of   animal
       -   The   susceptibility  of   particular
           populations or lifestages.
                   85     The focus of the analysis is to describe, if
                   86 possible, mechanistic pathways that lead to a
                   87 health effect. These pathways encompass:
          This document is a draft for review purposes only and does not constitute Agency policy.

                                               xxii             DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1     -   Toxicokinetic processes of absorption,
 2         distribution,     metabolism,     and
 3         elimination   that   lead  to   the
 4         formation of an active agent and its
 5         presence at the site of initial biologic
 6         interaction.
 7     -   Toxicodynamic processes that lead to
 8         a health effect at this or another site
 9         (also known as a mode of action}.
10     For each effect, the assessment discusses
11  the  available  information  on its  modes of
12  action and associated key events (key events
13  being  empirically  observable,  necessary
14  precursor steps or biologic markers  of such
15  steps; mode of action being a series of key
16  events  involving  interaction   with  cells,
17  operational  and  anatomic  changes,   and
18  resulting in disease). Pertinent information
19  may  also come from studies of metabolites
20  or of compounds that are structurally similar
21  or  that act through  similar mechanisms.
22  Information on  mode  of  action  is  not
23  required for a conclusion that the agent is
24  causally related to an effect [U.S. EPA. 2005a.
25  §2.5).
26     The  assessment   addresses   several
27  questions about each hypothesized mode of
28  actionfU.S. EPA. 2005a. §2.4.3.4).
29  1)  Is the  hypothesized mode of action
30     sufficiently supported in test animals?
31     Strong support  for a key  event being
32     necessary to a mode of action can come
33     from  experimental  challenge  to  the
34     hypothesized mode of action, in which
35     studies  that  suppress  a  key  event
36     observe   suppression   of   the  effect.
37     Support  for   a  mode  of  action is
38     meaningfully strengthened by consistent
39     results in different experimental models,
40     much   more   so   than  by  replicate
41     experiments  in  the same  model.  The
42     assessment  may   consider   various
43     aspects  of causation in addressing this
44     question.
45  2)  Is the  hypothesized mode of action
46     relevant to humans? The  assessment
47     reviews the key events to identify critical
48     similarities and differences between the
49     test    animals    and   humans.    Site
50     concordance  is  not assumed between
51     animals and humans, though it may hold
52     for certain effects or modes of action.
53     Information   suggesting   quantitative
54     differences in doses where effects would
55     occur   in   animals  or   humans   is
56     considered   in   the    dose-response
57     analysis.  Current  levels   of  human
58     exposure are not used to rule out human
59     relevance, as IRIS assessments may be
60     used in evaluating new or unforeseen
61     circumstances that may  entail  higher
62     exposures.
63  3)  Which  populations or lifestages  can
64     be  particularly  susceptible  to  the
65     hypothesized  mode of  action?  The
66     assessment reviews the  key events to
67     identify populations and lifestages that
68     might be susceptible to their occurrence.
69     Quantitative differences  may  result in
70     separate toxicity  values for susceptible
71     populations or lifestages.
72     The assessment discusses the likelihood
73  that  an agent operates  through multiple
74  modes of action. An uneven level of support
75  for different  modes  of action can reflect
76  disproportionate       resources      spent
77  investigating      them     [U.S.      EPA.
78  2005a. §2.4.3.3). It should be  noted that in
79  clinical reviews, the credibility  of a series of
80  studies is  reduced if evidence  is limited to
81  studies funded by  one  interested sector
82  fGuyattetal.. 2008a).
83     For cancer, the  assessment  evaluates
84  evidence of a mutagenic mode of action to
85  guide extrapolation  to lower doses  and
86  consideration of  susceptible lifestages.  Key
87  data  include the ability of the agent or a
88  metabolite  to react with or bind to DNA,
89  positive results in multiple test systems, or
90  similar properties  and  structure-activity
91  relationships to mutagenic carcinogens [U.S.
92  EPA.  2005a.§2.3.5).
          This document is a draft for review purposes only and does not constitute Agency policy.

                                              xxiii            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1  5.5.  Characterizing the overall weight
 2       of the evidence

 3     After evaluating the human, animal, and
 4  mechanistic evidence pertinent to an effect,
 5  the assessment answers the question: Does
 6  the agent  cause  the  adverse effect? [NRG.
 7  2009. 19831.  In doing this, the assessment
 8  develops a  narrative that  integrates the
 9  evidence pertinent to  causation. To provide
10  clarity   and   consistency,   the   narrative
11  includes a standard hazard  descriptor. For
12  example, the following standard descriptors
13  combine epidemiologic,  experimental,  and
14  mechanistic   evidence  of  carcinogenicity
15  fU.S. EPA. 2005a. 32.51.
16  Carcinogenic  to  humans:   There   is
17     convincing epidemiologic evidence  of a
18     causal  association  (that  is,  there  is
19     reasonable    confidence   that   the
20     association cannot be fully explained by
21     chance, bias, or confounding); or there is
22     strong human evidence of cancer or its
23     precursors, extensive  animal  evidence,
24     identification of key precursor events in
25     animals,  and  strong  evidence that they
26     are anticipated to occur in humans.
27  Likely to be carcinogenic to humans: The
28     evidence   demonstrates   a  potential
29     hazard to humans but does not meet the
30     criteria for carcinogenic. There may be a
31     plausible   association   in   humans,
32     multiple positive results in animals, or a
33     combination of human, animal, or other
34     experimental evidence.
35  Suggestive   evidence   of  carcinogenic
36     potential: The evidence raises concern
37     for effects in humans but is not sufficient
38     for  a   stronger  conclusion.   This
39     descriptor covers  a  range of evidence,
40     from a  positive  result in  the only
41     available study to a single positive result
42     in an extensive database that includes
43     negative results in other species.
44  Inadequate    information    to   assess
45     carcinogenic   potential:   No   other
46     descriptors apply. Conflicting evidence
47     can   be   classified    as   inadequate
48     information  if  all positive  results are
49     opposed by  negative  studies of equal
50     quality  in the  same  sex  and strain.
51     Differing  results,  however,   can  be
52     classified  as  suggestive evidence or as
53     likely to be carcinogenic.
54  Not likely to be carcinogenic to humans:
55     There is robust evidence for concluding
56     that there is no basis for concern. There
57     may be no effects in both sexes of at least
58     two  appropriate animal species; positive
59     animal  results and  strong,  consistent
60     evidence that each mode  of  action in
61     animals does not operate in humans; or
62     convincing evidence that effects are not
63     likely by a particular exposure route or
64     below a defined dose.
65     Multiple descriptors may be used if there
66  is evidence that carcinogenic effects differ by
67  dose range or  exposure  route  [U.S.  EPA.
68  2005a.32.5).
69     Another example of standard descriptors
70  comes  from the EPA's  Integrated Science
71  Assessments,  which  evaluate  causation for
72  the  effects of  the  criteria  pollutants  in
73  ambient air [U.S. EPA. 2010. §1.6).
74  Causal relationship: Sufficient  evidence to
75     conclude  that   there  is   a  causal
76     relationship.    Observational    studies
77     cannot  be   explained  by  plausible
78     alternatives,  or they are  supported  by
79     other lines  of evidence,  for  example,
80     animal    studies   or    mechanistic
81     information.
82  Likely   to  be   a   causal  relationship:
83     Sufficient   evidence   that   a  causal
84     relationship   is  likely, but  important
85     uncertainties   remain.  For  example,
86     observational    studies    show    an
87     association but co-exposures are difficult
88     to address or other lines of evidence are
89     limited  or   inconsistent;   or   multiple
90     animal    studies    from    different
91     laboratories  demonstrate  effects  and
92     there are limited or no human data.
          This document is a draft for review purposes only and does not constitute Agency policy.

                                               xxiv             DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
 1  Suggestive of a  causal relationship:  At
 2     least  one  high-quality  epidemiologic
 3     study shows an  association  but other
 4     studies are inconsistent.
 5  Inadequate to infer a causal relationship:
 6     The studies do not permit a conclusion
 7     regarding the presence or absence of an
 8     association.
 9  Not likely  to  be  a causal relationship:
10     Several adequate studies, covering the
11     full  range   of human   exposure  and
12     considering susceptible populations, are
13     mutually consistent in not showing  an
14     effect at any level of exposure.
15     The EPA is investigating and may  on a
16  trial  basis  use these  or other  standard
17  descriptors  to characterize  the   overall
18  weight of the evidence for effects other than
19  cancer.

    6.  Selecting studies for derivation
       of toxicity values

20     For each effect where there is  credible
21  evidence of an association with the agent,
22  the  assessment derives toxicity values  if
23  there   are  suitable  epidemiologic   or
24  experimental data. The decision to derive
25  toxicity values  may be linked to the hazard
26  descriptor.
27     Dose-response     analysis     requires
28  quantitative measures of dose and response.
29  Then, other factors being equal:
30
31
32
33
34

35
36
37
38

39
40
41
Epidemiologic studies are preferred
over animal studies, if quantitative
measures of exposure  are available
and effects can be attributed to the
agent
Among experimental animal  models,
those that respond most like humans
are preferred, if the comparability of
response can be determined.
Studies   by  a  route  of  human
environmental     exposure      are
preferred,  although   a  validated
                                      42
                                      43

                                      44
                                      45
                                      46
                                      47
                                      48

                                      49
                                      50
                                      51
                                      52

                                      53
                                      54
                                      55
                                      56
                                      57
          toxicokinetic model  can be used to
          extrapolate across exposure routes.
          Studies of longer exposure duration
          and  follow-up  are  preferred,  to
          minimize uncertainty about whether
          effects are representative of lifetime
          exposure.
          Studies with multiple exposure levels
          are  preferred  for  their ability to
          provide information  about the shape
          of the exposure-response curve.
          Studies  with  adequate  power  to
          detect  effects  at lower  exposure
          levels are preferred,  to minimize the
          extent  of extrapolation to  levels
          found in the environment.
58     Studies with non-monotonic exposure-
59  response relationships are  not necessarily
60  excluded from  the  analysis. A diminished
61  effect at higher exposure  levels  may  be
62  satisfactorily explained by  factors  such as
63  competing toxicity, saturation of absorption
64  or metabolism, exposure misclassification,
65  or selection bias.
66     If a large number of studies are suitable
67  for dose-response analysis,  the assessment
68  considers  the study characteristics in  this
69  section to focus on the  most informative
70  data. The  assessment explains the reasons
71  for not analyzing other groups of studies. As
72  a check on the selection of studies for dose-
73  response  analysis,   the  EPA  asks  peer
74  reviewers  to identify studies that were not
75  adequately considered.

    7. Deriving toxicity values

76  7.1.  General framework for dose-
77       response analysis
78     The EPA uses a two-step approach that
79  distinguishes analysis of the observed dose-
80  response data from inferences about lower
81  doses [U.S. EPA. 2005a. 331.
82     Within the observed range, the preferred
83  approach is to use modeling to incorporate a
          This document is a draft for review purposes only and does not constitute Agency policy.

                                               xxv            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1  wide range of data into the analysis. The
 2  modeling yields  a point of departure  (an
 3  exposure level near  the lower end  of  the
 4  observed   range,   without    significant
 5  extrapolation to lower doses) (Sections 7.2-
 6  7.3).
 7     Extrapolation to lower  doses  considers
 8  what is known about the modes of action for
 9  each effect (Sections  7.4-7.5). If response
10  estimates at lower doses are not required, an
11  alternative  is to  derive reference  values,
12  which are calculated by applying  factors to
13  the  point of departure in order to account
14  for  sources  of uncertainty and  variability
15  (Section 7.6).
16     For  a  group  of agents that  induce an
17  effect through a common mode of action, the
18  dose-response analysis may derive a relative
19  potency factor for each agent. A  full dose-
20  response analysis is conducted for one well-
21  studied index chemical in the group, then the
22  potencies of other members are expressed in
23  relative terms based on relative toxic effects,
24  relative  absorption  or metabolic   rates,
25  quantitative structure-activity relationships,
26  or receptor binding characteristics (U.S. EPA.
27  2005_a, §3.2.6; 2000, §4.4).
28     Increasingly, the EPA is basing toxicity
29  values  on combined  analyses of multiple
30  data sets or  multiple responses.  The EPA
31  also   considers   multiple   dose-response
32  approaches if they can be supported by
33  robust data.

34  7.2. Modeling dose to sites of biologic
35       effects

36     The  preferred approach for analysis of
37  dose is toxicokinetic modeling because of its
38  ability to incorporate a wide range of data.
39  The preferred dose metric would refer to the
40  active agent at the site of its biologic effect or
41  to a close, reliable surrogate measure. The
42  active  agent  may  be  the  administered
43  chemical or a metabolite. Confidence in the
44  use  of a toxicokinetic model depends  on the
45  robustness  of its validation process and on
46  the  results of sensitivity analyses  (U.S. EPA,
47  2006a: 2005a. §3.1: 1994b. §4.31
48     Because   toxicokinetic   modeling  can
49  require  many parameters  and more  data
50  than are typically available, the EPA has
51  developed standard approaches that can be
52  applied to typical data sets. These standard
53  approaches also facilitate comparison across
54  exposure patterns and species.

55     -  Intermittent  study   exposures are
56        standardized to a daily average over
57        the duration of exposure. For chronic
58        effects, daily exposures are averaged
59        over the lifespan. Exposures during a
60        critical  period,  however,  are not
61        averaged over a longer duration (U.S.
62        EPA. 2005a. 33.1.1: 1991. §3.2).

63     -  Doses are standardized to equivalent
64        human    terms     to     facilitate
65        comparison of results from different
66        species.

67     -  Oral doses are scaled allometrically
68        using mg/kg3/4-day as the equivalent
69        dose    metric   across    species.
70        Allometric   scaling   pertains   to
71        equivalence   across   species,  not
72        across  lifestages, and is not used  to
73        scale doses  from adult humans  or
74        mature animals to infants or children
75        fU.S. EPA. 2011: 2005a. §3.1.31

76     -  Inhalation  exposures   are   scaled
77        using  dosimetry models that  apply
78        species-specific  physiologic   and
79        anatomic  factors    and   consider
80        whether the  effect occurs at the site
81        of first  contact or  after  systemic
82        circulation   (U.S.    EPA.   2012a:
83        1994b. §3).

84     It can be informative to convert  doses
85  across exposure routes.  If this is  done, the
86  assessment describes the  underlying data,
87  algorithms,  and assumptions  (U.S.   EPA.
88  2005a. 33.1.4).
89     In the absence of study-specific data on,
90  for example, intake rates or body weight, the
91  EPA has developed recommended values for
92  use in dose-response analysis  (U.S.  EPA.
93  1988).
          This document is a draft for review purposes only and does not constitute Agency policy.

                                               xxvi            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1  7.3. Modeling response in the range
 2      of observation

 3     Toxicodynamic  ("biologically  based")
 4  modeling can incorporate data on biologic
 5  processes leading to an effect. Such models
 6  require sufficient data to ascertain a mode of
 7  action and to quantitatively support model
 8  parameters associated with its key events.
 9  Because  different  models   may  provide
10  equivalent fits  to  the  observed  data but
11  diverge substantially at lower doses, critical
12  biologic parameters  should  be measured
13  from laboratory studies, not by model fitting.
14  Confidence  in  the  use  of a  toxicodynamic
15  model  depends  on the  robustness  of its
16  validation  process and  on the results of
17  sensitivity analyses.  Peer  review of the
18  scientific basis and performance of a model
19  is essential [U.S. EPA. 2005a. §3.2.2).
20     Because  toxicodynamic  modeling  can
21  require   many  parameters    and  more
22  knowledge  and data  than  are typically
23  available, the EPA has developed a standard
24  set  of empirical  ("curve-fitting")  models
25  (http://www.epa.gov/ncea/bmds/) that can
26  be applied  to  typical data sets, including
27  those that are  nonlinear.  The EPA has  also
28  developed  guidance on  modeling  dose-
29  response data,  assessing model fit, selecting
30  suitable models,  and reporting modeling
31  results  (U.S.   EPA.  2012bl   Additional
32  judgment or alternative analyses are used if
33  the procedure  fails to yield reliable results,
34  for example, if the fit is poor, modeling may
35  be restricted to the lower  doses, especially if
36  there is competing toxicity at higher  doses
37  (U.S. EPA. 2005a. §3.2.3).
38     Modeling is used to  derive a point of
39  departure fU.S. EPA. 2012b:  2005a. §3.2.4).
40  (See Section 7.6 for alternatives if a point of
41  departure cannot be derived by modeling.):
42
43
44
45
46
47
48
49
50
51

52
53
54
55

56
57
58
59
60
61
62
63
64
65
66
67
68
If  linear  extrapolation  is   used,
selection   of   a  response   level
corresponding  to  the   point   of
departure is not highly influential, so
standard values near the low end of
the observable range are generally
used (for example, 10% extra risk for
cancer   bioassay  data,  1%   for
epidemiologic  data, lower for rare
cancers).
For  nonlinear  approaches,  both
statistical       and       biologic
considerations   are   taken   into
account
For dichotomous data,  a response
level of 10% extra risk is generally
used for minimally adverse effects,
5% or lower for more severe effects.
For continuous data, a response level
is  ideally based on  an established
definition of biologic significance. In
the absence of such definition, one
control standard deviation from  the
control   mean  is  often  used  for
minimally adverse effects, one-half
standard deviation for more  severe
effects.
69     The point of departure is the 95% lower
70  bound  on  the  dose  associated  with the
71  selected response level.

72  7.4.  Extrapolating to lower doses and
73       response levels

74     The purpose of extrapolating to lower
75  doses is to estimate responses at exposures
76  below  the  observed   data.   Low-dose
77  extrapolation, typically used for cancer data,
78  considers what is  known about  modes of
79  action fU.S. EPA. 2005a. §3.3.1 and §3.3.21
80  1)  If a biologically based model  has been
81     developed and  validated for  the agent,
82     extrapolation may use the fitted model
83     below the observed  range if significant
84     model uncertainty can be ruled out with
85     reasonable confidence.
          This document is a draft for review purposes only and does not constitute Agency policy.

                                              xxvii            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1   2)  Linear extrapolation is used if the dose-
 2      response curve is expected to have a
 3      linear component below  the point of
 4      departure. This includes:
 5     -   Agents or their metabolites that are
 6         DNA-reactive    and   have   direct
 7         mutagenic activity.
 8     -   Agents or their metabolites for which
 9         human exposures  or body burdens
10         are near doses associated with key
11         events leading to an effect
12       Linear extrapolation  is also used when
13       data are insufficient to establish  mode
14       of  action and   when  scientifically
15       plausible.
16       The  result of linear extrapolation is
17       described by an oral  slope factor or an
18       inhalation unit risk, which is the slope
19       of the  dose-response curve at lower
20       doses or concentrations, respectively.
21  3)  Nonlinear   models   are   used   for
22     extrapolation if there are sufficient data
23     to ascertain the mode of action and to
24     conclude  that it is not linear at lower
25     doses,  and  the   agent  does    not
26     demonstrate mutagenic or other activity
27     consistent with linearity at lower doses.
28     Nonlinear approaches  generally should
29     not be used in cases where mode  of
30     action has not ascertained.  If nonlinear
31     extrapolation  is   appropriate  but  no
32     model is developed, an alternative is to
33     calculate reference values.
34  4)  Both  linear and  nonlinear approaches
35     may be used if there a multiple modes of
36     action. For example, modeling to a low
37     response   level  can  be   useful  for
38     estimating the response at doses where a
39     high-dose mode of action would be less
40     important
41     If linear  extrapolation  is  used,  the
42  assessment  develops  a  candidate  slope
43  factor or unit risk for each suitable data set
44  These results are  arrayed, using common
45  dose metrics, to show  the  distribution  of
46  relative  potency across various effects and
47  experimental systems. The assessment then
48  derives or selects an overall slope factor and
49  an overall unit risk for the agent, considering
50  the  various  dose-response  analyses,  the
51  study preferences discussed in Section 6,
52  and the  possibility of basing a more robust
53  result on multiple data sets.

54  7.5.  Considering susceptible
55       populations and lifestages
56     The  assessment  analyzes the available
57  information  on  populations and  lifestages
58  that may be  particularly susceptible to each
59  effect. A tiered  approach is used [U.S. EPA.
60  2005a. 33.51.
61  1) If an  epidemiologic or experimental
62     study reports quantitative results for a
63     susceptible population or lifestage, these
64     data  are  analyzed to derive  separate
65     toxicity    values    for    susceptible
66     individuals.
67  2) If data on risk-related parameters allow
68     comparison  of the general population
69     and  susceptible individuals, these data
70     are used to adjust the general-population
71     toxicity   values   for   application  to
72     susceptible individuals.
73  3) In the absence of chemical-specific data,
74     the EPA  has developed age-dependent
75     adjustment factors for early-life exposure
76     to potential carcinogens that  have a
77     mutagenic  mode of action.  There  is
78     evidence  of early-life susceptibility to
79     various carcinogenic  agents, but most
80     epidemiologic   studies   and    cancer
81     bioassays  do  not   include  early-life
82     exposure. To address the potential  for
83     early-life    susceptibility,   the   EPA
84     recommends [U.S. EPA. 2005b. §5):
85     -  10-fold  adjustment  for  exposures
86        before age 2 years.
87     -  3-fold  adjustment   for  exposures
88        between ages 2 and 16 years.
          This document is a draft for review purposes only and does not constitute Agency policy.

                                              xxviii            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1  7.6.  Reference values and uncertainty
 2       factors

 3     An oral reference dose or an inhalation
 4  reference concentration is an estimate of an
 5  exposure    (including    in    susceptible
 6  subgroups) that is likely to be without an
 7  appreciable  risk of adverse  health  effects
 8  over   a  lifetime  (U.S.  EPA.  2002.54.21
 9  Reference values are typically calculated for
10  effects other than cancer and for suspected
11  carcinogens if a well characterized mode of
12  action indicates that a  necessary key event
13  does   not occur below   a  specific  dose.
14  Reference values provide  no  information
15  about risks at higher exposure levels.
16     The  assessment characterizes  effects
17  that form the basis for reference values as
18  adverse,  considered to be  adverse, or a
19  precursor  to   an   adverse   effect   For
20  developmental    toxicity,     reproductive
21  toxicity, and neurotoxicity there is  guidance
22  on adverse effects and their biologic markers
23  fU.S. EPA. 1998.1996.19911
24     To   account  for   uncertainty   and
25  variability in the derivation of a lifetime
26  human exposure where adverse effects  are
27  not anticipated to occur, reference values are
28  calculated by applying a series of uncertainty
29  factors to the point of departure. If a point of
30  departure cannot be derived by modeling, a
31  no-observed-adverse-effect   level    or   a
32  lowest-observed-adverse-effect level is used
33  instead. The assessment discusses  scientific
34  considerations  involving several  areas  of
35  variability or uncertainty.
36  Human variation. The  assessment accounts
37     for variation in susceptibility across  the
38     human  population  and  the possibility
39     that  the  available data  may not  be
40     representative  of  individuals  who  are
41     most susceptible to  the effect A factor of
42     10  is generally used to account for this
43     variation. This factor  is reduced  only if
44     the  point  of departure is derived  or
45     adjusted  specifically  for  susceptible
46     individuals (not for a general population
47     that includes both susceptible and non-
48     susceptible   individuals)   (U.S.   EPA,
49     2002.54.4.5:    1998.54.2:    1996.54:
50     1994b. 54.3.9.1: 1991.53.41

51  Animal-to-human extrapolation. If animal
52     results  are used to  make inferences
53     about humans, the  assessment adjusts
54     for cross-species  differences. These may
55     arise from differences in toxicokinetics
56     or  toxicodynamics.  Accordingly,  if  the
57     point of departure  is standardized to
58     equivalent human terms or is based on
59     toxicokinetic or dosimetry  modeling,  a
60     factor of 101/2 (rounded to 3) is applied
61     to account for the remaining uncertainty
62     involving      toxicokinetic        and
63     toxicodynamic    differences.    If    a
64     biologically based model adjusts fully for
65     toxicokinetic     and    toxicodynamic
66     differences across species, this factor is
67     not used. In most other cases, a factor of
68     10   is   applied     (U.S.   EPA.  2011:
69     2002.54.4.5:    1998.54.2:    1996.54:
70     1994b, §4.3.9.1; 1991, §3.4).

71  Adverse-effect  level  to   no-observed-
72     adverse-effect  level.  If  a  point  of
73     departure   is  based  on  a  lowest-
74     observed-adverse-effect    level,     the
75     assessment must infer  a  dose where
76     such effects are not expected. This can be
77     a matter of great uncertainty, especially
78     if there  is no evidence available at lower
79     doses.   A factor  of  10  is  applied  to
80     account for the uncertainty in making
81     this inference. A factor  other  than 10
82     may  be  used,   depending   on   the
83     magnitude and nature of the response
84     and the  shape  of  the  dose-response
85     curve (U.S. EPA. 2002. §4.4.5; 1998. §4.2;
86     1996. 54: 1994b. 54.3.9.1: 1991. 53.4).

87  Subchronic-to-chronic exposure. If a point
88     of  departure  is  based  on subchronic
89     studies,   the    assessment   considers
90     whether lifetime exposure  could  have
91     effects  at lower  levels of  exposure. A
92     factor of 10 is applied to account for the
93     uncertainty in using subchronic studies
94     to  make  inferences  about  lifetime
          This document is a draft for review purposes only and does not constitute Agency policy.

                                               xxix            DRAFT—DO NOT CITE OR QUOTE

-------
                                            Toxicological Review of Trimethylbenzene
 1     exposure.  This   factor  may  also  be
 2     applied    for     developmental    or
 3     reproductive effects if exposure covered
 4     less than the full critical period. A factor
 5     other than 10 may be used, depending
 6     on the duration of the  studies and the
 7     nature of the response  (U.S. EPA. 2002.
 8     §4.4.5; 1998. §4.2; 1994b. §4.3.9.1).
 9  Incomplete  database.  If an  incomplete
10     database  raises  concern that  further
11     studies might identify  a more sensitive
12     effect,  organ system,  or lifestage, the
13     assessment  may  apply  a  database
14     uncertainty    factor     (U.S.     EPA.
15     2002. §4.4.5:   1998. §4.2:    1996. §4:
16     1994b. §4.3.9.1; 1991. §3.4).  The size of
17     the factor depends on the nature of the
18     database deficiency. For example, the
19     EPA typically follows the suggestion that
20     a factor  of  10  be  applied  if  both  a
21     prenatal  toxicity  study  and  a  two-
22     generation   reproduction   study  are
23     missing and a factor of 101/2 if either is
24     missing (U.S. EPA. 2002. §4.4.5).
25     In this way, the  assessment  derives
26  candidate values for  each  suitable data set
27  and effect that is credibly associated with the
28  agent  These  results  are  arrayed, using
29  common dose metrics, to show where effects
30  occur across  a  range of exposures (U.S. EPA.
31  1994b, §4.3.9).
32     The  assessment  derives  or selects an
33  organ- or system-specific reference value for
34  each  organ or system affected by the agent
35  The assessment explains the rationale for
36  each  organ/system-specific reference value
37  (based  on, for example, the highest quality
38  studies,  the  most sensitive outcome, or  a
39  clustering of values). By providing these
40  organ/system-specific reference values, IRIS
41  assessments      facilitate      subsequent
42  cumulative risk  assessments that consider
43  the combined effect of multiple agents acting
44  at  a common site  or through  common
45  mechanisms (NRC. 2009).
46     The assessment then selects an overall
47  reference  dose  and  an  overall  reference
                                       48  concentration for the  agent to represent
                                       49  lifetime  human  exposure  levels  where
                                       50  effects are not anticipated to occur. This is
                                       51  generally the  most sensitive organ/system-
                                       52  specific     reference     value,     though
                                       53  consideration   of  study  quality   and
                                       54  confidence  in each value  may  lead  to a
                                       55  different selection.

                                       56  7.7. Confidence and uncertainty in the
                                       57      reference values
                                       58     The  assessment  selects  a  standard
                                       59  descriptor   to  characterize the  level  of
                                       60  confidence in  each reference value, based on
                                       61  the likelihood that the value would change
                                       62  with further testing. Confidence in reference
                                       63  values is based on quality of the studies used
                                       64  and completeness of the database, with more
                                       65  weight given to  the  latter. The  level  of
                                       66  confidence is  increased for reference values
                                       67  based  on human data  supported by animal
                                       68  data fU.S. EPA. 1994b. §4.3.9.21
                                       69  High confidence: The reference value is not
                                       70     likely to  change with  further  testing,
                                       71     except for mechanistic studies that might
                                       72     affect the interpretation of  prior  test
                                       73     results.
                                       74  Medium confidence:  This is  a matter  of
                                       75     judgment,   between   high   and   low
                                       76     confidence.
                                       77  Low  confidence: The reference value  is
                                       78     especially vulnerable  to  change  with
                                       79     further testing.
                                       80     These   criteria  are   consistent  with
                                       81  guidelines   for   systematic  reviews   that
                                       82  evaluate the quality of evidence. These  also
                                       83  focus on whether further research would be
                                       84  likely to change confidence in the estimate of
                                       85  effect (Guyattetal.. 2008b).
                                       86     All assessments discuss the  significant
                                       87  uncertainties  encountered  in the analysis.
                                       88  The   EPA   provides    guidance    on
                                       89  characterization  of  uncertainty  [U.S.  EPA.
                                       90  2005a. §3.6).  For example, the  discussion
                                       91  distinguishes  model uncertainty (lack  of
                                       92  knowledge  about  the  most  appropriate
This document is a draft for review purposes only and does not constitute Agency policy.

                                     xxx            DRAFT—DO NOT CITE OR QUOTE

-------
                                                          Toxicological Review of Trimethylbenzene
 1  experimental  or   analytic   model)   and
 2  parameter uncertainty  (lack  of knowledge
 3  about   the   parameters  of  a   model).
 4  Assessments  also discuss  human variation
 5  (interpersonal   differences    in   biologic
 6  susceptibility or in exposures  that  modify
 7  the effects of the agent).
 8       Note: The  Preamble reflects methods
 9       that will be employed once all the 2011
10       MAS recommendations have been fully
11       implemented.  As  this   Toxicological
12       Review was created during a period in
13       which the NAS recommendations were
14       being  incorporated  into  the  IRIS
15       process, the  methods utilized  in  the
16       assessment may not completely reflect
17       those  detailed in  the  Preamble.  For
18       further  information on  which specific
19       NAS  recommendations   have   been
20       implemented  in  this  document, please
21       refer to Appendix D (Documentation of
22       Implementation  of the  2011 National
23       Research Council Recommendations) in
24       the Supplemental Information.

    Preamble References
       CDC (Centers for Disease Control and
          Prevention). (2004). The health
          consequences of smoking: A report of the
          Surgeon General. Washington, DC: U.S.
          Department of Health and Human Services.
          http://www.surgeongeneral.gov/librarv/smok
          ingconsequences/

       Guvatt GH: Oxman. AD: Kunz. R: Vist GE:
          Falck-Ytter. Y: Schiinemann. HJ. (2008a).
          GRADE: What is "quality of evidence" and
          why is it important to clinicians? [Review].
          BMJ 336: 995-998.
          http://dx.doi.org/10.1136/bmi.39490.551019.
          BE
Guvatt GH: Oxman. AD: Vist GE: Kunz. R:
   Falck-Ytter. Y: Alonso-Coello. P:
   Schiinemann. HJ. (2008b). GRADE: An
   emerging consensus on rating quality of
   evidence and  strength of recommendations.
   BMJ 336: 924-926.
   http://dx.doi.org/10.1136/bmi.39489.470347.
   AD

HEW (U.S. Department of Health, Education
   and Welfare). (1964). Smoking and health:
   Report of the  advisory committee to the
   surgeon general of the public health service.
   Washington, DC: U.S. Department of Health,
   Education, and Welfare.
   http ://profiles .nlm. nih. gov/ps/retrieve/Resour
   ceMetadata/NNBBMQ

Hill. AB. (1965). The environment and disease:
   Association or causation? Proc R Soc Med
   58: 295-300.

IARC (International Agency for Research on
   Cancer). (2006). Preamble to the IARC
   monographs. Lyon, France.
   http://monographs.iarc.fr/ENG/Preamble/

IOM (Institute of Medicine). (2008). Improving
   the presumptive disability decision-making
   process for veterans. In JM Samet; CC
   Bodurow (Eds.). Washington, DC: National
   Academies Press.
   http://www.nap.edu/openbook.php7record id
   =11908

NRC (National Research Council). (1983). Risk
   assessment in the federal government:
   Managing the process. Washington, DC:
   National Academies Press.
   http://www.nap.edu/openbook.php7record id
   =366&page=Rl

NRC (National Research Council). (2009).
   Science and decisions: Advancing risk
   assessment. Washington, DC: National
   Academies Press.
   http://www.nap.edu/catalog/12209.html

Rothman. KJ: Greenland. S. (1998). Modern
   epidemiology (2nd ed.). Philadelphia, PA:
   Lippincott, Williams, & Wilkins.
U.S. EPA (U.S. Environmental Protection
   Agency). (1986a). Guidelines for
   mutagenicity risk assessment [EPA Report].
   (EPA/630/R-98/003). Washington, DC.
   http://www.epa.gov/iris/backgrd.html
           This document is a draft for review purposes only and does not constitute Agency policy.

                                                  xxxi             DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
U.S. EPA (U.S. Environmental Protection
   Agency). (1986b). Guidelines for the health
   risk assessment of chemical mixtures. Fed
   Reg 51: 34014-34025.

U.S. EPA (U.S. Environmental Protection
   Agency). (1988). Recommendations for and
   documentation of biological values for use in
   risk assessment [EPA Report]. (EPA/600/6-
   87/008). Cincinnati, OH.
   http ://cfpub. epa. gov/ncea/cfm/recordisplav. cf
   m?deid=34855

U.S. EPA (U.S. Environmental Protection
   Agency). (1991). Guidelines for
   developmental toxicity risk assessment [EPA
   Report]. (EPA/600/FR-91/001). Washington,
   DC: U.S. Environmental Protection Agency,
   Risk Assessment Forum.
   http://www.epa.gov/raf/publications/guidelin
   es-dev-toxicitv-risk-assessment, htm

U.S. EPA (U.S. Environmental Protection
   Agency). (1994b). Methods for derivation of
   inhalation reference concentrations and
   application of inhalation dosimetry [EPA
   Report]. (EPA/600/8-90/066F). Research
   Triangle Park, NC.
   http ://cfpub. epa. gov/ncea/cfm/recordisplav. cf
   m?deid=71993

U.S. EPA (U.S. Environmental Protection
   Agency). (1996). Guidelines for reproductive
   toxicity risk assessment [EPA Report].
   (EPA/630/R-96/009). Washington, DC.
   http://www.epa.gov/raf/publications/pdfs/RE
   PRO51.PDF

U.S. EPA (U.S. Environmental Protection
   Agency). (1998). Guidelines for
   neurotoxicity risk assessment  [EPA Report].
   (EPA/630/R-95/001F). Washington, DC.
   http://www.epa.gov/raf/publications/pdfs/NE
   UROTOX.PDF

U.S. EPA (U.S. Environmental Protection
   Agency). (2000). Supplementary guidance
   for conducting health risk assessment of
   chemical mixtures [EPA Report].
   (EPA/630/R-00/002). Washington, DC.
   http ://cfpub. epa. gov/ncea/cfm/recordisplav. cf
   m?deid=20533
U.S. EPA (U.S. Environmental Protection
   Agency). (2002). A review of the reference
   dose and reference concentration processes
   [EPA Report]. (EPA/630/P-02/002F).
   Washington, DC: Risk Assessment Forum,
   U.S. Environmental Protection Agency.
   http ://cfpub. epa. gov/ncea/cfm/recordisplav .cf
   m?deid=51717

U.S. EPA (U.S. Environmental Protection
   Agency). (2005a). Guidelines for carcinogen
   risk assessment [EPA Report]. (EPA/630/P-
   03/00IF). Washington, DC: Risk Assessment
   Forum.
   http://www.epa.gov/cancerguidelines/

U.S. EPA (U.S. Environmental Protection
   Agency). (2005b). Supplemental guidance
   for assessing susceptibility from early-life
   exposure to carcinogens [EPA Report] (pp.
   1125-1133). (EPA/630/R-03/003F).
   Washington, DC.
   http://www.epa.gov/cancerguidelines/guideli
   nes-carcinogen-supplement.htm

U.S. EPA (U.S. Environmental Protection
   Agency). (2006a). Approaches for the
   application of physiologically based
   pharmacokinetic (PBPK) models and
   supporting data in risk assessment (Final
   Report)  [EPAReport]. (EPA/600/R-
   05/043F). Washington, DC.
   http ://cfpub. epa. gov/ncea/cfm/recordisplav. cf
   m?deid= 157668

U.S. EPA (U.S. Environmental Protection
   Agency). (2006b). A framework for
   assessing health risk of environmental
   exposures to children [EPA Report].
   (EPA/600/R-05/093F). Washington, DC.
   http ://cfpub. epa. gov/ncea/cfm/recordisplav .cf
   m?deid=158363

U.S. EPA (U.S. Environmental Protection
   Agency). (2009). EPAs Integrated Risk
   Information System: Assessment
   development process [EPA Report].
   Washington, DC.
   http ://epa. gov/iris/process. htm

U.S. EPA (U.S. Environmental Protection
   Agency). (2010). Integrated science
   assessment for carbon monoxide [EPA
   Report]. (EPA/600/R-09/019F).  Research
   Triangle Park, NC.
   http ://cfpub. epa. gov/ncea/cfm/recordisplav .cf
   m?deid=218686
    This document is a draft for review purposes only and does not constitute Agency policy.

                                             xxxii              DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
U.S. EPA (U.S. Environmental Protection
   Agency). (2011). Recommended use of body
   weight 3/4 as the default method in
   derivation of the oral reference dose [EPA
   Report]. (EPA/100/R11/0001). Washington,
   DC.
   http://www.epa.gov/raf/publications/interspe
   cies-extrapolation.htm

U.S. EPA (U.S. Environmental Protection
   Agency). (2012a). Advances in inhalation
   gas dosimetry for derivation of a reference
   concentration (rfc) and use in risk assessment
   [EPA Report]. (EPA/600/R-12/044).
   Washington, DC.
   http ://cfpub. epa. gov/ncea/cfm/recordisplav. cf
   m?deid=244650
U.S. EPA (U.S. Environmental Protection
   Agency). (2012b). Benchmark dose technical
   guidance. (EPA/100/R-12/001). Washington,
   DC: Risk Assessment Forum.
   http ://www .epa. gov/raf/publications/pdfs/ben
   chmark  dose  guidance.pdf
    This document is a draft for review purposes only and does not constitute Agency policy.

                                            xxxiii             DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
     EXECUTIVE SUMMARY
                                  Occurrence and Health Effects

 1                 Trimethylbenzenes are a commercially available mixture of three individual
 2          isomers: 1,2,3-, 1,2,4-, and 1,3,5-trimethylbenzene (TMBs). 1MB isomers are
 3          produced during petroleum refining and production of aromatic hydrocarbons with
 4          nine carbons (i.e., C9 aromatic fraction). As the vast majority of the C9 fraction is
 5          used as a component of gasoline, vehicle emissions are expected to be the major
 6          anthropogenic source of TMBs. TMBs are volatile hydrocarbons, and thus humans
 7          are exposed to these isomers primarily through breathing air containing TMB
 8          vapors, although ingestion through food or drinking water is also possible.
 9                 Effects on the nervous system, respiratory system, and hematological
10          system (i.e., blood) have been reported in occupationally- and residentially-exposed
11          humans, but these effects were observed following exposure to complex mixtures
12          containing TMB isomers, thus making it difficult to determine the contribution of
13          each TMB isomer to the observed health effects. Health effects that are roughly
14          analogous to those seen in humans have been observed in animals exposed to the
15          individual isomers. Effects on the nervous system, including cognitive effects and
16          decreased pain sensitivity, are the most widely observed effects in animals. Effects
17          on other organ systems, including the respiratory and hematological systems, have
18          also been observed in animals. Both 1,2,4-TMB and 1,3,5-TMB have been observed
19          to elicit effects on pregnant animals and developing fetuses, but at exposure levels
20          greater than those that cause effects on the nervous system. There is inadequate
21          information to evaluate the carcinogenicity of TMBs.

     1. Effects Other Than Cancer Following Inhalation Exposure
22          The relationship between exposure to 1,2,3-TMB, 1,2,4-TMB, 1,3,5-TMB and health effects
23   has been evaluated in studies of (1) exposed human adults, (2) animals exposed via inhalation for
24   acute, short-term, and subchronic durations, and (3) animals exposed gestationally via inhalation.
25          Human studies included occupational exposure to various solvent mixtures containing
26   TMBs. Health effects noted in these studies were eye irritation, neurological  (hand tremble,

            This document is a draft for review purposes only and does not constitute Agency policy.

                                              xxxiv           DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene
 1   abnormal fatigue, lack of coordination), and hematological effects [Chenetal.. 1999: Norsethetal..
 2   1991: Battigetal.. 1958: Battigetal.. 1956]. Also, residential exposure to mixtures containing
 3   1,2,4-TMB were observed to result in asthma [Billionnetetal.. 2011]. However, as these studies
 4   involved exposures to mixtures containing multiple TMB isomers and other volatile organic
 5   compounds (VOCs], it is difficult to ascertain the specific contribution of each TMB isomer to the
 6   specific health effects reported. Controlled human exposures to individual isomers also exist,
 7   although these studies generally report little or no effect on respiratory or sensory irritation [Tones
 8   etal.. 2006: Tarnbergetal.. 1997a: Tarnbergetal.. 1997b: Kostrzewskietal.. 1997: Tarnbergetal..
 9   1996: Kostrewski and Wiaderna-Brycht. 1995]. One controlled human exposure study reported
10   some deficits in attention following exposure to white spirit (WS], a complex mixture containing
11   1,2,4-TMB fLammers etal.. 20071.
12          Animal inhalation studies [Wiadernaetal., 2002: Gralewicz and Wiaderna, 2001: Wiaderna
13   etal.. 1998: Gralewicz etal.. 1997b: Gralewicz etal.. 1997a: Korsak etal.. 19951 included acute and
14   short-term studies of TMBs that reported respiratory irritation (decreased respiration rates] and
15   neurological effects (decreased pain sensitivity, altered cognitive function, and decreased anxiety
16   and/or increased motor function] that are consistent with effects seen in human studies. Four
17   subchronic inhalation studies for 1,2,3-TMB and 1,2,4-TMB observed exposure-response effects in
18   multiple organ systems, including the nervous, hematological, and respiratory systems (Korsak et
19   al., 2000a, bj Korsak etal., 1997: Korsak and Rydzynski, 1996]. In these studies, disturbances in
20   central nervous system (CNS] function, including decreased pain sensitivity and decreased
21   neuromuscular function and coordination, appear to be the most sensitive endpoints following
22   exposure to 1,2,3-TMB or 1,2,4-TMB. No subchronic studies were found that investigated exposure
23   to 1,3,5-TMB. One developmental toxicity study (Saillenfaitetal.. 2005] observed similar levels of
24   maternal and fetal toxicity (i.e., decreased maternal weight gain and fetal weight] following
25   exposure to either 1,2,4-TMB or 1,3,5-TMB; other indices of fetal toxicity (i.e., fetal death and
26   malformations] were not affected by exposure.
27          Table ES-1 summarizes the RfCs derived for all three TMB isomers, and the sections that
28   follow provide details on the RfC derivation for each isomer.

            Table ES-1. Summary of inhalation reference concentrations (RfCs)
Isomer
1,2,4-TMB
1,2,3-TMB
1,3,5-TMB
Source
Decreased pain sensitivity
Decreased pain sensitivity
Adopted from 1,2,4-TMB based on
sufficient similarity of these isomers
Reference value
(mg/m3)
5 x 10"2
5 x 10"2
5 x 10"2
Confidence
Low-to-medium
Low-to-medium
Low
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                xxxv            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
     2. Inhalation Reference Concentration (RfC) for 1,2,4-TMB for Effects Other
       Than Cancer
            Table ES-2. Summary of reference concentration (RfC) derivation for
                       1,2,4-TMB
Critical effect
Decreased pain sensitivity
90 day male rat study
Korsak and Rydzyriski (1996)
Point of departure
PODHEc(mg/m3) = 15.8
Uncertainty
factor
300
Chronic RfC
(mg/m3)
5 x 10"2
 1          Decreased pain sensitivity was observed in multiple studies of acute, short-term, and
 2   subchronic durations [Gralewicz and Wiaderna. 2001: Gralewicz etal.. 1997b: Korsak and
 3   Rydzynski. 1996: Korsak etal.. 1995). Given the consistency of this effect and the determination
 4   that decreased pain sensitivity is an appropriate adverse effect with which to derive reference
 5   values (see Section 2.1.5 of this Toxicological Review), in accordance with the U.S. EPA's Guidelines
 6   for Neurotoxicity Risk Assessment [1998], decreased pain sensitivity was selected as the critical
 7   effect and Korsak and Rydzynski [1996] was selected as the principal study for derivation of the
 8   RfC for 1,2,4-TMB.
 9          The RfC calculation is summarized in Table ES-2. The available rat PBPK model fHissink et
10   al.. 2007] was used to convert the external concentrations (in mg/m3] from the animal study to the
11   internal blood metric of weekly average venous 1,2,4-TMB concentration (in mg/L]. These internal
12   blood metrics were then used as the dose inputs for benchmark dose  (BMD] modeling.
13   A benchmark response [BMR] equal to a 1 standard deviation change in the control mean for
14   decreased pain sensitivity was used. A BMDLiso of 0.086 mg/L was estimated for decreased pain
15   sensitivity in male rats exposed to 1,2,4-TMB via inhalation for 90 days (6 hours/day, 5 days/week]
16   [data used in model: [Korsak and Rydzynski. 1996]].
17          The available human PBPK model [Hissink et al.. 2007] was then used to estimate a human
18   equivalent concentration [HEC] of 15.8 mg/m3 from the BMDLiso of 0.086 mg/L. This HEC was used
19   as the PODnEc with which to derive the RfC. A composite uncertainty factor (UF] of 300 was applied:
20   3 to account for uncertainty in extrapolating from laboratory animals to humans  (interspecies
21   variability], 10 to account for variation in susceptibility among members of the human population
22   (interindividual variability],  3 to  account for subchronic-to-chronic extrapolation due to the  use of a
23   subchronic study with effects observed to recover within weeks of exposure termination, and 3 to
24   account for deficiencies in the database (no two-generation reproductive/developmental toxicity or
            This document is a draft for review purposes only and does not constitute Agency policy.

                                               xxxvi            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   developmental neurotoxicity studies were available). Dividing the PODHEc by the composite UF of
 2   300 yielded a chronic RfC of 5 x 1Q-2 mg/m3 for 1,2,4-TMB.

     3. Confidence in the Chronic Inhalation RfC for 1,2,4-TMB
 3          A confidence level of high, medium, or low is assigned to the study used to derive the RfC,
 4   the overall database, and the RfC itself, as described in Section 4.3.9.2 of EPA's Methods for
 5   Derivation of Inhalation Reference Concentrations and Application of Inhalation Dosimetry [U.S. EPA.
 6   1994b].
 7          Confidence in the study from which the critical effect was identified, Korsak and Rydzyhski
 8   [1996], is low to medium. This peer-reviewed study was well designed, using three dose groups
 9   plus untreated controls and a typical number of animals per dose group for evaluating
10   neurotoxicity following subchronic exposure.
11          One area of uncertainty regarding this study is the lack of reported actual concentrations.
12   However, as the methods by which the test atmosphere was generated and analyzed were reported
13   in sufficient detail, and given the fact that this laboratory has used this methodology in subsequent
14   studies [Korsak et al.. 2000a. b] and achieved appropriate actual concentrations (i.e., within 10% of
15   target concentrations), the concern regarding the lack of reported actual concentrations is minimal.
16   Another source of uncertainty is the fact that Korsak and Rydzyhski [1996] does not explicitly state
17   that the reported measures  of variance in Table 1 of that reference are standard deviations.
18   However, careful analysis of the reported levels of variance and magnitude  of statistical significance
19   reported indicate that the measures of variance are standard deviations. Supporting this
20   conclusions is the observation that all other papers by Korsak et al. [2000a. b; 1997: 1995] report
21   variance as standard deviations. The critical effect on which the RfC is based is well-supported as
22   the evidence for  1,2,4-TMB-induced neurotoxicity is coherent across multiple animals species (i.e.,
23   human, mouse, and rat] and consistent across multiple exposure durations (i.e., acute, short-term,
24   and subchronic]  [Gralewicz and Wiaderna, 2001: Chenetal., 1999: Wiaderna et al., 1998: Gralewicz
25   etal.. 1997b: Gralewicz etal.. 1997a: Korsak and Rydzyhski. 1996: Norseth  etal.. 19911
26          The database for 1,2,4-TMB includes acute, short-term, subchronic, and developmental
27   toxicity studies in rats and mice. However, confidence in the database is low to medium because it
28   lacks chronic, multi-gene ration reproductive/developmental, and developmental neurotoxicity
29   studies, and the studies supporting the critical effect predominantly come from the same research
30   institute. Consequently, the  overall confidence in the RfC for 1,2,4-TMB is low to medium.
            This document is a draft for review purposes only and does not constitute Agency policy.

                                               xxxvii            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
     4. Inhalation Reference Concentration (RfC) for 1,2,3-TMB for Effects Other Than
        Cancer
            Table ES-3. Summary of reference concentration (RfC) derivation for
                       1,2,3-TMB
Critical effect
Decreased pain sensitivity
90 day male rat study
Korsak and Rydzyriski
(1996)
Point of departure
PODHEc(mg/m3) = 16.3
Uncertainty
factor
300
Chronic RfC
(mg/m3)
5 x 10"2
 1          Decreased pain sensitivity was observed in multiple studies of acute, short-term, and
 2   subchronic durations [Lutz etal.. 2010: Wiaderna et al.. 1998: Korsak and Rydzynski. 1996). Given
 3   the consistency of this effect and the determination that decreased pain sensitivity is an adverse
 4   effect, in accordance with the U.S. EPA's Guidelines for Neurotoxicity Risk Assessment [U.S. EPA.
 5   1998]. decreased pain sensitivity was selected as the critical effect and Korsak and Rydzynski
 6   [1996] was selected as the principal study for derivation of the RfC for 1,2,3-TMB.
 7          The RfC calculation is summarized in Table ES-3. BMD modeling was used in order to
 8   identify the POD for decreased pain sensitivity. A BMR equal to a 1 standard deviation change in the
 9   control mean was used. A BMDLiso of 17.36 mg/m3 was estimated for decreased pain sensitivity in
10   male rats exposed to 1,2,3-TMB via inhalation for 90 days (6 hours/day, 5 days/week] [Korsak and
11   Rydzynski. 19961
12          As no PBPK model was available for 1,2,3-TMB, default dosimetry methodologies were used
13   to estimate the HEC of 16.3 mg/m3, based on the ratio of the human and animal blood:air partition
14   coefficients [U.S. EPA, 1994b]. This PODHEc was used to derive the RfC. A composite uncertainty
15   factor (UF] of 300 was applied: 3 to account for uncertainty in extrapolating from laboratory
16   animals to humans (interspecies variability], 10 to account for variation in susceptibility among
17   members of the human population (interindividual variability], 3 to account for subchronic-to-
18   chronic extrapolation due to the use of a subchronic study, and 3 to account for deficiencies in the
19   database (no two-generation reproductive/developmental toxicity, developmental toxicity, or
20   developmental neurotoxicity studies were available]. Dividing the PODHEc by the composite UF of
21   300 yielded a chronic RfC of 5 x 1Q-2 mg/m3 for 1,2,3-TMB.

     5. Confidence in the Chronic Inhalation RfC for 1,2,3-TMB
22          Confidence in the study from which the critical effect was identified, Korsak and Rydzynski
23   [1996] is low to medium. This peer-reviewed study was well designed, using three dose groups
24   plus untreated controls and a typical number of animals per dose group for evaluating

            This document is a draft for review purposes only and does not constitute Agency policy.

                                              xxxviii           DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   neurotoxicity following subchronic exposure. One area of uncertainty regarding this study is the
 2   lack of reported actual concentrations. However, as the methods by which the test atmosphere was
 3   generated and analyzed were reported in sufficient detail, and given the fact that this laboratory
 4   has used this methodology in subsequent studies [Korsak et al., 2000a, b] and achieved appropriate
 5   actual concentrations (i.e., within 10% of target concentrations), the concern regarding the lack of
 6   reported actual concentrations is minimal. Another source of uncertainty is the fact that Korsak and
 7   Rydzyhski [1996] does not explicitly state that the reported measures of variance in Table 1 of that
 8   reference are standard deviations. However, careful analysis of the reported levels of variance and
 9   magnitude of statistical significance reported indicate that the measures of variance are standard
10   deviations. Supporting this conclusions is the observation that all other papers by Korsak et al.
11   [2000a, b; 1997: 1995] report variance as standard deviations.The critical effect on which the RfC is
12   based is well-supported as the evidence for 1,2,3-TMB-induced neurotoxicity is coherent across
13   multiple animals species (i.e., mouse, and rat] and consistent across multiple exposure durations
14   (i.e., acute, short-term, and subchronic] (Lutz etal.. 2010: Wiaderna et al.. 1998: Korsak and
15   Rydzynski. 19961
16          The database for 1,2,3-TMB includes acute, short-term, and subchronic toxicity studies in
17   rats and mice. However, confidence in the database is low to medium because it lacks chronic,
18   multi-generation reproductive/developmental, developmental toxicity, or developmental
19   neurotoxicity studies, and the studies supporting the critical effect predominantly come from the
20   same research institute. Consequently, the overall confidence in the RfC for 1,2,3-TMB is low to
21   medium.

     6. Inhalation Reference Concentration (RfC) for 1,3,5-TMB for Effects Other Than
        Cancer
22          No chronic or subchronic studies exist that would support the derivation of an RfC for
23   1,3,5-TMB, however one developmental toxicity study (Saillenfaitetal.. 2005] was identified as a
24   potential study from which to identify a critical effect for RfC derivation.
25          The use of decreased maternal weight gain observed in Saillenfait et al. (2005] as the critical
26   effect for RfC derivation would result in an RfC 20-fold greater than that derived for 1,2,4-TMB (1
27   mg/m3 vs. 5 x 1Q-2 mg/m3]. This large difference is not consistent with the rest of the toxicological
28   database for 1,2,4-TMB and 1,3,5-TMB, which demonstrates that the two isomers are similar to one
29   another with regard to respiratory and developmental toxicity in acute and developmental studies
30   fSaillenfaitetal..2005: Korsak and Rydzynski. 1996: Korsak etal.. 19951 The 1,3,5-TMB isomer
31   was observed to induce some measures of neurotoxicity (e.g., passive and active avoidance] at
32   lower doses than 1,2,4-TMB,  in short-term studies (Wiaderna etal.. 2002: Gralewicz and Wiaderna.
33   2001: Gralewicz etal., 1997b]. Additionally, available toxicokinetic data regarding blood:air
34   partition coefficients, respiratory uptake, and absorption into the bloodstream in humans and rats


            This document is a draft for review purposes only and does not constitute Agency policy.

                                                xxxix            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   do not suggest any appreciable differences can be expected between the two isomers [Meulenberg
 2   and Vijverberg. 2000: TarnbergetaL 1996: Dahletal.. 1988).
 3          Therefore, the chronic RfC of 5 x 10 2 mg/m3 derived for 1,2,4-TMB was adopted as the
 4   RfC for 1,3,5-TMB. This is based on the determination of sufficient similarity with regard to
 5   chemical properties, kinetics, and toxicity between the two isomers (see Section 2.3.5).

     7. Confidence in the Chronic Inhalation  RfC for 1,3,5-TMB
 6          The chronic RfC for 1,2,4-TMB was adopted as the RfC for 1,3,5-TMB; thus, confidence in the
 7   study from which the critical effect was identified, Korsak and Rydzyhski [1996). is low to medium
 8   (see above). The database for 1,3,5-TMB  includes acute, short-term, and developmental toxicity
 9   studies in rats and mice. However, confidence  in the database is low to medium because it lacks
10   chronic, subchronic, multi-generation reproductive/developmental toxicity, and developmental
11   neurotoxicity studies and most of the studies supporting the critical effect come from the same
12   research institute.
13          Reflecting the confidence in the study and the database  and the uncertainty surrounding the
14   adoption of the RfC derived for 1,2,4-TMB as the RfC for 1,3,5-TMB, the overall confidence in the
15   RfC for 1,3,5-TMB is low.

     8. Effects Other Than Cancer Observed Following Oral Exposure
16          Only one subchronic study was identified that examined the effects of oral exposure to
17   1,3,5-TMB. Effects in the hematological system, including changes in clinical chemistry parameters
18   and differential white blood cell numbers, were observed following exposure to 1,3,5-TMB via oral
19   gavage. Ultimately, the Koch Industries [1995b) study was determined to not be suitable for RfD
20   derivation following an external peer review of the study (see Appendix F). No other subchronic
21   studies were found that investigated the  effects of oral exposure to 1,2,4-TMB or 1,2,3-TMB, and no
22   chronic oral studies were found that investigated noncancer effects of any of the TMB  isomers.
23          A series of studies utilizing single exposures (oral gavage or i.p. injection) were identified
24   that investigated the acute neurotoxic effects of TMBs (Tomasetal.. 1999a: Tomasetal.. 1999b:
25   Tomas etal.. 1999c). In these studies, exposed rats demonstrated changes in electrocortical arousal,
26   altered EEC activity in the cortical and hippocampal regions of the brain, and altered locomotor
27   activity in open field tests. As these effects were only observed in studies investigating acute
28   exposures, they were considered insufficient for derivation of oral toxicity reference values.
29          Therefore, given that Koch Industries study was not suitable for RfD derivation and effects
30   from acute studies generally are not suitable for derivation of chronic health values, RfDs were
31   derived for 1,2,4-TMB using route-to-route extrapolation and for 1,2,3-TMB and 1,3,5-TMB based
32   on sufficient similarity.
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                 xl             DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
 1           Table ES-4 below summarizes the RfDs derived for all three 1MB isomers, and the sections
 2    that follow provide details on the derivation of the RfD for each isomer.

            Table ES-4. Summary of reference doses (RfDs) for TMB isomers
Isomer
1,2,4-TMB
1,2,3-TMB
1,3,5-TMB
Source
Route-to-route extrapolation from
RfC for 1,2,4-TMB
Adopted from 1,2,4-TMB based on
sufficient similarity of these isomers
Adopted from 1,2,4-TMB based on
sufficient similarity of these isomers
Reference value
2 x 10"2
2 x 10"2
2 x 10"2
Confidence
Low
Low
Low
     9. Oral Reference Dose (RfD) for 1,2,4-TMB for Effects Other Than Cancer
            Table ES-5. Summary of reference dose (RfD) derivation for 1,2,4-TMB
Critical effect
Decreased pain sensitivity
90 day male rat study
Korsak and Rydzyriski
(1996)
Point of departure
Route-to-route
extrapolation using Korsak
and Rydzyriski (1996)
subchronic inhalation study
in Wistar rats
PODHED (mg/kg-day) = 6.3
Uncertainty
factor
300
Chronic RfD
(mg/kg-day)
2xl02
 3           A human PBPK model [Hissinketal.. 2007]. modified by EPA to include an oral
 4    compartment, was available for estimating the oral dose that would yield a blood concentration
 5    equal to the blood concentration at the POD used in the derivation of the RfC for 1,2,4-TMB (Section
 6    B.3.3.5, Appendix B). The RfD calculation is summarized in Table ES-5. Under the assumption of
 7    constant oral ingestion and 100% absorption of 1,2,4-TMB via constant infusion rate into the liver,
 8    a PODHED of 6.3 mg/kg-day was derived. Hepatic first-pass metabolism was also evaluated in
 9    humans using the modified PBPK model: following 50 days of low daily doses, inhalation doses
10    were estimated to result in steady state venous blood concentrations 4-fold higher than blood
11    concentrations resulting from equivalent oral doses due to hepatic first pass metabolism (see
12    Figure B-18, Appendix B). The same composite UF of 300 used for the RfC derivation was applied: 3
13    to account for uncertainty in extrapolating from laboratory animals to humans (interspecies
           This document is a draft for review purposes only and does not constitute Agency policy.

                                               xli            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   variability), 10 to account for variation in susceptibility among members of the human population
 2   (interindividual variability), 3 to account for subchronic-to-chronic extrapolation due to the use of a
 3   subchronic study, and 3 to account for deficiencies in the database (no multi-generation
 4   reproductive/developmental toxicity or developmental neurotoxicity studies). Dividing the PODHED
 5   by the composite UF of 300 yielded a chronic RfD of 2 x 10-2 mg/kg-day for 1,2,4-TMB.

     10. Confidence in the Chronic Oral RfD for 1,2,4-TMB
 6          A PBPK model was utilized to perform a route-to-route extrapolation to determine a POD
 7   for the derivation of the RfD from the Korsak and Rydzyhski [1996) inhalation study and
 8   corresponding critical effect. The confidence in the study from which the critical effect was
 9   identified, Korsak and Rydzyhski [1996), is low to medium (see above). The database for 1,2,4-TMB
10   includes acute, short-term, subchronic, and developmental toxicity studies in rats and mice.
11   However, confidence in the database for  1,2,4-TMB is low to medium because it lacks chronic,
12   multi-generation reproductive/developmental and developmental neurotoxicity studies, and the
13   studies supporting the critical effect predominantly come from the same research institute.
14          Reflecting the confidence in the study and the database and the uncertainty surrounding the
15   application of the available PBPK model for the purposes of a route-to-route extrapolation, the
16   overall confidence in the RfD for 1,2,4-TMB is low.

     11. Oral Reference Dose (RfD) for 1,2,3-TMB for Effects Other Than Cancer
17          The oral database is  inadequate to derive an RfD for 1,2,3-TMB. No chronic, subchronic, or
18   short-term oral exposure studies were found in the literature. However, as discussed in Sections
19   1.1.6 and B.2, the toxicokinetic and toxicity similarities between 1,2,3-TMB and 1,2,4-TMB support
20   adopting the RfD for 1,2,4-TMB as the RfD for 1,2,3-TMB. 1,2,3-TMB is observed to elicit the same
21   neurotoxic effects in rats (decreased pain sensitivity) as 1,2,4-TMB following subchronic inhalation
22   exposures, and the calculated RfCs for these two  isomers are equal: 5 x 1Q-2 mg/m3. In addition to
23   the outlined similarities in toxicokinetics, the qualitative metabolic profiles for the two isomers are
24   similar such that first-pass metabolism through the liver is not expected to differ greatly between
25   1,2,4-TMB and 1,2,3-TMB. Therefore, the chronic RfC of 2 x 10-2 mg/kg-day derived for
26   1,2,4-TMB was adopted as the RfD for  1,2,3-TMB based on the determination of sufficient
27   similarity between the two isomers with regard to chemical properties, toxicokinetics, and toxicity.

     12. Confidence in the Chronic Oral RfD for 1,2,3-TMB
28          The chronic oral RfD for 1,2,4-TMB was adopted as the chronic oral RfD for 1,2,3-TMB; thus,
29   confidence in the study from which the critical effect was identified, Korsak and Rydzyhski (1996).
30   is low to medium (see above). The database for 1,2,3-TMB includes acute, short-term, and
31   subchronic studies in rats and mice. However, confidence in the database is low to medium because


            This document is a draft for review purposes only and does not constitute Agency policy.

                                                xlii            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   it lacks chronic, multi-generation reproductive/developmental, developmental toxicity, or
 2   developmental neurotoxicity studies, and the studies supporting the critical effect predominantly
 3   come from the same research institute. Reflecting the confidence in the study and the database and
 4   the uncertainty surrounding the adoption of the RfD derived for 1,2,4-TMB as the RfD for
 5   1,2,3-TMB, the overall confidence in the RfD for 1,2,3-TMB is low.

     13. Oral Reference Dose (RfD) for 1,3,5-TMB for Effects Other Than Cancer
 6          The oral database is inadequate to derive an RfD for 1,3,5-TMB. No chronic, oral exposure
 7   study was found in the literature. However, one subchronic oral gavage study was identified that
 8   observed effects on multiple clinical chemistry parameters and monocyte levels [Koch Industries,
 9   1995b]. However, following an external peer review of this study (see Appendix F), it was
10   concluded that the Koch Industries [1995b] study was not suitable as the basis for quantifying the
11   noncancer human health risk following oral exposure. The most critical shortcoming of this study
12   was its failure to investigate the neurotoxicity of 1,3,5-TMB.
13          However, as determined for the RfC derivation for 1,3,5-TMB, the toxicokinetic and
14   toxicological similarities between 1,3,5-TMB and 1,2,4-TMB demonstrate sufficient similarity
15   between the two isomers to support adopting the RfD for 1,2,4-TMB for the RfD for 1,3,5-TMB. In
16   addition to the previously discussed similarities in toxicokinetics, the qualitative metabolic profiles
17   for the two isomers are similar to such a degree that first-pass metabolism through the liver is not
18   expected to differ greatly between 1,2,4-TMB and 1,3,5-TMB. Therefore, the chronic RfD of 2 x 10
19   2 mg/kg-day derived for 1,2,4-TMB was adopted as the RfD for 1,3,5-TMB based on the
20   determination of sufficient similarity between the two isomers with regard to chemical properties,
21   toxicokinetics, and toxicity.
        Confidence in the Chronic Oral RfD for 1,3,5-TMB
22          The chronic oral RfD for 1,2,4-TMB was adopted as the chronic oral RfD for 1,3,5-TMB; thus
23   confidence in the study from which the critical effect was identified, Korsak and Rydzyhski [1996].
24   is low to medium (see above). The database for 1,3,5-TMB includes acute, short-term, and
25   developmental toxicity studies in rats and mice. However, confidence in the database is low to
26   medium because it lacks chronic, multi-gene ration reproductive/developmental and
27   developmental neurotoxicity studies, and the studies supporting the critical effect predominately
28   come from the same research institute. Reflecting the confidence in the study and the database and
29   the uncertainty surrounding the adoption of the RfD derived for 1,2,4-TMB as the RfD for
30   1,3,5-TMB, the overall confidence in the RfD for 1,3,5-TMB is low.

     14. Evidence of Carcinogenicity
31          Under EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA.  2005a). there is
32   "inadequate information to assess carcinogenic potential" of TMBs. No chronic inhalation studies

            This document is a draft for review purposes only and does not constitute Agency policy.

                                               xliii            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   that investigated cancer outcomes were identified in the literature for 1,2,3-TMB, 1,2,4-TMB, or
 2   1,3,5-TMB. One cancer study in which rats were exposed to 1,2,4-TMB via oral gavage at one
 3   experimental dose of 800 mg/kg-day observed marginal increases in total malignant tumors and
 4   head tumors (e.g., neuroesthesioepitheliomas), but provided no statistical analyses of the results
 5   [Maltoni et al., 1997]. A number of methodological issues limit the utility of this study (e.g., only one
 6   dose group and no discussion of histopathological analyses). When Fisher's exact test was
 7   performed by EPA on the incidences calculated from the reported percentages of animals bearing
 8   tumors in the control and 800 mg/kg dose groups, no statistically significant elevations were
 9   observed. Therefore, a quantitative cancer assessment for TMBs was not conducted.

     15. Susceptible Populations and Lifestages
10          No chemical-specific data that would allow for the identification of populations or lifestages
11   with increased susceptibility to TMB exposure exist. However, some inferences can be made based
12   on the toxicokinetics of TMB isomers. TMB isomers are metabolized via side-chain oxidation to
13   form alcohols and aromatic carboxylic/mercapturic acids or by hydroxylation to form phenols,
14   which are then conjugated with glucuronic acid, glycine, or sulfates for urinary excretion. The
15   activities of multiple cytochrome P450  (GYP P450) mono-oxygenase isozymes and rates of
16   glucuronidation and sulfation conjugation are reduced in children up to 1 year in age, and renal
17   clearance is reduced in infants up to 2 months of age (Ginsberg et al., 2004]. Therefore, as GYP P450
18   mono-oxygenase activities, the  rate of glucuronidation and sulfation, and renal clearance appear to
19   be decreased in early life, newborns and young infants may experience higher and more persistent
20   blood concentrations of 1,2,3-TMB,  1,2,4-TMB, 1,3,5-TMB, and/or their respective metabolites
21   compared with adults at similar exposure levels. Additionally, those with pre-existing respiratory
22   diseases (e.g., asthma] may be more sensitive to the respiratory irritative and inflammatory effects
23   resulting from exposure to TMB isomers.

     16. Key Issues Addressed in the Assessment: Adoption of 1,2,4-TMB Toxicity Values
        for the 1,3,5- and 1,2,3-TMB Isomers
24          The toxicity database for 1,3,5-TMB was inadequate for derivation of either a reference
25   concentration or a reference dose. The chemical, toxicokinetic, and toxicological properties of the
26   individual isomers are sufficiently similar to one another to support adoption of 1,2,4-TMB's
27   reference values for 1,3,5-TMB  (see Section 2.3.5]. Both isomers are similar in their (1] chemical
28   properties (e.g., blood:tissue partition coefficients],  (2] toxicokinetic properties (i.e., absorption,
29   metabolism, and excretion profiles], and (3] toxicity profiles across studies utilizing multiple
30   durations of exposure and multiple  endpoints (i.e., neurological, respiratory, maternal, and fetal
31   effects]. Therefore, given these  similarities,  the RfC and RfD derived for 1,2,4-TMB were adopted as
32   the RfC and RfD for 1,3,5-TMB.
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                xliv             DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1          The toxicity database for 1,2,3-TMB was inadequate for derivation of a reference dose. No
 2   chemical-specific PBPK model is available for 1,2,3-TMB, and therefore, no route-to-route
 3   extrapolation can be performed on which to derive an RfD from the RfC for 1,2,3-TMB. The
 4   chemical, toxicokinetic, and toxicological properties of the individual isomers are sufficiently
 5   similar to one another to support adoption of 1,2,4-TMB's reference dose for 1,2,3-TMB (see
 6   Section 2.5.2). Both isomers are similar in their (1) chemical properties (e.g., blood:air and
 7   tissue:air partition coefficients), (2) toxicokinetic properties (i.e., the degree of absorption into the
 8   bloodstream between the two isomers indicates the internal blood dose metrics for 1,2,3-TMB
 9   would be similar to those calculated for 1,2,4-TMB by that isomer's available PBPK model), and (3)
10   toxicity profiles (i.e., the observation that both isomers affected pain sensitivity to an equal degree
11   and that the two isomer's RfCs for this effect were equal). Therefore, given these similarities, the
12   deficiencies in the 1,2,3-TMB oral database, and the lack of a 1,2,3-TMB PBPK model with which to
13   perform a route-to-route extrapolation, the RfD derived for 1,2,4-TMB was adopted as the RfD for
14   1,2,3-TMB.
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                 xlv             DRAFT—DO NOT CITE OR QUOTE

-------
                                                    Toxicological Review of Trimethylbenzene
     LITERATURE SEARCH STRATEGY  |  STUDY
     SELECTION AND EVALUATION
 1          The literature search strategy used to identify primary, peer-reviewed literature pertaining
 2   to TMBs was conducted using the databases and keywords listed in Table LS-1. References from
 3   health assessments developed by other national and international health agencies were also
 4   examined. Other peer-reviewed information, including review articles, literature necessary for the
 5   interpretation of TMB-induced health effects, and independent analyses of the health effects data
 6   were retrieved and included in the assessment where appropriate. EPA requested public
 7   submissions of additional information in April 2008; no submissions in response to the data call-in
 8   were received. A comprehensive literature search was last conducted in December 2011.
 9          Figure LS-1 depicts the literature search and study selection strategy and the number of
10   references obtained at each stage of the literature screening. Approximately 4,300 references were
11   obtained from the chemical name, keyword, and metabolite searches for 1,2,4-TMB,  1,2,3-TMB, and
12   1,3,5-TMB including references retrieved from specific literature searches necessary for the
13   interpretation of TMB-induced health effects (e.g., literature on specific modes of action, PBPK
14   analysis). From this full list of references, there were 218 references that were considered for
15   inclusion in the Toxicological Review.
16          Selection of studies for inclusion in the Toxicological Review was based on consideration of
17   the extent to which the study was informative and relevant to the assessment and general study
18   quality considerations. In general, the relevance of health effect studies was evaluated as outlined in
19   the Preamble and EPA Guidance (/I Review of the Reference Dose and Reference Concentration
20   Processes [U.S. EPA, 2002] and Methods for Derivation of Inhalation Reference Concentrations and
21   Application of Inhalation Dosimetry [U.S. EPA. 1994b]]. From the list of "considered"  references,
22   161 full text publications were identified as providing relevant information for use in the
23   development of this document, and included 30 studies in humans (e.g., occupational epidemiologic
24   studies, workplace exposure studies, and controlled human exposures), 63 inhalation or oral
25   animal studies, and 68 other studies (e.g., studies that provided supporting information on mode of
26   action, chemical properties, and susceptible subpopulations).
           This document is a draft for review purposes only and does not constitute Agency policy.

                                              xlvi            DRAFT—DO NOT CITE OR QUOTE

-------
                                                             Toxicological Review of Trimethylbenzene

1            The references that are cited in the document, as well as those that were considered but not
2    included in the Toxicological Review of TMBs, can be found within the Health and Environmental
3    Research Online (HERO) website3. This site contains HERO links to lists of references, including
4    bibliographic information and abstracts, which were considered for inclusion in the Toxicological
5    Review of TMBs.

             Table LS-1: Details of the search strategy employed for TMBs
             Databases
             EBSCO
             DISCOVERY
             SERVICE:
             HERO
             SCI
             NLM
             TOXLINE
             WOS
Keywords3'
Chemical name, CASRN, and synonym search:
1,2,4-trimethylbenzene, OR pseudocumene, OR 95-63-6;
1,2,3-trimethylbenzene, OR hemimellitene, OR 526-73-8;
1,3,5-trimethylbenzene, OR mesitylene, OR 108-67-8
Keyword search: neurotoxicity, genotoxicity, developmental toxicity, inflammation,
irritation, toxicokinetics, pbpk, mode of action, white spirit, C9, C9 fraction, JP-8
Additional search on specific metabolites:
2,3-dimethylbenzoic acid, OR 26998-80-1;
2,3-dimethylhippuric acid, OR 187980-99-0;
2,4-dimethylbenzoic acid, OR 611-01-8;
2,4-dimethylhippuric acid OR 41859-41-0;
2,5-dimethylbenzoic acid OR 610-72-0;
2,5-dimethylhippuric acid OR 41859-40-9;
2,6-dimethylbenzoic acid OR 632-46-2;
2,6-dimethylhippuric acid OR 187980-98-9;
3,4-dimethylbenzoic acid OR 619-04-5;
3,4-dimethylhippuric acid OR 23082-12-4;
2,4,5-trimethylphenol OR 496-78-6;
2,3,5-trimethylphenol OR 697-82-5;
2,3,6-trimethylphenol OR 2416-94-6;
2,4,6-trimethylphenol OR 527-60-6;
3,5-dimethylbenzoic acid OR 499-06-9;
3,5-dimethylhippuric acid OR 23082-14-6
             Potentially relevant publications on TMBs were identified through a literature search conducted with the EBSCO
              Discovery Service feature of Health and Environmental Research Online (HERO), a meta-search engine with access
              to numerous databases including the Science Citation Index (SCI), Toxicology Literature Online (TOXLINE), The
              National Library of Medicine (NLM, PubMed/Medline), and Web of Science (WOS).
             bLiterature search was performed using related words (i.e., lemmatization) of included search terms. Search terms
              were entered into the EBSCO Discovery Service portal with no qualifiers and the results from individual search
              engines were returned and exported to HERO.
      HERO is a database of scientific studies and other references used to develop EPA's risk assessments aimed at
     understanding the health and environmental effects of pollutants and chemicals. It is developed and managed in
     EPA's Office of Research and Development (ORD) by the National Center for Environmental Assessment (NCEA).
     The database includes more than 600,000 scientific articles from the peer-reviewed literature. New studies are added
     continuously to HERO.

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                     xlvii             DRAFT—DO NOT CITE OR QUOTE

-------
                                                            Toxicological Review of Trimethylbenzene
                    References identified based on initial keyword search (see Table LS-1): "4300
                                   V     L
                                            Reference excluded based on preliminary manual screen of
                                            titles/abstracts:  "4100
                                            Reasons for excluding references included the following:
                                            •  Published in non-relevant journals (i.e., Journal of Chemical
                                              Catalysis
                                              Topics not relevant to TMB toxicity
      Reference considered for inclusion in the Toxicological Review: 226
      Human Studies: 39
      Animal Studies: 78
      Other supporting studies: 144
      Including:
      • Reviews
      • Background and physical/chemical  properties
      « Studies related to mode-of-action
      • Studies related to toxicokinetics or PBPK analyses
                                             References excluded based on manual review of
                                             papers/abstracts: 65
                                             Types of papers evaluated and not considered further for
                                             Inclusion in the Toxicological Review:
                                             * Studies of exposure assessment
                                             • Studies of chemicals other than TMBs (including studies of
                                              complex solvent mixtures containing TMBs)
                                             * Non-relevant human and animal toxicity studies
                                             • C ase reports
                                             • Ecotoxicity studies
                                             • In vitro studies
                                             • Not available in English
      References cited in the Toxicological Review
      Human studies/reports: 30
      •Controlled Human Exposures: 10
      • Epidemiological Studies: 11
      •Hematatological: 3
      • Neurotoxicity: 9
      • Respiratory: 5
      •Toxicokinetic Studies: 14
      •Workplace Exposure studies: 10
Animal Studies: 63
• Cancer/Genotoxicity:
• Developmental: 5
•Hematological: 8
• Nephro toxicity: 1
• Neurotoxicity: 24
• Respiratory: 6
• Toxicokinetic: 20
Other supporting studies: 68
• Chemical properties: 6
•BlvlE) modeling: 6
• Exposure assessment: 4
• Mode of action: 22
•PBPIC27
•Susceptible populations: 3

Guidance documents: 41
•EPA: 23
• Non-EPA: 18
Note: Some references may provide information on more than one topic, and therefore, may be included in more than one
study type. Accordingly, the sum of the references for subcategories of studies is not expected to equal the number of
references for the larger category.

Figure LS-1. Literature search and study selection strategy for TMBs.
        This document is a draft for review purposes only and does not constitute Agency policy.

                                                   xlviii              DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
     l.HAZARD  IDENTIFICATION
     1.1. Synthesis of Evidence

     1.1.1.  Neurological Effects
 1          There is evidence in humans and animals that inhalation exposure to trimethylbenzenes
 2   (TMBs) induces neurotoxic effects. The human evidence comes from occupational studies involving
 3   complex volatile organic compound (VOC) mixtures that include TMBs; thus, effects cannot be
 4   attributed to any TMB isomer specifically. Prevalence rates of neuropsychological symptoms
 5   increased with exposure duration in dockyard painters, with symptoms related to motor
 6   coordination exhibiting the strongest association [Chen etal.. 1999): similarly, a significant
 7   association between exposure and impaired performance in short term memory (symbol digit
 8   substitution) and motor speed/ coordination (finger tapping) tests was observed in shipyard
 9   painters exposed to TMBs (isomers were not specified) and other solvents (Lee etal., 2005). A
10   significant, positive association between exposure symptoms (e.g., abnormal fatigue) and
11   1,2,4-TMB exposure, but not exposure to lower levels of 1,2,3-TMB or 1,3,5-TMB, was reported in
12   asphalt workers (Norsethetal.. 1991). Nervousness, tension, headaches, vertigo, and anxiety were
13   reported in paint shop workers exposed to 49-295 mg/m3 of a solvent mixture containing 50%
14   1,2,4-TMB, 30% 1,3,5-TMB, and unspecified amounts of 1,2,3-TMB  (listed as possibly present)
15   (Battig et al. (1956). as reviewed by MOE (2006) and Baettig et al. (1958)).
16          Additional evidence suggests damage or dysfunction of the  inner ear and increased
17   occurrence of vertigo following exposure to TMBs  and other organic solvents in paint and varnish
18   factories (Sulkowski etal.. 2002). Increased reaction time was significantly and consistently
19   associated with exposure in controlled, acute volunteer studies in which humans were exposed to
20   mixtures containing 1,2,4-TMB (Lammers etal.. 2007). although it is unclear whether 1,2,4-TMB or
21   other constituents within the mixtures were responsible for the observed effects. Uptake of TMBs
22   was reported in human volunteers exposed for 2 hours to either: 300 mg/m3 white spirit (WS,
23   corresponding to 11 mg/m31,2,4-TMB); 11 or 123 mg/m31,2,4-TMB; 123 mg/m31,2,3-TMB; or
24   123 mg/m31,3,5-TMB. However, effects on the central nervous system (CNS)  were based on
25   measures of overt CNS depression (heart rate and  pulmonary ventilation) and a subjective rating of
26   CNS symptoms (i.e., headache, fatigue, nausea, dizziness, and intoxication) (Tarnbergetal..  1997a:
27   Jarnbergetal.. 1996). For full details of the epidemiologic and controlled human exposures studies

            This document is a draft for review purposes only and does not constitute Agency policy.

                                              1-1            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   (including human subjects research ethics procedures), see individual study summary tables in
 2   Appendix B.
 3          In two studies examining the toxicokinetics of TMBs following controlled human exposures
 4   to 5-150 mg/m31,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB, no neurological abnormalities in routine
 5   clinical examinations were reported following exposure, although results data or details regarding
 6   the specific tests performed were not provided [Kostrzewskietal.. 1997: Kostrewski and
 7   Wiaderna-Brycht. 1995). Studies identifying an association between occupational exposure to TMB
 8   isomers and neurological effects are limited due to an inability to attribute effects due to 1,2,3-TMB,
 9   1,2,4-TMB, or 1,3,5-TMB individually versus those due to the other isomers or additional
10   constituents within the mixture. The studies detailing controlled exposures to human volunteers
11   are also limited for evaluating neurotoxicity to TMBs due to a lack of methods to adequately assess
12   CNS function and a lack of no-exposure controls, short exposure duration, and exposure of
13   individual subjects to different concentrations of TMB isomers.
14          In animals, there is consistent evidence of neurotoxicity following inhalation exposure, and
15   to a lesser extent following oral exposure, to either 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB; a summary
16   of the evidence pertaining to neurotoxic effects for TMBs is shown in Tables 1-1 and 1-2 for
17   inhalation and oral exposures, respectively. This information is presented graphically in Figures 1-1
18   to 1-4.
        Pain sensitivity
19          Decreased pain sensitivity has been observed following inhalation exposure to TMBs in
20   multiple studies conducted in male Wistar rats (Table 1-1; Figures 1-1 - 1-3). To test pain
21   responses following TMB exposure, animal studies have employed the hot plate test In this test, a
22   thermal stimulus is applied to determine pain sensitivity, as indicated by the animals' latency to
23   paw-lick following introduction of the stimulus. In short-term exposure studies, the animals were
24   subjected to an additional environmental challenge, namely foot shock, which itself decreases pain
25   sensitivity. Short-term TMBs exposure without the foot shock challenge did not result in
26   statistically significant effects on pain sensitivity in the hot plate test several weeks after exposures
27   had ended. Decreases in pain sensitivity have been observed at concentrations > 123 mg/m3 or >
28   492 mg/m3 following subchronic exposure to 1,2,4-TMB or 1,2,3-TMB, respectively (Wiaderna et
29   al., 2002: Gralewicz and Wiaderna, 2001: Korsak and Rydzynski, 1996). Decreased pain sensitivity
30   after a foot shock challenge was  observed at concentrations > 492 mg/m3 following short-term
31   exposure to 1,2,4-TMB  (Gralewicz and Wiaderna. 2001: Gralewicz etal.. 1997b). 1,3,5-TMB
32   (Wiaderna etal.. 2002: Gralewicz and Wiaderna. 2001). or 1,2,3-TMB (Wiaderna etal.. 1998:
33   Korsak and Rydzynski. 1996). although changes were not observed at 492 mg/m31,2,3-TMB
34   (latencies 75% longer than controls were not statistically significant) in another short-term
35   exposure study (Gralewicz and Wiaderna, 2001).
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-2             DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1          In the subchronic study [Korsak and Rydzynski. 1996}. inhalation of 1,2,4-TMB or
 2   1,2,3-TMB resulted in reduced pain sensitivity which occurred in a concentration-dependent
 3   manner. In short-term studies that examined a range of concentrations [Wiaderna et al.. 2002.
 4   1998: Gralewicz etal., 1997b], decreases in pain sensitivity after foot shock challenge following
 5   exposure to TMB isomers were non-monotonic. Differences in experimental design (discussed
 6   below) may account for the lack of monotonicity in these short-term studies, in contrast to the
 7   observations in Korsak and Rydzynski [1996]. Similar to the subchronic study, acute exposures to
 8   1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB induced concentration-dependent decreases in pain sensitivity,
 9   with ECso values of 4,172, 5,682,and 5,963 mg/m3 for increased latency to paw-lick compared to
10   controls, respectively [Korsak and Rydzynski. 1996: Korsak et al.. 1995].
11          The decreases in pain sensitivity measured in the subchronic and acute studies were
12   observed immediately after exposure [Korsak and Rydzynski, 1996: Korsak etal.,  1995], with no
13   significant effects persisting 2 weeks after subchronic exposures were terminated (i.e., increases in
14   latency were reduced from 95 to 12% or from 78 to 13% of controls at 1,230 mg/m31,2,4- or
15   1,2,3-TMB, respectively] fKorsak and Rydzynski. 1996: Korsak etal.. 19951 In contrast,
16   performance in the hot plate test after foot shock challenge was significantly impaired following
17   short-term exposure to the TMB isomers when tested 51 days after exposure (Wiaderna etal..
18   1998] (Wiaderna etal.. 2 002: Gralewicz and Wiaderna. 2001: Gralewicz et al.. 1997b], indicating a
19   persistence of these pain sensitivity- related effects.
20          The addition of a foot shock challenge to the hot plate tests following short-term (i.e., 4-
21   week], inhalation exposure to TMB isomers makes these experiments somewhat distinct from
22   those performed following subchronic exposure, as the foot shock challenge can elicit a cognitive
23   response from the animals in later hot plate test trials (see below] (Wiaderna etal.. 2002: Gralewicz
24   and Wiaderna. 2001: Wiaderna etal.. 1998: Gralewicz etal.. 1997b]. In the short-term studies
25   (Wiaderna etal.. 2 002: Gralewicz and Wiaderna. 2001: Wiaderna etal.. 1998: Gralewicz etal..
26   1997b], treatment-related, statistically significant  changes at > 492 mg/m31,2,3-TMB, 1,2,4-TMB,
27   or 1,3,5-TMB were observed 24 hours after rats were given a foot shock;  no consistent, significant
28   effects at any concentration were observed immediately following foot shock. Additionally, no
29   statistically significant effects were observed prior to foot shock at 50 days post-exposure; studies
30   did tend to observe increases in latency in non-shocked rats that were not statistically significant at
31   > 492 mg/m31,2,4-TMB  (up to 206% longer than controls], 1,3,5-TMB (up to 215% longer than
32   controls], or 1,2,3-TMB (up to 95% longer than controls], but these responses were highly variable
33   and not consistently observed across studies. As foot shock alone is known to cause transient
34   reductions in pain sensitivity, these findings suggest that inhalation exposure to TMBs prolongs
35   foot shock-induced reductions in pain sensitivity. However, although a lengthening of the foot
36   shock-induced decrease in pain sensitivity by TMBs exposure is the most likely reason for the
37   observed effects, and, accordingly, these responses are discussed in this context herein, this is not

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-3             DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   the only possible explanation. It is also plausible that cognitive effects resulting from TMBs
 2   exposure could contribute to the responses observed 24 hours after foot shock. Specifically, control
 3   groups may better associate the hot plate environment with the previously-applied aversive
 4   stimulus and more quickly withdraw their paws than their TMB-exposed counterparts, who may
 5   exhibit a decreased fear response or shorter retention of that fear-associated memory.
 6   Alternatively, since this test paradigm can cause the hot plate test apparatus to become associated
 7   with the effects of foot shock, inducing stress-related responses in the shocked animal such that
 8   subsequent exposure to the hot plate test apparatus alone can reduce sensitivity to pain (possibly
 9   via the release of endogenous opioids), prior TMBs exposure could amplify this effect. From the
10   data available, the relative contribution(s) of these behaviors to the observed effects cannot be
11   easily distinguished. Despite the possible overlap between contributing neurological processes in
12   this test paradigm, these observations are still regarded as significant and adverse, and clearly
13   indicate a persistence of neurological effects long after TMBs exposures have ceased.
14          Substantial differences in study design between short-term and subchronic studies make it
15   impossible to distinguish the particular aspects of the pain sensitivity phenotype that appear to be
16   latent and only manifest with an environmental challenge from those that appear to be reversible.
17   Regardless, the ability of male Wistar rats to respond to a thermal stimulus in the hot plate test was
18   consistently impaired following inhalation exposure to TMBs. The overall database indicates that
19   TMB isomers are similar in their capacity to decrease pain sensitivity following inhalation exposure
20   (Table 1-1; Figures 1-1 - 1-3). Pain sensitivity was not examined following oral exposure.
        Neuromuscular function and coordination
21          Human exposures to solvent mixtures containing 1,2,4-TMB (Lammers etal.. 2007] or
22   multiple TMB isomers [(Battig etal.. 19561. as reviewed by MOE (20061 and  (Lee etal.. 2005:
23   Sulkowskietal..20Q2: Battig etal.. 19581] result in effects that suggest alterations to
24   neuromuscular function and balance, including increased reaction time and vertigo. Animal studies
25   using rotarod performance, which tests motor coordination, balance, and overall neuromuscular
26   function, indicate  that inhalation of TMB isomers can affect neuromuscular system function (Table
27   1-1; Figures 1-1 and 1-2). Significant decreases in rotarod performance were observed at 1,230
28   mg/m31,2,4-TMB and > 493 mg/m31,2,3-TMB when tested immediately after  exposure for 13
29   weeks (Korsak and Rydzynski, 1996]: significant decreases in performance were also observed at
30   1,230 mg/m3after 4 or 8 weeks of exposure to 1,2,3-TMB or 1,2,4-TMB, respectively.  This impaired
31   function was still evident at 2 weeks post-exposure and, while not statistically significant for
32   1,2,4-TMB, may indicate long-lasting neuromuscular effects of subchronic exposures to 1,2,4-TMB
33   and 1,2,3-TMB. Acute inhalation exposure studies support this observation. Effects such as loss of
34   reflexes and righting responses, have been observed following acute inhalation exposure to 1,250-
35   45,000 mg/m31,2,4-TMB (MOE. 2006: Henderson. 20011. Similarly, acute exposure to 1,2,3-TMB,
36   1,2,4-TMB, or 1,3,5-TMB resulted in decreased performance in rotarod tests  immediately following

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                 1-4             DRAFT—DO  NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   exposure, with ECso values of 3,779 mg/m3,4,693 mg/m3, and 4,738 mg/m3, respectively [Korsak
 2   and Rydzynski. 1996: Korsak etal.. 1995). These results indicate that 1,2,4-TMB and 1,3,5-TMB are
 3   similar in their ability to impair neuromuscular function, balance, and coordination while
 4   1,2,3-TMB exposure may elicit effects at lower concentrations compared to the other two isomers.
 5   No studies evaluating oral exposure to TMB isomers address this endpoint.
 6          The neurobehavioral tests administered (i.e., hot plate and rotarod) in the subchronic and
 7   acute studies by Korsak and Rydzynski, [1996) and Korsak et al. [1995) appear to have been
 8   conducted on the same days; however, it is unclear whether the tests were performed sequentially
 9   in the same cohorts of animals. Performing the hot plate test immediately following the rotarod test
10   could introduce a potential confounder, as shock alone (such as that used as negative reinforcement
11   following rotarod failure, see Table B-30, Appendix B) can cause reductions in pain sensitivity.
12   Thus, if the tests were performed sequentially in the same animals, TMB-exposed animals failing
13   more often in the rotarod test may exhibit increases in paw-lick latency unrelated to treatment, as
14   compared to controls receiving less shock reinforcement However, the observations by Korsak and
15   Rydzynski, (1996) and Korsak et al. (1995) are supported by 2- to  3-fold increases in latency to
16   paw-lick that, although not statistically significant, were observed  50 days after termination of
17   short-term exposures to 492 mg/m31,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB (Gralewicz and Wiaderna.
18   2001): increases of this magnitude were not present in the studies evaluating multiple
19   concentrations of the isomers (Wiaderna et al., 2002,1998: Gralewicz etal., 1997b).
        Motor function and/or anxiety
20          Effects in open field testing have been consistently reported in oral and inhalation studies of
21   exposure to 1,2,4-TMB and 1,3,5-TMB, but not 1,2,3-TMB, in male rats (Table 1-1; Figures 1-1-1-3);
22   however, open field locomotion following injections with the stimulant, amphetamine, were
23   amplified by prior 1,2,3-TMB exposure, but not by prior 1,2,4-TMB exposure (Lutz etal., 2010)
24   Altered behaviors and locomotion in open field tests can be attributed to anxiety responses due to
25   open spaces and bright light, as well as changes to motor system function. Factors other than
26   anxiety and motor function (e.g., interpretation of olfactory or visual cues) may also contribute to
27   alterations in open field behavior, but these are unlikely to be drivers of the responses. As all but
28   one of the studies (Lutz etal.. 2010) observed animals for only  5 or 10 minutes, these experimental
29   tests are insufficient to identify the relative contribution(s) of the various neurological responses to
30   the observed increases in open field behaviors. Thus, EPA has concluded that decreased anxiety
31   and/or increased motor function are the most likely explanations for the TMB-induced effects.
32          Decreased anxiety and/or increased motor function at > 492 mg/m31,2,4-TMB or
33   1,3,5-TMB has been reported in short-term studies, as evidenced by increases in horizontal
34   locomotion or grooming activities (Lutz etal.. 2010: Gralewicz and Wiaderna. 2001: Gralewicz etal..
35   1997b). Statistically significant increases in horizontal locomotion were observed in short-term
36   studies assessing open field behavior following inhalation exposure to 1,2,4-TMB or 1,3,5-TMB

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                 1-5            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   [Lutz etal.. 2010: Gralewicz and Wiaderna. 2001). Non-monotonic increases in grooming were
 2   reported following short-term exposure to 1,2,4-TMB, although changes in horizontal locomotion
 3   were not statistically significant (increases of 3-35% were also non-monotonic) [Gralewicz etal..
 4   1997b]. No statistically significant effects on open field activity have been observed following short-
 5   term exposure of male rats to 1,2,3-TMB [Lutz etal., 2010: Gralewicz and Wiaderna, 2001:
 6   Wiaderna et al.. 1998). Open field locomotion following injections with the stimulant amphetamine
 7   was amplified by previous short-term exposure to 1,2,3-TMB, but not 1,2,4-TMB (which actually
 8   tended to inhibit amphetamine-induced increases in activity at 492 mg/m3), suggesting possible
 9   effects of 1,2,3-TMB on sensitization-type responses. As open field testing was conducted 14 or 25
10   days after termination of exposure in these studies and TMB isomers are cleared rapidly from the
11   body following the end of inhalation exposures (Section B.2, Appendix B),  the results suggest
12   persistence of the effects of 1,2,4-TMB and 1,3,5-TMB on anxiety and/or motor function following
13   clearance of the toxic moiety from the nervous system.
14          Slight, transient increases in locomotor activity were also observed in open field tests
15   immediately following acute, oral exposure to the TMB isomers (Table 1-2; Figure 1-4). Significant
16   increases in locomotor activity—measured as number of squares crossed  after exposure compared
17   with prior to exposure—were observed at 3,850 mg/kg for 1,2,4-TMB and 1,2,3-TMB, and at >
18   1,920 mg/kg for 1,3,5-TMB, with minimal concentration-effect or time-effect relationships and
19   negligible differences in the magnitude of the change in activity between isomers (Tomas etal.,
20   1999b]. Increases in locomotor activity were biphasic in nature. At early timepoints following
21   exposure, increased locomotor activity was associated with perturbed motor coordination and
22   tremor, whereas after 90 minutes, this apparent motor ataxia progressed to hind limb paralysis, full
23   immobility, and respiratory distress (e.g., tachypnea), leading to several deaths by 24 hours (Tomas
24   etal.. 1999b].
25          As mentioned previously, open field tests cannot easily distinguish between anxiety-related
26   responses and changes in motor activity. However, effects on motor activity were observed
27   following inhalation exposure to elevated concentrations of TMBs in several acute studies, although
28   the results are inconsistent with observations in open field tests. Decreased motor activity was
29   observed in male rats immediately after exposure to 5,000 mg/m31,2,4-TMB (McKee etal.. 2010].
30   Decreased motor activity was also reported in rats acutely exposed via inhalation to a mixture
31   containing TMB isomers (Lammers etal., 2007], but the use of a mixture precludes a determination
32   of the toxicity specifically associated with individual isomers. As biphasic changes in activity are
33   frequently observed following exposures to solvents, it is likely that the timing of the evaluations
34   conducted in the short-term versus acute studies, as well as the differing isomer concentrations,
35   may influence the consistency of these results.
36          Overall, exposure to 1,2,4-TMB and 1,3,5-TMB affects anxiety and/or motor function at
37   concentrations above 492 mg/m3, although the exact, potentially biphasic, concentration-response

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                 1-6             DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   relationship remains unclear. The results for 1,2,3-TMB are difficult to interpret, as no effects were
 2   observed following short term inhalation exposure while acute oral exposure elicited responses
 3   consistent with 1,2,4-TMB and 1,3,5-TMB. Although an explanation for this disparity is lacking,
 4   these data highlight a potential difference between 1,2,3-TMB and the other isomers, regarding
 5   altered motor function and/or anxiety.
        Cognitive function
 6          Cognitive function following exposure to TMB isomers alone has not been evaluated in
 7   humans or following oral exposure in animals; controlled exposure of human volunteers to
 8   mixtures containing TMBs did not indicate any effects on short-term learning and memory tests
 9   [Lammers etal., 2007]. Similarly, short-term spatial memory (radial maze performance) was
10   unaffected by exposure to either 1,2,4-TMB or 1,3,5-TMB via inhalation in animal studies
11   fWiaderna etal.. 2002: Gralewicz and Wiaderna. 2001: GralewlczetaL 1997bl Similarly, although
12   one study indicates a significant decrement in radial maze performance following exposure to
13   123 mg/m31,2,3-TMB t(Wiaderna etal.. 1998). higher concentrations had no effect[Wiadernaetal..
14   1998). preventing interpretations regarding the significance of this finding. In contrast, effects on
15   cognitive function in passive and active avoidance tests of conditioning behaviors were consistently
16   observed across multiple studies in male rats 6-8 weeks following short-term inhalation exposure
17   to the TMB isomers, although clear concentration-effect relationships were not observed (Table 1-
18   1; Figures 1-1-1-3). Comparing the results of the behavioral tests reveals that there are differences
19   in cognitive effects reported for each TMB isomer, as well as differences in the concentrations at
20   which the cognitive effects were observed.
21          In the passive avoidance tests, rats were conditioned to avoid stepping down from a small,
22   elevated platform (the impulse of rats is to step down in order to escape the bright light and
23   constrained, elevated space of the platform) through the use of a brief series of foot shocks applied
24   on the lower level.  It is important to clarify that these tests are distinct from tests of pain sensitivity
25   and that observations  of decreased step down latency in these tests do not contrast with the
26   increases in paw lick latency observed in hot plate tests; in fact, they may be complementary (see
27   below; note: the foot shocks used are of a much shorter duration than those used to induce
28   decreased pain sensitivity in the hot plate tests). Decreases in step-down latency in passive
29   avoidance tests, particularly at 7 days following foot shock conditioning, were observed 6-7 weeks
30   after short-term inhalation exposure to > 123 mg/m31,2,3-TMB and 1,3,5-TMB or > 492 mg/m3
31   1,2,4-TMB (Wiaderna etal..  2002: Gralewicz and Wiaderna. 2001: Wiaderna etal.. 1998: Gralewicz
32   etal.. 1997b). Differences in latency prior to foot shock were not observed. Decreases in latency
33   were consistently observed and similar in magnitude across all studies at 7 days post foot shock,
34   although the decreases were not statistically significant for 1,2,4-TMB or 1,2,3-TMB in the study by
35   Gralewicz and Wiaderna (2001). At 3 days post-foot shock, decreases in latency were less
36   consistent (i.e.,  statistically significant decreases were observed at 123 mg/m31,2,3-TMB and at

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                 1-7             DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   492 mg/m31,2,4-TMB, but not at other concentrations, and were not observed following exposure
 2   to 1,3,5-TMB), and only 123 mg/m31,2,3-TMB was shown to have an effect at 1 day post-foot shock.
 3   In these tests, the effects occurring several days following conditioning with foot shock are most
 4   likely attributable to a reduced ability to inhibit motor reactions (or a lowered motor threshold) in
 5   response to the fear-inducing environment. Alternative explanations involve possible contributions
 6   of the following in the TMBs exposed rats: a diminished fear response to the foot shock; decreased
 7   pain sensitivity leading to a less effective negative reinforcement by the (less painful) foot shock; or
 8   diminished retention of the fear-associated memory (i.e., from the foot shock). However, as
 9   statistically significant changes were  observed < 24 hours following foot shock only after exposure
10   to 123 mg/m31,2,3-TMB, neither diminished fear responses to the foot shock nor decreases in pain
11   sensitivity are likely to be the sole driver (s) of these effects. This suggests that, in this particular
12   test paradigm, TMBs exposure causes latent effects on neurological functions associated with the
13   persistence of adaptive behaviors to a fear-inducing stimulus. Despite the consistency of the results
14   at 7 days post-foot shock, these tests  are insufficient to pinpoint whether the effects of TMBs
15   exposure are specific to diminished memory retention, increased impulsivity, and/ or decreased
16   motor control.
17          Reduced performance in two-way active avoidance tests was observed in male rats
18   following short-term  inhalation exposure to > 492 mg/m31,2,4-TMB (Gralewicz and Wiaderna,
19   2001: Gralewicz etal.. 1997b). >  123  mg/m31.3.5-TMB (Wiaderna etal.. 2002: Gralewicz and
20   Wiaderna. 2001). and at 492 mg/m31,2,3-TMB (Wiaderna etal.. 1998):. The effects of TMBs were
21   particular to the learning component of the test (acquisition/ training session), rather than the
22   memory component (retention session 7 days later) (Wiaderna etal.. 2002: Gralewicz and
23   Wiaderna. 2001:  Wiaderna etal.. 1998). The conditioning or training of active avoidance behaviors
24   was based on avoiding a painful foot shock (the unconditioned stimulus) upon presentation of a
25   tone (conditioned stimulus). Similar to the interpretation of results from passive avoidance tests, it
26   is unclear whether and to what extent potential alterations in locomotor activity (rats had to shuttle
27   between compartments) and/ or pain sensitivity following exposure to TMB  isomers could
28   contribute to learning deficits in these tests.
29          Acute inhalation exposure studies provide some support for the observed effects of TMB
30   isomers on learned behaviors. Significant increases in response latency in psychomotor tasks,
31   observed immediately after exposure (effects did not persist to 24 hours later), were reported in
32   male rats following acute exposure to 5,000 mg/m31,2,4-TMB (McKee etal..  2010) or to 4,800
33   mg/m3 of a mixture containing TMBs (Lammers etal., 2007). The effects on active and passive
34   avoidance behaviors indicate that learning and/or long-term memory processes are affected by
35   exposure to the TMB  isomers. The data suggest that 1,3,5-TMB may be a more potent inducer of
36   toxic effects on cognitive function than 1,2,4-TMB and 1,2,3-TMB, as the effects following exposure
37   to 1,3,5-TMB were more consistent and sometimes occurred at lower concentrations than those

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-8            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   reported following exposure to the other two isomers. Overall, however, these differences were
 2   slight
 3          Controlled human exposure studies suggest that exposures of < 123 mg/m3 of the 1MB
 4   isomers do not cause overt CNS depression (measured as heart rate and respiration) (Tarnberg et
 5   al., 1996], although symptoms related to this effect (e.g., lightheadedness, fatigue) have been
 6   reported in workers occupationally exposed to mixtures containing TMBs. In mice, CNS depression
 7   has been observed following acute inhalation exposure to > 25,000 mg/m31,3,5-TMB, with similar
 8   effect levels for 1,2,4-TMB fACGIH. 20021
        Electrocortical activity
 9          Neurophysiological evidence from short-term inhalation studies in animals, as well as
10   supportive evidence from acute oral and injection studies, suggests that exposures to 1MB isomers
11   at lower concentrations (at least for 1,2,4-TMB) may affect parameters associated with brain
12   excitability. Decreases in a particular component of electrocortical arousal  (i.e., spike-wave
13   discharge, SWD, bursts in recordings from cortical-hippocampal electroencephalograms, EEGs)
14   were observed in male rats 120 days after short-term exposure to > 492 mg/m31,2,4-TMB
15   (statistically significant at 1,230 mg/m3), suggesting persistent functional changes in the rat CNS
16   (Gralewicz etal., 1997a). Altered EEC patterns can be induced by anesthetics as well as stimuli that
17   produce arousal, and may precede other measures of neurotoxicity (U.S. EPA. 1998). In recordings
18   from rats that were awake, but immobile (not exhibiting pronounced exploratory activity, as
19   determined by EEC morphology),  statistically significant decreases in the frequency of SWD
20   episodes were observed at 24 hours following short-term exposure to 492  mg/m31,2,4-TMB
21   (decreases that were not statistically significant were also observed at >492 mg/m31,2,4-TMB at
22   30 and 120 days after exposure) (Gralewicz et al.. 1997a).
23          Complementing these findings, dose-related decreases in the duration and number of SWD
24   bursts (termed high-voltage spindles) were observed at > 240 mg/kg of the TMB isomers
25   subsequent to acute oral exposure (Tomas etal.. 1999a] (Table 1-2; Figure 1-4). The stronger and
26   more persistent effects on electrocortical activity followed a pattern of 1,2,3-TMB > 1,3,5-TMB >
27   1,2,4-TMB (Tomas etal.. 1999a). Similarly, electrophysiological alterations in cortical and
28   hippocampal EEGs were more pronounced following i.p. injection of 1,2,3-TMB, with 1,2,4-TMB and
29   1,3,5-TMB exerting lesser effects (Tomas etal., 1999c). Although it is unclear whether these
30   changes affect related processes such as memory and seizure initiation/propagation, the observed
31   EEC abnormalities following inhalation (Gralewicz etal.. 1997a). oral (Tomas etal.. 1999a). and i.p.
32   (Tomas etal.. 1999c] exposure to TMB isomers provide supportive evidence of possible acute CNS
33   depression by TMB isomers (Tomas etal.. 1999a: Tomas etal.. 1999c] and indicate persistent (up
34   to 120 days post-exposure) (Gralewicz etal.. 1997a] alterations in CNS activity that may reflect an
35   adaptive response to TMB exposure.
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-9             DRAFT—DO NOT CITE OR QUOTE

-------
                                              Toxicological Review of Trimethylbenzene
   Neurological effects: Inhalation
Table 1-1. Evidence pertaining to neurological effects of TMBs in animals —
           inhalation exposures
Study Designa'b and Reference
Assay and Results (as response relative to control)
1,2,4-TMB
Pain sensitivity
0, 123, 492, 1,230 mg/m3, (recovery:
1,230 mg/m3 at 2 wks post-exposure)
90 day; Rat, Wistar, male, N = 10
Korsak and Rydzyriski (1996) Table B-30C
0, 492 mg/m3
4 wks; Rat, Wistar, male, N = 11
Gralewicz and Wiaderna (2001), Table B-26
0, 123, 492, 1,230 mg/m3
4 wks; Rat, Wistar, male, N = 15
Gralewicz et al. (1997b), Table B-24
Hot plate- exposure-dependent increase in paw-lick latency which
recovers by 2 weeks post-exposure:
Response after exposure: 0, 18, 79*, 95*%
Response at 2 weeks post-exposure: 0, ND, ND, 12%
Hot plate- increased paw-lick latency 24 hr after foot shock:
Response at 50 days post-exposure: 0, 206%
Response at 50 days post-exposure seconds after foot shock: 0, 25%
Response at 51 days post-exposure 24hr after foot shock: 0, 191*%
Hot plate- increased paw-lick latency 24 hr after foot shockd:
Response at 50 days post-exposure: 0, -6, 7, -9%
Response at 50 days post-exposure seconds after foot shock:
0,-8, 17, -11%
Response at 51 days post-exposure 24 hr after foot shock:
0, 2, 74*, 33*%
Neuromuscular function and coordination
0, 123, 492, 1,230 mg/m3, (recovery:
1,230 mg/m3 at 2 wks post-exposure)
90 day; Rat, Wistar, male, N = 10
Korsak and Rydzyriski (1996), Table B-30
Rotarod- exposure-dependent increase in failures at 13 weeks which
does not recover by 2 weeks post-exposure:
Response after 13 weeks of exposure: 0, 10, 20, 40*%
Response at 2 weeks post-exposure: 0, ND, ND, 30%
Motor function and/or anxiety
0, 123, 492, 1,230 mg/m3
4 wks; Rat, Wistar, male, N = 15
Lutzetal. (2010), Table B-35
0, 492 mg/m3
4 wks; Rat, Wistar, male, N = 11
Gralewicz and Wiaderna (2001), Table B-26
Open field- increased horizontal locomotion (distance traveled); no
overall effects with amphetamine challenge6:
Response at 2 weeks post-exposure with no challenge: 0, 100, 84, 154*%
Response to single amphetamine injection challenge: 0, 90, -25, 69%
Response to challenge after conditioning: 0, 43, -50, 31%
Open field- increased horizontal locomotion (number of crossings):
Response at 25 days post-exposure: 0, 61*%
Wo change in exploration (rearings) or grooming episodes
        This document is a draft for review purposes only and does not constitute Agency policy.

                                         1-10
                                                       DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene
Table 1-1. (Continued): Evidence pertaining to neurological effects of TMBs in
                            animals — inhalation exposures
Study Design '  and Reference
Assay and Results (as response relative to control)
0, 123, 492, 1,230 mg/m3
4 wks; Rat, Wistar, male, N = 15
Gralewicz et al. (1997b), Table B-24
Open field- increased grooming at middle concentration:
Response at 25 days post-exposure: 0, 82,147*, 76%
Wo change in horizontal locomotion (number of crossings) or exploration
Cognitive function
0, 492 mg/m3
4 wks; Rat, Wistar, male, N = 1
Gralewicz and Wiaderna (2001), Table B-26
Passive avoidance- decreased step-down latency 7 days post-foot shockf:
Response at 39 days post-exposure prior to foot shock: 0, 34%
Response at 42 days post-exposure 1 day after foot shock: 0, -23%
Response at 44 days post-exposure 3 days after foot shock: 0,-51 %
Response at 48 days post-exposure 7 days after foot shock: 0, -43%
[Note: statistical significance 7 days after foot shock was noted after the
highest and lowest responder from each group was excluded]
Active avoidance- decreased performance during training (learning):
Trials to reach avoidance criteria at 54-60 days post-exposure: 0, 58*%
Wo differences were noted during retraining (retention)
Radial maze- no notable change in performance 14-18 days
post-exposure
0,123, 492, or 1,230 mg/m3
4 wks; Rat, Wistar, male, N = 15
Gralewicz et al. (1997b), Table B-24
Passive avoidance- decreased step-down latency 3-7 days post-foot
shock:
Response at 39 days post-exposure prior to foot shock:
0, 26, 41, -31%
Response at 42 days post-exposure 1 day after foot shock:
0, 95, -28, -87%
Response at 44 days post-exposure 3 days after foot shock:
0, 7, -67*, -36%
Response at 48 days post-exposure 7 days after foot shock:
0, -20, -79*, -47*%
Active avoidance- decreased performance during acquisition (learning)8:
Slower increases in avoidance performance across trials: p < 0.003
Non-significant decrease in total avoidance responses: p = 0.08
Radial maze- no notable change in performance 14-18 days
post-exposure
Electrocortical activity
0, 123, 492, 1,230 mg/m3
4 wks; Rat, Wistar, male, N = 9
Gralewicz et al. (1997a), Table B-25
EEG recordingsh- decreased spike wave discharge (SWD) bursts/ hr:
Response at 120 days post-exposure: 0,13, -35, -55*%
Wo change in global arousal level or in SWD/hr at 1 or 30 days
post-exposure
       This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-11             DRAFT—DO NOT CITE OR QUOTE

-------
                                              Toxicological Review of Trimethylbenzene
Table 1-1. (Continued): Evidence pertaining to neurological effects of TMBs in
                       animals — inhalation exposures
Study Designa'b and Reference
Assay and Results (as response relative to control)
1,2,3-TMB
Pain sensitivity
0, 123, 492, 1,230 mg/m3, (recovery:
1,230 mg/m3 at 2 wks post-exposure)
90 days; Rat, Wistar, male, N = 10
Korsak & Rydzyriski (1996), Table B-30
0, 492 mg/m3
4 weeks; Rat, Wistar, male, N = 11
Gralewicz and Wiaderna (2001), Table B-26
0, 123, 492, 1,230 mg/m3
4 weeks; Rat, Wistar, male, N = 15
Wiaderna et al. (1998), Table B-42
Hot plate- exposure-dependent increase in paw-lick latency which
recovers by 2 weeks post-exposure:
Response after exposure: 0, 22*, 68, 78*%
Response at 2 weeks post-exposure: 0, ND, ND, 13%
Hot plate- no statistically significant change in paw-lick latency:
Response at 50 days post-exposure: 0, 95%
Response at 50 days post-exposure seconds after foot shock: 0, -1%
Response at 51 days post-exposure 24 hr after foot shock: 0, 75%
Hot plate- increased paw-lick latency 24 hr after foot shock at middle
concentration:
Response at 50 days post-exposure: 0, -28, -13, -12%
Response at 50 days post-exposure seconds after foot shock: 0, -9, -16, -
15%
Response at 51 days post-exposure 24 hr after foot shock: 0, -19, 45*, 8%
Neuromuscular function and coordination
0, 123, 492, 1,230 mg/m3, (recovery:
1,230 mg/m3 at 2 weeks post-exposure)
90 days; Rat, Wistar, male, N = 10
Korsak and Rydzyriski (1996), Table B-30
Rotarod- exposure-dependent increase in failures at 13 weeks which
does not recover by 2 weeks post-exposure:
Response after 13 weeks of exposure: 0, 20, 40*, 70*%
Response at 2 weeks post-exposure: 0, ND, ND, 50*%
Motor function and/or anxiety
0, 123, 492, 1,230 mg/m3
4 weeks; Rat, Wistar, male, N = 15
Lutzetal. (2010), Table B-35
0, 492 mg/m3
4 wks; Rat, Wistar, male, N = 11
Gralewicz and Wiaderna (2001), Table B-26
Open field- statistically significant increase in horizontal locomotion
(distance traveled) only after amphetamine challenge6:
Response at 2 weeks post-exposure with no challenge: 0, 96, 85, 115%
Response to single amphetamine injection challenge: 0, 15, 198*, 111%
Response to challenge after conditioning: 0, -21, 103*, 41%
Open field- no change in horizontal locomotion (crossings):
Response at 25 days post-exposure: 0, -9%
Wo change in exploration (rearings), or grooming
      This document is a draft for review purposes only and does not constitute Agency policy.

                                       1-12            DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene
Table 1-1. (Continued): Evidence pertaining to neurological effects of TMBs in
                            animals — inhalation exposures
Study Design '  and Reference
Assay and Results (as response relative to control)
0, 123, 492, 1,230 mg/m3
4 wks; Rat, Wistar, male, N = 15
Wiaderna et al. (1998), Table B-42
Open field- no significant change in horizontal locomotion (crossings):
Response at 25 days post-exposure: 0,19, 51, 37%
Wo statistically significant change' in exploration (rearings) or grooming
Cognitive function
0, 492 mg/m3
4 wks; Rat, Wistar, male, N = 11
Gralewicz and Wiaderna (2001), Table B-26
Active avoidance- decreased performance during training (learning):
Trials to reach avoidance criteria at 54-60 days post-exposure: 0, 53*%
Wo differences were noted during retraining (retention)
Passive avoidance- no significant change in step down latency':
Response at 39 days post-exposure prior to foot shock: 0, -39%
Response at 42 days post-exposure 1 day after foot shock: 0, -40%
Response at 44 days post-exposure 3 days after foot shock: 0,-23 %
Response at 48 days post-exposure 7 days after foot shock: 0, -28%
Radial maze- no notable change in performance 14-18 days
post-exposure
0,123, 492, or 1,230 mg/m
4 wks; Rat, Wistar, male, N = 15
Wiaderna et al. (1998), Table B-42
Passive avoidance- decreased step-down latency after foot shock:
Response at 39 days post-exposure prior to foot shock: 0, -41, -37,19%
Response at 42 days post-exposure 1 day after foot shock: 0, -74*, -52, -
43%
Response at 44 days post-exposure 3 days after foot shock: 0, -54*, -49, -
14%
Response at 48 days post-exposure 7 days after foot shock: 0, -50*, -62*,
-37%
Active avoidance- decreased performance during training (learning):
Trials to reach avoidance criteria at 54-60 days post-exposure: 0, 3, 41*,
14%
Wo statistically significant differences noted during retraining (retention)
Radial maze- decreased performance at low concentration':
Increased errors on trial day 3: 0, 32*, -28, -4% & day 5: 0, 30*, -16,1%
Wo notable change in trial duration at any day (14-18 days
post-exposure)
       This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-13              DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene
Table 1-1. (Continued): Evidence pertaining to neurological effects of TMBs in
                           animals — inhalation exposures
Study Design '  and Reference
Assay and Results (as response relative to control)
1,3,5-TMB
Pain sensitivity
0, 492 mg/m3
4 wks; Rat, Wistar, male, N = 11
Gralewicz and Wiaderna (2001), Table B-26
Hot plate- increased paw-
Response at 50 days post-
Response at 50 days post-
Response at 51 days post-
lick latency 24 hr after foot shock:
exposure: 0, 215%
exposure seconds after foot shock: 0, 26%
exposure 24 hr after foot shock: 0,  246*%
0, 123, 492, 1,230 mg/m3
4 wks; Rat, Wistar, male, N = 12
Wiaderna et al. (2002), Table B-43
Hot plate- increased paw-lick latency 24 hr after foot shock at middle
concentration:
Response at 50 days post-
Response at 50 days post-
4%
exposure: 0, -6, 36, 24%
exposure seconds after foot shock: 0, -14, 8, -
                                      Response at 51 days post-exposure 24 hr after foot shock: 0, -4, 68*, 18%
Motor function and/or anxiety
0, 492 mg/m3
4 wks; Rat, Wistar, male, N = 11
Gralewicz and Wiaderna (2001), Table B-26
Open field- increased horizontal locomotion (number of crossings):
Response at 25 days post-exposure: 0, 65*%
Wo change in exploration (rearings) or grooming
Cognitive function
0, 123, 492, 1,230 mg/m3
4 wks; Rat, Wistar, male, N = 12
Wiaderna et al. (2002), Table B-43
Passive avoidance- decreased step-down latency 7 days post-foot shock:
Response at 39 days post-exposure prior to foot shock: 0, -5,146, 40%
Response at 42 days post-exposure 1 day after foot shock: 0, 99,108,
113%
Response at 44 days post-exposure 3 days after foot shock: 0, -32, -41, -
40%
Response at 48 days post-exposure 7 days after foot shock: 0, -47*, -53*,
-43*%
Active avoidance- decreased performance during training (learning):
Trials to reach avoidance criteria at 54-60 days post-exposure: 0,40*,
35*, 50*%
Radial maze- no notable change in performance 14-18 days post-
exposure
       This document is a draft for review purposes only and does not constitute Agency policy.

                                               1-14             DRAFT—DO NOT CITE OR QUOTE

-------
                                                          Toxicological Review of Trimethylbenzene

Table 1-1. (Continued): Evidence pertaining to neurological effects of TMBs in
                             animals — inhalation exposures
Study Design ' and Reference
Assay and Results (as response relative to control)
Cognitive function (continued)
0, 492 mg/m3
4 wks; Rat, Wistar, male, N = 11
Gralewicz & Wiaderna (2001), Table B-26
Passive avoidance- decreased step-down latency 7 days post-foot shock8:
Response at 39 days post-exposure prior to foot shock: 0, -3%
Response at 42 days post-exposure 1 day after foot shock: 0, -61%
Response at 44 days post-exposure 3 days after foot shock: 0,-65%
Response at 48 days post-exposure 7 days after foot shock: 0, -57*%
[Note: statistical significance 3 days after foot shock was noted after the
highest and lowest responder from each group was excluded]
Active avoidance- decreased performance during training (learning):
                                         Trials to reach avoidance criteria at 54-60 days post-exposure: 0, 65*%
                                         Radial maze- no notable change in performance 14-18 days
                                         post-exposure
*Significantly different from controls (p< 0.05).
Notes: For studies other than Korsak and Rydzynski (1996). % change from control calculated from digitized data using Grab
It! XP software; ND= Not determined
aRotarod and hot plate tests were administered immediately after termination of exposure or following a 2 week recovery
period by Korsak and Rydzynski (1996). EEC recordings were acquired prior to exposure and one, 30, or 120 days after
exposure by Gralewicz et al. (1997a). Motor behavior in an open field (tested for 30 min) was assessed 14 days after
exposure and re-tested following single and multiple (to induce sensitization) injections with amphetamine for 120 min by
Lutz et al. (2010). For the remaining studies (Wiaderna etal.. 2002: Gralewicz and Wiaderna. 2001: Wiaderna etal.. 1998:
Gralewicz etal.. 1997b): radial maze tests were administered prior to exposure and on days 14-18 after exposure; open field
activity (tested for 5-10 minutes) was assessed prior to exposure and on day 25 after exposure; passive avoidance was tested
on days 35-48 after exposure; hot plate sensitivity was assessed on days 50 and 51 after exposure; and active avoidance tests
were administered on or after day 54 post-exposure.
bln instances where authors reported exposures in ppm, EPA converted these values to mg/m3. See Appendix B (Table B-l)
for conversion factor, and individual study summary tables for ppm values.
cTables referenced in Study Design and Reference column correspond to study summary tables in Appendix B
dObservations of hot plate latency were made prior to  (LI); immediately following (L2); or 24hr after foot shock (L3). Values
for L3 in Graleiwicz et al. (1997b) were determined from reported values for LI and the ratio of L3/L1 x 100.
eNo challenge= prior to amphetamine challenge, evaluated for 30 min, and reported as Block 1: statistical significance
indicated in study text only; amphetamine challenge-induced activity was measured following a single injection or following
a single injection challenge after conditioning with 5 daily injections and evaluated for 120 min
fResults of passive avoidance tests in Gralewicz and Wiaderna (2001) may reflect adjusted data where, due to large individual
differences, 2 rats (the highest and lowest responders  to foot shock) in each group were excluded. As a result, the exact
magnitude of change is assumed to be somewhat inaccurate and statistical comparisons of the modified groups are provided
in the above evidence table only as notes.
sAt 54 days post-exposure, TMB-exposed rats were slower to increase the percentage of avoidance responses across blocks
(1 block = 5 trials). This reduction in avoidance responses across blocks appeared  to be lowest (although not statistically
significant) at 1,230 mg/m3. Rats were also observed to have a lower (p = 0.08) number of avoidance responses in the whole
30-trial session.
hElectroencephalograms (EEGs) were recorded at electrodes implanted in the fronto-parietal cortex and the dorsal
hippocampus (one recording from each region was analyzed for each rat).
'Dose-dependent increases in exploration and nonlinear increases in grooming were not statistically significant
iData represents % change  relative to control in same trial day, but statistical significance determined by the authors based
on comparison to trial day 1 responses within the same group.
        This document is a draft for review purposes only and does not constitute Agency policy.

                                                  1-15               DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene
   Neurological effects: Oral
Table 1-2. Evidence pertaining to neurological effects of 1,2,3-TMB, 1,2,4-TMB, or
             1,3,5-TMB in animals — oral exposures
            ,a,b
Study Design3'  and Reference
Assay and Results (as response relative to control)
1,2,4-TMB
Motor function and/or anxiety
0, 960,1,920, 3,850 mg/kg single oral gavage
Rat, Wag/Rij, male, N = 10
Tomas et al. (1999b), Table B-40
Open field- transient increases in locomotor activity:
Response at 20 min after exposure relative to pre-injection controls: 0,
34.1, 57.8, 60.6*%
Wo significant changes were reported at 10, 30, 40, 50, 60, or 70 min
Electrocortical activity
0, 240, 960, 3,850 mg/kg, single oral gavage
Rat, Wag/Rij, male, N = 6
Tomas et al. (1999a), Table B-39
EEG recordings - inhibition of the duration and number of high voltage
spindle episodes (response relative to vehicle control):
                                         Duration
                                         Number
                                                   20 min
          0, -72, -58, -83%
          0, -26, -44,-62*%
                           40 min
0, -80*, -97*, -45%
0, -53*,-88*,-73*%
                   60 min
0, 11, -67, -45%
0, 7, -53*, -22%
1,2,3-TMB
Motor function and/or anxiety
0, 960,1,920, 3,850 mg/kg single oral gavage
Rat, Wag/Rij, male, N = 10
Tomas et al. (1999b), Table B-40
Open field- transient increases in locomotor activity:
Response at 20 or 30 min after exposure relative to pre-injection controls:
0, 30.9, 26.5, 56.1*% (increased 65.6*% at 30 min in at the highest
concentration
Wo significant changes were noted at 10, 40, 50, 60, or 70 min
Electrocortical activity
0, 960, 3,850 mg/kg, single oral gavage
Rat, Wag/Rij, male, N = 6
Tomas et al. (1999a), Table B-39
EEG recordings - inhibition of the duration and number of high voltage
spindle episodes (response relative to vehicle control):
                                         Duration
                                         Number
                                                   20 min
         0, -86, -97*, -76*%
         0,-71*,-86*,-48%
                           40 min
0, -95, -98*, -97*%
0, -84*,-93*,-86*%
                   60 min
0, -81, -94*, -99*%
0, -70*,-99*,-96*%
         This document is a draft for review purposes only and does not constitute Agency policy.
                                                 1-16
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

Table 1-2 (Continued): Evidence pertaining to neurological effects of 1,2,3-TMB,
 	1,2,4-TMB, or 1,3,5-TMB in animals — oral exposures	
Study Design3' and Reference
Assay and Results (as response relative to control)
1,3,5-TMB
Motor function and/or anxiety
0, 960,1,920, 3,850 mg/kg single oral gavage
Rat, Wag/Rij, male, N = 10
Tomas et al. (1999b), Table B-40
Open field- transient increases in locomotor activity:
Response at 20 min after exposure relative to pre-injection controls: 0, 0,
46.7*, 42.4*% (increased 65-70% at 40-60 min at the highest
concentration
Wo significant changes were noted at 10, 30, or 70 min
Electrocortical activity
0, 240, 960, 3,850 mg/kg, single oral gavage
Rat, Wag/Rij, male, N = 6
Tomas et al. (1999a), Table B-39
EEG recordings - inhibition of the duration and number of high voltage
spindle episodes (response relative to vehicle control):
                                          Duration
                                          Number
                                                    20 min
          0, -76*, -79,-86%
          0, -57,- 67, -77%
                           40 min
0, -85*,-97*,-95*%
0,-52*,-93*,-91*%
                   60 min
0, -66*,-94*,-88*%
0,-49*,-91*,-89*%
*Significantly different from controls (p < 0.05).
Note: % change from control calculated from digitized data using Grab It! XP software.
aLocomotor activity in open field tests and electrocortical arousal were assessed prior to exposure and immediately after
exposure every 10 minutes for up to 70 minutes.
bln instances where authors reported exposures in ppm, EPA converted these values to mg/m3. See Appendix B (Table B-l) for
conversion factor, and individual study summary tables for ppm values.
cTables referenced in Study Design and Reference column correspond to study summary tables in Appendix B.
dElectroencephalograms (EEGs) were recorded prior to exposure and at 20,40, and 60 minutes after exposure via electrodes
implanted in the fronto-parietal cortex.
       This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-17             DRAFT—DO NOT CITE OR QUOTE

-------
                                                    Toxicological Review of Trimethylbenzene
   Exposure Response Arrays
EffectSummary Duration
4< Pain Sensitivity Sub-chronic
4, Pain Sen sitivity (after .
t i i i i ii i Short-term
foot shock challenge,!
Impaired Neuro- . .
h Sub-chronic
muscular Function
T Motor Activity
an d/ori Anxiety Short-term
Altered Cognitive Function Short-term
4- Electro-cortical Activity Short-term
1,2,4-TMB
End point
t- Paw-lick latency
(hot pi ate) (a)
f1 Paw-lick latency
(hot plate) (d)
T Paw-lick latency
(hotplate) (b)
T Rotarod failure (a)
•T- Horizontal
Locomotion (e)
N/CAnnph,- Induced
Locomotion (e)
•T- Horizontal
Locomotion (d)
t Grooming Activity (b)
N/C Passive
Avoidance (d)
4/ Active Avoidance
W
-,J,- Passive
Avoidance (b)
4- Active
Avoidance (b)
N/C Radial Maze (b)
N/C Radial Maze (d)
4- SWD bursts (c)
i- LOW Dose -i High Dose 4 NOAEL • LOAEL

•
«|
1
[

•
• II
•
•
f
|
•
I

                                                       20      500      ;50     1,000     1,30
                                                             Concentration [mg/mSj


Note: Solid lines represent range of exposure concentrations, (a) Korsak and Rydzynski (1996): (b) Gralewicz et al.
(I997b); (c) Gralewicz et al. (I997a); (d) Gralewicz and Wiaderna (2001): (e) Lutz et al. (2010). All effects are in male
Wistar rats.

Figure 1-1. Exposure response array of neurological effects following inhalation
            exposure to 1,2,4-TMB.
       This document is a draft for review purposes only and does not constitute Agency policy.

                                            1-18            DRAFT—DO NOT CITE OR QUOTE

-------
                                                    Toxicological Review of Trimethylbenzene
Effect Summary Duration
•^ Pain Sensitivity Sub-chronic
\|/ Pain Sensitivity [after ,
t t ti ' Short-term
footshock challenge)
impaired Neuro- Sub.chronic
muscular Function
N/C Motor Activity
,, „ . ' Short-term
and/or Anxiety
Altered Cognitive Function Short-term
1,2,3-TMB
Endpoint
•t Paw-lick latency
(hot plate) (a)
N/C Paw- lick latency
(hot plate) (b)
•f Paw-lick latency
(hot plate) (c)
1* Rotarod failure (a)
N/C Horizontal
Locomotion (d)
1* Arnph,- Induced
Locomotion (d)
N/C Open Field
Activity(b)
N/C Open Field
Activity (c)
N/C Passive
Avoidance (b)
\|/ Passive
Avoidance (c)
J/ Active Avoidance
(b)
-J, Active
Avoidance (c)
N/C Radial Maze (b)
4, Radial Maze (c)
i- Low Dose -i High Dose * NOAEL • LOAEL
• i
•
«i
• i
i
• I
•
i
•
• i
•
• I
•
• i

250      500      750      1,000

       Concentration (mg/m3)
                                                                                       1,250
Note: Solid lines represent range of exposure concentrations, (a) Korsak and Rydzynski (1996) ; (b) Gralewicz and
Wiaderna (2001): (c) Wiaderna et al. (1998): (d) Lutz et al. (2010). All effects are in male Wistar rats.

Figure 1-2. Exposure response array of neurological effects following inhalation
            exposure to 1,2,3-TMB.
       This document is a draft for review purposes only and does not constitute Agency policy.


                                            1-19             DRAFT—DO NOT CITE OR QUOTE

-------
                                                   Toxicological Review of Trimethylbenzene
EffectSummary Duration
4 Pain Sensitivity [after
, . i iii, . Short-term
foot shock challenge)

T- Motor Activity
an d/ori Anxiety Short-term
Altered Cognitive Function Short-term




1,3,5-TMB
End point
11 Paw-lick latency
(hot plate) (a)
T" Paw-lick latency
(hot plate) (b)
T Horizontal
Locomotion (a)
4, Passive
Avoidance (a)
J, Passive
Avoidance (b)
4, Active
Avoidance (a)
4r Active
Avoidance (b)
N/C Radial Maze (a)
N/C Radial Maze(b)
i- Low Dose -i High Dose * NOAEL • LOAEL
•
«i

•
•
• 1

•


*

                                                      250      500      750     1,000     1,350
                                                             Concentration (mg/m3)


Note: Solid lines represent range of exposure concentrations, (a) Gralewicz and Wiaderna (2001): (b) Wiaderna et al.
(2002). All effects are in male Wistar rats.

Figure 1-3. Exposure response array of neurological effects following inhalation
            exposure to 1,3,5-TMB.
       This document is a draft for review purposes only and does not constitute Agency policy.

                                            1-20            DRAFT—DO NOT CITE OR QUOTE

-------
                                                    Toxicological Review of Trimethylbenzene
1,2,4-TMB, 1,3,5-TMB,
Isomer Effect Summary
1,2,4-TMB t Motor Activity
and/ or 4/ Anxiety
•4/ Electro-
cortical Activity
•t Motor Activity


and/or \|/ Anxiety
4- Electro-
cortical Activity
•t Motor Activity
and/or \|/ Anxiety
4- Electro-
cortical Activity
or 1,2,3-TMB
Duration Endpoint
"t Horizontal
Acute . .
Locomotion (a)
4- HVS Episodes
Acute .,r.
and Duration (b)
"t Horizontal
Acute
Locomotion (a)
4- HVS Episodes
Acute , ' ...
and Duration (b)
•t Horizontal
Acute ,
Locomotion (a)
4, HVS Episodes
Acute , '
and Duration (b)
i- Low Dose H High Dose • NOAEL • LOAEL
I
• 1
I
• i
• i
• i


                                                      0   500  1,000 1,500 2,000  2,500 3,000  3,500 4,OOC
                                                                    Dose (mg/kg)


Note: Solid lines represent range of exposure concentrations, (a) Tomas et al. (I999a); (b) Tomas et al. (I999b). All effects
are in male WAG/Rij (Tomas et al. (I999a)) or Wistar (Tomas et al. (I999b)) rats.

Figure 1-4. Exposure response array of neurological effects following oral exposure
            to 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB.
       This document is a draft for review purposes only and does not constitute Agency policy.

                                            1-21             DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene

     1.1.1.1.  Mode of Action Analysis - Neurological Effects
 1          The observation of neurotoxicity following acute-, short-term-, and subchronic-duration
 2   exposure to 1MB [Lutz etal.. 2010: Lammers etal.. 2007: Wiadernaetal.. 2002: Gralewicz and
 3   Wiaderna. 2001: Wiadernaetal.. 1998: Gralewicz etal.. 1997b: Gralewicz etal.. 1997a: Korsakand
 4   Rydzynski. 1996: Korsak etal.. 1995] may indicate thatTMB perturbs normal neurotransmission in
 5   exposed animals, although the specific key events necessary for TMB-induced neurotoxicity are not
 6   established. Although mechanistic and mode-of-action data is lacking for TMBs, structurally similar
 7   compounds like toluene and xylene have been more thoroughly characterized and it is reasonably
 8   assumed that TMBs would operate through a similar mechanism in producing the resultant
 9   neurotoxicological effects. Aromatic hydrocarbons are known to interact with catecholaminergic
10   systems [Kyrklund. 1992}. Inhalation exposures to toluene and xylene have been shown to
11   significantly change concentration and turnover rate of both dopamine and norepinephrine in
12   various regions of the rat brain [Reaetal.. 1984: Anderssonetal.. 1983: Anderssonetal.. 1981:
13   Andersson et al.. 1980]. These changes have been hypothesized to be due to potential metabolites
14   with affinity to catecholamine receptors that would, in turn, influence the uptake and release of
15   neurotransmitters [Anderssonetal., 1983: Anderssonetal., 1981: Anderssonetal., 1980].
16          Catecholaminergic changes with toluene have been reported and are similar to that
17   observed with TMBs which would therefore increase the plausibility that the mechanisms of
18   neurotoxicity are similar between the two compounds. For example, subchronic inhalation
19   exposures of rats to low concentrations of toluene (as low as 80 ppm [300 mg/m3]] have been
20   shown to decrease spatial learning and memory, increase dopamine-mediated locomotor activity,
21   increase the number of dopamine D2 receptors, and increase dopamine D2 agonist receptor
22   binding (Hillefors-Berglund etal.. 1995: vonEuler etal.. 1994: vonEuler etal.. 1993]. These effects
23   were observed to persist up to four weeks after the termination of the toluene exposure.
24          Activation of the dopaminergic system may also result in an inability to inhibit locomotor
25   responses normally suppressed by punishment (Tackson and Westlind-Danielsson. 1994]. Direct
26   application of dopamine to the nucleus accumbens of rats has been observed to result in
27   retardation of the acquisition of passive avoidance learning at concentrations that also stimulated
28   locomotor activity [Braes etal.. 1984]. Increases in catecholaminergic neurotransmission (through
29   exposure to norepinephrine or dopamine agonists] result in dose-dependent reductions in the
30   duration of spike wave discharges in rats (Snead, 1995: Warter etal., 1988]. These observations
31   and findings are in concordance with those resulting from exposure to TMBs (Wiadernaetal.. 2002:
32   Gralewicz and Wiaderna. 2001: Tomas et al.. 1999a: Tomas et al.. 1999c: Gralewicz etal.. 1997b:
33   Gralewicz etal.. 1997a]. Additionally, with regards to toluene  and related aromatic hydrocarbons, it
34   is known that there is direct interaction with these compounds on various ion channels (ligand and
35   voltage gated] that are present in the central nervous system (Bowenetal., 2006: Balster, 1998].
36   There is not enough information to ascertain the specific molecular sites and how the changes

            This document is a draft for review purposes only and does not constitute Agency policy.

                                               1-22             DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   correlate to the observed neurotoxicological effects. However, it is widely believed that the
 2   interactions with the neuronal receptors in the brain (e.g., ion channels, catecholaminergic
 3   systems) may influence these changes.
 4          Aromatic hydrocarbons may also affect the phospholipids in the nerve cell membrane
 5   [Andersson et al., 1981]. Perturbation of the phospholipids on the cell membrane could indirectly
 6   affect the binding of neurotransmitters to the catecholamine or other receptors and potentially lead
 7   to alterations in receptor activity or uptake-release mechanisms. Uneven distribution of
 8   metabolites within differing regions of the brain, or spatial variations in phospholipid composition
 9   of nerve cell membranes, may explain the differential effects seen in regard to catecholamine levels
10   and turnover [Andersson etal.. 1981]. Based on effect levels with other related solvents (e.g.,
11   toluene - see Balster (1998]], it is hypothesized that with TMBs there may be an initial interaction
12   with the neuronal receptors (e.g., catecholaminergic systems, ion channels] followed by, at much
13   higher exposures, interaction with the lipid membrane when the available sites on the neuronal
14   receptors are completely occupied.
15          Additional mechanisms that may play a role in 1MB neurotoxicity include production of
16   reactive oxygen species (ROS]. Myhre et al. (2000] observed increased respiratory burst in
17   neutrophils after 1,2,4-TMB exposure demonstrated by fluorescence spectroscopy, hydroxylation of
18   4-hydroxybenzoic acid, and electron paramagnetic resonance spectroscopy. The authors suggest
19   that the observation of solvent-induced ROS production may be relevant to brain injury, as
20   microglia cells have a respiratory burst similar to neutrophils. Stronger evidence of potential ROS-
21   related mechanisms of neurotoxicity was observed in a related study by Myhre and Fonnum (2001]
22   in which rat neural synaptosomes exposed to 1,2,4-TMB produced a dose-dependent increase in
23   reactive oxygen and nitrogen species demonstrated by the formation of the fluorescence of 2'7'-
24   dichlorofluorescein. This observation of ROS production in rat synaptosomes may potentially
25   explain the observed TMB-induced neurotoxicity in acute, short-term, and subchronic inhalation
26   studies.

     1.1.1.2.  Summary of Neurological Effects
27          Neurotoxicity is associated with exposure to TMBs based on evidence in humans exposed to
28   mixtures containing TMBs and in animals exposed to individual TMB isomers. All three TMB
29   isomers are taken up in humans (Tarnbergetal., 1998,1997a:  Tarnbergetal., 1996], and
30   occupational studies involving exposure to TMBs and other VOCs show neuropsychological effects
31   (Chen etal., 1999], deficits in short term memory and reduced motor speed/coordination (Lee et
32   al.. 2005], abnormal fatigue (Norsethetal.. 1991], and nervousness, anxiety,  and/or vertigo [(Battig
33   etal.. 1956], as reviewed by MOE (2006] and (Battig etal.. 1958]]. These effects, however, cannot
34   be attributed to any specific compound. None of the available human studies have addressed the
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-23            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   potential for latent neurological effects and no studies examined the potential for neurological
 2   effects in sensitive populations.
 3          There is strong, consistent evidence of neurotoxicity in male Wistar rats exposed to any
 4   TMB isomer via inhalation across multiple concentrations and multiple durations; however, the
 5   studies were all conducted at the same institute [Wiadernaetal., 2002: Gralewicz and Wiaderna,
 6   2001: Wiadernaetal.. 1998: Gralewicz etal.. 1997b: Gralewicz etal.. 1997a: Korsakand Rydzynski.
 7   1996: Korsak etal.. 1995). By gavage, similar effects were observed (e.g., altered EEC recordings;
 8   increased locomotor activity in open field tests) [Tomas etal.. 1999a: Tomas et al.. 1999b). although
 9   testing by this route was not as extensive as by inhalation.
10          The spectrum of observed effects suggests that TMBs affect multiple, possibly overlapping,
11   CNS systems rather than a single brain region or neuronal nuclei (suggested by the solvent activity
12   of the compounds). Almost all tests (including pain sensitivity) involve a contributing component of
13   motor system function. It is notable that none of the identified studies on individual TMB isomers
14   employed protocols capable of distinguishing effects on motor activity alone (e.g., the majority of
15   studies used open field tests 5-10 minutes in duration); thus, it remains to be determined whether
16   TMBs exposure specifically affects motor system function. Some endpoints exhibited clear
17   exposure-response relationships, including measures of pain sensitivity and neuromuscular
18   function, when tested immediately after exposure. Most other endpoints did not show a clear
19   concentration-effect relationship, although the direction and magnitude of responses was relatively
20   consistent across studies. In most cases, effects at 1,230 mg/m3 were less robust than those
21   observed at lower TMB concentrations (i.e., responses were nonlinear). However, nonlinear
22   relationships are not uncommon for solvents and, as they were  observed across multiple studies
23   using each of the three isomers, they are considered to be biologically-relevant observations rather
24   than experimental artifacts. Latent neurological effects following TMBs exposure were consistently
25   observed, but were difficult to characterize as deficits in a single neurological function. For
26   example, latent measures of pain sensitivity following TMBs exposure, although consistent, were
27   only statistically significant when the rats were challenged with a foot shock on the prior day. The
28   most likely explanation for this observation is that TMBs exposure extends the duration of foot
29   shock-induced decreases in pain sensitivity, since the immediate response to foot shock was similar
30   across groups; yet, it cannot be ruled out that TMBs exposure could alter cognitive function,
31   resulting in the observed responses. In summary, the evidence supports a determination that TMBs
32   are neurotoxic following inhalation or oral exposure, based on consistency and coherency of effects
33   in animals and humans, biological plausibility, evidence of delayed-onset and/ or latent
34   neurological effects in animals several weeks following exposure, and observed exposure-response
35   relationships in animals tested immediately after exposure.
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-24            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

     1.1.2. Respiratory Effects
 1          There is evidence in humans and animals that inhalation exposure to TMBs induces
 2   respiratory toxicity. The human evidence comes from occupational and residential studies
 3   involving complex VOC mixtures that include TMBs; thus, effects cannot be attributed to any TMB
 4   isomer specifically. TMB isomers are associated with increased measures of respiratory irritation,
 5   such as laryngeal and/or pharyngeal irritation [Norseth etal.. 1991] and asthmatic bronchitis
 6   [[Battigetal.. 1956]. as reviewed in MOE [2006] and Baettig et al. [1958]] following occupational
 7   exposures. Residential exposures have demonstrated significant associations between 1,2,4-TMB
 8   and asthma [Billionnetetal., 2011]. Controlled human exposures [Tones etal., 2006: Tarnbergetal.,
 9   1997a: Tarnbergetal., 1996] have failed to observe substantial irritative symptoms following acute
10   (less than 4 hours] inhalation exposures to TMB isomers of up to 25 ppm (123 mg/m3]. For full
11   details of the epidemiologic and controlled human exposures studies (including human subjects
12   research ethics procedures], see individual study summary tables in Appendix B.
13          In animals, there is consistent evidence of respiratory toxicity following inhalation exposure
14   of rodents to the TMB isomers (Table 1-3; Figure 1-5]. Markers of inflammation and irritation in the
15   lungs of rats have been observed following subchronic inhalation exposures of Wistar rats to
16   1,2,4-TMB or 1,2,3-TMB. Increases in immune and inflammatory cells in bronchoalveolar lavage
17   (BAL] fluid have been observed following subchronic exposures of male Wistar rats to 1,2,4-TMB at
18   concentrations > 123 mg/m3 (Korsaketal.. 1997]. Specifically, the number of cells in the BAL fluid
19   of exposed rats was increased for both total cells (> 123 mg/m3] and macrophages (> 492 mg/m3].
20   However, some attenuation of these effects was observed at high concentrations (i.e., at 1,230
21   mg/m3] compared to lower concentrations. For example, the number of macrophages was
22   increased 2.7-fold relative to control at 492 mg/m3, but only 2.2-fold at 1,230 mg/m3. This may
23   indicate either adaptation to the respiratory irritation effects of 1,2,4-TMB, saturation of metabolic
24   pathways, or immune suppression at higher doses. Subchronic exposure of male Wistar rats also
25   significantly increased the BAL fluid content of polymorphonuclear leukocytes and lymphocytes;
26   however the specific concentrations eliciting these significant increases were not reported by study
27   authors. A small, but not significant, decrease in cell viability (all cells] was observed following
28   subchronic exposure to 1,2,4-TMB at > 123 mg/m3 (Korsaketal.. 1997].
29          In addition to increases in immune and inflammatory cells in BAL fluid following exposure
30   to 1,2,4-TMB, histopathological alterations characterized by increases in lymphatic tissue  in the
31   lower respiratory tract have also been observed following subchronic exposures of male and female
32   Wistar rats to 1,2,4-TMB or 1,2,3-TMB (Korsak etal.. 2000a. b]. Significant proliferation of
33   peribronchial lymphatic tissue was observed in male rats exposed to 123 mg/m31,2,3-TMB or 492
34   mg/m31,2,4-TMB and female rats exposed to 123 and 492 mg/m31,2,3-TMB, although trend
35   analysis demonstrated that these increases were not concentration-dependent. Non-concentration
36   dependent increases in interstitial lymphocytic infiltrations were also observed in male rats

           This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-25            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   exposed to 492 mg/m31,2,4-TMB. However, statistically significant increases in interstitial
 2   lymphocytic infiltrations observed in male and female rats exposed to 1,230 mg/m31,2,3-TMB or
 3   1,2,4-TMB, respectively, were concentration-dependent based on trend analysis.
 4          In some 1,2,4-TMB or 1,2,3-TMB-exposed rats exhibiting peribronchial lymphatic
 5   proliferation, the bronchial epithelium lost its cuboidal shape and formed lymphoepithelium.
 6   However, this formation of lymphoepithelium was apparently non-monotonic and not dependent
 7   on concentration. Alveolar macrophages were increased in both sexes exposed to 1,230 mg/m3
 8   1,2,4-TMB (significant only for males), with trend analysis demonstrating concentration-
 9   dependence across the entire concentration range. Goblet cells were statistically significantly
10   increased in a concentration-dependent manner in female rats exposed to > 492 mg/m31,2,3-TMB.
11   When the incidences of all pulmonary lesions were analyzed in aggregate, lesions were significantly
12   increased in males at 492 mg/m31,2,4-TMB, but not at any concentration in females. However,
13   trend-analysis demonstrated significant increases in aggregate pulmonary lesions in both sexes
14   across the entire concentration range. In rats exposed to 1,2,3-TMB, the aggregate incidences of
15   pulmonary lesions were not statistically significantly increased at any single concentration in males
16   or females. Male rats, however, did exhibit a concentration-dependent increase in aggregate lesions
17   according to trend analysis. Studies on the respiratory effects of subchronic exposures to 1,3,5-TMB
18   were not available.
19          Additional effects on clinical chemistry including increased total protein (37% increase at
20   exposures of both 123 and 492 mg/m3), decreased mucoprotein (13% decrease at 123 mg/m3
21   exposure), increased lactate dehydrogenase (170%  and 79% increase at 123 and 492 mg/m3,
22   respectively) and increased acid phosphatase activity (47-75% increase at > 123 mg/m3) were
23   observed in animals exposed to 1,2,4-TMB, suggesting pulmonary irritation or inflammation. All of
24   these effects also exhibited either some attenuation of effect at high concentrations compared to
25   lower concentrations. Therefore, some adaptation to the respiratory irritation effects of 1,2,4-TMB
26   may be occurring.
27          Decreased respiration, a symptom of sensory irritation, has been observed in male BALB/C
28   mice during acute inhalation exposures to the TMB isomers for 6 minutes. These acute exposures
29   were observed to result in dose-dependent depression of respiratory rates, with the maximum
30   decrease in respiration occurring in the first 1 or 2 minutes of exposure (Korsaketal.. 1997: Korsak
31   etal.. 1995). The concentration of 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB that was observed to result
32   in a 50% depression in the respiratory rate (RDso) was similar between the three isomers: 578, 541,
33   or 519 ppm (2,844, 2,662, or 2,553 mg/m3), respectively.
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-26            DRAFT—DO NOT CITE OR QUOTE

-------
                                              Toxicological Review of Trimethylbenzene
Table 1-3. Evidence pertaining to respiratory effects of TMBs in animals —
           inhalation exposures
Study design3 and reference
Results
1,2,4-TMB
Pulmonary inflammation/irritation
0, 123, 492, 1,230 mg/m3, 90 days
(6 hr/day, 5 days/week)
Rat, Wistar, male, N = 6-7
Korsak et al. (1997), Table B-31
0, 123, 492, 1,230 mg/m3, 90 days
(6 hr/day, 5 days/week)
Rat, Wistar, male, N = 6-7
Korsak et al. (1997), Table B-31
0, 123, 492, 1,230 mg/m3, 90 days
(6 hr/day, 5 days/week)
Rat, Wistar, male and female, N = 10
Korsak et al. (2000a), Table B-32
Increased total bronchoalveolar cell count with evidence of attenuation
at high exposure.
Response relative to control: 0, 202***, 208**, 131*%
Increased macrophage count with evidence of attenuation at high
exposure.
Response relative to control: 0, 107, 170**, 116**%
Increase in number of pulmonary lesions.
Response relative to control: Incidences not reported, thus calculation of
response relative to control not possible; authors report statistically
significant increases at 492 and 1,230 mg/m3.
Clinical chemistry effect
0, 123, 492, 1,230 mg/m3, 90 days
(6 hr/day, 5 days/week)
Rat, Wistar, male, N = 10
Korsak et al. (1997), Table B-31
Increased acid phosphatase activity with evidence of attenuation at
exposure.
Response relative to control: 0, 47*, 74*, 45*%
high
Sensory irritation (decreased respiration)
1,245, 3,178, 5,186, 6,391, 9,486 mg/m3,
6 min
Mouse, BALB/C, male, N = 8-10
Korsak et al. (1997); Korsak et al. (1995),
Tables B-31 and B-29
Decreased respiratory rate as measured during first minute of exposure.
Response relative to control: RD50 = 2,844
        This document is a draft for review purposes only and does not constitute Agency policy.

                                          1-27
                                                       DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
Table 1-3 (Continued): Evidence pertaining to respiratory effects of TMBs in animals
                          — inhalation exposures
Study design3 and reference
Results
1,2,3-TMB
Pulmonary inflammation/irritation
0, 123, 492, 1,230 mg/m3, 90 days
(6 hr/day, 5 days/week)
Rat, Wistar, male and female, N = 10
Korsak et al. (2000b), Table B-33
Increase in number of pulmonary lesions.
Response relative to control: Incidences not reported, thus calculation
of response relative to control not possible; authors report statistically
significant increases at 492 and 1,230 mg/m3.
Sensory irritation (decreased respiration)
1,255, 2,514, 4,143, 7,828 mg/m3, 6 min
Mouse, BALB/C, male, N = 8-10
Korsak et al. (1997); Tables B-31
Decreased respiratory rate as measured during first minute
exposure.
Response relative to control: RD50 = 2,662
of
1,3,5-TMB
Sensory irritation (decreased respiration)
1,348, 2,160, 2,716, 3,597, 4,900 mg/m3,
6 min
Mouse, BALB/C, male, N = 8-10
Korsak et al. (1997), Table B-31
Decreased respiratory rate as measured during first minute
exposure.
Response relative to control: RD50 = 2,553
of
*, **, *** Statistically different from controls at p < 0.05, p < 0.01, and p < 0.001, respectively.
aln instances where authors reported exposures in ppm, EPA converted these values to mg/m3. See Appendix B (Table B-l) for
  conversion factor, and individual study summary tables for ppm values.
 Tables referenced in Study Design and Reference column correspond to study summary tables in Appendix B
       This document is a draft for review purposes only and does not constitute Agency policy.

                                              1-28             DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
1,2,4-TMB, 1,2,3-TMB,
Isomer Effect Summary
Pulmonary
1,2,4-TMB Inflammation/
Irritation
Clinical
Chemistry
Effects
Sensory
Irritation
Pulmonary
1,2,3-TMB Inflammation/
Irritation
Sensory
Irritation
1,3,5-TMB f^
Irritation
or 1,3,5-TMB
i-
Duration Endpoint
^ BAL Total
Subchronic
Cell Count (a)
1" BALMacro-
phage Count (a)
't' Pulmonary
Inflammatory
Lesions (b)
I"1 Acid
Subchronic Phosphatase
Activity (a)
d. Respiratory
Acute , i
Rate (a, c)
T1 Pulmonary
Subchronic Inflammatory
Lesions (d)
>!• Respiratory
Acute ' ,. , ,
Rate (a)
4/ Respiratory
Acute , ,
Rate (a)
Low Dose H High Dose • NOAEL • LOAEL A RD50


• i
• i





LOO 1000 1000(
Concentration (mg/m3)
Note: Solid lines represent range of exposure concentrations, (a) Korsak et al. [1997]: (b) Korsak et al. [2000a]: (c) Korsak
et al. [1995]: [d] Korsak [2000b]. Y-axis is displayed on a logarithmic scale. All subchronic effects are in male Wistar rats,
except for increased pulmonary lesions, which occur in both male and female Wistar rats; acute effects are in BALB/c
mice.

Figure 1-5. Exposure response array of respiratory effects following inhalation
            exposure to 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB.


       This document is a draft for review purposes only and does not constitute Agency policy.


                                             1-29             DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene
     1.1.2.1.  Mode of Action Analysis - Respiratory Effects
 1          Data regarding the potential mode of action for the respiratory effects resulting from TMB
 2   inhalation exposures are limited and the key events for TMB-induced respiratory toxicity are not
 3   established. However, the available toxicity data suggest that TMB isomers act as potent acute
 4   respiratory irritants and induce inflammatory responses following longer exposures (i.e.,
 5   subchronic) in animals. Korsak et al. [1995] and Korsak et al. [1997] have suggested that decreased
 6   respiratory rate following TMB inhalation exposure is indicative of irritation, and proposed that
 7   respiratory irritants such as TMB may activate a "sensory irritant receptor" on the trigeminal nerve
 8   ending in the nasal mucosa leading to an inflammatory response. Korsak et al. [1997: 1995] further
 9   suggested that activation of this irritant receptor follows either adsorption of the agonist, or
10   adsorption and chemical reaction with the receptor. The authors referenced a proposed model for
11   the receptor protein that includes two main binding sites for benzene moieties and a thiol group.
12   Further, they suggested that in the case of organic solvents (i.e., toluene, xylene, and TMB], a
13   correlation between the potency of the irritating effect and the number of methyl groups is likely
14   given the observation that RDso values for depressed respiratory rates following exposure to TMB
15   isomers is approximately 8-fold lower than toluene and 4-fold lower than xylene.
16          Following subchronic inhalation exposure of rats to 1,2,4-TMB, inflammatory cell (i.e.,
17   macrophages, polymorphonuclear leukocytes, and lymphocytes] numbers were increased along
18   with markers of their activation (i.e., total lactate dehydrogenase and acid phosphatase activity in
19   BAL] (Korsak et al.. 1997]. further indicating the inflammatory nature of responses in the
20   respiratory tract of TMB-exposed animals. Inflammatory pulmonary lesions were also observed
21   following subchronic inhalation exposures in rats. However, many of these effects were not
22   observed to be concentration-dependent in repeat exposure studies (i.e., no progression of effect
23   over an order of magnitude of concentrations], suggesting that there may be adaptation to
24   respiratory irritation that occurs following extended inhalation exposure to TMB. The processes
25   responsible for the respiratory inflammatory responses observed in subchronically exposed
26   animals are unknown. However,  a major inflammatory mediator, interleukin 8 (IL-8], was
27   increased following exposure of porcine and human macrophages to secondary organic aerosol
28   (SOA] particles derived from 1,3,5-TMB (Gaschenetal.. 2010]. The observation that IL-8 levels
29   increase  following exposure to 1,3,5-TMB-derived SOA is noteworthy as a major function of IL-8 is
30   to recruit immune cells to sites of inflammation. Therefore, the observation of inflammatory lesions
31   involving immune cells (i.e., macrophages and leukocytes] may be partially explained by increases
32   in inflammatory cytokines following TMB exposures. Additionally, ROS-generation has been
33   observed in cultured neutrophil granulocytes and rat neural synaptosomes exposed to TMB (Myhre
34   andFonnum, 2001: Myhre etal.,  2000], and the related compounds benzene and toluene have been

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-30             DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene

 1   shown to induce oxidative stress in cultured lung cells [Mogeletal.. 2011). Although pulmonary
 2   ROS-generation has not been observed following in vivo or in vitro 1MB exposures, there is
 3   suggestive evidence that it could play a role in the irritative and inflammatory responses seen in
 4   exposed animals.
 5          In a study investigating jet fuel-induced cytotoxicity in human epidermal keratinocytes
 6   (HEK), aromatic hydrocarbons were more potent inducers of cell death than aliphatic constituents,
 7   even though the aromatic compounds only accounted for less than one-fourth of aliphatic
 8   constituents [Chou etal.. 2003). Of the single aromatic ring hydrocarbons, 1,2,4-TMB and xylene
 9   were the most lethal to HEK. Increased cytotoxicity may explain the small, but insignificant,
10   decrease in BAL cell viability observed in Korsak et al. [1997].

     1.1.2.2.  Summary of Respiratory Effects
11          Respiratory toxicity is associated with inhalation exposure to TMBs based on evidence in
12   humans and animals. All three 1MB isomers are taken up by humans [Tarnbergetal.. 1998.1997a:
13   Jarnbergetal.. 1996]. and occupational and residential studies involving exposure to TMBs and
14   other VOCs suggest an association between 1MB exposure and asthmatic symptoms [Billionnet et
15   al., 2011: Battigetal., 1956] and sensory irritation [Norseth etal., 1991]. These effects, however,
16   cannot be attributed to any specific compound.
17          There is strong, consistent evidence of respiratory toxicity in male and female Wistar rats
18   exposed to any TMB isomer via inhalation across multiple concentrations and multiple durations,
19   although the studies were conducted at the same institute [Korsak etal.. 2000a. b; Korsak etal..
20   1997: Korsak etal.. 1995]. Some endpoints (i.e., BAL macrophages and alkaline phosphatase]
21   showed concentration-dependence at low- and mid-exposures, all effects were observed to exhibit
22   some attenuation of effect at high doses, potentially indicating either adaptation to the respiratory
23   irritation effects, saturation of metabolic and/or toxicity pathways, or immune suppression at
24   higher doses. In summary, the evidence supports a determination that TMBs are respiratory
25   toxicants following inhalation exposure, based on consistency and coherency of effects observed in
26   humans and animals, biological plausibility, and observed exposure-response relationships.

     1.1.3. Reproductive and Developmental Effects
27          There are no studies in humans that investigated the reproductive or maternal toxicity of
28   the TMB isomers by any route of exposure. Maternal toxicity in the form of decreased corrected
29   body weight (i.e., maternal body weight minus the weight of the gravid uterus] was observed in
30   Sprague-Dawley rat dams following inhalation exposure during gestation to 1,2,4-TMB or
31   1,3,5-TMB [Saillenfaitetal..20Q5] (Table 1-4; Figure 1-6]. Dams exposed to 2,952 mg/m3
32   1,2,4-TMB gained only 50% of the weight gained by control animals, whereas dams exposed to
33   2,952 mg/m31,3,5-TMB gained only 25% of the weight gained by controls. Decreased maternal

            This document is a draft for review purposes only and does not constitute Agency policy.

                                               1-31            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene

 1    food consumption (across GD6-GD21) was also observed at > 2,952 mg/m31,2,4-TMB and > 1,476
 2    mg/m31,3,5-TMB, although the magnitude of the difference compared to controls (88-83% and 92-
 3    75% of controls, respectively) was modest relative to the observed decreases in maternal weight
 4    gain. The decrease in food consumption at 1,476 mg/m31,3,5-TMB (92% relative to controls) was
 5    not considered to be a marker of adversity given no accompanying decrease in maternal weight
 6    gain was observed at that concentration.
 7           There are no studies in humans that investigated the developmental toxicity of either
 8    1,2,4-TMB or 1,3,5-TMB by any route of exposure. Developmental toxicity (reported as decreased
 9    fetal body weight) has been observed in male and female rats following gestational exposure to
10    1,2,4-TMB and 1,3,5-TMB on gestational days 6 through 20 via inhalation for 6 hours a day
11    (Saillenfaitetal., 2005) (Table  1-4). Fetal body weights were decreased (statistically significantly)
12    by 5-13% at concentrations of > 2,952 mg/m3 of 1,2,4-TMB and 1,3,5-TMB. No adverse effects were
13    noted on embryo/fetal viability and no increase in skeletal, visceral, or external morphology (i.e.,
14    teratogenesis) was observed up to the highest concentrations for either isomer. Studies on the
15    developmental or reproductive effects of 1,2,3-TMB by any route of exposure were not available.
            This document is a draft for review purposes only and does not constitute Agency policy.

                                               1-32            DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
Table 1-4. Evidence pertaining to reproductive and developmental effects of
             1,2,4-TMB and 1,3,5-TMB in animals — inhalation exposures
Study Design3 and Reference
1,2,4-TMB
Developmental toxicity
0, 492, 1,476, 2,952, 4,428 mg/m3,
GD6-GD20 (6 hr/day)
Rat, Sprague-Dawley, female and male0
Saillenfait et al. (2005), Table B-38
Maternal toxicity
0, 492, 1,476, 2,952, 4,428 mg/m3,
GD6-GD20 (6 hr/day)
Rat, Sprague-Dawley, female, N = 24-25
dams
Saillenfait et al. (2005), Table B-38
1,3,5-TMB
Developmental toxicity
0, 492, 1,476, 2,952, 5,904 mg/m3,
GD6-GD20 (6 hr/day)
Rat, Sprague-Dawley, female and malea'c
Saillenfait et al. (2005), Table B-38
Maternal Toxicity
0, 492, 1,476, 2,952, 5,904 mg/m3,
GD6-GD20 (6 hr/day)
Rat, Sprague-Dawley, female, N = 24-25
dams
Saillenfait et al. (2005), Table B-38
Results


Decreased fetal body weight of male and female fetuses.
Response relative to control:
Male:0, -l,-2, -5*, -11**%
Female: 0, -1, -3, -5*, -12**%

Decreased corrected maternal weight gain.
Response relative to control: 0, +7, -7, -51**, -100**%
(weight gain = 0g)


Decreased fetal body weight of male and female.
Response relative to control:
Male:0, -1, -5, -7*, -12**%
Female: 0, -1, -4, -6, -13**%

Decreased corrected maternal weight gain.
Response relative to control: 0, +3, -31,- 76**, -159**%
weight gain = -12 g)
*, ** Statistically significantly different from controls at p < 0.05 and p < 0.01, respectively.
aln instances where authors reported exposures in ppm, EPA converted these values to mg/m3. See Appendix B (Table B-l) for
  conversion factor, and individual study summary tables for ppm values.
 Tables referenced in Study Design and Reference column correspond to study summary tables in Appendix B.
cNumber of fetuses analyzed not reported.
       This document is a draft for review purposes only and does not constitute Agency policy.

                                             1-33             DRAFT—DO NOT CITE OR QUOTE

-------
                                                  Toxicological Review of Trimethylbenzene
1,2,4-TMB, or 1,3,5-TMB
Isomer Effect Summary Endpoint Sex
1,2,4-TMB D^'°Pmental ^Fetal Male;
Toxicity Body Weight
Females
. \1/ Maternal
Maternal
Body Weight Female
T™^ Gain
1,3,5-TMB D^P™^1 4, Fetal ^
Toxicity Body Weight
Females
4/ Maternal
Maternal '...,,,,-.,
Body Weight Female
Toxicity
Gain
i- Low Dose -i High Dose • NOAEL • LOAEL
I I
i I
i i
i i
i i
i i
I • I
I * \


                                                      1,000    2,000   3,000   4,000   5,000   6,0011
                                                             Concentration (mg/m3)
Note: Solid lines represent range of exposure concentrations. All effects from Saillenfait et al. (2005).

Figure 1-6. Exposure response array of reproductive and developmental effects
            following inhalation exposure to 1,2,4-TMB or 1,3,5-TMB.
      This document is a draft for review purposes only and does not constitute Agency policy.

                                          1-34            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene

     1.1.3.1.  Summary of Reproductive and Developmental Effects
 1          The database for reproductive and developmental toxicity following inhalation exposure to
 2   1,2,4-TMB and 1,3,5-TMB is limited to one animal developmental study; no studies in humans are
 3   available. Thus, these isomers may cause developmental toxicity, although this is based on only one
 4   study that demonstrated clear, exposure-related effects on fetal and maternal body weights.

     1.1.4. Hematological and Clinical Chemistry Effects
 5          There is limited evidence in humans, and stronger evidence in animals, that exposure to
 6   TMB isomers via inhalation induces hematological toxicity and alterations in clinical chemistry
 7   parameters. Alterations in blood clotting and anemia in workers exposed to a paint solvent
 8   containing 50% 1,2,4-TMB, 30% 1,3,5-TMB, and unspecified amounts of 1,2,3-TMB (listed as
 9   possibly present) was reported by Battig et al. [1956]. as reviewed by MOE [2006]: effects observed
10   at 295 mg/m3. However, as workers were exposed to  a solvent mixture containing multiple TMB
11   isomers and other VOCs, effects cannot be attributed to any TMB isomer specifically.
12          In animals, there is evidence of hematological toxicity following subchronic inhalation
13   exposure to 1,2,4-TMB or 1,2,3-TMB and short-term inhalation exposure to 1,3,5-TMB  (Table 1-5;
14   Figures 1-7 and 1-8). Subchronic exposures to 1,2,4-TMB or 1,2,3-TMB have been shown to result
15   in hematological effects and changes in serum chemistry in rats [Korsak et al.. 2000a. b). In male
16   rats exposed to 1,230 mg/m31,2,4-TMB or 1,2,3-TMB, red blood cells (RBC) counts were
17   significantly decreased 23 and 15%, respectively. The observed alterations in RBCs were
18   concentration-dependent as determined by trend analysis. Exposure to  1,2,4-TMB or 1,2,3-TMB did
19   not significantly decrease RBCs in female rats, but trend analysis demonstrated that decreases in
20   RBC counts in female rats exposed to  1,2,3-TMB were concentration dependent, with a maximum
21   decrease of 9% at 1,230 mg/m3. RBCs in both sexes were observed to still be depressed relative to
22   controls 2 weeks following termination of exposure to both isomers, but these decreases were not
23   statistically significant.
24          White blood cell (WBC) counts were significantly increased 80% in male rats and increased
25   30% (not statistically significant) in female rats exposed to  1,230 mg/m31,2,4-TMB. After a two-
26   week follow-up after termination of exposure, WBC counts had returned to normal in female rats
27   and were slightly depressed (18%) in male rats. WBC numbers were unchanged in male rats
28   exposed to 1,2,3-TMB, but were increased (not statistically significant) 22% in female rats exposed
29   to 1,230 mg/m3. After two weeks following termination of exposure, WBC counts in male and
30   female rats had fallen to roughly 60% of controls.
31          Significant decreases in reticulocytes (71% decrease relative to controls) and clotting time
32   (37% decrease relative to controls) were observed in female rats exposed to 1,230 mg/m3 and 492
33   mg/m31,2,4-TMB, respectively. Both  of these effects were concentration-dependent across the
34   entire-range of concentrations as determined by trend-analysis; animals fully recovered within 2

            This document is a draft for review purposes only and does not constitute Agency policy.

                                               1-35            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   weeks after termination of exposure. Reticulocyte numbers were statistically significantly
 2   increased 60% in male rats exposed to 1,230 mg/m31,2,3-TMB, with reticulocyte numbers even
 3   further increased (150%) two weeks following the termination of exposure. Reticulocyte numbers
 4   in females exposed to 1,2,3-TMB were significantly increased 77% and 100% at 123 and 492
 5   mg/m3, and increased 69% (not statistically significant) at 1,230 mg/m3. Reticulocyte numbers
 6   were still increased in males and females 2 weeks after the termination of exposure to 1,2,3-TMB.
 7   Segmented neutrophils were statistically significantly decreased 29% in male rats exposed to 1,230
 8   mg/m31,2,3-TMB; statistically significant decreases of 29% and 48% were observed in female rats
 9   exposed to 492 and 1,230 mg/m31,2,3-TMB. Lymphocytes were statistically increased 11% and
10   15% in male and female rats exposed to 1,230 mg/m3, respectively. Numbers of segmented
11   neutrophils and lymphocytes returned to control values 2 weeks after termination of exposure.
12          Exposure to TMB isomers was also observed to have an effect on clinical chemistry markers
13   that possibly indicate hepatic injury. Sorbitol dehydrogenase was increased at > 123 mg/m3 in male
14   rats exposed to 1,2,4-TMB (18-23% relative to controls) and at 1,230 mg/m3 in male rats exposed
15   to 1,2,3-TMB (69% relative to controls)(Korsaketal.. 2000a. b). However, the increases following
16   exposure to 1,2,4-TMB were not concentration-dependent Sorbitol dehydrogenase activity was
17   also higher in female rats exposed to 1,2,4-TMB (19-23% relative to controls) but the increases in
18   activity were not significantly higher when compared to controls. Sorbitol dehydrogenase activity
19   was not affected in female rats exposed to 1,2,3-TMB. Alanine aminotransferase was decreased
20   (23% relative to controls) and alkaline phosphatase was increased (42-45% relative  to controls) at
21   1,230 mg/m3 and > 492 mg/m3 (respectively) in female rats exposed to 1,2,3-TMB. Absolute iver
22   weights were only observed to increase (9%) in male rats exposed to 1,230 mg/m31,2,3-TMB, and
23   no histopathological changes were observed in either sex exposed to  1,2,3-TMB or  1,2,4-TMB.
24   Therefore, the adversity of the observed changes in clinical chemistry parameters is unclear.
25          An increase (30% relative to controls) in aspartate aminotransferase, but no other
26   substantial hematological effects, was observed in rats 14 days following short-term exposure (6
27   hours/day, 6 days/week for 5 weeks) (Wiglusz etal.. 1975b: Wiglusz etal.. 1975a). The adversity of
28   aspartate aminotransferase is uncertain given the lack of a clear pattern in temporality (effects at
29   some days post-exposure, but not others) and the lack of accompanying liver histopathology.
30          Acute inhalation exposures of male Wistar rats to 1,500-6,000 mg/m31,3,5-TMB for 6
31   hours did not result in substantial effects on hemoglobin or RBC or WBC count (Wiglusz etal.,
32   1975b). However, the number of segmented neutrophilic granulocytes was increased in 1,3,5-TMB-
33   exposed rats up to 28 days following exposure (statistics not reported). The greatest increase in
34   granulocyte numbers (100%) was observed the day of exposure and  1 day following in rats
35   exposed to 6,000 mg/m3, although attenuation was seen 7-28 days following exposure, possibly
36   indicating induction of metabolizing enzymes or saturation of toxicity pathways. Investigation of
37   clinical chemistry parameters in rats acutely exposed to 300-3,000 mg/m3 for 6 hours did not

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-36            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1    reveal any consistent pattern in the levels of aspartate or alanine aminotransferases, although
 2    alkaline phosphatase was statistically increased 84% in rats 7 days following exposure to 3,000
 3    mg/m3 [Wigluszetal.. 1975a).
 4           Slight alterations in clinical chemistry parameters and differential white blood cell counts
 5    were also observed in rats following subchronic, oral exposure to 1,3,5-TMB (Table 1-6; Figure 1-9)
 6    [Koch Industries. 1995b). While no hematological parameters (i.e., RBC counts, hematocrit) were
 7    observed to differ between exposed rats and controls, the number of monocytes were observed to
 8    increase (100-200% increase) in male rats exposed to > 200 mg/kg-day 1,3,5-TMB. Additionally, a
 9    number of clinical chemistry parameters were altered in exposed rats. In female rats exposed to
10    600 mg/kg-day, sodium and chloride levels were statistically significantly decreased (2.3 and 2.7%,
11    respectively) relative to controls, and cholesterol and phosphorus were statistically significantly
12    increased (41% and 23%, respectively). In male rats, exposure to 600 mg/kg-day resulted in a
13    significant decrease (19%) in glucose levels, and significant increases in phosphorus levels and
14    alkaline phosphatase activity (17% and 46%, respectively). In a related, preliminary study (Koch
15    Industries.  1995a). hematological and clinical chemistry effects were also observed following 14
16    days of oral exposure. Female Sprague Dawley rats exposed to either 150 or 600 mg/kg-day
17    1,3,5-TMB had increased cholesterol levels, and high-dose males exhibited increased white blood
18    cell counts with corresponding increased neutrophil and lymphocyte numbers.
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-37            DRAFT—DO NOT CITE OR QUOTE

-------
                                              Toxicological Review of Trimethylbenzene
Table 1-5. Evidence pertaining to hematological and clinical chemistry effects of
           1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB in animals — inhalation exposures
Study Design3 and Reference
Results
1,2,4-TMB
Hematological toxicity
0, 123, 492, 1,230 mg/m3, 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000a), Table B-32
0, 123, 492, 1,230 mg/m3, 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000a), Table B-32
0, 123, 492, 1,230 mg/m3, 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000a), Table B-32
0, 123, 492, 1,230 mg/m3, 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000a), Table B-32
Decreased red blood cells in males only.
Response relative to control: 0, -1, -15, -23**%
(recovery = 24% decrease)
Increased white blood cells in males only.
Response relative to control: 0, 2, 4, 80**%
(recovery = 18% decrease)
Decreased reticulocytes in females only.
Response relative to control: 0, -51, -49, -71*%
(recovery = 65% increase)
Decreases in clotting time in females only.
Response relative to control: 0, -23, -37**, -27*%
(recovery = 60% increase)
Clinical chemistry effects
0, 123, 492, 1,230 mg/m3, 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000a), Table B-32
Non-monotonic increases in sorbitol dehydrogenase in males
only.
Response relative to control: 0, 73**, 74*,73**%
        This document is a draft for review purposes only and does not constitute Agency policy.

                                         1-38
                                                      DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene
Table 1-5 (Continued): Evidence pertaining to hematological and clinical chemistry
                           effects of 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB in animals —
  	inhalation exposures	
Study Design3 and Reference
Results
1,2,3-TMB
Hematological toxicity
0,123, 492,1,230 mg/m3, 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000b), Table B-33
Decreased red blood cells in males only.
Response relative to control: 0, 8, 6, -15*%
(recovery = 9% decrease)
0,123, 492,1,230 mg/m3, 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000b), Table B-33
Decreased segmented neutrophils in males and females.
Response relative to control:
Males: 0, 2, -17, -29*% (recovery = 11% increase)
Females: 0, -15, -29*, -48*% (recovery = 15% decrease)
0,123, 492,1,230 mg/m3, 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000b), Table B-33
Increased lymphocytes in males and females.
Response relative to control:
Males: 0,1, 6,11**% (recovery = 11% decrease)
Females: 0, 6,10,15**% (recovery = 3% increase)
0,123, 492,1,230 mg/m3, 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000b), Table B-33
Increased reticulocytes in males and females (non-monotonic).
Response relative to control:
Males: 0, -25, 36, 61*% (recovery = 146**% increase)
Females: 0, 77*,  100**, 69% (recovery = 162**% increase)
Clinical chemistry effects
0,123, 492,1,230 mg/m , 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000b), Table B-33
Decreased alanine aminotransferase in females only.
Response relative to control: 0, -1, -6, -23*%
0,123, 492,1,230 mg/m3, 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000b), Table B-33
Increased alkaline phosphatase in females only.
Response relative to control: 0, 20,45*, 42*%
0,123, 492,1,230 mg/m3, 90 days (6 hr/day,
5 days/week)
Rat, Wistar, female and male, N = 10
Korsak et al. (2000b), Table B-33
Increased sorbitol dehydrogenase in males only.
Response relative to control: 0,44, 56, 69*%
       This document is a draft for review purposes only and does not constitute Agency policy.

                                               1-39             DRAFT—DO NOT CITE OR QUOTE

-------
                                                   Toxicological Review of Trimethylbenzene

Table 1-5 (Continued): Evidence pertaining to hematological and clinical chemistry
                         effects of 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB in animals —
                         inhalation exposures
Study Design3 and Reference
Results
1,3,5-TMB
Hematological toxicity
1,500, 3,000, 6,000 mg/m3, 6 hr
Samples collected 0, 1, 7, 14, and 28 days post
exposure
Rat, Wistar, male, N = 5.8
Wiglusz et al. (1975b), Table B-44
Increased segmented neutrophilic granulocytes (1-28 days
post-exposure).
Response relative to control:
Day 0: 0, 59, 118, 95%
Day 1: control response not reported
Day 7: control response not reported
Day 14: 0, 15, 184, 94%
Day 28: 0, -20, 124, 1%
Clinical chemistry effects
3,000 mg/m3, 5 weeks (6 hr/day, 6 days/week)
Samples collected 1, 3, 7, 14, and 28 days during
exposure
Rat, Wistar, male, N = 6
Wiglusz et al. (1975a), Table B-45
300-3,000 mg/m3, 6 hr, Samples collected 0, 2, 7, 14
and 28 days post exposure
Rat, Wistar, male, N = 6
Wiglusz et al. (1975a), Table B-45
Increased aspartate aminotransferase on day 14.
Response relative to control (day 14): 12*%
Increased alkaline phosphatase on day 7 post-exposure.
Response relative to control (on day 7 :0, -0.1, 0.03, 84*%
*, ** Statistically different from controls at p < 0.05 and p < 0.01, respectively.
aln instances where authors reported exposures in ppm, EPA converted these values to mg/m3. See Appendix B
  (Table B-l) for conversion factor, and individual study summary tables for ppm values.
bTables referenced in Study Design and Reference column correspond to study summary tables in Appendix B.
       This document is a draft for review purposes only and does not constitute Agency policy.

                                           1-40             DRAFT—DO NOT CITE OR QUOTE

-------
                                                  Toxicological Review of Trimethylbenzene
Table 1-6. Evidence pertaining to hematological and clinical chemistry effects of
            1,3,5-TMB in animals — oral exposures
Study Design and Reference
Results
1,3,5-TMB
Hematological toxicity
0, 50, 200, 600 mg/kg-day, 90 days (once
5 days/week)
Rat, Sprague-Dawley, female and male, N
Koch Industries (1995b), Table B-28a
daily,
= 10
Increased monocyte levels in males only
Response relative to control:
Male: 0, 100, 200*, 100*% (recovery = 100% increase)
Clinical chemistry effects
0, 50, 200, 600 mg/kg-day, 90 days (once
5 days/week)
Rat, Sprague-Dawley, female and male, IN
Koch Industries (1995b), Table B-28
0, 50, 200, 600 mg/kg-day, 90 days (once
5 days/week)
Rat, Sprague-Dawley, female and male, IN
Koch Industries (1995b), Table B-28
0, 50, 200, 600 mg/kg-day, 90 days (once
5 days/week)
Rat, Sprague-Dawley, female and male, N
Koch Industries (1995b), Table B-28
0, 50, 200, 600 mg/kg-day, 90 days (once
5 days/week)
Rat, Sprague-Dawley, female and male, IN
Koch Industries (1995b), Table B-28
0, 50, 200, 600 mg/kg-day, 90 days (once
5 days/week)
Rat, Sprague-Dawley, female and male, N
Koch Industries (1995b) Table B-28
0, 50, 200, 600 mg/kg-day, 90 days (once
5 days/week)
Rat, Sprague-Dawley, female and male, IN
Koch Industries (1995b), Table B-28
daily,
= 10
daily,
= 10
daily,
= 10
daily,
= 10
daily,
= 10
daily,
= 10
Increased phosphorus levels in males and females
Response relative to control:
Male1 038 17*% (recovery ~~ 11% decrease)
Female: 0, 0, 5, 23*% (recovery = 13% decrease)
Decreased sodium levels in females only
Response relative to control: 0, 0, 0, -2*%
(recover = 1% decrease)
Decreased chloride levels in females only
Response relative to control: 0, 0, 0, -3*%
(recovery = 1% increase)
Increased cholesterol levels in females only
Response relative to control: 0, -3, 7, 41*%
(recovery = 21% decrease)
Decreased glucose levels in males only
Response relative to control: 0, -10, -9, -19*%
(recovery = 12% increase)
Increased alkaline phosphatase activity in males only
Response relative to control: 0, 5, 13, 46*%
(recovery = 28% decrease)
*, ** Statistically different from controls at p < 0.05 and p < 0.01, respectively.
a Tables referenced in Study Design and Reference column correspond to study summary tables in Appendix B.
       This document is a draft for review purposes only and does not constitute Agency policy.

                                           1-41            DRAFT—DO NOT CITE OR QUOTE

-------
                                                            Toxicological Review of Trimethylbenzene
                          1,2,4-TMB and 1,3,5-TMB

         Isomer  Effect Summary    Duration     Endpoint       Sex
        1,2,4-TMB
        1,3,5-TMB
Hematological
  Toxidty
                   Clinical
                   Chemistry
                   Effects
Hematological
  Toxidty
                   Clinical
                   Chemistry
                   Effects
                               Sub-chronic
          4/Red Blood
           Cells (a)
                                          f White Blood
                                            Cells (a)
                                          •I- Reticulocytes
                                           4- Clotting
                                            Time (a)
                                                         Males
                                       Males
                                                        Females
                                       Females
                       ^ Sorbitol Dehydro-
              'jub-:hn:inic           '       Males
                          genase(a)
         1" Segmented
Acute       Neutrophilic       Males
         granulocytes(b)
                       't'AspartateAmino-
              Short-term   transferaseonday     Males
                       11 post-exposure (c)
                                           t Alkaline
                                 Acute     Phosphataseonday     Males
                                         7 post-exposure (c)
                                                      i- Low Dose H High Dose • NOAEL • LOAEL
                                                                100                1000
                                                                           Concentration (mg/m3)
                                                                                                  10001
Note: Solid lines represent range of exposure concentrations, (a) Korsak et al. [2000a]: (b) Wiglusz et al. (1975b); (c)
Wiglusz et al. [I975a]. Y-axis is displayed on a logarithmic scale.

Figure 1-7. Exposure response array of hematological and clinical chemistry effects
              following inhalation exposure to 1,2,4-TMB or 1,3,5-TMB.
        This document is a draft for review purposes only and does not constitute Agency policy.

                                                   1-42               DRAFT—DO NOT CITE OR QUOTE

-------
                                                 Toxicological Review of Trimethylbenzene
Effect Summary
Hematological
Toxicity
Clinical
Chemistry
Effects
1,2,3-TMB
Duration Endpoint Sex
4/ Red Blood
Subchromc .. , , Males
Cells (a)
4/ Segmented
Neutrophils (a)
\1/ Segmented
Females
Neutrophils (a)
t Lymphocytes ^
't' Lymphocytes
' Females
*}'• Reticulocytes
(a) MalS!
f- Reticulocytes
, , Females
(a)
\1/ AlanineAmino-
Subchronic Females
transferase (a)
t Alkaline
, , Females
Phosphatase(a)
t Sorbitol
. . . . . Male;
dehydrogenase(a)
i- Low Dose H High Dose • NOAEL • LOAEL
1
I
• i
I
i
1 A
• 1
1
1
«|
I

                                                          250    500    750    1,000    1,250
                                                              Concentration (mg/mD)
Note: Solid lines represent range of exposure concentrations, (a) Korsaketal. (200Qb).

Figure 1-8. Exposure response array of hematological and clinical chemistry effects
           following inhalation exposure to 1,2,3-TMB.
      This document is a draft for review purposes only and does not constitute Agency policy.

                                          1-43            DRAFT—DO NOT CITE OR QUOTE

-------
                                                  Toxicological Review of Trimethylbenzene
Isomer Effect Summary
Hematological
Subchromc
Toxicity
Clinical
Chemistry Subchronic
Effects
1,3,5-TMB
Duration Sex
f Monocytes(a) Males
t Phosphorus (a) Males
t Phosphorus (a) Females
4- Sodium (a) Females
4/ Chloride (a) Females
t Cholesterol (a) Females
4/ Glucose (a) Males
t Alkaline
, / v Males
Fhosphatase(a)
i- Low Do SB -i High Dose * HOAEL • LOAEL
• i

i
i



i


                                            0  50  100  150 200 250 300 350 400 450 500  550  600  650  700
                                                             Dose (mg/kg-day)
Note: Solid lines represent range of exposure concentrations, (a) Koch Industries f!995b).

Figure 1-9. Exposure response array of hematological and clinical chemistry effects
            following oral exposure to 1,3,5-TMB.
       This document is a draft for review purposes only and does not constitute Agency policy.

                                           1-44            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
     1.1.4.1.  Mode of Action Analysis - Hematological and Clinical Chemistry Effects.
 1          The mode of action for TMB-induced hematological and clinical chemistry effects has not
 2   been established. Increased sorbitol dehydrogenase activity is a marker for hepatic injury
 3   [Ramaiah. 2007] and therefore, underlying hepatotoxicity could explain its increase in rats exposed
 4   to 1,2,4-TMB or 1,2,3-TMB. However, absolute and relative liver weights were not observed  to
 5   increase with inhalation exposure to 1,2,4-TMB, and microscopic histopathological analysis  of the
 6   liver did not demonstrate any observable changes following exposure to either isomer. Similarly,
 7   although increased cholesterol levels and alkaline phosphatase levels could indicate hepatic
 8   dysfunction, no gross or histopathological lesions were observed in animals orally exposed to
 9   1,3,5-TMB. The increases in WBC counts in exposed animals could be secondary to the observed
10   respiratory irritative and inflammatory effects of 1,2,4-TMB exposure in Korsak et al. [2000a:
11   19971

     1.1.4.2.  Summary of Hematological and Clinical Chemistry Effects
12          Hematological and clinical chemistry toxicity was observed following inhalation and oral
13   exposure to TMBs based on evidence in humans and animals. The information regarding
14   hematological toxicity in humans is limited to one study involving exposure to a complex VOC
15   mixture containing both 1,2,4-TMB and 1,3,5-TMB [Battigetal.. 1956). as reviewed in MOE [2006]
16   and Baettig et al. [1958]. Although this study reported hematological effects (alterations in clotting
17   and anemia], exposure was to a mixture of TMB isomers and other VOCs. Therefore, it is impossible
18   to attribute the effects to any TMB isomer. There is evidence of hematological effects in male and
19   female Wistar rats following inhalation exposure [Korsak etal., 2000a, b], that are roughly
20   analogous to those observed in humans. Additionally, there is some evidence  of hematological and
21   clinical chemistry effects in male and female Sprague-Dawley rats following oral exposure [Koch
22   Industries. 1995bl
23          In summary, the evidence supports a determination that 1,2,4-TMB and 1,2,3-TMB result in
24   hematological toxicity following inhalation exposure, based on consistency and coherency of effects
25   across species (human and rats]. The general lack of data on hematological effects following
26   exposure to 1,3,5-TMB precludes a determination of hazard to humans for this isomer, although it
27   is reasonably anticipated given the observed effects following 1,2,4-TMB or 1,2,3-TMB exposure.

     1.1.5. Carcinogenicity
28          There are no studies in humans that investigated the carcinogenic potential of the TMB
29   isomers by any route of exposure. One animal study was identified that investigated the association
30   of chronic oral exposure (via gavage] to 1,2,4-TMB and cancer endpoints (Maltoni et al., 1997]. Male


            This document is a  draft for review purposes only and does not constitute Agency policy.

                                               1-45            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   and female Sprague-Dawley rats were exposed to a single dose of 800 mg/kg-day of 1,2,4-TMB in
 2   olive oil by stomach tube for 4 days/week starting at 7 weeks of age. Exposures were terminated at
 3   the end of 104 weeks (i.e., at 111 weeks of age) and the animals were keptunder observation until
 4   natural death. The authors report that chronic oral exposure to 1,2,4-TMB resulted in an
 5   "intermediate" reduction of survival in male rats and a "slight" reduction in females (no
 6   quantitative information on survival was reported). A slight increase in total malignant tumors in
 7   both sexes of rats was observed, with the incidence of head cancers being specifically increased in
 8   male rats. The predominant type of head cancer identified was neuroesthesioepithelioma, which
 9   arises from the olfactory neuroepithelium and is normally rare in Sprague-Dawley rats. Other head
10   cancers observed included those in the Zymbal gland, ear duct, and nasal and oral cavities. No tests
11   of statistical significance were reported for these data. When EPA performed the Fisher's exact test
12   on the incidences calculated from the reported percentages of animals bearing tumors in the
13   control and exposed animals, no statistically significant elevations in tumor incidence relative to
14   controls were observed.
15          Janik-Spiechowicz et al. [1998) investigated the genotoxicity of TMB isomers by measuring
16   three genotoxic endpoints: mutation frequency in bacteria, micronucleus formation in mice, and
17   sister chromatid exchanges in mice. Neither 1,2,4-TMB or 1,3,5-TMB induced gene mutations in any
18   Salmonella typhimurium strain tested (TA102, TA100, TA98,  and TA97a). However, 1,2,3-TMB
19   induced gene mutations in all four strains in absence of rat S9 fraction. When cells were incubated
20   in the presence of S9,1,2,3-TMB did not induce gene mutation, indicating possibly that 1,2,3-TMB
21   itself is the primary mutagen. No isomer induced the formation of micronuclei in Imp:BALB/c mice
22   following i.p. injection. Males in the high-dose groups for 1,2,4-TMB and 1,3,5-TMB, but not
23   1,2,3-TMB, exhibited a statistically significant reduction in the ratio of polychromatic erythrocytes
24   to normochromatic erythrocytes, indicating bone marrow cytotoxicity. All three isomers
25   significantly increased the frequency of sister chromatid exchanges (SCEs) in Imp:BALB/c mice
26   following i.p. injection, with 1,2,4-TMB eliciting the more significant response. These results appear
27   to have occurred at doses that did not induce significant bone marrow cytotoxicity.
28          In summary, very little genotoxicity data are available on TMBs. Janik-Spiechowicz et al.
29   [1998) observed varying results in the Ames mutation assay in Salmonella, with 1,2,3-TMB, but not
30   1,2,4-TMB or 1,3,5-TMB, inducing gene mutations. Results for the in vivo assays for micronucleus
31   and SCE formation  were consistent across isomers: TMB isomers were observed to induce SCEs, but
32   not micronuclei in mouse bone marrow cells. Increased frequency of SCEs indicates that DNA
33   damage has occurred as a result of exposure to these isomers, but it does not provide a specific
34   indication of mutagenic potential, as there is no known mechanistic association between SCE
35   induction and a transmissible genotoxic effect. With only one isomer (1,2,3-TMB) demonstrating a
36   positive result for gene mutation and positive SCE results for all three isomers, there is inadequate
37   evidence to conclude that any isomer is directly genotoxic.

            This document is a draft for review purposes only and does not constitute Agency policy.

                                               1-46             DRAFT—DO NOT CITE  OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

     1.1.6. Similarities Among TMB Isomers Regarding Observed Inhalation and Oral
            Toxicity
 1          In the existing toxicological database for 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB, important
 2   similarities have been observed in the potency and magnitude of effect resulting from exposure to
 3   these three isomers in male and female Wistar rats, although some important differences also exist
 4   (Table 1-7).
 5          In acute studies investigating respiratory irritative effects (i.e., decreased respiratory rate),
 6   the RDso for the three isomers were very similar, ranging from 2,553 to 2,844 mg/m3 (Korsaketal.,
 7   1997). Measures of acute inhalation neurotoxicity, namely ECso values for decreases in rotarod
 8   performance (4,694 and 4,738 mg/m3) and pain sensitivity (5,683 5,963 mg/m3), were also similar
 9   for 1,2,4-TMB and 1,3,5-TMB, respectively (Korsak and Rydzynski. 1996). However, the ECso values
10   for both measures were lower following exposure to 1,2,3-TMB (3.779 and 4,172 mg/m3,
11   respectively). The observation that 1,2,3-TMB may be slightly more neurotoxic than 1,2,4-TMB or
12   1,3,5-TMB was also observed following acute oral and injection exposures. Although all three
13   isomers were observed to result in altered EEC readings, stronger and more persistent effects
14   followed a pattern of 1,2,3-TMB > 1,3,5-TMB > 1,2,4-TMB following oral exposures (Tomasetal..
15   1999al and 1,2,3-TMB > 1,2,4-TMB > 1,3,5-TMB following i.p. injections fTomas etal.. 1999cl Acute
16   exposure to both 1,2,4-TMB and 1,2,3-TMB affected motor function and/or anxiety at similar
17   exposure levels, whereas 1,3,5-TMB appeared to be slightly more potent, although the magnitude of
18   the response across isomers suggests that this difference is negligible (Tomas etal.. 1999b).
19          In short-term neurotoxicity studies, a qualitatively similar pattern of effects (inability to
20   learn passive and/or active avoidance and decreased pain sensitivity following foot shock
21   challenge) indicating altered neurobehavioral function was observed for TMBs, although some
22   quantitative differences were noted (Wiadernaetal..20Q2: Gralewicz and Wiaderna. 2001:
23   Wiaderna et al.. 1998: Gralewicz etal.. 1997b). Exposure to any isomer resulted in statistically
24   significant decreases in pain sensitivity following foot shock challenge at the same concentration,
25   although the magnitude of effect and  consistency across studies was greater for 1,3,5-TMB and
26   1,2,4-TMB compared to 1,2,3-TMB (Wiaderna etal.. 2002: Gralewicz and Wiaderna. 2001:
27   Wiaderna etal.. 1998: Gralewicz etal.. 1997b). 1,2,4-TMB and 1,3,5-TMB were also observed to
28   increase motor function and/or decrease anxiety in open field tests, whereas 1,2,3-TMB was
29   observed to have no statistically significant effects fLutz etal.. 2010: Wiaderna etal.. 2002.1998:
30   Gralewicz etal.. 1997b). In contrast, increased locomotor activity elicited by amphetamine was
31   amplified following exposure to 1,2,3-TMB, but not 1,2,4-TMB (Lutz etal.. 2010). All three isomers
32   elicited effects on cognitive function,  as measured by learning decrements in two-way active
33   avoidance or by decreased fear responses in a passive avoidance test paradigm(Wiaderna et al.,
34   2002: Gralewicz and Wiaderna. 2001: Wiaderna etal.. 1998: Gralewicz et al.. 1997b). 1,3,5-TMB
35   was observed to be the most potent isomer in this regard, eliciting effects on both passive and

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-47             DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   active avoidance at > 123 mg/m3.1,2,3-TMB and 1,2,4-TMB affected passive avoidance
 2   performance at > 123 and > 492 mg/m3, respectively, and both 1,2,3-TMB and 1,2,4-TMB affected
 3   the ability to learn active avoidance at 492 mg/m3. For all isomers, short-term exposure to 1,230
 4   mg/m3 TMB was nearly always less effective (or ineffective), as compared to lower TMB
 5   concentrations, at eliciting responses (i.e., responses were nonlinear).
 6          Following subchronic exposure to either 1,2,4-TMB or 1,2,3-TMB, both decreased pain
 7   sensitivity and decreased rotarod performance were observed. With regard to decreased pain
 8   sensitivity, although 1,2,3-TMB was observed to decrease pain sensitivity at a lower concentration
 9   than 1,2,4-TMB, the magnitude of effect was similar between isomers at every concentration
10   (Korsak and Rydzynski. 1996). For either isomer, effects on pain sensitivity appeared to be
11   reversible at 1,230  mg/m3 TMB; lower concentrations were nottested. 1,2,3-TMB was more potent
12   than 1,2,4-TMB in reducing rotarod performance. Specifically, 1,2,3-TMB elicited effects at a lower
13   concentration and caused a greater magnitude of effect at each concentration, as well as following a
14   period of recovery (Korsak and Rydzynski. 1996).
15          Similarities were also observed in 1,2,4-TMB- and 1,3,5-TMB-induced developmental and
16   maternal effects (Saillenfaitetal.. 2005). Male fetal weights were significantly reduced in animals
17   exposed gestationally to 2,952 mg/m31,2,4-TMB (5% decrease) or 1,3,5-TMB (7% decrease).
18   1,2,4-TMB also significantly decreased female fetal weights by approximately 5% in animals
19   exposed to the same concentration. Although, 1,3,5-TMB significantly reduced female fetal weights
20   by 13% in animals exposed to 5,904 mg/m3, female fetal weights were decreased at 2,952 mg/m3 to
21   a similar degree (6%) as animals exposed to the same concentration of 1,2,4-TMB. Maternal
22   toxicity, measured as decreased corrected maternal weight gain, was significantly decreased in
23   animals exposed to 2,952 mg/m31,2,4-TMB or 1,3,5-TMB. However, 1,3,5-TMB exposure resulted
24   in a 75% reduction of maternal weight gain compared to controls, whereas 1,2,4-TMB exposure
25   reduced maternal weight gain by 50%.
26          Lastly, 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB were observed to elicit hematological toxicity
27   in exposed animals. Although all three isomers were observed to qualitatively affect similar
28   hematological parameters, the direction and magnitude of effect often differed between isomers.
29   Red blood cells were significantly decreased in male rats exposed to 1,230 mg/m31,2,3-TMB (23%
30   decrease) or 1,2,4-TMB (15% decrease) (Korsak et al.. 2000a. b). Reticulocyte numbers were also
31   altered in rats following exposure to these isomers, although 1,2,4-TMB was observed to
32   significantly decrease reticulocytes in male rats at 1,230 mg/m3 (71% decrease), while exposure to
33   1,2,3-TMB increased reticulocytes in male rats at 1,230  mg/m3 (61% increase) and female rats at
34   123 and 492 mg/m3 (77% and 100% increases, respectively).  1,2,3-TMB and 1,2,4-TMB were also
35   altered the numbers of white blood cells in exposed animals following subchronic exposures. In
36   male rats exposed to 1,230 mg/m31,2,4-TMB, white blood cell numbers were significantly
37   increased by 80%. Exposure to 1,230 mg/m31,2,3-TMB also increased lymphocyte numbers by

            This document is a draft for review purposes only and does not constitute Agency policy.

                                               1-48            DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
 1    11% and 15% in male and female rats, respectively. Exposure to 1,230 mg/m31,2,3-TMB decreased
 2    segmented neutrophils by 29% in male rats, whereas exposure to 492 mg/m3 and 1,230 mg/m3
 3    decreased neutrophil numbers in female rats by 29% and 48%, respectively. Acute exposure (six
 4    hours) to 1,500 - 6,000 mg/m31,3,5-TMB was also reported to result in increased numbers of
 5    segmented neutrophils that persisted for up to 28 days post exposure [Wiglusz etal., 1975b]. A
 6    summary of these comparisons across isomers is presented below in Table 1-7.

     Table 1-7. Similarities between 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB regarding
                observed inhalation and oral toxicity
Health Outcome Measure
Pain Sensitivity
Pain Sensitivity following foot
shock challenge
Neuromuscular Function
Motor Function / Anxiety
Sensitization
Cognitive Function
Electrocortical activity
Respiratory Effects
Developmental Effects
Hematological Effects
Exposure Duration
acute
subchronic
short-term
acute
subchronic
short-term
short-term
short-term
acute
acute
gestational
subchronic
TMB Isomer Potency
1,2,3-TMB > 1,2,4-TMB * 1,3,5-TMB
1,2,4-TMB * 1,2,3-TMB
1,2,4-TMB * 1,3,5-TMB > 1,2,3-TMB
1,2,3-TMB > 1,2,4-TMB * 1,3,5-TMB
1,2,3-TMB > 1,2,4-TMB
1,2,4-TMB * 1,3,5-TMB » 1,2,3-TMB
1,2,3-TMB > 1,2,4-TMB
1,3,5-TMB > 1,2,4-TMB * 1,2,3-TMB
1,2,3-TMB » 1,3,5-TMB > 1,2,4-TMB
1,2,4-TMB * 1,3,5-TMB * 1,2,3-TMB
1,2,4-TMB = 1,3,5-TMB
1,2,4-TMB * 1,2,3-TMB
     1.1.7.  Similarities Among TMB Isomers Regarding Toxicokinetics
 7          In the existing toxicokinetic database for 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB, important
 8   similarities have been observed in the chemical properties and absorption, distribution,
 9   metabolism, and excretion profiles for these isomers in animals and humans, although some
10   important differences also exist
11          All three isomers have very similar Log Kow values (3.42-3.78), and blood:air partition
12   coefficients reported for humans and rats in the literature are similar: 43.0 and 55.7 for 1,2,4-TMB,
13   66.5 and 62.6 for 1,2,3-TMB, and 59.1 and 57.7 for 1,3,5-TMB fMeulenberg and Viiverberg. 20001
14   This gives an indication that the three isomers would partition into the blood in a similar fashion.
           This document is a draft for review purposes only and does not constitute Agency policy.

                                              1-49            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   Supporting this is the observation that 1,2,4-TMB and 1,3,5-TMB absorb equally into the
 2   bloodstream of exposed humans (6.5 and 6.2 |iM, respectively), although the absorption for 1,2,3-
 3   1MB was observed to be higher (7.3 [iM) (Jarnbergetal.. 1998.1997a: Jarnbergetal.. 1996). Also,
 4   the net respiratory uptake of 1,2,3-TMB, 1,2,4-TMB and 1,3,5-TMB was similar among humans (48-
 5   60%), and the respiratory uptake for 1,2,4-TMB was similar across humans and rats (50-
 6   60%)(Tarnbergetal.. 1996: Dahl etal.. 1988). Although no data exist regarding the distribution of
 7   TMB isomers in humans, experimentally-derived tissue-specific partition coefficients were similar
 8   for all three isomers across a number of organ systems (Meulenberg and Vijverberg. 2000).
 9   strongly suggesting that the individual isomers can be  expected to distribute similarly to these
10   various organ systems. Distribution of the 1,2,4-TMB and 1,3,5-TMB throughout the body is
11   qualitatively similar in animals, although it appears that liver and kidney concentrations for
12   1,2,4-TMB are greater than those for 1,3,5-TMB after both acute and short-term inhalation
13   exposures fSwiercz etal.. 2006: Swiercz etal.. 2003: Swiercz etal.. 20021 Although 1,2,4-TMB was
14   observed to distribute to the brain (Swiercz etal.. 2003: Eide and Zahlsen. 1996}. distribution of
15   1,3,5-TMB to the brain was not experimentally measured in any study. However, the predicted
16   brain:air partition coefficient was similar between 1,2,4-TMB and 1,3,5-TMB for both humans (206
17   vs. 199) and rats (552 vs. 535) (Meulenberg and Vijverberg. 2000). This strongly suggests that
18   1,2,4-TMB and 1,3,5-TMB can be expected to distribute similarly to the brain in both humans and
19   rats. Detailed information regarding the distribution of 1,2,3-TMB following inhalation exposure is
20   lacking. However, similar tissue-specific partition coefficients for 1,2,3-TMB compared to 1,2,4-TMB
21   and 1,3,5-TMB indicate a similar pattern of distribution can be reasonably anticipated (Meulenberg
22   and Vijverberg. 2000).
23          All three TMB isomers were observed to primarily metabolize to benzoic and hippuric acids
24   in humans and rats (Tarnbergetal., 1996: Huoetal., 1989: Mikulski and Wiglusz, 1975), although
25   the amounts of inhaled TMB recovered as hippuric acid metabolites following exposure to 1,2,3-
26   TMB, 1,2,4-TMB,  or 1,3,5-TMB was dissimilar in humans (11%, 22%, and 3%, respectively) and rats
27   (10%, 24-38%, and 59%, respectively)  flarnberg etal.. 1996: Mikulski and Wiglusz. 19751 Greater
28   amounts of urinary benzoic acid and hippuric acid metabolites (73%) were observed after
29   exposure to higher amounts of 1,3,5-TMB (up to 30.5 ppm) for 8 hours (Kostrzewskietal.. 1997:
30   Kostrewski and Wiaderna-Brycht. 1995). Other terminal metabolites included mercapturic acids
31   (~14-19% total dose), phenols (~12% total dose), and glucuronides and sulphuric acid conjugates
32   (4-9% total dose) for 1,2,4-TMB; mercapturic acids (-5% total dose), phenols (
-------
                                                      Toxicological Review of Trimethylbenzene

 1          In humans, the half-lives of elimination from blood were observed to be similar for all
 2   isomers in the first three phases of elimination: 1,2,4-TMB (1.3 ± 0.8 min, 21 ± 5 min, 3.6 ± 1.1 hr),
 3   1,2,3-TMB (1.5 ± 0.9 min, 24 ± 9 min, 4.7 ± 1.6 hr), and 1,3,5-TMB (1.7 ± 0.8 min, 27 ± 5 min, 4.9 ±
 4   1.4 hr) (Tarnbergetal., 1996). The half-life of elimination for 1,3,5-TMB in the last and longest
 5   phase is much greater than those for 1,2,4-TMB or 1,2,3-TMB (120 ± 41 hr vs. 87 ± 27 and 78 ± 22
 6   hr, respectively). Urinary excretion of unchanged parent compound was extremely low (<0.002%)
 7   for all three isomers (Tanasik et al.. 2008: Tarnbergetal.. 1997b). The difference observed in half-
 8   lives between the three isomers in the last elimination phase may be due to small sample sizes and
 9   difficulties in measuring slow elimination phases rather than a true difference in half-lives. All three
10   isomers were eliminated via exhalation: 20-37% of the absorbed dose of 1,2,4-TMB, 1,2,3-TMB, or
11   1,3,5-TMB was eliminated via exhalation during exposure to 123 mg/m3 (25 ppm) for 2 hours
12   (Tarnbergetal., 1996). At low concentrations in rats, half-life of elimination from the blood was
13   greater for 1,2,4-TMB compared to 1,3,5-TMB (3.6 vs. 2.7 hours). This difference became much
14   greater with increasing doses (17.3 hours for 1,2,4-TMB and 4 hours for 1,3,5-TMB following
15   exposure to 1,230 mg/m3 for 6 hours) fSwiercz etal..  2003: Swiercz etal.. 20021 For a full
16   discussion of the chemical properties and toxicokinetics 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB see
17   Appendices B.I and B.2.

     1.2.  Summary and Evaluation

     1.2.1. Weight of Evidence for Effects Other Than Cancer
18          In both humans and animals, inhalation exposure to TMBs has been shown to result in
19   toxicity in multiple organ systems, including the nervous, respiratory, and hematological systems.
20   In addition, developmental toxicity has been observed in animals exposed to either 1,2,4-TMB or
21   1,3,5-TMB. Generally, the information regarding inhalation toxicity in humans is limited for a
22   number of reasons, including that the majority of human studies involved exposure to complex VOC
23   mixtures containing several TMB isomers and other VOCs, and not the individual isomers
24   themselves. Therefore, the observed health effects cannot be attributed to specific TMB isomers.
25   However, these studies observe effects in exposed human populations that are generally analogous
26   to effects observed in animal toxicity studies, and provide qualitative, supportive evidence for
27   hazard identification. Currently, no human studies exist that investigate the oral toxicity of any TMB
28   isomer. Potential limitations in the animal inhalation and oral toxicity database for TMBs include
29   the lack of a chronic study and the fact that all of the available inhalation animal studies were
30   conducted by the same research group: The Nofer Institute of Occupational Medicine, Lodz  Poland.
31          The most strongly and widely supported manifestation of toxicity in humans and animals
32   following inhalation exposure to 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB is neurotoxicity. In humans
33   exposed to TMB-containing VOC mixtures, a multitude of effects, including neuropsychological

            This document is a draft for review purposes only and does not constitute Agency policy.

                                               1-51            DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1   effects [Chenetal.. 1999). deficits in short-term memory and reduced motor speed/coordination
 2   [Lee etal.. 2005]. abnormal fatigue [Norseth et al.. 1991]. dysfunction of the inner ear/vertigo
 3   [Sulkowskietal.. 2002]. and nervousness, anxiety, and/or vertigo [Battig et al. [1956]. as reviewed
 4   by MOE [2006] and Baettig et al. [1958]], have been observed. None of the available human studies
 5   have addressed the potential for latent neurological effects and no studies examined the potential
 6   for neurological effects in sensitive populations. Although the reported human symptoms do not
 7   directly parallel the animal data, exposure of male Wistar rats to the TMB isomers has been shown
 8   to consistently result in a multitude of neurotoxic effects, including decreased pain sensitivity,
 9   impaired neuromuscular function and coordination, altered cognitive function, decreased anxiety
10   and/or increased motor function, and neurophysiological effects (e.g., decreased electrocortical
11   activity] across multiple concentrations and durations [Wiadernaetal., 2002: Gralewicz and
12   Wiaderna. 2001: Wiadernaetal.. 1998: Gralewicz etal.. 1997b: Gralewicz etal.. 1997a: Korsakand
13   Rydzynski. 1996: Korsak etal.. 19951
14          The effects observed in the animal neurotoxicity studies are recognized in the U.S. EPA's
15   Guidelines for Neurotoxicity Risk Assessment [U.S. EPA. 1998] as possible indicators of neurotoxicity.
16   The effects observed include concentration-dependent decrements in pain sensitivity in hot plate
17   tests and neuromuscular function in rotarod tests following subchronic exposure. Although effects
18   on pain sensitivity appeared to be reversible atthe highest concentration (i.e., 1,230 mg/m3],
19   reversible effects occurring in occupational settings may be of high concern, particularly if they
20   diminish a person's ability to survive or adapt to the environment [(U.S. EPA. 1998]. pg.8]; such is
21   the case for exposure to TMBs in occupations with dangerous surroundings and/ or heavy
22   equipment, such as dockyard painters or asphalt workers. These effects are supported by
23   additional data from short-term exposure studies that consistently identified latent effects of TMBs
24   exposure on pain sensitivity in hot plate tests following an environmental challenge (i.e., foot
25   shock], alongside reproducible learning decrements in passive and active avoidance experiments,
26   altered EEC patterns, and increased locomotor activity in open field tests. Further, the data from
27   these short-term studies clearly indicated a persistence of neurological effects several weeks after
28   TMB exposures had ended and identified a consistent nonlinearity in many of the TMB-elicited
29   responses (e.g., 1,230 mg/m3 was nearly always substantially less effective than 123 or 492
30   mg/m3]. The neurotoxic effects are biologically plausible and analogous to  effects that could occur
31   in humans. Thus, the evidence for TMBs identifies neurotoxicity as a toxicity hazard based on
32   consistency and coherency of effect across multiple studies and durations of exposure.
33          Three acute oral studies (Tomas etal., 1999a: Tomas et al., 1999b: Tomas etal., 1999c]
34   observe similar effects as observed in the available inhalation neurotoxicity studies (i.e., increased
35   locomotor activity and altered brain wave activity]. However, these studies are also limited with
36   regard to the range of endpoints investigated, and as such, no weight of evidence determination can
37   be made regarding the chronic oral toxicity of the TMB isomers.

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-52            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1          In addition to neurotoxicity, both respiratory and hematological toxicity have been
 2   observed in human populations and animals exposed to TMBs, or to mixtures containing the three
 3   isomers. In humans, occupational and residential exposure to VOC mixtures containing 1MB
 4   isomers have resulted in number of effects characterized as respiratory toxicity, including
 5   asthmatic bronchitis [[Battigetal.. 1956], as reviewed in MOE [2006] and Baettig et al. [1958]].
 6   asthma [Billionnetetal.. 2011]. or laryngeal/pharyngeal irritation [Norseth etal.. 1991].
 7   Additionally, workers exposed to a VOC mixture containing 1,2,4-TMB and 1,3,5-TMB, and possibly
 8   1,2,3-TMB, were reported to exhibit hematological effects including alterations in clotting time and
 9   anemia [[Battigetal.. 1956]. as reviewed in MOE [2006] and Baettig et al. [1958]]. Again, as
10   workers were exposed to complex VOC mixtures containing TMB isomers, the observed health
11   effects cannot be attributed to any single TMB isomer.
12          The observation of respiratory irritation and inflammation in Wistar rats and BALB/C mice
13   following exposure to 1,2,4-TMB was consistent across multiple concentrations, and subchronic
14   and acute exposure durations [Korsaketal.. 2000a: Korsaketal.. 1997: Korsaketal.. 1995].
15   Respiratory toxicity was also observed in multiple studies involving exposure to 1,2,3-TMB [Korsak
16   etal.. 2000b: Korsaketal.. 1995]. Although the reported symptoms in humans (laryngeal and/or
17   pharyngeal irritation, asthmatic bronchitis, and asthma] do not directly parallel the effects
18   observed in animal studies, the observation of irritative and/or inflammatory responses in multiple
19   species (including humans] demonstrates a consistency in TMB-induced respiratory toxicity.
20   Additionally, multiple measures of hematological toxicity have been observed in rats subchronically
21   exposed to 1,2,4-TMB or 1,2,3-TMB, including decreased RBCs, increased WBCs, decreased clotting
22   time, and decreased reticulocytes (1,2,4-TMB] and decreased RBCs, decreased segmented
23   neutrophils, increased lymphocytes and  increased reticulocytes (1,2,3-TMB] [Korsak et al.. 2000a.
24   b]. At least two of these effects, decreased RBCs and decreased clotting time, are roughly analogous
25   to the hematological effects (alterations in clotting and anemia] observed in occupationally exposed
26   humans, thereby demonstrating a consistency and coherency of effect across species. Therefore, the
27   respiratory and hematological effects observed in animals are biologically plausible and analogous
28   to effects that could occur in exposed human populations. The available evidence for 1,2,4-TMB and
29   1,2,3-TMB identified respiratory and hematological toxicity as a hazard.
30          Currently, no human studies exist that investigate the reproductive or developmental
31   toxicity of 1,2,3-TMB, 1,2,4-TMB, or 1,3,5-TMB. However, one animal study (Saillenfait etal.. 2005]
32   observed effects on fetal body weights and maternal body weight gains due to gestational exposure
33   to 1,2,4-TMB or 1,3,5-TMB. Although the weight of evidence regarding developmental toxicity is not
34   as strong compared to other measures of toxicity in the TMB database, these effects observed in
35   animals are considered biologically plausible and potentially analogous to effects that could occur
36   in humans. The available evidence for 1,2,4-TMB and 1,3,5-TMB identifies maternal and
37   developmental toxicity as a hazard.

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-53            DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

     1.2.2. Weight of Evidence for Carcinogenicity
 1          Under the Guidelines for Carcinogen Risk Assessment [2005a], the database for the TMBs
 2   provides "inadequate information to assess carcinogenic potential" of these isomers. This
 3   characterization is based on the fact that there is no information regarding the carcinogenicity of
 4   TMB in humans and that the only animal study available on the carcinogenicity of 1,2,4-TMB
 5   observed no statistically significant carcinogenic effects. No studies regarding the carcinogenicity of
 6   1,2,3-TMB or 1,3,5-TMB were identified in the available scientific literature.
 7          In the animal carcinogenicity study identified [Maltoni et al.. 1997]. involving exposure to
 8   1,2,4-TMB by oral gavage, an increased incidence of total malignant tumors in both sexes and head
 9   cancers (predominantly neuroethesioepithelioma) in males was observed in exposed rats, no
10   statistical analyses were reported. When EPA independently performed the Fisher's exact test on
11   the reported data, no statistically significant effects were observed.
12          Additionally, in the only study investigating the genotoxicity of TMB isomers, Janik-
13   Spiechowicz et al. [1998] observed negative results in in vitro genotoxicity assays (i.e., Ames
14   mutation assay in Salmonella) involving 1,2,4-TMB and 1,3,5-TMB. However,  1,2,3-TMB was
15   observed to induce gene mutations in all Salmonella typhimurium strains tested. All three isomers
16   failed to induce micronuclei in mouse bone marrow cells. Janik-Spiechowicz et al. [1998] observed
17   an increased incidence of SCE in mice exposed to all three TMB isomers (individually); however,
18   this observation does not provide a specific indication of mutagenic potential. Given the findings
19   regarding the in vitro genotoxicity of the TMB isomers, and increased frequency SCEs does not
20   provide specific indication of mutagenic potential, the evidence is inadequate to conclude that any
21   TMB isomer is genotoxic.

     1.2.3. Susceptible Populations and Lifestages
22          Although there are  no chemical-specific data that would allow for the  identification of
23   susceptible populations and lifestages, the reduced metabolic and elimination capacities in children
24   relative to adults may be a source of susceptibility [Ginsberg et al.. 2004). TMB isomers are
25   metabolized following inhalation and oral exposure via side-chain oxidation to form alcohols and
26   aromatic carboxylic/mercapturic acids or by hydroxylation to form phenols, which are then
27   conjugated with glucuronic acid, glycine, or sulfates for urinary excretion. The activities of multiple
28   cytochrome P450 (GYP P450) mono-oxygenase isozymes have been shown to be reduced in
29   children up to 1 year of age compared to adult activities [Ginsberg et al.. 2004). Additionally, the
30   rate of glucuronidation and sulfation is decreased in children. Therefore, as both GYP P450 mono-
31   oxygenase activities and the rate of glucuronidation and sulfation appear to be decreased in early
32   life, newborns and young infants may experience higher and more persistent blood concentrations
33   of the TMB isomers, and/or their respective metabolites compared with adults at similar exposure
34   levels. Reduced renal clearance in children may be another important source  of potential

            This document is a draft for review purposes only and does not constitute Agency policy.

                                                1-54            DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene

1   susceptibility. 1MB isomers and their metabolites are excreted in the urine of exposed laboratory
2   animals and occupationally exposed humans. Data indicating reduced renal clearance for infants up
3   to 2 months of age [Ginsberg et al.. 2004] may suggest a potential to affect 1MB excretion, thus
4   possibly prolonging its toxic effects. Additionally, those with pre-existing respiratory diseases (e.g.,
5   asthma) may be more sensitive to the respiratory irritative and inflammatory effects of TMB
6   isomers.
7
           This document is a draft for review purposes only and does not constitute Agency policy.

                                               1-55            DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
     2. DOSE-RESPONSE ANALYSIS
     2.1. Inhalation Reference Concentration for Effects Other Than Cancer
          for 1,2,4-TMB
 1          The RfC (expressed in units of mg/m3) is defined as an estimate (with uncertainty spanning
 2   perhaps an order of magnitude) of a continuous inhalation exposure to the human population
 3   (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects
 4   during a lifetime. It can be derived from a NOAEL, LOAEL, or the 95% lower bound on the
 5   benchmark concentration (BMCL), with UFs generally applied to reflect limitations of the data used.

     2.1.1.  Identification of Studies and Effects for Dose-Response Analysis for 1,2,4-TMB
 6          The nervous, respiratory, hematological systems, as well as pregnant animals and the
 7   developing fetus, are the primary targets of inhaled 1,2,4-TMB in humans and experimental
 8   animals, and effects in these systems have been identified as hazards following inhalation exposure
 9   to 1,2,4-TMB.
10          The selection of studies and general procedures for dose-response analysis are outlined in
11   Sections 6 and 7 of the Preamble. Human data are preferred over animal data for deriving reference
12   values when possible because the use of human data is more relevant in the assessment of human
13   health and avoids the uncertainty associated with interspecies extrapolation introduced when
14   animal data serve as the basis for the reference value. In this case, while literature exists on the
15   effects of 1,2,4-TMB exposure in humans, including neurological, respiratory, and hematological
16   toxicities, no human studies are available that would allow for dose-response analysis. The human
17   studies evaluated TMB exposures occurring as complex solvents or VOC mixtures, and this
18   confounding along with other uncertainties including high imprecision in effect measures due to
19   low statistical power, lack of quantitative exposure assessment, and lack of control for
20   co-exposures, limit their utility in  derivation of quantitative human health toxicity values. However,
21   these studies provide supportive evidence for the neurological, respiratory, and hematological
22   toxicity of TMB isomers in humans and indicate a coherency of effects  in both humans and
23   laboratory animals.
24          Several studies investigating 1,2,4-TMB effects in experimental animal models were
25   identified in the literature. No chronic studies were available, although acute, short-term,

                 This document is a draft for review purposes only and does not constitute Agency policy.
                                                2-1
                                                              DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene
 1   subchronic, and developmental toxicity studies were identified. 1,2,4-TMB-induced toxicity was
 2   observed across several organ systems in three subchronic studies by Korsak et al., [2000a: 1997]
 3   and Korsak and Rydzyhski [1996]. and in pregnant animals and developing fetuses in a
 4   developmental toxicity study by Saillenfait et al. [2005]. These four studies were the only
 5   subchronic or developmental studies identified in the peer-reviewed literature. Data from these
 6   studies pertaining to the primary hazards observed in humans and animals identified in Chapter 1
 7   (neurological, respiratory, and hematological toxicity] or in animals only (maternal and
 8   developmental toxicity] were considered as candidate critical effects for the purpose of
 9   determining the point of departure (POD] for derivation of the inhalation RfC for 1,2,4-TMB.
10   Neurotoxicity was also observed in both acute  and short-term inhalation studies and respiratory
11   toxicity was also observed in acute studies. However, the high concentrations used in acute studies
12   and the short exposure durations employed in both acute and short-term studies limit their utility
13   for the quantitation of chronic human health effects. Nevertheless, as with  the human mixture
14   studies, these studies provide  qualitative information regarding hazard identification, especially the
15   observation of the consistency and coherency of these effects across the 1,2,4-TMB database.
16          The three subchronic studies by Korsak et al., (2000a: 1997] and Korsak and Rydzyhski
17   (1996]. and the developmental toxicity study by Saillenfait et al. (2005]. adequately supported dose
18   response analysis. All four studies exposed rats, a common model for human response, by
19   inhalation, to 1,2,4-TMB (reported as > 97-99% pure [impurities not reported]]. All studies used at
20   least three exposure levels, spaced approximately threefold apart All controls were exposed under
21   similar conditions to untreated air. The durations of exposure, subchronic  or gestational, were
22   suitable for the effects under evaluation:  neurological, developmental, and short-term general
23   toxicity. In addition, the persistence of some outcomes after termination of exposure was
24   investigated.  Typical numbers of animals per exposure group for these study designs were used: at
25   least 10/group for the subchronic studies [Korsak et al., (2000a: 1997]. Korsak and Rydzyhski
26   (1996]]:  and  25/group for the developmental study (Saillenfait et al. (2005]. Regarding exposure
27   characterization, Korsak et al.  (2000a] and Saillenfait et al. (2005] reported actual concentrations,
28   as measured by gas chromatography, to be within 10%  of target concentrations. This increases the
29   confidence in the overall adequacy of these studies. Although Korsak and Rydzyhski (1996] and
30   Korsak et al. (1997] did not report actual, measured concentrations, these  studies used the same
31   exposure methodology as Korsak et al. (2000a]: suggesting that it is likely that the actual
32   concentrations in these studies were also within 10% of target concentrations. Target and actual
33   concentrations are presented in Table 2-1.
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                   2-2
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                          Toxicological Review of Trimethylbenzene
     Table 2-1. Target and actual inhalation concentrations, and internal blood dose
                 metrics of 1,2,4-TMB calculated using the available rat PBPK model
                 (Hissinketal., 2007)
Reference
Korsak and
Rydzyriski
(1996)
Korsak et al.
(1997)
Korsak et al.
(2000a)
Saillenfait et
al. (2005)
Species/
sex
Rat, male
Rat, male
Rat, male
Rat, female
Rat:
Female
(pregnant
dam); Male
and female
(fetuses)
Body
weight (kg)a
0.387
0.404
0.403
0.383
0.409
0.416
0.390
0.399
0.389
0.243
0.230
0.229
--
--
--
--
Exposure concentration
(mg/m3)b
123
492
1,230
123
492
1,230
123 (129)
492 (492)
1,230 (1,207)
123 (129)
492 (492)
1,230 (1,207)
492 (492)
1,476 (1,471)
2,952 (2,913)
4,428 (4,408)
Internal dose - average
weekly venous blood
concentration (mg/L)
0.1272
0.8666
5.4424
0.1272
0.8661
5.4274
0.1339
0.8671
5.2481
0.1335
0.8899
5.5189
n/a
n/a
n/a
n/a
     aFor Korsak et al. (2000a; 1997), exposure group-specific terminal body weights from those studies were used to calculate
      internal dose metrics; for Korsak and Rydzyriski (1996) the average of the exposure group-specific body weights reported in
      Korsak et al. (2000a; 1997) were used in internal dose metric calculations. For Saillenfait et al. (2005), body weights were not
      provided so the PBPK model was not used to derive internal dose metrics for this study
     b For Korsak et al. (2000a) and Saillenfait et al. (2005), values in parentheses are actual concentrations, as measured by gas
      chromatography
     Rat PBPK model (Hissinket al., 2007)
1           These subchronic and developmental toxicity studies examined 1,2,4-TMB-induced toxicity
2    in multiple organ systems and neurological, respiratory, hematological, maternal, or developmental
3    toxicity endpoints that demonstrated statistically significant increases or decreases relative to
4    control were considered  for the derivation of the RfC for 1,2,4-TMB (Table 2-2). The endpoints
5    included decreased pain  sensitivity in male rats [Korsak and Rydzynski. 1996]. increased BAL total
6    cells in male rats [Korsak etal.. 1997]. increased inflammatory lung lesions, decreased RBCs, and
7    increased WBCs in male rats and decreased reticulocytes and clotting time in female rats [Korsak et
                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                     2-3
                                                                   DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene
 1   al., 2000a], and decreased fetal weight (males and females) and decreased maternal weight gain
 2   [Saillenfaitetal.. 2005). Increases in BAL polymorphonuclear leukocytes and lymphocytes
 3   observed in the Korsak et al. [1997] study were not considered for RfC derivation due to a lack of
 4   reporting of exposures at which statistically significant increases occurred. Additionally, Korsak et
 5   al. [1997] reported that 123 mg/m3 was the LOAEL for increased BAL total cells, butthe NOAEL for
 6   increased BAL macrophages. Therefore, increased BAL macrophages were not considered for RfC
 7   derivation as these effects were not observed at concentrations that elicited an increase in total
 8   BAL cells. Changes in BAL protein and enzyme activity level were not considered due to non-
 9   monotonically increasing dose-responses, and increases in sorbitol dehydrogenase were not
10   further considered due to the lack of accompanying hepatocellular histopathological alterations in
11   exposed animals.
12          Impaired neuromuscular function and coordination, measured as performance deficits on
13   the rotarod apparatus, was also observed in rats exposed to 1,2,4-TMB. The use of rotarod data
14   from Korsak and Rydzyhski [1996] was initially considered as a candidate critical effect for
15   1,2,4-TMB. However, upon critical evaluation of the exposure-response information in the study, it
16   was determined that rotarod performance was reported in a manner that reduced the confidence in
17   the observed effect levels. The most widely used and accepted measure of rotarod performance in
18   rodents is latency to fall from the rotating rod [Brooks and Dunnett. 2009: Kaspar etal.. 2003: Bogo
19   etal., 1981], typically with an arbitrary upper limit on the maximum latency allowed to prevent
20   confounding by fatigue. The primary limitation for these data was that rotarod performance was
21   presented as percent of failures to last 2 minutes on the apparatus. Although the quantal percent
22   failures data can provide useful information, these measures require an arbitrary selection of the
23   length of time required for successful performance; there is no scientific consensus on an optimal
24   time for this parameter. In addition, when identifying effect levels based on the data presented by
25   Korsak and Rydzyhski [1996]. latencies on the rod of 1 second versus 119 seconds would be
26   treated identically as failures when, in fact, they indicate very different levels of neurological
27   dysfunction [Bogo etal.. 1981]. This adds uncertainty when trying to extrapolate to a concentration
28   associated with a minimally adverse effect Finally, this quantal presentation of data does not allow
29   for interpretations related to intra-rat and intra-group variability in performance. Due to these
30   reporting limitations, impaired neuromuscular function and coordination, measured as
31   performance deficits on the rotarod apparatus, was considered to be less informative than the data
32   supporting decreases in pain sensitivity, and thus, was excluded from consideration for derivation
33   of the RfC for 1,2,4-TMB.
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                  2-4
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                               Toxicological Review of Trimethylbenzene

Table 2-2. Endpoints considered for the derivation of the RfC for 1,2,4-TMB
Endpoint
Neurological endpoints
Decreased pain sensitivity
(measured as latency to paw-lick,
in seconds)b
Hematological endpoints
Decreased RBCs (106/mm3)c
(10s cells per 100 ul)
Increased WBCs (103/mm3)c
(103 cells per 100 ul)
Decreased reticulocytes (%)c
Decreased clotting time (s)c
Respiratory endpoints
Increased BAL total cells
(106/cm3)d
Increased inflammatory
lung lesions0
Developmental endpoints
Decreased fetal weight (g)f'g
Species/
sex

Rat,
male

Rat,
male
Rat,
female

Rat,
male

Rat,
male
Rat,
female
Exposure concentration (mg/m3)
0
15.4 ±
5.8a
(n = 9)
0
9.98 ±
1.68
(n = 10)
8.68 ±
2.89
(n = 10)
3.5 ±2.6
(n = 10)
30 ±10
(n = 10)
0
1.93 ±
0.79
(n = 6)
e
(n = 10)
0
5.86 ±
0.34
5.57 ±
0.33
123
18.2 ±
5.7
(n = 10)
123
9.84 ±
1.82
(n = 10)
8.92 ±
3.44
(n = 10)
1.7 ±2.0
(n = 10)
23 ±4
(n = 10)
123
5.82 ±
1.32***
(n = 6)
e
(n =10)
492
5.79 ±
0.30
5.51 ±
0.31
492
27.6 ±
3.2**
(n = 9)
492
8.50 ±
1.11
(n = 10)
8.30 ±
1.84
(n = 10)
1.8 ±0.9
(n = 10)
19 ± 5**
(n = 10)
492
5.96 ±
2.80**
(n = 7)
e
(n = 10)
1,476
5.72 ±
0.49
5.40 ±
0.45
1,230
30.1 ±7.9**
n = 10)
1,230
7.70+1.38**
(n = 10)
15.89 ± 5.74**
(n = 10)
1.0 ±0.6*
n = 10)
22 ±7*
(n =10)
1,230
4.45 ± 1.58*
(n = 7)
e
n = 10)
2,952 4,428
0.48* 5-20±°-42**
C TQ _|_
0.40* 4.92 ±0.40-
        This document is a draft for review purposes only and does not constitute Agency policy.

                                          2-5
                                                       DRAFT—DO NOT CITE OR QUOTE

-------
                                                          Toxicological Review of Trimethylbenzene
Table 2-2  (Continued): Endpoints considered for the derivation of the RfC for
                             1,2,4-TMB
Endpoint
Maternal endpoints
Decreased maternal weight
gain (g)f
Species/
sex

Rat,
female
Exposure concentration (mg/m3)
0
29 ±12
(n = 24)
492
31 ±14
(n = 22)
1,476
27 ±12
(n = 22)
2,952
15 ± 17**
(n=22)
4,428
0 ± 14**
(n = 24)
*p<0.05; **p<0.01; ***p< 0.001.
aValues are expressed as mean ± 1 SD. Korsak and Rydzyriski (1996) does not explicitly state that the reported measures of
variance in Table 1 of that reference are standard deviations. However, independent analysis conducted by EPA confirms that
the reported measures of variance are standard deviations.
b Adapted from Korsak and Rydzyriski (1996)
c Adapted from Korsak et al. (2000a)
d Adapted from Korsak et al. (1997)
Incidences for individual exposure groups not reported; however, based on qualitative information reported in the study (i.e.,
  that female rats exhibited a statistically significant increase in inflammatory lung lesions at 492 mg/m3), a NOAEL of
  123 mg/m3 was identified.
f Adapted from Saillenfait et al. (2005)
5 Numbers of fetuses not explicitly reported. See maternal weight gain for number of litters.
               This document is a draft for review purposes only and does not constitute Agency policy.
                                                     2-6
                                                                     DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

     2.1.2. Methods of Analysis for 1,2,4-TMB
 1          This assessment uses PBPK model estimates of internal blood dose metrics coupled with
 2   the benchmark dose (BMD) approach, when possible, to estimate a POD for the derivation of an RfC
 3   for 1,2,4-TMB (see Section B.3  of Appendix B and Section C.I of Appendix C for details regarding
 4   PBPK model estimates and BMD modeling, respectively). As dosimetry can often be non-linear due
 5   to metabolic saturation, and internal dose metrics are expected to correlate more closely to toxic
 6   response than external concentrations [Mclanahanetal.. 2012]. this assessment used the PBPK
 7   model-estimated internal dose metrics for dose-response modeling.
 8          A deterministic rat PBPK model [Hissinketal., 2007] was used to convert non-continuous
 9   external inhalation concentrations (in mg/m3] of 1,2,4-TMB to the internal blood dose metric of
10   average weekly venous blood concentration (in mg/L] of 1,2,4-TMB for Korsak et al., (2000a: 1997]
11   and Korsak and Rydzyhski (1996] only (see Table 2-1]. Weekly average venous blood 1,2,4-TMB
12   concentration was chosen as the internal dose metric on which to base the POD as it is assumed
13   that the parent compound is the toxic moiety of interest and that average venous blood
14   concentration of 1,2,4-TMB is assumed to adequately represent the target tissue dose across the
15   multiple tissues of interest. The use of concentration of parent compound in venous blood as the
16   relevant dose metric in non-metabolizing, non-first pass organs is recommended by Aylward et al.
17   (2011]. Furthermore, toluene-induced neurological effects in the brain are provided by Aylward et
18   al. (2011] as an example of a chemically induced toxic endpoint for which this dose metric is
19   relevant As discussed in Section 1 (Mode of Action Analysis - Neurotoxic Effects], 1,2,4-TMB is
20   reasonably expected to have a mode of action for neurotoxic effects similar to toluene, further
21   supporting the selection of venous blood concentration as the relevant internal dose metric.
22          One consequence of using PBPK model-estimated internal dose metrics as the dose inputs
23   for BMD modeling was the necessity of dropping the high exposure group in all datasets modeled.
24   During the validation and optimization of the animal PBPK model (Hissinketal., 2007] against
25   available animal toxicokinetic datasets, the model accurately reproduced venous blood
26   concentrations of 1,2,4-TMB following repeated (6 hours/day, 5  days/week, 4 weeks] exposures to
27   123 or 492 mg/m3 (see Section B.3.3.2, Appendix B]. However, the PBPK model consistently
28   overpredicted venous blood concentrations following exposure to 1,230 mg/m3. It was concluded
29   that the optimized animal PBPK model produces acceptable simulations of venous blood 1,2,4-TMB
30   concentrations for chronic exposures of up to 100 ppm [492 mg/m3] in rats following inhalation
31   exposure to 1,2,4-TMB (Section B.3.3.2, Appendix B]. Therefore, as the model-estimated internal
32   blood dose metrics at the high concentration are not representative of empirically observed blood
33   concentrations, using the high-dose model estimates as dose inputs for BMD modeling is not
34   appropriate. The decision to drop the high concentration results  in a loss of information regarding
35   dose-response characteristics at high concentrations and a reduction in the number of available

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-7
                                                               DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1   dose-response models to fit to the data (due to the number of model parameters > exposure
 2   groups). However, this methodology is preferred over inclusion of demonstrably inaccurate
 3   internal blood dose metrics that result from high exposure concentrations. Additionally, this
 4   methodology still allows for BMD modeling of these endpoints, which is preferred over use of the
 5   NOAEL/LOAEL approach.
 6           After calculation of internal blood dose metrics, those dose metrics were used as the dose
 7   inputs for BMD modeling. As the Hissink et al. [2007] PBPK model was not parameterized for
 8   pregnant animals and did not include a fetal compartment, internal dose metrics were not
 9   calculated from Saillenfait et al. [2005]. Instead, actual exposure concentrations were used for these
10   endpoints.
11           The BMD approach involves fitting a suite of mathematical models to the observed dose-
12   response data using EPA's Benchmark Dose Software (BMDS, version 2.2]. Each fitted model
13   estimates a BMD and its associated 95% lower confidence limit (BMDL] corresponding to a selected
14   benchmark response (BMR]. For continuous data (i.e., decreased pain sensitivity, increased BAL
15   total cells, decreased RBCs, decreased reticulocytes, and decreased clotting time] from the Korsak
16   and Rydzyhski [1996] and Korsak et al. [2000a: 1997] studies, and maternal weight gain from
17   Saillenfait et al. [2005]. no information is available regarding the change in these responses that
18   would be considered biologically significant, thus a BMR equal to a 1 standard deviation change in
19   the control mean was used in modeling these endpoints, consistent with EPA's Benchmark Dose
20   Technical Guidance [U.S. EPA. 2012b]. For the decreased male and female fetal body weight
21   endpoints identified from the Saillenfait et al. [2005] study, a BMR of 5% relative deviation from
22   the control mean was selected. A 5% decrease in fetal body weight relative to control was
23   determined to be a minimal, biologically significant response. This determination is based on the
24   fact that decreased body weight gain in fetuses and/or pups is considered indicative of altered
25   growth, which has been identified by EPA as one of the four major manifestations of developmental
26   toxicity [U.S. EPA, 1991]. In addition, a 10% decrease in adult body weight in animals is generally
27   recognized as a biologically significant response associated with identifying a maximum tolerated
28   dose, but since fetuses and/or pups are generally recognized as a susceptible lifestage, and thus are
29   assumed to be more greatly affected  by decreases in body weight than adult animals, a 5% decrease
30   in fetal body weight is considered a biologically significant response. Finally, in humans, reduced
31   birth weight is associated with a series of adverse effects including neonatal and postnatal
32   mortality, coronary heart disease, arterial hypertension, chronic renal insufficiency, and diabetes
33   mellitus [Barker, 2007: Reyes and Manalich, 2005]. For these reasons, the selection of a BMR of 5%
34   for decreased fetal body weight was considered reasonable. Additionally, a BMR equal to a
35   1 standard deviation change in the control mean was also selected for the BMD modeling of both
36   fetal body weight and maternal body weight gain to facilitate comparisons across assessments  [see
37   EPA's Benchmark Dose Technical Guidance [2012b]].

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-8
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene

 1          Some endpoints for 1,2,4-TMB were not modeled for a variety of reasons, including equal
 2   responses at all exposure groups (e.g., increased BAL total cells and decreased reticulocytes),
 3   responses only in the high exposure group with no changes in responses in lower exposure groups
 4   (e.g., increased WBCs), and absence of incidence data (e.g., increased inflammatory lung lesions).
 5   Additionally, some datasets were modeled, but no model provided estimated BMDLs that were
 6   considered to be biologically plausible (e.g., decreased clotting time). In cases where BMD modeling
 7   was not feasible or modeling failed to appropriately describe the dose-response characteristics, the
 8   NOAEL/LOAEL approach was used to identify a POD. Detailed modeling results are provided in
 9   Section C.I of Appendix C.
10          Because an RfC is a toxicity value that assumes continuous human inhalation exposure over
11   a lifetime, data derived from inhalation studies in animals dose metrics need to be adjusted to
12   account for the noncontinuous exposures used in these studies. This is addressed by calculation of
13   internal dose metrics for the Korsak et al., (2000a: 1997) and Korsak and Rydzyhski (1996) studies.
14   For the Saillenfait et al. (2005) study,  rats were exposed to 1,2,4-TMB for 6 hours/day for 15
15   consecutive days (GD6-GD20). Therefore, the duration-adjusted PODs for developmental/maternal
16   effects were calculated as follows:
17          PODADj (mg/m3) = POD (mg/m3) x hours exposed per day/24 hours
18   For example, for decreased fetal weight in males, the PODADj would be calculated as follows:
19          PODADj (mg/m3) = 1,640.07 mg/m3 x 6 hours/24 hours
20          PODADj (mg/m3) = 410 mg/m3
21          For the derivation of an RfC based upon animal data, the calculated PODADj values are
22   converted to human equivalent concentrations (HECs) using the  available human PBPK model
23   (Hissinketal..20Q7) for the selected endpoints from the Korsak et al., (2000a: 1997) and Korsak
24   and Rydzyhski (1996) studies. The human PBPK model was run (as described in Appendix B),
25   assuming a continuous (24 hours/day, 7 days/week) exposure, to estimate a human PODHEc that
26   would result from the same weekly average venous blood concentration reflected in the PODADj in
27   animals (Table 2-3). As the selected endpoints from Saillenfait et al. (2005) (i.e., decreased fetal
28   body weight, and maternal body weight gain) are assumed to result primarily from systemic
29   distribution of 1,2,4-TMB, and the Hissink et al. (2007) PBPK model is not parameterized for
30   pregnant animals and did not include a fetal compartment, the human equivalent concentration
31   (HEC) for 1,2,4-TMB was calculated by the application of the appropriate dosimetric adjustment
32   factor (DAF) for systemically acting gases (i.e., Category 3 gases), in accordance with the EPA's RfC
33   Methodology (U.S. EPA. 1994b). DAFs are ratios of animal and human physiologic parameters, and
34   are dependent on the nature of the contaminant (i.e., particle or gas) and the target site (i.e.,
35   respiratory tract or remote to the portal-of-entry [i.e., systemic])  (U.S. EPA.  1994b). For gases with

                 This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-9
                                                               DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene

 1    systemic effects, the DAF is expressed as the ratio between the animal and human blood:air
 2    partition coefficients:
 3           DAF = (Hb/g)A/(Hb/g)H
 4           where:
 5           (Hb/g)A = the animal bloochair partition coefficient
 6            (Hb/g)H = the human bloochair partition coefficient
 7           DAF = 57.7 flarnberg and Tohanson. 19951/59.1 fMeulenberg and Viiverberg. 20001
 8           DAF = 0.98
 9           In cases where the animal blood:air partition coefficient is lower than the human value,
10    resulting in a DAF < 1, the calculated value is used for dosimetric adjustments [U.S. EPA, 1994b].
11    For example, the HEC for decreased female fetal body weight (reported in Saillenfait et al. [2005]]
12    is calculated as follows:
13           PODHEc = PODADj (mg/m3) x DAF
14           PODHEC = PODADj (mg/m3) x 0.98
15           PODHEC = 403.2  mg/m3 x 0.98
16           PODHEc = 395.1  mg/m3
17           The calculated PODHEc (mg/m3] values for all endpoints considered for candidate value
18    derivation are presented in Table 2-3.
                 This document is a draft for review purposes only and does not constitute Agency policy.

                                                 2-10
                                                               DRAFT—DO NOT CITE OR QUOTE

-------
                                                    Toxicological Review of Trimethylbenzene

Table 2-3. Summary of derivation of points of departure for 1,2,4-TMB
Endpoint/Reference
Species/sex
Model; BMRor
NOAEL/LOAEL
PODa
Candidate
PODADJa
Candidate
PODHEc
(mg/m3)
Neurological endpoints
Decreased pain sensitivity
(Korsakand Rydzyriski, 1996)

Rat, male
Exponential M4;
ISO
0.086
0.086
15.8
Hematological endpoints
Decreased RBCs
(Korsaketal.,2000a)
Increased WBCs
(Korsaketal.,2000a)
Decreased reticulocytes
(Korsaketal.,2000a)
Decreased clotting time
(Korsaketal.,2000a)
Rat, male
Rat, male
Rat, female
Rat, female
Linear; ISO
NOAELb
NOAELb
NOAELb
0.499
0.867
0.890
0.134
0.499
0.867
0.890
0.134
83.9
131.5
134.0
24.4
Respiratory endpoints
Increased BAL total cells
(Korsaketal., 1997)

inflammatory lung lesions (Korsak et
al., 2000a)

Rat, male
Rat, male
LOAELb
NOAELb
0.127
0.134
0.127
0.134
23.2
24.4
Developmental endpoints
Decreased fetal weight
Saillenfait et al. (2005)
Rat, male
Rat, female
Linear, 5% RD
Linear, 5% RD
1,640.07
1,612.89
410
403.2
401.8
395.1
Maternal endpoints
Decreased maternal weight gain
(Saillenfait et al., 2005)
Rat, female
Exponential M3,
ISO
2,247.99
562
550.8
a Values are weekly average venous blood 1,2,4-TMB concentration (mg/L) for Korsak et al. (2000a; 1997) and
   Korsak and Rydzyriski (1996). See Appendix B for details on PBPK modeling, Values are in mg/m3 for
   Saillenfait etal. (2005)
b No model was able to fit data adequately, or data were not modeled. NOAEL/LOAEL method used to identify a
   POD
             This document is a draft for review purposes only and does not constitute Agency policy.

                                               2-11
                                                              DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

     2.1.3. Derivation of Candidate RfC Values for 1,2,4-TMB
 1          Under EPA's A Review of the Reference Dose and Reference Concentration Processes [[U.S.
 2   EPA, 2002] §4.4.5], also described in the Preamble, five possible areas of uncertainty and variability
 3   were considered in deriving the candidate RfC values for 1,2,4-TMB. An explanation of these five
 4   possible areas of uncertainty and variability and the values assigned to each as a designated
 5   uncertainty factor (UF) to be applied to the candidate PODnEc are as follows:
 6          An interspecies uncertainty factor, UFA, of 3 (101/2 = 3.16, rounded to 3) was applied to
 7   account for uncertainty in characterizing the toxicokinetic and toxicodynamic differences between
 8   rats and humans following inhalation exposure to 1,2,4-TMB. In this assessment, the use of a PBPK
 9   model to convert internal doses in rats to administered doses in humans reduces toxicokinetic
10   uncertainty in extrapolating from the rat to humans, but does not account for interspecies
11   differences due to toxicodynamics. A default UFA of 3 was thus applied to account for this remaining
12   toxicodynamic and any residual toxicokinetic uncertainty not accounted for by the PBPK model.
13          An intraspecies uncertainty factor, UFH, of 10 was applied to account for potentially
14   susceptible individuals in the absence of data evaluating variability of response in the human
15   population following inhalation of 1,2,4-TMB. No information is currently available to predict
16   potential variability in human susceptibility, including variability in the expression of enzymes
17   involved in 1,2,4-TMB metabolism.
18          A LOAEL to NOAEL uncertainty factor, UFi, of 1 was applied because the current approach is
19   to address this factor as one of the considerations in selecting a BMR for BMD modeling. In this
20   case, a BMR equal to a 1 standard deviation change in the control mean for modeled endpoints was
21   selected under the assumption that this BMR represents a minimal, biologically significant change
22   for these effects. For endpoints that could not be modeled, a LOAEL to NOAEL uncertainty factor of
23   1 was applied as a NOAEL was used, except for increased BAL cells to which a uncertainty factor of
24   10 was applied due to the use of a LOAEL for this endpoint
25          A subchronic to chronic uncertainty factor, UFS, of 3 (101/2 = 3.16, rounded to 3) was applied
26   to account for extrapolation from a subchronic exposure duration study to derive a chronic RfC, for
27   all endpoints except decreases in fetal weight, to which an UFS of 1 was applied. The 3-fold
28   uncertainty factor is applied to the POD identified from the subchronic study on the assumption
29   that effects observed in a similar  chronic study would be observed at lower concentrations for a
30   number of possible reasons, including potential cumulative damage occurring over the duration of
31   the chronic study or an increase in the magnitude or severity of effect with increasing duration of
32   exposure. For example, in the case of neurotoxicity, chronic exposures may overwhelm the adaptive
33   responses observed after termination of subchronic exposure, potentially resulting in more severe
34   and/or irreversible  changes in neurological function. A full subchronic to chronic uncertainty factor
35   of 10 was not applied in this case as there was evidence of reversibility of not only neurotoxic
36   effects, but also hematological effects in rats exposed to 1,2,4-TMB for subchronic durations. Also,

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-12
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1   the respiratory effects appeared to be inflammatory in nature. Although reversibility was not
 2   investigated for these endpoints, it is possible that adaptive mechanisms may alleviate these effects
 3   following the termination of exposure.
 4          A database uncertainty factor, UFD, of 3 (101/2 = 3.16, rounded to 3) was applied to account
 5   for database deficiencies. Strengths of the database  include the three well-designed subchronic
 6   studies that observe exposure-response effects in multiple organ systems (nervous, respiratory,
 7   and hematological systems), in Wistar rats exposed to 1,2,4-TMB via inhalation. An additional
 8   strength of the database is the well-designed developmental toxicity study that investigated
 9   standard measures of maternal and fetal toxicity in  a different strain of rat (Sprague-Dawley).
10   However, the lack of a multi-generation reproductive/developmental toxicity study is a weakness
11   of the database. EPA's A Review of the Reference Dose and Reference Concentration Processes [U.S.
12   EPA, 2002] recommends that the database uncertainty factor take into consideration whether there
13   is concern from the available toxicology database that the developing organism may be particularly
14   susceptible to effects in specific organ systems. TMBs (unspecified isomer) are able to cross the
15   placenta (Cooper etal.. 2001:  Dowty etal.. 1976}: therefore, as neurotoxicity is observed in adult
16   animals, there is the concern that exposure to 1,2,4-TMB may result in neurotoxicity in the
17   developing organism. EPA's Guidelines for Neurotoxicity Risk Assessment (U.S. EPA. 1998] identifies
18   specific effects observed in adult animals (e.g., cognitive and motor function] that can also affect the
19   developing organism exposed in utero. The Neurotoxicity Guidelines (U.S. EPA, 1998] also indicate
20   that neurotoxicants may have greater access to the nervous system in developing organisms due to
21   an incomplete blood-brain barrier and immature metabolic detoxifying pathways. Lastly, EPA's A
22   Review of the Reference Dose and Reference Concentration Processes (U.S. EPA. 2002] also states that
23   effects that may be mild or reversible in adults may produce more robust or permanent effects in
24   offspring following developmental exposures. Therefore, there is some concern that the lack of a
25   developmental neurotoxicity study is a deficiency in the database and that inclusion of such a study
26   would potentially result in a lower POD than the POD for neurotoxicity identified from the available
27   1,2,4-TMB toxicity database. In summary, a 3-fold database UF was applied to account for the lack
28   of both a multi-generation reproductive/developmental toxicity study and a developmental
29   neurotoxicity study in the available database for 1,2,4-TMB.
30          Table 2-4 is a continuation of Table 2-3, and summarizes the application of UFs s to each
31   POD to derive a candidate RfC value for each data set The candidate RfC values presented in Table
32   2-4 are preliminary to the derivation of the organ/system-specific RfC values. These candidate RfC
33   values are considered individually in the selection of a representative inhalation reference RfC
34   value for a specific hazard and subsequent overall RfC for 1,2,4-TMB. Figure 2-1 presents
35   graphically these candidate RfC values, uncertainty  factors, and points of departure, with each bar
36   corresponding to one data set described in Tables 2-3 and 2-4.
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                   2-13
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                           Toxicological Review of Trimethylbenzene

Table 2-4. Effects and corresponding derivation of candidate RfC values for
              1,2,4-TMB
Endpoint/Reference
HEC
(mg/m3)3
UFA
UFH
UFL
UFS
UFD
Composite
UF
Candidate RfC
value (mg/m3)b
Neurological endpoints
Decreased pain sensitivity
(Korsakand Rydzyhski, 1996)
15.8
3
10
1
3
3
300
5.27 x 10"2
Hematological endpoints
Decreased RBCs,
(Korsaketal.,2000a)
Increased WBCs
(Korsaketal.,2000a)
Decreased reticulocytes
(Korsaketal.,2000a)


Decreased clotting time
(Korsaketal.,2000a)
83.9
131.5
134.0
24.4
3
3
3
3
10
10
10
10
1
1
1
1
3
3
3
3
3
3
3
3
300
300
300
300
2.80 x 10"1
4.38 x 10"1
4.47 x 10"1
8.13xlO"2
Respiratory endpoints
Increased BAL total cells
(Korsaketal., 1997)
Increased inflammatory lung lesions
(Korsaketal.,2000a)
23.2
24.4
3
3
10
10
10
1
3
3
3
3
3,000
300
n/ac
8.13xlO"2
Developmental endpoints
Decreased fetal weight
(Saillenfait et al., 2005)

Maternal endpoints
rat, male
(rat, female)
401.8
395.1
3
3
10
10
1
1
1
1
3
3
100
100
4.02
3.95

Decreased maternal weight gain
(Saillenfait et al., 2005)
550.8
3
10
1
3
3
300
1.84
aHuman equivalent concentration.
 As calculated by application of uncertainty factors, not rounded to 1 significant digit.
cEndpoint excluded for further consideration due to a UFCOMPOSITE of 3,000. In the report, "A Review of the Reference Dose and
   Reference Concentration Processes" (U.S. EPA, 2002) the RfD/RfC Technical Panel concluded that, in cases where
   maximum uncertainty exists in four or more areas of uncertainty, or when the composite uncertainty factor is 10,000 or
   more, it is unlikely that the database is sufficient to derive a reference value. Although, 3,000 is generally recognized as
   the maximum composite uncertainty factor for RfC derivation, a candidate RfC based on the data for increased BAL total
   cells was not derived due to the fact that the uncertainty surrounding this endpoint was much higher than for any other
   endpoint.
               This document is a draft for review purposes only and does not constitute Agency policy.

                                                     2-14
                                                                      DRAFT—DO NOT CITE OR QUOTE

-------
                                                         Toxicological Review of Trimethylbenzene
         01

         D.
         _O
         01
         >
         01
         Q
         01
         cc
         E
         01
         3:
O fT3
= -^
d, 00
               Maternal body weight gain
               (Saillenfaitetal., 2005)
               Fetal body weight, females
               (Saillenfaitetal., 2005)
       4- Fetal body weight, males
         (Saillenfaitetal., 2005)
               Inflammatory lung lesions
                (Korsaketal, 2000a)
               t BAL Total eel Is
              (Korsaketal, 1997)
                     4- Clotting time
                   (Korsaketal., 2000a)
                     4- Reticulocytes
                   (Korsaketal., 2000a)
                 N|/WBCs
            (Korsaketal., 2000a)
                        N|/ RBCs
                   (Korsaketal., 2000a)
                  \|/ Pa in Sensitivity
             (Korsakand Rydzynski, 1996)
                                         Composite UF

                                       A Candidate value

                                       • POD(HEC)
                                   0.001      0.01       0.1        1        10
                                                         Concentrations (mg/m3)
                                                                              100
1000
     Figure 2-1. Candidate RfC values with corresponding POD and composite UF for
                 1,2,4-TMB.
     2.1.4.  Derivation of Organ/System Specific Reference Concentrations for 1,2,4-TMB
1           Table 2-5 distills the candidate RfC values from Table 2-4 into a single value for each organ
2    or system. The single RfC value selected for a particular organ system was preferably chosen using
3    biological and toxicological information regarding that endpoint. If no compelling biological
4    information exists with which to select the primary hazard, the lowest RfC value for that organ
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                    2-15
                                                                   DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene
 1    system was selected. These organ- or system-specific reference concentrations may be useful for
 2    subsequent cumulative risk assessments that consider the combined effect of multiple agents acting
 3    at a common site. The individual organs and systems for which specific RfC values were derived
 4    were the neurological, hematological, and respiratory systems, along with specific RfCs derived for
 5    the pregnant animal (maternal) and developing fetus (developmental). The RfC value for the
 6    neurological system, based on decreased pain sensitivity, was selected for the proposed overall RfC
 7    for 1,2,4-TMB (see Section 2.1.5 for details). The RfC values for the hematological and respiratory
 8    systems, based on decreased clotting time and increased inflammatory lung lesions, respectively,
 9    are only slightly higher than the RfC derived for neurological effects (8 x 10~2 vs. 5 x 10~2 mg/m3),
10    indicating that effects in these organ systems may also be of concern. However, effects to pregnant
11    animals and the developing fetus may be of less concern as the RfCs for these types of effects (based
12    on decreased maternal weight gain and decreased male and female fetal weight, respectively) are
13    much higher than those derived for other organ systems.

     Table 2-5. Organ/system-specific RfCs and proposed overall RfC for 1,2,4-TMB
Effect
Hematological
Respiratory
Maternal
Developmental
Proposed
overall RfC
(Neurological)
Basis
Decreased clotting time
Increased inflammatory lung
lesions
Decreased maternal weight
gain
Decreased fetal weight (males
and females)
Decreased pain sensitivity
RfC
(mg/m3)
8 x 10"2
8 x 10"2
2
4
5 x 10"2
Exposure
description
Subchronic
Subchronic
Gestational
Gestational
Subchronic
Confidence
Low to medium
Low to medium
Low to medium
Low to medium
Low to medium
     2.1.5. Selection of the Proposed Overall Reference Concentration for 1,2,4-TMB
14          Neurotoxicity is the most consistently observed endpoint in the toxicological database for
15   1,2,4-TMB. According to EPA's Guidelines for Neurotoxicity Risk Assessment (U.S. EPA. 19981 many
16   neurobehavioral changes are regarded as adverse, and the observation of correlated and replicated
17   measures of neurotoxicity strengthen the evidence for a hazard. Decreased pain sensitivity,
18   measured as an increased latency to paw-lick in hot plate tests, represents an alteration in

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-16
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1   neurobehavioral function [U.S. EPA. 1998). Decreased pain sensitivity or decreased pain sensitivity
 2   following a foot shock challenge was observed in multiple studies across multiple exposure
 3   durations [Gralewicz and Wiaderna.  2001: Gralewicz etal.. 1997b: Korsak and Rydzynski. 1996:
 4   Korsak etal., 1995], and in the presence of other measures of altered neurobehavior, including
 5   impaired neuromuscular function and altered cognitive function. Additionally, neurological
 6   symptoms (e.g., hand tremble, weakness) were observed in worker populations exposed to
 7   complex VOC mixtures containing 1,2,4-TMB (notably, pain sensitivity has not been tested in
 8   humans), suggesting a consistency and coherency of neurotoxic effects in humans and animals
 9   following exposure to 1,2,4-TMB.
10          EPA's Guidelines for Neurotoxicity Risk Assessment (U.S. EPA. 1998] note that effects that are
11   reversible in minutes, hours, or days after the end of exposure and appear to be associated with the
12   pharmacokinetics of the agent and its presence in the body may be of less concern than effects that
13   persist for longer periods of time after the end of exposure. Pain sensitivity was observed to return
14   to control levels 2 weeks after termination of subchronic 1,2,4-TMB exposure  at 1,230 mg/m3 in
15   one study (Korsak and Rydzynski. 1996]. However, the Neurotoxicity Guidelines also indicate that
16   reversible effects occurring in occupational settings may be of high concern, particularly if they
17   diminish a person's ability to survive or adapt to the environment (U.S. EPA. 1998] (pg. 8]; such is
18   the case for exposure to 1,2,4-TMB in occupations with dangerous surroundings and/ or heavy
19   equipment, such as dockyard painters or asphalt workers.
20          In several short-term studies of TMBs, there is evidence indicating that decreased pain
21   sensitivity in the presence  of an additional environmental challenge (i.e., foot shock] is not rapidly
22   reversible and is not associated with clearance of the chemical from the body.  TMB isomers have
23   been observed to clear rapidly from blood and nervous tissues (Section B.2, Appendix B], and
24   decreased pain sensitivity following foot shock persisted 51 days  after termination of short-term
25   exposures (Wiaderna etal., 2002: Gralewicz and Wiaderna, 2001: Gralewicz etal., 1997b]. As
26   pointed out in A Review of the Reference Dose and Reference Concentration Process (U.S. EPA, 2002],
27   "[i]t is also important to keep in mind that effects that may initially appear to be reversible may re-
28   appear later or be predictive of later adverse outcomes." (pg. 4-16]. Additionally, the Neurotoxicity
29   Guidelines (U.S. EPA. 1998] state that "latent effects (those that become evident only after an
30   environmental challenge [e.g., in this case, footshock]] have a high level of concern." The hot plate
31   test is a relatively simple assessment that may not be sensitive enough to detect subtle changes
32   (U.S. EPA, 1998], suggesting that the large changes observed immediately after 1,2,4-TMB exposure
33   may represent gross effects. It is possible that, at longer durations after exposure, an environmental
34   challenge is necessary for the more subtle perturbations that persist to become manifest at a
35   detectable level. The latent decrements in pain sensitivity following foot shock appear to reflect a
36   lengthening of the numbing effects of foot shock following exposure to 1,2,4-TMB weeks earlier, as
37   the immediate increases in latency due to foot shock were unchanged by prior 1,2,4-TMB exposure.

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-17
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1   Although these measures may be complicated by less likely, but possible, effects on cognition, the
 2   results suggest that some aspect(s) of the altered pain sensitivity phenotype fail to resolve
 3   following termination of exposure. No environmental challenge was applied in the subchronic study
 4   by Korsak and Rydzyhski [1996]: such an experiment may have uncovered similar latent responses.
 5   Conversely, the short-term 1,2,4-TMB exposure studies testing pain sensitivity failed to analyze hot
 6   plate latency with a foot shock challenge shortly after exposure, as these evaluations only occurred
 7   at > 50 days post-exposure.
 8          Uncertainty regarding the reversibility of pain sensitivity in non-shocked rats at all tested
 9   1,2,4-TMB concentrations also exists. Reversibility of the pain sensitivity phenotype following
10   subchronic exposure was only tested at the highest concentration of 1,2,4-TMB used in any study
11   (i.e., 1,230 mg/m3). In multiple other tests of neurological function (including pain sensitivity
12   following a foot shock challenge), it was clearly shown that exposure to 1,2,4-TMB elicits nonlinear
13   effects when tested some period of time after exposure, with 1,230 mg/m31,2,4-TMB usually
14   resulting in no response or a substantially reduced response as compared to lower 1,2,4-TMB
15   concentrations (e.g., 492 mg/m3). Thus, from the data available, a determination regarding the
16   reversibility of 1,2,4-TMB-induced decreases in pain sensitivity at other concentrations (i.e., 492
17   mg/m3) at two weeks post-exposure cannot be made with confidence.
18          Although it is important to consider the potential for reversibility of neurological effects,
19   "for chronic lifetime exposures, designation of an effect as irreversible or reversible is academic, as
20   exposure is presumed to be lifetime (i.e., there is no post-exposure period)" (U.S. EPA. 2002] (pg. 3-
21   27). In other words, the nature of an RfC precludes the possibility of recovery of the critical effect.
22   This supports the choice of the principal study even were all aspects of the pain sensitivity
23   phenotype identified as transient, which, notably, does not appear to be the case. Taken as a whole,
24   the database supports the characterization of decreased pain sensitivity associated with exposure
25   to 1,2,4-TMB as being an effect of high concern. Given the consistency of observations from hot
26   plate tests with or without foot shock challenge across several studies from the same research
27   group using multiple durations of exposure in male Wistar rats, as well as the evidence and
28   biological plausibility of similarities in neurological effects between rats and humans, there is
29   strong evidence that neurotoxicity is the primary hazard associated with exposure to 1,2,4-TMB.
30   Based  on the above considerations, decreased pain sensitivity measured immediately after
31   subchronic exposure is identified as an adverse neurotoxic effect and thus is an appropriate effect
32   on which to base the RfC. Therefore, the candidate RfC for neurotoxicity based on decreased pain
33   sensitivity was selected as the RfC for 1,2,4-TMB.
34          A PODnEc of 15.8 mg/m3 for decreased pain sensitivity (Korsak and Rydzynski. 1996] was
35   used as the POD from which to derive the chronic RfC for 1,2,4-TMB (see Table 2-4). The
36   uncertainty factors (UFs), selected and applied in accordance with the procedures described in
37   EPA's A Review of the Reference Dose and Reference Concentration Processes [(U.S.  EPA. 2002]

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                   2-18
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   (Section 4.4.5 of the report)], were discussed previously in Section 2.1.3. Application of the
 2   composite UF of 300 to the PODHEc yields the following chronic RfC for 1,2,4-TMB:

 3                  RfC = PODHEC * UF = 15.8 mg/m3 4- 300 = 0.05 mg/m3 = 5 x 1Q-2 mg/m3
 4                                (rounded to one significant digit)

     2.1.6. Uncertainties in the Derivation of the Reference Concentration for 1,2,4-TMB
 5          As presented above, the UF approach, following EPA practices and RfC guidance [U.S. EPA.
 6   2002.1994b). was applied to the PODHEc in order to derive the chronic RfC for 1,2,4-TMB. Factors
 7   accounting for uncertainties associated with a number of steps in the analyses were adopted to
 8   account for extrapolation from animals to humans, a diverse human population of varying
 9   susceptibilities, duration of exposure, POD determination methodologies (NOAEL, LOAEL, or
10   BMDL), and database deficiencies.
11          The critical effect selected, decreased pain sensitivity, does not introduce substantial
12   uncertainty into the RfC calculation as selection of alternative hematological or respiratory effects
13   would result in similar RfCs that would be equivalent when rounding to one significant digit (i.e.,
14   2 x ID-2 mg/m3, see  Figure 2-2). Some uncertainty exists regarding the selection of the BMRs for use
15   in BMD modeling due to  the absence of information to determine the biologically significant level of
16   response associated with the endpoints. However in cases such as this, the selection of a BMR of 1
17   standard deviation for continuous endpoints is supported by EPA guidance [U.S. EPA. 2012b). In
18   addition, there is uncertainty in the estimated standard deviation for decreased pain sensitivity
19   [Korsak and Rydzynski. 1996). which was two- to threefold higher than that estimated in the
20   parallel evaluation of 1,2,3-TMB in the same publication. Given the lack of information concerning a
21   biologically significant level of response for pain sensitivity, the concurrently estimated standard
22   deviation was judged to be most relevant for characterizing this response to 1,2,4-TMB.
23          Uncertainty regarding the selection of particular models for individual endpoints exists as
24   selection of alternative models could decrease or increase the estimated POD and consequently, the
25   RfC. The selection criteria for model selection was based on a practical approach as described in
26   EPA's Benchmark Dose Technical Guidance (U.S. EPA. 2012bl Uncertainty may exist in the PBPK
27   model estimates of internal blood dose metrics for the rat, and subsequent HEC calculations for
28   humans, including parameter uncertainty, but such uncertainties would apply equally to all
29   endpoints.

     2.1.7. Confidence Statement for 1,2,4-TMB
30          A confidence level of high, medium, or low is assigned to the study used to derive the RfC,
31   the overall database, and the  RfC itself, as described in Section 4.3.9.2 of EPA's Methods for
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                  2-19
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   Derivation of Inhalation Reference Concentrations and Application of Inhalation Dosimetry [U.S. EPA.
 2   1994b).
 3          Confidence in the study from which the critical effect was identified, Korsak and Rydzyhski
 4   [1996] is low to medium. The study is a peer-reviewed study that utilized three dose groups plus
 5   untreated controls and employed an appropriate number of animals per dose group. However,
 6   sources of uncertainty exist that reduce confidence in this study.
 7          One area of uncertainty regarding this study is the lack of reported actual concentrations.
 8   However, as the methods by which the test atmosphere was generated and analyzed were reported
 9   in sufficient detail, and given the fact that this laboratory has used this methodology in subsequent
10   studies [Korsak et al.. 2000a. b) and achieved appropriate actual concentrations (i.e., within 10% of
11   target concentrations), the concern regarding the lack of reported actual concentrations is minimal.
12   The critical effect on which the RfC is based is well-supported as the weight of evidence for
13   1,2,4-TMB-induced neurotoxicity is coherent across species (i.e., human and rat) and consistent
14   across multiple exposure durations (i.e., acute, short-term, and subchronic) [Gralewicz and
15   Wiaderna. 2001: Chen etal.. 1999: WiadernaetaL 1998: Gralewicz etal.. 1997b: Gralewicz etal..
16   1997a: Korsak and Rydzynski. 1996: Norseth etal.. 1991).
17          The database for 1,2,4-TMB includes acute, short-term, subchronic, and developmental
18   toxicity studies in rats and mice. However, confidence in the overall database is low to medium
19   because it lacks chronic, multi-generation reproductive/developmental, and developmental
20   neurotoxicity studies, and the studies supporting the critical effect predominantly come from the
21   same research institute. The overall confidence in the RfC for 1,2,4-TMB is low to medium.

     2.2.  Inhalation Reference Concentration for Effects Other Than Cancer
           for 1,2,3-TMB

     2.2.1. Identification of Studies and Effects for Dose-Response Analysis for 1,2,3-TMB
22          The nervous, hematological, and respiratory systems are the primary targets of inhaled
23   1,2,3-TMB in humans and experimental animals, and effects in these systems have been identified
24   as hazards following inhalation exposure to 1,2,3-TMB. Although literature exists on the effects of
25   1,2,3-TMB exposure in humans, including neurological, hematological, and respiratory toxicities, no
26   human studies are available that would allow for dose-response analysis. The human studies
27   evaluated TMB exposures occurring as complex solvents or VOC mixtures, and this consideration
28   along with other uncertainties including high imprecision in effect measures due to low statistical
29   power, lack of quantitative exposure assessment, and lack of control for co-exposures, limit their
30   utility in derivation of quantitative human health toxicity values. However, these studies provide
31   supportive evidence for the neurological, hematological, and respiratory toxicity of TMB isomers in
32   humans and indicate a coherency of effects in both humans and laboratory animals.

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                 2-20
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1          Several studies investigating 1,2,3-TMB effects in experimental animal models were
 2   identified in the literature. No chronic studies were available, although several acute, short-term,
 3   and subchronic studies were identified. 1,2,3-TMB-induced toxicity was observed across several
 4   organ systems in two subchronic studies by Korsak et al. [2000b] and Korsak and Rydzyhski
 5   [1996]. These were the only subchronic studies identified in the peer-reviewed literature. Data
 6   from these studies pertaining to the primary hazards observed in humans and animals identified
 7   previously in Chapter  1 (neurological, hematological, and respiratory toxicity) were considered as
 8   candidate critical effects for the purpose of determining the point of departure (POD) for derivation
 9   of the inhalation RfC for 1,2,3-TMB. Neurotoxicity was also observed in both acute and short-term
10   inhalation studies and respiratory toxicity was also observed in acute studies. However, the high
11   concentrations used in acute studies and the short exposure durations employed in both acute and
12   short-term studies limit their applicability for quantitation of chronic human health effects.
13   Nevertheless, as with the human in which subjects were exposed to mixtures containing 1,2,3-TMB,
14   these studies provide qualitative information regarding the consistency and coherency of these
15   effects across the 1,2,3-TMB database.
16          The two subchronic studies by Korsak et al. (2000b) and Korsak and Rydzyhski (1996) are
17   adequate for dose-response analysis. Both studies exposed rats, a common model for human
18   response, by inhalation, tol,2,3-TMB (reported as > 97% pure [impurities not reported]). The
19   studies used three exposure levels spaced two- to threefold apart, facilitating dose-response
20   analysis and utilized sham-exposed controls. The subchronic durations of exposure were suitable
21   for the effects under evaluation. In addition, the persistence of some outcomes after termination of
22   exposure was investigated. Typical numbers of animals per exposure group for subchronic studies
23   were used, at least 10/group. Regarding exposure characterization, Korsak et al. (2000b) reported
24   actual concentrations, as measured by gas chromatography, to be within 10% of target
25   concentrations. This increases the confidence in the overall evaluation and adequacy of this study.
26   Although Korsak and Rydzyhski (1996) do not report actual, measured concentrations, this study
27   uses the same exposure methodology as Korsak et al. (2000b): suggesting that it is likely that the
28   actual concentrations in this study were also within 10% of target concentrations. Target and actual
29   concentrations for these studies are presented in Table 2-6.
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                  2-21
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
     Table 2-6. Target and actual exposure concentrations used in BMD modeling of
                 1,2,3-TMB endpoints considered for the derivation of the RfC
Reference
Korsak and
Rydzyriski (1996)
Korsak et al.
(2000b)
Species/
sex
Rat, male
Rat, male
Rat, female
Target exposure
concentration (mg/m3)
123
492
1,230
123
492
1,230
123
492
1,230
Actual exposure
concentration (mg/m3)
n/a
n/a
n/a
128
523
1,269
128
523
1,269
 1           These two subchronic studies examined 1,2,3-TMB-induced toxicity in multiple organ
 2    systems and the neurological, hematological, and respiratory endpoints that demonstrated
 3    statistically significant increases or decreases relative to control were considered for the derivation
 4    of the RfC for 1,2,3-TMB (Table 2-7). These endpoints included decreased pain sensitivity in male
 5    rats [Korsak and Rydzynski, 1996], and decreased RBCs and increased reticulocytes in male rats,
 6    decreased segmented neutrophils and increased lymphocytes in male and female rats, and
 7    increased inflammatory lung lesions in female rats [Korsak et al.. 2000b). Changes in liver organ
 8    weights and clinical chemistry parameters from Korsak et al. [2000b] were not further considered
 9    due to the lack of accompanying hepatocellular histopathological alterations in exposed animals.
10    Changes in splenic organ weights were similarly not considered further due to a lack of any
11    observed histopathological changes in that organ. Increases in reticulocytes in females were not
12    further considered due to non-monotonicity in response (increases in high concentration animals
13    that were not statistically significant). Increased lymphocytes were excluded from further
14    consideration due to the unusually high standard deviations reported in the high-concentration
15    group.
                 This document is a draft for review purposes only and does not constitute Agency policy.

                                                 2-22
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                          Toxicological Review of Trimethylbenzene

     Table 2-7. Endpoints considered for the derivation of the RfC for 1,2,3-TMB
Endpoint
Species/
sex
Exposure concentration (mg/m3)3
0
123
492
1,230
Neurological endpoints
Decreased pain sensitivity (measured as
latency to paw-lick in seconds)b
Rat, male
9.7 ±2.1
(n=30)
11.8 ±3.8*
(n = 20)
16.3 ±6.3°
(n = 10)
17.3 ±3.4**
(n = 10)
Hematological endpoints
Decreased RBCs (106/mm3)d
(10s cells per 100 ul)
Decreased segmented neutrophils (%)d
Increased reticulocytes (%)d
Rat, male
Rat, male
Rat, female
Rat, male
9.49 ± 2.03
(n = 10)
24.8 ±4.5
(n = 10)
23.1 ±6.1
(n = 10)
2.8 ±1.3
(n = 10)
10.25 ± 1.29
(n = 10)
25.4 ±5.8
(n = 10)
19.7 ±3.4
(n = 10)
2.1 ±1.7
(n = 10)
10.11 ±1.27
(n = 10)
20.7 ±5.8
(n = 10)
16.4 ±4.2*
(n = 10)
3.8 ±2.1
(n = 10)
8.05 ± 1.38*
(n = 10)
17.7 ±8.3*
(n = 10)
11.9 ±7.1**
(n = 10)
4.5 ± 1.8*
(n = 10)
Respiratory Endpoints
Increased inflammatory lung lesionsd
Rat, female
e
(n = 10)
e
(n =10)
e
(n = 10)
e
(n = 10)
     p<0.05;   p<0.01.
     aValues are expressed as mean ± 1SD. Korsak and Rydzyriski (1996) does not explicitly state that the reported
        measures of variance in Table 1 of that reference are standard deviations. However, independent analysis
        conducted by EPA confirms that the reported measure of variance are standard deviations.
     b Adapted from Korsak and Rydzyriski (1996)
     c Level of significance not reported in Table 1 from Korsak and Rydzyriski (1996), however the results of an ad-hoc t-
        test (performed by EPA) indicated significance at p < 0.01.
     d Adapted from Korsak et al. (2000b)
     e Incidences for exposure groups not reported; however, based on qualitative information reported in the study
        (i.e., that female rats exhibited a statistically significant increase in inflammatory lung lesions at 492 mg/m3),
        a NOAEL of 123 mg/m3 was identified.


1           Impaired neuromuscular function and coordination, measured as performance on the
2    rotarod apparatus, was also observed in rats exposed to 1,2,3-TMB. See Section 2.1.1 for a detailed
3    discussion of the uncertainties surrounding the use of this endpoint for derivation of an RfC. Due to
4    these uncertainties, this endpoint was excluded from consideration for the derivation of the RfC for
5    1,2,3-TMB.
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                    2-23
                                                                    DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

     2.2.2.  Methods of Analysis for 1,2,3-TMB
 1          As discussed above in Section 2.2.1, endpoints observed in Korsak et al. [2000b] and Korsak
 2   and Rydzynski [1996] that demonstrated statistically significant (p < 0.05 level) increases or
 3   decreases relative to control for at least one exposure group were considered for the derivation of
 4   the RfC for 1,2,3-TMB; these effects are listed in Table 2-7. This assessment used the BMD approach,
 5   when possible, to estimate a POD for the derivation of an RfC for 1,2,3-TMB (see Section C.I of
 6   Appendix C for detailed methodology). The BMD approach involves fitting a suite of mathematical
 7   models to the observed dose-response data using EPA's BMDS (version 2.2). Each fitted model
 8   estimates a BMD and its associated BMDL corresponding to a selected BMR. For continuous data
 9   (i.e., decreased pain sensitivity, decreased RBCs, decreased segmented neutrophils, increased
10   reticulocytes) from the Korsak and Rydzynski (1996) and Korsak et al. (2000b) studies, no
11   information is available regarding the change in these responses that would be considered
12   biologically significant, and thus a BMR equal to a 1 standard deviation change in control mean was
13   used in modeling the endpoints, consistent with the Benchmark Dose Technical Guidance Document
14   (U.S. EPA. 2012b). The estimated BMDL is then used as the POD for deriving the RfC (Table 2-8).
15          The suitability of the above methods to determine a POD is dependent on the nature of the
16   toxicity database for a specific chemical. Some endpoints for 1,2,3-TMB were  not modeled for a
17   variety of reasons, including responses only in the high exposure group with no changes in
18   responses in lower exposure groups (e.g., decreased RBCs) and absence of incidence data (e.g.,
19   increased inflammatory lung lesions). In cases where BMD modeling was not feasible, the
20   NOAEL/LOAEL approach was used to identify a POD. Additionally, for decreased pain sensitivity,
21   the reported SD of 3.4 in the high exposure group resulted in an inability of the variance power
22   model to fit the data adequately. For this reason, the high exposure group was dropped in order to
23   facilitate model fitting. Detailed modeling results are provided in Section C.I of Appendix C.
24          Because an RfC is a toxicity value that assumes continuous human inhalation exposure over
25   a lifetime, data derived from inhalation studies in animals need to be adjusted to account for the
26   noncontinuous exposures used in these studies. In the Korsak et al. (2000b) and Korsak and
27   Rydzynski (1996) studies, rats were exposed to 1,2,3-TMB for  6 hours/day, 5 days/week for 3
28   months. Because no PBPK model exists for  1,2,3-TMB, the duration-adjusted PODs for effects in rats
29   were calculated as follows:
30          PODADj (mg/m3) = POD (mg/m3) x hours exposed per day/24 hours x days exposed
31          per week/7 days
32          Therefore, for example, for decreased pain sensitivity from Korsak and Rydzynski (1996).
33   the PODADj would be calculated as follows:
34          PODADj (mg/m3) = 97.19 mg/m3x  6 hours/24 hours x 5 days/7 days

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                 2-24
                                                               DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1          PODADj (mg/m3) = 17.36 mg/m3
 2          Because the majority of the selected endpoints for consideration as the critical effect
 3    (decreased pain sensitivity, decreased RBCs, decreased segmented neutrophils, increased
 4    reticulocytes) result primarily from systemic distribution of 1,2,3-TMB, and no available PBPK
 5    model exists for 1,2,3-TMB, the human equivalent concentration (HEC) for 1,2,3-TMB was
 6    calculated by the application of the dosimetric adjustment factor (DAF) for systemically acting gases
 7    (i.e., Category 3 gases), in accordance with the U.S. EPA RfC Methodology (U.S. EPA. 1994bl
 8    Additionally, although the observation of lung lesions would normally indicate portal-of-entry
 9    effects, the observation that the overwhelming majority of 1,2,3-TMB-induced effects are systemic
10    in nature supports the determination that 1,2,3-TMB is a Category 3 gas. Other factors also support
11    that 1,2,3-TMB is a systemically-acting toxicant, including the isomer's relatively low water-
12    solubility and non-reactivity. Gases with these properties are expected to preferentially distribute to
13    the lower regions of the respiratory tract where larger surface areas and thin alveolar-capillary
14    boundaries facilitate uptake. Respiratory absorption of 1,2,3-TMB into the bloodstream has been
15    observed to be relatively high (~60%) following inhalation exposures to humans (Tarnbergetal..
16    1996]. Therefore, increased inflammatory lung lesions are assumed to result from systemic
17    distribution of 1,2,3-TMB in the bloodstream of exposed animals. DAFs are ratios of animal and
18    human physiologic parameters, and are dependent on the nature of the contaminant (particle or
19    gas) and the target site (e.g., respiratory tract or remote to the portal-of-entry [i.e., systemic]) (U.S.
20    EPA. 1994b). For gases with systemic effects, the DAF is expressed as the ratio between the animal
21    and human blood:air partition coefficients:
22          DAF = (Hb/g)A/(Hb/g)H
23          where:
24          (Hb/g)A = the animal bloochair partition coefficient
25           (Hb/g)H = the human bloochair partition coefficient
26          DAF = 62.6 (Tarnberg and Tohanson. 19951/66.5 (Meulenberg and Viiverberg. 2000)
27          DAF = 0.94
28          In cases where the animal blood:air partition coefficient is lower than the human value,
29    resulting in a DAF < 1, the calculated value is used for dosimetric adjustments (U.S. EPA. 1994b).
30    For example, the HEC for decreased pain sensitivity reported in Korsak and Rydzyhski (1996) is
31    calculated as follows:
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                  2-25
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1           PODHEc = PODADj (mg/m3) x DAF
 2           PODHEC = PODADj (mg/m3) x 0.94
 3           PODHEC = 17.36 mg/m3 x 0.94
 4           PODHEc = 16.32 mg/m3
 5           Table 2-8 presents the calculated HECs for the candidate critical effects, selected
 6    uncertainty factors (UFs), and the resulting derivation of candidate RfCs from the two subchronic
 7    toxicity studies [Korsak et al.. 2000b: Korsak and Rydzynski. 1996).

     Table 2-8. Summary of derivation of points of departure for 1,2,3-TMB
Endpoint/Reference
Species/sex
Model; BMR or
NOAEL/LOAEL
POD
(mg/m3)
Candidate
PODADJ
(mg/m3)
Candidate
PODHEC
(mg/m3)
Neurological endpoints
Decreased pain sensitivity
(Korsak and Rydzyriski, 1996)

Rat, male
Linear; 1 SD
97.19
17.36
16.32
Hematological endpoints
Decreased RBCs
(Korsak etal.. 2000b)
Increased segmented
neutrophils
(Korsak etal., 2000b)

Increased reticulocytes
(Korsak etal., 2000b)

Rat, male
Rat, male
Rat, female
Rat, male
NOAELa
Exponential M2;
1SD
Hill; 1 SD
Linear; 1 SD
523
534.81
99.21
652.90
93.39
95.50
17.72
116.58
87.79
89.77
16.66
109.58
Respiratory endpoints
inflammatory lung lesions
(Korsak etal., 2000b)
Rat, male
NOAELa
128
22.86
21.49
     ' No model was able to fit data adequately, or data were not modeled. NOAEL/LOAEL method used to identify
        a POD.
 9
10
2.2.3.  Derivation of Candidate RfC Values for 1,2,3-TMB
       Under EPA's A Review of the Reference Dose and Reference Concentration Processes [[U.S.
EPA. 2002] §4.4.5], also described in the Preamble, five possible areas of uncertainty and variability
were considered in deriving the candidate RfC values for 1,2,4-TMB. An explanation of these five
                 This document is a draft for review purposes only and does not constitute Agency policy.

                                                 2-26
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1   possible areas of uncertainty and variability and the values assigned to each as a designated
 2   uncertainty factor (UF) to be applied to the candidate PODHEc are as follows:
 3          An interspecies uncertainty factor, UFA, of 3 (101/2 = 3.16, rounded to 3) was applied to
 4   account for uncertainty in characterizing the toxicokinetic and toxicodynamic differences between
 5   rats and humans following inhalation exposure to 1,2,3-TMB. In this assessment, the use of a DAF to
 6   extrapolate external concentrations from rats to humans reduces toxicokinetic uncertainty in
 7   extrapolating from the rat data, but does not account for the possibility that humans may be more
 8   sensitive to 1,2,3-TMB than rats due to toxicodynamic differences. A default UFA of 3 was thus
 9   applied to account for this remaining toxicodynamic and residual toxicokinetic uncertainty not
10   accounted for in the DAF.
11          An intraspecies uncertainty factor, UFH, of 10 was applied to account for potentially
12   susceptible individuals in the absence of data evaluating variability of response in the human
13   population following inhalation of 1,2,3-TMB. No information is currently available to predict
14   potential variability in human susceptibility, including variability in the expression of enzymes
15   involved in 1,2,3-TMB metabolism.
16          A LOAEL to  NOAEL uncertainty factor, UFL, of 1 was applied because the current approach is
17   to address this factor as one of the considerations in selecting a BMR for BMD modeling. In this
18   case, a BMR equal to a 1 standard deviation change in the control mean for modeled endpoints was
19   selected under the assumption that this BMR represents a minimal, biologically significant change
20   for these effects. For endpoints that could not be modeled, a LOAEL to NOAEL uncertainty factor of
21   1 was applied as a NOAEL was used.
22          A subchronic to chronic uncertainty factor, UFs, of 3 (101/2 = 3.16, rounded to 3) was applied
23   to account for extrapolation from a subchronic exposure duration study to derive  a chronic RfC. The
24   3-fold uncertainty factor is applied to the POD identified from the subchronic study on the
25   assumption that effects observed in a similar chronic study would be observed at lower
26   concentrations for a number of possible reasons, including potential cumulative damage occurring
27   over the duration of the chronic study or an increase in the magnitude or severity of effect with
28   increasing duration of exposure. For example, in the case of neurotoxicity, chronic exposures may
29   overwhelm the adaptive responses observed after termination of subchronic exposure, potentially
30   resulting in more severe and/or irreversible changes in neurological function. A full subchronic to
31   chronic uncertainty factor of 10 was not applied in this case as there was evidence of reversibility
32   of not only neurotoxic effects, but also hematological effects in rats exposed to 1,2,4-TMB for
33   subchronic durations. Also, the respiratory effects appeared to be inflammatory in nature. Although
34   reversibility was not investigated for these endpoints, it is possible that adaptive mechanisms may
35   alleviate these effects following the termination of exposure.
36          A database uncertainty factor, UFo, of 3 (101/2 = 3.16, rounded to 3) was applied to account
37   for database deficiencies. Strengths of the database include the two well-designed subchronic

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-27
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1   studies that observe exposure-response effects in multiple organ systems (i.e., neurological,
 2   hematological, and respiratory effects) in Wistar rats exposed to 1,2,3-TMB via inhalation.
 3   However, the lack of a either a multi-generational reproductive/developmental toxicity study or a
 4   developmental toxicity study investigating effects due to 1,2,3-TMB exposure is a weakness of the
 5   database. Normally, the lack of both of these types of studies in a toxicity database would warrant
 6   the application of a full, 10-fold UFo in accordance with EPA's A Review of the Reference Dose and
 7   Reference Concentration Processes [2002]. Although there is no developmental toxicity study for
 8   1,2,3-TMB, Saillenfait et al. [2005] investigates the developmental toxicity of the other two TMB
 9   isomers (1,2,4-TMB and 1,3,5-TMB] and observes developmental toxicity at levels much higher
10   than those eliciting neurotoxicity, hematotoxicity, and respiratory toxicity in adult animals (Korsak
11   studies]. Given that toxic effects were observed at lower concentrations in adult animals exposed
12   1,2,4-TMB and 1,3,5-TMB compared with rats exposed in utero and the similarities in toxicity
13   profiles amongst the three isomers, it is unlikely that the inclusion of a developmental toxicity study
14   for 1,2,3-TMB would result in a POD that is lower than the POD associated with neurotoxicity for
15   this isomer. Thus, the application of an UF to account for the lack of a developmental toxicity study
16   is not warranted.
17          EPA's A Review of the Reference Dose and Reference Concentration Processes [U.S. EPA. 2002]
18   also recommends that the database uncertainty factor take into consideration whether there is
19   concern from the available toxicology database that the developing  organism may be particularly
20   susceptible to effects in specific organ systems. TMBs (unspecified isomer] are able to cross the
21   placenta [Cooper et al.. 2001: Dowtyetal.. 1976]: therefore, as neurotoxicity is observed in adult
22   animals, there is concern that exposure to 1,2,3-TMB may result in neurotoxicity in the developing
23   organism. EPA's  Guidelines for Neurotoxicity Risk Assessment [U.S. EPA. 1998] identifies specific
24   effects observed in adult animals (e.g., cognitive and motor function] that can also affect the
25   developing organism exposed in utero. The Neurotoxicity Guidelines [U.S. EPA, 1998] also indicate
26   that neurotoxicants may have greater access to the nervous system  in developing organisms due to
27   an incomplete blood-brain barrier and immature metabolic detoxifying pathways. Lastly, EPA's A
28   Review of the Reference Dose and Reference Concentration Processes  [U.S. EPA. 2002] also states  that
29   effects that may be mild or reversible in adults may produce more robust or permanent in offspring
30   following developmental exposures. Therefore, there is some concern that the lack of a
31   developmental neurotoxicity study is a deficiency in the database and that the inclusion of such a
32   study would potentially result in a lower POD than the POD for neurotoxicity identified from the
33   available 1,2,3-TMB toxicity database. In summary, a 3-fold database UF was applied to  account for
34   the lack of both a multi-generation reproductive/developmental toxicity study and a developmental
35   neurotoxicology study in the available database for 1,2,3-TMB.
36          Table 2-9 is a continuation of Table 2-8, and summarizes the application of UFs to each POD
37   to derive a candidate value for each data set. The candidate values presented in Table 2-9 are

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                   2-28
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
1
2
3
4
5
                                                    Toxicological Review of Trimethylbenzene

preliminary to the derivation of the organ/system-specific values. These candidate values are
considered individually in the selection of a representative inhalation reference value for a specific
hazard and subsequent overall RfC for 1,2,3-TMB. Figure 2-2 presents graphically these candidate
values, uncertainty factors, and points of departure, with each bar corresponding to one data set
described in Tables 2-8 and 2-9.

Table 2-9. Effects and corresponding derivation of candidate RfC values for
            1,2,3-TMB
Endpoint/Reference
HEC
(mg/m3)3

UFA

UFH

UFL

UFS

UFD
Composite
UF
Candidate
value (mg/m3)b
Neurological endpoints
Decreased pain sensitivity
(Korsakand Rydzyriski, 1996)



16.32


3


m


1


3


3


300


5.44 x 10"2


Hematological endpoints
Decreased RBCs

(Korsak et al., 2000b)
Decreased segmented
neutrophils,
(Korsak etal.,2000b)
Increased reticulocytes
(Korsak etal.,2000b)



male

female



87.79

89.77

16.66
109.58


3

3

3
3


m

10

10
10


1

1

1
1


3

3

3
3


3

3

3
3


300

300

300
300


2.93 x 10"1

2.99 x 10"1

5.55 x 10 2
3.65 x 10"1

Respiratory endpoints
Increased inflammatory lung lesions
(Korsak etal.,2000b)
21.49
3
10
1
3
3
300
7.16 x 10"2
     aHuman equivalent concentration.
     bAs calculated by application of uncertainty factors, not rounded to 1 significant digit.
                 This document is a draft for review purposes only and does not constitute Agency policy.

                                                   2-29
                                                                  DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene
       6  ro
                   Inflammatory lung lesions
                    (Korsaketal, 2000b)
                         T" Reticulocytes
                       (Korsaketal., 2000b)
         —   4- Segmented neutrophils,females
         .y        (Korsaketal., 2000b)
              4- Segmented neutrophils, males
                  (Korsaketal., 2000b)
                            v|/ RBCs
                       (Korsaketal., 2000b)
                     \|/ Pain Sensitivity
                 (Korsakand Rydzynski, 1996)
  Composite UF
A Candidate value
• POD(HEC)
                                      0.001     0.01      0.1       1       10
                                                         Concentrations (mg/m3)
 100
1000
     Figure 2-2. Candidate RfC values with corresponding POD and composite UF for
                 1,2,3-TMB.
     2.2.4. Derivation of Organ/System Specific Reference Concentrations for 1,2,3-TMB
 1          Table 2-10 distills the candidate values from Table 2-9 into a single value for each organ or
 2   system. The single RfC value selected for a particular organ system was preferably chosen using
 3   biological and toxicological information regarding that endpoint. If no compelling biological
 4   information exists with which to select the primary hazard, the lowest RfC value for that organ
 5   system was selected. These organ- or system-specific reference concentrations may be useful for
 6   subsequent cumulative risk assessments that consider the combined effect of multiple agents acting
 7   at a common site. The individual organs and systems for which specific RfC values were derived
 8   were the neurological, hematological, and respiratory systems. The RfC value for the neurological
 9   system, based on decreased pain sensitivity, was selected for the proposed overall RfC for
10   1,2,3-TMB (see Section 2.2.5 for details). The RfC values for the hematological and respiratory
11   systems, based on decreased segmented neutrophils and increased inflammatory lung lesions, were
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                   2-30
                                                                  DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene
 1   only marginally higher than the RfC derived for neurological effects (6 x 1Q-2 and 7 x 1Q-2 mg/m3 vs.
 2   5 x IQ-2 mg/m3), indicating that effects in these organ systems may also be of concern.

     Table 2-10. Organ/system-specific RfCs and proposed overall RfC for 1,2,3-TMB
Effect
Hematological
Respiratory
Proposed
overall RfC
(Neurological)
Basis
Decreased segmented
neutrophils
Increased inflammatory
lung lesions
Decreased pain sensitivity
Rfc (mg/m3)
6 x 10"2
7 x 10"2
5 x 10"2
Exposure
description
Subchronic
Subchronic
Subchronic
Confidence
Low to medium
Low to medium
Low to medium
     2.2.5. Selection of the Proposed Overall Reference Concentration for 1,2,3-TMB
 3          Neurotoxicity is the most consistently observed endpoint in the toxicological database for
 4   1,2,3-TMB. According to EPA's Guidelines for Neurotoxicity Risk Assessment [U.S. EPA, 1998], many
 5   neurobehavioral changes are regarded as adverse, and the observation of correlated and replicated
 6   measures of neurotoxicity strengthen the evidence for a hazard. Decreased pain sensitivity, as
 7   measured as an increased latency to paw-lick in hot plate tests, represents an alteration in
 8   neurobehavioral function [U.S. EPA. 1998]. Decreased pain sensitivity or decreased pain sensitivity
 9   following a foot shock challenge was observed in two studies investigating short-term and
10   subchronic exposure durations [Wiadernaetal.. 1998: Korsak and Rydzynski. 1996] and in the
11   presence of other metrics of altered neurobehavior, including impaired neuromuscular function
12   and altered cognitive function. Additionally, neurological symptoms (e.g., hand tremble, weakness]
13   are observed in human worker populations exposed to complex VOC mixtures containing
14   1,2,3-TMB (notably, pain sensitivity has not been tested in humans] indicating a consistency and
15   coherency of neurotoxic effects in humans and animals following exposure to 1,2,3-TMB.
16          See Section 2.1.5 for a detailed discussion of U.S. EPA's Guidelines for Neurotoxicity Risk
17   Assessment (U.S. EPA. 1998] and the use of reversible and/or latent neurotoxicological endpoints in
18   the derivation of reference values. The issues pertaining to the observed 1,2,3-TMB neurotoxicity
19   are the same as those identified for 1,2,4-TMB. For example, although 1,2,3-TMB-induced pain
20   sensitivity was observed to return to control levels two weeks after termination of subchronic
21   inhalation exposure in one study (Korsak and Rydzynski. 1996]. the Neurotoxicity Guidelines note
22   that reversible effects occurring in occupational settings may be of high concern, particularly if they

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                 2-31
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1   diminish a person's ability to survive or adapt to the environment [U.S. EPA. 1998] (pg. 8).
 2   Additionally, the "designation of an effect as irreversible or reversible is academic, as exposure is
 3   presumed to be lifetime (i.e., there is no post-exposure period)" [U.S. EPA. 2002] (pg. 3-27]. In other
 4   words, the nature of an RfC precludes the possibility of recovery from the critical effect Lastly, the
 5   issues surrounding the use of an environmental challenge (i.e., foot shock] in short-term
 6   neurotoxicity studies of 1,2,3-TMB are the same as those discussed for 1,2,4-TMB in Section 2.1.5.
 7          Taken as a whole, the database supports the characterization of decreased pain sensitivity
 8   associated with exposure to 1,2,3-TMB as being an effect of high concern. Given the consistency of
 9   observations from hot plate tests with or without foot shock challenge across several studies from
10   the same research group using multiple durations of exposure in male Wistar rats, as well as the
11   evidence and biological plausibility of similarities in neurological effects between rats and humans,
12   there is strong evidence that neurotoxicity is the primary hazard associated with exposure to
13   1,2,3-TMB. Based on these considerations, decreased pain sensitivity observed immediately after
14   subchronic exposure is identified as an adverse neurotoxic effect and thus is an appropriate effect
15   on which to base the RfC. Therefore, the candidate RfC for neurotoxicity based on decreased pain
16   sensitivity was selected as the proposed overall RfC for 1,2,3-TMB.
17          A PODHEc of 16.3 mg/m3 for decreased pain sensitivity (Korsak and Rydzynski. 1996] was
18   used as the POD to derive the chronic RfC for 1,2,3-TMB. The uncertainty factors (UFs], selected and
19   applied in accordance with the procedures described in EPA's A Review of the Reference Dose and
20   Reference Concentration Processes (U.S. EPA. 2002] (Section 4.4.5 of the report], were discussed
21   previously in Section 2.2.3. Application of this composite UF of 300 to the PODnEc yields the
22   following chronic RfC for 1,2,3-TMB:
23          RfC = PODHEC * UF = 16.3 mg/m3  4- 300 = 0.05 mg/m3 = 5 x 10 2 mg/m3 (rounded to one
24          significant digit)

     2.2.6. Uncertainties in the Derivation of the Reference Concentration for 1,2,3-TMB
25         As presented above, the UF approach following EPA practices and RfC guidance (U.S. EPA.
26   2002.1994b]. was applied to the PODHEc in order to derive the chronic RfC for 1,2,3-TMB. Factors
27   accounting for uncertainties associated with a number of steps in the analyses were adopted to
28   account for extrapolation from animals to humans, a diverse human population of varying
29   susceptibilities, duration of exposure, POD determination methodologies (NOAEL, LOAEL, or
30   BMDL], and database deficiencies.
31         The critical effect selected, decreased pain sensitivity, does not introduce substantial
32   variability into the RfC calculation as selection of alternative hematological  or respiratory effects
33   would result in similar RfCs that would be equivalent when rounding to one significant digit (i.e.,
34   2 x ID-2 mg/m3, see Figure 2-4]. Some uncertainty exists regarding the selection of the BMRs for use
35   in BMD modeling due to the absence of information to determine the biologically significant level  of
                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-32
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1   response associated with the endpoints. However in cases such as this, the selection of a BMR of
 2   1 standard deviation for continuous endpoints is supported by EPA guidance [U.S. EPA. 2012b]. In
 3   addition, there is uncertainty in the estimated standard deviation for decreased pain sensitivity
 4   [Korsak and Rydzynski, 1996], which was two- to threefold lower than that estimated in the
 5   parallel evaluation of 1,2,4-TMB in the same publication (see Section 2.1.6.). Given the lack of
 6   information concerning a biologically significant level of response for pain sensitivity, the
 7   concurrently estimated standard deviation was judged to be most relevant for characterizing this
 8   response to 1,2,3-TMB.
 9         Uncertainty regarding the selection of particular models for individual endpoints exists as
10   selection of alternative models could decrease or increase the estimated POD and consequently, the
11   RfC. The criteria for model selection was based on a practical approach as described in EPA's
12   Benchmark Dose Technical Guidance [U.S. EPA, 2012b]. Uncertainty may exist in the default
13   dosimetry methods used to calculate HEC estimates, but such uncertainties would apply equally to
14   all endpoints.

     2.2.7. Confidence Statement for 1,2,3-TMB
15          Confidence in the study from which the critical effect was identified, Korsak and Rydzynski
16   [1996] is low to medium. The study is a peer-reviewed study that utilized three dose groups plus
17   untreated controls, employed an appropriate number of animals per dose group, and appropriately
18   performed statistical analyses. However, sources of uncertainty exist that reduce confidence in this
19   study.
20          One area of uncertainty regarding this study is the lack of reported actual concentrations.
21   However, as the methods by which the test atmosphere was generated and analyzed were reported
22   in sufficient detail, and given the  fact that this laboratory has used this methodology in subsequent
23   studies [Korsak et al.. 2000a. b] and achieved appropriate actual concentrations (i.e., within 10% of
24   target concentrations], the concern regarding the lack of reported actual concentrations is minimal.
25   Another source of uncertainty is the fact that Korsak and Rydzynski (1996] does not explicitly state
26   that the reported measures of variance in Table 1 of that reference are standard deviations.
27   However, careful analysis of the reported levels of variance and magnitude of statistical significance
28   reported indicate that the measures of variance are standard deviations. Supporting this
29   conclusions is the observation that all other papers by Korsak et al. (2000a, b; 1997: 1995] report
30   variance as standard deviations. The critical effect on which the RfC is based is well-supported as
31   the weight of evidence for 1,2,3-TMB-induced neurotoxicity is coherent across multiple animals
32   species (i.e., mouse, and rat] and consistent across multiple exposure durations (i.e., acute, short-
33   term, and subchronic] (Lutz etal.. 2010: Wiadernaetal.. 1998: Korsak and Rydzynski. 1996].
34          The database for 1,2,3-TMB includes acute, short-term, and subchronic toxicity studies in
35   rats and mice. However, confidence in the database is low to medium because it lacks chronic,

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                   2-33
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   multi-generation reproductive/developmental, developmental toxicity, or developmental
 2   neurotoxicity studies, and the studies supporting the critical effect predominantly come from the
 3   same research institute. The overall confidence in the RfC for 1,2,3-TMB is low to medium.

     2.3.  Inhalation Reference Concentration for Effects Other Than Cancer
           for 1,3,5-TMB

     2.3.1. Identification of Studies and Effects for Dose-Response Analysis for 1,3,5-TMB
 4          The nervous, hematological, and respiratory systems are the primary targets of toxicity for
 5   inhaled 1,3,5-TMB in humans, whereas the nervous system in adults, pregnant females, and
 6   developing organism are the primary targets of inhaled 1,3,5-TMB in experimental animals. Effects
 7   in these systems have been identified as hazards following inhalation exposures to 1,3,5-TMB.
 8   Although literature exists on the effects of 1,3,5-TMB exposure in humans, including neurological,
 9   hematological, and respiratory toxicities, no human studies are available that would allow for dose-
10   response analysis. The human studies evaluated TMB exposures occurring as complex solvents or
11   VOC mixtures, and this consideration along with similar uncertainties as discussed for 1,2,4-TMB
12   and 1,2,3-TMB limit their utility in derivation of quantitative human health toxicity values. As for
13   the other two isomers, the human studies provide supportive evidence for the neurological toxicity
14   of 1,3,5-TMB in humans and indicate a consistency and coherency of this effect in humans and
15   laboratory animals.
16          Several studies investigating 1,3,5-TMB effects in experimental animals models were
17   identified in the literature. No chronic or subchronic inhalation studies were identified that
18   investigated effects in adult animals. One developmental toxicity study investigating maternal and
19   fetal toxicity was identified in the literature [Saillenfaitetal.. 2005). Data from this study pertaining
20   to the primary hazards observed animals (maternal/developmental effects) was considered as
21   candidate critical effects for the purpose of determining the point of departure (POD) for derivation
22   of the inhalation RfC for 1,3,5-TMB. Neurotoxicity and respiratory toxicities were also observed in
23   acute and short-term inhalation studies investigating effects in adult animals. However, the high
24   exposure concentrations used in acute studies and the short exposure durations employed in both
25   acute and short-term studies  limit their utility for the quantitation of chronic human health effects.
26   Nevertheless, as with the human in which subject were exposed to mixtures containing 1,3,5-TMB,
27   these studies provide qualitative information regarding hazard identification, especially the
28   observation of the consistency and coherency of these effects across the 1,3,5-TMB database.
29          The developmental toxicity study by Saillenfait et al. [2005) is adequate for dose-response
30   analysis. This study exposed rats, a common laboratory animal for developmental studies, by
31   inhalation to 1,3,5-TMB (reported as 99% pure [impurities not reported]). The four exposure
32   groups covered just over an order of magnitude, with the higher three groups spaced about twofold

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-34
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene
 1    apart Typical numbers of animals per exposure group were used, 25/group. Regarding exposure
 2    characterization, Saillenfait et al. [2005] reported actual concentrations, as measured by gas
 3    chromatography, to be within 10% of target concentrations. This increases the confidence in the
 4    overall evaluation and adequacy of this study. Target and actual concentrations are provided in
 5    Table 2-11.

     Table 2-11. Target and actual exposure concentrations used in BMD modeling of
                 1,3,5-TMB endpoints considered for the derivation of the RfC
Reference
Saillenfait et al.
(2005)
Species/sex
Rat, female
(pregnant dam);
male and female
(fetuses)
Target exposure
concentration (mg/m3)
492
1,476
2,952
5,904
Actual exposure
concentration (mg/m3)
497
1,471
2,974
5,874
 6          The Saillenfait et al. [2005] study examined 1,3,5-TMB-induced toxicity in both the
 7   pregnant animal and the developing fetus, and the observed effects that demonstrated statistically
 8   significant decreases relative to control were considered for the derivation of the RfC for 1,3,5-TMB
 9   (Table 2-12]. These endpoints included decreased male and female fetal weights and decreased
10   maternal weight gain (minus gravid uterine weight]. Changes in serum chemistry parameters in
11   rats exposed to 1,3,5-TMB in a short-term (five weeks] inhalation study (Wiglusz etal.. 1975a]
12   were not considered for derivation of the RfC due to inconsistent temporal patterns of effect and
13   the lack of accompanying histopathology.
                 This document is a draft for review purposes only and does not constitute Agency policy.

                                                 2-35
                                                               DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene
     Table 2-12. Endpoints considered for the derivation of the RfC for 1,3,5-TMB
Endpoint
Species/sex
Exposure concentration (mg/m3)
0
492
1,476
2,952
5,904
Developmental endpoints
Decreased fetal weight
(g)a
Rat, male
Rat, female
5.80 ±
0.41b'c
5.50 ±
0.32
5.76 ±
0.27
5.74 ±
0.21
5.50 ±
0.31
5.27 ±
0.47
5.39 ±
0.55*
5.18±
0.68
5.10±
0.57**
4.81
0.45**
Maternal endpoints
Decreased maternal
weight gain (g)a
Rat, female
29 ±14
(n = 21)d
30 ±9
(n = 22)
20 ±12
(n = 21)
7 ±20*
(n = 17)
-12 ± 19**
(n = 18)
      p<0.05;  p<0.01
     "Adapted from Saillenfait et al. (2005).
     bl\lumbers of live fetuses not explicitly reported.
     "Values are expressed as mean ± 1SD.
     dl\lumber of dams with live litters.
     2.3.2. Methods of Analysis for 1,3,5-TMB
 1          As discussed above in Section 2.3.1, endpoints observed in Saillenfait et al. [2005] that
 2   demonstrated statistically significant (p < 0.05) increases or decreases relative to control for at
 3   least one exposure group were considered for the derivation of the RfC for 1,3,5-TMB; these effects
 4   are listed in Table 2-12. This assessment used the BMD approach, when possible, to estimate a POD
 5   for the derivation of an RfC for 1,3,5-TMB (see Section C.I of Appendix C for detailed methodology).
 6   The BMD approach involves fitting a suite of mathematical models to the observed dose-response
 7   data using EPA's BMDS (version 2.2), and then selecting the best fitting model. Each best-fit model
 8   estimates a BMD and its associated BMDL (i.e., a 95% lower bound on the BMD) corresponding to a
 9   selected BMR.
10          For maternal weight gain identified from the Saillenfait et al. (2005) study,  no information
11   is available regarding the change in these responses that would be considered biologically
12   significant, thus a BMR equal to a 1 standard deviation change in the control mean was used in
13   modeling these endpoints, consistent with EPA's Benchmark Dose Technical Guidance (U.S. EPA.
14   2012b). For the decreased male and female fetal body weight endpoints identified from the
15   Saillenfait et al. (2005) study, a BMR of 5% relative deviation from the control mean was selected
16   (see Section 2.1.2 for a detailed discussion for the rationale for this choice). Additionally, a BMR
17   equal to a 1 standard deviation change in the control mean was also selected for the BMD modeling

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                   2-36
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene

 1   of fetal body weight to facilitate comparisons across assessments [U.S. EPA. 2012b]. The estimated
 2   BMDL is then used as the candidate POD (Table 2-13).
 3          The suitability of the above methods to determine a POD is dependent on the nature of the
 4   toxicity database for a specific chemical. In the Saillenfait et al. [2005] study, although decreased
 5   fetal body weight in males and females was considered for BMD modeling, BMDS was unable to
 6   adequately model the variance in response for this endpoint Consequently, the NOAEL/LOAEL
 7   approach was used to identify a POD. Detailed modeling results are provided in Section C.I of
 8   Appendix C.
 9          Because an RfC is a toxicity value that assumes continuous human inhalation exposure over
10   a lifetime, data derived from inhalation studies in animals need to be adjusted to account for the
11   noncontinuous exposures used in these studies. In the Saillenfait et al. [2005] study, rats were
12   exposed to 1,3,5-TMB for 6 hours/day for 15 consecutive days (GD6-GD20). Therefore, the
13   duration-adjusted PODs for developmental/ maternal effects were calculated as follows:
14          PODADj (mg/m3) = POD (mg/m3) x  hours exposed per day/24 hours
15   For example, for decreased fetal weight in males, the PODADj would be calculated as follows:
16          PODADj (mg/m3) = 2,974 mg/m3 x 6 hours/24 hours
17          PODADj (mg/m3) = 744 mg/m3
18          Because the selected endpoints for consideration as the critical effect (i.e., decreased fetal
19   body weight, and maternal body weight gain] are assumed to result primarily from systemic
20   distribution  of 1,3,5-TMB, and no available PBPK model exists for 1,3,5-TMB, the human equivalent
21   concentration [HEC] for 1,3,5-TMB was calculated by the application of the appropriate dosimetric
22   adjustment factor (DAF] for systemically acting gases (i.e., Category 3 gases], in accordance with the
23   EPAs RfC Methodology [U.S. EPA. 1994b]. DAFs are ratios of animal and human physiologic
24   parameters,  and are dependent on the nature of the contaminant (i.e., particle or gas] and the target
25   site (i.e., respiratory tract or remote to the portal-of-entry [i.e., systemic]] [U.S. EPA, 1994b]. For
26   gases with systemic effects, the DAF is expressed as the ratio between the animal and human
27   blood:air partition coefficients:
28          DAF = (Hb/g)A/(Hb/g)H
29          where:
30          (Hb/g)A = the animal blood:air partition coefficient
31          (Hb/g)H = the human blood:air partition coefficient
32          DAF = 55.7 [Tarnberg and Tohanson. 1995]/43 [Meulenberg and Vijverberg. 2000]
33          DAF = 1.3

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                 2-37
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene
 1           In cases where the animal blood:air partition coefficient is higher than the human value,
 2    resulting in a DAF > 1, a default value of 1 is substituted [U.S. EPA. 1994b]. For example, the HEC for
 3    decreased female fetal body weight (reported in Saillenfait et al. [2005]] is calculated as follows:
 4           PODHEc = PODADj (mg/m3) x DAF
 5           PODHEc = PODADj (mg/m3) x 1.0
 6           PODHEc = 744 mg/m3 x 1.0
 7           PODHEc = 744 mg/m3
 8           Table 2-13 presents the calculated HECs for the candidate critical effects, selected
 9    uncertainty factors (UFs], and the resulting derivation of candidate RfCs from the Saillenfait et al.
10    [2005] developmental toxicity study.

     Table 2-13. Summary of derivation of points of departure for 1,3,5-TMB
Endpoint/Reference
Species/sex
Model; BMRor
NOAEL/LOAEL
POD
(mg/m3)
Candidate
PODADJ
(mg/m3)
Candidate
PODHEC
(mg/m3)
Developmental endpoints
Decreased fetal body weight
(Saillenfait etal., 2005)

Rat, male
Rat, female
NOAELa
NOAELa
2,974
2,974
744
744
744
744
Maternal endpoints
Decreased maternal body
weight gain
(Saillenfait etal. ,2005)
Rat, female
Power; 1 SD
1,302
326
326
     ' No model was able to fit data adequately, or data were not modeled.
     2.3.3. Derivation of Candidate RfC Values for 1,3,5-TMB
11           Under EPA's A Review of the Reference Dose and Reference Concentration Processes [[U.S.
12    EPA. 2002]. §4.4.5], also described in the Preamble, five possible areas of uncertainty and
13    variability were considered in deriving the candidate RfC values for 1,2,4-TMB. An explanation of
14    these five possible areas of uncertainty and variability and the values assigned to each as a
15    designated uncertainty factor [UF] to be applied to the candidate PODHEc are as follows:
16           An interspecies uncertainty factor, UFA, of 3 (101/2 = 3.16, rounded to 3] was applied to
17    account for uncertainty in characterizing the toxicokinetic and toxicodynamic differences between
18    rats and humans following inhalation exposure to  1,3,5-TMB. In this assessment, the use of a DAF to

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-38
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1   extrapolate external concentrations from rats to humans reduces toxicokinetic uncertainty in
 2   extrapolating from the rat data, but does not account for the possibility that humans may be more
 3   sensitive to 1,3,5-TMB than rats due to toxicodynamic differences. A default UFA of 3 was thus
 4   applied to account for this remaining toxicodynamic uncertainty and any residual toxicokinetic
 5   uncertainty.
 6          An intraspecies uncertainty factor, UFn, of 10 was applied to account for potentially
 7   susceptible individuals in the absence of data evaluating variability of response in the human
 8   population following inhalation of 1,3,5-TMB. No information is currently available to predict
 9   potential variability in human susceptibility, including variability in the expression of enzymes
10   involved in 1,3,5-TMB metabolism.
11          A LOAEL to NOAEL uncertainty factor, UFL, of 1 was applied because the current approach is
12   to address this factor as one of the considerations in selecting a BMR for BMD modeling. In this
13   case, a BMR equal to a 1 standard deviation change in the control mean for modeled endpoints was
14   selected under the assumption that this BMR represents a minimal, biologically significant change
15   for these effects. For endpoints that could not be modeled, a LOAEL to NOAEL uncertainty factor of
16   1 was applied as a NOAEL was used.
17          A subchronic to chronic uncertainty factor, UFS, of 3 (101/2 = 3.16, rounded to 3) was applied
18   to decreased maternal weight to account for extrapolation from a subchronic (albeit gestational)
19   exposure duration study to derive a chronic RfC. The 3-fold uncertainty factor is applied to the POD
20   identified from the subchronic study on the assumption that effects observed in a similar chronic
21   study would be observed at lower concentrations for a number of possible reasons, including
22   potential cumulative damage occurring over the duration of the chronic study or an increase in the
23   magnitude or severity of effect with increasing duration of exposure. A full subchronic to chronic
24   uncertainty factor of 10 was not applied in this case as there was no observed decrease in adult
25   body weights in rats exposed to either 1,2,4-TMB or 1,2,3-TMB for longer durations (i.e., 90 days).
26   For decreases in fetal weight, a UFS of 1 was applied.
27          A database uncertainty factor, UFD, of 3 (101/2 = 3.16, rounded to 3) was applied to account
28   for database deficiencies. Strengths of the database include the well-designed developmental
29   toxicity study that investigated standard measures of maternal and fetal toxicity in Sprague-Dawley
30   rats. However, the lack of a multi-generational reproductive/developmental toxicity study
31   investigating effects due to  1,3,5-TMB exposure is a weakness of the database. EPA's^l Review  of the
32   Reference Dose and Reference Concentration Processes (U.S. EPA, 2002] also recommends that the
33   database uncertainty factor take into consideration whether there is concern from the available
34   toxicology database that the developing organism may be particular susceptible to effects in
35   specific organ systems. TMBs (unspecified isomer) are able to cross the placenta (Cooper etal..
36   2001: Dowty etal.. 1976]: therefore, as neurotoxicity is observed in adult animals in the available
37   short-term 1,3,5-TMB inhalation studies, there is concern that exposure to 1,3,5-TMB may result in

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                   2-39
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene

 1   neurotoxicity in the developing organism. EPA's Guidelines for Neurotoxicity Risk Assessment [U.S.
 2   EPA. 1998] identifies specific effects observed in adult animals (e.g., cognitive and motor function)
 3   that can also affect the developing organism exposed in utero. The Neurotoxicity Guidelines [U.S.
 4   EPA, 1998] also indicate that neurotoxicants may have greater access to the nervous system in
 5   developing organisms due to an incomplete blood-brain barrier and immature metabolic
 6   detoxifying pathways. Therefore, there is some concern that the lack of a developmental
 7   neurotoxicity study is a deficiency in the database and that the inclusion of such a study would
 8   potentially result in a lower POD than the POD for maternal effects identified from the available
 9   1,3,5-TMB toxicity database. In summary, a 3-fold database UF was applied to account for the lack
10   of both a multi-generation reproductive/developmental toxicity study and a developmental
11   neurotoxicity study in the available database for 1,3,5-TMB.
12          Table 2-14 is a continuation of Table 2-13, and summarizes the application of UFs to each
13   POD to derive a candidate value for each data set. The candidate values presented in Table 2-14 are
14   preliminary to the derivation of the organ/system-specific values. These candidate values are
15   considered individually in the selection of a representative inhalation reference value for a specific
16   hazard and subsequent overall RfC for 1,3,5-TMB. Figure 2-3 presents graphically these candidate
17   values, uncertainty factors, and points of departure, with each bar corresponding to one data set
18   described in Tables 2-13 and 2-14. Additionally, the RfC values for 1,2,4-TMB and 1,2,3-TMB are
19   shown for comparative purposes

     Table 2-14. Effects and corresponding derivation of candidate RfC values for
                 1,3,5-TMB
Endpoint/Reference
HEC
(mg/m3)3
UFA
UFH
UFL
UFS
UFD
Composite
UF
Candidate RfC
value (mg/m3)b
Developmental endpoints
Decreased fetal body weight, male
(Saillenfait et al., 2005)

Decreased fetal body weight, female
(Saillenfait et al., 2005)

744
744
3
3
10
10
1
1
1
1
3
3
100
100
7.44
7.44
Maternal endpoints
Decreased maternal body weight gain
(Saillenfait et al., 2005)
326
3
10
1
3
3
300
1.09
     aHuman equivalent concentration.
      As calculated by application of uncertainty factors, not rounded to 1 significant digit.
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                   2-40
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                        Toxicological Review of Trimethylbenzene
                         Pain Sensitivity
           E   xl/Maternal body weight gain
           •£     (Saillenfaitetal., 2005)
                 Fetal body weight, females
                 (Saillenfaitetal., 2005)
         gj -53
                xl/ Fetal body weight, males
                 (Saillenfaitetal., 2005)
 Composite UF
A Candidate value
• POD(HEC)
                                   0.001      0.01      0.1         1        10
                                                        Concentrations (mg/m3)
                                                                                  100
                                                                                           1000
     Figure 2-3. Candidate RfC values with corresponding POD and composite UF for
                 1,3,5-TMB.
     2.3 A. Derivation of Organ/System Specific Reference Concentrations for 1,3,5-TMB
 1          Table 2-15 distills the candidate values from Table 2-14 into a single value for each organ or
 2   system. The single RfC value selected for a particular organ system was preferably chosen using
 3   biological and toxicological information regarding that endpoint. If no compelling biological
 4   information exists with which to select the primary hazard, the lowest RfC value for that organ
 5   system was selected. These organ- or system-specific reference concentrations may be useful for
 6   subsequent cumulative risk assessments that consider the combined effect of multiple agents acting
 7   at a common site. The individual organs and systems for which specific RfC values were derived
 8   were the pregnant animal (maternal) and developing fetus (developmental). The RfC value for
 9   maternal effects was the lowest of the derived specific RfCs using 1,3,5-TMB data. The RfC value for
10   developmental effects was greater than that for maternal effects, indicating this effect may be of
11   less concern. However, effects to pregnant animals and the developing fetus may be of less concern
12   in general as the RfC values for these types of effects (based on decreased maternal weight gain and
13   decreased male and female fetal weight, respectively) are much greater than the RfC value derived
14   for 1,2,4-TMB based on decreased pain sensitivity (see Section 2.3.5 for details).
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                   2-41
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene
     Table 2-15. Organ/system-specific RfCs and proposed overall RfC for 1,3,5-TMB
Effect
Developmental
Maternal
Proposed
overall RfC
(Neurological)
Basis
Decreased fetal weight (male
and female
Decreased maternal weight
gain
Decreased pain sensitivity
(based on RfC derived for
1,2,4-TMB)
RfC
(mg/m3)
7
1
5 x 10"2
Exposure
description
Gestational
Gestational
Subchronic
Confidence
Low to medium
Low to medium
Low to medium
     2.3.5. Selection of the Proposed Overall Reference Concentration for 1,3,5-TMB
 1          Decreased maternal weight gain was identified as the most sensitive endpoint in the
 2   1,3,5-TMB database. A PODHEc of 326.0 mg/m3 for decreased maternal weight gain from Saillenfait
 3   et al. [2005] was used to derive a candidate chronic RfC for 1,3,5-TMB as shown in Table 2-14.
 4   Uncertainty factors, selected and applied in accordance with the procedures described in EPA's A
 5   Review of the Reference Dose and Reference Concentration Processes [U.S. EPA, 2002], were
 6   previously discussed in Section 2.3.2. Application of this 300-fold composite UF yields the
 7   calculation of the chronic RfC for 1,3,5-TMB as follows:
 8          RfC = PODHEC * UF = 326 mg/m3 4- 300 = 1.09 mg/m3 = 1 mg/m3 (rounded to one
 9          significant digit)
10          However, while Saillenfait et al. [2005] is a well-conducted developmental toxicity study
11   that evaluates a wide range of fetal and maternal endpoints resulting from 1,3,5-TMB inhalation
12   exposure, a number of other factors lessens its suitability for use in deriving an RfC for 1,3,5-TMB.
13   First, although maternal and fetal toxicities were observed in this study, it is important to note that
14   the candidate RfC for 1,3,5-TMB, derived based on the critical effect of decreased maternal body
15   weight gain  (corrected for gravid uterine weight], is 20-fold higher than the RfC derived for
16   1,2,4-TMB, which is based on altered CNS function measured as decreased pain sensitivity. As
17   discussed in Section 1.1.6, the available toxicological database for 1,2,4-TMB and 1,3,5-TMB, across
18   all exposure durations, indicates there are important similarities in the two isomers' neurotoxicity
19   that are supportive of an RfC for 1,3,5-TMB that is not substantially different than the RfC derived
20   for 1,2,4-TMB. Also supporting this conclusion is the observation that 1,2,4-TMB and  1,3,5-TMB
21   display important similarities in regard to chemical properties and toxicokinetics, including
22   similarities in blood:air partition coefficients, respiratory uptake, and absorption into the

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-42
                                                                 DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   bloodstream (see Section 1.1.7 and Appendices B.I and B.2). These similarities support the
 2   conclusion that internal blood dose metrics for 1,3,5-TMB would be comparable to those calculated
 3   for 1,2,4-TMB using the available PBPK model.
 4          Given these considerations, the use of 1,3,5-TMB-specific data for derivation of an RfC was
 5   not considered to be scientifically supported. Thus, the chronic RfC of 5 x 1Q-2 mg/m3 derived
 6   for 1,2,4-TMB was adopted as the RfC for 1,3,5-TMB based on the conclusion that the two
 7   isomers were sufficiently similar based on chemical properties, toxicokinetics, and toxicity.

     2.3.6. Uncertainties in the Derivation of the Reference Concentration  for 1,3,5-TMB
 8          Uncertainties exist in adopting the RfC derived for 1,2,4-TMB based on altered CNS function
 9   (i.e., decreased pain sensitivity) as the RfC for 1,3,5-TMB. While the available database for
10   1,3,5-TMB was considered sufficient to derive an RfC, if the most sensitive endpoint from the only
11   adequate study in the 1,3,5-TMB database [i.e., decreased maternal weight gain; Saillenfait et al.
12   (2005)] was used for the RfC derivation, an RfC 20-fold higher would be derived for 1,3,5-TMB vs.
13   that derived for 1,2,4-TMB (1 vs. 5 x 1Q-2 mg/m3, respectively). Although uncertainty exists in
14   adopting the 1,2,4-TMB RfC for 1,3,5-TMB RfC, both isomers share multiple commonalities and
15   similarities regarding their chemical, toxicokinetic, and toxicological properties that support the
16   adoption of the value of one isomer for the other. The majority of uncertainty regarding 1,3,5-TMB's
17   database involves the lack of a chronic, subchronic, or multi-generational reproductive study for
18   this isomer. Given the similarities in toxicity from the developmental toxicity study, and
19   neurotoxicity and respiratory toxicity observed in the available acute and short-term studies, there
20   is strong evidence that the two isomer's toxicity resulting from subchronic exposure can be
21   expected to be similar. Therefore, while uncertainty exists in the derivation of 1,3,5-TMB's RfC, the
22   available information regarding sufficient chemical, toxicokinetic, and toxicological similarity
23   between the two isomers supports adopting the RfC for 1,2,4-TMB as the RfC for 1,3,5-TMB.

     2.3.7.  Confidence Statement for 1,3,5-TMB
24         The chronic RfC for 1,2,4-TMB was  adopted as the RfC for 1,3,5-TMB; thus, confidence in the
25   study from which the critical effect was identified, Korsak and Rydzynski (1996). is low to
26   medium (see above). The database for 1,3,5-TMB  includes acute, short-term, and developmental
27   toxicity studies in rats and mice. However, confidence in the overall database is low to medium
28   because it lacks chronic, subchronic, multi-generation reproductive/developmental toxicity, and
29   developmental neurotoxicity studies and  most of the studies supporting the critical effect come
30   from the same research institute. Reflecting the confidence in the study and the database and the
31   uncertainty surrounding the adoption of the RfC derived for 1,2,4-TMB as the RfC  for 1,3,5-TMB, the
32   overall confidence in the RfC for 1,3,5-TMB is low.
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                  2-43
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene
     2.4.  Oral Reference Dose for Effects Other Than Cancer for 1,2,4-TMB
 1         The RfD (expressed in units of mg/kg-day) is defined as an estimate (with uncertainty
 2   spanning perhaps an order of magnitude) of a daily exposure to the human population (including
 3   sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a
 4   lifetime. It can be derived from a NOAEL, a LOAEL, or a 95% lower bound on the benchmark dose
 5   (BMDL), with uncertainty factors (UFs) generally applied to reflect limitations of the data used.

     2.4.1. Identification of Studies and Effects for Dose-Response Analysis for 1,2,4-TMB
 6          No chronic or subchronic studies were identified for 1,2,4-TMB that utilized the oral route
 7   of exposure. Therefore, the available oral database for 1,2,4-TMB is minimal as defined by EPA
 8   guidance (i.e., there is no human data available nor any adequate oral animal data) (U.S. EPA. 2002).
 9   and thus this database is inadequate for the derivation of an RfD.

     2.4.2. Methods of Analysis for 1,2,4-TMB
10          Even though the available oral database for 1,2,4-TMB is inadequate to derive an RfD, a
11   route-to-route extrapolation from inhalation to oral for the purposes of deriving an RfD is possible
12   using the existing inhalation data and the available 1,2,4-TMB PBPK model (Hissinketal.. 2007).
13   The Hissink model was chosen as an appropriate model because it was the only published
14   1,2,4-TMB model that included parameterization for both rats and humans, the model code was
15   available,  and the model adequately predicted experimental data in the dose range of interest.
16   Using route-to-route extrapolation via application of PBPK models is supported by EPA guidance
17   (U.S. EPA, 2002,1994b) given enough data and the ability to interpret that data with regard to
18   differential metabolism and toxicity between different routes of exposure. The available database
19   for 1,2,4-TMB supports the use of route-to-route extrapolation; sufficient evidence exists that
20   demonstrates similar qualitative profiles of metabolism (i.e., observation of dimethylbenzoic and
21   hippuric acid metabolites) and patterns of parent compound distribution across exposure routes
22   (Section B.2, Appendix B). Further, no evidence exists that would suggest toxicity profiles would
23   differ to a substantial degree between oral and inhalation exposures. In fact, in acute oral studies in
24   rats (Tomas etal., 1999a: Tomas etal., 1999b), the observed neurotoxic effects of exposure to 1,2,4-
25   TMB (i.e.,  alterations in motor function and electrocortical activity) are similar to effects observed
26   following  short-term exposures to 1,2,4-TMB via inhalation.
27          Therefore, assuming oral exposure would result in the same systemic effect as inhalation
28   exposure  (i.e., altered CNS function, measured as decreased pain sensitivity (Korsak and Rydzynski.
29   1996)). an oral exposure component was added to the Hissink et al. (2007) PBPK model by EPA
30   (Section B.3.3.5, Appendix B), assuming continuous oral ingestion and 100% absorption of the
31   ingested 1,2,4-TMB by constant infusion of the oral dose into the liver. This is a common
32   assumption when information about the oral absorption of the compound is unknown. The

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                 2-44
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   contribution of the first-pass metabolism in the liver for oral dosing was evaluated by simulating
 2   steady-state venous blood levels (at the end of 5 0 days continuous exposure) for a standard human
 3   at rest (70 kg) for a range of concentrations and doses; at low daily doses (0.1-10  mg/kg-day),
 4   equivalent inhalation concentrations result in steady state blood concentrations 4-fold higher than
 5   those resulting from oral doses, indicating the presence of first-pass metabolism following oral
 6   exposure (see Figure B-18, Appendix B). This difference became insignificant for daily doses
 7   exceeding 50 mg/kg-day.
 8          The human PBPK model inhalation dose metric (weekly average blood concentration,
 9   mg/L) for the PODADj (0.086 mg/L) for decreased pain sensitivity was used as the  target for the oral
10   dose metric. The human PBPK model was run to determine what oral exposure would yield  an
11   equivalent weekly average blood concentration, and then the resulting value of 6.3 mg/kg-day was
12   used as the human equivalent dose POD (PODHED) for the RfD derivation.

     2.4.3. Derivation of the Reference Dose for 1,2,4-TMB
13          A PODHED of 6.3 mg/kg-day was derived for the oral database using route-to-route
14   extrapolation based on the neurotoxic effects (i.e., decreased pain sensitivity) observed by Korsak
15   and Rydzyhski (1996) following inhalation exposure to 1,2,4-TMB. Thus, the same uncertainty
16   factors applied to derive the RfC (see Section 2.1.5) were also applied to derive the RfD. The
17   uncertainty factors, selected and applied in accordance with the procedures described in EPA's A
18   Review of the Reference Dose and Reference Concentration Processes (U.S. EPA. 2002) (Section 4.4.5
19   of the report), address five areas of uncertainty resulting in a composite UF of 300.
20          Application of this 300-fold composite UF yields the calculation of the chronic RfD for
21   1,2,4-TMB as follows:

22          RfD = PODHED * UF = 6.3 mg/kg-day 4- 300 = 0.02 mg/kg-day = 2 x 10 2 mg/kg-day
23          (rounded to one significant digit)
                  This document is a draft for review purposes only and does not constitute Agency policy.

                                                  2-45
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                      Toxicological Review of Trimethylbenzene

     2.4.4. Uncertainties in the Derivation of the Reference Dose for 1,2,4-TMB
 1          As the oral RfD for 1,2,4-TMB was based on a route-to-route extrapolation in order to
 2   determine the oral dose that would result in the same effect (i.e., decreased pain sensitivity) as
 3   inhalation exposure in Korsak and Rydzyhski [1996]. the uncertainties regarding this derivation
 4   are the same as those for the RfC for 1,2,4-TMB (see  Section 2.1.6), with the exception of the
 5   uncertainty surrounding the route-to-route extrapolation. The model used to perform this route-to-
 6   route extrapolation is a well-characterized model considered appropriate for the purposes of this
 7   assessment. One source of uncertainty regarding the route-to-route extrapolation is the assumption
 8   of 100% bioavailability, that is, 100% of the ingested 1,2,4-TMB would be absorbed and pass
 9   through the liver. If not all of the compound is bioavailable, a lower blood concentration would be
10   expected compared to the current estimate, and thus, a higher RfD would be calculated.

     2.4.5. Confidence Statement for 1,2,4-TMB
11          A PBPK model was utilized to perform a route-to-route extrapolation to determine a POD
12   for the derivation of the RfD from the Korsak and Rydzyhski (1996) inhalation study and
13   corresponding critical effect. The confidence in the study from which the critical effect was
14   identified, Korsak and Rydzyhski (1996), is low to medium (see Section 2.1.7). The inhalation
15   database for 1,2,4-TMB includes acute, short-term, subchronic, and developmental toxicity studies
16   in rats and mice. However, confidence in the database for 1,2,4-TMB is low to medium because it
17   lacks chronic, multi-generation reproductive/developmental and developmental neurotoxicity
18   studies, and the studies supporting the critical effect predominantly come from the same research
19   institute. Reflecting the confidence in the study and the database and the uncertainty surrounding
20   the application of the available PBPK model for the purposes of a route-to-route extrapolation, the
21   overall confidence in the RfD for 1,2,4-TMB is low.

     2.5.  Oral Reference Dose for Effects Other Than Cancer for 1,2,3-TMB

     2.5.1. Identification of Studies  and Effects for Dose-Response Analysis for 1,2,3-TMB
22          No chronic or subchronic studies were identified for 1,2,3-TMB that utilized the oral route
23   of exposure. Therefore, the available oral database for  1,2,3-TMB is minimal as defined by EPA
24   guidance (i.e., there is no human data available nor any adequate oral animal data) (U.S. EPA. 2002).
25   and thus this database is inadequate for the derivation of an RfD.

     2.5.2. Methods of Analysis and Derivation of the Reference Dose for 1,2,3-TMB
26          The available oral database is inadequate to derive an RfD for 1,2,3-TMB. No chronic,
27   subchronic, or short-term oral exposure studies were found in the literature. However, as discussed
28   in Section 1.1.6, there are sufficient similarities between isomers regarding observed toxicological

                 This document is a draft for review purposes only and does not constitute Agency policy.
                                                 2-46
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   effects that support adopting the RfD for 1,2,4-TMB as the RfD for 1,2,3-TMB. Specifically, the
 2   qualitative pattern of neurotoxic effects following short-term and subchronic inhalation exposures
 3   is similar between 1MB isomers. Particularly important to this determination is that, although
 4   1,2,3-TMB is observed to decrease pain sensitivity at lower concentrations than 1,2,4-TMB (LOAEL
 5   values of 123 vs. 492 mg/m3, respectively), the magnitude of decreased pain sensitivity is similar
 6   for 1,2,4-TMB and 1,2,3-TMB, especially at the low- and mid-concentrations. This similarity of effect
 7   in the low-dose region of the dose-response curve is exhibited by equal RfC values derived from
 8   isomer-specific data: 5 x 10~2 mg/m3. Additionally, given that similar patterns of neurotoxicity are
 9   observed following acute oral and subchronic inhalation exposures to 1,2,4-TMB, it is reasonable to
10   assume that neurotoxicity profiles would not differ substantially between oral and inhalation
11   exposures to  1,2,3-TMB. Although a PBPK model exists for 1,2,4-TMB that allows for route-to-route
12   extrapolation from inhalation to oral exposure, no such model exists for 1,2,3-TMB. However,
13   similarities in blood:air and tissue:air partition coefficients and degree of absorption into the
14   bloodstream between 1,2,4-TMB and 1,2,3-TMB support the conclusion that internal blood dose
15   metrics for 1,2,3-TMB would be similar to those calculated for 1,2,4-TMB using that isomer's
16   available PBPK model. Also, the qualitative metabolic profiles for the two isomers are similar, with
17   dimethylbenzyl hippuric acids being the major terminal metabolite for both isomers, such that first-
18   pass metabolism through the liver is not expected to differ greatly between 1,2,4-TMB and
19   1,2,3-TMB. Therefore, given the similarities in  chemical properties, toxicokinetics, and
20   toxicity, the RfD derived for 1,2,4-TMB, 2 x 10 2 mg/kg-day was adopted as the RfD for
21   1,2,3-TMB.

     2.5.3. Uncertainties in the Derivation of the Reference Dose for 1,2,3-TMB
22          The uncertainties regarding adopting the  RfD for 1,2,4-TMB as the RfD for 1,2,3-TMB
23   encompass previous areas of uncertainty involved in the derivation of the RfC for 1,2,3-TMB and
24   the RfD for 1,2,4-TMB (see Sections 2.1.6 and 2.2.6). Additionally, there is uncertainty in this
25   adoption regarding the assumptions made about the similarity in toxicokinetics and toxicity
26   between the two isomers. However, as  discussed  above in Sections 1.1.6 and 1.1.7 and in Appendix
27   B (Section B.2), there is strong evidence that both isomers share multiple commonalities and
28   similarities regarding their toxicokinetic and toxicological properties that support adopting one
29   isomer's value for the other.

     2.5.4. Confidence Statement for 1,2,3-TMB
30          The chronic RfD for 1,2,4-TMB was adopted as the RfD for 1,2,3-TMB; thus, confidence in
31   the study from which the critical effect was identified, Korsak and Rydzyhski (1996). is low to
32   medium (see  above). The inhalation database for  1,2,3-TMB includes acute, short-term, and
33   subchronic studies in rats and mice. However, confidence in the database is low to medium because

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-47
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   it lacks chronic, multi-generation reproductive/developmental, developmental toxicity, or
 2   developmental neurotoxicity studies, and the studies supporting the critical effect predominantly
 3   come from the same research institute. Reflecting the confidence in the study and the database and
 4   the uncertainty surrounding the adoption of the RfD derived for 1,2,4-TMB as the RfD for
 5   1,2,3-TMB, the overall confidence in the RfD for 1,2,3-TMB is low.

     2.6.  Oral Reference Dose for Effects Other Than Cancer for 1,3,5-TMB

     2.6.1. Identification of Studies and Effects for Dose-Response Analysis for 1,3,5-TMB
 6          Only one subchronic study [Koch Industries, 1995b] investigating 1,3,5-TMB's toxicity was
 7   located that utilized the oral route of exposure. As this study was not located in the peer-reviewed
 8   literature (it was submitted to EPA under a TSCA 4(a) test rule), EPA sought an independent
 9   external peer review to assess the study's reliability and suitability for use as the basis of an RfD
10   derivation [Versar. 2013). Ultimately, the results of the external peer review led EPA to conclude
11   that this study was not suitable to serve as a principal study with which to derive human health
12   reference doses (see Appendix F). The most critical shortcoming noted in the external peer review
13   of the Koch Industries (1995b] study was its lack of investigation of neurotoxicity endpoints, as
14   these effects (e.g., decreased pain sensitivity, altered cognitive ability) have been demonstrated to
15   be the most sensitive endpoints following inhalation exposure to other TMB isomers. Given the
16   conclusion that the Koch Industries (1995b) study is insufficient for use in RfD derivation, the
17   available oral database for 1,3,5-TMB is minimal as defined by EPA guidance (i.e., there is no human
18   data available nor any adequate oral animal data) (U.S. EPA. 2002). and thus this database is
19   inadequate for the derivation of an RfD.

     2.6.2. Methods of Analysis and Derivation of the Reference Dose for 1,3,5-TMB
20          The available oral database is inadequate to derive an RfD for 1,3,5-TMB. The only identified
21   oral toxicity study was judged to be unsuitable for derivation of the RfD. However, as outlined in the
22   RfC Derivation for 1,3,5-TMB, the chemical, toxicokinetic, and toxicological similarities between
23   1,3,5-TMB and 1,2,4-TMB support adopting the RfC for 1,2,4-TMB as the RfC for 1,3,5-TMB. These
24   considerations also apply to the oral reference value, thus the RfD for 1,2,4-TMB was adopted for
25   1,3,5-TMB. 1,3,5-TMB and 1,2,4-TMB are observed to elicit similar neurotoxic effects in rats in acute
26   and short-term oral and inhalation studies, and therefore the selected critical effect for 1,2,4-TMB,
27   altered CNS function, is relevant to observed 1,3,5-TMB-induced toxicity. Further,  no evidence
28   exists to suggest that toxicity profiles would differ substantially between oral and inhalation
29   exposures to 1,3,5-TMB. In fact, in acute oral studies in rats (Tomas etal.. 1999a: Tomas etal..
30   1999b). the observed neurotoxic effects of exposure to 1,3,5-TMB (i.e., alterations in motor
31   function) are similar to effects observed following short-term exposures via inhalation. Similarities

                  This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-48
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                       Toxicological Review of Trimethylbenzene

 1   in blood:air and tissue:air partition coefficients and absorption into the bloodstream between the
 2   two isomers support the conclusion that internal blood dose metrics for 1,3,5-TMB would be
 3   similar to those calculated for 1,2,4-TMB using the available PBPK model. Also, the qualitative
 4   metabolic profiles for the two isomers are similar, with dimethylbenzyl hippuric acids being the
 5   major terminal metabolite for both isomers, so that first-pass metabolism through the liver is not
 6   expected to differ greatly between 1,2,4-TMB and 1,3,5-TMB. Therefore, given the similarities in
 7   chemical properties, toxicokinetics, and toxicity, the RfD derived for 1,2,4-TMB of 2 x 1Q-2
 8   mg/kg-day was adopted as the RfD for 1,3,5-TMB.

     2.6.3. Uncertainties in the Derivation of the Reference Dose for 1,3,5-TMB
 9          The uncertainties regarding adopting the RfD for 1,2,4-TMB as the RfD for 1,3,5-TMB
10   encompass previous areas of uncertainty involved in the derivation of the RfC for 1,3,5-TMB and
11   the RfD for 1,2,4-TMB (see Sections  2.3.6 and 2.4.4). There is uncertainty regarding this  adoption.
12   However, as discussed above in Section 2.3.3, both isomers share multiple commonalities and
13   similarities regarding their chemical, toxicokinetic, and toxicological properties that support
14   adopting one isomer's value for the other. Additionally, as the RfD derivation for 1,2,4-TMB was
15   based on a route-to-route extrapolation, the uncertainties in that toxicity value's derivation (see
16   Section 2.4.3) apply to the derivation of the RfD for 1,3,5-TMB.

     2.6.4. Confidence Statement for 1,3,5-TMB
17          The chronic RfD for 1,2,4-TMB was adopted as the RfD for 1,3,5-TMB; thus confidence in the
18   study from which the critical effect was identified, Korsak and Rydzyhski (1996). is low  to medium
19   (see above). The inhalation database for 1,3,5-TMB includes acute, short-term, and developmental
20   toxicity studies in rats and mice. However, confidence in the database is low to medium  because it
21   lacks chronic, multi-generation reproductive/developmental and developmental neurotoxicity
22   studies, and the studies supporting the critical effect predominantly come from the same research
23   institute. Reflecting the confidence in the study and the database and the uncertainty surrounding
24   the adoption of the RfD derived for 1,2,4-TMB as the RfD for 1,3,5-TMB, the overall confidence in
25   the RfD for 1,3,5-TMB is low.

     2.7.  Cancer Risk Estimates for 1,2,3-TMB, 1,2,4-TMB, and 1,3,5-TMB
26          Under the U.S. EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA. 2005a). the
27   database for 1,2,4-TMB, 1,2,3-TMB, and 1,3,5-TMB provides "inadequate information to  assess
28   carcinogenic potential". This characterization is based on the limited and equivocal genotoxicity
29   findings, and the lack of data indicating carcinogenicity in experimental animal species via any
30   route of exposure. Information available on which to base a quantitative cancer assessment is
31   lacking, and thus, no cancer risk estimates for either oral or inhalation exposure are derived.
                 This document is a draft for review purposes only and does not constitute Agency policy.
                                                  2-49
                                                                DRAFT—DO NOT CITE OR QUOTE

-------
                                                     Toxicological Review of Trimethylbenzene
REFERENCES
   ACGIH (American Conference of Governmental Industrial Hygienists). (2002). Trimethyl benzene isomers.
      In Documentation of the threshold limit values and biological exposure indices (7 ed.). Cincinnati, OH.
      http://www.acgih.org/Store/ProductDetail.cfm?id=1311

   Andersson. K: Fuxe. K: Nilsen. OG: Toftgard. R: Eneroth. P: Gustafsson. IA. (1981). Production of discrete
      changes in dopamine and noradrenaline levels and turnover in various parts of the rat brain following
      exposure to xylene, ortho-, meta-, and para-xylene, and ethylbenzene. Toxicol Appl Pharmacol 60: 535-
      548.

   Andersson. K: Fuxe. K: Toftgard. R: Nilsen. OG: Eneroth. P: Gustafsson. IA. (1980). Toluene-induced
      activation of certain hypothalamic and median eminence catecholamine nerve terminal systems of the
      male rat and its effects on anterior pituitary hormone secretion. Toxicol Lett 5: 393-398.
      http://dx.doi.org/10.1016/0378-4274f80)90021-l

   Andersson. K: Nilsen. OG: Toftgard. R: Eneroth. P: Gustafsson. IA: Battistini. N: Agnati. LF. (1983).
      Increased amine turnover in several hypothalamic noradrenaline nerve terminal systems and changes
      in prolactin secretion in the male rat by exposure to various concentrations of toluene.
      Neurotoxicology 4: 43-55.

   Aylward. LL: Becker. RA: Kirman. CR: Hays. SM. (2011). Assessment of margin of exposure based on
      biomarkers in blood: an exploratory analysis. Regul Toxicol Pharmacol 61: 44-52.
      http://dx.doi.0rg/10.1016/j.yrtph.2011.06.001

   Balster. RL. (1998). Neural basis of inhalant abuse [Review]. Drug Alcohol Depend 51: 207-214.

   Barker. DIP. (2007). The origins of the developmental origins theory. J Intern Med 261: 412-417.
      http://dx.doi.0rg/10.llll/i.1365-2796.2007.01809.x

   Battig. K: Grandjean. E: Rossi. L: Rickenbacher. I.  (1958). Toxicologische untersuchungen uber
      trimethylbenzol. Archiv fuer Gewerbepathologie und Gewerbehygiene 16: 555-566.

   Battig. K: Grandjean. E: Turrian. V. (1956). [Health damage after continuous exposure to trimethyl
      benzene in a painting workshop].  Soz Praventivmed 1: 389-403.
      http://dx.doi.org/10.1007/BF02031676

   Billionnet. C: Gay. E: Kirchner. S: Leynaert. B: Annesi-Maesano. I. (2011). Quantitative assessments of
      indoor air pollution and respiratory health in a population-based sample of French dwellings. Environ
      Res 111: 425-434. http://dx.doi.0rg/10.1016/j.envres.2011.02.008

   Bogo. V: Hill. TA: Young. RW. (1981).  Comparison of accelerod and rotarod sensitivity in detecting
      ethanol- and acrylamide-induced performance decrement in rats: Review of experimental
      considerations of rotating rod systems. Neurotoxicology 2: 765-787.

   Bowen. SE: Batis. 1C: Paez-Martinez. N: Cruz. SL. (2006). The last decade of solvent research in animal
      models of abuse: Mechanistic and behavioral  studies [Review]. Neurotoxicol Teratol 28: 636-647.
      http://dx.doi.0rg/10.1016/j.ntt.2006.09.005

   Braes. PU: Gregory. P: lackson. DM. (1984). Passive avoidance in rats: Disruption by dopamine applied to
      the nucleus accumbens. Psychopharmacology 83: 70-75.

   Brooks. SP: Dunnett. SB. (2009). Tests to assess motor phenotype in mice: A user's guide [Review]. Nat
      Rev Neurosci 10: 519-529. http://dx.doi.org/10.1038/nrn2652
            This document is a draft for review purposes only and does not constitute Agency policy.

                                                R-1            DRAFT—DO NOT CITE OR QUOTE

-------
                                                   Toxicological Review of Trimethylbenzene

CDC (Centers for Disease Control and Prevention). (2004). The health consequences of smoking: A report
   of the Surgeon General. Washington, DC: U.S. Department of Health and Human Services.
   http://www.surgeongeneral.gov/library/smokingconsequences/

Chen. R: Dick. F: Seaton. A. (1999). Health effects of solvent exposure among dockyard painters: Mortality
   and neuropsychological symptoms. Occup Environ Med 56: 383-387.
   http://dx.doi.0rg/10.1136/oem.56.6.383

Choi. DW: Moon. KW: Byeon. SH: Lee. El: Sul. DG: Lee. IH: Oh. EH: Kim. YH. (2009). Indoor volatile organic
   compounds in atopy patients houses in South Korea. Indoor Built Environ 18:144-154.
   http://dx.doi.org/10.1177/1420326X08101945

Chou. CC: Riviere. IE: Monteiro-Riviere. NA. (2003). The cytotoxicity of jet fuel aromatic hydrocarbons and
   dose-related interleukin-8 release from human epidermal keratinocytes. Arch Toxicol 77: 384-391.
   http://dx.doi.org/10.1007/s00204-003-0461-z

Cooper. SP: Burau. K: Sweeney. A: Robison. T: Smith. MA: Symanski. E: Colt.  IS: Laseter. I: Zahm. SH.
   (2001). Prenatal exposure to pesticides: a feasibility study among migrant and seasonal farmworkers.
   Am JInd Med 40: 578-585.

Dahl. AR: Damon. EG: Mauderly. IL: Rothenberg. SI: Seller. FA: Mcclellan. RO. (1988). Uptake of 19
   hydrocarbon vapors inhaled by F344 rats. 10: 262-269. http://dx.doi.org/10.1016/0272-
   0590(88)90310-7

Dowty. Bl: Laseter. IL: Storer. I. (1976). The transplacental migration and accumulation in blood of volatile
   organic constituents. Pediatr Res 10: 696-701. http://dx.doi.org/10.1203/00006450-197607000-
   00013

Eide. I: Zahlsen. K. (1996). Inhalation experiments with mixtures of hydrocarbons.  Experimental design,
   statistics and interpretation of kinetics and possible interactions. Arch Toxicol 70: 397-404.
   http://dx.doi.org/10.1007/s002040050291

Gaschen. A: Lang. D: Kalberer. M: Savi. M: Geiser. T: Gazdhar. A: Lehr. CM: Bur. M: Dommen. I:
   Baltensperger. U: Geiser. M. (2010). Cellular responses after exposure of lung cell cultures to secondary
   organic aerosol particles. Environ Sci Technol 44:1424-1430. http://dx.doi.org/10.1021/es902261m

Ginsberg. G: Ir.  SW: Bruckner. I: Sonawane. B. (2004). Incorporating children's toxicokinetics into a risk
   framework  [Review]. Environ Health Perspect 112: 272-283. http://dx.doi.org/10.1289/ehp.6013

Gralewicz. S: Wiaderna. D. (2001). Behavioral effects following subacute inhalation exposure to m-xylene
   or trimethylbenzene in the rat: A comparative study. Neurotoxicology 22: 79-89.
   http://dx.doi.org/10.1016/S0161-813XrOO)00003-6

Gralewicz. S: Wiaderna. D: Tomas. T. (1997a). Retardation of the age-related increase in spontaneous
   cortical spike-wave discharges (SWD) in rats after a 28-day inhalation (SWD) in rats after a 28-day
   inhalation exposure to an industrial solvent, pseudocumene (1,2,4-trimethylbenzene). Int J Occup Med
   Environ Health 10: 213-222.

Gralewicz. S: Wiaderna. D: Tomas. T: Rydzynski. K. (1997b). Behavioral changes following 4-week
   inhalation exposure to pseudocumene (1,2,4-trimethylbenzene) in the rat. Neurotoxicol Teratol 19:
   327-333. http://dx.doi.org/10.1016/S0892-0362r97)00001-9

Guo. H: Kwok. NH: Cheng. HR: Lee. SC: Hung. WT: Li. YS. (2009). Formaldehyde and volatile organic
   compounds in Hong Kong homes: Concentrations and impact factors. Indoor Air 19: 206-217.
   http://dx.doi.0rg/10.llll/i.1600-0668.2008.00580.x

Guyatt. GH: Oxman. AD: Kunz. R: Vist. GE: Falck-Ytter. Y: Schiinemann. HI. (2008a). GRADE: What is
   "quality of evidence" and why is it important to clinicians? [Review]. BMJ 336: 995-998.
   http://dx.doi.org/10.1136/bmi.39490.551019.BE
          This document is a draft for review purposes only and does not constitute Agency policy,

                                             R-2            DRAFT—DO NOT CITE OR QUOTE

-------
                                                   Toxicological Review of Trimethylbenzene

Guyatt. GH: Oxman. AD: Vist. GE: Kunz. R: Falck-Ytter. Y: Alonso-Coello. P: Schiinemann. HI. (2008b).
   GRADE: An emerging consensus on rating quality of evidence and strength of recommendations. BMJ
   336: 924-926. http://dx.doi.org/10.1136/bmi.39489.470347.AD
Henderson. RF. (2001). Aromatic hydrocarbons: Benzene and other alkylbenzenes. In E Bingham; B
   Cohrssen; CH Powell (Eds.), Patty's toxicology (5 ed., pp. 231-301). New York, NY: John Wiley and Sons.

HEW (U.S. Department of Health, Education and Welfare). (1964). Smoking and health: Report of the
   advisory committee to the surgeon general of the public health service. Washington, DC: U.S.
   Department of Health, Education, and Welfare.
   http://profiles.nlm.nih.gov/ps/retrieve/ResourceMetadata/NNBBMQ

Hill.AB. (1965). The environment and disease: Association or causation? Proc R Soc Med 58: 295-300.

Hillefors-Berglund. M: Liu. Y: von Euler. G. (1995). Persistent, specific and dose-dependent effects of
   toluene exposure on dopamine D2 agonist binding in the rat caudate-putamen. Toxicology 100:185-
   194.http://dx.doi.org/10.1016/0300-483Xr95)03084-S

Hissink. AM: Kriise. I: Kulig. BM: Verwei. M: Muijser. H: Salmon. F: Leenheers. LH: Owen. DE: Lammers. IH:
   Freidig. AP: McKee. RH. (2007). Model studies for evaluating the neurobehavioral effects of complex
   hydrocarbon solvents III. PBPK modeling of white spirit constituents as a tool for integrating animal
   and human test data. Neurotoxicology 28: 751-760. http://dx.doi.Org/10.1016/j.neuro.2007.03.005

HSDB (Hazardous Substances Data Bank). (2011a). 1,2,3-trimethylbenzene. Bethesda, MD: National
   Library of Medicine.

HSDB (Hazardous Substances Data Bank). (2011b). 1,2,4-Trimethylbenzene [Database]. Bethesda, MD:
   National Library of Medicine. Retrieved from http://toxnet.nlm.nih.gov

HSDB (Hazardous Substances Data Bank). (2011c). 1,3,5-Trimethylbenzene [Database]. Bethesda, MD:
   National Library of Medicine. Retrieved from http://toxnet.nlm.nih.gov

Huo. IZ: Aldous. S: Campbell. K: Davies. N. (1989). Distribution and metabolism of 1,2,4-trimethylbenzene
   (pseudocumene) in the rat. Xenobiotica 19: 161-170. http://dx.doi.org/10.3109/00498258909034688

IARC (International Agency for Research on Cancer). (2006). Preamble to the IARC monographs. Lyon,
   France. http://monographs.iarc.fr/ENG/Preamble/

IOM (Institute of Medicine). (2008). Improving the presumptive disability decision-making process for
   veterans. In JM Samet; CC Bodurow (Eds.). Washington, DC:  National Academies Press.
   http://www.nap.edu/openbook.php7record id=11908

lackson. DM: Westlind-Danielsson. A. (1994). Dopamine receptors: Molecular biology, biochemistry and
   behavioural aspects [Review]. Pharmacol Ther 64: 291-370. http://dx.doi.org/10.1016/0163-
   7258(94)90041-8

lanasik. B: lakubowski. M: laiowiecki. P.  (2008). Excretion of unchanged volatile organic compounds
   (toluene, ethylbenzene, xylene and mesitylene) in urine as result of experimental human volunteer
   exposure. Int Arch Occup Environ Health 81: 443-449. http://dx.doi.org/10.1007/s00420-007-0233-9

lanik-Spiechowicz. E: Wyszynska. K: Dziubaltowska.  E. (1998).  Genotoxicity evaluation of
   trimethylbenzenes. Mutat Res Genet Toxicol Environ Mutagen 412:  299-305.
   http://dx.doi.org/10.1016/S1383-5718f97)00202-7

larnberg. I: lohanson. G. (1995). Liquid/air partition coefficients of the trimethylbenzenes. Toxicol Ind
   Health 11: 81-88. http://dx.doi.org/10.1177/074823379501100107

larnberg. I: lohanson. G: Lof. A. (1996). Toxicokinetics of inhaled trimethylbenzenes in man. Toxicol Appl
   Pharmacol 140: 281-288. http://dx.doi.org/10.1006/taap.1996.0223

larnberg. I: lohanson. G: Lof. A: Stahlbom. B. (1997a). Inhalation toxicokinetics of 1,2,4-trimethylbenzene
   in volunteers: Comparison between exposure to white spirit and 1,2,4-trimethylbenzene alone. Sci
   Total Environ 199: 65-71. http://dx.doi.org/10.1016/S0048-9697(97)05482-X
          This document is a draft for review purposes only and does not constitute Agency policy.

                                             R-3            DRAFT—DO NOT CITE OR QUOTE

-------
                                                   Toxicological Review of Trimethylbenzene

larnberg. I: lohanson. G: Lof. A: Stahlbom. B. (1998). Toxicokinetics of 1,2,4-trimethylbenzene in humans
   exposed to vapours of white spirit: Comparison with exposure to 1,2,4-trimethylbenzene alone. Arch
   Toxicol 72: 483-491. http://dx.doi.org/10.1007/s002040050532

larnberg. I: Stahlbon. B: lohanson. G: Lof. A. (1997b). Urinary excretion of dimethylhippuric acids in
   humans after exposure to trimethylbenzenes. Int Arch Occup Environ Health 69: 491-497.
   http://dx.doi.org/10.1007/s004200050179

liun-Horng. T: Kuo-Hsiung. L: Chih-Yu. C: Nina. L: Sen-Yi. M: Hung-Lung. C. (2008). Volatile organic
   compound constituents from an integrated iron and steel facility. J Hazard Mater 157: 569-578.
   http://dx.doi.0rg/10.1016/j.jhazmat.2008.01.022

lones. K: Meldrum. M: Baird. E: Cottrell. S: Kaur. P: Plant. N: Dyne. D: Cocker. I. (2006). Biological
   monitoring for trimethylbenzene exposure: A human volunteer study and a practical example in the
   workplace. Ann Occup Hyg 50: 593-598. http://dx.doi.org/10.1093/annhyg/mel016

Kaspar. BK: Llado. I: Sherkat. N: Rothstein. ID: Gage. FH. (2003). Retrograde viral delivery of IGF-1
   prolongs survival in a mouse ALS model. Science  301: 839-842.
   http://dx.doi.org/10.1126/science.1086137

Koch Industries (Koch Industries, Incorporated). (1995a). 14-day oral gavage toxicity study of 1,3,5-
   trimethylbenzene in rats with a recovery group, with cover letter dated 2/7/95. (44616). Wichita, KS.
   http://www.ntis.gov/search/product.aspx?ABBR=OTS0558836

Koch Industries (Koch Industries, Incorporated). (1995b). 90-day oral gavage toxicity study of 1,3,5-
   trimethylbenzene in rats with a recovery group. (44618). Wichita, KS: Koch Industries, Inc.

Korsak. Z: Rydzyriski. K. (1996). Neurotoxic effects of acute and subchronic inhalation exposure to
   trimethylbenzene isomers (pseudocumene, mesitylene, hemimellitene) in rats. Int J Occup Med
   Environ Health 9: 341-349.

Korsak. Z: Rydzynski. K: lajte. I. (1997). Respiratory irritative effects of trimethylbenzenes: An
   experimental animal study. Int J Occup Med Environ Health 10: 303-311.

Korsak. Z: Stetkiewicz. I: Majcherek. W: Stetkiewicz. I: lajte. I: Rydzynski. K. (2000a). Sub-chronic
   inhalation toxicity of 1,2,4-trimethylbenzene (pseudocumene) in rats. Int J Occup Med Environ Health
   13: 155-164.

Korsak. Z: Stetkiewicz. I: Majcherek. W: Stetkiewicz. I: lajte. I: Rydzynski. K. (2000b). Subchronic inhalation
   toxicity of 1,2,3-trimethylbenzene (hemimellitene) in rats. Int J Occup Med Environ Health 13:  223-
   232.

Korsak. Z: Swiercz. R: Rydzynski. K. (1995). Toxic effects of acute inhalation exposure to 1,2,4-
   trimethylbenzene (pseudocumene) in experimental animals. Int J Occup Med Environ Health 8: 331-
   337.

Kostrewski. P: Wiaderna-Brycht. A.  (1995). Kinetics of elimination of mesitylene and 3,5-dimethylbenzoic
   acid after experimental human exposure. Toxicol Lett 77: 259-264. http://dx.doi.org/10.1016/0378-
   4274f95103305-X

Kostrzewski. P: Wiaderna-Brycht. A: Czerski. B. (1997). Biological monitoring of experimental human
   exposure to trimethylbenzene. Sci Total Environ 199: 73-81. http://dx.doi.org/10.1016/S0048-
   9697(97)05504-6

Kyrklund. T. (1992). The use of experimental studies to reveal suspected neurotoxic chemicals as
   occupational hazards: Acute and chronic exposures to organic solvents [Review]. Am J Ind Med 21:15-
   24. http://dx.doi.org/10.1002/aiim.4700210105

Lammers. IH: Emmen. HH: Muijser.  H: Hoogendijk. EM: McKee. RH: Owen. DE: Kulig. BM. (2007). Model
   studies for evaluating the neurobehavioral effects of complex hydrocarbon solvents  II.
   Neurobehavioral effects of white spirit in rat and human. Neurotoxicology 28: 736-750.
   http://dx.doi.0rg/10.1016/j.neuro.2007.03.003
          This document is a draft for review purposes only and does not constitute Agency policy.

                                             R-4            DRAFT—DO NOT CITE OR QUOTE

-------
                                                   Toxicological Review of Trimethylbenzene

Lee. CR: leong. KS: Kim. Y: Yoo. CI: Lee. IH: Choi. YH. (2005). Neurobehavioral changes of shipyard painters
   exposed to mixed organic solvents. Ind Health 43: 320-326.

Lutz. P: Gralewicz. S: Wiaderna. D: Swiercz. R: Grzelinska. Z: Majcherek. W. (2010). Contrasting effects of
   4-week inhalation exposure to pseudocumene or hemimellitene on sensitivity to amphetamine and
   propensity to amphetamine sensitization in the rat. Int J Occup Med Environ Health 23: 85-94.
   http://dx.doi.org/10.2478/vl0001-010-0005-8

Maltoni. C: Ciliberti. A: Pinto. C: Soffritti. M: Belpoggi. F: Menarini. L. (1997). Results of long-term
   experimental carcinogenicity studies of the effects  of gasoline, correlated fuels, and major gasoline
   aromatics on rats. Ann N Y Acad Sci 837:15-52. http://dx.doi.Org/10.llll/j.1749-
   6632.1997.tb56863.x

Martins. EM: Arbilla. G: Gatti. LV. (2010). Volatile organic compounds in a residential and commercial
   urban area with a diesel, compressed natural gas and oxygenated gasoline vehicular fleet. Bull Environ
   Contam Toxicol 84:175-179. http://dx.doi.org/10.1007/s00128-009-9886-2

McKee. RH: Lammers. IH: Muijser. H: Owen. DE: Kulig. BM. (2010). Neurobehavioral effects of acute
   exposure to aromatic hydrocarbons. Int J Toxicol 29: 277-290.
   http://dx.doi.org/10.1177/1091581810365089

Mclanahan. ED: El-Masri. HA: Sweeney. LM:  Kopylev. LY: Clewell. HI: Wambaugh. IF: Schlosser. PM. (2012).
   Physiologically based pharmacokinetic model use in risk assessment-why being published is not
   enough. Toxicol Sci 126: 5-15. http://dx.doi.org/10.1093/toxsci/kfr295

Meulenberg. C: Vijverberg. H. (2000). Empirical relations predicting human and rat tissue: Air partition
   coefficients of volatile organic compounds. Toxicol Appl Pharmacol 165: 206-216.
   http://dx.doi.org/10.1006/taap.2000.8929

Mikulski.  PI: Wiglusz. R. (1975). The comparative metabolism of mesitylene, pseudocumene, and
   hemimellitene in rats. Toxicol Appl Pharmacol 31: 21-31. http://dx.doi.org/10.1016/0041-
   008X(75)90048-4

MOE (Ontario Ministry of the Environment). (2006). Rationale for the development of Ontario air
   standards for trimethylbenzenes: 1,2,3-Trimethylbenzene. Ontario, Canada.

Mogel. I: Baumann. S: Bohme. A: Kohajda. T: von Bergen. M: Simon. 1C: Lehmann. I. (2011). The aromatic
   volatile organic compounds toluene, benzene and styrene induce COX-2 and prostaglandins in human
   lung epithelial cells via oxidative stress and p38 MAPK activation. Toxicology 289: 28-37.
   http://dx.doi.0rg/10.1016/i.tox.2011.07.006

Myhre. 0: Fonnum. F. (2001). The effect of aliphatic, naphthenic, and aromatic hydrocarbons on
   production of reactive oxygen species and reactive nitrogen species in rat brain synaptosome fraction:
   the involvement of calcium, nitric oxide synthase, mitochondria, and phospholipase A. Biochem
   Pharmacol 62:119-128. http://dx.doi.org/10.1016/S0006-2952(01)00652-9

Myhre. 0: Vestad. TA: Sagstuen. E: Aarnes. H: Fonnum. F. (2000). The effects of aliphatic (n-nonane),
   naphtenic (1,2,4-trimethylcyclohexane), and aromatic (1,2,4-trimethylbenzene) hydrocarbons on
   respiratory burst in human neutrophil granulocytes. Toxicol Appl Pharmacol 167: 222-230.
   http://dx.doi.org/10.1006/taap.2000.9008

Norseth. T: Waage. I: Dale. I. (1991). Acute effects and exposure to organic compounds in road
   maintenance workers exposed to asphalt. Am J Ind Med 20: 737-744.
   http://dx.doi.org/10.1002/ajim.4700200604

NRC (National Research Council). (1983). Risk assessment in the federal government: Managing the
   process. Washington, DC: National Academies Press.
   http://www.nap.edu/openbook.php7record id=366&page=Rl

NRC (National Research Council). (2009). Science and decisions: Advancing risk assessment. Washington,
   DC: National Academies Press, http://www.nap.edu/catalog/12209.html

          This document is a draft for review purposes only and does not constitute Agency policy.

                                             R-5            DRAFT—DO NOT CITE OR QUOTE

-------
                                                   Toxicological Review of Trimethylbenzene

NRG (National Research Council). (2011). Review of the Environmental Protection Agency's draft IRIS
   assessment of formaldehyde. Washington, DC: National Academies Press.
   http://www.nap.edu/catalog/13142.html

OSHA (Occupational Safety & Health Administration). (1996). Occupational safety and health guideline for
   trimethylbenzene. Available online at
   http://www.osha.gov/SLTC/healthguidelines/trimethylbenzene/recognition.html (accessed August 1,
   2007).

Ramaiah. SK. (2007). A lexicologist guide to the diagnostic interpretation of hepatic biochemical
   parameters [Review]. Food Chem Toxicol 45:1551-1557. http://dx.doi.Org/10.1016/j.fct.2007.06.007

Rea. TM: Nash. IF: Zabik. IE: Born. GS: Kessler. WV. (1984). Effects of toluene inhalation on brain biogenic
   amines in the rat. Toxicology 31: 143-150. http://dx.doi.org/10.1016/0300-483X(84)90006-4

Reyes. L: Manalich. R. (2005). Long-term consequences of low birth weight [Review]. Kidney Int
   SupplS107-Slll. http://dx.doi.0rg/10.llll/i.1523-1755.2005.09718.x

Rothman. Kl: Greenland. S. (1998). Modern epidemiology (2nd ed.). Philadelphia, PA: Lippincott, Williams,
   &Wilkins.

Saillenfait. AM: Gallissot. F: Sabate. IP: Morel. G. (2005). Developmental toxicity of two trimethylbenzene
   isomers, mesitylene and pseudocumene, in rats following inhalation exposure. Food Chem Toxicol 43:
   1055-1063. http://dx.doi.0rg/10.1016/i.fct.2005.02.008

Snead. PC. III. (1995). Basic mechanisms of generalized absence seizures [Review].  Ann Neurol 37:146-
   157. http://dx.doi.org/10.1002/ana.410370204

Sulkowski. Wl: Kowalska. S: Matyja. W: Guzek. W: Wesolowski. W: Szymczak. W: Kostrzewski. P. (2002).
   Effects of occupational exposure to a mixture  of solvents on the inner ear: A field study. Int J Occup
   Med Environ Health 15: 247-256.

Swiercz. R: Rydzyriski. K: Wasowicz. W: Majcherek. W: Wesolowski. W. (2002). Toxicokinetics and
   metabolism of pseudocumene (1,2,4-trimethylbenzene) after inhalation exposure in rats. Int J Occup
   Med Environ Health 15: 37-42.

Swiercz. R: Wasowicz. W: Majcherek. W. (2006). Mesitylene (1,3,5-trimethylbenzene) in the liver, lung,
   kidney, and blood  and 3,5-dimethylbenzoic acid in the liver, lung, kidney and urine of rats after single
   and repeated inhalation exposure to mesitylene. Pol J Environ Stud 15: 485-492.

Swiercz. R: Wiaderna. D: Wasowicz. W: Rydzynski. K. (2003). Pseudocumene in brain, liver, lung and blood
   of rats after single and repeated inhalation exposure. Int J Occup Med Environ Health 16: 61-66.

Tomas. T: Lutz. P: Wiaderna. D. (1999a). Changes in electrocortical arousal following acute
   trimethylbenzene  administration in rats. Int J Occup Med Environ  Health 12: 67-78.

Tomas. T: Swiercz. R: Wiaderna. D: Gralewicz. S. (1999b). Effects of acute exposure  to aromatic
   hydrocarbons C 9 on locomotor activity in rats. Trimethylbenzene isomers. Int J Occup Med Environ
   Health 12: 331-343.

Tomas. T: Wiaderna. D: Swiercz. R. (1999c). Neurotoxicity assessment of selected organic solvents based
   on spontaneous and evoked cortical and hippocampal activity in rats. Int J Occup Med Environ Health
   12: 73-84.

TRI (Toxic Release Inventory). (2008). Toxic Release Inventory [Database]: U.S. Environmental Protection
   Agency.

Tsujimoto. Y: Noda. T: Shimizu. M: Moriwaki. H: Tanaka. M. (1999). Identification of the dimethylbenzyl
   mercapturic acid in urine of rats treated with  1,2,3-trimethylbenzene. Chemosphere 39: 725-730.

Tsujimoto. Y: Noda. T: Shimizu. M: Moriwaki. H: Tanaka. M. (2000). Identification of the dimethylbenzyl
   mercapturic acid in urine of rats administered with 1,2,4-trimethylbenzene. Chemosphere 40: 893-
   896. http://dx.doi.org/10.1016/S0045-6535(99)00467-l
          This document is a draft for review purposes only and does not constitute Agency policy.

                                             R-6             DRAFT—DO NOT CITE OR QUOTE

-------
                                                   Toxicological Review of Trimethylbenzene

Tsujimoto. Y: Warashina. M: Nam. VD: Noda. T: Shimizu. M: Yamaguchi. Y: Moriwaki. H: Morimoto. T:
   Kakiuchi. K: Maeda. Y: Tanaka. M. (2005). Determination of urinary phenolic metabolites from rats
   treated with 1,2,3-and 1,3,5-trimethylbenzenes. J Occup Health 47: 337-339.

U.S. Congress. (2011). Consolidated Appropriations Act, 2012. (Pub. L. No. 112-74; 125 STAT. 786). 112th
   U.S. Congress. http://www.gpo.gov/fdsys/pkg/PLAW-112publ74/pdf/PLAW-112publ74.pdf

U.S. EPA (U.S. Environmental Protection Agency). (1986a). Guidelines for mutagenicity risk assessment
   [EPA Report]. (EPA/630/R-98/003). Washington, DC. http://www.epa.gov/iris/backgrd.html

U.S. EPA (U.S. Environmental Protection Agency). (1986b). Guidelines for the health risk assessment of
   chemical mixtures [EPA Report]. (EPA/630/R-98/002). Washington, DC.
   http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=22567

U.S. EPA (U.S. Environmental Protection Agency). (1987). Health effects assessment for trimethylbenzenes
   [EPA Report]. (EPA/600/8-88/060). Cincinnati, OH.
   http://nepis.epa.gov/Exe/ZyPU RL.cgi?Dockey=2 OOOT8ZG.txt

U.S. EPA (U.S. Environmental Protection Agency). (1988). Recommendations for and documentation of
   biological values for use in risk assessment [EPA Report]. (EPA/600/6-87/008). Cincinnati, OH.
   http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=34855

U.S. EPA (U.S. Environmental Protection Agency). (1991). Guidelines for developmental toxicity risk
   assessment [EPA Report]. (EPA/600/FR-91/001). Washington, DC: U.S. Environmental Protection
   Agency, Risk Assessment Forum, http://www.epa.gov/raf/publications/guidelines-dev-toxicity-risk-
   assessmenthtm

U.S. EPA (U.S. Environmental Protection Agency). (1994a). Chemical summary for 1,2,4-trimethylbenzene
   [EPA Report]. (EPA/749/F-94/022A). Washington, DC. http://www.epa.gOV/chemfact/s trimettxt

U.S. EPA (U.S. Environmental Protection Agency). (1994b). Methods for derivation of inhalation reference
   concentrations and application of inhalation dosimetry [EPA Report]. (EPA/600/8-90/066F). Research
   Triangle Park, NC. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=71993

U.S. EPA (U.S. Environmental Protection Agency). (1996). Guidelines for reproductive toxicity risk
   assessment [EPA Report]. (EPA/630/R-96/009). Washington, DC.
   http://www.epa.gov/raf/publications/pdfs/REPR051.PDF

U.S. EPA (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk assessment
   [EPA Report]. (EPA/630/R-95/001F). Washington, DC.
   http://www.epa.gov/raf/publications/pdfs/NEUROTOX.PDF

U.S. EPA (U.S. Environmental Protection Agency). (2000). Supplementary guidance for conducting health
   risk assessment of chemical mixtures [EPA Report]. (EPA/630/R-00/002). Washington, DC.
   http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=20533

U.S. EPA (U.S. Environmental Protection Agency). (2002). A review of the reference dose and reference
   concentration processes [EPA Report]. (EPA/630/P-02/002F). Washington, DC: Risk Assessment
   Forum, U.S. Environmental Protection Agency.
   http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=51717

U.S. EPA (U.S. Environmental Protection Agency). (2005a). Guidelines for carcinogen  risk assessment [EPA
   Report]. (EPA/630/P-03/001F). Washington, DC: Risk Assessment Forum.
   http://www.epa.gov/cancerguidelines/

U.S. EPA (U.S. Environmental Protection Agency). (2005b). Supplemental guidance for assessing
   susceptibility from early-life exposure to carcinogens [EPA Report] (pp. 1125-1133). (EPA/630/R-
   03/003F). Washington, DC. http://www.epa.gov/cancerguidelines/guidelines-carcinogen-
   supplement.htm
          This document is a draft for review purposes only and does not constitute Agency policy,

                                             R-7            DRAFT—DO NOT CITE OR QUOTE

-------
                                                   Toxicological Review of Trimethylbenzene

U.S. EPA (U.S. Environmental Protection Agency). (2006a). Approaches for the application of
   physiologically based pharmacokinetic (PBPK) models and supporting data in risk assessment (Final
   Report) [EPA Report]. (EPA/600/R-05/043F). Washington, DC.
   http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=157668

U.S. EPA (U.S. Environmental Protection Agency). (2006b). A framework for assessing health risk of
   environmental exposures to children [EPA Report]. (EPA/600/R-05/093F). Washington, DC.
   http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=158363

U.S. EPA (U.S. Environmental Protection Agency). (2009). EPAs Integrated Risk Information System:
   Assessment development process [EPA Report]. Washington, DC. http://epa.gov/iris/process.htm

U.S. EPA (U.S. Environmental Protection Agency). (2010). Integrated science assessment for carbon
   monoxide [EPA Report]. (EPA/600/R-09/019F). Research Triangle Park, NC.
   http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=218686

U.S. EPA (U.S. Environmental Protection Agency). (2011). Recommended use of body weight 3/4 as the
   default method in derivation of the oral reference dose [EPA Report]. (EPA/100/R11/0001).
   Washington, DC. http://www.epa.gov/raf/publications/interspecies-extrapolation.htm

U.S. EPA (U.S. Environmental Protection Agency). (2012a). Advances in inhalation gas dosimetry for
   derivation of a reference concentration (rfc) and use in risk assessment [EPA Report]. (EPA/600/R-
   12/044). Washington, DC. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=244650

U.S. EPA (U.S. Environmental Protection Agency). (2012b). Benchmark dose technical guidance.
   (EPA/100/R-12/001). Washington, DC: Risk Assessment Forum.
   http://www.epa.gov/raf/publications/pdfs/benchmark dose guidance.pdf

U.S. EPA (U.S. Environmental Protection Agency). (2012c). EPA announces NAS' review of IRIS Assessment
   development process. Available online at
   http://yosemite.epa.gOV/opa/admpress.nsf/0/lce2a7875daf093485257aOOOQ54df5470penDocumen
   t

Versar (Versar Inc.). (2013). Peer review report: External peer review of the 1995 Koch Industries study
   report: 90-day oral gavage toxicity study of 1,3,5-trimethylbenzene in rats with a recovery group. (EP-
   C-12-045). Springfiled, VA: Versar, Inc.

von Euler. G: Ogren. SO: Eneroth. P: Fuxe. K: Gustafsson. IA. (1994). Persistent effects of 80 ppm toluene on
   dopamine-regulated locomotor activity and prolactin secretion in the male rat. Neurotoxicology 15:
   621-624.

von Euler. G: Ogren. SO: Li. XM: Fuxe. K: Gustafsson. IA. (1993). Persistent effects of subchronic toluene
   exposure on  spatial learning and memory, dopamine-mediated locomotor activity and dopamine D2
   agonist binding in the rat. Toxicology 77: 223-232. http://dx.doi.org/10.1016/0300-483X(93)90162-L

Warter. IM: Vergnes. M: Depaulis. A: Tranchant. C: Rumbach. L: Micheletti. G: Marescaux. C. (1988). Effects
   of drugs affecting dopaminergic neurotransmission in rats with spontaneous petit mal-like seizures.
   Neuropharmacology 27: 269-274. http://dx.doi.org/10.1016/0028-3908r88190043-3

Wiaderna. D: Gralewicz. S: Tomas. T. (1998). Behavioral changes following a four-week inhalation
   exposure to hemimellitene (1,2,3-trimethylbenzene) in rats. Int J Occup Med Environ Health 11: 319-
   334.

Wiaderna. D: Gralewicz. S: Tomas. T. (2002). Assessment of long-term neurotoxic effects of exposure to
   mesitylene (1,3,5-trimethylbenzene) based on the analysis of selected behavioral responses. Int J
   Occup Med Environ Health 15: 385-392.

Wiglusz. R. (1979). The effect of 1, 3, 5-trimethylbenzene inhalation exposure on the glucuronic acid
   pathway and activity of some xenobiotic-metabolizing enzymes. Bull Inst Marit Trop Med Gdynia 30:
   189-196.


          This document is a draft for review purposes only and does not constitute Agency policy.

                                              R-8            DRAFT—DO NOT CITE OR QUOTE

-------
                                                   Toxicological Review of Trimethylbenzene

Wiglusz. R: Delag. G: Mikulski. P. (1975a). Serum enzymes activity of mesitylene vapour treated rats. Bull
   Inst MaritTrop Med Gdynia 26: 303-313.

Wiglusz. R: Kienitz. M: Delag. G: Galuszko. E: Mikulski. P. (1975b). Peripheral blood of mesitylene vapour
   treated rats. Bull Inst Marit Trop Med Gdynia 26: 315-321.
          This document is a draft for review purposes only and does not constitute Agency policy,

                                             R-9            DRAFT—DO NOT CITE OR QUOTE

-------