Scholarly article on topic 'Carcinogenicity and Mode of Action Evaluation for Alpha-Hexachlorocyclohexane: Implications for Human Health Risk Assessment'

Carcinogenicity and Mode of Action Evaluation for Alpha-Hexachlorocyclohexane: Implications for Human Health Risk Assessment Academic research paper on "Biological sciences"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{Alpha-hexachlorocyclohexane / Hexachlorocyclohexane / "Mode of action" / "Human relevance" / "Reference dose" / "Liver carcinogenesis" / Nonlinear / Threshold / Dose-response}

Abstract of research paper on Biological sciences, author of scientific article — Ann E. Bradley, Joanna L. Shoenfelt, Judi L. Durda

Abstract Alpha-hexachlorocyclohexane (alpha-HCH) is one of eight structural isomers that have been used worldwide as insecticides. Although no longer produced or used agriculturally in the United States, exposure to HCH isomers is of continuing concern due to legacy usage and persistence in the environment. The U.S. Environmental Protection Agency (EPA) classifies alpha-HCH as a probable human carcinogen and provides a slope factor of 6.3 (mg/kg-day)−1 for the compound, based on hepatic nodules and hepatocellular carcinomas observed in male mice and derived using a default linear approach for modeling carcinogens. EPA's evaluation, last updated in 1993, does not consider more recently available guidance that allows for the incorporation of mode of action (MOA) for determining a compound's dose-response. Contrary to the linear approach assumed by EPA, the available data indicate that alpha-HCH exhibits carcinogenicity via an MOA that yields a nonlinear, threshold dose-response. In our analysis, we conducted an MOA evaluation and dose-response analysis for alpha-HCH–induced liver carcinogenesis. We concluded that alpha-HCH causes liver tumors in rats and mice through an MOA involving increased promotion of cell growth, or mitogenesis. Based on these findings, we developed a threshold, cancer-based, reference dose (RfD) for alpha-HCH.

Academic research paper on topic "Carcinogenicity and Mode of Action Evaluation for Alpha-Hexachlorocyclohexane: Implications for Human Health Risk Assessment"

Accepted Manuscript

Carcinogenicity and Mode of Action Evaluation for Alpha-Hexachlorocyclohexane: Implications for Human Health Risk Assessment

Ann E. Bradley, Joanna L. Shoenfelt, Judi L. Durda

PII: S0273-2300(15)30142-2

DOI: 10.1016/j.yrtph.2015.12.007

Reference: YRTPH 3472

To appear in: Regulatory Toxicology and Pharmacology

Received Date: 5 October 2015 Revised Date: 16 December 2015 Accepted Date: 17 December 2015

Please cite this article as: Bradley, A.E., Shoenfelt, J.L., Durda, J.L., Carcinogenicity and Mode of Action Evaluation for Alpha-Hexachlorocyclohexane: Implications for Human Health Risk Assessment, Regulatory Toxicology and Pharmacology (2016), doi: 10.1016/j.yrtph.2015.12.007.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

Carcinogenicity and Mode of Action Evaluation for Alpha-Hexachlorocyclohexane: Implications for Human Health Risk Assessment

Ann E. Bradley1

Joanna L. Shoenfelt1

Judi L. Durda1

1 Integral Consulting Inc.

Corresponding author and contact information: Ann E. Bradley Integral Consulting Inc. 61 Broadway, Suite 1601 New York, NY 10006 abradley@integral-corp.com +1-212-440-6703 (office phone) +1-410-940-9627 (cellular phone) +1-212-962-4302 (fax)

Word counts: Abstract - 193; Text - 6593; References - 2189

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

Abstract

Alpha-hexachlorocyclohexane (alpha-HCH) is one of eight structural isomers that have been used worldwide as insecticides. Although no longer produced or used agriculturally in the United States, exposure to HCH isomers is of continuing concern due to legacy usage and persistence in the environment. The U.S. Environmental Protection Agency (EPA) classifies alpha-HCH as a probable human carcinogen and provides a slope factor of 6.3 (mg/kg-day)-1 for the compound, based on hepatic nodules and hepatocellular carcinomas observed in male mice and derived using a default linear approach for modeling carcinogens. EPA's evaluation, last updated in 1993, does not consider more recently available guidance that allows for the incorporation of mode of action (MOA) for determining a compound's dose-response. Contrary to the linear approach assumed by EPA, the available data indicate that alpha-HCH exhibits carcinogenicity via an MOA that yields a nonlinear, threshold dose-response. In our analysis, we conducted an MOA evaluation and dose-response analysis for alpha-HCH-induced liver carcinogenesis. We concluded that alpha-HCH causes liver tumors in rats and mice through an MOA involving increased promotion of cell growth, or mitogenesis. Based on these findings, we developed a threshold, cancer-based, reference dose (RfD) for alpha-HCH.

Key Words: alpha-hexachlorocyclohexane, hexachlorocyclohexane, mode of action, human relevance, reference dose, liver carcinogenesis, nonlinear, threshold, dose-response

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

1 1. Introduction

2 Alpha-hexachlorocyclohexane (alpha-HCH) is one of eight structural isomers that have been used

3 worldwide as insecticides. It is a component of technical-grade HCH and a byproduct of gamma-HCH

4 (commonly called Lindane) synthesis. Prior to the late 1970s, technical-grade HCH, a mixture of the

5 alpha, beta, gamma, delta, and epsilon isomers, was used on a wide variety of food crops, seeds, and

6 ornamental plants. After that time only the gamma isomer, the isomer with the most significant

7 insecticidal activity, was used. HCH production and use have declined over time and the final remaining

8 approved uses of Lindane in the United States were cancelled by the U. S. Environmental Protection

9 Agency (EPA) in 2006. Lindane has similarly been phased out in Europe, Mexico, and Canada, although

10 it is still produced in India and possibly Russia (USEPA 2006).

11 Although no longer produced or used agriculturally in the United States, exposure to HCH isomers is of

12 continuing concern due to the historical levels of HCH use, their persistence in the environment, and

13 presence at approximately 10% of sites proposed for inclusion on the National Priority List (NPL; i.e.,

14 Superfund) (ATSDR 2005). EPA recently included alpha-HCH on its fourth Draft Contaminant

15 Candidate List (CCL 4) (USEPA 2015a), and proposed that its occurrence in public water supplies be

16 monitored as part of the 4th Unregulated Contaminant Monitoring Rule (UCMR) program (USEPA

17 2015b). By way of the UCMR alpha-HCH may require regulation under the Safe Water Drinking Act

18 (SWDA) in the future.

19 In its Integrated Risk Information System (IRIS), EPA classifies alpha-HCH as a probable human

20 carcinogen and provides a slope factor of 6.3 (mg/kg-day)-1 for the compound, based on hepatic nodules

21 and hepatocellular carcinomas observed in male mice (Ito et al. 1973a). The slope factor was derived

22 using EPA's default, linear low-dose extrapolation approach for modeling carcinogens. Last updated in

23 1993, EPA's evaluation does not consider more recently available guidance (e.g., USEPA 2005) that

24 allows for the incorporation of mode of action (MOA) for determining a compound's dose-response.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

Contrary to the linear approach assumed by EPA in their last evaluation of alpha-HCH, the available data indicate that alpha-HCH exhibits carcinogenicity via an MOA that yields a nonlinear, threshold dose-response.

The potential for alpha-HCH to cause carcinogenic effects in humans is not well studied, and animal data provide the primary basis with which to assess effects. Long-term dosing with high doses of alpha-HCH has been found to induce hyperplastic nodules and carcinomas in the livers of rats and mice (Hanada et al. 1973; Ito et al. 1973a,b; Ito et al. 1975; Ito et al. 1976; Nagasaki et al. 1975; Schulte-Hermann and Parzefall 1981; Tryphonas and Iverson 1983). Lower-doses of alpha-HCH have not been found to induce tumor formation (Hanada et al. 1973; Ito et al. 1973a,b; Ito et al. 1975). Alpha-HCH does not consistently demonstrate mutagenicity in short-term assays (Masuda et al. 2001; Puatanachokchai et al. 2006; Schroter et al. 1987; Fukushima et al. 2005; RIVM 2001; USEPA 1987) and therefore has been classified as a nongenotoxic carcinogen. Initiation-promotion studies have further demonstrated that alpha-HCH does not initiate carcinogenic activity but causes tumor formation via promotion of preneoplastic lesions (Schroter et al. 1987; Siglin et al. 1995). Patterns that suggest a growth-promoting MOA, including long time-to-tumor and reversibility of tumor formation upon cessation of exposure (USEPA 2005), have also been demonstrated for alpha-HCH (Ito et al. 1976; Schulte-Hermann and Parzefall 1981; Tryphonas and Iverson 1983).

Dose-response data for alpha-HCH further suggest the existence of a threshold below which there is no increased risk of cancer following alpha-HCH exposure. Ito et al. (1973a,b) observed increases in carcinomas in the livers of mice dosed with greater than or equal to 250 ppm (i.e., 37 mg/kg-day) alpha-HCH, but no increase in the livers of animals dosed with 50 and 100 ppm alpha-HCH. Ito et al. (1975) observed carcinomas in the liver of rats dosed with 1,000 and 1,500 ppm alpha-HCH, but no increase in rats dosed with 500 ppm alpha-HCH. Threshold and hormetic responses have also been observed for known precursors to tumor formation. For example, Puatanachokchai et al. (2006) showed a hormetic dose-response for the development of preneoplastic lesions in the liver. Specifically, the number and size

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

50 of preneoplastic lesions increased in initiated rats administered higher doses of alpha-HCH. Rats initiated

51 with diethylnitrosamine (DEN) and dosed with 0.05 ppm alpha-HCH had statistically fewer and smaller

52 preneoplastic lesions in the liver compared to rats initiated with DEN and dosed with 0.01 ppm

53 alpha-HCH. Moreover, it is generally recognized that nongenotoxic carcinogens exhibit a threshold

54 below which there is no increased risk of cancer (Butterworth 2006; Melnick et al. 1996; Williams 2008).

55 In this analysis, we conducted an MOA evaluation and dose-response analysis of the alpha-HCH data set.

56 We used guidance available from EPA and the International Programme on Chemical Safety (IPCS) on

57 evaluating the relevance of animal tumors for human health risk assessment (USEPA 2005, Meek et al.

58 2003), to frame our analysis. We used the body of literature on alpha-HCH carcinogenicity and liver

59 toxicity to describe the progression of biological events that occurs with the formation of liver tumors in

60 rats and mice, the underlying dose-response of those events, and the relevance of the effect for humans.

61 The results of the MOA analysis were used to select the modeling scheme (e.g., linear or

62 nonlinear/threshold) for the dose-response analysis. Based on the findings of the MOA analysis, we

63 developed a threshold, cancer-based reference dose (RfD) for alpha-HCH.

64 2. Methods

65 2.1 MOA Analysis

66 EPA's guidelines for carcinogen risk assessment (2005) emphasize the use of MOA in the assessment of

67 potential carcinogens. Specifically, the guidelines recognize that understanding the MOA may provide

68 important insight for determining whether a cancer hazard exists, and may help inform appropriate

69 consideration of the dose-response relationship below the range of observable tumor response. An MOA

70 for a toxicological effect is defined as a biologically plausible sequence of key events, starting with

71 interaction of an agent with a cell, proceeding through biochemical, functional, and anatomical changes,

72 that ultimately results in cancer formation or noncancer effects (Boobis et al. 2006, 2009; Meek 2008;

73 USEPA 2005). A key event is an empirically observable precursor step that is itself a necessary element

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

74 of the MOA or is a biologically based marker for such an element (USEPA 2005). This definition is

75 broader than that described in Andersen et al. (2014), which limits the definition of a key event to an

76 empirically observable precursor step and defines any reliable indicators or markers of key events as

77 "associative events." The former definition was applied for this evaluation.

78 We identified an MOA, including key events (based on EPA's definition), for alpha-HCH liver

79 carcinogenesis. The MOA and its human relevance was evaluated under the MOA framework proposed

80 by EPA (USEPA 2005), based in part on Meek et al. (2003), and applied and discussed by others (e.g.,

81 Boobis et al. 2009; Butterworth 2006; Meek 2008; Elcombe et al. 2014). The two central questions of the

82 MOA evaluation were (1) is the weight of evidence (WOE) sufficient to establish an MOA in animals?,

83 and (2) is the MOA relevant to humans? In addition, we evaluated the biochemical and physiological

84 processes underlying each key event, and available dose-response data for each key event, to determine

85 whether each is threshold based. This information was used in the dose-response assessment.

86 We evaluated our hypothesized MOA in animals by using the modified Hill criteria, including strength,

87 consistency, specificity of association; dose-response concordance; temporal relationship; and biological

88 plausibility and coherence (USEPA 2005). The possibility that an alternative MOA is responsible for

89 liver carcinogenicity in animals was also evaluated.

90 We also assessed the relevance of the MOA for humans. Using the human relevance framework (HRF)

91 proposed by Meek et al. (2003), two categories of information were evaluated to inform this analysis:

92 (1) alpha-HCH-specific data, and (2) generic information pertinent to each key event but not derived

93 from alpha-HCH specifically. Both qualitative and quantitative differences in key events between

94 animals and humans were considered.

95 2.2 Dose-Response Evaluation and Toxicity Criterion Development

96 2.2.1 General Approach

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

97 The results of the MOA evaluation were used to determine the dose-response scheme selected for

98 establishing a toxicity criterion for alpha-HCH. When a chemical acts via an MOA that is sufficiently

99 established in animals, exhibits a threshold dose-response, and is relevant to humans, EPA guidance

100 (USEPA 2005) allows for a nonlinear modeling approach to quantitatively characterize cancer risk.

101 Based on our evaluation of the MOA for liver carcinogenicity in rats and mice, we followed EPA's cancer

102 guidelines (USEPA 2005) for modeling nonlinear cancer dose-response.

103 Defining a point of departure (POD) for the critical effect is the first step in deriving a toxicity criterion

104 under EPA's MOA determination approach. EPA defines the critical effect as the first adverse effect, or

105 its known precursor, that occurs in the most sensitive species as the dose rate of an agent increases

106 (USEPA 2015c). The POD can be the lower bound on dose for an estimated incidence, a change in

107 response level from a dose-response model, or a no-observed-adverse-effect level (NOAEL) or

108 lowest-observed-adverse-effect level (LOAEL) for an observed incidence or change in level of response.

109 We selected candidate endpoints for the POD by identifying toxicological endpoints from the available

110 experimental studies that appropriately reflect or are related to liver tumors. We limited candidate

111 endpoints to those evaluated in studies of subchronic or chronic duration. Our evaluation focused on low-

112 dose studies, defined for this evaluation as studies with at least one treatment dose of 10 mg/kg bw-day

113 or less.

114 Benchmark response (BMR) levels, NOAELs, and LOAELs were all considered in the POD

115 determination. Although use of BMRs offers some advantages over NOAELs or LOAELs, which are

116 limited to experimental treatment doses, not all data types and sets are amenable to benchmark dose

117 (BMD) modeling (USEPA 1995, 2012). By considering each of these types of response levels, data that

118 were not suitable for BMD modeling were still considered in the determination of the POD.

119 Average daily doses were calculated, where necessary from dietary doses, using body weight and food

120 consumption data included in the study, or in the absence of such data, default values available from

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

USEPA (1988). Average daily animal doses (mg/kg-day) were then converted to human equivalent doses (HEDs) on the basis of three-quarters body weight scaling, which is equivalent to mg/kg%-d (milligrams of the agent normalized by the three-quarter power of body weight per day) (USEPA 2011). The default human body weight of 80 kg (USEPA 2014) was assumed:

The human equivalent dose {HED; mg/kg-day) = animal dase (mg/kg-day)

(BWh/BWa)1^

2.2.2 BMD Modeling

To model dose response we used EPA's benchmark dose software (BMDS; version 2.6.0.86) (USEPA 2012). For the low-dose studies, absolute and relative liver weights, DNA content, foci area, and foci number were selected as candidate endpoints for the POD, where data suitable for the BMDS program were available. For continuous data, EPA's standard approach is to define the BMR based on the level of change in the endpoint at which the effect becomes biologically significant (USEPA 2012). When it is not known at what level a response is considered to be adverse, a change in the mean equal to one standard deviation from the control mean may be used as the BMR (USEPA 2012). For the modeled data, a 10% increase was used as the BMR for absolute and relative liver weight, while the BMR was set at 1 standard deviation from the control mean for DNA content, foci area, and foci number. The 95% lower confidence limit on the BMD (BMDL) at the defined BMR was considered the POD. The MOA established for alpha-HCH does not prescribe any particular dose-response model for selection, nor does it reject any. Therefore, all available models for continuous data that support a nonlinear endpoint within the BMDS program were run (e.g., Hill, exponential, power, polynomial) to aid in selecting a model that best describes the data. Default program parameters were used, with the exception of constant variance. When the statistical test for variance failed, a nonconstant variance approach was employed. The p-value to determine test acceptance or rejection was set at 0.1 in accordance with EPA recommendations (USEPA 2012). Model fit was assessed considering p-value for goodness-of-fit, the Akaike information

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

criterion (AIC) value, scaled residuals near the range of the BMD, and visual inspection of the dose-response curves in the low-dose range. Among the models with adequate fit to the data, we selected the model with the lowest AIC as the basis of the BMDL if the BMDLs were within a factor of three. If the BMDLs for models with adequate data fit were not within a factor of 3, we selected the model with the lowest BMDL (USEPA 2012).

2.2.3 Selection of NOAELs and LOAELs

In determining the POD, we also considered the lowest statistically significant NOAEL or LOAEL among relevant endpoints. Where available, NOAELs were preferred and selected. For studies and endpoints for which a NOAEL was not available LOAELs were considered. Only statistically significant effect levels (i.e., p < 0.05) were selected for LOAELs. Effect levels that were statistically significant but not part of an overall dose-response trend were not selected for the LOAEL.

2.2.4 Uncertainty Factors

The last step in determining an RfD is selecting and applying uncertainty factors (UFs) to the POD to account for uncertainties associated with the available data and variability between the test species and sensitive human populations. We selected and applied the following UFs commonly used in human health risk assessment: (1) intraspecies extrapolation factor (intended to account for the variation in sensitivity among members of the human population), (2) interspecies extrapolation factor (intended to account for uncertainty involved in extrapolating from animal data to human data), (3) subchronic-to-chronic duration factor (intended to account for uncertainty involved in extrapolating from less-than-chronic NOAELs to chronic), (4) LOAEL-to-NOAEL factor (intended to account for the uncertainty involved in extrapolating from a LOAEL to a NOAEL), and (5) database factor (intended to account for the potential of deriving an underprotective RfD as the result of an incomplete characterization of the chemical's toxicity). We also considered the need for any additional modifying factors (MFs) to account for scientific uncertainties not explicitly treated by other UFs.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk Assessment December 2015 3. Results

3.1 MOA Evaluation 3.1.1 Hypothesized MOA

Using an MOA framework, the WOE supports a nonlinear dose-response with alpha-HCH causing liver tumors in rats and mice through an MOA involving increased promotion of cell growth, or mitogenesis. The key events underlying its carcinogenic action are: (1) absorption of alpha-HCH in the liver, (2) cytochrome P450 (CYP; P450) induction via receptor-mediated mechanisms, (3) increased cell proliferation, ultimately resulting in (4) benign and malignant tumor formation. Table 1 summarizes the available primary literature for alpha-HCH toxicology (aside from mutagenicity studies, which are presented in Table 3). Table 2 summarizes the evaluation of this MOA against the modified Hill criteria.

3.1.1.1 Key Event 1 - Absorption of Alpha-HCH in the Liver

Alpha-HCH is readily absorbed into the liver and has been detected in liver and other tissues after subchronic or chronic dietary exposure (Fitzhugh et al. 1950; Schroter et al. 1987). Alpha-HCH has also been isolated from mouse liver deoxyribonucleic acid (DNA), ribonucleic acid (RNA), and/or protein following a single oral bolus or intraperitoneal dose (Iverson et al. 1984; Sagelsdorff et al. 1983). The detection of alpha-HCH in the liver supports an association between exposure and the development of liver tumors.

3.1.1.2 Key Event 2 - Cytochrome P450 Induction via Receptor-Mediated Mechanisms

Liver tumor formation in rodents after exposure to nongenotoxic agents is often associated with the selective induction of hepatic microsomal P450 enzymes. This induction is triggered through activation of receptor-mediated mechanisms that lead to enhanced gene transcription. Important nuclear receptors involved in the induction of CYP1A, 2B, 3A, and 4A enzymes are, respectively, the aryl hydrocarbon, constitutive androstane receptor (CAR), the pregnane X receptor (PXR), and the peroxisome proliferator-activated receptor (PPAR-alpha). Activation of these receptors in rodents produces a cascade of

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

192 alterations in gene transcription that leads to increased hepatocellular proliferation, a critical event in the

193 development of liver tumors (Elcombe et al. 2014).

194 Upon absorption in the liver, alpha-HCH induces hepatic P450 enzymes. Increased P450 protein and

195 isozyme activity are consistently demonstrated in rats exposed to alpha-HCH, as shown in studies of

196 varying experimental design (Masuda et al. 2001; Puatanachokchai et al. 2006; Schroter et al. 1987;

197 Schulte-Hermann and Parzefall 1980; Schulte-Hermann and Parzefall 1981; Sumida et al. 2007; Werle-

198 Schneider et al. 2006). P450 activity and protein are increased in treated systems, with CYP2B and 3A

199 showing the greatest increases (Masuda et al. 2001; Puatanachokchai et al. 2006). Induction of CYP2B

200 and 3A indicate activation of the orphan nuclear receptors CAR and/or PXR. Both CAR and PXR

201 heterodimerize with the retinoid X receptor (RXR) to regulate transcription. Many of the molecules that

202 can activate CAR may also activate PXR, producing reciprocal activation of CYP2B and 3A genes

203 (Moore et al. 2000). Nicotinamide adenine dinucleotide phosphate (NADPH) P450 reductase activity has

204 been shown to increase along with total P450 levels following alpha-HCH exposure in rats (Barros et al.

205 1991; Puatanachokchai et al. 2005).

206 Increases in P450 activity and protein are both dose- and time-dependent (Masuda et al. 2001;

207 Puatanachokchai et al. 2006). Masuda et al. (2001) report increased CYP2B1 and CYP3A2 protein

208 expression and activity in animals dosed with greater than 15 ppm alpha-HCH. Puatanachokchai et al.

209 (2006) observed a hormetic dose-response curve for HCH-mediated total P450 and P450 reductase levels,

210 with significant decreases in rats initiated with DEN and dosed with 0.05 ppm alpha-HCH, and significant

211 increases in rats initiated with DEN and treated with 500 ppm alpha-HCH. Increases in CYP isoform

212 activity diminish following cessation of exposure, supporting the role of alpha-HCH as a tumor promoter,

213 for which an ultimate tumorigenic effect is due to sustained cellular change mediated by sustained

214 exposure (Schulte-Hermann and Parzefall 1981).

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

P450 induction is a threshold-based, receptor-mediated process that is regulated largely at the level of transcription. CYP inducers normally bind as ligands to the nuclear receptor; however, CAR can be activated without direct ligand binding by an indirect or ligand-independent mechanism. This indirect mechanism involves a dephosphorylation reaction that signals through the epidermal growth factor receptor and leads to the nuclear translocation of CAR. Regardless of whether nuclear receptors are activated through direct or indirect mechanisms, the effectiveness of receptor-mediated transcription induction depends on an array of factors such as the affinity of the xenobiotic for receptors and the presence of co-activators or co-repressors (Kohn and Melnick 2002). In such a multifactorial process there are doses of inducer at which no measurable response would occur, as has been demonstrated for alpha-HCH.

Increases in total P450, P450 isoform protein/activity, and P450 reductase activity have been observed in parallel with increased proliferation or increased liver weight, and hepatic foci formation (Barros et al. 1991; Masuda et al. 2001; Puatanachokchai et al. 2006; Sumida et al. 2007).

3.1.1.3 Key Event 3 - Cellular Proliferation

Exposure to alpha-HCH results in increased hepatocellular proliferation, as evidenced by increased liver DNA synthesis, hypertrophy, and hyperplasia, leading to increased relative liver weight in both rats and mice. DNA synthesis is increased in hyperplastic and, in some cases, normal liver tissue from exposed mice (Gerlyng et al. 1994; Schulte-Hermann et al. 1981; Siglin et al. 1991, 1995; Tryphonas and Iverson 1983) and is observed after single oral doses or long-term dietary exposure to alpha-HCH (Schroter et al. 1987; Schulte-Hermann et al. 1981, 1983).

The observed increase in DNA synthesis occurs in hepatocytes of differing chromosomal content; however, the proportion of binuclear hepatocytes decreases during mitogenesis (Gerlyng et al. 1994). Decreased binucleation is indicative of loss of terminal differentiation and entrance into an aberrant pattern of cell growth (Guidotti et al. 2003).

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

239 Following long-term alpha-HCH exposure, observed increases in hypertrophy, gamma-glutamyl

240 transpeptidase (GGT) positive foci formation, DNA content, and liver weight are dose-dependent

241 (Fitzhugh et al. 1950; Goto et al. 1972; Ito et al. 1973a,b; Luebeck et al. 1975; Masuda et al. 2001;

242 Schroter et al. 1987). Alpha-HCH-mediated increases in centrilobular hyperplasia or hypertrophy, liver

243 DNA and RNA content, and liver weight regress after cessation of exposure (Angsubhakorn et al. 1981;

244 Ito et al. 1976; Kraus et al. 1981; Schulte-Hermann and Parzefall 1981), further supporting the tumor-

245 promoting effect of the chemical.

246 3.1.1.4 Key Event 4 - Benign and Malignant Tumor Formation

247 Chronic administration of high doses of alpha-HCH has been shown to produce both benign and

248 malignant tumors in mice and rats (Hanada et al. 1973; Ito et al. 1973a,b; Ito et al. 1975; Ito et al. 1976;

249 Nagasaki et al. 1975; Schulte-Hermann and Parzefall 1981; Tryphonas and Iverson 1983), with increased

250 sensitivity for mice (Nagasaki et al. 1975). Tumor formation exhibits a threshold response. For example,

251 Ito et al. (1973a,b) observed increases in carcinomas in the livers of mice dosed with greater than or equal

252 to 250 ppm alpha-HCH, but no increase in the livers of animals dosed with 50 and 100 ppm alpha-HCH.

253 Ito et al. (1975) observed carcinomas in the liver of rats dosed with 1,000 and 1,500 ppm alpha-HCH, but

254 no increase in rats dosed with 500 ppm alpha-HCH.

255 3.1.1.5 Alternative MOAs

256 From a human risk perspective, the most crucial potential alternative MOA for alpha-HCH is

257 mutagenicity. Although some evidence of genotoxicity has been observed, the lack of a consistent

258 positive response in the short-term bioassays conducted in a variety of in vitro and in vivo systems,

259 evaluating a variety of endpoints associated with DNA damage, does not support the notion that alpha-

260 HCH is mutagenic. Table 3 presents a summary of the available short-term bioassays for mutagenic

261 potential of alpha-HCH.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

262 Results of four in vitro assays testing for gene mutation at a variety of concentrations, both with and

263 without metabolic activation, were negative. Alpha-HCH has shown the ability to bind to DNA in both in

264 vitro and in vivo test systems, however the levels of DNA binding were weak (Iverson et al. 1984;

265 Sagelsdorff et al. 1983). Sagelsdorff et al. (1983) characterize it as "minute DNA binding," stating that

266 "the level of binding is more than three orders of magnitude lower than would be expected if the

267 mechanism of tumor induction was genotoxicity." Several assays measuring DNA damage or

268 fragmentation, or repair of such damage, showed mixed results. Kalantzi et al. (2004) report positive

269 results for a comet assay for DNA fragmentation performed with high doses of alpha-HCH but note that

270 at lower concentrations, no comet-forming effects were observed (data not shown by authors). Mattioli et

271 al. (1996) found alpha-HCH-induced, dose-dependent strand breaks in vitro in rat and human

272 hepatocytes; however, no DNA strand breaks were observed in mouse hepatocytes. DNA synthesis,

273 indicative of lesion repair or general proliferation, was not increased in rat or human hepatocytes despite

274 observed DNA fragmentation, which suggests that any induced damage was not sufficient to elicit a

275 detectable repair response (Mattioli et al. 1996). Venkat et al. (1995) provide results on a relative scale of

276 activity for the induction of gene pathways involved in DNA repair. The study reports that alpha-HCH

277 showed activity levels ranging from one-tenth to one-fourth of 4-Nitro-quinoline oxide (4-NQO), which is

278 considered a direct-acting mutagen. The single study we reviewed for chromosomal abnormalities reports

279 positive results in liver cells from Donryu rats given a single dose of 600 ppm in vivo (Hitachi et al.

280 1975).

281 Liver tumors can be induced in rodents by cytotoxic agents; however, examination of the literature

282 reveals that induced tumors in rats and mice are observed at dose levels that are not associated with any

283 evidence of marked hepatotoxicity.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

284 3.1.2 Human Relevance

285 In deriving a cancer slope factor (CSF) for alpha-HCH-induced liver cancer, EPA assumes that

286 alpha-HCH is a probable human carcinogen (i.e., classification B2). Given EPA's current assumption,

287 the animal cancer response and MOA were assumed to be relevant for humans for the purposes of this

288 dose-response evaluation, including the derivation of a toxicity criterion for the protection of public

289 health. As such, we developed a threshold-based RfD for the protection of alpha-HCH-induced cancer.

290 The current state of knowledge regarding the human relevance of the defined MOA for alpha-HCH

291 hepatocarcinogenicity is presented below.

292 Activation of nuclear receptors and subsequent induction of P450 is a well-known MOA for rodent

293 hepatocarcinogenesis. Activation of CAR, PXR, and PPAR-alpha produces a cascade of alternations in

294 gene transcription that leads to increased hepatocellular proliferation, a critical event in the development

295 of liver tumors (Elcombe et al. 2014; Klaunig et al. 2003; Cohen et al. 2010; Williams and Iatropoulos

296 2002; Hall et al. 2012).

297 Compound-specific evidence regarding the human relevance of this MOA for alpha-HCH-induced

298 hepatocarcinogenesis is sparse; however, information from other well studied compounds informs the

299 human relevance evaluation for alpha-HCH. For example, phenobarbital (PB) is a well studied liver

300 carcinogen that operates via a receptor-mediated mechanism that alters gene transcription, leading to

301 increased cellular proliferation and eventually to the development of liver tumors (Whysner et al. 1996);

302 similarities in the toxic activity of alpha-HCH and PB toward the liver have been observed in Schroter et

303 al. (1987), Schulte-Hermann et al. (1981), Schulte-Hermann et al. (1983), Werle-Schneider et al. (2006),

304 and Fukushima et al. (2005). Key events for PB-induced liver carcinogenesis in rodents have been

305 described by Elcombe et al. (2014) to include (1) activation of CAR, (2) altered gene expression specific

306 to CAR activation, (3) increased cell proliferation, (4) clonal expansion leading to altered foci, and (4)

307 formation of liver adenomas/carcinomas. Exposure to PB in humans likewise leads to activation of CAR

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

and PXR and to the induction of P450; however, a different pattern of response is induced in humans compared to rodents (reviewed in Elcombe et al. 2014; Lake 2009). No evidence of increased hepatocellular proliferation in humans or in primary human hepatocytes in vitro exists (Elcombe et al. 2014; Lake 2009; Parzefall et al. 1991). A single in vitro study also reports that alpha-HCH does not induce hepatocellular proliferation in human hepatocytes (Parzefall et al. 1991). Experiments with transgenic mice, in which mouse receptors have been replaced with their human counterparts, have shown conflicting results. Elcombe et al. (2014) provide a summary of two such studies. In the first, wild-type mice treated with PB-induced P450 enzymes produced hepatocellular hypertrophy, increased DNA synthesis, and increased liver weight, whereas humanized mice, in which the CAR/PXR receptors were replaced with hCAR/hPXR induced P450 enzymes and produced hepatocellular hypertrophy but did not increase replicative DNA synthesis. In the second study, treatment of hCAR mice with PB resulted in increases in P450 mRNA levels, relative liver weight, and cell proliferation. Elcombe et al. (2014) hypothesize that the difference in results may be explained by differences in the models developed or in the treatment regimens for the two studies; however, firm conclusions regarding the human relevance of the receptor-mediated pathway for PB -induced liver carcinogenesis cannot be drawn. There is no evidence of liver cancer in humans following pharmaceutical administration of PB (as summarized by Elcombe et al. 2014).

Beyond alpha-HCH- and PB- specific information, experts have opined, more generally, on what constitutes a relevant adverse effect in the context of hepatocellular hypertrophy. An expert panel convened by the European Society of Toxicologic Pathology (ESTP) concluded that observations in chemically exposed laboratory rodents of increased liver weight, hepatocellular hypertrophy, and cell proliferation in the absence of overt hepatotoxicity and mediated via CAR, PXR, or PPAR-alpha should be considered a "non-adverse" effect for human exposures (Hall et al. 2012). Under these conditions, hepatocellular changes are considered to be fully reversible and do not compromise viability or functional integrity (Hall et al. 2012).

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

333 3.2 Dose-Response Analysis and Toxicity Criterion Development

334 Increased incidence of liver adenomas and carcinomas (alpha-HCH) and gross histological changes to the

335 liver (gamma- and beta-HCH) have been observed at relatively high doses above 150 ppm (Ito et al.

336 1973a,b, 1975, 1976; Hanada et al. 1973; Nagasaki et al. 1975; Schulte-Hermann and Parzefall 1981).

337 The lowest administered doses for which tumor incidence was observed are equivalent to average daily

338 doses on the order of 45 mg/kg-day (Ito et al. 1973a,b) and an initial dose of 100 mg/kg followed by 18.4

339 mg/kg-day for approximately 13.5 months (Schulte-Hermann and Parzefall 1981). Lower-dose studies,

340 many of which were designed to measure the progression of and/or mechanisms responsible for alpha-

341 HCH toxicity, demonstrate that known or hypothesized precursors to hepatocarcinogenicity or

342 hepatotoxicity occur at lower doses. Our evaluation focused on low-dose studies, defined here as studies

343 with at least one animal treatment dose of 10 mg/kg-day or less.

344 Endpoints considered for the POD included increased liver weight (absolute and relative measures),

345 increased DNA content, and increased number or area of preneoplastic foci. Each of these endpoints

346 potentially reflects the occurrence of key events associated with alpha-HCH hepatocarcinogenicity and

347 may be thought of as precursors to alpha-HCH-induced hepatocarcinogenicity. Nonspecific, early

348 markers of exposure, including P450 levels and activity, were not considered as endpoints in the POD

349 determination.

350 Table 4 summarizes the findings of the BMD modeling. The lowest statistically significant NOAEL that

351 was appropriate for the POD determination from each study was also considered. For studies for which

352 no NOAEL was available, a LOAEL was selected. Table 5 summarizes the available NOAELHEDs and

353 LOAELHEDs considered in determining the POD.

354 3.2.1 BMD Modeling Results

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

355 Four studies (Puatanochokchai et al. 2006; Schroter et al. 1987; Masuda et al. 2001; and Sumida et al.

356 2007) contained endpoints that were suitable for BMD modeling. All selected endpoints had at least one

357 model that provided an adequate fit to the data, and the BMDLHEDs from the best-fit models selected for

358 each modeled endpoint ranged from 0.049 to 2.2 mg/kg-day. BMDLHEDs associated with foci formation

359 (increased number of foci, and increased area of foci ) fell in the lower end of this range (0.049 to 0.072

360 mg/kg-day), while those for DNA content and liver weight fell in the higher end of the range (0.22 to 2.2

361 mg/kg-day) (Table 4).

362 3.2.2 NOAELs/LOAELs

363 Fitzhugh et al. (1950) established a LOAEL of 3.7 mg/kg-day and a NOAEL of 0.74 mg/kg-day,

364 equivalent to a NOAELhed of 0.20 mg/kg-day, in males for slight microscopic changes and increased

365 liver weight. A slightly higher effect level was measured in female rats (LOAEL of 4.2 mg/kg-day and a

366 NOAEL of 0.84 mg/kg-day). The NOAEL of 0.74 mg/kg-day provided the basis for one of the two RfDs

367 derived by EPA in the 2006 evaluation of other HCH isomers, which was completed as part of the

368 reregistration eligibility decision (RED) for Lindane (USEPA 2006). Sumida et al. (2007) established a

369 slightly lower LOAEL of 2 mg/kg-day (equivalent to a LOAELHED of 0.38 mg/kg-day) for increased

370 relative liver weight and alanine aminotransferase (ALT) following 28 days of exposure to alpha-HCH in

371 male rats. The LOAEL was observed at the lowest dose tested, and no NOAEL was established for these

372 endpoints.

373 Masuda et al. (2001), Puatanachokchai et al. (2006), and Schroter et al. (1987) established LOAELs for

374 changes in foci formation (i.e., number and area of hepatic foci) in rats. In all three studies, animals were

375 dosed with alpha-HCH after exposure to a known initiator, and then changes were measured after

376 subchronic or nearly subchronic exposures. Although the study designs are variable, there is an

377 observable consistency in the LOAELHEDs identified across the three studies measuring foci formation,

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

378 with values ranging from 0.75 to 2.8 mg/kg-day, equivalent to LOAELHEDs from 0.21 to 0.70 mg/kg-day.

379 The range of NOAELHEDs associated with these effects is 0.014 to 0.12 mg/kg-day.

380 3.2.3 Synthesis

381 The BMD modeling results and NOAELs/LOAELs were considered for selecting the POD for alpha-

382 HCH. The collective results demonstrate that endpoints, including the incidence of hepatic foci and foci

383 area, were the most sensitive, while changes in liver weight and DNA content were less sensitive. All of

384 the effects measured in the low-dose range were significantly lower than dose levels at which tumor

385 formation was observed. Effect levels associated with foci formation were therefore considered to be

386 conservatively appropriate for the basis of a POD.

387 BMD modeling established a range of BMDLHEDs of 0.049 to 0.072 mg/kg-day for increased foci area

388 and number, based on findings from Schroter et al. (1987) and Parzefall (2010). Three studies reported

389 NOAELs for these effects. The lowest NOAELhed from the three studies that measured an endpoint for

390 foci development was 0.014 mg/kg-day (Puatanachokchai et al. 2006); however, as evidenced by a

391 comparison of LOAELs across the three studies (Puatanachokchai et al. 2006; Masuda et al. 2001;

392 Schroter et al. 1987) (Table 5), the significantly lower NOAELHED from Puatanachokchai is likely an

393 effect of the large intervals in the low end of the dosing regimen. The NOAELHEDs from Masuda et al.

394 (2001) and Schroter et al. (1987) were an order of magnitude higher, at 0.11 and 0.12 mg/kg-day,

395 respectively.

396 Data from Masuda et al. (2001) and Puatanachokchai et al. (2006) for foci number and area were not

397 suitable for BMD modeling because these data were presented only in graphical format. Given the

398 recognized advantages that the BMD method holds over the use of NOAELs and LOAELs in establishing

399 PODs (USEPA 2012), preference was given to the BMDLHEDs established from Schroter et al. (1987),

400 i.e., 0.049 and 0.072 mg/kg-day for increases in foci area and foci number, respectively (Table 4; Figures

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

401 1 and 2). Based on the average of these two results, the HED POD determined for alpha-HCH is 0.061

402 mg/kg-day.

403 3.2.4 Application of Uncertainty Factors to the POD

404 The following UFs were applied to the POD in determining the cancer-based RfD for alpha-HCH:

405 • Intraspecies Extrapolation Factor - A default value of 10 was selected for this factor to

406 account for the variation in sensitivity among members of the human population. The default

407 value was selected in the absence of any data on the toxicity of alpha-HCH in humans.

408 • Interspecies Extrapolation Factor - A value of 3 was selected for this factor. One standard

409 approach is to apply an UF of 10 to account for toxicokinetic and toxicodynamic differences in

410 the equivalent dose between animals and humans. However, the allometric scaling recommended

411 in EPA's cancer guidelines (USEPA 2005) and employed in the estimation of the POD accounts

412 for differences in toxicokinetics between humans and animals and also for some toxicodynamic

413 variability (USEPA 2006). In situations where such allometric adjustments are employed, EPA

414 recommends that an UF of 3 be applied to the allometrically scaled dose to account for

415 toxicodynamic differences between animals and humans (USEPA 2005, 2006).

416 • Subchronic-to-Chronic Duration Factor - A value of1 was selected for this factor. Although

417 the data selected for the POD were derived from subchronic and nearly subchronic studies of 28

418 days and 20 weeks, the selected endpoint for the critical effect is not a toxic manifestation, but

419 rather precursors that occur at both lower doses and earlier time points compared to the final toxic

420 manifestation of tumor formation determined for alpha-HCH. Given that longer exposures to

421 similar doses of alpha-HCH have not been shown to manifest toxic lesions in mice or rats, this

422 precursor event is considered to be a conservative basis for the toxicity criterion and intended to

423 be protective against cancer effects. Therefore, no UF to account for effects seen at less than

424 chronic durations was applied.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

• LOAEL-to-NOAEL Factor - A value of 1 was selected for this factor, as the final POD was not based on a LOAEL but rather a BMDL.

• Database Uncertainty Factor - A value of 1 was selected for this factor, as the available database was considered complete. Significant data gaps that would affect the determination of the critical effect and the POD for that critical effect were not identified.

• Additional Modifying Factors - No additional modifying factors were determined to be necessary for the derivation of the RfD.

3.2.5 Alpha-HCH RfD

Based on the MOA evaluation and dose-response evaluation presented, we recommend an RfD for alpha-HCH of 0.002 mg/kg-day. The value is based on a POD of 0.061 mg/kg-day for development (incidence and size) of preneoplastic hepatic foci in rats, and a combined UF of 30 (10 for intraspecies extrapolation and 3 for interspecies extrapolation).

4. Discussion and Conclusions

We completed an MOA evaluation and dose-response assessment for alpha-HCH-induced liver carcinogenesis and developed a supportable toxicity criterion for the protection of human health in line with established guidelines and published frameworks (e.g., USEPA 2005, Meek et al. 2003). We conclude that the WOE supports a nonlinear dose-response with alpha-HCH causing liver tumors in rats and mice through an MOA involving increased promotion of cell growth, or mitogenesis. The key events underlying its carcinogenic action are: (1) absorption of alpha-HCH in the liver, (2) P450 induction via receptor-mediated mechanisms, (3) increased cell proliferation, ultimately resulting in (4) benign and malignant tumor formation. Although the human relevance of this MOA remains a matter of scientific debate, EPA currently treats alpha-HCH as a human carcinogen to regulate human exposures.

The nonlinear RfD for alpha-HCH that we established in this paper of 0.002 mg/kg-day is approximately 100 to 10,000 times greater than the risk specific dose (RSD) values corresponding with EPA's IRIS

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

449 excess cancer risks of 10-4 and 10-6, respectively. Although U.S. regulatory policy suggests that there are

450 different definitions of what constitutes de minimis risk, in general, the point of little-to-no concern is at

451 the 10-6 risk level, ranging upward to 10-4, which corresponds to EPA's target risk range in the

452 Comprehensive Environmental Response, Compensation, and Liability Act program, under the National

453 Oil and Hazardous Substances Pollution Contingency Plan, (also known as the National Contingency

454 Plan). Environmental standards established upon the basis of risk would thus be several orders of

455 magnitude higher than those currently employed based on a linear cancer model should the resultant RfD

456 established here be applied.

457 Although sufficient data are available to demonstrate that alpha-HCH is a rodent carcinogen that operates

458 via a threshold based mechanism/MOA there are some uncertainties in the underlying toxicological

459 studies (Table 1) that could be reduced with further study. Studies investigating gene expression for

460 alpha-HCH alongside other known CAR and PXR inducers would help to bolster the defined MOA.

461 Studies measuring early indicators of tumor formation in rodents, including foci growth, would help to

462 reduce uncertainties in the RfD established. Further, additional data to help establish the human

463 relevance of CAR and PXR induced effects for alpha-HCH along with other similarly acting chemicals

464 would help to inform whether this compound should be regarded as a potential human carcinogen and

465 address the on-going scientific debate regarding the significance of cancers mediated via receptor

466 mediated mechanisms (Elcombe et al. 2014, Hall et al. 2012).

467 Exposure to alpha-HCH is of continuing concern due to the historical levels of HCH use, and its

468 persistence in the environment. EPA recently included alpha-HCH on the CCL4 (USEPA 2015a) and

469 proposed it for inclusion in the 4th UCMR program (USEPA 2015b), suggesting the possibility that alpha-

470 HCH may require regulation under the SWDA in the future. Given the regulatory focus on alpha-HCH

471 regulatory agencies should review the scientific evidence that supports treating alpha-HCH as a threshold

472 based toxicant.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

473 Precedent for establishing threshold based criteria for evaluating the carcinogenicity of a compound based

474 on the compound's MOA and EPA's cancer guidelines (USEPA 1996, 2005) exists. EPA uses a

475 nonlinear, threshold approach for characterizing cancer risks to chloroform (USEPA 2001, 2015b) based

476 on the compound's MOA. The WOE supports that chloroform does not produce rodent tumors via a

477 mutagenic MOA, but rather carcinogenic responses observed in animals are associated with regenerative

478 hyperplasia that occurs in response to cytolethality and cytolethality occurs only at exposure levels above

479 some critical dose level (USEPA 2001). Also based on the compounds' MOAs, EPA has determined that

480 a threshold for carcinogenicity exists for captan (69 Fed. Reg. 68357) and ethylene glycol monobutyl

481 ether (EGBE) (USEPA 2010, 2015c).

482 Acknowledgements

483 Funding for the preparation of this article was provided by Syngenta Crop Protection, Inc., Honeywell,

484 Olin Corporation, and FMC Corporation

485 References

486 Andersen, M.E., R.J. Preston, A. Maier, A.M. Willis, and J. Patterson. 2014. Dose-response approaches

487 for nuclear receptor-mediated modes of action for liver carcinogenicity: results of a workshop. Crit Rev

488 Toxicol. 44(1):50-63.

489 Angsubhakorn, S., N. Bhamarapravati, K. Romruen, S. Sahaphong, W. Thamavit, and M. Miyamoto.

490 1981. Further study of a benzene hexachloride inhibition of aflatoxin B1 hepatocarcinogenesis in rats.

491 Br. J. Cancer. 43:881-883.

492 ATSDR. 2005. Toxicological profile for alpha-, beta-, gamma-, and delta-hexachlorocyclohexane. U.S.

493 Department of Health and Human Services, Public Health Service, Agency for Toxic Substances and

494 Disease Registry. 377 pp. August.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

495 Barros, S.B.M., K. Simizu, and V.B.C. Junqueira. 1991. Liver lipid peroxidation-related parameters

496 after short-term administration of hexachlorocyclohexane isomers to rats. Toxicol. Letters. 56:137-144.

497 Boobis, A.R., S.M. Cohen, V. Dellarco, D. McGregor, M.E. (Bette) Meek, C. Vickers, D. Willcocks, and

498 W. Farland. 2006. IPCS framework for analyzing the relevance of a cancer mode of action for humans.

499 Crit. Rev. Toxicol. 36:781-792.

500 Boobis, A.R., G.P. Daston, R.J. Preston, and S.S. Olin. 2009. Application of key events analysis to

501 chemical carcinogens and noncarcinogens. Crit. Rev. FoodSci. Nutrition. 49(8):690-707.

502 Butterworth, B.E. 2006. A classification framework and practical guidance for establishing a mode of

503 action for chemical carcinogens. Reg. Toxicol. Pharma. 45:9-23.

504 Cohen, S. 2010. Evaluation of possible carcinogenic risk to humans based on liver tumors in rodent

505 assays: the two-year bioassay is no longer necessary. Toxicol. Path. 38:487-501.

506 Elcombe, C.R., R.C. Peffer, D C. Wolf, J. Bailey, R. Bars, D. Bell, R.C. Cattley, S.S. Ferguson, D. Geter,

507 A. Goetz, J.I. Goodman, S. Hester, A. Jacobs, C.J. Omiecinski, R. Schoeny, W. Xie, and B. G. Lake.

508 2014. Mode of action and human relevance analysis for nuclear receptor-mediated liver toxicity: A case

509 study with phenobarbital as a model constitutive androstane receptor (CAR) activator. Crit Rev Toxicol.

510 44(1):64-82.

511 Fitzhugh, O.G., A.A. Nelson, and J.P. Frawley. 1950. The chronic toxicities of technical benzene

512 hexachloride and its alpha, beta, and gamma isomers. 59-66.

513 Fukushima, S., A. Kinoshita, R. Puatanachokchai, M. Kushida, H. Wanibuchi, and K. Morimura. 2005.

514 Hormesis and dose-response-mediated mechanisms in carcinogenesis: evidence for a threshold in

515 carcinogenicity of non-genotoxic carcinogens. Carcinogenesis. 26(11): 1835-1845.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

Gerlyng, P., T. Grotmol, and P.O. Seglen. 1994. Effect of 4-acetylaminofluorene and other tumour promoters on hepatocellular growth and binucleation. Carcinogenesis. 15(2):371-379.

Goto, M., M. Hattori, T. Miyagawa, and M. Enomoto. 1972. Beitrage zur okologischen chemie. II. Hepatoma-bildung in mausen nach Verabreichung von hch-isomeren in hohen dosen. Chemosphere. 6:279-282.

Guidotti, J.-E., O. Bregerie, A. Robert, P. Debey, C. Brechot, and C. Desdouets. 2003. Liver cell polyploidization: A pivotal role for binuclear hepatocytes. J. Biol. Chem. 278(21): 19095-19101.

Hanada, M., C. Yutani, and T. Miyaji. 1973. Induction of hepatoma in mice by benzene hexachloride. Gann. 64(5):511-513.

Hall, A.P., C R. Elcombe, J.R. Foster, T. Harada, W. Kaufmann, A. Knippel, K. Kuttler, D.E. Malarkey, R.R. Maronpot, A. Nishikawa, T. Nolte, A. Schulte, V. Strauss, and M.J. York. 2012. Liver hypertrophy: A review of adaptive (adverse and non-adverse) changes - conclusions from the 3rd International ESTP Expert Workshop. Toxicol Path. 40:971-994.

Hitachi, P.M., K. Yamada, and S. Takayama. 1975. Cytologic changes induced in rat liver cells by short-term exposure to chemical substances. J. Natl. Cancer Inst. 54:1245-1247.

Ito, N., H. Nagasaki, M. Arai, S. Sugihara, and S. Makiura. 1973a. Histologic and ultrastructural studies on the hepatocarcinogenicity of benzene hexachloride in mice. J. Natl Cancer Inst. 51:817-826.

Ito, N., H. Nagasaki, M. Arai, S. Makiura, S. Sugihara, and K. Hirao. 1973b. Histopathologic studies on liver tumorigenesis induced in mice by technical polychlorinated biphenyls and its promoting effect on liver tumors induced by benzene hexachloride. J. Natl Cancer Inst. 51(5): 1637-1646.

Ito, N., H. Nagasaki, H. Aoe, S. Sugihara, Y. Miyata, M. Arai, and T. Shirai. 1975. Development of hepatocellular carcinomas in rats treated with benzene hexachloride. J. Natl Cancer Inst. 54:801-805.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

Ito, N., M. Hananouchi, S. Sugihara, T. Shirai, H. Tsuda, S. Fukushima, and H. Nagasaki. 1976. Reversibility and irreversibility of liver tumors in mice induced by the a isomer of 1,2,3,4,5,6-hexachlorocyclohexane. Cancer Res. 36:2227-2234

Ito, N., H. Tsuda, R. Hasegawa, and K. Imaida. 1983. Comparison of the promoting effects of various agents in induction of preneoplastic lesions in rat liver. Environ. Hlth. Perspect. 50:131138.

Iverson, F., J.J. Ryan, R. Lizotte, and S.L. Hierlihy. 1984. In vivo and in vitro binding of a- and y-hexachlorocyclohexane to mouse liver macromolecules. Toxicol. Letters. 20:331-335.

Kalantzi, O.I., R. Hewitt, K.J. Ford, L. Cooper, R.E. Alcock, G O. Thomas, J.A. Morris, T.J. McMillan, K.C. Jones, and F.L. Martin. 2004. Low dose induction of micronuclei by lindane. Carcinogenesis. 25(4):613-622.

Klaunig, J.E., M.A. Babich, K.P.Baetcke, J.C. Cook, J.C. Corton, R.M. David, J.G. DeLuca. D.Y. Lai, R.H.McKee, J.M. Peters, R.A. Roberts, P.A. Fenner-Crisp. 2003. PPAR-alpha agonist-induced rodent tumors: modes of action and human relevance. Crit. Rev. Toxicol. 33(6):655-780.

Kohn, M.C., and R.L. Melnick. 2002. Biochemical origins of the non-monotonic receptor-mediated dose-response. J. Mol. Endocrinology. 29:113-123.

Kraus, P., B. Gross, and H-D. Kloft. 1981. The elevation of rat liver glutathione-S-transferase activity by alpha-hexachlorocyclohexane. Biochem. Pharm. 30:355-361.

Lake, B.G. 2009. Species differences in the hepatic effects of inducers of CYP2B and CYP4A subfamily forms: relationship to rodent liver tumour formation. Xenobiotica. 39(8):582-596.

Lee, C.H., and A.M. Edwards. 2001. Stimulation of DNA synthesis by tumor promoters in primary rat hepatocytes is not mediated by arachidonic acid metabolites. J. Cellular Physiol. 187:336-344.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

560 Luebeck, E.G., B. Grasl-Kraupp, I. Timmermann-Trosiener, W. Bursch, R. Schulte-Hermann, and S.H.

561 Moolgavkar. 1995. Growth kinetics of enzyme-altered liver foci in rats treated with phenobarbital or a-

562 hexachlorocyclohexane. Toxicol. Appl. Pharmacol. 130:304-315.

563 Masuda, C., H. Wanibuchi, K. Otori, M. Wei, S. Yamamoto, T. Hiroi, S. Imaoka, Y. Funae, and S.

564 Fukushima. 2001. Presence of a no-observed effect level for enhancing effects of development of the a-

565 isomer of benzene hexachloride (a-BHC) on diethylnitrosamine-initiated hepatic foci in rats. Cancer

566 Letters. 163:179-185.

567 Mattioli, F., L. Robbiano, D. Adamo, R. Federa, A. Martelli, and G. Brambilla. 1996. Genotoxic effects

568 of a-hexachlorocyclohexane in primary cultures of rodent and human hepatocytes. Mutagenesis.

569 11(1):79-83.

570 Meek, M.E. Bette. 2008. Recent developments in frameworks to consider human relevance of

571 hypothesized modes of action for tumours in animals. Environ. Molecular Mutagenesis. 49:110-116.

572 Meek, M.E. (Bette), J.R. Bucher, S.M. Cohen, V. Dellarco, R.H. Hill, L.D. Lehman-McKeeman, D.G.

573 Longfellow, T. Pastoor, J. Seed, and D.E. Patton. 2003. A framework for human relevance analysis of

574 information on carcinogenic modes of action. Crit. Rev. Toxicol. 33(6):591-653.

575 Melnick, R.L., M.C. Kohn, and C.J. Portier. 1996. Implications for risk assessment of suggested

576 nongenotoxic mechanisms of chemical carcinogenesis. Environ. Hlth Perspect. 104(Suppl 1): 123-134.

577 Moore, L.B, D.J. Parks, S.A. Jones, R.K. Bledsoe, T.G. Consler, J. B. Stimmel, B. Goodwin, C. Liddle,

578 S.G. Blanchard, T.M. Wilson, J.L. Collins, and S.A. Kliewer. 2000. Orphan nuclear receptors constitutive

579 androstane receptor and pregnane X receptor share xenobiotic and steroid ligands. J. Biol. Chem.

580 275(20): 15122-15127.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

Moriya, M., T. Ohta, K. Watanabe, T. Miyazawa, K. Kato, and Y. Shirasu. 1983. Further mutagenicity studies on pesticides in bacterial reversion assay systems. Mut. Res. 116:185-216.

Nagasaki, H., H. Kawabata, Y. Miyata, K. Inoue, K. Hirao, H. Aoe, and N. Ito. 1975. Effect of various factors on induction of liver tumors in animals by the a-isomer of benzene hexachloride. Gann. 66:185191.

Parzefall, W., E. Erber, R. Sedivy, and R.Schulte-Hermann. 1991. Testing for induction of DNA synthesis in human hepatocyte primary cultures by rat liver tumor promoters. Canc. Res. 51:1143-1147.

Parzefall, W. 2010. Personal communication (email to R. Schulte-Hermann, and J. Shoenfelt, Integral Consulting Inc., Annapolis, MD, dated April 13, 2010, regarding data from Schroter et al. 1987 published study). Medical University Vienna, Vienna, Austria.

Puatanachokchai, R., K. Morimura, H. Wanibuchi, M. Oka, A. Kinoshita, F. Mitsuru, S. Yamaguchi, Y. Funae, and S. Fukushima. 2006. Alpha-benzene hexachloride exerts hormesis in preneoplastic lesion formation of rat hepatocarcinogenesis with the possible role for hepatic detoxifying enzymes. Cancer Letters. 240:102-113.

RIVM. 2001. Re-evaluation of human-toxicological maximum permissible risk levels. RIVM Report 711701 025. Rijks Instituut Voor Volksgezondheid en Milieu, National Institute of Public Health and the Environmental, Bilthoven, The Netherlands. 297 pp. March.

Sagelsdorff, P., W.K. Lutz, and C. Schlatter. 1983. The relevance of covalent binding to mouse liver DNA to the carcinogenic action of hexachlorocyclohexane isomers. Carcinogenesis. 4(10): 1267-1273.

Schroter, C., W. Parzefall, H. Schroter, and R. Schulte-Hermann. 1987. Dose-response studies on the effects of a-, P-, and y-hexachlorocyclohexane on putative preneoplastic foci, monooxygenases and growth in rat liver. Cancer Res. 47:80-88.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

603 Schulte-Hermann, R., and W. Parzefall. 1980. Adaptive responses of rat liver to the gestagen and anti-

604 androgen cyproterone acetate and other inducers. I. Induction of drug-metabolizing enzymes. Chem.

605 Biol. Interactions. 31:279-286.

606 Schulte-Hermann, R., and W. Parzefall. 1981. Failure to discriminate initiation from promotion of liver

607 tumors in a long-term study with the phenobarbital-type inducer a-hexachlorocyclohexane and the role of

608 sustained stimulation of hepatic growth and monooxygenases. Cancer Res. 41:4140-4146.

609 Schulte-Hermann, R., G. Odhe, J. Schuppler, and I. Timmermann-Trosiener. 1981. Enhanced

610 proliferation of putative preneoplastic cells in rat liver following treatment with the tumor promoters

611 phenobarbital, hexachlorocyclohexane, steroid compounds, and nafenopin. Cancer Res. 41:2556-2562.

612 Schulte-Hermann, R., J. Schuppler, I. Timmermann-Trosiener, G. Ohde, W. Bursch, and H. Berger.

613 1983. The role of growth of normal and preneoplastic cell populations for tumor promotion in rat liver.

614 Environ. Hlth Perspect. 50:185-194.

615 Shahin, M.M., and R.C. von Borstel. 1977. Mutagenic and lethal effects of a-benzene

616 hexachloride, dibutyl phthalate and trichloroethylene in Sacchaoromyces cerevisiae. Mut. Res.

617 48:173-180.

618 Siglin, J.C., C.M. Weghorst, and J.E. Klaunig. 1991. Role of hepatocyte proliferation in a-

619 hexachlorocyclohexane and phenobarbital tumor promotion in B6C3F1 mice. Chemically Induced Cell

620 Proliferation: Implications for Risk Assessment. Prog. Clin. Biol. Res. 369:407-416.

621 Siglin, J.C., C.M. Weghorst, D.E. Rodwell, and J.E. Klaunig. 1995. Gender-dependent differences in

622 hepatic tumor promotion in diethylnitrosamine initiated infant B6C3F1 mice by alpha-

623 hexachlorocyclohexane. J. Toxicol. Environ. Hlth. 44:235-245.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

624 Sumida, K., K. Saito, K. Oeda, Y. Yakabe, M. Otsuka, H. Matsumoto, M. Sekijima, K. Nakayama, Y.

625 Kawano, and T. Shirai. 2007. A comparative study of gene expression profiles in rat liver after

626 administration of alpha-hexachlorocyclohexane and lindane. J. Toxicol. Sci. 32(3):261-288.

627 Tanooka, H. 1977. Development and applications of Bacillus subtilis test systems for mutagens,

628 involving DNA-repair deficiency and suppressible auxotrophic mutations. Mut. Res. 42:19-32.

629 Thamavit, W., Y. Hiasa, N. Ito, and N. Bhamarapravati. 1974. The inhibitory effects of a-

630 benzenehexachloride on 3'-methyl-4-dimethylaminoazobenzene and DL-ethionine

631 carcinogenesis in rats. Cancer Res. 34:337-340.

632 Tryphonas, L., and F. Iverson. 1983. Sequential histopathologic analysis of alpha-

633 hexachlorocyclohexane-induced hepatic megalocytosis and adenoma formation in the HPB mouse. J.

634 Natl. Cancer Inst. 71:1307-1318.

635 Tsukada, H., M. Gotoh, Y. Mochizuki, and K. Furukawa. 1979. Changes in peroxisomes in

636 preneoplastic liver and hepatoma of mice induced by a-benzene hexachloride. Cancer Res. 39:1628637 1634.

638 USEPA. 1987. Health and environmental effects profile for hexachlorocyclohexanes. EPA/600/X-

639 88/248. U.S. Environmental Protection Agency, Environmental Criteria and Assessment Office, Office of

640 Health and Environmental Assessment, Office of Research and Development, Cincinnati, OH. December.

641 USEPA. 1988. Recommendations for and Documentation of Biological Values for Use in Risk

642 Assessment. EPA/600/6-87/008. U.S. Environmental Protection Agency, Environmental Criteria and

643 Assessment Office, Cincinnati, OH.

644 USEPA. 1995. The Use of the Benchmark Dose Approach in Health Risk Assessment. EPA/630/R-

645 94/007. U.S. Environmental Protection Agency, Risk Assessment Forum, Washington, DC. 93 pp.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

646 USEPA. 1996. Proposed guidelines for carcinogen risk assessment. EPA/600/P-92/003C. U.S.

647 Environmental Protection Agency, Office of Research and Development, Washington, DC. 142 pp.

648 USEPA. 2001. Toxicological review of chloroform. EPA/635/R-01/001. U.S. Environmental Protection

649 Agency. Washington, DC. 82 pp.

650 USEPA. 2005. Guidelines for carcinogen risk assessment. EPA/630/P-03/001F. U.S. Environmental

651 Protection Agency, Risk Assessment Forum, Washington, DC. 166 pp.

652 USEPA. 2006. Assessment of lindane and other hexachlorocyclohexane isomers. U.S. Environmental

653 Protection Agency, Office of Prevention, Pesticides, and Toxic Substances. February.

654 USEPA. 2010. Toxicological review of ethylene glycol monobutyl ether (EGBE). EPA/635/R-08/006F.

655 U.S. Environmental Protection Agency. Washington, DC. 123 pp.

656 USEPA. 2011. Recommended Use of Body Weight 3/4 as the Default Method in Derivation of the Oral

657 Reference Dose. Final. EPA/100/R11/0001. U.S. Environmental Protection Agency, Risk Assessment

658 Forum, Washington, DC. 50 pp.

659 USEPA. 2012. Benchmark Dose Technical Guidance. EPA/100/R-12/001. U.S. Environmental

660 Protection Agency, Risk Assessment Forum, Washington, DC. 99 pp.

661 USEPA. 2014. Memorandum. Human Health Evaluation Manual, Supplemental Guidance: Update of

662 Standard Default Exposure Factors. OSWER Directive 9200.1-120. U.S. Environmental Protection

663 Agency, Office of Solid Waste and Emergency Response, Washington, DC.

664 USEPA. 2015a. Summary of nominations for the fourth contaminant candidate list. EPA 815-R-15-001.

665 draft ccl list. U.S. Environmental Protection Agency.

REVISED - Carcinogenicity and MOA Evaluation for alpha-HCH: Implications for Human Health Risk

Assessment

December 2015

666 USEPA. 2015b. Fourth unregulated contaminant monitoring rule. Available

667 at: http://www.epa.gov/dwucmr/fourth-unregulated-contaminant-monitoring-rule. U.S. Environmental

668 Protection Agency. Office of Groundwater and Drinking Water, Cincinnati, OH.

669 USEPA. 2015c. Integrated Risk Information System. Available at: http://www.epa.gov/iris. U.S.

670 Environmental Protection Agency.

671 Venkat, J.A., S. Shami, K. Davis, M. Nayak, J.R. Plimmer, R. Pfeil, and P.P. Nair. 1995. Relative

672 genotoxic activities of pesticides evaluated by a modified S0S microplate assay. Environ. Mole.

673 Mutagenesis. 25:67-76.

674 Werle-Schneider, G., A. Wolfelschneider, M.C. von Brevern, J. Scheel, T. Storck, D. Muller, R.

675 Glockner, H. Bartsch, and M. Bartelmann. 2006. Gene expression profiles in rat liver slices exposed to

676 hepatocarcinogenic enzyme inducers, peroxisome proliferators, and 17a-ethinylestradiol. Int. J. Toxicol.

677 25:379-395.

678 Whysner, J., P.M. Ross, and G.M. Williams. 1996. Phenobarbital mechanistic data and risk assessment:

679 enzyme induction, enhanced cell proliferation, and tumor promotion. Pharmacol. Ther. 71:153-191.

680 Williams, G.M. and M.K. Iatropoulos. 2002. Alteration of liver cell function and proliferation:

681 differentiation between adaptation and toxicity. Toxicol Pathol. 30: 41-53.

682 Williams, G.M. 2008. Application of mode-of-action considerations in human cancer risk assessment.

683 Toxicol. Letters. 180:75-80.

Summary of Findings

Major Study Limitations

■erlyng et al. (1994) Rat (Wistar), male

Duration: up to 35 weeks ex and 30 week recovery Sample Size: 3-8/group Route: dietary, ad libitum Dose Levels: 0, 500 ppm Duration: 15 or 30 days Sample Size: 6-22/group Route: dietary, ad libitum Dose Levels: 0, 20 ppm

Duration: approximately 107 weeks

Sample Size: 10/sex/group, 20/sex/group controls

Route: dietary, ad libitum

Dose Levels: 0, 10, 50, 100, 800 ppm

Duration: 50 hours Sample Size: 2-19/group Route: oral gavage Dose Levels: 0, 150 mg/kg

•oto et ai. Mouse (ICR-JCL), male Duration: 26 weeks

1972) Sample Size: 10/group

Route: dietary (unknown if ad libitum) Dose Levels: 600 ppm

period (incidence

of 5/5 after 35 weeks). Treatment-related centrilobular hypertrophy regressed: incidence of 0/7 after the recovery period.

Incidence of foci of cellular alterations was 1/5 after 35 weeks and 1/7 after 35 weeks plus 30 week recovery. No nodules or HCC observed.

Increased total P450 levels at 15 days, further in

■e at 30 days (both significant).

Relative liver weight significantly increased in the 50, 100, and 800 ppm groups (dose-dependent).

Rats exposed to 800 ppm had decreased body weight gain and decreased sumval compared to controls, MTD exceeded.

A single dose of HCH did not alter hepatocyte ploidy. DNA labeling index (BrdU incorporation) was maximal ~30 hours after a single HCH dose and was increased in both mononuclear and binuclear hepatocytes. DNA labeling was significantly increased in diploid, tetraploid, and octaploid hepatocytes following daily 150 mg/kg oral doses of HCH, the proportion of binuclear cells decreased, suggesting aberrant proliferation.

Hepatoma (10/10) consisting of areas of atypical proliferation, nodules, and tu Relative liver weight was increased.

. Hepatoma incidence in control animals not reported. No fibrosis.

Small sample size. Only males evaluated. Only one dos evaluated. Mortality/general toxicity not reported. No sta evaluation.

Increased P450 reductase at 30 days. Increased TBARS formation in liver homogenates and microsomes after 15 and 30 days. Increased microsomal superoxide production at 15 days, further increase at 30 days. Increased SOD activity at 15 days, decreased at 30 days relative to 15 days (but higher than control). Increased glutathione reductase at 30 days, increased catalase at 15 and 30 days. All changes were statistically significant and were generally time-dependent.

No microscopic changes.

No gross tumors reported. Necrotic foci (<1 mm diameter) and other degenerative changes observed in highest dose (800 ppm) group.

Small sample size. Only one dose level evaluated. Only evaluated. Unclear mortality.

Small sample size. Minimal details on histopathology. H overall mortality in the study, evaluations were based eith moribund or found dead animals. Inadequate discussion mortality/general toxicity. Data were not stratified by sex.

Only males tested. Small sample size. Only one dose le evaluated. The number of repeated doses was not speci for the binucleation experiment.

Only one dose level evaluated. Small sample size. Only tested. No statistical analysis. Inadequate characterizat histopathological changes. Mortality not reported. Incidei benign and malignant tumors not reported. Inadequate translation from German did not allow for comprehensive

Hanada et al. (1973)

Duration: 32 weeks plus 5-6 weeks recovery Sample Size: 10-11/sex/group, 20-21/sex/group controls

Route: dietary, ad libitum

Dose Levels: 0, 100, 300, 600 ppm

Average tumor size increased with increasing dose in exposure plus recovery group. No microscopic peritoneal proliferation in liver (hypertrophic foci, associated with liver cell damage) noted in all treated exposure plus recovery mice except females Incidence of 8/8 100 ppm males, 7/7 300 ppm males, 3/3 300 ppm females, 7/7 600 ppm males, and 8/8 600 ppm females.

One 600 ppm female had mammary

at the week 26 laparotomy. Hepatoma observed after exposure plus Small sample size. No statistical analysis. Apparent incr

in mortality in treated animals that was not dose-depende General toxicity data were not reported. No evaluation d Atypical the end of the 32 week exposure period, regression of ch i ppm. could not be evaluated.

Duration: 24 weeks Sample Size: 20-40/group Route: dietary, ad libitum Dose Levels: 0, 100, 250, 500 ppm

Duration: 24 weeks

Sample Size: 26-30/group, 20/group controls

Route: dietary, ad libitum

Dose Levels: 0, 50, 100, 250 ppm

Dose-dependent in

e in HCC (0/20, 0/20, 10/38, 17/20)

Dose-dependent increase in liver nodular hyperplasia in treated mice (0/20, 30/38, and 20/20). No metastatic changes or tumors in other organs were noted upon gross examination. Dose-dependent increase in relative liver weight. Severe liver cell hypertrophy observed in 250 and 500 ppm groups, less severe at 100 ppm. Necrotic or fatty change rarely noted. Increased smooth endoplasmic reticulum in carcinomas and non-cancerous tissue. Body weight not affected.

Nodule incidence of 23/30 and carcinoma incidence of 8/30 in the 250 ppm group. No nodules or carcinoma in 50 or 100 ppm groups. Centrilobular hypertrophy observed in the 100 and 250 ppm groups (dose-dependent increase in severity). No cirrhosis or metastases. Relative liver weight was increased (dose-dependent). Body weight not affected.

Only males evaluated. No statistical evaluation. Only examined liver histologically. Mortality not reported.

No statistical evaluation. Only males evaluated. Unclear extra-hepatic tumors/metastases were evaluated microscopically. Mortality not reported.

Rat (Wistar), male Duration: 72 weeks, interim sacrifices Sample Size: 5-16/group Route: dietary, ad libitum Dose Levels: 0, 500, 1000, 1500 ppm

HCC observed only in 1000 and 1500 ppm groups at 72 weeks (incidence of 1/16 and 3/13, respectively).

Control animals sacrificed at different time than treated animals. Mortality not reported. Unclear if metastases w

Incidence of nodular hyperplasia as follows: 1000 ppm - 5/12 (48 weeks), 12/16 (72 weeks), 1500 ppm - 10/13 (72 weeks). None in control or 500 ppm evaluated grossly or microscopically. Insufficient descrip groups. general toxicity. Only males evaluated. Small sample si;

statistical evaluation.

Increased relative liver weight in all dose groups at all time points (dose-dependent). Hepatic hypertrophy observed, dose- and time-dependent

Mouse (DDY), male Duration: 72 weeks, interim sacrifices and recovery. Centrilobular hypertrophy observed after 16 weeks, regressed following

Sample Size: 12-20/group continuous exposure (25% after 16 weeks, 70% after 20 weeks, 100% after 24 weeks),

Route: dietary, ad libitum Increased relative liver weight over time, increases regressed following exposure

Dose Levels: 0, 500 ppm were not observed microscopically.

Incidence of liver tumors increased progressively with i tumors regressed following exposure cessation. Metastases to regional lymph nodes, lungs, or kidneys

to et al. (1983) Rat (Fisher 344), s

not reported

After 24 weeks, m

s were nodular hyperplasia. At 60-72 weeks, most tumors were HCC. it of smooth endoplasmic reticulum was observed on electron microscopy in the hyperplastic cells.

Duration: Initiation with DEN and partial hepatectomy No increase in number or area of hepatic hyperplastic nodules, in number of degenerated hyperplastic nodules, or in number of HCCs was observed followed by 6 weeks of alpha-HCH exposure. Some compared to control over the 50 week experimental period. Hyperplastic nodule number and area were significantly increased in DEN-initiated, partially rats sacrificed at the end of the 6 weeks, other groups hepatectomized rats who received 6 weeks of dietary alpha-HCH and were immediately sacrificed. were periodically sacrificed over a 50 week total duration.

Sample Size: 8-34/group for some endpoints, not

reported for other endpoints.

Route: dietary

Dose Levels: 0, 1000 ppm

Duration: 16 days, interim sacrifices Sample Size: 5-10/group Route: intraperitoneal or gavage Dose Levels: 0, 3, 10, 30, 50, 100, 200 mg/kg

Only one dose level evaluated. Only males evaluated. N statistical analysis. Apparent increase in mortality over ti and with longer exposure.

Lack of methodological details, including animal sex and size. No mortality or toxicity data.

GST activity was significantly increased 2, 4, and 6 days after a single intraperitoneal dose (multiple GST substrates), GST after a single oral dose, except for 6 days post-dose when an HCH metabolite was used as the substrate. GST increases w increases were dose-dependent and significant at doses at or above 30 mg/kg (multiple substrates). Relative liver weights were significantly increased 2, 6, and 10 days after a single oral dose and 4 and 6 days after a single intraperitoneal dose, increases were transient in the intraperitoneal group.

generally not increased Only males evaluated. Small sample size. Potentially GST activity irrelevant route of exposure (intraperitoneal).

GST activity and relative liver w in 3 day old rats.

ights w

significantly increased 6 days after a single ip dose of 200 mg/kg in animals 14, 21, and 42 days old but not

ee and Edwards 2001)

Rat (Wistar), male hepatocytes

Duration: 6 hours Prostaglandin E2 release w

Sample Size: 2 cultures Route: in vitro Dose Levels: 0, 30 microM Rat (Wistar) female Duration: NNM initiation, 8 week recovery, then 10 or Volume fraction and m 28 weeks alpha-HCH followed by 2, 6, or 21 week recovery

Sample Size: 3-7/group/time point Route: dietary

Dose Levels: 0, 20 mg/kg bw

ot increased in treated hepatocytes.

in number of hepatic foci increased over an 18 or 36 week exposure and decreased upon

Small sample size. Only one concentration evaluated. V little data presented on HCH, including DNA synthesis da Cell viability and treatment cytotoxicity were not reported.

No mortality or toxicity data. Only female ra

Masuda et al. (2001) Rat (F344), male

Duration: 6 weeks after initiatio hepatectomy Sample Size: 15/group Route: dietary, ad libitum Dose Levels: 0, 0.01, 0.1, 0.5, 60, 125, and 500 ppm

by DEN and partial GST-P-positive foci increased in dose-related m

n groups receiving 0.5 ppm or more. Numbers of GST-P positive foci significantly increased in Non-isoform-specific P450 substrate. The effect of HCH

2, 4, 7.5, 15, 30,

groups treated with 2 ppm and higher, with the exception of 4 ppm. Areas of GST-P positive foci significantly increased in groups treated with 7.5 ppm on foci formation was not evaluated. Potential confoundi and higher, with the exception of 15 ppm. effect of partial hepatectomy. Only males evaluated. Sm

sample size.

Dose-dependent, significant increases in CYP2B protein from 60 ppm were seen. Testosterone 16B-hydroxylation activity was significantly increased in a dose-related manner from 30 ppm. CYP3A protein significantly increased (dose-dependent) from 4 ppm and testosterone hydroxylation significantly increased (dose-dependent) from 15 ppm.

Relative liver weight significantly increased in the 7.5, 60, 125, and 500 ppm groups. Body weight significantly decreased in 15, 30, 60, 125, and 500

lagasaki et al. Mouse (DDY), male, Duration: 24 weeks

1975) Rat (Wistar), male, Sample Size: 20 rats, 16 hamsters, 36 mice, 48

Hamster (Golden mice in the second experiment Syrian), male Route: dietary, ad libitum Dose Level: 0, 500 ppm

Mouse (DDY), Mouse Duration: 24 weeks

(CH3/He), Mouse Sample Size: 13-29/sex/strain (DBA/2), Mouse (ICR), Route: dietary, ad libitum

Mouse (C57BL/6), Dose Level: 0, 500 ppm

Two experiments conducted (One for species comparison, and one for comparison of alpha-HCH +/- v.

s other compounds). Only one dose level evaluated. Only males evaluated. s

sample size. No statistical analysis, standard deviation fo

Nodular hyperplasia (20/20 mice and 13/19 mice in first and second experiments, respectively) and HCC observed (6/20 mice and 8/19 mice in first and body weights not reported. Only livers examined. Morta second experiments, respectively).

not reported. No evaluation of

mice. No cirrhosis. No tumors in ight gain in rats and hamsters. Co-t

Centrilobular hypertrophy seen in all three species, most pronounced ir weight in all three species, most pronounced in mice. Reduced body w inducers reduced the incidence of mouse liver tumors.

Strain comparison: Increased relative liver weight in treated males and females of multiple mou increase. Centrilobular hypertrophy and oval cells observed in males and females of multiple m in incidence of nodular hyperplasia (16.7-100%) and HCC (0-65%). In general, males were mor

. or hamsters. Increased relative liver : with 3-MC but not other enzyme

strains, strain differences in the degree of the Only livers examined histologically. Only one dose level

se strains. Strain- and gender-dependent differences evaluated. Mortality not reported. No statistical analysis susceptible than females.

Duration: 10 weeks following initiation with DEN

Sample Size: 12/group

Route: dietary (unknown if ad libitum)

Dose Levels: 0, 0.01, 0.05, 0.1, 1, 50, 500 ppm

Dose-dependent increase in number and area of GST-P positive foci (significant at high doses), foci number and area at 0.05 ppm were significantly decreased. The proportion of proliferating cells (i.e., PCNA positive) within GST-P positive foci decreased and then increased (dose-dependent, significant at highest dose). Foci were observed in all treated rats.

Total P450 content and P450 reductase activity were significantly decreased at 0.05 ppm but significantly increased at 500 ppm. P450 reductase protein level significantly increased at 50 and 500 ppm. Some significant increases in P450 activity at 50 and 500 ppm. Dose-dependent increases CYP2B, 2C, 2E, and 3A protein levels, increases were significant at 50 and 500 ppm.

Liver 8-OHdG levels significantly decreased at 0.1 and 1 ppm but significantly increased at 500 ppm. GST activity significantly increased at 500 ppm. Decreased body weight gain, significantly increased relative liver weight at 500 ppm. Adenomas and HCCs observed only at 500 ppm (mean of 2.8 tumors per rat).

Small sample size. Only males evaluated. Mortality not reported. Oxidative DNA damage repair data difficult to interpret due to high variation in OGG1 message. P450 a analyses were not isoform-specific, testosterone was the substrate used for all isoforms. The effect of HCH alone in without initiation, was not evaluated.

■chroter et al. (1987) Rat (Wistar), female Duration: 17 weeks (initiation), 15-20 weeks following Initiation Study: No in

n by NNM (promotion) Sample Size: 3-8/group (initiation study), 4/group (promotion study)

Route: gavage or dietary, ad libitum

Dose Level: single oral bolus dose of 200 mg/kg

(initiation) or 0-20 mg/kg-day in the feed (promotion)

e in GGT-positive foci in partially hepatectomized rats given a single oral bolus dose followed by 15 weeks of phenobarbital Small sample size. Only females evaluated. Not all data

statistically evaluated. Mortality not reported. Only liver evaluated. The effect of HCH alone, without initiation, ws evaluated in the promotion study.

Promotion Study: Dose- and time-dependent increases in foci number and area were obseived after 15 and 20 weeks. Foci area was significantly increased relative to control at mid- to high-doses. Dose-dependent increases in liver mass, liver DNA (both significant at highest dose tested), and P450 activities (not significant) were observed. P450 induction and liver weight increases were not predictive of foci formation. NOELs calculated.

Species, Sex

ngsubhakorn et al 981)

et al. (1991)

tzhugh et al. (1950) Rat (Wistar), male/female

Mouse (dd), male/female

1/5 300 ppm male, 1/4 300 ppm female, and 2/4 600 ppm males had liver recovery: males 0/14, 1/8, 7/7, 7/7 and females 0/15, 0/8, 2/3, 6/8.

o et al. (1973a)

o et al. (1973b)

o et al. (1975)

o et al. (1976)

Increased

et al. (1981)

uebeck et al 995)

of exposure

agasaki et al 975)

uatanachokchai et Rat (F344), male . (2006)

1. Alpha-HCH Carcinogenicity and Mode of Action for Hepatocarcinogenicity.

Summary of Findings

Major Study Limitations

■chulte-Hermann and Rat (Wistar), female ■arzefall (1981)

Duration: 6 days

Sample Size: 5-6/group

Route: oral gavage

Dose Levels: 0, 200 mg/kg

Duration: 24.5 months, interim sacrifices and

recovery.

Sample Size: 2-5/group

Route: oral gavage and/or dietary (unknown if ad libitum)

Dose Levels: 0 and initial oral dose of 100 mg/kg followed by 18.4 mg/kg-day in the diet, 420 mg/kg ii week intervals; or 200 mg/kg in 2 week intervals.

e increased following a single oral dose.

CYP1A, CYP2B, CYP2A, and CYP3A enzyme

Significantly decreased body weight after 4.5 and 23.5 months of continuous exposure, also decreased after 21.5 months of interval exposure (n=2). Significantly increased relative liver weight after 4.5, 13.5, or 23.5 months of dietary exposure and after 11.5 months of interval exposure, no clear temporal trend.

After 11.5 months of interval exposure followed by recovery period, relative liver weights and body weights were similar to control.

3 Significantly increased RNA and DNA in liver after continuous or interval regressed in the 11.5 month interval exposure plus recovery group.

Significantly increased cytochrome P450 activity following 4.5, 13.5, and 23.5 months of trend. The increases regressed in the 11.5 month interval exposure plus recovery group.

up to 23.5 months, no clear temporal trend. The

or 11.5 months interval

Small sample size. Only females tested. Inconsistent d> regimen. High incidence of microscopic foci in controls (3 1/3, and 5/6). Body weight decrease was severe (~20%) 21.5 month interval treatment and 23.5 month c t groups. Mortality not reported.

species, Sex

chulte-Hermann and Rat (Wistar), female arzefall (1980)

no clear temporal

Pronounced increase in GST activity after 11.5 months of interval treatment or 13.5 months of continuous treatment.

Time-dependent increase in incidence of macroscopic and microscopic liver lesions (foci, nodules/tumors, and HCC). Low incidence of HCC (1/6 and 1/8 continuous and interval-treated rats after at least 20 months). High incidence of liver nodules (5/6 and 6/8 continuous and interval-treated rats after at least 20 months, 2/4 and 1/4 continuous and interval-treated rats after at least 11.5 months, 3/8 interval-treated rats after 11 months).

■chulte-Hermann et Rat (Wistar), female l. (1981)

Duration: 8 weeks after DEN, interim sacrifice Sample Size: 5-6 rats/group Route: oral gavage

Dose Levels: 0, 150-200 mg/kg each +/- DEN

After 2 HCH doses at 5 and 8 weeks after DEN in

n, the proportion of GGT-positive foci of larger size increased significantly.

DNA synthesis in GGT positive and normal cells was significantly increased in HCH-treated animals relative to the same cell type in control animals. Among HCH-treated animals initiated with either DEN or NNM, DNA synthesis was significantly higher in GGT positive cells compared to normal cells.

For one experiment, DEN was given for 40 days followed 25 days later by a single 200 mg/kg oral bolus dose of HCH. DNA synthesis in GGT positive cells was greater than in normal hepatocytes. DNA labeling was not different in GGT positive islands of different sizes (40 days of DEN followed 25 days later by 200 mg/kg HCH). Mitotic index was increased in DEN initiated rats given a single 200 mg/kg dose of HCH 3 or 11 months later. No GGT-positive islands were found in rats treated with HCH alone.

Small sample size. Only females evaluated. Inconsisten dosing regimen. Potential confounding effect of initiator administration.

■chulte-Hermann et Rat (Wistar), female Duration: unclear (gavage study), 28 weeks (dietary DNA synthesis and m

Sample Size: not specified

Route: oral gavage or dietary, ad libitum

Dose Level: 0, 200 mg/kg (gavage), 0, 20 mg/kg

(dietary)

e increased in GGT-positive liver cells in initiated ra

Mean GGT-positive island size increased after 28 weeks of dietary HCH exposure in

> after a single oral dose of HCH. 1-initiated rats relative to NNM treatment alone.

Sample size not reported. Unclear duration. No statistic, analysis. Only females evaluated. The effect of HCH ale island formation was not evaluated.

Study Design

Summary of Findings

Major Study Limitations

Duration: 14 days or 28 weeks (separate experiments)

Sample Size: 15/sex/group

Route: dietary, ad libitum

Dose Levels: 0, 250 ppm each +/- DEN init

increased after 24 weeks of HCH. Adenoma number overall was higher in male m dietary exposure (significant at 7 and 14 days) in foci and surrounding tissue in n< labeling was significantly decreased after 14 days in DEN-initiated males.

. Progressively increasing DNA labeling w

ir 14 days of labeling data were not differentiated according to cell type

i-initiated males and females and in DEN-initiated females. DNA foci vs. normal hepatocytes).

Duration: 14 days or 28 weeks (separate experiments)

Significantly increased relative liver weight in males and females initiated with DEN and exposed to HCH for 24 weeks, no in

Only liver examined. Only one dose level evaluated. Sm

Significantly decreased body weight in HCH-only and DEN+HCH males. Foci incidence of 4/15 males and 1/15 females receiving HCH only sample size. Only liver examined. Differential susceptibil

Sample Size: 15/sex (promotion study), 10/sex (DNA for 24 weeks (no initiation). Foci incidence 100% in males and females treated with DEN only. Hepatocellular adenoma incidence of 4/15 HCH-only

synthesis study) Route: dietary, ad libitum Dose Levels: 0, 250 ppm each +/- DEN initiation

males, 15/15 DEN-only and DEN+HCH males, 0/15 HCH-only females, 1/15 DEN-only females, and 11/15 DEN+HCH females (significant increase) after 28 weeks. DNA labeling after 14 days was significantly increased in normal liver from DEN+HCH males and females compared to DEN only, but not in foci. DNA labeling in general was higher in foci.

infant male mice to DEN-mediated adenoma formation. Questionable relevance of initiation due to high incidence foci in DEN-only animals and high incidence of adenoma DEN-only males.

Species, Sex

glin et al. (1991) Mouse (B6C3F1),

male/female

No foci or adenomas observed in HCH-only or control mice after 28 weeks. Very high number of adenomas in DEN-initiated male mice (no HCH), number significantly decreased after 24 weeks of HCH exposure. In females, adenoma number in DEN-only mice was very low and significantly

Only one dose level evaluated. Small sample size. Only examined. No body weight or mortality data reported. DN

Questionable relevance of initiation due to differentially hi adenoma incidence in DEN-only males.

glin et al. (1995) Mouse (B6C3F1),

male/female

■umida et al. (2007) Rat (F344), male

Duration: 28 days, interim sacrifices Sample Size: 4/group Route: oral gavage Dose Levels: 0, 2, 20 mg/kg-day

Hepatocellular hypertrophy seen in high-dose animals (0/4, 0/4, 4/4) .

n the 2 mg/kg-day group w

Progressive and significant (except day 3) time-dependent decrease in ALP in 20 mg/kg-day animals beginning 1 day post-dose.

Increased GST and P450 isoform expression was seen at the 28 day time point. Some increases also noted at 1 and 3 days, but there was no clear

hamavit et al. (1974) Rat (Fisher), male

Tryphonas and verson (1983)

Duration: 6 months, 2 month interim sacrifice and 5 month exposure plus one month recovery Sample Size: 3-6/group Route: dietary, ad libitum Dose Level: 0, 0.06% (600 ppm)

Mouse (HPB), male Duration: 50 weeks, interim sacrifices

Sample Size: 75 -treatment group, 48 -control group, 4-9/group interim sacrifices Route: dietary, ad libitum Dose Levels: 0, 500 ppm

No abnormal histology, nodules, or

at 2 months or 6 months. Small sample size. Only males evaluated. Only one dos

evaluated. No statistical evaluation. Only evaluated the months, very slight increase at 6 months. Decreased body weight gain after 6 months, no change after 2 Mortality not reported. Animals not sacrificed at 5 month'

assess the effect of the 1 month recovery period.

sukada et al. (1979) Mouse (DD), male

Duration: 36 weeks, interim sacrifices Sample Size: 6/group Route: dietary, ad libitum Dose Levels: 0, 500 ppm

Gross (29% incidence) and microscopic (57% incidence) nodules of the liver observed beginning at 21 weeks of exposure, incidence was 100% at 33 weeks. All nodules were benign (adenoma).

No gross evidence of metastases in the lungs.

Increased relative liver weight (time-dependent) and megalocytosis observed in exposed mice. Increased mitotic index in megalocytic and nodular hepatocytes. Single cell necrosis, lipid accumulation, and nodules arising from areas of megalocytic cells were observed microscopically. Reduced body weight gain in treated mice after 50 weeks.

Only one dose level evaluated. Small sample size. Only tested. No statistical analysis. Only examined the liver a lungs. Emaciation noted in mice with severe liver enlarge or large tumors. Initial body weights not reported.

Proliferation of smooth endoplasmic reticulum was noted, peroxisome proliferation was not observed at 16-20 weeks.

Only one dose level evaluated. Small sample size. Only evaluated. Mortality and body weights not reported. Background tumor incidence could not be evaluated in co mice (due to short duration of inclusion in the study).

Only males evaluated. Small dose groups. Results not confirmed with PCR. Inconsistency in most changes ov

Significantly increased relative liver weight after 3 days at 20 mg/kg-day, no response evident but no clear temporal trend.

Centrilobular hypertrophy obseived beginning at 16 weeks. Periportal atrophy obseived. Hyperplastic nodules: 1/6 mice at 16 weeks, 5/6 mice at 20 weeks. 2/6 mice had hepatomas at 28 weeks, 3/6 mice had hepatomas at 32-36 weeks.

^erle-Schneider et Rat (Wistar), male Duration: 24 hours Some dose-dependent changes in expression are suggested (e.g., ubiquitin, P-glycoprotein, GAPDH, retinoblastoma, p53, UGT isoforms, ERK, GST Results not confirmed by PCR. I

l. (2006) liver slices Sample Size: 8 liver slices from 4 rats isoforms, P450 isoforms). over time.

Route: in vitro

Dose Levels: 0, 0.1, 1, 10, 100 microM

stency of

-OHdG = 8-hiyaray deoxyguanosine

LP = alkaline phosphatase

rdU = Oromo-deoxyuridine

YP = cytochrome P450

EN = diethynitrosamine

na = deoxyribonucleic acid

RK = extracellular signal-regulated kinase

APDH = glyceraldehyae phosphate dehydrogenase

■GT = gamma-glutamyl transpeptidase

■ST = glutathione-S-transferase

■ST-P = glutathione-S-transferase, pi isoform

CC = hepatocellular carcinoma

CH = hexachlorocyclohexane (alpha isomer)

lOA = mode of action

ITD = maximum tolerated dose

NM = N-nitrosomorpholine

OEL = no-oOseived-effect level

450 = cytochrome P450

CNA = proliferating cell nuclear antigen

CR = polymerase chain reaction

OD = superoxide dismutase

GT = UDP-glucuronosyl transferase

Table 2. Evaluation of the Hypothesized MOA using EPA's Causation Criteria

MANUSCRIPT

Criteria_Evaluation Findings

- P450-related parameters such as isozyme protein and activity, total P450, and P450 reductase were consistently increased in inc experimental design. Statistically significant increases in total P450 and P450 reductase activity and/or protein were seen in 2/2 st Statistically significant increases in P450 isozyme activity were seen in 3/4 studies which examined this effect; in the fourth study, i data were not statistically evaluated. This trend is supported by two microarray studies which found increased P450 isozyme exprf were not conclusive. The small sample size in some studies, and the lack of data in multiple species (effects have not been evalu evaluating the role of P450 induction.

- Six independent studies evaluated parameters related to oxidative stress; however, most of the parameters were not measured s peroxidation, oxidative DNA damage, superoxide anion production, and antioxidant enzyme activities (catalase, SOD, and GRed) w Significant increases in these four endpoints were observed. In addition, increased GST activity was seen in 2/2 studies which exa statistically significant in one study and was not statistically evaluated in the second study. Potential increases in GST expression studies; however, the microarray data were not conclusive.

- Increased cell growth and proliferation have been consistently observed in numerous independent studies of varying experimenta relative liver weight, hypertrophy, foci formation, DNA labeling) were increased in almost every study in which these endpoints weri high incidence (e.g., hypertrophy) or were statistically significant (e.g., relative liver weight) in most cases. Only two studies did not of cell proliferation; the lack of an effect may be due to short study duration or species/strain differences in response.

Strength, Specificity, and Consistency of the Association

- Consistent dose-dependent increases in markers of P450 parameters, oxidative stress markers, and cell proliferation were obser dose levels/concentrations were evaluated, with the number of studies different for each endpoint and therefore providing different Dose-Response Concordance consistent observation across endpoints was that of no increase at low doses, suggesting a threshold, and, in some cases, hormes

observed in one study where the incidence of response was 100% at all doses, which illustrates the potency of the doses used. Tl effects, particularly cell proliferation, strengthens the causal association between alpha-HCH exposure and liver tumor formation.

- Temporal relationships between exposure and P450 parameters, oxidative stress parameters, and increased cell proliferation we Specifically, two types of temporal patterns were observed: increased response over time and regression of response Temporal Association following cessation of exposure. However, the number of studies that looked at a temporal response for some endpoints was sma

stress endpoints). None of the studies evaluated was inconsistent with a temporal pattern. Regression of increases in P450 activity and hypertrophy were seen in several studies following cessation of exposure, which further supports the role of alpha-HCH as a ti

- The reported MOA is consistent with what is known about carcinogenesis. The roles of P450 induction, oxidative stress, and incr

. . ...... . carcinogenesis are well established. Other chemicals including phenobarbital have been shown to elicit similar effects, most notai

Biological Plausibility and Coherence , ,, » ■ ■ , . ■ . ... .......

increased cell proliferation. The role of oxidative stress in carcinogenesis for this compound is suggested, although the link is not i

is internally consistent in supporting the reported MOA.

ACCEPTED MANUSCRIPT

Table 3. Summary of Mutagenic Assays for Alpha-HCH.

Test System

Reference In Vitro/ In Vivo Species/Strain/ Cell Type Assay/Test Treatment Result Comments

Mutation Moriya et al. (1983) In vitro Salmonella typhimurium TA98, TA100, TA1535, TA1537, TA1538 Ames assay up to 5,000 pg/plate w/ and w/out activation Negative

Escherichia coli WP2 Reverse mutation assay up to 5,000 pg/plate w/ and w/out activation Negative

Shahin and von Borstel (1977) In vitro Saccharomyces cerevisiae XV185-14C Reverse mutation assay 0.1-200 ng/ml w/ and w/out activation Negative

Tanooka (1977) In vitro Bacillus subtilis TKJ5211 Spot test 5,000 mg/plate Negative

DNA Binding Iverson et al. (1984) In vitro Calf (thymus DNA) 1 pm Weakly positive Low levels of DNA binding only consistent with a non-genotoxic mechanism for neoplastic response.

In vivo Mouse liver 25 mg/kg Weakly positive Low levels of DNA binding only consistent with a non-genotoxic mechanism for neoplastic response.

Sagelsdorff et al. (1983) In vivo NMRI mouse liver HPLC analysis of nucleosides 6.2-8.5 mg/kg Weakly positive "Minute DNA binding". Binding is more than three orders of magnitude lower than would be expected if the mechanism of tumor induction was genotoxicity mediated by DNA binding.

DNA Damage, Fragmentation, and Repair Kalantzi et al. In vitro Human MCF-7 breast (2004) carcinoma cells Comet assay 10-4 M Positive At lower concentrations no comet-forming effects were observed.

Human PC-3 prostate carcinoma cells Comet assay 10-4 M Positive At lower concentrations no comet-forming effects were observed.

Mattioli et al. (1996) In vitro Human hepatocytes Comet assay 0.056-0.32 mM Positive

Rat hepatocytes Comet assay 0.056-0.32 mM Positive Modest, dose-dependent increase in DNA breaks Co-treatment with metyrapone ( inhibitor of CYP450) resulted in a reduction in the frequency of DNA breaks, and is suggestive of several potential mechanisms: 1) alpha-HCH may be transformed into a reactive species by CYP450 dependent reaction; 2) alpha-HCH may interact with CYP450s to generate ROS that cause damage.

Mouse hepatocytes Comet assay 0.056-0.32 mM Negative

Venkat et al. (1995) In vitro Escherichia coli PQ37 SOS microplate assay NA See comment Results provide a relative scale of activity. Alpha-HCH had levels of activity that ranged 1/10 to 1/4 (dependent on dosing vehicle) that of 4-NQO, which is considered to be a direct acting mutagen.

Hitachi et al. (1975) In vivo Liver cells from Donryu rats Chromosomal alterations -Inspection for cell distribution - % by ploidy 600 ppm Positive

Notes: 4-NQO

= 4-Nitro-quinoline oxide = cytochrome P450 DNA = deoxyribonucleic acid

HCH = hexachlorocyclohexane

HPLC = high performance liquid chromatography

NA = not available, dose not specified or unclear

ROS = reactive oxygen species

-- = specific test name not provided. Only endpoint is provided.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

Table 4. BMD Modeling Results for Critical Effects

BMDhed BMDLhed Scaled

Study Endpoi nt/C ritical Effect BMRF Model (mg/kg-day) (mg/kg-day) AIC p-value a residual b Visual fit c

Schroter et al. 1987 Foci number 1 SD Polynomial (third order) 0.15 0.072 82 0.28 0.12 Good

Power 0.39 0.26 83 0.13 1.3 Good

Schroter et al. 1987 Foci area 1 SD Exponential - Model 4 0.16 0.11 -77 0.15 0.09 Good

Polynomial (second order) 0.2 0.1 -77 0.17 0.13 Good

Polynomial (third order) 0.096 0.049 -79 0.47 -0.31 Good

Power 0.17 0.11 -77 0.15 0.11 Good

Schroter et al. 1987 DNA content 1 SD Exponential - Model 2 2.5 1.6 23 0.8 -0.57 Good

Exponential - Model 3 3.6 1.7 24 0.82 0.0012 Good

Exponential - Model 4 2.4 1.5 25 0.46 -0.63 Good

Polynomial (second order) 3.5 1.1 24 0.82 0.0068 Good

Power 3.6 1.5 24 0.82 0.0012 Good

Masuda et al. 2001 Relative liver weight 10% relative change Exponential - Model 2 2.4 2.2 -130 0.26 -0.82 Good

Polynomial (second order) 2.3 1.8 -130 0.14 -0.85 Good

Power 2.4 1.7 -130 0.12 -0.77 Good

Sumida et al. 2007 Relative liver weight 10% relative change Exponential - Model 2 1.7 1.5 -33 0.12 1.1 Good

Day 28 Power 1.6 1.4 -34 0.15 1.1 Good

Sumida et al. 2007 Absolute liver weight 10% relative change Exponential - Model 2 2.3 1.7 -2.1 0.66 -0.023 Good

Day 28 Power 2.2 0.22 -2.1 0.68 -0.025 Good

Puatanachokchai et al. 2006 Relative liver weight 10% relative change Polynomial (third order) 0.72 0.71 -140 0.45 0.082 Good

Puatanachokchai et al. 2006 Absolute liver weight 10% relative change Polynomial (third order) 0.72 0.71 67 0.71 0.000038 Good

Notes:

AIC = Akaike information criterion

BMDhed = human equivalent benchmark dose

BMDLhed = human equivalent benchmark dose lower bound

BMRF = benchmark response factor

a Goodness-of-fit test, p -value > 0.1 is deemed a good fit to the data and appropriate to model.

b Scaled residual near the range of the BMD. Scaled residuals need to be within range for model consideration (i.e., absolute value < 2). c Visual fit at low doses. Rating is classified as poor, moderate or good. Good fit indicates scaled residuals at low doses are within an absolute value of 2.

The selected model for each study and endpoint is shown in bold text.

Table 5. Summary of NOAELs and LOAELs for Critical Effects

Fitzhugh et al. (1950)

Endpoint/Critical Effect Slight microscopic changes and increased liver weight

Masuda et al. (2001) Increased number of GST-P

positive foci

Increased number of GST-P positive foci

Increased relative liver weight and increased liver

Puatanachokchai et al. (2006)

Sumida et al. (2007)

Schroter et al. (1987)

Increased area of foci

LOAELhed (mg/kg-day) Males - 1.0

Females - 1.0 0.21 a

NOAELhed (mg/kg-day) Males - 0.20

Females - 0.21 0.11

a Masuda et al. reports an increase in the number of GST-P positive foci at 0.2 mg/kg-day; however, the effect dropped off at the next highest dose tested and only began to display a positive dose-response relationship at 0.75 mg/kg-day equivalent to a LOAELhed of 0.21 mg/kg-day.

; 1 2 ; i i

Figure 1

Figure 2

Figure 1. BMDS Polynomial (3rd Order) Model Results for Foci Area. Model results for foci area based on Schroter et al. 1987 using a BMR of 1 standard deviation for the BMD and 95% lower confidence limit for the BMDL. BMD = benchmark dose, BMDL = benchmark dose lower bound, BMDS = Benchmark Dose Software, BMR = benchmark response.

Figure 2. BMDS Polynomial (3rd Order) Model Results for Foci Number. Model results for foci number are based on Schroter et al. 1987 using a BMR of 1 standard deviation for the BMD and 95% lower confidence limit for the BMDL. BMD = benchmark dose, BMDL = benchmark dose lower bound, BMDS = Benchmark Dose Software, BMR = benchmark response.

Highlights

• The WOE supports a nonlinear, threshold dose-response for alpha-HCH carcinogenicity.

• Alpha-HCH causes liver tumors in rats and mice through an MOA involving increased promotion of cell growth, or mitogenesis.

• We establish a nonlinear RfD for alpha-HCH of 0.002 mg/kg-day that is protective of chronic exposure to humans.

• Given the regulatory focus on alpha-HCH, agencies should review the evidence that supports treating alpha-HCH as a threshold based toxicant.