Trichloroethylene: Mechanistic, Epidemiologic and Other Supporting Evidence of Carcinogenic Hazard

Abstract

The chlorinated solvent trichloroethylene (TCE) is a ubiquitous environmental pollutant. The carcinogenic hazard of TCE was the subject of a 2012 evaluation by a Working Group of the International Agency for Research on Cancer (IARC). Information on exposures, relevant data from epidemiologic studies, bioassays in experimental animals, and toxicity and mechanism of action studies was used to conclude that TCE is carcinogenic to humans (Group 1). This article summarizes the key evidence forming the scientific bases for the IARC classification. Exposure to TCE from environmental sources (including from hazardous waste sites and contaminated water) is common throughout the world. While workplace use of TCE has been declining, occupational exposures remain of concern, especially in developing countries. Strongest human evidence is from studies of occupational TCE exposure and kidney cancer. Positive, although less consistent, associations were reported for liver cancer and non-Hodgkin's lymphoma. TCE is carcinogenic at multiple sites in multiple species and strains of experimental animals. The mechanistic evidence includes extensive data on the toxicokinetics and genotoxicity of TCE and its metabolites. Together, available evidence provided a cohesive database supporting the human cancer hazard of TCE, particularly in the kidney. For other target sites of carcinogenicity, mechanistic and other data were found to be more limited. Important sources of susceptibility to TCE toxicity and carcinogenicity were also reviewed by the Working Group. In all, consideration of the multiple evidence streams presented herein informed the IARC conclusions regarding the carcinogenicity of TCE.

1. Trichloroethylene is a ubiquitous environmental and occupational pollutant

Trichloroethylene (TCE) is a common environmental contaminant of hazardous waste sites, as well as water supplies. A volatile organic compound, TCE predominantly (97.7%) partitions to air from emitted sources (Boutonnet, et al., 1998). Both ambient and indoor air have been identified as important exposure routes in the general population (Wu & Schaum, 2000). TCE can be also released into the soil, surface and ground water supplies through industrial discharges and leaching from landfills or underground storage tanks. TCE occurs frequently at low levels in water supplies and in groundwater, owing to its widespread use and persistence in the environment (ATSDR, 1997). The 1999-2000 NHANES survey reported a mean blood level of TCE of 0.013 ng/ml in 290 subjects, with 88% below the detection limit (Jia, et al., 2012). In exhaled air samples from 361 subjects, mean TCE was 3.48 μg/m³ and 76% were below the detection limit (< 0.01 ng/ml). There was a moderate association between air and blood concentrations in this study. In the most recent 2005-2006 NHANES survey, all 3178 subjects had levels below the detection limits (Jia et al., 2012). The measured levels in the 1999-2000 NHANES survey indicated that around 10% of the population at large had a low exposure to TCE. The absence of measurable concentrations in the 2005-2006 survey clearly indicates that exposure of the general population had further decreased to a level no longer detectable; however, TCE remains to be an environmental contaminant of high concern at Superfund and other hazardous waste sites.

Commercial-scale production of TCE began in the 1920s in Germany and the United States, reaching hundreds of thousands of tons per year between 1950s and 1970s (NICNAS, 2000). Demand for very high production volumes of TCE was generated mainly by vapor degreasing uses, and by the growth of the dry cleaning industry. Although perchloroethylene replaced TCE as the predominant dry cleaning solvent in the 1950s, use as a spotting agent lasted longer and TCE continues to be used as a solvent (primarily in degreasing processes) and as a feedstock material in chemical manufacturing. By 1990, metal cleaning uses accounted for between 80 and 95% of the trichloroethylene produced in the United States, Canada, Japan and Western Europe (CCME, 1999; Mertens, 2007).

Concerns about environmental and human health hazards have led to a considerable (estimated as an order of magnitude) decline in both production volumes and commercial uses of trichloroethylene worldwide in the past 20 years (ECSA, 2012). Currently, the major regions generating demand for this chemical are the United States, Western Europe, China, and Japan (27%, 24%, 18%, and 13%, respectively) (Glauser & Ishikawa, 2008). In developed countries, the number of workers in occupations with likely trichloroethylene exposure (e.g., producing metal products, electrical and other machinery, appliances, transport equipment, etc.) has declined (Mirabelli & Kauppinen, 2005; Nagasawa, et al., 2011; von Grote, et al., 2003). However, data are lacking to support a similar trend in the developing world.

2. Evidence for cancer hazard of trichloroethylene

2.1. Experimental animal studies

The International Agency for Research on Cancer (IARC) Monographs Working Group concluded that there is sufficient evidence in experimental animals for the carcinogenicity of TCE (Guha, et al., 2012). Notably, in rats, modest increases in the incidence of kidney tumours were observed in both sexes of multiple strains exposed to TCE by either inhalation, or corn oil gavage (Maltoni, et al., 1988; NTP, 1988; NTP, 1990); however, statistical significance was only achieved in male F344/N rats (NTP, 1988). Nonetheless, the rarity of renal tumor occurrence in unexposed rats (no kidney tumors observed in unexposed rats in all of these studies) supports the biological significance of the findings. Other tumor findings in rats exposed to TCE include leukemia in male Sprague-Dawley and female August rats, and testicular tumors of the Leydig cell type with inhalation exposure in the Sprague-Dawley strain (Maltoni, et al., 1988; NTP, 1988). TCE by gavage also increased testicular interstitial cell tumors in Marshall rats, but was without effect on testicular tumorigenesis in other rat strains with much higher or lower background tumor rates [>75% in ACI, August, and F344/N, and 3% in Osborne-Mendel strains] (NTP, 1988). The inter-strain variability in background incidence of this type of tumor suggests that susceptibility factors may exist.

The primary tumor finding with TCE exposure in mice is statistically significant increases in liver tumors, reported in multiple inhalation and gavage bioassays of male Swiss and male and female B6C3F1 mouse strains (Anna, et al., 1994; Bull, et al., 2002; Herren-Freund, et al., 1987; Maltoni, et al., 1988; NCI, 1976; NTP, 1990). TCE decreased mouse liver tumor latency in male B6C3F1 mice. Additional tumor findings in mice include increased malignant lymphomas observed in female B6C3F1 mice exposed to TCE via corn oil gavage (NTP, 1990). Alveolar/bronchiolar adenomas were also induced by TCE inhalation exposure in mice (Fukuda, et al., 1983; Maltoni, et al., 1988), but not rats or hamsters (Fukuda, et al., 1983; Henschler, et al., 1980; Minowa, et al., 2000).

The Preamble to the IARC Monographs (http://monographs.iarc.fr/ENG/Preamble/) states that sufficient evidence of carcinogenicity in experimental animals is specified when “a causal relationship has been established between the agent and an increased incidence of malignant neoplasms or of an appropriate combination of benign and malignant neoplasms in (a) two or more species of animals or (b) two or more independent studies in one species carried out at different times or in different laboratories or under different protocols.” In addition, “an increased incidence of tumours in both sexes of a single species in a well-conducted study, ideally conducted under Good Laboratory Practices” also would classify an agent in this category. Based on these criteria, positive findings of liver neoplasms in two or more independent studies in mice support the conclusion of sufficiency for the carcinogenicity of TCE in experimental animals. Other findings corroborate this conclusion, but are not necessary to meet the IARC criteria.

2.2. Epidemiological studies

TCE was evaluated for the first time by IARC in 1995 (IARC, 1995). The overall epidemiological evidence was limited and based on a relatively small number of studies. Several new cohorts and case-control studies on exposure to TCE and risk of cancer from the US and Europe have been published since that time, including reviews and meta-analyses. In particular, the focus has been on cancers of the kidney, liver, lung, cervix and non-Hodgkin's lymphoma, multiple myeloma and leukemia.

The Preamble to the IARC Monographs (http://monographs.iarc.fr/ENG/Preamble/) states that sufficient evidence of carcinogenicity in humans is ascribed when “a positive relationship has been observed between the exposure and cancer in studies in which chance, bias and confounding could be ruled out with reasonable confidence.” The 2012 IARC Monographs Working Group concluded that based on the epidemiologic data there is sufficient evidence for the association between exposure to TCE and human kidney cancer. Positive associations between occupational TCE exposure and non-Hodgkin's lymphoma and cancer of the liver were also been observed, but the overall epidemiological evidence was less consistent, and thus characterized as limited for these two cancers (Guha, et al., 2012). Finally, although statistically significant elevated risks for cancers of the lung, cervix and oesophagus were observed in a few isolated studies, the evidence of carcinogenicity for these sites was inadequate.

2.2.1. Kidney Cancer

The strongest and most consistent evidence regarding the carcinogenicity of TCE in humans is from occupational studies of kidney cancer. In the previous evaluation by IARC (IARC, 1995), information regarding the carcinogenicity of kidney cancer came from only five relative small cohort studies. Three were from Europe and based on incidence (Anttila, et al., 1995; Axelson, et al., 1994; Henschler, et al., 1995), and two were from the US and based on mortality (Garabrant, et al., 1988; Spirtas, et al., 1991). Since that time, the follow-up of the cohort by Spirtas et al. (Spirtas, et al., 1991) has been extended (Blair, et al., 1998; Radican, et al., 2008). Also, new US cohorts, almost all with mortality as outcome, of aircraft and aerospace workers potentially exposed to TCE, have become available (Boice, et al., 2006; Boice, et al., 1999; Lipworth, et al., 2011; Morgan, et al., 1998; Zhao, et al., 2005). Cohorts of other types of TCE exposed workers followed-up for incidence (Hansen, et al., 2001; Henschler, et al., 1995; Raaschou-Nielsen, et al., 2003) and mortality (Bahr, et al., 2011) have recently been published. A majority of these cohort studies reported modestly elevated relative risks for kidney cancer, with an indication of exposure-response in one study (Raaschou-Nielsen, et al., 2003).

It should be noted that none of these cohorts have been adjusted for the effect of tobacco smoking, an established cause of kidney cancer. However, if tobacco smoking was a serious confounder in the cohort studies, an increased risk of lung cancer and coronary risk should also be expected, which was in general not the case. The group of three relative small similar independent cohort studies comprises biologically monitored workers from Finland, Denmark and Sweden shows little evidence of increased risk of kidney cancer, based on in total 16 cases of kidney cancer (Anttila, et al., 1995; Axelson, et al., 1994; Hansen, et al., 2001). These studies, however, may have had limited ability to detect a modest increase in risk. Overall, two recent meta-analyses of occupational cohort studies of TCE exposure, including the three Nordic studies of bio-monitored workers, reported relative risks of 1.26 (95% CI 1.02-1.56) (Karami, et al., 2012) and 1.16 (95% CI: 0.96-1.40) (Scott & Jinot, 2011), respectively.

Several case-control studies of kidney cancer in different countries have included information on occupational exposure to TCE (Bruning, et al., 1998; Charbotel, et al., 2006; Christensen, et al., 2013; Dosemeci, et al., 1999; Greenland, et al., 1994; Moore, et al., 2010; Pesch, et al., 2000; Vamvakas, et al., 1998). In contrast to the cohort studies, the case-control studies did adjust for tobacco use and other confounders. Adjustments had little influence on the risk estimates, suggesting no major confounding from either tobacco smoking or from occupational exposures.

The risk estimates from the case-control studies were, in general, stronger than those from the cohort studies. Two recent studies provided detailed exposure assessments, one in France (Charbotel, et al., 2006) and one in Eastern Europe (Moore, et al., 2010). Both studies demonstrated exposure-response relationship. The French study was conducted in an area with high prevalence of occupational exposure to TCE. An odds ratio of 1.64 (95% CI 0.95-2.84) was reported for the ever vs. never occupational TCE exposure, adjusted for tobacco smoking and body mass index. An odds ratio of 2.73 (95% CI 1.06-7.07) was reported for workers in the category with highest estimated TCE exposure. When, the data were adjusted for exposure to cutting fluids and to other petroleum oils the odds ratio was 2.63 (95% CI 0.79-8.83), suggesting no major confounding. The adjusted odds ratios in the East European study of any exposure to TCE were 1.6 (95% CI 1.04-2.54) and in the highest category of exposure intensity it was 2.3 (95% CI 1.05-2.51).

Because the risk estimates for TCE exposure and kidney cancer were modest and because most studies were small and had limited statistical power, the working group also weighted two recent, meta-analyses based on virtually all existing studies of kidney cancer. Overall, they estimated statistically significant meta-relative risks of 1.3 to 1.4 for kidney cancer and TCE (Karami, et al., 2012; Scott & Jinot, 2011). One meta-analysis that not only reported overall meta-RR, but also reported a meta-RR for highest exposure to TCE of RR of 1.6 (95% CI 1.3-2.0), thus indicating some dose-response relationship (Scott & Jinot, 2011). Overall, the relative risk of kidney cancer is only modestly increased but at the approximate same level as many, but not all, other occupational chemicals and exposure that IARC classifies as carcinogenic to humans (Group 1) (Rushton et al. 2012). The epidemiological evidence for the association with kidney cancer is, however, relatively robust because no single study appeared overly influential, the meta-RR estimates were not highly sensitive to alternate RR estimate selections, and there was no major heterogeneity across the studies.

2.2.2. Non-Hodgkin's Lymphoma

Information on occupational TCE exposure and non-Hodgkin's lymphoma was provided from in total of 16 studies (8 cohorts and 8 case-control studies). The overall epidemiologic evidence was less strong and consistent than for kidney cancer. The majority of studies have been published after the 1995 IARC evaluation. A large cohort study of TCE exposed workers from Denmark (Raaschou-Nielsen, et al., 2003), as well as most cohort studies of aircraft and aerospace workers in the United States, reported modestly elevated relative risks for non-Hodgkin's lymphoma (Boice, et al., 1999; Lipworth, et al., 2011; Radican, et al., 2008). The three cohort studies of biologically-monitored workers from Nordic countries show evidence of increased risk for non-Hodgkin's lymphoma in the range from 1.5 to 3.1, based on a total of 21 cases (Anttila, et al., 1995; Axelson, et al., 1994; Hansen, et al., 2001). Several case-control studies showed modestly increased odds ratios, but were limited in interpretation due to the use of different classification systems for NHL.

A recent meta-analysis of existing studies of non-Hodgkin's lymphoma (Scott & Jinot, 2011) reported meta-relative risks of 1.2 (95% CI 1.1-1.4) for any exposure to TCE and 1.4 (95% CI 1.1-1.8) for higher exposure. There was heterogeneity between studies, and also some indication of publication bias.

2.2.3. Liver Cancer

Nine cohort studies have examined the relationship between occupational TCE exposure and risk of liver cancer. A majority of the cohorts reported only results for the combination of cancers of the liver and gall bladder or biliary passages. Although positive associations were observed in some studies, the results were somewhat inconsistent, there was no overall indication of an exposure-response relationship, and none of the studies provided information on potential confounders, such as alcohol drinking. The only case-control study available had only one exposed case. A recent meta-analysis reported a meta-RR of 1.3 (95% CI 1.1-1.6) for the overall TCE exposure based on the nine cohorts. This value was similar when results from eight studies that provided information on high exposure were analysed (1.3, 95% CI 0.9-1.8) (Scott & Jinot, 2011).

2.2.4. Other potential human cancer target sites

Statistically significant excess risks of cancers of the lung, cervix and esophagus and multiple myeloma and leukemia were observed in isolated studies, but due to relatively few observations for each site and some inconsistency in reported results, the database was considered to be inadequate with respect to evidence of human carcinogenicity (Guha, et al., 2012).

3. Metabolism and Genotoxicity of TCE

3.1. Metabolic pathways and main toxic intermediates

A comprehensive database exists to characterize the absorption, distribution, metabolism and excretion of TCE in humans and in experimental animals. The data support a conclusion that qualitative similarities in toxicokinetics of TCE are evident between humans and rodents. TCE is readily absorbed through pulmonary (Monster, et al., 1976), oral (Bruning, et al., 1998; Yoshida, et al., 1996) and dermal (Kezic, et al., 2001) routes in humans, as well as in other species (Dallas, et al., 1991; Greenberg, et al., 1999; Mahle, et al., 2007; Simmons, et al., 2002). Virtually complete absorption has been demonstrated following oral administration of TCE in rodents (Dekant, et al., 1986a; Prout, et al., 1985). Rapid systemic distribution has been characterized in humans and animals. Because of its high lipophilicity, TCE distributes widely to all tissues, especially those with high lipid content, in both humans and animals (Abbas & Fisher, 1997; Coopman, et al., 2003; Dehon, et al., 2000; Fisher, et al., 1998; Pellizzari, et al., 1982; Simmons, et al., 2002). Faster and more extensive uptake has been shown with an aqueous vehicle, as compared to oil (Withey, et al., 1983).

Metabolism is critical to the various adverse effects of TCE in biological systems and its adverse effects in target tissues can largely be attributed to specific metabolites. The basic metabolic pathways for TCE have been established over many years and have been summarized, including in schematic representation of the enzymes and metabolites, in several reviews (Chiu, et al., 2009; Lash, et al., 2000a). Two major metabolic pathways of TCE have been characterized in both humans and laboratory animals, cytochrome P450 (CYP)-dependent oxidation and glutathione (GSH) conjugation.

The major pathway for TCE metabolism is CYP-mediated oxidation, resulting in formation of a variety of short- and long-lived species. In addition to CYP2E1, the major CYP enzyme that metabolizes organic solvents, the other human CYP enzymes reported to have some activity with TCE as substrate include CYP1A1/2, CYP2A6, and CYP3A4 (Hissink, et al., 2002; Lash, et al., 2000a). In rodents, Cyp2e1, Cyp1a1/2, Cyp2b1/2, Cyp2c11 (Nakajima, et al., 1992a) and Cyp2f2 (Forkert, et al., 2005) have been shown to metabolize TCE in liver and other tissues. The initial step in TCE metabolism by CYP enzymes results in formation of an enzyme-bound chemically unstable intermediate (TCE-O-CYP) which can be converted to TCE-epoxide (TCE-O), N-hydroxy-acetyl-aminoethanol, or chloral/chloral hydrate (CHL/CH). TCE-O spontaneously generates dichloroacetyl chloride (DCAC), another chemically unstable and reactive species, or oxalic acid, which is a stable product found in urine. DCAC undergoes spontaneous dechlorination to dichloroacetic acid (DCA). CHL/CH can be reduced by aldehyde dehydrogenase or CYPs to trichloroethanol (TCOH), or oxidized by ALDH to trichloroacetic acid (TCA). TCOH can also be oxidized by CYPs to yield TCA, or can undergo glucuronidation by UDP-glucuronosyltransferases to produce trichloroethanol glucuronide. While oxidative metabolism of TCE occurs predominantly in the liver, it can also occur in the lungs (Forkert, et al., 2006; Odum, et al., 1992), kidneys (Cummings & Lash, 2000), and testes (Forkert, et al., 2003; Forkert, et al., 2002). Most of the products of TCE oxidation, except for chemically unstable TCE-O-CYP, TCE-O and DCAC, are systemically available and can be distributed to, and cause adverse effects in, other tissues.

GSH conjugation is another important metabolic pathway resulting in formation of short-lived and reactive metabolites. An S_N2 nucleophilic displacement reaction of TCE with GSH generates Cl^– ion and S-(1,2-dichlorovinyl)glutathione (DCVG) as products. Although this initial GSH conjugation step can occur in many tissues, it occurs primarily in the liver owing to the high hepatic content of glutathione S-transferases. Subsequent processing of DCVG derived in liver or other tissues occurs primarily in the kidneys by a sequence of two hydrolytic enzymes on the proximal tubular brush-border membrane, γ-glutamyltransferase and cysteinylglycine dipeptidase (DP). The resulting cysteine conjugate, S-(1,2-dichlorovinyl)-L-cysteine (DCVC) (Lash, et al., 2000a; Lash, et al., 2000b), is further metabolized via three pathways. First, DCVC can be N-acetylated by the microsomal cysteine conjugate N-acetyltransferase to form the mercapturate N-acetyl-S-(1,2-dichlorovinyl)-L-cysteine (NAcDCVC). NAcDCVC can convert back to DCVC via deacetylation within the renal proximal tubular cell by aminoacylase (Uttamsingh & Anders, 1999). Second, DCVC can generate the reactive thiolate S-(1,2-dichlorovinyl)-thiol (DCVT) via action of cysteine conjugate β-lyase. DCVT spontaneously rearranges to form either chlorothioketene or chlorothionoacetyl chloride (Volkel & Dekant, 1998), both of which are chemically unstable and reactive and form covalent adducts with nucleic acids (Muller, et al., 1998), proteins (Hayden, et al., 1991), and phospholipids (Hayden, et al., 1992). Third, DCVC can yield a reactive S-(1,2-dichlorovinyl)-L-cysteine sulfoxide by flavin-containing monooxygenase. Except for DCVG, DCVC and NAcDCVC, few of the products of TCE GSH conjugation are systemically available and can be distributed to, or cause adverse effects in, other tissues. Only NAcDCVC has been recovered in urine of both experimental animals (Bernauer, et al., 1996) and humans (Bernauer, et al., 1996; Birner, et al., 1993) exposed to TCE or DCVC.

In all species, TCA and TCOH/TCOG are formed in vastly larger amounts than other oxidative metabolites. Evidence exists supporting quantitative differences in the extent of oxidative TCE metabolism among species at higher exposures, but at lower exposures oxidation is hepatic blood-flow limited. In addition, in rodents, data suggest that the amount of metabolites derived from the glutathione conjugation pathway is orders of magnitude less than those formed via oxidation (Kim, et al., 2009). In humans, limited data exist to characterize the extent of GSH conjugation (Lash, et al., 1999).

Route and dose of TCE exposure have a significant impact on the pattern of excretion due to a major role of hepatic oxidative metabolism. When inhaled, a significant fraction is expired un-metabolized in both animals and humans, but this is a major route of elimination from oral exposures only at higher doses (Chiu, et al., 2007; Dekant, et al., 1986a; Prout, et al., 1985). Urinary excretion of metabolites (primarily TCA, TCOH/TCOG and NAcDCVC) has been shown in both human and rodent studies as the other major excretory route (Dekant, et al., 1984; Prout, et al., 1985). Fecal elimination represents a minor pathway (Kim & Ghanayem, 2006; Prout, et al., 1985). TCE and its metabolites have been shown to cross the placenta and to be excreted in breast milk (Laham, 1970; Pellizzari, et al., 1982).

3.2. Genotoxicity of TCE and its metabolites

TCE and its major systemically available metabolites (TCA, DCA, CH, DCVC, DCVG, and TCOH) have been evaluated to varying degrees for their genotoxic activity in several in vitro systems such as bacteria, yeast, and mammalian cells, as well as in in vivo systems (Tabrez & Ahmad, 2009). Overall, there is strong evidence from TCE itself and from its metabolites to conclude that, following metabolism, TCE can be genotoxic, particularly in the kidney where in situ metabolism occurs (Guha, et al., 2012).

3.2.1. Genotoxicity studies of TCE

Carefully controlled studies evaluating TCE itself (without mutagenic stabilizers and without metabolic activation) found it to be incapable of inducing gene mutations in most standard bacterial mutagenesis assays [reviewed in (EPA, 2011)]. Therefore, it appears unlikely that TCE is a direct-acting mutagen. Interestingly, a recent study showed that TCE-spiked water is mutagenic in the Ames test with metabolic activation at high concentrations of TCE (Tabrez & Ahmad, 2012). TCE is positive in some, but not all, fungal and yeast mutagenicity systems. Some positive data comes primarily from studies in yeast strains with high CYP content (Callen, et al., 1980), although the S9 fraction showed no effect on the genotoxicity of TCE in yeast (Koch, et al., 1988).

Human studies provide inconclusive evidence of a genotoxic effect of TCE with some reports of an elevated frequency of structural aberrations and sister chromatid exchanges (SCEs) in a subset of exposed individuals (Kumar, et al., 2009; Seiji, et al., 1990). Several studies have evaluated TCE in relation to mutations of the von-Hippel-Lindau tumor suppressor gene (VHL). VHL mutations have been reported in a relatively large percentage of renal cell carcinoma (RCC) cases. One research group has suggested that VHL mutation frequency is associated with occupational exposure to TCE in subjects with histologically confirmed RCC who were employed in metal degreasing in Germany (Brauch, et al., 1999; Brauch, et al., 2004; Bruning, et al., 1997b). However, two recent large studies showed that VHL mutations are generally observed with high frequency in RCC, irrespective of exposure to TCE and other chlorinated solvents (Charbotel, et al., 2007; Moore, et al., 2011).

Exposure to TCE can lead to binding to nucleic acids and proteins, and such binding appears to be due to conversion to one or more reactive metabolites (Miller & Miller, 1983). For instance, increased binding was observed in samples bio-activated with mouse and rat microsomal fractions (Mazzullo, et al., 1992; Stott, et al., 1982). DNA binding is consistent with the ability to induce DNA and chromosomal perturbations. Several studies reported that TCE exposure resulted in induction of micronuclei in vitro (Hu, et al., 2008; Robbiano, et al., 2004; Wang, et al., 2001) and in vivo (Hrelia, et al., 1994; Kligerman, et al., 1994; Robbiano, et al., 2004). It was reported that TCE may cause SCE induction in vitro (Galloway, et al., 1987; Gu, et al., 1981), but not in vivo (Kligerman, et al., 1994; Nagaya, et al., 1989).

3.2.2. Genotoxicity studies of oxidative metabolites of TCE

Strong evidence is available to suggest that CH/CHL may cause genotoxicity. Numerous studies have shown that CH/CHL is genotoxic both in vivo and in vitro, in mammalian and other test systems, including studies with and without metabolic activation (Beland, 1999; Fellows, et al., 2011; Giller, et al., 1995; Harrington-Brock, et al., 1998; Liviac, et al., 2010; Pant, et al., 2008; Zhang, et al., 2012). The types of genotoxic damage detected encompass mutations, chromosomal aberrations, micronuclei, and cell transformation. One study found that in infants exposed to CH orally, a significant increase in micronuclei in peripheral blood lymphocytes was observed (Ikbal, et al., 2004). In bacterial and fungal genotoxicity test systems, positive results have been observed in substitution mutation assays (Beland, 1999; Giller, et al., 1995; Haworth, et al., 1983), in yeast (Albertini, 1990) and Drosophila (Beland, 1999; Zordan, et al., 1994).

Evidence suggests that TCA is not genotoxic. In vitro, in cultured human lymphocytes or in a lymphoblast cell line, no genotoxicity (chromosomal aberrations or strand breaks) was observed with neutralized TCA (Chang, et al., 1992; Mackay, et al., 1995). In mammalian systems in vivo, inconsistent evidence is available that TCA affects DNA strand breaks, micronuclei or chromosomal aberrations with most studies reporting no effects (Chang, et al., 1992; Mackay, et al., 1995; Nelson, et al., 1989; Styles, et al., 1991). In mammalian in vitro studies, TCA had no genotoxic effects (Harrington-Brock, et al., 1998; Zhang, et al., 2010). Several studies that examined mutational profile of mouse liver tumors following TCA administration found no differences from those in spontaneous tumors (Bull, et al., 2002; Ferreira-Gonzalez, et al., 1995). TCA was found to be overwhelmingly negative in bacterial and fungal test systems (DeMarini, et al., 1994; Giller, et al., 1997; Kargalioglu, et al., 2002; Waskell, 1978).

Weak to moderate evidence is available to suggest that DCA may cause genotoxicity. There is only one study in a human lymphoblast cell line in vitro in which no DNA strand breaks have been observed (Chang, et al., 1992). In mammalian systems, gene mutations were reported from in vivo experiments with DCA (Leavitt, et al., 1997), and limited evidence exists for increased mutations after treatment in vitro (Harrington-Brock, et al., 1998; Zhang, et al., 2010). DCA was positive for induction of chromosomal aberrations in mouse lymphoma cells (Harrington-Brock, et al., 1998), but not in CHO cells (Fox, et al., 1996). With regards to micronuclei, there are conflicting results (Fuscoe, et al., 1996; Harrington-Brock, et al., 1998). Inconsistent evidence exists to suggest that DCA can cause DNA damage (e.g., DNA unwinding) in bone marrow and blood leukocytes in in vivo animal studies (Chang, et al., 1992; Fuscoe, et al., 1996; Nelson & Bull, 1988). In addition, several studies have found specific H-ras codon 61 mutations in liver tumors following DCA administration, distinct from those in spontaneous tumors (Anna, et al., 1994; Ferreira-Gonzalez, et al., 1995). In bacterial and fungal genotoxicity test systems, positive results have been observed only in substitution mutation assays [reviewed in (Plewa, et al., 2002)].

Limited studies are available on the effect of TCOH on genotoxicity, as TCOH has not been evaluated in most recommended genotoxicity screening assays. TCOH was negative in the S. typhimurium assay using the TA100 strain (DeMarini, et al., 1994; Waskell, 1978), and positive in one S. typhimurium strain TA104 study in the presence of exogenous metabolic activation (Beland, 1999).

3.2.3. Genotoxicity studies of GSH conjugation metabolites of TCE

DCVG has not been evaluated in most recommended genotoxicity screening assays. The mutagenicity of DCVG was evaluated in S. typhimurium strain TA2638, using kidney subcellular fractions for metabolic activation and a beta-lyase inhibitor (Vamvakas, et al., 1988b). DCVG exhibited direct-acting mutagenicity, with kidney mitochondria, cytosol, or microsomes enhancing the effects and a beta-lyase inhibitor diminishing, but not abolishing the effects. Addition of liver subcellular fractions did not enhance the mutagenicity of DCVG, consistent with in situ metabolism playing a significant role in the genotoxicity of DCVG-derived species in the kidney.

DCVC has demonstrated a strong, direct-acting mutagenicity both with and without the presence of mammalian activation enzymes, including those derived from the kidney, in bacterial mutagenesis tests (Dekant, et al., 1986b; Irving & Elfarra, 2013; Vamvakas, et al., 1988b). The genotoxicity of DCVC is further supported by the predominantly positive results in other available in vitro and in vivo assays. The observed effects include DNA strand breaks (Clay, 2008; Jaffe, et al., 1985) and unscheduled DNA synthesis (Vamvakas, et al., 1989; Vamvakas, et al., 1988a), but not micronuclei.

S-(1,2-dichlorovinyl)-l-cysteine sulfoxide (DCVCS), a product of sulfoxidation of DCVC, was found to be mutagenic in Ames Salmonella TA100 strain (Daoud & Irving, 1977). NAcDCVC was also shown to exhibit direct-acting mutagenicity in the absence of exogenous metabolic activation, with kidney cytosol enhancing the effects and a beta-lyase inhibitor diminishing, but not abolishing the effects (Vamvakas, et al., 1987).

4. Non-genotoxic mechanisms of TCE carcinogenesis

Mechanistic data can aid in interpreting the observed cancer effects of TCE, including evidence of a multisite carcinogenicity in rodents, and of association with elevated risk of several tumor types in epidemiologic studies of occupational cohorts. Target organs of carcinogenicity that have been identified in these studies include kidney, liver, lung, immune system and testis. The non-genotoxic mechanistic data concerning the cancers that have been associated with TCE exposure are reviewed below.

4.1. Kidney

Overall, the evidence for kidney as a target tissue for TCE carcinogenicity is strong. Supporting evidence exists for three non-genotoxic mechanisms of carcinogenesis: α2u-globulin-associated nephropathy, cytotoxicity not associated with α2u-globulin accumulation, and peroxisome proliferator activated protein (PPAR)α activation. However, the overall database in each case is limited, none of the mechanisms have been tested experimentally, and the evidence for each lacks specificity with respect to species and dose.

4.1.1. α2u-Globulin-associated nephropathy

α2u-Globulin accumulation is a histopathological phenomenon elicited by chronic chemical exposure in male rat kidney. This phenomenon, if supported by the data, may be used to conclude lack of relevance of male rat kidney tumor observations to human health. The available evidence for TCE was found insufficient to meet the criteria defined by IARC, all of which must be satisfied for male rat kidney tumors to be attributed to an α2u-globulin-associated response (Capen, et al., 1999). In particular, TCE and certain of its metabolites are genotoxic, as discussed above, and a lack of genotoxic activity (of the agent and/or metabolite) is one of the IARC criteria. Additionally, direct evidence that TCE induces the characteristic histopathology, α2u accumulation or reversible binding to α2u globulin in a male rat-specific manner, is lacking. To the contrary, Goldsworthy et al. (Goldsworthy, et al., 1988) reported that TCE did not induce increases in this urinary protein, nor did it stimulate cellular proliferation in rats. Indirect evidence supporting an increase in α2u-globulin and the associated histopathological changes is provided by a separate study of TCOH (Green, et al., 2003). A dose-related increase in both the incidence and severity of hyaline droplet accumulation was observed when rats were exposed to TCOH in drinking-water. However, as posited by the authors, the observed increases in α2u-globulin were insufficient to account for the renal pathology.

4.1.2. Cytotoxicity not associated with α2u-globulin accumulation

Several human studies provide evidence for cytotoxicity in the kidneys associated with TCE exposure. Such evidence, in the form of increased excretion of nephrotoxicity markers N-acetyl-beta-D-glucosaminidase, urinary protein, albumin or Kim-1, has been presented at low (Vermeulen, et al., 2012), occupational (Green, et al., 2004), and high (Bolt, et al., 2004; Bruning, et al., 1999a; Bruning, et al., 1999b) levels of TCE exposure. Likewise, there is substantial evidence that TCE is nephrotoxic in both rats and mice by all routes of exposure. Such evidence comes from chronic bioassays (NCI, 1976; NTP, 1988; NTP, 1990), as well as sub-chronic studies (Chakrabarti & Tuchweber, 1988; Cojocel, et al., 1989; Green, et al., 1997a; Messing, et al., 2003).

Several mechanisms have been suggested to account for cytotoxicity of TCE in the kidneys. In vitro studies with GSH conjugates of TCE show unequivocally that DCVC and its metabolites (e.g., DCVCS) are cytotoxic to primary human proximal tubular cells (Cummings & Lash, 2000; Lash, et al., 2005; Lock, et al., 2006). Similar observations were made in rodent cells (Lash, et al., 1986; Stevens, et al., 1986). In addition, increased formation and urinary excretion of formic acid mediated by the oxidative metabolites TCA or TCOH (Dow & Green, 2000; Green, et al., 2003) have also been posited to contribute to the observed nephrotoxicity of TCE. However, these oxidative metabolites do not appear sufficient to explain the range of renal effects observed after TCE exposure. Other mechanisms of cytotoxicity, including alteration of calcium ion homeostasis and mitochondrial dysfunction, have been identified in vitro in kidney cells (Vamvakas, et al., 1992; van de Water, et al., 1994).

The primary limitation to the evidence supporting the role of this mechanism in kidney cancer is that nephrotoxicity is observed in both mice and rats, in some cases with nearly 100% incidence in all dose groups, but kidney tumors are only observed at low incidences in rats at the highest tested doses (NCI, 1976; NTP, 1990). In rats carrying the Eker mutation, Mally et al. (Mally, et al., 2006) reported TCE-induced increase in cell proliferation, but no evidence of clonal expansion or tumorigenesis in the form of increased pre-neoplastic or neoplastic lesions, as compared to controls. Therefore, data demonstrating a causal link between compensatory proliferation and the induction of kidney tumors are lacking.

4.1.3. PPARα activation

No study addressed the potential for PPARα-activation mechanism in human kidney; however, in vitro transactivation studies that are not specific to the kidney have shown that human PPARα is activated by TCA and DCA, while TCE itself is relatively inactive (Maloney & Waxman, 1999; Zhou & Waxman, 1998). Human hepatocytes transfected with mouse PPARα displayed increased expression of PPARα, and increased peroxisome proliferator response element-reporter activity when treated with TCA and DCA (Walgren, et al., 2000). Limited evidence exists to suggest that TCE exposure induces peroxisome proliferation in the kidney of exposed rodents. Peroxisome proliferation in the kidney has been evaluated by only one study of TCE (Goldsworthy & Popp, 1987). Increases in renal palmitoyl-CoA oxidation activity were observed in rats and mice treated with TCE. No significant increases in kidney/body weight ratios were observed in either species.

4.2. Liver

The evidence for liver as a target tissue for TCE carcinogenicity is strong and it is likely that multiple genotoxic and non-genotoxic mechanisms are operational in the liver. TCE and its oxidative metabolites have been shown to induce several non-genotoxic effects that may contribute to hepatocellular tumors. These include epigenetic alterations; cytotoxicity and secondary oxidative stress; alteration of proliferation and apoptosis, and clonal expansion; and PPARα activation.

4.2.1. Epigenetic alterations

Several studies tested the hypothesis that hypomethylation of DNA may be related to the carcinogenicity of TCA and DCA in mice. In an initiation (i.p. injection of N-methyl-N-nitrosourea)-promotion (administration of TCA or DCA in drinking-water) study design, DNA methylation in the resulting hepatocellular adenomas and carcinomas was about half that observed in noninvolved tissue from the same animal or from animals treated with only the initiating agent (Tao, et al., 1998). Sub-chronic drinking-water exposure to TCA or DCA also resulted in a decreased total liver DNA methylation (Tao, et al., 1999). An association between hypomethylation and cell proliferation in liver of TCA- or DCA-exposed mice was demonstrated by Ge et al. (Ge, et al., 2001). Nonetheless, experimental evidence of DNA hypomethylation by TCE is limited to studies of TCA and DCA in mouse liver. Together with the fact that methylation changes represent common early molecular events in most tumors (Pogribny & Rusyn, 2013), these data support the plausibility of a hypothesis that dysregulation of gene methylation may play a role in TCE-induced tumorigenesis. However, no data from human or experimental animal studies are available specifically testing this hypothesis for TCE.

4.2.2. Cytotoxicity and secondary oxidative stress

Several cohort studies in humans reported significant changes in serum liver function tests, or plasma or serum bile acid changes in subjects with occupational exposure to TCE [reviewed in (EPA, 2011)]. Case reports of liver damage and cirrhosis (Kamijima, et al., 2007; Thiele, et al., 1982) in workers exposed to high amounts of TCE have also been published. However, the overall database of this mechanism in vivo in humans is limited. Numerous animal studies have demonstrated that TCE is hepatotoxic, as evidenced by the elevation of serum enzymes and bile acids (Hamdan & Stacey, 1993; Ramdhan, et al., 2008). A few studies examined TCA- and DCA-induced hepatic oxidative stress and demonstrated small, albeit significant, increases in lipid peroxidation and oxidative DNA damage (Austin, et al., 1996; Parrish, et al., 1996).

4.2.3. Alteration of proliferation and apoptosis, and clonal expansion

This mechanism has not been addressed in human in vivo or in vitro studies of either TCE or its metabolites. Most of the experimental in vivo data on these mechanisms is available from studies in mice, including PPARα-null mice. All studies that examined hepatocellular proliferation in response to TCE exposure reported a significant effect, albeit at different doses and time points (Channel, et al., 1998; Dees & Travis, 1993; Mirsalis, et al., 1989). Sano et al. (Sano, et al., 2009) observed mitotic figures in the livers of TCE-treated mice, but not in rats but no quantitative analysis was reported.

With regards to changes in apoptosis, Dees and Travis (Dees & Travis, 1993) is the only study that reported a qualitative increase in liver apoptotic bodies in B6C3F1 mice exposed to TCE. Channel et al. (Channel, et al., 1998), and Sano et al. (Sano, et al., 2009), found no differences in liver apoptosis between control or TCE-exposed animals.

The effects of TCE on cell proliferation were examined in PPARα-null mice. BrdU incorporation, a measure of DNA synthesis that may reflect cell division, polyploidization, or DNA repair, was observed to be diminished in null mice as compared to wild-type mice (Laughter, et al., 2004). BrdU incorporation in null mice was still about threefold higher than controls, although it was not statistically significantly different due to the small number of animals, high variability, and the two- to threefold higher baseline levels of BrdU incorporation in control null mice as compared to control wild-type mice. Liver weight changes have also been examined in PPARα-null mice in response to TCE; however, the two available studies present conflicting results (Laughter, et al., 2004; Nakajima, et al., 2000).

4.2.4. PPARα activation

As detailed in section 4.1.3, in vitro transactivation studies have shown that human (as well as murine) versions of PPARα are activated by TCA and DCA, while TCE itself is relatively inactive. Numerous studies have reported that TCE administered to mice and rats leads to proliferation of peroxisomes in hepatocytes. Some studies have measured changes in the volume and number of peroxisomes as measures of peroxisome proliferation (Channel, et al., 1998; Elcombe, 1985; Nakajima, et al., 2000), while others have measured peroxisomal enzyme activity (Elcombe, 1985; Goldsworthy & Popp, 1987; Laughter, et al., 2004; Melnick, et al., 1987). Two studies (Laughter, et al., 2004; Nakajima, et al., 2000) found diminished responses in PPARα-null mice in terms of increased peroxisomal volume and peroxisomal enzyme activities, although there was some confounding due to baseline differences between null and wild-type control mice in several measures. No TCE-specific study has directly tested the hypothesis that TCE-induced PPARα-activation, along with its sequelae, are key or causative events in TCE-induced hepatocarcinogenesis (e.g., bioassays with knockout mice or involving the blocking of hypothesized key events).

The hypothesis that TCE induces tumors through PPARα-activation is based on the fact that TCA is a major metabolite that may activate PPARα and induce peroxisome proliferation and hepatocyte proliferation. However, several data gaps reduce the confidence in the conclusion that TCA induces hepatocarcinogenesis solely through a PPARα-activation mechanism. First, while TCA induces peroxisome proliferation (a marker for PPARα-agonism) in both rats and mice, to date, TCA has been shown to be tumorigenic in B6C3F1 mice but not F344 rats (DeAngelo, et al., 1997). Second, the tumor phenotype of TCA-induced mouse liver tumors has been reported to have a different pattern of H-ras mutation frequency from DCA and other peroxisome proliferators (Bull, et al., 2002; Fox, et al., 1990). Other effects of TCA, including increased c-myc expression and hypomethylation of DNA, are not specific to the PPARα-activation mechanism, and other data also contribute uncertainty as to whether PPARα independent mechanisms may be involved in TCA-induced tumors in mice.

4.3. Immune system

Human and animal studies of TCE provide strong evidence for a role of TCE in autoimmune disease (Cooper, et al., 2009). Evidence supports a specific type of generalized hypersensitivity syndrome, as well as in alteration of immune response. Several molecular epidemiology studies have evaluated the effect of TCE exposure on immune marker concentrations in exposed individuals and its potential to result in immunosuppressive effects. Most have been of workers occupationally exposed to TCE and unexposed control workers, and have been conducted using a cross-sectional design with an exposure monitoring period preceding blood or other specimen collection. All studies showed significant effects on white blood cell counts or serum cytokine levels (Hosgood, et al., 2011; Iavicoli, et al., 2005; Lan, et al., 2010; Zhang, et al., 2013). In addition, there have been a large number of case reports of a severe hypersensitivity skin disorder, distinct from contact dermatitis and often accompanied by hepatitis, associated with occupational exposure to TCE (Kamijima, et al., 2007; Li, et al., 2007).

Experimental animal studies provide additional support for the immunotoxicity of TCE. Numerous studies have demonstrated accelerated autoimmune responses in autoimmune-prone mice (Blossom & Doss, 2007; Griffin, et al., 2000). With shorter exposure periods, effects include changes in cytokine levels similar to those reported in human studies. More severe effects, including autoimmune hepatitis, inflammatory skin lesions, and alopecia, manifest at longer exposure periods. Immunotoxic effects, including increases in anti-dsDNA antibodies in adult animals, decreased thymus weights have been also reported in B6C3F1 mice, which do not have a known particular susceptibility to autoimmune disease (Keil, et al., 2009; Peden-Adams, et al., 2006). Evidence of a treatment-related increase in delayed hypersensitivity response has been observed in guinea-pigs (Tang, et al., 2002) and in mice (Peden-Adams, et al., 2006). Evidence of localized immunosuppression, as measured by pulmonary response to bacterial challenge, was observed in acute exposure studies in CD-1 mice (Selgrade & Gilmour, 2010).

There is a well-established connection between immune status and carcinogenesis. In particular, an increased risk of lymphoma has been associated with a history of immunosuppressive medication use, with certain chronic infections such as HIV, and with certain autoimmune diseases and lifestyle factors which result in immune alterations and abnormalities (Whiteside, 2006). Furthermore, more subtle changes in immune functioning, including imbalances in Th1/Th2 responses resulting from cytokine alterations, have in general been implicated in the oncogenic process via regulation of transcriptional factors and of tumor growth, angiogenesis, and cell differentiation and survival (Tan & Coussens, 2007). However, evidence defining the precise mechanisms by which the immunotoxic effects of TCE may ultimately lead to carcinogenesis has not been identified.

4.4. Lung

The hypotheses concerning non-genotoxic mechanisms for TCE-induced lung tumors are limited to a cytotoxicity mechanism. The human evidence of the cytotoxicity of TCE is very limited. Two reports of a study of gun-manufacturing workers (Cakmak, et al., 2004; Saygun, et al., 2007) indicate pulmonary toxicity of smoking and exposure to solvents, with smoking having the most important effect on asthma-related symptoms. Smoking, but not solvent exposure, was shown as a statistically significant predictor of lung function decrements.

Acute cytotoxicity in the bronchiolar Clara cell and transient cell proliferation following TCE exposure were observed in mouse studies (Buckpitt, et al., 1995; Forkert & Forkert, 1994; Green, et al., 1997b; Villaschi, et al., 1991). In sub-chronic studies, pulmonary fibrosis was observed 90 days following i.p. administration of a single dose of TCE (Forkert & Forkert, 1994). The effects were in the lung parenchyma, not in the bronchioles where Clara cell damage has been observed after acute exposure. Because the cell type (or types) of origin for the observed lung tumors in mice has not been determined, the contribution to carcinogenicity of Clara cell toxicity and subsequent regenerative cell division is largely unknown. Similarly, dichloroacetyl-lysine protein adducts have only been studied in i.p. exposure paradigms of short duration (Forkert, et al., 2006), and the contribution of these adducts to the tumor response has not been investigated.

Single inhalation experiments (Odum, et al., 1992; Villaschi, et al., 1991) suggest that the Clara cell is also not the target for TCE exposure in rats. In a continuous exposure (5 days) study in rats (Le Mesurier, et al., 1980), vacuolation in Type 1 alveolar cells, but not in Clara cells, and abnormalities in the endothelium and minor morphological changes in Type 2 alveolar cells were observed. Histopathology in these studies was recorded up to 15 days post-exposure, and it is not possible to discern whether the lack of reported Clara cell damage in rats following repeated exposure was due to recovery or lack of toxicity in this particular experiment.

4.5. Male Reproductive System

Hormonal disruption was suggested as a mechanism for tumorigenesis in the testis. A number of human and laboratory animal studies suggest that TCE exposure has the potential for male reproductive toxicity [reviewed in (EPA, 2011)]. Two human studies have reported TCE exposure to be associated with increased sperm density and decreased sperm quality, altered sexual drive or function, and altered serum levels of testosterone and other hormones (Chia, et al., 1997; Goh, et al., 1998). Adverse effects on male reproduction have also been reported, including histopathological lesions in the testes or epididymis (Kumar, et al., 2000; Kumar, et al., 2001). Overall, the database to support the mechanisms of carcinogenesis in the testes is weak.

5. Susceptibility to TCE toxicity and carcinogenicity

5.1. Inter-individual variability in TCE metabolism

Carcinogenicity and toxicity of TCE are associated with its metabolism. Polymorphisms in metabolism genes in both oxidative (e.g., CYP2E1, ADH, ALDH) and glutathione conjugation (e.g., GSTs) pathways have been studied in connection to susceptibility to TCE toxicity and carcinogenicity. The oxidative metabolism of TCE is largely dependent on CYP2E1 and clinically-relevant genetic polymorphisms are known to exist in this gene (Daly, 2012), but the functional significance of these variants is still unclear. The only indirect evidence of the potential role of CYP2E1 polymorphisms in TCE toxicity has been reported in a study by Povey et al. (Povey, et al., 2001) who reported that the CYP2E1*3 allele may be associated with susceptibility to TCE-associated scleroderma. Relevance of this finding to cancer outcomes is difficult to ascertain.

While it has not been firmly established what GST enzymes are responsible for TCE metabolism, allelic polymorphisms of GSTs in humans have been associated with variations in enzyme activity (Katoh, et al., 2008). Brüning et al. (Bruning, et al., 1997a) investigated the potential for an association between polymorphisms of GSTM1 and GSTT1 and risk of renal cell cancer in workers with long-term high occupational exposure to TCE. Odds ratios for renal cell cancer were 2.7 for GSTM1+ individuals (95%CI: 1.2–6.3) and 4.2 for GSTT1+ individuals (95%CI: 1.2–14.9). The data from this study and from additional control subjects were re-evaluated by Wiesenhütter et al. (Wiesenhutter, et al., 2007). These authors found no genetic influences on the development of renal cancer due to TCE (e.g., no effect of deletion polymorphisms of GSTT1 and GSTM1, or of NAT2 rapid/slow acetylator genotype). The authors did find that renal cell cancer cases displayed a somewhat higher proportion of the homozygous GSTP1 313A wild type (GSTP1*A), although this was not statistically significant.

Moore et al. (Moore, et al., 2010) found a significant association among TCE-exposed subjects with at least one intact GSTT1 allele (active genotype), but not among subjects with two deleted alleles (null genotype). Similar associations for all exposure metrics including average intensity were observed among GSTT1-active subjects, but not among GSTT1 nulls.

A recent study in a panel of inbred mouse strains showed that TCE metabolism through oxidative and conjugative pathways varied considerably among strains (Bradford, et al., 2011). Peroxisome proliferator-activated receptor-mediated pathways, consisting of the metabolism genes known to be induced by TCE, were identified as some of the most pronounced molecular effects of TCE treatment in mouse liver that were dependent on genetic background. Conversely, cell death, liver necrosis, and immune-mediated response pathways, which were altered by TCE treatment in liver, were found to be genetic background independent.

The effect of ADH/ALDH genotype on formation of TCOH and TCA was examined (Bronley-DeLancey, et al., 2006) in 13 human hepatocyte samples. While large variation in Vmax values was observed among individuals, disposition of CH into downstream metabolites was found to be relatively constant. The authors concluded that cellular factors other than genotype may contribute to the observed variability in metabolism of chloral hydrate in human liver. CH is an inhibitor of ALDH (Wang, et al., 1999), thus suggesting that production of TCA from CH may not increase in linear fashion with dose. An inhibitory effect of CH on liver ADH was also reported in studies in the mouse (Sharkawi, et al., 1983).

Several kidney transporters (e.g., OAT1 and OAT3) that are likely responsible for the uptake and cellular accumulation of DCVG and DCVC are known to be polymorphic in humans (Urban, et al., 2006). Thus, different subpopulations of humans may have a markedly different capacity to accumulate DCVG or DCVC, which may affect their susceptibility to nephrotoxicity.

5.2. Life stage susceptibility

With respect to the life stage susceptibility, data are available to suggest that both early (especially pre- and neo-natal) and later life stages may be more susceptible to adverse health outcomes of exposure to TCE. This evidence includes life stage-specific differences in exposure routes (e.g., placental transfer and breast milk in early life stages) or toxicokinetics, or dissimilarities in adverse health outcomes (e.g., developmental cardiac abnormalities).

It has been shown that TCE can be transferred to the fetus by placenta in all mammalian species studied, including humans. Infants that are fed breast milk by mothers exposed to TCE have been shown to receive appreciable amounts of TCE (Abbas & Fisher, 1997). While TCE has been found in the blood and tissues of fetuses born to exposed mothers, studies disagree on whether such concentrations are higher or lower than those found in maternal tissues (Withey & Karpinski, 1985). Another factor contributing to higher early life exposure may be the higher consumption by children of dairy products, which have been found to contain TCE (Wu & Schaum, 2000). Because children have higher ventilation rate than adults, as well as increased alveolar surface area for the first 2 years of life, greater absorption of TCE is expected in early life stages. However, no data comparing TCE absorption in children and adults to directly corroborate these assumptions are available.

Several early life stage-specific adverse health outcomes of TCE exposure during pregnancy or neonatal development have been reported in humans and other species. These include cardiac birth defects, neural tube defects, oral clefts, and choanal atresia (Bove, et al., 2002). It should be noted, that a large number of epidemiological and experimental studies have not observed significant associations between TCE and these developmental abnormalities (Watson, et al., 2006). Several studies evaluated the potential for susceptibility to cancer outcomes in early life stages. Most studies have found no evidence that children may be more susceptible than adults; however, studies contained small number of cases and had poor exposure characterization.

Limited evidence exists to suggest that TCE exposures in adults of advanced age (>65) may lead to greater adverse health effects of TCE. Some have suggested that toxicokinetic parameters in later life stages are different from those in young adults (Benedetti, et al., 2007); however, there is little evidence to suggest that expression of CYP2E1 or GSTs differs with age in adults. While several studies have documented significant age-related declines in their amounts, most liver functions in humans appear to be well preserved with age (Schmucker, 2001). In addition to metabolism, it has been suggested that drug metabolism-related clearance may also be age-dependent (Hilmer, 2008). However, the evidence for TCE is limited.

5.3. Gender differences

Gender differences have been noted in adverse health outcomes associated with exposure to TCE. Gender-specific susceptibility has been linked to differences in exposure, physiological factors (e.g., hormonal status), and toxicokinetics.

Several studies examined gender differences in toxicokinetic parameters of TCE using physiologically-based pharmacokinetic modeling. Sato et al. (Sato, et al., 1991) reported that absorption, distribution, metabolism and excretion of TCE vary according to the different anatomical features of men and women. In a follow up study, Sato (Sato, 1993) concluded that there is a sex difference in the pharmacokinetic profiles of TCE and although retention of TCE in the body is greater in men than in women, the blood concentration of TCE in women is higher than in men 16 hours after exposure. Fisher et al. (Fisher, et al., 1998) evaluated gender-specific differences in uptake and metabolism of TCE in humans using human exposure data. This study concluded that only minor gender differences exist in TCE toxicokinetics.

Lash et al. (Lash, et al., 2006) evaluated metabolism and tissue distribution of orally administered trichloroethylene in male and female rats. Higher levels of TCE were generally observed in tissues of males at lower doses. Higher concentrations of oxidative metabolites were observed in livers of males than in females, whereas the opposite pattern was observed in kidneys. DCVG was recovered in liver and kidneys of female rats only and in blood of both males and females, with generally higher amounts found in females. DCVC was recovered in male and female liver, female kidneys, male blood, and in urine of both males and females. However, Nakajima et al. (Nakajima, et al., 1992b) found no sex difference in TCE metabolism to CH in a study with liver microsomes from 3- and 18-week-old male and female rats.

Two human studies evaluating kidney effects have concluded that females may be more susceptible to kidney disease and diabetes in association with TCE exposure (Davis, et al., 2005); however, males were reported to be more sensitive to renal toxicity in studies in rats (Lash, et al., 2001). Gender differences in cancer susceptibility to TCE are well established. In rats, exposure to TCE by inhalation or gavage caused kidney cancer (tubular adenocarcinoma) only in males. Leukemia was also observed in males, albeit low survival of the animals was noted as a challenge in interpreting this study (Maltoni, et al., 1988). In mouse studies, no gender difference in the incidence of liver or lung tumors was observed. Lymphomas were reported only in female mice (Henschler, et al., 1980).

Raaschou-Nielsen et al. (Raaschou-Nielsen, et al., 2003) evaluated cancer risk among workers at Danish companies using trichloroethylene in a cohort study. No significant gender differences in the incidence of tissue-specific cancer were observed between genders in this study. Most other epidemiological studies of cancer and TCE exposure also did not report gender differences.

5.4. Life style factors and nutrition (e.g., alcohol drinking)

Exposure to alcohol and other chemicals may impact TCE metabolism, while nutrition or obesity may affect the internal concentrations of TCE and its metabolites. Overall, while mechanistically plausible, the associations between TCE exposure, disease outcomes and such additional factors has not received much attention in experimental studies and needs to be investigated further.

TCE is metabolized to chloral hydrate and further to trichloroacetic acid by aldehyde dehydrogenase (ALDH) and to trichloroethanol by alcohol dehydrogenase (ADH). ALDH and ADH are known to be polymorphic in humans, and these polymorphisms are well known to have major impact on cancer susceptibility in humans who consume ethanol-containing alcoholic beverages, especially in Asian countries (Baan, et al., 2007). It has been, therefore, suggested that polymorphisms in ALDH and ADH metabolic pathways may yield subpopulations with greater than expected formation of trichloroacetic acid with associated enhanced risk of adverse health effects after exposure to chloral hydrate or other chlorinated solvents.

The effects of TCE, likely through chloral hydrate, on alcohol and acetaldehyde metabolism have been suggested as a mechanism for dramatic effects of co-exposures to chlorinated solvents and alcohol. First, such co-exposures lead to more that additive sedative effects in rodents (Sharkawi, et al., 1983) and humans (Sellers, et al., 1972). Furthermore, adverse health effects indicative of elevated blood levels of acetaldehyde have been described as “degreaser's flush” (Stewart, et al., 1974).

Because alcohol exposure leads to increased liver activity of CYP2E1 (Bradford, et al., 2005; Powell, et al., 2010), TCE metabolism via oxidation pathway may be increased. Indeed, Nakajima et al. (Nakajima, et al., 1992a) observed an increase in TCE metabolism in rat liver microsomes from alcohol-pre-treated animals.

6. Summary and Conclusions

TCE has been extensively studied for its carcinogenicity. The database includes occupational studies providing sufficient evidence of an association between TCE exposure and kidney cancer. Two independent meta-analyses of case-control and cohort studies of kidney cancer reported statistically significant risks of modest magnitude. Positive associations, supported by a meta-analysis, have also been observed between occupational TCE exposure and risks for non-Hodgkin's lymphoma and cancer of the liver but this epidemiological evidence was characterized as limited. Statistically significant excess risks of cancers of the lung, cervix and esophagus were also observed in a few isolated studies, supporting a conclusion of inadequate evidence of carcinogenicity for these sites. Multiple experimental animal bioassays have identified mouse liver and rat kidney as sites of carcinogenic action of TCE in both sexes. Other reported tumors include mouse lung and rat testis, with more limited evidence of leukemias and lymphomas in rodents.

TCE is metabolized via two main pathways to multiple toxic, mutagenic and carcinogenic metabolites that are likely to contribute to the carcinogenicity of the parent compound. Supporting mechanistic findings include evidence of genotoxicity of TCE and its metabolites, with strongest evidence of metabolites formed from the glutathione pathway in the kidney. A recent epidemiologic study reported that kidney cancer risk was attenuated in individuals lacking glutathione conjugation gene GSTT1. This study provides evidence in support of the hypothesis that glutathione conjugation plays a critical role, and contributes significantly to an overall cohesive mechanistic database supporting the classification of TCE as carcinogenic in humans by IARC (Guha, et al., 2012).

The metabolism and fate of TCE, and the mechanisms by which TCE causes carcinogenicity continue to be the subject of scientific investigation. In the kidney, genotoxicity was concluded to be a contributing mechanism while the evidence was not adequate to support the non-genotoxic mechanisms of α2u, cytotoxicity, or PPARα. The data were also not adequate to support conclusions concerning the non-genotoxic effects contributing to liver carcinogenicity. It is likely that multiple mechanisms, potentially including immune dysregulation, epigenetic alterations, cytotoxicity and secondary oxidative stress, alteration of proliferation and/or apoptosis, may contribute to hepatocarcinogenesis. Few data are available to inform hypotheses for other targets of carcinogenicity, particularly the non-Hodgkin lymphomas in humans.

Differences in susceptibility to TCE may be due to life-stage, gender, life-style, genetics, and other factors. Evidence of inter-individual variability has been provided in humans and experimental animals, with differences in metabolism having been shown to affect outcome. These findings underscore the critical role of TCE metabolism in its carcinogenicity. Ongoing investigations are aimed at further elucidating the metabolites and mechanisms that influence TCE carcinogenicity, and the impact of co-exposures to carcinogens. Additionally, epidemiologic investigations are likely to provide additional insight into the target organs, mechanisms, and susceptibility factors of TCE carcinogenicity in exposed individuals.

Overall, the conclusion that TCE should be classified as carcinogenic to humans (Group 1) was based on convincing evidence for a positive association between exposure to TCE and (i) kidney cancer in humans (e.g., sufficient evidence for the carcinogenicity of TCE in humans) and (ii) liver neoplasms in more than two independent studies in mice (e.g., sufficient evidence for the carcinogenicity of TCE in experimental animals). Additional human epidemiology, animal bioassay, mechanistic, and toxicokinetic data corroborate and provide biological plausibility to a causal link between TCE exposure and cancer.