Abstract
Perchloroethylene (PCE) is used as a solvent and chemical intermediate. Following chronic inhalation exposure, PCE selectively induced liver tumors in mice. Understanding the mode of action (MOA) for PCE carcinogenesis in mice is important in defining its possible human cancer risk. The proposed MOA is based on the extensive examination of the peer-reviewed studies that have assessed the mouse liver effects of PCE and its major oxidative metabolite trichloroacetic acid (TCA). Similar to PCE, TCA has also been demonstrated to liver tumors selectively in mice following chronic exposure. The Key Events (KE) of the proposed PCE MOA involve oxidative metabolism of PCE to TCA [KE 1]; activation of the peroxisome proliferator-activated receptor alpha (PPARα) [KE 2]; alteration in hepatic gene expression including cell growth pathways [KE 3]; increase in cell proliferation [KE 4]; selective clonal expansion of hepatic preneoplastic foci [KE 5]; and formation of hepatic neoplasms [KE 6]. The scientific evidence supporting the PPARα MOA for PCE is strong and satisfies the requirements for a MOA analysis. The PPARα liver tumor MOA in rodents has been demonstrated not to occur in humans; thus, human liver cancer risk to PCE is not likely.
Keywords
Introduction
Tetrachloroethylene (CAS No. 127-18-4), also known as perchloroethylene (PCE) is a volatile chlorinated hydrocarbon that is used primarily as a feedstock in the production of refrigerants, but also as a solvent in a number of applications including dry cleaning and metal degreasing, and in oil refining catalyst regeneration (ATSDR, 2019). Historically, PCE has also been used as an anthelminthic for hookworms, and as a grain protectant and fumigant (ATSDR, 2019). Following chronic treatment in rodents, PCE has been reported to increase the incidence of hepatocellular adenomas and/or carcinomas in mice (JISA, 1993; NCI, 1977; NTP, 1986). In these studies, both male and female mice showed an increase in liver tumors, while similarly treated rats had no increase in liver tumor incidence. Based on these chronic treatment study results in rodents, PCE has been classified as “reasonably anticipated to be a human carcinogen” by the National Toxicology Program (NTP, 2022) and “probably carcinogenic to humans” by the International Agency for Research on Cancer (IARC, 2014).
The potential mechanisms for PCE-induced mouse liver tumors have been previously discussed by the U.S. Environmental Protection Agency (EPA) for its Integrated Risk Information System (IRIS) (USEPA, 2012) and more recently in its Toxic Substance Control Act (TSCA) Risk Evaluation (USEPA, 2020). In its Risk Evaluation (USEPA, 2020), the EPA concluded that “PCE likely induces liver tumors through multiple modes of action mediated largely by metabolites. TCA [trichloroacetic acid] appears to be an important hepatic metabolite but is probably not the only metabolite involved in hepatic effects of PCE.” These USEPA documents stated that the primary MOA for PCE-induced liver tumors is not PPARα activation, although it may have some influencing in the reported liver effects from PCE exposure (USEPA, 2020). However, these USEPA documents failed to provide an alternative scientifically accepted liver tumor MOA for the observed PCE-induced liver tumors in rodents. The USEPA documents did not provide a specific reason for ruling out PPARα activation as the mechanism for the PCE-induced mouse liver tumors other than to suggest that PCE induces other adverse effects in the liver after acute term exposure. This manuscript provides a MOA analysis for PCE-induced rodent liver tumors including a review of the scientific evidence supporting and against the involvement of PPARα activation in the PCE-induced mouse liver tumors.
The metabolism of PCE has been reviewed by Chiu and Ginsberg (2011), Chiu et al. (2007), and Lash and Parker (2001) as well as in several regulatory documents (ATSDR, 2019; USEPA, 2010, 2020). PCE appears to be metabolized through two pathways: oxidation by cytochrome P-450 (predominantly via Cyp2E1) and glutathione conjugation via glutathione S-transferase. PCE metabolism is similar in humans and rodents (Lash and Parker, 2001). However, the predominant pathway of metabolism and the relative amount of metabolism has been shown to vary by dose, exposure pathway and species. The primary route of PCE metabolism in the liver in rodents and in humans involves CYP450 oxidation in the liver (specifically mediated by CYP2E1) that results in the formation of trichloroacetic acid (TCA) as the major metabolite (Lash and Parker, 2001). Oxidative metabolism is postulated to occur in the liver, lung, and kidney (Chiu and Ginsberg, 2011). CYP2E1 is the major isozyme involved in oxidative metabolism in the liver (Lash and Parker, 2001). Oxidation of PCE appears to result in the formation trichloroacetyl chloride which is then hydrolyzed to TCA (Chiu and Ginsberg, 2011). The oxidative pathway appears to be the predominant pathway involved in liver toxicity and carcinogenesis in rodents. The glutathione conjugation pathway produces trichlorovinyl glutathione and S-trichlorovinyl-L-cysteine (TCVC) (Chiu and Ginsberg, 2011; Lash and Parker, 2001) and appears to be the reactive metabolites responsible for the observed renal toxicity and carcinogenicity in PCE-exposed rats (Chiu and Ginsberg, 2011; Lash and Parker, 2001).
Similar to PCE, TCA has been shown to induce liver tumors selectively in mice following chronic exposure (IARC, 2014). Given that TCA is the primary metabolite of PCE and that both compounds selectively induce liver tumors in mice, a common MOA would be likely for tumor induction. Therefore, a review of the pertinent studies for both PCE and TCA-induced rodent liver tumor studies is included in this manuscript. The MOA for TCA-induced mouse liver tumors has been previously discussed (Bull, 2000; Corton, 2008), and it was concluded based on a weight of evidence approach to involve PPARα activation (Corton, 2008).
MOA framework for carcinogens
Our knowledge of the carcinogenesis process has advanced considerably in the last half century allowing for better understanding of the fundamental biological events and pathways involved in taking a normal cell through multistage sequential changes to a neoplasm. Understanding these pathways of the chemical carcinogenesis process in experimental animals is important in the application of the scientifically based assessment of human risk. In concert with the development of the revised EPA’s 2005 Guidelines for Carcinogen Risk Assessment, the concept of a MOA framework was developed (USEPA, 2005). In these guidelines the MOA is defined as “a sequence of Key Events and processes, starting with interaction of an agent with a cell, proceeding through operational and anatomical changes, and resulting in cancer formation.” A “Key Event” is an observable necessary precursor in development of the tumor (Cohen et al., 2004). The MOA analysis begins with the identification of a specific neoplasm (usually in rodents) that is linked to exposure of a specific chemical. Based on the pharmacology, toxicology, and pathology of the chemical and the neoplasm induced, a hypothesized rodent cancer MOA using defined Key Events is developed that explains how exposure to the chemical leads to the observed neoplasm. Besides the Key Events, other important biological processes linked to chemical exposure may be observed. While not required for the cancer MOA these other endpoints can be used as an indicator or biomarker for a Key Event (Associate Events). In addition, modulating Factors, including external or internal (host) factors, can modulate the dose–response relationship of one or more of the Key Events, thereby changing the probability and/or magnitude of the end result (Cohen et al., 2004).
Liver carcinogen MOA framework
Liver cancer MOA involves a multistep process involving several biologically based Key Events that are expressed in a temporal manner. Therefore, no single Key Event is sufficient for the neoplasm to occur. The elimination of a Key Event in the multistep process will prevent the outcome (liver cancer). It is important to also note that the underlying biological process of each KEY Event exhibits dose response characteristics.
In performing a MOA analysis for rodent liver tumorigenesis, it is important to understand and appreciate the well-defined multistep, multistage dose-dependent process of liver tumor formation (Figure 1). The initial stage is the formation of a preneoplastic cell either spontaneously or through chemical treatment. A second stage occurs where subsequent selective cell proliferation of preneoplastic cells leads to the formation of preneoplastic focal lesions some of which can progress to neoplastic lesions (adenomas and subsequently carcinomas). The presence of spontaneous liver tumors in rodent livers (as demonstrated by the high incidence of liver tumors in untreated control animals) is important in understanding the susceptibility of rodents (in particular mice) to nongenotoxic carcinogens since spontaneous preneoplastic cells are already present in the liver of untreated rodents. While there is strain variability, most mouse strains have a high rate of spontaneous background liver tumors and hence have a higher level of preneoplastic cells that can be “promoted” by nongenotoxic compounds to develop neoplasms (Klaunig and Wang, 2018). Multistage Hepatic Carcinogenesis. Diagram of the multistage nature of liver carcinogenesis. Starting with a normal cell, a progression to a preneoplastic cells from either spontaneously acquired genetic modification or from chemical interaction. The preneoplastic cell may sit quiescent until intrinsic or extrinsic growth signals leading to a proliferation of the preneoplastic cell to a focal lesion. The focal lesion is still under endogenous growth control in the liver. However, further growth stimulus allows for the preneoplastic cells to acquire independent growth leading to a neoplasm. It is important to note that within each step of the progression from a normal cell to a neoplasm, control mechanisms are in place (DNA repair and apoptosis) to prevent or reverse this progression.
Possible modes of action for rodent liver carcinogens.
Chronic rodent studies with PCE
Incidence of liver tumors in mice exposed to PCE for two years.
aValues reflect the incidence (percentage) of animals with tumors.
Incidence of liver tumors in rats exposed to PCE for two years.
aValues reflect the incidence (percentage) of animals with tumors.
b4-5 days/week for 141 weeks.
A second chronic inhalation study with PCE was performed by the Japan Industrial Safety Association (JISA, 1993). In this study male and female F344DuCrj rats and male and female Crj:BDF1 mice were treated via inhalation at PCE exposure levels of 0, 50, 200, or 600 ppm for rats and 0, 10, 50, or 250 ppm for mice for a total of 104 weeks (6 hours/day, 5 days/week). In the treated rats, there was no increase in liver tumors in either sex at any of the concentration of PCE (Table 3). In mice, there was a dose-related increase in the incidence of hepatocellular adenoma in males: the incidences were 7/50 (0 ppm), 13/50 (10 ppm), 8/50 (50 ppm), and 26/50 (250 ppm), with the highest exposure level (250 ppm) being the only significant effect (Table 2). In female mice an increase in adenomas was seen at 50 ppm (7/49) and 250 ppm (16/49), the two highest exposure levels. For hepatic carcinoma incidence, a significant increase was seen only in males at the highest exposure of 250 ppm (25/50) compared to the controls. The authors reported that no hepatic carcinomas were found in female mice at 0, 10, or 50 ppm PCE treatment, but as with the male mice, there was a significant increase in hepatic carcinoma incidence at 250 ppm (14/29). The increase in incidence of hepatocellular adenomas and carcinomas at 250 ppm PCE was reported to be statistically significant in both sexes.
A third chronic PCE carcinogenicity study was performed by the NCI in 1977 (NCI, 1977). Male and female Osborne-Mendel rats and male and female B6C3F1 mice were treated with PCE in corn oil by oral gavage 5 days/week for 78 weeks, followed by an observation period of 32 weeks (rats) and 12 weeks (mice). Adjusted doses of PCE were calculated to be 471 and 941 mg/kg/day for male rats, 474 and 949 mg/kg/day for female rats, 536 and 1072 mg/kg/day for male mice, and 386 and 772 mg/kg/day for female mice. An increase in mortality related to compound-induced toxic nephropathy in the rats deemed the study to be inadequate for evaluation of carcinogenesis. In contrast, the survival in the mice was sufficient to allow for the evaluation of liver tumors. Only hepatic carcinomas were quantitated. The results showed a significant increase in hepatocellular carcinomas in the PCE-treated mice of both sexes. Incidences in the untreated control, vehicle control, low-dose, and high-dose groups were 2/17, 2/20, 32/49, and 27/48, respectively, in male mice, and 2/20, 0/20, 19/48, and 19/48, respectively, in female mice (Table 2). There are limitations in the NCI study including early death of the rats, control group numbers smaller that treated groups, potential confounding effect from trace amounts of epichlorohydrin present in the PCE test material as a stabilizer, and a nonstandard treatment protocol. While the NCI study is considered inadequate for assessing carcinogenicity, the overall results from this study agreed with the two inhalation studies.
The results of these three chronic studies taken together show that the liver in mice (both male and female) is a target organ of PCE carcinogenicity following chronic exposure, with the liver tumor incidence being dose responsive. The rat liver was refractory to the tumorigenic effects of PCE. In addition to the studies above, additional carcinogenicity studies by Rampy et al. (Rampy et al., 1978) and Maltoni and Cotti (Maltoni and Cotti, 1986) have shown a lack of a tumor response in the rat following chronic exposure to PCE (see Table 3). The negative liver tumor response in the rat is valuable in examining possible mechanisms for PCE-induced mouse liver tumors. Further discussion of the negative finding in the rat liver is below.
Chronic rodent studies with TCA
Hepatocellular tumors in rodents exposed to trichloroacetic acid (TCA) in drinking water studies.
bNR = not reported.
aValues reflect the incidence (percentage) of animals with tumors.
PPARα MOA for rodent liver tumors
PPARα rodent liver tumor MOA.
Proposed MOA for PCE-induced mouse liver tumors
Perchloroethylene (PCE) MOA for rodent liver tumors.
Key event 1: Metabolism of PCE to TCA
As noted above, PCE is metabolized in rodents and humans by two pathways: (1) a cytochrome P-450 (CYP450)-mediated oxidation pathway; and (2) a glutathione conjugation pathway (Chiu and Ginsberg, 2011, Chiu et al., 2007; Lash and Parker, 2001). In rodent liver the primary route of metabolism involves CYP450 (CYP2E) oxidation in the liver, with TCA being the primary metabolite. The CYP450 oxidation pathway, specifically involving CYP2E has been generally associated with PCE-induced liver toxicity, whereas the metabolites formed from further metabolism of the glutathione conjugate of PCE have been postulated to mediate PCE-induced kidney toxicity. While the linkage between PCE metabolism to TCA with subsequent neoplastic effects seen in the mouse liver appears to be directly dependent on TCA, questions have been raised whether the levels of TCA produced in the liver following PCE exposure are sufficient to produce the neoplastic effects observed in the mouse bioassays (Clewell et al., 2005). However, subsequent analysis of data from a TCA drinking water bioavailability study and TCA blood levels from a PCE mouse 5-day inhalation study using physiologically based pharmacokinetic (PBPK) models and dose–response modeling showed that TCA levels reported in PCE-exposed mice are sufficient to explain the incidence of liver tumors (Sweeney et al., 2009). Therefore, TCA levels after PCE inhalation exposure are similar to TCA concentrations directly associated with hepatic tumorigenicity. The reason for lack of a liver tumor response in rats compared to similarly treated mice has been addressed below. However, with regard to this Key Event one reason for the species difference may be linked to TCA blood levels after PCE exposure. An increase in blood TCA levels was dose-dependent in the PCE-exposed mice, but not in the rats was reported by Green (2003). Green (2003) measured TCA blood levels in both mice and rats following inhalation exposure to PCE. In this study, TCA blood levels were 8- to 9-times higher in mice than in rats after exposure to 200 ppm PCE 6 h/d for 5 days. The TCA blood levels in mice exposed to 10 ppm PCE were significantly higher than the TCA blood levels seen in 400 ppm-exposed rats. Blood TCA levels were dose-dependent in the PCE-exposed mice but were not seen in rats.
Key event 2: Activation of PPARα
The second key event in the PCE MOA for mouse liver tumors is the activation of PPARα. This step is a foundational Key Event for the PPARα MOA. Activation of the PPARα (not necessarily receptor binding agonism) results in multiple downstream effects including modification of lipid metabolism, increase in cell growth, peroxisome proliferation, and induction of CYP4a1 (Corton et al., 2018). The measurement of activation of the PPARα can be accomplished by measuring the receptor activation directly or indirectly or through the measurement of downstream surrogate targets, including induction of Cyp4a1 (a specific CYP induced by PPARα), increase in palmitoyl CoA oxidase (PCO) enzyme activity and/or protein, or peroxisome proliferation (via electron microscopy) (Corton et al., 2014; Klaunig et al., 2003;). These are Associated Events. In support of the proposed PCE MOA, multiple studies have demonstrated the activation of PPARα by PCE or its oxidative metabolite TCA (see below).
Activation of PPARα by PCE
Maloney and Waxman examined the activation of PPARα by PCE using COS-1 cells transiently transfected with human or mouse PPARα expression plasmids in a PPRE-luciferase reporter cloned from mouse and human tissues. PCE treatment at a relatively high concentration (5 mM) did not activate mouse or human PPARα in this in vitro system. In contrast, activation of PPARα was observed with TCA (Maloney and Waxman, 1999; Zhou and Waxman, 1998). These results support the premise that the metabolite of PCE (TCA) and not the parent compound is responsible for the PPARα activation. In vivo, PCE-treated mice showed an increase in the number of peroxisomes and PCO enzyme activity, both of which are indicative of activation of PPARα (Goldsworthy and Popp, 1987; Odum et al., 1988). In male B6C3F1 mice and Sprague-Dawley rats given oral gavage doses of 1000 mg/kg PCE for 10 days, peroxisome proliferation was seen in the mouse liver, but not in rat liver, as indicated by increased PCO activity (Goldsworthy and Popp, 1987). In a study by Odum et al. (1988), an increase in both the number of peroxisomes and PCO enzyme activity (almost 4-fold) was seen in the livers of male and female B6C3F1 mice following inhalation treatment with 200 or 400 ppm PCE for 28 days. In contrast, liver PCO activity was only slightly increased (1.3-fold) in similarly exposed male and female Sprague-Dawley rats. Zhou et al. (2017) reported a dose-dependent induction of peroxisomal fatty acid β-oxidation gene expression in the liver of male B6C3F1 mice 24 h after a single gavage dose of PCE (30 to 1000 mg/kg) that correlated the PCE dose and liver TCA levels (Zhou et al., 2017).
Activation of PPARα by TCA
As noted above, mouse and human PPARα were activated by millimolar concentrations of TCA (1-5 mM) in a transfection assay involving COS-1 cells in vitro (Maloney and Waxman, 1999; Zhou and Waxman, 1998). Similarly, TCA at a single high dose (4 mM) was also shown to activate mouse PPARα that had been transfected in human hepatocyte cell lines (Walgren et al., 2000). In the Walgren study, PPARα levels in six human liver tissues and also a long-term cultured human hepatocyte cell line was compared. In the human liver samples, PPARα levels showed significant variation of expression between the individual donors. The PPARα levels detected in the human liver samples were much lower than that seen in the mouse liver. In untreated in vitro cultured cell lines, human hepatocyte cell lines displayed PPARα levels slightly lower than similarly cultured mouse hepatocytes cell lines. In mouse liver cell lines transfected with human PPARα no significant PPARα activation was seen even after high-dose treatment with TCA. In contrast human hepatocytes transfected with mouse PPARα showed an increase in PPARα expression after TCA treatment. Interpretation of those results needs to be done with consideration of the models used. A caveat of trans-activation assays is that the systems used are artificial and don’t necessarily reconstruct the molecular environment found in vivo. While full-length receptors have been used (Maloney and Waxman, 1999; Zhou and Waxman, 1998), other receptor constructs are hybrids consisting of an estrogen receptor DNA binding domain linked to a PPARα ligand binding domain or a mutant full-length form of mouse PPARα with decreased background activity (Walgren et al., 2000). The assays also do not consider the f species-specific interactions between PPARα and the PP response elements (PPRE). This topic has been reviewed previously (Corton, 2008; Corton et al., 2000).
TCA has been demonstrated by multiple laboratories to activate PPARα as measured by peroxisome proliferation. PPARα-dependent endpoints including the induction of PCO enzyme activity, acyl-CoA oxidase (ACO) protein expression, and/or Cyp4a1 induction have been reported in the mouse liver following TCA treatment (DeAngelo et al., 1989; Elcombe, 1985; Goldsworthy and Popp, 1987). Laughter et al. (2004) showed an induction of Cyp4a1 and ACO protein expression and PCO activity in male wild-type mice after exposure to TCA in drinking water exposure for 7 days, while concurrently treated PPARα-null mice were negative for Cyp4a1 and ACO protein expression and PCO activity. DeAngelo et al. (1989) similarly showed an increase in ACO protein expression as well as an increase in peroxisome number and volume in the liver of male B6C3F1 mice treated with TCA (31 mM) for 14 days in drinking water. Parrish et al. (1996) reported a dose-related increase in ACO enzyme activity in male B6C3F1 mice treated with 0, 0.1, 0.5, or 2.0 g/L TCA in drinking water for 3 or 10 weeks. All dose groups at both times of treatment showed a significant increase in ACO activity. Elcombe (1985) reported a dose-dependent increase in PCO activity (4.8-fold) and peroxisome volume densities after 10 days of gavage dosing with 50 to 200 mg/kg TCA. PCO activity and relative liver weights were increased 285% in both male and female B6C3F1 mice dosed by oral gavage with 500 mg/kg TCA (Goldsworthy and Popp, 1987). PCO activity was also increased (352 to 1890%) in the livers of male B6C3F1 mice given 4.5 g/L TCA in the drinking water after 15, 30, 45, and 104 weeks of exposure (DeAngelo et al., 2008). In contrast to mice, DeAngelo et al. (1989) showed there was no increase in PCO activity in male Sprague-Dawley rats, and only a modest increase in male F344 and Osborne-Mendel rats when given 31 mM TCA in drinking water for 14 days. A slight increase PCO activity was also reported in rats following short-term gavage treatment of TCA (Elcombe, 1985; Goldsworthy and Popp, 1987).
Key event 3. Alteration in hepatic gene expression including cell growth pathways
As discussed above, activation of the PPARα in rodents results in an increase in hepatic gene expression that encompasses three major pathways: (1) lipid metabolism genes, (2) peroxisomal genes, and (3) growth regulatory genes (Corton, 2008; Corton et al., 2014; Klaunig et al., 2003). In the PPARα MOA for rodent liver carcinogenesis, two of these pathways—lipid metabolism genes and peroxisomal genes—are considered Associative Event endpoints while the modification of growth regulatory genes is a Key Event for tumor development. Therefore, Key Events 3 and 4, while separate in function, are linked by Associative Event endpoints.
PCE
In a study by Zhou et al. (2017), the liver effects of PCE were examined in male B6C3F1/J mice (3 mice per treatment group) with a single dose of PCE via oral gavage (0, 30, 100, 300, and 1000 mg/kg) and sampled 24 h after gavage treatment. In the liver, TCA was measured, and RNA sequencing was performed. At sampling TCA was detected in the liver of the PCE-treated mice in a dose-dependent manner. Dose-dependent effects were seen on the transcriptome that correlated between the PCE dose and the liver concentration of TCA. Pathway analysis on the liver showed that the pathways associated with PPARα signaling correlated with the dose of PCE. They also noted a dose-dependent induction of peroxisomal fatty acid-oxidation gene expression in the liver. This induction of peroxisomal fatty acid β-oxidation and associated pathways was among the most robust and dose-responsive effects for PCE. In addition, they noted that the dose-responsive differential gene expression in liver correlated between the PCE external dose and the internal TCA dose levels. Although the most significant changes seen related to PPARα signaling pathway analysis showed that mitochondria-related transcriptional pathways were also affected, in the liver in a dose–response manner. Changes in cell proliferation, apoptosis, and oxidative stress in the pathway analysis were not seen possibly due to the short acute treatment and sampling protocol. The authors noted that the transcriptional changes seen with acute treatment and sampling makes pathway analysis challenging, since the acute effects may not be related to chronic responses.
TCA
Ge et al. (2001) reported hypomethylation of the c-myc proto-oncogene after 4 days in the liver of female B6C3F1 mice treated with a single gavage dose of 500 mg/kg TCA. Concurrent with the activation of c-myc was an increase in cell proliferation in the liver. This correlated with previous reports by Tao et al. (2000a, 2000b) of TCA-induced hypomethylation of DNA and activation of c-myc in mouse liver. Tao’s group also showed that the addition of methionine (a methyl group donor) to the diet of the mice lessened the hypomethylation in TCA-induced methylation of c-myc (Tao et al., 2000b).
Key event 4. Increase in hepatocyte proliferation
PCE
Schumann et al. (1980) noted a dose-related increase in liver DNA synthesis in B6C3F1 mice, but not in Sprague-Dawley rats, given doses of 100, 250, 500 or 1000 mg/kg PCE by oral gavage for 11 days. Similarly, a dose-dependent increase in hepatic incorporation of [3H]thymidine in Swiss-Webster mice following 7 days of oral gavage dosing with 150, 500 or 1000 mg/kg PCE was reported by Philip et al. (2007). The level of [3H]-thymidine in the liver remained elevated in the 500 and 1000 mg/kg groups after 14 days of daily dosing but returned to control levels after 30 days of dosing. The mice receiving the two highest doses of PCE (500 and 1000 mg/kg) exhibited significantly greater number of hepatocytes in the S-phase after 7 and 14 days of dosing compared to controls.
TCA
Liver cell proliferation has been reported in mice following TCA treatment in drinking water. In a study by Sanchez and Bull (1990) male B6C3F1 mice and male and female Swiss-Webster mice were given drinking water containing 0, 0.3, 1, or 2 g/L TCA for up to 14 days. [3H]thymidine was used to measure DNA synthesis. A dose of 2 g/L TCA on day 5 and 14 significantly increased liver DNA synthesis in all of the treated mice. Dees and Travis (1994) showed a similar dose-dependent increase in hepatocyte [3H]thymidine labeling in male and female B6C3F1 mice following oral gavage doses of 0, 100, 250, 500, or 1000 mg/kg TCA for 11 days. In a longer-term treatment study, Stauber and Bull (1997) gave male B6C3F1 mice 2 g/L TCA (in drinking water) for up to 52 weeks. Mice sampled early in the treatment period at days 14 and 28 showed a significant increase in hepatocyte division rates. After 52 weeks of treatment, rates of cell division in the normal hepatocytes of mice treated with TCA were significantly decreased relative to control mice while cell division within altered hepatic preneoplastic foci and tumors were significantly higher. These findings showed the selective effect of TCA on cell proliferation in preneoplastic foci during chronic exposure.
Key event 5. Selective clonal expansion of hepatic preneoplastic foci cells
Increased cell proliferation ultimately leads to selective clonal expansion of altered focal hepatocytes leading to tumors. The origin of hepatic preneoplastic cells has been attributed to two possibilities. The first and most evident are the spontaneous preneoplastic cells present in rodents (Felter et al.,. 2018; Holsapple et al., 2006; Klaunig and Wang, 2018). In particular, there is a large contingent of these cells in the mouse as evident by the presence of histologically detectable focal lesions and tumors in untreated control animals (Cohen and Arnold, 2011; Klaunig and Wang, 2018). It has been hypothesized that nongenotoxic compounds that induce cell proliferation (i.e., PCE and TCA) “selectively promote” already present spontaneous preneoplastic cells to focal lesions and eventually to tumors. An alternate hypothesis has also been proposed in which spontaneous preneoplastic cells are initially formed due to the pressure from cell proliferation induced by exposure to a nongenotoxic compound. Under these conditions mistakes can occur in the DNA, which are fixed permanently and inherited by the cells during DNA replication in the pluripotential cells, either under normal circumstances generating one daughter cell and a committed cell or when duplicating itself into two pluripotential cells (Cohen and Arnold, 2011; Greenfield et al., 1984). These preneoplastic cells (whether from the spontaneous population or induced through replication mistakes) can then be selectively promoted by further cell proliferation pressure to form focal lesions (Klaunig and Wang, 2018) (Figure 1).
PCE
No studies examining preneoplastic liver foci growth or tumor promotion were located for PCE.
TCA
In support of Key Event 5, Stauber et al. (1998) showed a selective increase in DNA synthesis in hepatic foci of TCA-treated mice (52 weeks treatment in drinking water). The Pereira laboratory, which published several articles using a two-stage initiation promotion model with methylnitrosourea (MNU) as the tumor initiator, showed that TCA both induced and increased preneoplastic foci in female B6C3F1 mice (Pereira, 1996; Pereira et al., 1997; Pereira and Phelps, 1996). In one study, the hepatic tumor promoting properties of TCA were examined in MNU-initiated mice. When MNU-initiated mice were given 20 mmol/L TCA in drinking water for 31 or 52 weeks, TCA promoted (increased) both the incidence and size of the preneoplastic foci compared to MNU-initiated mice with no TCA treatment. Preneoplastic lesions promoted by TCA appeared at 31 weeks of treatment and were basophilic in appearance; at the 52-week sampling time both preneoplastic lesions and hepatic adenomas were present in these mice. In mice where TCA treatment was stopped at 37 weeks and then continued with untreated drinking water until the 52-week sampling time, there was a decrease in number and incidences of hepatocellular neoplasms compared to the mice receiving TCA in the drinking water for the full 52 weeks. These findings support a tumor promotion mechanism for TCA whereby foci present at 37 weeks required continual TCA treatment to produce the neoplasms (Pereira and Phelps, 1996). In a second study, Pereira et al. (1997) MNU-initiated mice received TCA (6.0 or 25 mmol/L) in drinking water from 6 weeks of age until sampling at 50 weeks of age). There was a linear increase in proliferative lesions (preneoplastic foci and hepatocellular adenomas) with TCA drinking water concentrations. The authors noted that these lesions were predominantly basophilic in appearance and negative for glutathione S-transferase-pi (GST-pi).
Key event 6. Formation of hepatic neoplasms
Both PCE and its oxidative metabolite TCA produce hepatic neoplasms in mice following chronic treatment. The three chronic studies performed with PCE (Table 2) showed a dose-dependent increase in liver tumors in mice following treatment. In contrast to the mice, PCE in these studies showed the rat to be refractory for liver tumors using the same dosing protocols (Table 3). Multiple studies have also examined rodent liver carcinogenicity following TCA administered in the drinking water (Table 4). Similar to that seen for PCE, TCA has been consistently shown the mouse to be the specific sensitive species for liver tumor induction.
Summary
Concordance of the induction of liver tumors and PPARα activation by PCE and TCA.
Comments: (+) the chemical was reported to be positive for the endpoint.
(−) the chemical was reported to be negative for the endpoint.
(+/
References: 1DeAngelo et al., 2008; 2Bull et al., 1990; 3Bull et al., 2002; 4Herren-Freund et al., 1987; 5Goldsworthy and Popp, 1987; 6DeAngelo et al., 1989; 7Laughter et al., 2004; 8Elcombe, 1985; 9Pereira, 1996; 10Parrish et al., 1996; 11DeAngelo et al., 1997; 12NTP, 1986; 13NCI, 1977; 14JISA, 1993; 15Odum et al., 1988; 16Zhou et al., 2017.
References for data supporting key events (KE) in mouse liver for TCA and PCE carcinogenicity.
References: 1Sweeney et al., 2009; 2Maloney and Waxman, 1999; 3DeAngelo et al., 1989; 4Elcombe, 1985; 5Goldsworthy and Popp, 1987; 6Laughter et al., 2004; 7Parrish et al., 1996; 8DeAngelo et al., 2008; 9Odum et al., 1988; 10Zhou et al., 2017; 11Dees and Travis, 1994; 12Stauber et al., 1998; 13Ge et al., 2001; 14Sanchez and Bull, 1990; 15Tao et al., 2000a, 16Tao et al., 2000b; 17Schumann et al., 1980; 18Philip et al., 2007; 19Herren-Freund et al., 1987; 20Bull et al., 1990; 21Pereira and Phelps, 1996; 22Bull et al., 2002; 23Pereira, 1996; 24NCI, 1977; 25JISA, 1993; 26NTP, 1986.
Lack of rat liver tumor induction by TCA or PCE
In chronic studies examining the carcinogenicity of TCA or PCE in the liver, the rat, unlike the mouse, did not show an increase in liver neoplasms (NTP, 1986; JISA, 1993; NCI, 1977; Rampy et al., 1978; Maltoni and Cotti, 1986; DeAngelo et al., 1997) (Table 7). To further define this species difference in tumor induction several studies examined the induction of the PPARα-related PCO activity in the rat after treatment with TCA (Table 7). DeAngelo et al. (1989) reported PCO activity was increased 163% in male F344 rats following exposure to 31 mM TCA in drinking water. DeAngelo et al. (1989) reported that PCO activity was increased 228% (corn oil) and 179% (water) after 10 consecutive daily oral gavage doses of 200 mg/kg TCA in corn oil and water, respectively. A direct comparison of the level of TCA in drinking water versus oral gavage showed a negligible difference in PCO activity (163% vs 179%, respectively). In contrast. Using a different rat strain (Alderley Park Wistar-derived), Elcombe (1985) reported liver PCO activity increasing 650% from control values following 10 consecutive daily doses of 200 mg/kg TCA in corn oil. Goldsworthy and Popp (1987) reported an increase of 284% PCO activity in male F344 rats following 10 consecutive daily doses of 500 mg/kg TCA in corn oil. DeAngelo et al. (1989) noted that corn oil alone can increase PCO activity. In a report by Green (2003) TCA blood levels were measured in mice and rats exposed by inhalation to PCE. The results showed that TCA blood levels were approximately 8- to 9-times higher in mice than in rats following inhalation exposure to 200 ppm PCE 6 h/d for 5 consecutive days. Mice exposed to 10 ppm PCE had higher TCA blood levels than in the 400 ppm-exposed rats (combined sexes: 17.52 mg/mL vs 11.17 mg/mL, respectively). The increase in blood TCA levels was dose-dependent in the PCE-exposed mice, but not in the rats. While at first glance these data suggest a disconnect between activation of PPARα (measured by PCO) and liver tumor formation in the rat, a review of the proposed MOA (Table 6) shows that activation of the of PPARα is only the first step (Key Event (1) in MOA. For the third Key Event, the induction of for tumor formation in the mouse liver additional steps are needed including induction of cell proliferation in the liver and selective clonal expansion of preneoplastic hepatocytes, neither of which has been demonstrated experimentally for TCA or PCE in rats. Therefore, the difference in liver tumor response between rats and mice can be explained by (1) the difference in metabolism of PCE to TCA (Key Event 1), (2) the lack of induction of cell proliferation by PCE and TCA in the rat (Key Event 3) and (3) the selective clonal expansion of focal preneoplastic cells by TCA in the mouse but not the rat (Key Event 5).
Human relevance
There is a preponderance of experimental and epidemiological evidence that have shown compounds that function through the PPARα MOA for rodent liver tumors are not human liver carcinogens (Corton et al., 2018). Several reviews have addressed this issue and have successfully rebutted concerns about both the PPARα MOA in chemically induced liver cancer in rodents and the human relevance of this MOA (Corton et al., 2014, 2018; Felter et al., 2018; Klaunig et al., 2003). The PPARα MOA has been generally accepted by the liver carcinogenesis scientific community to not be of human relevance. Extensive reviews on the mechanisms supporting the PPARα MOA have shown that in the rodent (rat and mice), a three-tier response is seen following activation of PPARα: (1) peroxisome proliferation, (2) cell growth modification (cell proliferation), and (3) lipid metabolism gene expression. In humans, only one tier of the three has been demonstrated—the lipid metabolism gene expression which accounts for the hypolipidemic effects of PPARα drugs. Modification of cell growth, required for tumor growth, is not seen in humans with PPARα activation (Corton et al., 2014, 2018; Felter et al., 2018; Klaunig et al., 2003). In further support, there have been several large retrospective epidemiological studies that have examined the chronic treatment with the PPARα-activating hypolipidemic drugs gemfibrozil and clofibrate (reviewed in Klaunig et al., 2003; Corton et al., 2018). These studies have shown no elevated risk of mortality from liver cancer associated with over a decade of chronic use of these hypolipidemic pharmaceuticals. The seminal reviews by Corton et al. (2014, 2018) provides a thorough assessment of the support for the PPARα MOA and its application to human liver cancer risk. However, several groups (Guyton et al., 2009; Lai, 2004; USEPA, 2012; USEPA, 2020) have questioned whether alternate modes of action (not PPARα) may be responsible for both the rodent liver tumors as well as potential human liver effects.
Alternative modes of action for PCE-induced liver tumors
Although the evidence strongly supports a PPARα MOA for the observed mouse liver tumors following chronic exposure to PCE, other possible modes of action need to be considered. Three other modes of action in the liver (see Table 1) are considered below including DNA reactivity/genotoxicity, cytotoxicity, and other nuclear receptor activation.
DNA reactivity/genotoxicity
PCE has been evaluated in a number of genotoxicity studies and this literature was recently reviewed (ATSDR, 2019; USEPA, 2012; USEPA, 2020). In general, PCE has not shown direct DNA reactivity, but induced positive responses in selected in vitro test systems under certain treatment conditions. There are four studies that specifically examined the effects of PCE on the mouse liver DNA—two investigated DNA binding (Mazullo et al., 1987; Schumann et al., 1980), one studied DNA strand breaks using the comet assay (Cederberg et al., 2010) and one study examined micronucleus induction (Murakami and Horikawa, 1995). No DNA binding was observed in the study by Schumann et al. (1980) following a 6-hour inhalation exposure up to 600 ppm PCE (three-times the tumorigenic concentration) or a single oral gavage dose of 500 mg/kg (tumorigenic dose). Although Mazzullo et al. (1987) reported DNA binding in the mouse liver following intraperitoneal injection of 1.4 mg/kg PCE, these results were attributed to the likely contamination of the DNA samples by RNA (USEPA, 2012).
In the Cederberg et al. (2010) publication, a marginal increase in DNA strand breakage was reported following oral dosing of mice with 1000 or 2000 mg/kg PCE (2 doses, 24 h apart). However, as discussed by the EPA (USEPA, 2012), the interpretation of these results was questioned in a number of publications based on statistical and biological considerations. Murakami and Horikawa (1995) identified a small (<2-fold), but significant, increase in micronucleated hepatocytes following intraperitoneal dosing of mice with 1000 or 2000 mg/kg PCE. However, the authors did not provide any information on their laboratory’s historical negative control data in order to determine whether the values observed among their treated groups were within or outside of the historical range. This information is critical given the small increase in the micronucleus frequencies observed in the treated groups. In summary, the above studies do not provide convincing evidence for the genotoxicity of PCE in the livers of mice.
The PCE liver metabolites that result from glutathione conjugation: S-(1,2,2-trichlorovinyl) glutathione (TCVG), N-acetyl-S-(1,2,2-trichlorovinyl)-L-cysteine (NAcTCVC), S-(1,2,2-trichlorovinyl)-L-cysteine (TCVC) and TCVC sulfoxide (TCVCS) plus the oxidative metabolites (PCE-oxide and trichloroacetyl chloride), were reported to induce mutagenicity in the Ames bacterial reverse mutation assay. It is worth pointing that TCVG was mutagenic only in the presence of subcellular fractions from rat kidney, but not rat liver, due to the very low or non-detectable levels of γ-glutamyltransferase (GGT) in the liver fractions (Vamvakas et al., 1989). Without GGT present in liver cells, TCVG cannot be metabolized to TCVC, which is necessary for the formation of the reactive (and potentially mutagenic) metabolites. No data exists on the ability of mouse liver microsomes to activate TCVG to a bacterial mutagen. The primary issue with the extrapolation of this in vitro data on the metabolites is that it is not known whether PCE-treated mice do indeed generate adequate quantities of the above metabolites to elicit a mutagenic response in this tissue. In addition, as discussed previously, assays for DNA binding in the livers of mice treated with PCE did not provide definitive evidence for this initial key event in the proposed MOA, raising the possibility that the concentrations of these metabolites might not be high enough to reach the nuclear DNA and elicit a mutagenic response in the liver.
The primary metabolite of PCE in the liver, TCA, has shown little to no genotoxic activity in vitro (reviewed in USEPA, 2012). In bacteria, TCA did not produce a positive response when examined in several Salmonella reverse mutation assays except for a positive finding in a fluctuation assay using strain TA100. In a mouse lymphoma forward mutation assay, a weak positive response was observed in the presence of S9 only at concentrations exceeding the limit dose of 10 mM as recommended in the current OECD test guidelines for in vitro assays. Furthermore, the significance of this positive response needs to be interpreted with caution since addition of TCA can perturb the pH of the culture medium potentially leading to false-positive responses. Treatment of human lymphocytes in culture with TCA in the presence of S9 induced chromosomal aberrations; however, no such effect was observed when the cultures were treated with neutralized TCA. TCA did not induce DNA strand breaks upon treatment of primary or established cell cultures when tested in the absence of S9. Both positive and negative findings were reported for DNA strand breakage in the livers and cytogenetic damage in the bone marrows of mice exposed to TCA by intraperitoneal and/or oral gavage administrations. Cytogenetic damage was also reported following TCA treatment of chickens and newts and there was also a report on the induction sperm head abnormalities (not a bona fide genotoxicity endpoint) in mice intraperitoneally treated with TCA. In general, the intraperitoneal route of treatment is no longer considered relevant for hazard assessment since it is not a physiologically relevant route. Furthermore, many of the studies on TCA were conducted using nonstandard protocols or non-validated test systems and thus limiting the value of these studies in the overall assessment of the genotoxicity of TCA.
In conclusion, the available data do not provide support of the key events in a mutagenic MOA (i.e., DNA reactivity and mutagenicity in the tumor target tissue). While some of the liver metabolites of PCE are in vitro mutagens/genotoxicants, it is uncertain whether these metabolites are generated at adequate levels in the tumor target tissue to damage nuclear DNA. Although there are no studies that interrogated the induction of mutations in the livers of mice exposed to PCE or TCA, the available data for a mutagenic MOA for PCE is relatively weak and, as discussed above, a compelling case can be made for an alternate MOA that does not involve DNA reactivity/mutagenesis as an early key event for mouse liver tumors.
Other receptor-mediated MOAs
Activation of other nuclear receptors have been linked to the induction of rodent liver tumors (Table 1). A review of the literature failed to locate any information on other nuclear receptors (other than PPARα) being activated by TCA or PCE. In the acute treatment pathway analysis study by Zhou et al. (2017) examining PCE, the pathway analysis did not reveal any additional receptor participation while supporting the PPARα pathway as the predominant pathway. The genomic signature was entirely consistent with PPARα activation and events associated with it.
Cytotoxicity
The cytotoxicity MOA for the liver requires that the chemical induces chronic cytotoxicity, resulting in chronic necrosis and compensatory hyperplasia. The resulting necrosis is accompanied by inflammation and significantly high liver serum enzymes. This pattern was not observed for either PCE OR TCA in the reported chronic studies or in short-term in vivo treatment protocols. One laboratory suggested that the cell proliferative response seen in a 30-day oral gavage study with PCE was not due to PPARα activation but rather to a regenerative response following cytotoxicity (Philip et al., 2007). These investigators reported an increase in CYP4A expression (an indicator of peroxisome proliferation) but only at the highest dose studied (1000 mg/kg/day). However, a peroxisome proliferation response was not sustained over the 30-day treatment period. Although “regenerative” cell proliferation was reported in this study, the level of necrosis and liver enzymes seen were not sufficient to qualify as a cytotoxic MOA. These results may simply reflect a phenomenon that is specific to a short-term, high-dose oral gavage treatment of PCE in the SW mouse strain and not to changes leading to carcinogenesis. These observations have not been seen in other reports and have not been repeated. The sensitivity of the SW mouse used in this study to PCE hepatocarcinogenicity is unknown, thus limiting the significance of these findings. It is also important to note that in the chronic PCE and TCA carcinogenicity studies, there was no chronic necrosis reported with regenerative compensatory hyperplasia. The available data do not support the cytotoxic MOA.
Other possible MOA
Besides the examination of the three established rodent liver MOA noted above, other less specific and poorly investigated MOAs have been proposed for PCE-induced mouse liver tumors (USEPA, 2012). These alternate MOAs are speculative and tend to be based on limited information from a single laboratory involving single-dose or short-term exposures. The hypomethylation of DNA has been proposed as an alternative MOA (USEPA, 2012). DNA hypomethylation is a common component of liver carcinogenesis that is not unique to any particular compound treatment and occurs “spontaneously” during the stages of tumorigenesis (Counts and Goodman, 1994). In the case of the PPARα MOA, hypomethylation has been reported as a consequence of increased gene expression seen with PPARα activation (Pogribny et al., 2008). These investigators compared two well established PPARα activating compounds known to induce liver tumors in rodents, WY-14,643 [4-chloro-6-(2,3-xylidino)-pyrimidynylthioacetic acid] and DEHP [di-(2-ethylhexyl)phthalate] for DNA hypomethylation. After exposure of F344 male rats to carcinogenic doses of WY-14,643 (0.1% (w/w) or DEHP (1.2% (w/w) for 5 months, liver DNA hypomethylation was measured. DNA hypomethylation was increased 2-fold in WY-14,643-treated rats compared to control rat liver, whereas there was no change in DNA methylation in the liver from DEHP-treated rats. The livers from WY-14,643-treated rats in the livers also had an increase in the accumulation of DNA single-strand breaks, increased cell proliferation, and diminished expression of DNA methyltransferase. In contrast, none of these endpoints were changed in the liver from the DEHP-treated rats. These results confirm that while changes in DNA methylation can be evident during nongenotoxic hepatic carcinogenesis in rodents, it is not, however, a required or Key Event. For liver tumor MOA such as PPARα activation, it should thus be considered an Associative Event and/or modifying factor.
Zhou et al. (2017) examined liver transcriptomic responses in B6C3F1 mice given a single oral gavage dose of PCE (30, 100, 300, and 1000 mg/kg) There was a dose-related upregulation of several genes including ones involved in PPARα signaling, as well as gene expression changes related to ABC transporters, mitochondrial pathways, and nucleotide metabolism. Analysis revealed that pathways associated with peroxisome proliferator-activated receptor (PPAR) signaling and oxidation/reduction were positively correlated with the administered dose in the liver. Positive dose–response transcriptional effects of PCE were largely related to fatty acid metabolism and membrane-associated transport (primarily ABC transporter family). Effects on the mitochondria and nucleotide metabolism pathways were negatively correlated with the administered dose of PCE. This study has led to some speculation that, in addition to PPARα activation, there may be multiple MOA involved in PCE-induced mouse liver cancer (USEPA, 2020). While this elegant and comprehensive study provides some insight to pathways involved in PCE treatment in mice several concerns are raised regarding application of these results to the carcinogenicity of PCE in mouse liver. One concern is the lack of clinical and histologic pathways to accompany the treatment protocols. No report on liver serum enzymes was presented and thus it is unclear if cytotoxic mechanisms may be involved at the doses used. In addition, caution should be made in taking the results from a single gavage treatment with a one-day exposure time and applying these results to multistage chronic carcinogenesis. Finally, the increase expression of multiple pathways in a transcriptomic study is not surprising and while the pathways noted above are involved in cellular functions and are frequently seen for other compounds, a specific MOA for these pathways has not been proposed with any supporting evidence.
Data gaps
The support of a PPARa MOA for PCE-induced rodent liver tumors is strong. As noted above, two additional studies involving PCE in vivo treatment to mice are lacking. A demonstration of preneoplastic foci formation and clonal expansion (Key Event 5) has not been reported. Similarly, studies examining the PPAR activation and downstream biological responses using PPARa-null mice have not been performed. While these studies have not been performed with PCE in vivo, studies with TCA on these two issues have been reported. The induction and expansion of preneoplastic foci in mice (tumor promotion) (Key event 5), has been reported by two laboratories (Stauber et al., 1998; Pereira, 1996; Pereira and Phelps, 1996; Pereira et al.,. 1997). In addition, Laughter et al. (2004) using PPARα-null mice showed a lack of an induction Cyp4a1 and ACO protein expression and PCO activity compared to male wild-type mice after seven-day treatment with TCA in drinking water. Given the essential first step involving metabolism of PCE to TCA these studies provide the needed experimental support for the proposed PCE mode of action (Table 6). Therefore, the suggested two data gaps with in vivo PCE treatment are negligible and offset by the strong experimental support for the TCA PPAR activation MOA in the mouse liver.
Overall conclusions
PCE is a mouse liver tumorigen that functions through a PPARα MOA. TCA, the oxidative metabolite of PCE, appears to be the metabolite responsible for the observed PPARα activation. The data supporting this MOA are extensive, show dose–response characteristics, are scientifically valid, and have been generated independently in a number of laboratories. The Key Events for the PPARα MOA have been studied extensively and demonstrated for a number of chemicals by multiple laboratories. The use of null PPARα mice has supported the essential role of this receptor as a major Key event (Corton et al., 2018). With regard to human relevance the fact that a Key Event of the PPARα MOA is the induction of cell growth pathways and eventual selective cell proliferation of preneoplastic lesions in the rodent liver, both of which do not occur in human liver following treatment with PPARα compounds. While the data supporting these conclusions appear strong, others have suggested that other potential mechanisms not involving PPARα activation in liver cancer induction cannot be ruled out (Guyton et al., 2014; Lai, 2004; USEPA, 2012; USEPA, 2020). However, no scientifically acceptable alternative MOA has been proposed to refute the PPARα MOA for liver carcinogens that activate this receptor. Epidemiological evidence in humans following chronic treatment with PPARα active hypolipidemic drugs show a lack of liver tumors. A considerable number of epidemiology studies have also been conducted on PCE-exposed workers and liver cancer (USEPA, 2012; USEPA, 2020). Yet, as stated by EPA in its conclusion of the cancer epidemiology in the recent TSCA Risk Evaluation on PCE, there is “little or no support for associations with kidney, esophagus, or liver cancer” (USEPA, 2020). The PPARα MOA for the mouse liver tumor induction by PCE is very strong. The lack of an association between liver cancer and PCE exposure in humans would be expected from the PPARa MOA for the PCE-induced mouse liver tumors, which is not considered to be relevant to humans by many. Therefore, given that the PPARα MOA has been accepted not to be of human relevance based on significant scientific studies, and the lack of an increase in human liver cancer from PCE exposure based on epidemiology, PCE does not represent a human liver cancer risk.
Footnotes
Acknowledgments
Financial support for this work, including the compensation paid to Drs. Klaunig and Gollapudi, was provided by the Halogenated Solvents Industry Alliance, Inc. (HSIA) and the American Petroleum Institute (API).
Declaration of competing interest
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr Bevan is the Science Director of HSIA, a trade association that represents producers and users of PCE.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Halogenated Solvents Industry Alliance, Inc. (HSIA), and American Petroleum Institute (API).
Data availability statement
This is a review manuscript, and all data is openly available in a public repository and openly available.