APPENDIX E
Peroxisome Proliferators and Liver Cancer

BACKGROUND

Peroxisomes, subcellular organelles found in the cytoplasm of mammalian and other cells, have important metabolic functions, in particular fatty acid oxidation. Several groups in the mid-1960s first reported the phenomenon of peroxisome proliferation when significant increases of hepatic peroxisomes were seen in response to administration of the hypolipidemic drug clofibrate to rats (de Duve 196). In addition to peroxisome proliferation and hepatomegaly, peroxisomal fatty acid oxidation is induced, and long term administration of this drug (and similar compounds) causes hepatocarcinogenesis (Reddy et al. 1976). A number of compounds were later identified that share the morphologic and biochemical response of clofibrate and were deemed “peroxisome proliferators” (see Table E-1). Peroxisome proliferators are a diverse group of chemicals that include the fibrate class of hypolipidemic drugs (e.g., clofibrate, ciprofibrate), Wy14,643 (often used as the prototypical peroxisome proliferator), commercially used plasticizers (phthalates, perfluorinated fatty acids), and endogenous fatty acids (see Table E-2).

Peroxisome proliferator-activated receptors (PPARs) are members of the nuclear receptor superfamily of transcription factors. They respond to specific factor ligands by altering gene expression in a cell-, developmental-, and sex-specific manner. Three subtypes of PPAR are expressed in different tissues, PPARα, PPARβ (also called PPARδ and NUC1), and PPARγ. Each receptor has the possibility of responding to different ligands, altering the



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 418
Assessing the Human Health Risks of Trichloroethylene: Key Scientific Issues APPENDIX E Peroxisome Proliferators and Liver Cancer BACKGROUND Peroxisomes, subcellular organelles found in the cytoplasm of mammalian and other cells, have important metabolic functions, in particular fatty acid oxidation. Several groups in the mid-1960s first reported the phenomenon of peroxisome proliferation when significant increases of hepatic peroxisomes were seen in response to administration of the hypolipidemic drug clofibrate to rats (de Duve 196). In addition to peroxisome proliferation and hepatomegaly, peroxisomal fatty acid oxidation is induced, and long term administration of this drug (and similar compounds) causes hepatocarcinogenesis (Reddy et al. 1976). A number of compounds were later identified that share the morphologic and biochemical response of clofibrate and were deemed “peroxisome proliferators” (see Table E-1). Peroxisome proliferators are a diverse group of chemicals that include the fibrate class of hypolipidemic drugs (e.g., clofibrate, ciprofibrate), Wy14,643 (often used as the prototypical peroxisome proliferator), commercially used plasticizers (phthalates, perfluorinated fatty acids), and endogenous fatty acids (see Table E-2). Peroxisome proliferator-activated receptors (PPARs) are members of the nuclear receptor superfamily of transcription factors. They respond to specific factor ligands by altering gene expression in a cell-, developmental-, and sex-specific manner. Three subtypes of PPAR are expressed in different tissues, PPARα, PPARβ (also called PPARδ and NUC1), and PPARγ. Each receptor has the possibility of responding to different ligands, altering the

OCR for page 418
Assessing the Human Health Risks of Trichloroethylene: Key Scientific Issues TABLE E-1 Characteristic Effects of Peroxisome Proliferator Compounds Morphologic changes Hepatomegaly Increase in number and size of peroxisomes Hepatocarcinogenesis (tumor promotion) Biochemical changes Decrease in serum lipids, triacylglycerols, and cholesterol Peroxisomal protein induction (acyl-CoA oxidase, bi/trifunctional protein, carnitine acetyltransferase) Mitochondrial protein induction (acyl-CoA dehydrogenase, carnitine palmitoyl transferase) Microsomal protein induction (CYP4A) Cytosol protein induction (acyl-CoA hydrolase, malic enzyme, fatty acid binding protein) Other characteristics of biochemical changes Species differences: (rat, mouse > hamster, guinea pig > rabbit, dog, monkey) Sex differences: (male > female) Target organ specificity (liver > kidney, heart, small intestine > other organs) TABLE E-2 Representative Peroxisome Proliferators Commercial Category Compound Hypolipidemic drug (approved in U.S. and elsewhere) Gemfibrozil (U.S.) Clofibrate (U.S. and others) Ciprofibrate (France) Fenofibrate (Other countries, not U.S.) Hypolipidemic drug (not approved) Wy-14,643 Nafenopin BR-931 Methylclofenepate Herbicide Lactofen Fomasafen 2,4-Dichlorophenoxyacetic acid 2,4,5-Trichlorophenoxyacetic acid Plasticizers and polymerizers Di-(2-ethylhexyl)phthalate Di-(2-ethylhexyl)adipate Di-n-butyl phthalate Perfluorooctanoic acid Perfluorooctanesulfonate Solvents Trichloroethylene Perchloroethylene Miscellaneous pharmaceutical Valproic acid (antimania, approved U.S. and elsewhere) LY-171,883 (leukotriene D4 receptor antagonist [not approved]) Dehydroepiandrosterone (dietary supplement and human adrenal steroid, approved in U.S. and elsewhere)

OCR for page 418
Assessing the Human Health Risks of Trichloroethylene: Key Scientific Issues expression of different target genes, and playing different biologic roles (Vanden Heuvel 1999a). The biologic effects of PPARs are mediated by altered gene expression. Peroxisome proliferator response elements, consisting of imperfect direct repeats of the sequence TGACCT spaced by a single base pair, have been identified in the upstream regulatory sequences of several PPAR target genes. The ligand 9-cis-retinoic acid enhances PPAR action by activating the retinoid X receptor, which forms a heterodimer with PPAR and binds to the peroxisome proliferator response elements to induce gene transcription. Of the three PPAR subtypes, PPARα is the key receptor for peroxisome proliferators. A prototypical marker of peroxisome proliferator action is induction of the peroxisomal enzyme acyl-CoA oxidase, which is elevated about 10-fold in the livers of treated rodents. Additional peroxisome-proliferator-responsive genes include other peroxisomal beta-oxidation enzymes and members of the cytochrome P-450 IVA family. PPARα activation mediates pleiotropic effects such as stimulation of lipid oxidation, alteration in lipoprotein metabolism, and inhibition of vascular inflammation. PPARα activators increase hepatic uptake and esterification of free fatty acids by stimulating the fatty acid transport protein and acyl-CoA synthetase expression. The carcinogenic effects of PPARα ligands are thought to be associated with altered cell proliferation resulting from the regulation of growth regulatory genes (IAS 2005). Since the mid-1970s, it has been demonstrated that peroxisome proliferators induce hepatomegaly and hepatocarcinogenicity in rodents (Moody and Reddy 1978; Reddy and Lalwai 1983; Klaunig et al. 2003). Although it has been hypothesized that peroxisome proliferation is causally linked to peroxisome proliferator-induced liver cancer, the direct link between peroxisome proliferation and liver cancer is uncertain. As described below, the two conclusively demonstrated causally linked mechanistic changes of PPARα agonist-induced hepatocellular carcinogenesis in rodents are increased activation of PPARα and PPARα-dependent hepatocyte proliferation. It is well accepted that these agents do not act by a genotoxic (DNA reactive) process. The role of PPARα in the hepatocarcinogenicity of peroxisome proliferators is clearly demonstrated by the fact that these chemicals do not induce hepatomegaly and hepatocarcinogenicity in PPARα null mice (Klaunig et al. 2003). In rodents, peroxisome proliferators lead to tumors that are histologically adenomas or carcinomas that are characterized as basophilic and have absence of γ-glutamyl transpeptidase expression (Kraupp-Grasl et al. 1990; Grasl-Kraupp et al. 1993). These findings suggest that the development of liver tumors by peroxisome proliferators involves amplification of a specific subtype of altered hepatic foci. Some studies have suggested that peroxisome proliferators and PPARα activation might result in tumors in extrahepatic organs. For example, certain peroxisome proliferators increase

OCR for page 418
Assessing the Human Health Risks of Trichloroethylene: Key Scientific Issues Leydig cell and pancreatic ademomas and carcinomas (Klaunig et al. 2003). However, whether these are PPARα dependent is unknown. Marked species differences are observed in response to peroxisome proliferators in terms of peroxisome proliferation and hepatocarcinogenesis. Rats and mice are the most sensitive, and hamsters show an intermediate response, whereas guinea pigs, monkeys, and humans appear to be relatively insensitive or nonresponsive at doses that produce a marked response in rodents (Lai 2004). Nonhuman primates and humans appear to be resistant to the induction of peroxisome proliferation and the development of liver cancer by PPARα agonist drugs (fibrates). To examine the mechanism determining species differences in peroxisome proliferator response between mice and humans, a PPARα-humanized mouse line was generated. The PPARαhumanized and wild-type mice responded to treatment with peroxisome proliferators, as shown by the induction of genes encoding peroxisomal and mitochondrial fatty-acid-metabolizing enzymes and a resultant decrease of serum triglyceride concentrations. However, only the wild-type mice (and not the PPARα-humanized mice) exhibited hepatocellular proliferation (IAS 2005). Thus far, epidemiologic studies on peroxisome proliferators have shown little evidence of carcinogenic effects in humans (Lai 2004). APPLICABILITY TO HUMAN HEALTH RISK ASSESSMENT There is considerable debate about the mechanisms by which peroxisome proliferators cause liver tumors in rodent models and whether these chemicals represent a human cancer risk. It has been well-established that human liver and hepatocytes in culture are less sensitive to the peroxisome proliferation effects of these chemicals; however, as stated above, this event is only associative and not causal to the development of tumors. Two clinical trials also have examined relative mortality and cancer rates in human males treated with fibrates. In the Helsinki Heart Study, 4,081 men aged 40-55 years with elevated serum cholesterol were treated with either gemfibrozil or placebo for 5 years (Frick et al 1987; Huttunen et al. 1994). Despite significant lowering of serum lipids, which prevented coronary heart disease in the gemfibrozil-treated group, no differences in total death rate or liver cancer incidence were observed between the groups. Liver cancer incidence was not reported as a separate end point; the incidence was reported either as total deaths from cancer or as deaths from liver, gallbladder, and intestinal cancers combined. Except for a borderline statistically significant difference (P = 0.062) in the incidence of basal cell carcinomas of the skin between 2,051 patients treated with gemfibrozil and 2,030 subjects receiving a placebo, no differences were found for other cancers. The incidence of cancer mortality in this study, for placebo and

OCR for page 418
Assessing the Human Health Risks of Trichloroethylene: Key Scientific Issues fibrate-treated patients, was less than 2% for each group, compared with virtually 100% in PPARα agonist-treated rodents (see above). The other randomized clinical trial was conducted by the World Health Organization to determine whether clofibrate would lower the incidence of ischemic heart disease in men. A group of 15,745 men were treated with clofibrate and two control groups (one with high cholesterol and one with low cholesterol) of about 5,000 men each were followed for an average of 5.3 years. Follow-up reports were provided 4.3 and 7.9 years after this period. Clofibrate was reported to cause a statistically significantly higher age-adjusted total mortality compared with the high cholesterol placebotreated control groups in this study. The excess mortality was due to a 25% increase in noncardiovascular causes—that is, diseases of the liver, gallbladder, pancreas, and intestines, including malignant neoplasms of these sites (Committee of Principal Investigators 1980). However, in the final follow-up study (5.3 years in the treatment phase with 7.9 years follow-up for a total of 13.2 years), neither the difference in the number nor the difference in the rate of cancer deaths between the clofibrate-treated group and the control groups was statistically significant (Committee of Principal Investigators 1984). The reason for the difference in mortality at the earlier time point is uncertain. Similar to the Helsinki Heart Study, no data on the incidence of liver cancer alone were provided. In this final follow-up study, there was a 12% excess of deaths from all causes other than ischemic heart disease compared with 25% in the previous studies. Furthermore, the proportional difference between the treated group and the control groups in the final follow-up study was diminished for malignant disease but increased for nonmalignant diseases. The results indicate that the excess in deaths from diseases other than ischemic heart disease was largely confined to the clofibrate-treatment period (average 5.3 years). However, 7.9 years posttreatment, there were 27 deaths associated with liver, gallbladder, and intestinal cancers in the clofibrate-treated group, compared with 18 and 11 deaths associated with the same end points in the high-cholesterol and low-cholesterol control groups, respectively. Similar to the Helsinki Heart Study, this incidence is less than 1% per group. Data concerning the human susceptibility to liver cancer from peroxisome proliferators come primarily from species comparisons of short-term responses, such as proliferation of peroxisomes in liver parenchymal cells, hepatomegaly, and the induction of various hepatic enzymes and of PPARα expression. Given the importance of PPARα in mediating the short- and long-term effects of PPARα agonist exposure in mice and rats, including liver cancer, cancer risk in humans has been gauged in part by comparing the properties of PPARα among susceptible rodent species and humans. Transient transfection studies using a human PPARα cDNA show that this receptor can transactivate reporter constructs, providing indirect evidence

OCR for page 418
Assessing the Human Health Risks of Trichloroethylene: Key Scientific Issues that the human PPARα is functional. Thus, it is not surprising that peroxisome proliferator response elements have been described in human genes that are transcriptionally regulated by PPARα, including human apo C-III, lipoprotein lipase, apo A-I, apo A-II, carnitine palmitoyltransferase-I, and acyl-CoA oxidase. Interestingly, large increases in the expression of marker mRNAs and proteins, including peroxisomal acyl-CoA oxidase, are not found in human and nonhuman primate hepatocytes treated with these chemicals in vitro. These observations are consistent with a recent study demonstrating the lack of an increase in acyl-CoA oxidase mRNA in human liver samples from 48 patients treated with one of several fibrates (bezafibrate, fenofibrate, or gemfibrozil), despite significant induction of hepatic apolipoprotein A-I mRNA and lowering of serum lipids after treatment. However, dose-dependent induction (<3-fold) of acyl-CoA oxidase activity has been observed in human hepatocytes treated with clofibrate and ciprofibrate, and treatment with perfluorodecanoic acid resulted in significant induction of peroxisomal density and increased acyl-CoA oxidase activity in human cells derived from glioblastoma. Human PPARα activation is reported to result in increased apolipoprotein A-II and lipoprotein lipase transcription and reduced apolipoprotein C-III, which is key to lowering serum triglycerides. Although data are available indicating that cultured human hepatocytes do not exhibit increased markers of cell proliferation in response to exposure to PPARα agonists while cultured rodent hepatocytes do, this model system might be inappropriate for evaluating this effect because primary hepatocytes do not proliferate substantially in vitro compared with hepatocytes in vivo. Combined, the preceding observations demonstrate that there is some overlap in target gene activation between humans and rodents and that further characterization is required to determine the reasons for the differences. Several possible explanations have been postulated to account for this disparity. There are dramatic differences in the expression levels or function of the expressed protein. Guinea pig liver also was reported to contain significantly less PPARα than mouse liver. However, given the relatively small number of human liver samples examined, further quantification of PPARα mRNA and protein in addition to other transcription factors to serve as good positive controls would be important. More recently, data from the humanized PPARα mouse suggest that inherent differences in receptor activation might contribute to the observed species difference in hepatocarcinogenicity, rather than receptor expression level. It appears that lower levels of PPARα expression might, in part, contribute to the species differences between rodents and humans, but expression of truncated or mutant PPARα, some through expression of alternatively spliced products, also has been described. No mutations or polymorphisms have been described to date in rodent species. A dominant-negative form of

OCR for page 418
Assessing the Human Health Risks of Trichloroethylene: Key Scientific Issues human PPARα has been described, and the presence of this protein could significantly inhibit PPARα activation and subsequent target gene modulation. Additionally, separate laboratories have described two different mutations (L162V and V227A) in human PPARα. The biologic significance of these mutant human isoforms of PPARα is unclear but has been linked to significant differences in serum apolipoproteins, serum cholesterol, and fibrate-induced changes in serum high density lipoprotein cholesterol. Some of the mutant PPARα proteins were shown to act as dominant-negative proteins in that they could prevent the normal PPARα from interacting with retinoid X receptor, binding to the peroxisome proliferator response element, and activating gene expression. These findings suggest that, in addition to reported lower expression of human PPARα, mutant variants, in particular a dominant-negative isoform, also might contribute to the apparent insensitivity in humans to PPARα agonists. However, because some humans respond to fibrate therapy, the hypothesis that altered PPARα protein accounts for the species differences is likely not true for all human cell types that express PPARα. The available epidemiologic and clinical studies are inconclusive but, nonetheless, do not provide strong evidence that PPARα agonists cause liver cancer in humans. Evidence from an in vivo model suggests that there could be considerably more similarities in PPARα target genes among humans and rodents. The weight of evidence suggests that this mode of action would be plausible in humans because they possess PPARα in sufficient quantities to mediate the human hypolipidemic response to therapeutic fibrate drugs. Thus, human PPARα is comparable to rat or mouse PPARα in its affinity for some, but not all, PPARα ligands, and evidence from the humanized PPARα mouse suggests that there are differences in target genes activated between human and rodent PPARα. However, a point in the rat and mouse key events cascade where the pathway is biologically precluded in humans in principle cannot be identified. Whereas the mode of action is plausible in humans, the weight of evidence suggests that this mode of action is not likely to occur in humans based on differences in several key steps when taking into consideration kinetic and dynamic factors. There is no convincing evidence for human tumorigenicity resulting from exposure to the PPARα agonists di(2-ethylhexyl)phthalate and various hypolipidemic fibrates. The studies of di-(2-ethylhexyl)phthalate are not considered adequate to demonstrate lack of tumor hazard but do constitute some evidence of absence. In contrast, there are extensive data for some of the hypolipidemic drugs, particularly clofibrate. The International Agency for Research on Cancer concluded that “the mechanism of liver carcinogenesis in clofibrate treated rats would not be operative in humans” based on the results of extensive epidemiologic studies (IARC 1996), particularly the World Health Organization trial on clofibrate comprising 208,000 man-years of observation (Committee of

OCR for page 418
Assessing the Human Health Risks of Trichloroethylene: Key Scientific Issues Principal Investigators 1980, 1984). Further, a meta-analysis of the results from six clinical trials on clofibrate found no excess cancer mortality. These human data are supported by evidence from nonhuman primate studies that show no evidence of tumors or focal lesions over 6-7 years of exposure. However, it is important to point out that many of the molecular and biochemical changes just described have not been evaluated in nonhuman primates with recently developed, high-affinity PPARα agonists. Most of the work to date has been performed and published using PPARα agonists with relatively low (di-[2-ethylhexyl]phthalate) or moderate (fibrate class of hypolipidemic drugs) affinity for activation of PPARα.

OCR for page 418
Assessing the Human Health Risks of Trichloroethylene: Key Scientific Issues This page intentionally left blank.