Fibrates, members of peroxisome proliferators and agonists of PPARα, are used to treat hyperlipidemia by reducing plasma triglyceride and cholesterol levels via accelerating lipid metabolism. In rodents, administration of fibrates can induce hepatomegaly and hepatocarcinoma, possibly due to the induction of cell proliferation and increased oxidative stress. Examination of PPARα-deficient mice demonstrated that the activation of PPARα is required exclusively for mediating both pharmacological (hypolipidemic) and toxic (carcinogenic) responses of fibrate administration [15
]. However, mechanisms of fibrate-induced hepatocarcinoma development and the potential risk of use of these drugs to humans remain unclear. Examination of gene expression profiles is an important approach that may help us better understand PPARα-mediated pleiotropic effects.
In this study, microarray analysis was applied to generate a molecular portrait of gene expression in mouse primary hepatocytes exposed to fibrates (Figure ). We treated mouse primary hepatocytes with three fibrates (bezafibrate, fenofibrate and WY-14,643) at multiple doses (0, 10, 30, and 100 μM). Although global gene analysis study was conducted in vitro [17
], the design of this study (i.e., treatment with multiple fibrates at multiple level doses) permitted us to detect whether changes in gene expressions are a class effect (e.g., genes are commonly regulated by multiple drugs) and whether changes are dose-dependent as well. Indeed, the majority of genes regulated by low and middle doses were also identified in high dose treatments. For example, 4/4, 9/9, and 25/31 genes that were regulated by 10 μM treatments of bezafibrate, fenofibrate and WY-14,643, respectively, were also found to be regulated in 100 μM treatments (Figure ). In addition, dose-response dependency in gene expressions was also observed for the genes commonly regulated at multiple doses (Figure ). The dose-dependent expression levels of genes altered by PPARα agonists allowed us to assess biological activity of this class of agents.
PPARα agonists have a therapeutic role in the management of fatty acid metabolism through their effects on β-oxidation and lipid transport. The gene expression changes in common across PPARα agonists may indicate those genes are directly regulated by PPARα stimulation. In this study, we demonstrated that 121 DEGs were altered in common by at least two of the three PPARα agonists tested. The Ingenuity Pathway Analysis was used to analyze gene functions and to provide pathway annotations. Based on this analysis, many of these genes (49 genes) are involved in the oxidation of fatty acids (Table ) as has been previously shown for this class of agents [4
]. For example, acyl-coA synthetase catalyzes the precursor step to β-oxidation (ligates CoA to a free fatty acid) and three members of the long chain acyl CoA synthestase family (Acsl1, Acsl4, and Acsl5) were increased. This observation is supported by the work of Schoonjans et al. who demonstrated that the expression of Acs is altered by fibrates and that there is a PPRE in the Acsl promoter [19
]. These findings also agree with those of Cornwall et al. who reported that the expression of Acsl was elevated in the liver of rats exposed to fenofibrate [20
]. The effects on the β-oxidation pathway also include the induction of the first enzyme of peroxisomal β-oxidation, acyl-CoA oxidase (Acox), as well as the next enzyme in the cascade, enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase (Ehhadh). The identification of a large number of lipid metabolizing genes following exposure to several PPARα agonists is in concordance with the known biochemical and molecular effects of these hypolipidemic agents to regulate lipid metabolism.
PPARα agonists are also considered to be nongenotoxic carcinogens in rodents. Oxidative stress has been proposed as a common pathway for many non-genotoxic carcinogens [21
]. In the present study, 10 genes involved in fatty acid β-oxidation were up-regulated upon exposure to PPARα agonists (Table ), which included Acox, the key enzyme of peroxisomal fatty acid β-oxidation system. The elevation of peroxisomal fatty acid β-oxidation such as occurs with PPARα agonist exposure in rodents results in the elevated generation of hydrogen peroxide [9
]. Substantial production of hydrogen peroxide causes oxidative stress and the induction of ROS. The increased ROS associated with elevated levels of Acox has been postulated to mediate the hepatocarcinogenesis resulting from PPARα exposure in rodents. We observed four genes (Aldh3a2, Apoc2, Cd36, and Slc25a10) associated with the production of ROS were up-regulated. A growing body of evidence indicates that Cd36 (CD36 antigen) is involved in the cytotoxicity associated with inflammation and is an essential mediator of the production of ROS [22
]. In addition, six genes classified as related to hepatic disorders were identified as being up-regulated (Table ). These observations support the hypothesis that increased peroxisome proliferation results in oxidative stress, which may be due to the disproportionate increase in the level of oxidation versus antioxidation enzyme activities [8
It is believed that the activation of PPARα and ensuing cascade effects are linked to both pharmacological and tumorigenic effects of PPARα agnoists [8
]. The carcinogenic response seems likely to be associated with both the induction of oxidative stress and the increased cell proliferation from peroxisome proliferation after treatment with these chemicals. In this study, we found that two genes (Acox1 and Pex11a) associated with cellular proliferation of peroxisomes were up-regulated about 3-fold. The level of peroxisomal biogenesis factor 11 (Pex11) correlates roughly with peroxisome abundance in the cell, and over-expression of Pex11 alone is sufficient to accelerate peroxisome division and to increase peroxisome abundance [23
]. It is thought that alteration in the balance between cell proliferation and apoptosis is causally related to the induction of liver tumors, and induced cell proliferation plays a key role in carcinogenesis in animals and humans [24
Based on the Ingenuity Pathway Analysis, 11 genes associated with liver cancer development were up-regulated by at least two PPARα agonists tested (Table ). For example, Bnip3 (BCL2 19 kDa-interacting protein 1), a pro-apoptotic factors of the Bcl-2-family, has been previously shown to be up-regulated in malignant tumors [26
]. Diazepam binding inhibitor (Dbi), interacts with hepatocyte nuclear factor-4 α that transcriptionally regulates the genes involved in both lipid and glucose metabolism [27
], was also increased. Previous studies indicated that Dbi levels are higher in hepatocellular carcinoma (HCC) patients [28
], and the elevation of Dbi expression is useful in evaluating malignancy and in diagnostic approaches of tumors in liver tissue [29
]. Fatty acid-binding proteins (Fabps) are involved in lipid metabolism by intracellular transport of long-chain fatty acids. Liver fatty acid-binding protein (Fabp1) is demonstrated immunohistochemically in human hepatocellular malignancies, suggesting that its immunoreactivity is a candidate for the tumor marker in hepatic cell malignancies [30
]. Fatty acid synthase (Fasn) is the key enzyme of de novo
fatty acid synthesis. The over-expression of Fasn is an early phenomenon presented in both hormonally and chemically induced rat hepatocarcinogenesis [31
]. Hypoxia inducible factor-1 alpha (Hif1a) regulates the expression of a myriad of genes involved in oxygen transport, glucose uptake, glycolysis and angiogenesis. The expression of Hif1a in HCC tissue is higher than that in paraneoplastic tissue or normal liver tissue, and Hif1a plays an important role in neovascularization in HCC [32
]. Lgals3 (lectin) has been demonstrated to be associated with assorted processes such as cell growth, tumor transformation and metastasis. It has been reported that Lgals3 expression was induced in cirrhotic liver and HCC, and that the expression of Lgals3 in proliferating cells possibly indicates an early neoplastic event [33
]. The plasminogen activation system, including PAI-1 (plasminogen activator inhibitor 1), plays a crucial role in the process of cancer. PAI-1 is increased in HCC, and its expression is related to the invasiveness, metastasis, and prognosis [34
]. Our findings support the observation that PPARα agonists increase proliferation of peroxisomes in rodent hepatocytes and alter lipid metabolism. In addition, the gene expression profiles indicate a number of leads toward understanding PPARα agonist-induced hepatocarcinogenesis in the mouse.