Administration of a methionine and choline deficient (MCD) diet to rodents causes progressive fibrosing steatohepatitis pathologically similar to human metabolic steatohepatitis. We have previously shown that the peroxisome proliferator-activated receptor-α (PPARα) agonist, Wy-14,643, prevented the development of MCD diet-induced steatohepatitis. We have now tested whether Wy-14,643 ameliorates established steatohepatitis and fibrosis. Male C57BL6 mice were fed the MCD diet for 51 days to induce severe steatohepatitis. They were then treated with Wy-14,643 together with the MCD diet for 5 or 12 days; positive controls continued on the MCD diet for 5 or 12 days. After 5 days of Wy-14,643 treatment, alanine aminotransferase (ALT) levels were significantly decreased, steatohepatitis less severe, and hepatic lipoperoxides significantly reduced. After 12 days, hepatic triglycerides were normalized and there was near resolution of histological changes. MCD dietary feeding was associated with increased expression of vascular cell adhesion molecule (VCAM)-1, and increased numbers of activated macrophages in the liver. Treatment with Wy-14,643 reduced VCAM-1 expression and macrophage numbers. MCD diet-fed mice developed hepatic fibrosis with increased hepatic collagen α1(I), tissue inhibitor of metalloproteinases (TIMP)-1, TIMP-2, and matrix metalloproteinase (MMP)-13 mRNA levels. After treatment with Wy-14,643, expression of these genes was reduced in a manner that paralleled the reduction in activated hepatic stellate cells and near resolution of liver fibrosis. In conclusion, the present study shows that MCD diet-induced fibrosing steatohepatitis can be reversed by treatment with Wy-14,643. It is likely that activation of PPARα reverses fibrosis indirectly by reducing stimuli, such as lipid peroxides, and activation of cells responsible for promoting hepatic fibrosis. (HEPATOLOGY 2004;39:1286–1296.)
The spectrum of nonalcoholic fatty liver disorders ranges from hepatic steatosis to steatohepatitis and fibrosis. While hepatic steatosis is generally considered benign, nonalcoholic steatohepatitis (NASH) can have serious consequences. At the time of diagnosis, between 30% and 42% of NASH patients have advanced bridging fibrosis or cirrhosis.1–3 Hepatic fibrosis is associated with poorer prognosis, as it is more likely to progress to cirrhosis and liver failure, or to hepatocellular carcinoma.4, 5 At present, the therapeutic options for NASH even at the early stages are limited, while liver transplantation is the only option for end-stage cirrhosis.
Mice fed a high-fat methionine and choline deficient (MCD) diet develop steatohepatitis that is histologically similar to liver disease observed in humans with NASH.6, 7 We have previously shown that coadministration of the fibrate, Wy-14,643 with the MCD diet prevents the development of steatohepatitis in this nutritional model.8 Wy-14,643 binds to and specifically activates peroxisome proliferator-activated receptor-α (PPARα), a transcription factor that increases hepatic uptake and breakdown of fatty acids by upregulating the expression of a suite of genes that include peroxisomal and mitochondrial β-oxidation enzymes, cytochrome P450 4A (CYP4A) enzymes, and fatty acid transport proteins.9 As a consequence, Wy-14,643 impedes lipid accumulation in the livers of MCD diet-fed mice,8 a phenomenon likely to be implicated in the prevention of steatohepatitis. Other possible pathways through which PPARα agonist could protect the liver include upregulation of CYP4A enzymes, which could enhance the capacity of the liver to catabolize proinflammatory lipid mediators; these include prostaglandins,10 leukotrienes,11 and end products of lipid peroxidation, such as 4-hydroxynonenal,12 all of which are substrates for CYP4A enzymes. PPARα also blocks the activation of the proinflammatory transcription factor nuclear factor-κB (NF-κB),13 although higher doses of Wy-14,643 can also activate NF-κB.14
Lipoperoxides are increased in human NASH livers, and are localized in areas of hepatic steatosis and fibrosis.15, 16 They also accumulate in the livers of mice fed the MCD diet.17 Malondialdehyde and 4-hydroxynonenal, the most abundant end-products of lipid peroxides, have proinflammatory18, 19 and profibrotic properties20, 21 and may play a role in the pathogenesis of steatohepatitis. Therefore, the lipid-depleting effects of PPARα activation could reduce the substrate for lipid peroxidation, while also increasing the breakdown of formed proinflammatory lipid products. In addition, increased hepatic turnover of triglycerides could be considered hepatoprotective, as fatty livers are more sensitive to oxidative stress that is increased after MCD dietary feeding partly because MCD dietary feeding leads to depletion of glutathione.22
Having shown that activation of PPARα pathways protects against de novo development of MCD diet-induced steatohepatitis by preventing intrahepatic lipid accumulation, we questioned here whether a PPARα agonist, such as Wy-14,643, could also be an efficient treatment for steatohepatitis. Because long-term administration of the MCD diet is associated with progressive pericellular hepatic fibrosis,23 the model also allowed us to study the effects of PPARα stimulation on hepatic fibrogenesis. The results show a dramatic effect of PPARα agonist stimulation in reversing established nutritional steatohepatitis, and shed insight into factors responsible for the reversible hepatic fibrosis that characterizes this model.
Animal breeding and care were conducted within the Animal Care Facility, Westmead Hospital. All animal studies complied with the highest International Criteria of Animal Experimentation, and protocols were approved by the Western Sydney Area Health Service Animal Ethics Committee. Breeding pairs of C57BL6/N mice were kindly donated by Professor Frank J. Gonzalez (National Cancer Institute, National Institute of Health, Bethesda, MD). Male mice, weighing 23–26 g at 8–10 weeks of age, were fed the MCD diet (cat no. 960439; ICN, Aurora, OH) for 51 days to induce fibrosing steatohepatitis. They were then divided into 4 groups and fed 1) MCD diet with Wy-14,643 (0.1% wt/wt) (Chemsyn Laboratories, Lenexa, KS) for 5 days, 2) MCD diet for 5 days, 3) MCD diet with Wy-14,643 (0.1% wt/wt) for 12 days, or 4) MCD diet for 12 days. Another group was fed the control diet (MCD diet supplemented with DL-methionine (3 g/kg) and choline chloride (2 g/kg; cat no. 960441; ICN) for 8 weeks. At the end of the specified feeding periods, mice were anesthetized (ketamine 100 mg/kg and xylazine 20 mg/kg, administered intraperitoneally) and blood and liver tissue collected.
Paraffin-embedded liver tissue was used for all histological analyses. Steatohepatitis was assessed in hematoxylin-eosin stained liver sections (4 μm thick), and blindly scored by an experienced pathologist (P.H.), as previously described.8 Hepatic steatosis was graded according to the percentage of lipid-laden hepatocytes as 0: 0%, 1: 0%–33%, 2: 33%–67%, 3: 67%–100%. Necroinflammation was graded as 0: none, 1: mild, 2: moderate, 3: severe based on the number and size of the necroinflammatory lesions in representative liver sections. Macrophages were visualized in liver sections stained using the diastase/periodic acid-Schiff method.24 Neutrophils were stained with nitrosylated pararosanilin (Sigma, St. Louis, MO) containing naphthol AS-D chloroacetate (0.028% wt/vol; Sigma) in 1M Sorensen's buffer pH 7.4, and counterstained with hematoxylin.25 Collagen was assessed in sirius red-stained liver sections.22 The area occupied by collagen was quantitated by morphometry using Optimas 6.5 software (Media Cybernetics, Silver Spring, MD) and expressed as a percentage of total cross-sectional area.22 Activated hepatic stellate cells (HSCs) were immunostained with antibody for α-smooth muscle actin (Dako, Glostrup, Denmark) as previously described.26 Slides were coded and the average number of α-smooth muscle actin positive cells calculated in 5 fields (400× magnification) per section, independently counted by two scientists (E.I. and I.L.).
Serum alanine aminotransferase (ALT) levels were measured using automated techniques within the Department of Clinical Chemistry, Westmead Hospital. Total liver triglyceride content was measured using a commercial kit (Wako E-test triglyceride kit; Wako Pure Chemical Industries, Osaka, Japan), while hepatic lipoperoxide levels were estimated as thiobarbituric acid-reactive substances (TBARS) (17).
Messenger RNA (mRNA) Analyses.
Total RNA was prepared from frozen liver using TRIzol reagent (Gibco BRL, Melbourne, Australia). Riboprobes for peroxisomal acyl-CoA oxidase (ACO), ketothiolase (KT), Cyp4a10, Cyp4a14, and liver-fatty acid binding protein (LFABP) were prepared as previously described.8 Hepatic mRNA for ACO and KT were detected by northern blotting with 18S as an internal control. Cyp4a10, Cyp4a14, and LFABP mRNA were measured by ribonuclease protection assay, with β-actin or cyclophilin as internal control, as previously described.8, 17
Complementary DNA (cDNA) was synthesized from 5 μg hepatic mRNA using SuperScript III reverse transcriptase and random hexamers (Invitrogen, Carlsbad, CA). Hepatic vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1), tissue inhibitor of metalloproteinase-1 (TIMP-1), TIMP-2, matrix metalloproteinase-2 (MMP-2), MMP-13, transforming growth factor-β1 (TGF-β1), connective tissue growth factor (CTGF), and collagen α1(I) mRNA were quantitated using real time polymerase chain reaction (PCR) with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as internal control. PCR reactions and analyses were carried out using the ABI Prism 7700 Sequence Detector and software (Applied Biosystems, Foster City, CA) using TaqMan® fluorogenic probes (GAPDH and collagen α1(I)) or SYBRgreen® detection (VCAM-1, ICAM-1, TIMP-1, TIMP-2, MMP-2, MMP-13, TGF-β1, CTGF). All primers and probes were designed using the Primer Express™ design software (Applied Biosystems); sequence details are given in Table 1. In order to quantify mRNA species, PCR product amplification curves for each sample were compared to those of standard dilution curves. The final result for each sample was normalized to the respective GAPDH value.
Table 1. Primer Pairs and Probe Sequences Used for Real Time PCR Reactions
Analysis of Hepatic Protein Expression by Western Blotting.
Frozen liver was homogenized in lysis buffer containing 50 mmol/L Hepes, 1.5 mmol/L MgCl2, 150 mmol/L NaCl, 1 mmol/L ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid, 10% glycerol, 0.1% Triton X-100, and a mixture of protease inhibitors, and total hepatic protein estimated using bovine serum albumin as a standard (DC protein assay; Biorad, Hercules, CA). Hepatic protein (20 μg) was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto polyvinylidine difluoride membranes (Immobilin-P; Millipore, Bedford, MA). Membranes were blocked using skim milk powder and then incubated overnight with specific primary antibodies against TGF-β1 (sc-146), VCAM-1 (sc-1504), or ICAM-1 (sc-8439; Santa Cruz Biotechnology, Santa Cruz, CA). Peroxidase conjugated anti-rabbit immunoglobulin (Ig)G was used as secondary antibody for the detection of TGF-β1, anti-goat IgG used for VCAM-1, and anti-mouse IgG used for ICAM-1 (Sigma). Proteins were visualised using enhanced chemiluminescence (Pierce Perbio, Rockford, IL).
Two-way ANOVA was used to analyze the effects of dietary regime on different parameters. When F was significant, the Student's t test was used to compare individual differences. When P was less than .05, differences were considered significant.
Mice fed the MCD diet for 8 or 9 weeks developed severe steatohepatitis (Fig. 1A, B) associated with elevated serum ALT levels (Fig. 2). After induction of steatohepatitis, treatment with Wy-14,643 for 5 and 12 days was associated with significant reduction in serum ALT levels (Fig. 2). Consistent with these findings, the severity of liver injury observed histologically was also reduced. In animals treated for 5 days, several large lipid droplets remained but inflammatory foci were reduced in size (Fig. 1C) compared to untreated MCD diet-fed mice. After 12 days of treatment, very few small fat droplets remained and inflammatory cells were almost completely absent (Fig. 1D). Histological grading of liver sections confirmed the significant improvement of hepatic injury achieved by Wy-14,643 treatment (Table 2).
Table 2. Effect of MCD Diet and Treatment With Wy-14,643 on Severity of Hepatic Steatosis and Necroinflammatory Lesions
Control (8 wk)
MCD (8 wk)
Wy-14,643 (5 d)
MCD (9 wk)
Wy-14,643 (12 d)
NOTE. The severity of hepatic steatosis and necroinflammation were each scored on a scale of 0–3 by an independent pathologist (PH), as described in Materials and Methods. Data are mean ± SD, n = 3–6/group.
P < .001 compared with control diet-fed mice
P < .05
P < .01
P < .001 compared with untreated mice fed the MCD diet for same total period of time.
Wy-14,643 Reduces Hepatic Triglyceride and Lipoperoxide Levels.
Administration of the MCD diet was associated with a two-fold increase in intrahepatic triglyceride levels at 8 and 9 weeks. Consistent with the histological findings of continuing steatosis (Fig. 1C), 5 day treatment with Wy-14,643 did not significantly alter liver triglyceride content compared to untreated MCD diet-fed mice (Table 3). However, after treatment for 12 days, Wy-14,643 reduced hepatic triglyceride to levels observed in control diet-fed mice (Table 3). In accordance with the proposal that this could be due to hepatic upregulation of PPARα-regulated genes that promote fatty acid disposal, transcript levels for the peroxisomal β-oxidation enzymes ACO, KT, and the fatty acid transport protein, LFABP, were increased 8, 9, and fivefold, respectively, in Wy-14,643-treated groups compared to untreated MCD-fed mice (Fig. 3).
Table 3. Effect of MCD Diet and Treatment With Wy-14,643 on Hepatic Triglyceride and Lipoperoxide Content
Control (8 wk)
MCD (8 wk)
Wy-14,643 (5 d)
MCD (9 wk)
Wy-14,643 (12 d)
NOTE. Hepatic lipoperoxide content was measured as thiobarbituric acid reactive substances. Data are mean ± SD, n = 3–6/group.
P < .01
P < 0.01 compared with control diet-fed mice
P < .05
P < .01 compared with untreated mice fed the MCD diet for same total period of time.
We previously hypothesized that PPARα-stimulated fatty acid breakdown may prevent inflammation by critically reducing lipid substrate for formation of proinflammatory lipoperoxides.8 As expected,17 hepatic lipoperoxide levels were increased 100-fold in mice fed the MCD diet compared to control diet-fed mice (Table 3). Treatment with Wy-14,643 significantly reduced lipoperoxide levels. Interestingly, this occurred after only 5 days of treatment, before the significant reduction in liver triglyceride levels observed at 12 days (Table 3). Enzymes from the Cyp4a family are transcriptionally regulated by PPARα and are known to play a role in the metabolism of lipid peroxides.12 It is therefore of interest that hepatic expression of Cyp4a10 and Cyp4a14 was increased eight- and seven-fold, respectively, after treatment with Wy-14,643 for 5 or 12 days (Fig. 3).
Effect of Wy-14,643 on Intrahepatic Recruitment and Activation of Inflammatory Cells and Adhesion Molecules.
Further analyses of liver sections from 8-week MCD diet-fed mice revealed large clusters of foamy macrophages (Fig. 4A). Treatment with Wy-14,643 for 5 days reduced the number and size of these clusters (Fig. 4B). After 12 days, remaining macrophages were small and occurred singly in perivenular regions (Fig. 4C). The number of polymorphonuclear leukocytes in liver sections was also elevated in mice fed the MCD diet for 8 or 9 weeks compared to controls (7.8 ± 2.2, 5.5 ± 2.8, and 0.28 ± 0.32 polymorphonuclear leukocytes/400× field respectively, P < .01 for MCD diet-fed mice compared to controls). Treatment with Wy-14,643 reduced polymorphonuclear leukocyte numbers after 5 days (3.5 ± 0.86 cells/400× field, P = .004); this effect appeared to be sustained after 12 days treatment (2.9 ± 1.6 cell/400× field, P = 0.08), though the difference was no longer significant.
In mice fed the MCD diet for 8 or 9 weeks, hepatic mRNA and protein levels of VCAM-1 and ICAM-1 were increased (Fig. 5). Treatment with Wy-14,643 significantly decreased VCAM-1 and ICAM-1 mRNA expression (Fig. 5A, B), and VCAM-1 protein levels (Fig. 5C). However, the effect, if any, of Wy-14,643 on ICAM-1 protein levels was minimal and not significant (Fig. 5D).
Rapid Reversal of Hepatic Fibrosis After Treatment With Wy-14,643.
Steatohepatitis in the MCD model gives rise to appreciable intraparenchymal pericellular fibrosis, as shown in sirius red-stained liver sections from mice fed the MCD diet for 8 and 9 weeks (Fig. 6A, B). Having shown that hepatic steatosis and inflammation were reversed after treatment with Wy-14,643, it was therefore of considerable interest to determine whether hepatic fibrosis was also reduced. Treatment with Wy-14,643 for 5 days caused impressive reduction in the amount of collagen fibers (Fig. 6C), and after 12 days, hepatic fibrosis was almost completely reversed (Fig. 6D). Quantitation of collagen in sirius-red stained liver sections (Fig. 6E), and assessment of hepatic collagen α1(I) gene expression (Fig. 6F), supported the observations that MCD dietary feeding increased, and treatment with Wy-14,643 reversed, liver fibrosis.
Effects of Wy-14,643 on Hepatic Stellate Cells.
Control diet-fed mice showed few activated HSCs, as indicated by paucity of α-smooth muscle actin positive-staining cells in liver sections (Fig. 7A). Consistent with presence of hepatic fibrosis, there were increased numbers of α-smooth muscle actin-positive, activated HSCs in mice fed the MCD diet for 8 weeks (Fig. 7B) and 9 weeks (not shown). Treatment with Wy-14,643 for 5 days reduced the number of activated HSCs (Fig. 7C) compared to mice fed the diet for a similar period of time, and after 12 days, HSCs were scanty (Fig. 7D).
Effects of Wy-14,643 on Hepatic Expression of Genes Involved in Fibrogenesis.
In order to clarify the mechanism through which Wy-14,643 rapidly reduced liver fibrosis, we examined molecular pathways involved with matrix remodelling and degradation to determine whether activation of PPARα stimulates matrix degradation at the transcriptional level. In control diet-fed mice, there was only low level hepatic expression of genes for inhibitors of fibrosis reversal (TIMP-1 and TIMP-2) and those involved in matrix degradation (MMP-2 and MMP-13; Table 4). Consistent with the presence of hepatic fibrosis,27, 28 MCD dietary feeding increased the expression of transcripts for these proteins. Treatment with Wy-14,643 for 12 days significantly reduced expression of TIMP-2, and there was a trend towards reduction in TIMP-1. MMP-13 expression was not increased, but was actually reduced after 12 days of treatment, while MMP-2 mRNA levels remained unaffected (Table 4). The delayed effect of Wy-14,643 on expression of these fibrogenic genes is in contrast to its rapid effects on genes involved in hepatic fatty acid oxidation (ACO, KT) and inflammation (VCAM-1 and ICAM-1), and suggests that PPARα does not directly affect these genes at the transcriptional level.
Table 4. Effect of MCD Diet and Treatment With Wy-14,643 on Hepatic mRNA Levels of Genes Involved in Hepatic Fibrogenesis
Control (8 wk)
MCD (8 wk)
Wy-14,643 (5 d)
MCD (9 wk)
Wy-14,643 (12 d)
NOTE. Hepatic mRNA levels of target genes were assessed using reverse transcription-real time PCR, standardized against internal control (GAPDH) and are expressed as fold difference compared with values obtained in mice fed the control diet. Data are mean ± SD, n = 3–6/group.
P < .05
P < .01
P < .001 compared with control diet-fed mice
P < .05
P < .01 compared with untreated mice fed the MCD diet for same total period of time.
Effects of Wy-14,643 on TGF-β1 and Connective Tissue Growth Factor.
TGF-β1 is an important profibrogenic cytokine that is upregulated in rats fed the MCD diet.22, 26 Hepatic mRNA levels of TGF-β1 were increased in mice fed the MCD diet compared to control diet-fed mice. Treatment with Wy-14,643 (for 5 or 12 days) did not affect hepatic TGF-β1 mRNA levels (Table 4). Because TGF-β1 is post-transcriptionally regulated, we also determined hepatic protein levels of the active form of TGF-β1 by Western blotting. Protein levels of the active form of TGF-β1 were actually lower in MCD diet-fed mice compared to those fed the control diet, and treatment with Wy-14,643 had no effect on active levels of this cytokine (Fig. 8). Therefore, PPARα-mediated reversal of hepatic fibrosis in MCD diet-fed mice is not due to suppression of hepatic mRNA levels or of activation of TGF-β1.
CTGF promotes fibrosis by stimulating fibroblast proliferation and expression of extracellular matrix components. Hepatic mRNA expression of CTGF was increased after MCD dietary feeding compared to control diet-fed mice, but Wy-14,643 did not affect CTGF mRNA levels (Table 4).
The pathogenesis of steatohepatitis caused by metabolic factors is not fully understood, but accumulation of fat in the liver (steatosis) is clearly an essential precondition.29, 30 In experimental steatohepatitis caused by a lipogenic diet deficient in methionine and choline, lipid peroxidation appears to be a key mediator of subsequent inflammatory events.8, 17, 22 We have previously shown that co-administration of the MCD diet with the PPARα agonist, Wy-14,643, prevented the development of steatohepatitis, presumably through the effects of PPARα stimulation on hepatic fatty acid turnover, which prevented the increase in hepatic triglycerides and lipoperoxides caused by methionine and choline deficiency.8 Because effective treatment of the human form for steatohepatitis associated with insulin resistance, diabetes, and obesity (NASH) is yet unclear, the paradigm of steatohepatitis treatment investigated in the present study provides important new information. Thus, reversing triglyceride and lipid peroxidation accumulation in the liver with Wy-14,643 rapidly reversed steatohepatitis. Further, such resolution was very effective, with near reversion of liver fibrosis within 12 days.
Activation of PPARα upregulates a set of target genes that promote fatty acid breakdown in the liver, including the peroxisomal β-oxidation enzymes, ACO and KT, fatty acid transport proteins such as LFABP, and Cyp4a enzymes. In the context of hepatic lipid overload, induced expression of these genes leads to resolution of hepatic steatosis.31, 32 We previously hypothesized that the reduced hepatic lipid load due to activation of PPARα pathways may additionally prevent inflammation by reducing lipid substrate for peroxidation.8 In the current study, treatment with Wy-14,643 led to a reduction in hepatic lipoperoxide content that actually preceded the decrease in hepatic triglycerides. It is therefore unlikely, in this scenario of preexisting steatohepatitis, that reduction in hepatic lipid is the only driving force behind reduced formation of these important mediators of hepatic inflammation and fibrosis. Rather, PPARα-mediated upregulation of Cyp4a and peroxisomal β-oxidation enzymes may increase the hepatic capacity to break down existing lipoperoxides, promoting their removal and thus countering a potent stimulus for hepatic inflammation and fibrosis.11
Another mechanism through which Wy-14,643 may promote recovery from steatohepatitis is by reducing the expression of molecules that facilitate transendothelial migration of leukocytes. We showed that hepatic VCAM-1 and ICAM-1 expression levels were increased in mice fed the MCD diet. Treatment with Wy-14,643 downregulated hepatic VCAM-1 mRNA and protein levels. This could be functionally important, as we also noted that the number and activation of hepatic macrophages was substantially reduced in treated animals. While treatment with Wy-14,643 was also associated with downregulation of ICAM-1 expression at the mRNA level, protein levels were not significantly reduced, consistent with other reports of the effects of PPARα agonists on ICAM-1 expression.33, 34 The implications of these findings are twofold. First, although overlapping in specificity, ICAM-1 recruits cells most often linked to an acute inflammatory response. Therefore, ICAM-1, but not VCAM-1, increases transendothelial migration of neutrophils. As treatment with Wy-14,643 reversed MCD diet-induced steatohepatitis without significantly affecting neutrophil numbers in the liver, this cell type is unlikely to play a critical pathogenic role in the established phase of this form of steatohepatitis, as do polymorphs in alcoholic liver disease.35 Second, activation of PPARα inhibits NF-κB-dependent gene transcription,13 providing another mechanism through which Wy-14,643 may effect recovery from steatohepatitis. Both VCAM-1 and ICAM-1 are downstream targets of NF-κB activation, but VCAM-1 (and not ICAM-1) expression can be reduced by PPARα through non-NF-κB-dependent mechanisms.33 Thus, the control exerted by PPARα over NF-κB pathways is not absolute, and the present results are not consistent with the concept that inhibition of NF-κB could be the primary mechanism behind PPARα-mediated reversal of MCD diet-induced hepatic inflammation.
A novel and exciting finding from the present study was that treatment with Wy-14,643 caused regression of hepatic fibrosis. In contrast to PPARγ and PPARβ, PPARα is not expressed in quiescent or activated HSCs and has not been shown to affect HSC biology in vitro or hepatic fibrosis.36–38 So the rapidity with which fibrotic recovery occurred in these experiments was unexpected. Consistent with previous studies in rats,22 we showed that collagen α1(I), TIMP-1, and TIMP-2 were upregulated in MCD diet-fed mice. As in other models of chronic hepatic fibrosis,27 expression of the matrix remodelling enzymes, MMP-2 and MMP-13, was also increased. Treatment with Wy-14,643 lead to downregulation of collagen α1(I), TIMP-2, and MMP-13 expression after 12 days, while MMP-2 expression was maintained. We interpret these findings as indicating that if an imbalance between MMPs and their inhibitors exists, it is likely to tend towards increased degradation. However, the delayed effect of the PPARα agonist on mRNA levels of these genes indicates that it is unlikely that Wy-14,643 directly affects their expression; rather the downregulation is more likely the result of reduced activation of HSCs and Kupffer cells. A similar gradual decrease in TIMP and MMP expression has been observed during spontaneous recovery from carbon tetrachloride-induced liver cirrhosis.39
TGF-β1 is an important profibrogenic cytokine. It is an activator and survival factor for HSCs, and alters gene expression to favor extracellular matrix deposition. Elevated levels of TGF-β1 have been detected in fibrotic human liver, cirrhotic rat liver,40 and in the livers of rats fed the MCD diet22 where they are likely to play an important role in the development of hepatic fibrosis.26 CTGF mediates the profibrogenic downstream effects of TGF-β1 activation, including fibroblast proliferation and modulation of gene expression. CTGF is chiefly, but not exclusively, induced by TGF-β1.41–43 Expression of CTGF is low or undetectable in normal liver, but is increased in humans with cirrhosis44 or NASH,45 and in rat models of liver fibrosis.44, 46 In the present study, MCD dietary feeding was associated with increased hepatic mRNA expression of TGF-β1 and CTGF, but that treatment with Wy-14,643 failed to alter these levels. This indicates that PPARα activation does not promote reversal of hepatic fibrosis by suppressing the expression of these profibrogenic cytokines. Taken together, our data support the hypothesis that activation of PPARα reverses fibrosis by reducing profibrogenic stimuli. These could include lipoperoxides, which are known regulators of collagen I gene expression,21 and/or diminished release of cytokines by macrophages and other leukocytes that stimulate HSC activity and increase the production of extracellular matrix components.47
In considering the possible relevance of the results of this study to humans, it is necessary to consider the species differences between PPARα activation in humans and rodents. PPARα is expressed in human liver at ten-fold lower levels than in rats or mice.48 Human PPARα does not upregulate peroxisomal β-oxidation enzymes, although it does increase the expression of mitochondrial β-oxidation enzymes.49 This is likely to be functionally relevant as fibrate therapy in humans is effective against hyperlipidemia, reducing hypertriglyceridemia and increasing HDL cholesterol. In vitro studies using human-derived primary and immortalized endothelial cells have shown that PPARα agonists downregulate cytokine-induced VCAM-1 expression, thereby reducing endothelial cell-leukocyte interactions.33, 34 PPARα agonist also partially inhibits the activation of human T lymphocytes, and reduces their cytokine production.50 While the in vivo implications and relevance to human fatty liver disease are yet to be determined, PPARα may have important anti-steatotic and anti-inflammatory effects in humans as well as rodents, in addition to its ability to reduce hyperlipidemia.
In summary, the present study shows the dynamic and potentially reversible nature of fibrosing steatohepatitis in this rodent dietary model. While reduction in hepatic lipid levels is likely to be important in promoting long-term protection against progression of steatohepatitis,8 a complete reduction in hepatic lipid appears to be unnecessary in order for recovery from inflammation (and fibrosis) to begin. If this interpretation is correct, it reinforces the fundamental principle that reduction in hepatic accumulation of lipid and concomitant reduction in proinflammatory lipid metabolites (including products of lipid peroxidation) are crucial in order to reverse steatohepatitis. The finding that PPARα agonist treatment reversed hepatic fibrosis without any obvious direct effects on major regulators of hepatic fibrogenesis suggests that targeted modulation of intrahepatic lipid metabolism and anti-inflammatory therapy can also reverse the hepatic fibrosis that results from steatohepatitis. The role of pharmacologic modulation of hepatic lipid turnover in the treatment of NASH is therefore worthy of further investigation.
The authors thank Prof. Frank J. Gonzalez (National Cancer Institute, National Institute of Health, Bethesda, MD) for supplying breeding pairs of C57BL6/N mice, Martine Stevens (Pathology Unit, Catholic University of Louvain, Belgium) for assistance with immunohistochemistry, and Sandy Bierach for coordination of animal care and breeding. We also thank Assoc. Prof. Jacob George for helpful discussion during the study and in the preparation of this manuscript.