Dysregulated Bile Acid Synthesis, Metabolism and Excretion in a High Fat-Cholesterol Diet-Induced Fibrotic Steatohepatitis in Rats

Nonalcoholic steatohepatitis (NASH) is a progressive form of nonalcoholic fatty liver disease (NAFLD), characterized by liver fatty deposition with various degrees of inflammation and fibrosis [1]. Although obesity and insulin resistance are considered as pathogenic factors, populations without these risks still could develop NASH [2, 3], as lifestyle-related factors such as food intake and food composition may play a role in disease progression [1]. Of these, dietary cholesterol is an important risk factor in the progression of steatosis, inflammatory recruitment and fibrosis in NASH patients [2, 4, 5], and in a wide variety of animal models [6‐8]. Interestingly, dietary cholesterol intake has been reported to be abundant in NASH patients without obesity or insulin resistance [2, 3].

An animal model with fibrotic steatohepatitis was established by our group by feeding a high fat-cholesterol (HFC) diet to stroke-prone spontaneously hypertensive 5/Dmcr (SHRSP5/Dmcr) rats [9‐11]. The most profound feature was severe fibrosis. This model recapitulated histological features of human NASH and exhibited a similar physiological condition to a subgroup of NASH patients without obesity or insulin resistance. In view of the extraordinary cholesterol content (5 % by weight) in the HFC diet, and the hepatic total cholesterol which progressively accumulated as liver disease developed, unlike triglycerides [9, 10], the accumulation of cholesterol, rather than triglycerides, may have played a critical role in the progression of fibrotic steatohepatitis in this dietary rat model.

In hepatocytes, cholesterol homeostasis pathways include cholesterol de novo synthesis, uptake in the form of low-density lipoprotein (LDL) and chylomicron remnants, excretion into the blood in the form of very-low-density lipoprotein, excretion and uptake through bile, and catabolism to bile acids (BAs) and their excretion [5]. BA synthesis is initially catalyzed by cholesterol 7α-hydroxylase (CYP7A1) in the classic pathway, followed by sterol 12α-hydroxylase (CYP8B1) to form cholic acid and thus control the ratio of cholic acid to chenodeoxycholic acid in the BA pool. The alternative pathway is catalyzed by sterol 27-hydroxylase (CYP27A1), followed by oxysterol 7α-hydroxylase (CYP7B1), and mainly produces chenodeoxycholic acid [12]. BAs are hepatotoxic when their concentrations reach abnormally high levels, because they can cause mitochondria damage, apoptosis or necrosis, and eventually fibrosis and cirrhosis [13]. Indeed, BA levels increased in the liver of NASH patients and significantly correlated with histological inflammation and fibrosis [14]. Taken together, the significance of dietary cholesterol in the development of NASH may be attributable to BA-induced liver injury.

Therefore, we extended our earlier study [11] and used the same liver samples to investigate the abnormalities of cholesterol and BA kinetics in the context of HFC diet-induced fibrotic steatohepatitis in rats.

We previously reported that the HFC diet time-dependently induced lipid and inflammatory cell accumulations, hepatocyte ballooning, necrosis and fibrosis in the liver of SHRSP5/Dmcr rats, although these rats were not obese or insulin resistant [9‐11]. The lipid accumulation in the liver might have resulted from cholesterol rather than triglycerides, because the level of the latter time-dependently decreased [11]. Thus, the major lipotoxic molecule promoting the pathogenesis of fibrotic steatohepatitis in this model might have been derived from cholesterol.

Feedback regulation of cholesterol biosynthesis mainly occurs in the liver and is exerted primarily on the HMG-CoA reductase that catalyzes the rate-limiting step of cholesterol biosynthesis [18]. Similar to previous reports on feedback inhibition by dietary cholesterol [19, 20], hepatic HMG-CoA reductase was down-regulated at 8 and 14 weeks to compensate for the increased absorption of dietary cholesterol. However, SREBP-2, a well-known transcription factor for cholesterol homeostasis, did not change, indicating that not only SREBP-2 but also many additional factors (hormones, cytokines, etc.) might be required for complete control of lipid metabolism [21]. LDLR and LRP1 are members of the LDLR family and contribute to controlling serum cholesterol levels by removing cholesterol-contained lipoproteins from the bloodstream [22]. Dietary cholesterol (1 %) increased hepatic LDLR and had no effect on the plasma cholesterol concentration in Wistar rats, but the effects were opposite in Sprague–Dawley rats [23]. We observed a downregulation of both LDLR and LRP1 in SHRSP5/Dmcr rats fed a 5 % cholesterol diet, and eventually hypercholesterolemia occurred. Such disagreements with previous reports [23] may derive from differences in dietary cholesterol level, strain, and liver condition. Together, cholesterol synthesis and uptake were suppressed as expected in response to cholesterol accumulation in liver with fibrotic steatohepatitis, indicating that these pathways may not be disorganized during the development of steatoheptitis.

We next focused on how dietary cholesterol induced BA accumulation, because it may contribute to understanding the pathogenesis of fibrotic steatoheptitis. Hepatic biotransformation of cholesterol into BAs constitutes the major pathway of cholesterol catabolism. Increased conversion into BAs via upregulation of microsomal CYP7A1 in response to cholesterol challenge was confirmed in rats [24, 25], but CYP8B1 did not change [25]. CYP27A1 was induced in the high-cholesterol (1 %) diet group of rats, but not in mice [25, 26], and stayed upregulated in rabbits with a 2 % cholesterol diet intake, regardless of the BA pool size [27]. As expected, in response to HFC diet-induced hepatic cholesterol accumulation in our model, CYP7A1 protein was upregulated over 2-fold, but CYP8B1 was suppressed in the classic synthesis pathway. And while CYP27A1 was downregulated, CYP7B1 was induced in the alternative pathway, especially at 8 and 14 weeks, when fibrotic steatohepatitis occurred. This suggested that an increased BA production, especially chenodeoxycholic acid rather than cholic acid, may be involved in fibrogenesis. Our further study indeed identified an increment in hepatic chenodeoxycholic acid rather than in cholic acid, especially during these periods, by analysis of BA profiles (Jia X et al., manuscript in submission). Downregulations of CYP27A1 and CYP8B1 were also observed in NAFLD patients [28] and the mouse model of NASH [7], but no upregulations of CYP7A1 and CYP7B1 were observed. CYP27A1 is a mitochondrial protein and its downregulation may reflect mitochondrial injury that is present in NASH development [28].

BA homeostasis is under the regulation of nuclear receptors. LXRα is a sterol sensor, and increased intercellular cholesterol drives the production of oxysterols, which are agonists of LXRα [29]. LXRα regulates cholesterol catabolism by acting as a positive regulator of CYP7A1, and mediates fatty acid synthesis by inducing SREBP-1c expression. However, in combination with our earlier study [11], neither the cholesterol-LXRα-CYP7A1 nor the cholesterol-LXRα-SREBP-1c signaling pathway was stimulated in our model, unlike with NAFLD patients [5, 29]. The reason for this apparent discrepancy requires further investigation. FXR is proposed to be a nuclear BA receptor and directly or indirectly (via SHP) inhibits CYP7A1. FXR was not activated in rats fed 2 % cholesterol for 1 week, where CYP7A1 was stimulated and the BA pool size was stable [30]. Likewise, CYP7A1 protein was continuously induced from 8 to 14 weeks in rats fed a 5 % cholesterol-contained HFC diet, but nuclear FXR was not. Nevertheless, hepatic chenodeoxycholic acid species, the most potent FXR ligand [31], was found to increase in HFC-fed rats throughout the study (Jia X et al., manuscript in submission). Therefore, the failure of FXR upregulation by BAs may suggest that there was a disturbed feedback regulation of BA synthesis in our model.

Biliary compounds such as BAs and bilirubin are cytotoxic when present in abnormally high concentrations. Both glucuronidation and sulfation become pivotal eliminating pathways in cholestasis. They transform hydrophobic, toxic substrates into more hydrophilic, less-toxic derivatives for biliary and urinary excretion, although glycine and taurine conjugation are primarily involved in the process [32, 33]. The most profound changes in the pathways of BA turnover in this dietary model were disturbed UGT-catalyzed glucuronidation, blunted SULT2A1-catalyzed sulfation, and impaired BSEP-mediated canalicular export to bile duct, especially observed at 8 and 14 weeks, when fibrosis progressed. All of these disorders, together with upregulations of BA synthetic enzymes, revealed hepatic accumulation of toxic biliary constituents, as evidenced by a direct analysis of BA profiles in liver (Jia X et al., manuscript in submission). Consequently, toxic biliary compound overload in liver may have contributed to the pathogenesis of liver damage in this model, especially severe fibrosis, as confirmed by the correlation analysis between specific BAs and parameters of liver injury (Jia X et al., manuscript in submission). HFC-diet feeding also induced MRP3-regulated basolateral excretion, an important alternative spillover route during BA overload, which was in agreement with the changes under a cholestatic condition [34]. These changes allow BA elimination from liver to the blood, but are not sufficient for a complete detoxification of the liver cells.

FXR, PXR and CAR protect against hepatotoxicity of BAs in a complementary manner by regulating genes involved in BA transport and detoxification, such as BSEP, UGT and SULT [33, 35]. Simultaneous suppression of nuclear FXR, PXR and CAR occurred with the development of fibrotic steatoheptitis, especially at 14 weeks. On one hand, the downward shift of nuclear FXR expression in HFC-fed rats may correlate with the suppressed transcription of BSEP; on the other hand, downregulated PXR and CAR may result in reduced UGT activity and SULT2A1 protein. PXR and CAR also regulate the transport of MRP3 [33]; however, regardless of the two nuclear regulators, MRP3 was spontaneously enhanced as a compensatory efflux pathway in response to reduced BA transport during cholestasis [34]. Therefore, upregulated MRP3 may not have been dependent on the downregulation of these nuclear receptors in our model.

Finally, we shed insight into the inconsistencies of gene expression in mRNA and protein levels, especially at some particular time points, such as LDLR (2 weeks), LRP1 (each time-point), LXRα (2 and 8 weeks), FXR (each time-point), PXR (2 and 8 weeks) and CAR (8 weeks). Our study analyzed mRNA and protein levels in triplicate and observed similar results. Therefore, we excluded experimental errors. Gene expression is a highly complex process that comprises many steps, from transcription to protein degradation. Following transcription, the pre-mRNA is processed and exported from the nucleus to the cytoplasm where additional steps of mRNA maturation take place. The mature mRNA is then either translated, stored, or degraded. Each step in gene expression is under elaborate regulation and different phases of the process are coupled [36]. Thus, the entire gene expression process can be viewed as a more complex network with feedback between coupled regulatory mechanisms. The discrepancy between mRNA and protein levels could be considered as reasonable phenomena. Additionally, the fatty acid type in HFC diet may affect the expression of mRNA for LDLR and LRP1, as fatty acids from the diet delivered to the liver in chylomicron remnants influenced the transcription of hepatic genes regulating their uptake, such as LDLR and LRP1 [37]. In the liver of the HFC diet-induced rat model, to maintain cholesterol homeostasis, LDL uptake may have been suppressed by inducing degradation of LDLR and LRP1 via protein ubiquitination [38, 39]. LXRα, FXR, PXR and CAR have functional connections in the regulation of cholesterol biosynthesis and metabolism to BAs in liver [40]. Their protein expressions tended to decrease in the HFC diet-induced rat model during the whole study, although significance relative to control was limited. Post-translational modification or trafficking might explain the differences between mRNA and mature protein [41], which warrants further investigation.

Many NAFLD patients are not obese, and many of them do not exhibit insulin resistance or increased visceral fat volume, which compel us to question what triggers and maintains the disease in non-obese patients [29]. Interestingly, a nutritional assessment of NAFLD patients revealed that dietary cholesterol intake was significantly higher in non-obese patients than in obese ones [3]. Indeed, in animal models, it has been reported that a high-cholesterol diet caused NASH without triggering obesity or insulin resistance [42]. Although we used a 5 % cholesterol-contained HFC diet in the present study, a 2.5 %-cholesterol diet could also induce fibrotic steatohepatitis, to a lesser extent, in SHRSP5/Dmcr rats (data not shown), suggesting that cholesterol dose-dependently induced the liver damage. Together, in combination with diverse changes in hepatic triglycerides and cholesterol with liver disease progression [11], cholesterol over-intake and accumulation in liver may be more potential than high fat intake and triglyceride deposition in our non-obese and non-diabetic model.

In conclusion, the alterations in hepatic cholesterol synthesis and catabolism enzymes, uptake receptors, BA transport systems, phase II detoxification enzymes, and nuclear receptors investigated in the liver of an HFC diet-induced fibrotic steatohepatitis rat model are summarized in Fig. 6. Primarily, cholesterol over-intake in this model accumulated cholesterol in rat livers, which further upregulated BA synthetic enzymes, CYP7A1 and CYP7B1, whereas it downregulated BSEP, UGT and SULT2A1, especially at 8 and 14 weeks. These changes may result in BA accumulation in the liver and mediate a transition towards fibrotic steatohepatitis. Therefore, this study suggests that cholesterol over-intake may be one of the potent inducers of fibrotic steatohepatitis, which is mainly achieved via hepatotoxicity of BA accumulation rather than a well-established cholesterol-LXRα-SREBP-1c pathway in humans, although hepatotoxicity of cholesterol accumulation may be considered.