Hepatitis C virus, steatosis and lipid abnormalities: clinical and pathogenic data

Authors


Correspondence
Francesco Negro, MD, Viropathology Unit, Divisions of Gastroenterology, Hepatology, and Clinical Pathology, University Hospitals, 24 rue Micheli-du-Crest, 1211 Geneva 14, Switzerland
Tel: +41 22 3729340
Fax: +41 22 3729366
e-mail: francesco.negro@hcuge.ch

Abstract

Abnormal accumulation of fat in the liver (steatosis) is commonly observed in hepatitis C virus (HCV) infection, and the severity of steatosis has been well correlated with the degree of hepatic fibrosis. In patients with chronic HCV infection, steatosis may occur in conjunction with other metabolic risk factors such as insulin resistance and the metabolic syndrome. This was observed primarily in patients infected with non-genotype 3 virus. Otherwise, in HCV-infected patients, especially those infected with genotype 3a, reductions in total cholesterol as well as high-density lipoprotein and low-density lipoprotein cholesterol are observed compared with matched controls, and the normalization of these parameters appears to be an important correlate of the response to antiviral therapy. In that setting, the pathogenic mechanisms involved in HCV-induced steatosis are mediated in large part by the HCV core protein, whose expression is associated with lipid droplet accumulation, changes in lipogenic gene expression and/or the activity of lipogenic proteins, and effects on mitochondrial oxidative function. The importance of genes such as peroxisome proliferator-activated receptor-α and the proteasome activator PA28-γ in HCV-mediated steatosis has been elucidated from studies in genetically altered mice, and the manipulation of these and other pathways may provide an avenue for therapeutic intervention.

Hepatitis C virus (HCV) and non-alcoholic fatty liver disease (NAFLD) are the two most common causes of chronic liver disease in North America (1). The abnormal accumulation of fatty deposits in the liver, or steatosis, is a common feature of HCV infection, occurring in about 50% or more of patients (1). The degree of steatosis has been correlated with the development of liver fibrosis in these patients (1–5). In certain respects, HCV infection is reminiscent of NAFLD and the more advanced condition known as non-alcoholic steatohepatitis (NASH), in that steatosis, serum dyslipidaemia and evidence of hepatic oxidative stress are observed in all these conditions (6). However, HCV also appears to have important effects on gene expression and intracellular signalling pathways that are distinct from NAFLD and NASH. As such, HCV can be regarded not only as a viral disease of the liver but also as a metabolic disease (6). Fibrosis, obesity, body mass index (BMI), type 2 diabetes, hypertriglyceridaemia, the presence of the metabolic syndrome and infection with HCV-3a have all been associated with the development of steatosis in HCV-infected patients; however, BMI and genotype 3a infection appear to be the most important independent risk factors (1, 3–5). Insulin resistance (IR) and an associated increase in the risk of type 2 diabetes is another feature commonly observed in HCV infection (7, 8). Thus, steatosis may occur in a sizable proportion of HCV-infected patients in conjunction with IR and other metabolic risk factors, all of which can lead to poor outcomes.

In this article, we discuss the key features of lipid abnormalities and steatosis associated with HCV infection, their potential importance for HCV pathogenesis and the molecular mechanisms associated with them.

Lipid abnormalities in steatosis and fatty liver

Non-alcoholic fatty liver disease generally refers to hepatic fat that is not a result of liver diseases such as alcoholism, viral infection, drugs, toxins or autoimmune disease (8, 9). Non-alcoholic fatty liver disease can occur both in individuals with obesity as well as in those who are lipoatrophic (8). Once the liver becomes fatty, it begins to overproduce many factors associated with cardiovascular risk, including very low-density lipoprotein (VLDL) and C-reactive protein. Additionally, the risk for type 2 diabetes is increased (8). Accordingly, IR is generally, if not always, a component of NAFLD, and type 2 diabetes increases the risk for both NAFLD and NASH (8). Evidence for an inter-relationship between fatty liver and IR also emerges from the finding that weight loss and insulin sensitizers can reduce liver fat content (8, 10–13). Studies have shown that thiazolidinediones such as rosiglitazone and pioglitazone can decrease liver fat to an extent similar to moderate weight loss. Despite these associations, it should be noted that steatosis per se does not always induce IR but depends on the aetiology of liver fat deposition. Firstly, in a study of 141 patients, homeostasis model assessment IR (HOMA-IR) was found to be significantly higher in patients with HCV-1-related steatosis than in those with HCV-3-related steatosis, or those without steatosis (14), and HOMA-IR was the only variable independently related to fatty liver in genotype 1 patients, whereas in genotype 3-infected patients, viral load was the only variable associated with fatty liver (14). Secondly, mice with a hepatocyte-specific deletion of the phosphatase and tensin homologue gene display histological steatohepatitis and lipid accumulation, but insulin hypersensitivity, as opposed to resistance (15). Thirdly, in familial hypobetalipoproteinaemia, defects in apolipoprotein B (Apo-B) expression are associated with steatohepatosis, but not glucose intolerance, as is seen in the matabolic syndrome (16). These findings thus suggest that lipid accumulation in the liver does not necessarily cause the hepatic IR that may occur in HCV-infected patients.

Multiple changes in liver fat occur in persons with NAFLD and NASH (Table 1). In a comparative study of patients with NAFLD (n=9), NASH (n=9) and controls matched for age, BMI, gender, ethnicity and glucose metabolism (n=9), total hepatic lipid content was significantly increased in the subjects with NAFLD and NASH (P<0.001), and this increase was driven primarily by an increase in the triacylglycerol (TAG) content (P<0.001) (17). Several other lipids were also increased in NAFLD and NASH compared with controls, whereas free fatty acids did not differ among the groups (Table 1). These investigators also found higher levels of diacylglycerol, increased free cholesterol and decreased phosphatidylcholine, as well as decreased levels of arachidonic acid, n-3 fatty acids, and an increase in the n-6:n-3 fatty acid ratio in NAFLD, all of which could play a role in the pathogenesis of this condition (17). Non-alcoholic fatty liver disease and NASH may also co-exist in patients with HCV infection. In one study, when patients with chronic HCV and co-existent NASH (n=22) were compared with those with only steatosis (n=49), and those without steatosis (n=49), the patients with co-existent NASH had significantly higher BMIs (30.64 vs 29.90 vs 27.33 respectively; P=0.008), had higher waist-to-hip ratios (0.97 vs 0.91 vs 0.87 respectively; P<0.001), were more commonly infected with genotype 3 (14 vs 12% vs 0% respectively; P=0.036) and had a higher degree of fibrosis (95.5 vs 75.5 vs 42.9% respectively; P<0.001) (3). These findings highlight the contribution of metabolic risk factors such as obesity and pre-existent fatty liver disease to the severity of HCV infection.

Table 1.   Hepatic lipid content (mean nmol/g of tissue) in control subjects, and patients with non-alcoholic fatty liver disease and non-alcoholic steatohepatitis
ParameterControls (n=9)NAFLD (n=9)NASH (n=9)
  • Adapted with permission from Puri et al. (17). Copyright © 2007 American Association for the Study of Liver Diseases.

  • *

    P<0.05 vs control (analysis of variance).

  • DAG, diacylglycerol; FFA, free fatty acid; FC, free cholesterol; LYPC, lysophosphatidylcholine; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; PC, phosphatidylcholine; PE, phosphatidylethanolamine; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; SFA, saturated fatty acid; TAG, triacylglycerol.

Total lipids15 97849 991*43 658*
Total SFA718922 889*18 756*
Total MUFA352719 61614 819*
Total PUFA51047328*9823*
Total n-39498621336
Total n-6413164248449*
FFA553359296115
DAG19224946*3304*
TAG13 609128 585*104 035*
FC753810 38212 862*
LYPC193618002239*
PC20 32115 322*16 874
PE14 59911 456*14 828

Hepatitis C virus effects on lipid abnormalities

Steatosis has been most closely linked to HCV-3. In our own study of chronic HCV-infected patients (n=70), we found steatotic patients to be more commonly infected with genotype 3 than with other genotypes (16 vs 8 patients; P=0.002), and the average steatosis scores were significantly higher for genotype 3 patients compared with genotype 1 patients (1.33 vs 0.172; P<0.001) (18). In this study, we also found a significant correlation between the intrahepatic HCV RNA level (but not the serum HCV RNA level) and the steatosis score for both the genomic (r=0.26; P=0.04) and the minus strand RNA (r=0.36; P=0.006), and the significance of the correlation with the minus (but not genomic) strand was further increased when patients with genotype 3 were considered (r=0.64; P=0.004), whereas it decreased when only genotype 1 patients were considered (r=0.32; P=NS) (18). Furthermore, we found that the response to α-interferon (IFN) treatment in genotype 3-infected patients (n=3) with moderate to severe steatosis was accompanied by a complete disappearance of steatosis. These findings suggest that steatosis is linked to the hepatic HCV RNA level and may represent the morphological expression of an HCV-related cytopathic effect (18). An association of steatosis with fibrosis has also been demonstrated in a meta-analysis of chronically infected patients with HCV (n=3068); in this large data set, we found a significant association of steatosis with fibrosis in multivariate analysis (P<0.001) (2). Similarly, when hepatic fibrosis was considered as the dependent variable, its presence was associated with steatosis (P<0.001), in addition to greater histological activity, male sex and older age. The association of steatosis and fibrosis remained across large subgroups of patients, including those with genotype 1 infection and those with a BMI of <25 kg/m2 (2). Taken together, these findings suggest a relationship among HCV replication level, the degree of steatosis and the degree of ensuing liver damage and fibrosis in patients infected with HCV.

Hepatic lipid abnormalities are a characteristic feature of HCV infection. In experimental studies using cultured cells, the expression of transfected HCV core proteins from various genotypes has been found to increase lipid accumulation (19). Lipid staining of transfected cells was significantly higher than that of non-transfected cells for all HCV genotypes tested; however, lipid droplet accumulation was most pronounced for genotype 3a (P<0.001 compared with other genotypes). In this study, triglyceride accumulation was significantly higher than that in non-transfected cells for genotypes 1b (P=0.015), 3a (P=0.003) and 3h (P=0.025), but not for other genotypes, and the level of triglyceride accumulation was again significantly higher for genotype 3a (19). These findings suggest that whereas core proteins from all HCV genotypes appear to induce lipid accumulation in hepatocytes, genotype-specific mechanisms may account for these effects, and selected genotypes (e.g. 3a) may be more efficient in promoting hepatic steatosis.

The effect of infection with various HCV genotypes is also evident on plasma lipids. In a study comparing lipid profiles of 155 patients with chronic HCV and 138 normal age- and sex-matched controls, it was found that chronic HCV patients exhibited significantly lower total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C) and LDL-cholesterol (LDL-C) compared with the controls (Fig. 1a) (20). This study also found that the mean serum LDL-C levels were significantly lower in patients with steatosis (95.6 mg/dl) compared with those without steatosis (108.7 mg/dl; P<0.05) and that genotype 3a patients had significantly lower serum TC, HDL-C and LDL-C levels compared with the other genotypes (Fig. 1b) (20). The results remained significant after controlling for age and sex, and genotype 3a was identified as a significant predictive variable for low cholesterol concentration in this study [odds ratio (OR), 6.96; P=0.0011]. These findings were also associated with a significantly higher incidence of steatosis in the genotype 3a patients relative to the other genotypes (61.7 vs 29.3%; P<0.0005). Another study of children with chronic HCV (n=16) found that TC levels were significantly lower than age- and sex-matched healthy controls (n=16) (21). These findings identify lower serum cholesterol concentrations as a hallmark feature of patients with chronic HCV infection.

Figure 1.

 (A) Serum lipid levels (mg/dl) in patients with chronic hepatitis C virus (HCV) vs healthy donors; (B) serum lipid levels (mg/dl) in HCV-3a patients vs other genotypes. TC, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol (20).

As noted earlier, among the genotypes of HCV, genotype 3a has been associated with a higher incidence of hepatocellular fat accumulation and steatosis. In a study of liver biopsies from 137 consecutive, IFN-naïve chronic HCV patients, hepatocellular fat was detected in significantly more patients with genotype 3a (n=47) compared with those with genotype 1 (n=67) or genotype 4 (n=23; Fig. 2) (22). The 3a group also had the highest incidence of micro/macrovascular steatosis, defined as >20% of hepatocytes containing fat droplets (Fig. 2). These alterations in lipid metabolism may be an important component in the pathogenesis and life cycle of HCV, as a response to antiviral therapy has been correlated with changes in selected lipid parameters in patients with chronic HCV. In the foregoing study, before the start of antiviral therapy, patients infected with HCV-3a had lower serum cholesterol levels than those infected with HCV-1 or HCV-4 (mean 146 vs 187 vs 172 mg/dl respectively), and at the end of therapy, serum cholesterol increased significantly in responders, whereas it remained unchanged in non-responders. Similar findings were observed at the end of follow-up, 6 months after the end of antiviral treatment (Fig. 3) (22).

Figure 2.

 Percentage of patients with hepatocellular fat (dark bars) and with marked micro/macrovascular steatosis (defined as >20% of hepatocytes containing fat droplets, light bars) according to hepatitis C virus genotype (22).

Figure 3.

 Change in serum cholesterol levels (mg/dl) in hepatitis C virus 3a patients in virological responders and non-responders at baseline, at the end of antiviral therapy and at the end of follow-up. NS, not significant (22).

In another recent retrospective study of 109 HCV-infected patients, a sustained response to antiviral treatment (IFN-α or pegylated IFN-α with or without ribavirin) occurred in 49% of the patients (n=53) (23). Subgroup analysis comparing responders with non-responders revealed significantly higher serum baseline TC, LDL-C and Apo-B levels in the former group (Fig. 4). In multivariate logistic regression, after adjustment for potential confounders, a 10 mg/dl increase in TC was found to confer a 3.02-fold higher chance of responding to treatment in this study (P<0.001). Similarly, the adjusted OR for Apo-B was 1.81 per 10 mg/dl increase (P<0.001). This study also found a lower probability of response to treatment with increasing BMI (adjusted OR, 0.73; P=0.03) and HCV-1 (adjusted OR, 0.20; P=0.008) (23). In the latter subgroup of patients infected with HCV-1, TC was the only significant predictive variable associated with treatment response (adjusted OR, 3.71; P=0.007). Although the potential mechanisms of TC facilitation of the antiviral response in this study are unclear, the investigators suggest that lipid-induced increases in certain cytokine levels, such as tumour necrosis factor α and interleukin 6, could further potentiate antiviral activity, and/or that a ‘regulatory effect’ of TC and Apo-B-rich lipoproteins could facilitate viral clearance by impeding HCV interaction with cell surface receptors (23).

Figure 4.

 Comparison of total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C) and apolipoprotein B (Apo-B) in hepatitis C virus patients who were responders (n=53) and non-responders (n=56) (23).

Another study has identified hypobetalipoproteinaemia as a risk factor for HCV-associated steatosis. In this study of 100 male patients with non-cirrhotic chronic HCV and without risk factors for steatosis, the cholesterol concentration was significantly lower (P<0.0001) compared with a reference male population, patients with chronic hepatitis B virus or patients with NAFLD (24). This study also found a low level of Apo-B in the study population (<70 mg/dl in 26% of the patients). In addition, 26% of the study population had hypobetalipoproteinaemia, as defined by cholesterol concentrations less than the 5th percentile of the reference population, and low Apo-B concentrations. The degree of steatosis was significantly higher in patients with hypobetalipoproteinaemia compared with those without (20 vs 7%; P<0.0001), and severe steatosis (i.e. involving >30% of hepatocytes) was observed significantly more frequently in patients with hypobetalipoproteinaemia than in other patients (38 vs 3%; P<0.0001) (24). With respect to HCV genotypes, both hypocholesterolaemia (P=0.005) and hypobetalipoproteinaemia (P=0.003) were significantly more common in patients with genotype 3a compared with 1a, 1b or other genotypes. Interestingly, following IFN therapy, prolonged responders (n=7) were found to show significant increases in both cholesterol and Apo-B levels at 1 year after the end of treatment, whereas no such increase was observed in non-responders (n=11) or patients who relapsed (n=11) (24). Taken together, these findings identify HCV-related changes in lipid metabolism as important components of HCV infection, particularly genotype 3a, and serum TC and Apo-B levels as important predictors of response to antiviral therapy.

Pathogenic mechanisms

Multiple mechanisms may account for the disruption of lipid metabolism in patients with chronic HCV infection. Genomic studies have identified numerous genes involved in the metabolism of lipids that can be affected by HCV infection. Transgenic mouse models and targeted gene disruption studies have further identified specific genes that are essential for steatosis and lipid accumulation induced by HCV; reactive oxygen species (ROS) appear to be important components of this process. Finally, molecular studies of the core protein itself are beginning to dissect the processes that may be involved in the association of genotype 3a with steatosis.

Hepatitis C virus effects on gene expression

A genome-wide analysis of host changes in gene expression (approximately 9000 genes) with HCV infection (genotype 1a) has been conducted in chimpanzees (25). The three animals examined in the study had different outcomes of infection (sustained clearance, transient clearance or persistence), and the gene expression analysis was conducted at multiple time points. Expression of >45 genes was positively correlated with viraemia onset, and many of these genes were involved in lipid metabolism. These included peroxisome proliferator-activated receptor-α (PPAR-α), whose expression was downregulated with an increase in viraemia, and lipid metabolism genes related to the sterol response element-binding protein (SREBP) signalling pathway, whose expression was upregulated during early viraemia (25). Other genes upregulated during early viraemia included lipase A, a gene involved in the hydrolysis of cholesterol esters and triglycerides, and UDP-glucose ceramide glucosyltransferase, which is involved in membrane glycosphingolipid biosynthesis (25). Fatty acid synthase was also expressed at higher levels in the chimpanzees, showing a sustained or a transient clearance in this study, although it was not correlated with HCV RNA levels. In another study, however, the expression of either genotype 3a or 1b core protein was found to activate the fatty acid synthase promoter, with the former genotype resulting in the greatest activation (26).

The impact of HCV infection on gene expression profiles has also been examined by transfection studies in hepatocytes. In HCV core protein (genotype 1b)-expressing cells, the triglyceride level was unchanged; however, the expressions of PPAR-α, multidrug resistance protein 3 (MDR3 in humans, Mdr2 in mice) and microsomal triglyceride transfer protein (MTP) were all significantly reduced at 48 h compared with mock-transfected cells (27). This pattern differed somewhat from that observed at 24 h, in that PPAR-α was unchanged, while PPAR-γ, MDR3, MTP and acyl coenzyme A (CoA) oxidase (ACO1 in humans, AOX in mice) were all upregulated compared with controls. By comparison, in vivo studies in mice demonstrated a 1.45-fold increase in the hepatic triglyceride content (P=0.044) and significantly reduced expression of PPAR-α, PPAR-γ, Mdr2, AOX and carnitine palmitoyl transferase-1 (CPT) with core expression, while MTP was not different (27). All these genes are involved in lipid metabolism, and taken together, these findings imply that the regulation of lipid metabolism by HCV core protein is important for the onset of HCV infection and in the HCV life cycle. They also provide an insight into the potential mechanisms of HCV-associated steatosis (25, 27).

Peroxisome proliferator-activated receptor-α

As described above, expression of the PPAR-α appears to be impaired with HCV infection. PPAR-α belongs to a family of nuclear receptors (28, 29), and in the liver, PPAR-α is known to modulate the expression of oxidative enzymes and fatty acid import into mitochondria through an induction of the carnitine palmitoyl acyl-CoA transferase 1 (CPT1A) gene (30, 31). In patients with HCV infection, expression of the PPAR-α gene in the liver was reduced by 86% compared with controls, and the expression of its target gene, CPT1A, was co-ordinately reduced by 85% (Fig. 5) (30). Expression of PPAR-α protein in HCV-infected livers was also significantly reduced compared with controls (P=0.009), whereas PPAR-γ, retinoid X receptor (RXR) and liver X receptor (LXR) were not different. Thus, hepatocytes infected with HCV display abnormally low levels of PPAR-α. Similarly, expression of the core protein in hepatoma cells was also found to reduce PPAR-α levels, but not the expression of the aforementioned receptors, and the induced expression of the CPT1A target gene by fenofibric acid was inhibited in core protein-expressing cells but not control cells (30). These findings suggest that HCV infection and expression of HCV core protein result in reduced PPAR-α expression, with a corresponding reduction in its transcriptional activity (as assessed by its target gene CPT1A).

Figure 5.

 mRNA expression of peroxisome proliferator-activated receptor-α (PPAR-α) and carnitine palmitoyl transferase (CPT) 1A genes in the liver of patients with hepatitis C virus infection (n=46) and healthy controls (n=40) (30). Levels were calculated by normalization to TATA box-binding protein mRNA expression.

Recent evidence of a direct role for PPAR-α in steatosis has been obtained from studies that combine mice with a homozygous deletion of PPAR-α (Ppara−/−), with mice expressing the HCV core protein in a liver-specific manner (HCVcpTg) (32). The results of these experiments found that only livers of Ppara+/+/HCVcpTg, but not those of Ppara+/−/HCVcpTg or Ppara−/−/HCVcpTg, developed marked triglyceride accumulation in the liver and steatosis, indicating that PPAR-α expression is essential for the development of steatosis (32). Additional results from this study indicated that the steatosis conferred by HCV core protein expression was associated with decreased mitochondrial fatty acid degradation caused by mitochondrial outer membrane breakdown and a disproportional increase in fatty acid uptake. Analysis of Ppara+/+/HCVcpTg mice also showed that PPAR-α was persistently and spontaneously activated, with corresponding activation of multiple target genes, and this was not observed in the Ppara+/−/HCVcpTg or Ppara−/−/HCVcpTg mice. Lastly, this study showed that expression of the core protein in and of itself was insufficient to cause steatosis, as evidenced by the lack of steatosis in mice heterozygous or homozygous for the deletion of PPAR-α (32).

Sterol response element-binding proteins

Sterol response element-binding proteins are transcription factors that are membrane bound to the endoplasmic reticulum and are involved in the regulation of enzymes that regulate cholesterol and fatty acid synthesis, as well as the cellular uptake of lipoproteins (33–35). At least three distinct isoforms have been described: SREBP1a, SREBP1c and SREBP2 (33). Studies examining the effect of HCV infection on SREBPs have shown that the genes for these proteins are transcriptionally induced, and their proteolytic cleavage is stimulated, by HCV infection in cultured cells (33). In addition, it has been shown that the phosphorylation and activation of these proteins via the mitogen-activated protein kinase and PI3-K-Akt signalling pathway is also stimulated by HCV infection, and the net result of this activation is the increased expression of target genes involved in hepatic lipid biosynthesis, such as fatty acid synthase (33). Other lipogenic genes induced by HCV infection in this model included hepatic hydroxymethyl glutaryl CoA reductase, squalene synthase, adenosine triphosphate citrate lyase, acyl-CoA carboxylase, fatty acid synthase and stearoyl-CoA desaturase, all of which were upregulated in infected cells relative to controls (33).

Activation of SREBP1c by HCV and steatosis further appears to require the presence of the proteasome activator PA28-γ. In mice with a targeted disruption of PA28-γ with HCV core gene expression (PA28γ−/−/CoreTg), there was an accumulation of core protein in the nucleus, and no steatosis was observed in the liver, in contrast to the livers of PA28γ+/+/CoreTg mice, in which a marked liver steatosis was observed, and that displayed predominantly cytoplasmic core protein expression (36). The results suggest that PA28-γ is required for steatosis and that HCV core protein is at least in part translocated to the nucleus during HCV infection and degraded in a PA28-γ-dependent manner. Further analysis of these mice revealed that increased transcription of the SREBP1c gene by HCV required the expression of the core gene as well as PA28-γ, and that the accumulation of core protein in the nucleus alone (i.e. as occurred in the PA28γ−/−/CoreTg mice) was insufficient to upregulate the SREBP1c gene (36). Lastly, this study showed that transcriptional upregulation of the SREBP1c promoter by the HCV core protein required complex formation between the nuclear receptors LXR-α and RXR-α, in addition to the expression of the PA28-γ gene. The results suggest a model whereby core protein potentiates the activation of the SREBP1c promoter by LXR-α/RXR-α in a PA28-γ-dependent fashion, the net result being activation of the SREBP1c fatty acid synthesis pathway and steatosis (36).

Microsomal triglyceride transfer protein

Steatosis in HCV-infected patients has also been associated with altered expression of Apo-B and altered MTP activity, which play an important role in the assembly and secretion of triglyceride-rich VLDL particles (37). Transgenic mice of the same lineage have been developed that express either HCV core protein (CoreTg), ApoAII (ApoAIITg) or both genes (CoreTg/ApoAIITg), and it has been shown that core-expressing mice have profoundly impaired VLDL-triglyceride and Apo-B secretion that can be abrogated by hepatic expression of ApoAII (37). In HCV core protein-expressing mice, VLDL particle size and abundance was markedly reduced compared with non-transgenic mice, whereas no such reduction was observed in the CoreTg/ApoAIITg mice. The CoreTg mice also exhibited a significant decrease in MTP protein activity vs non-transgenic mice (without altering MTP expression), and lipid peroxidation was increased in CoreTg mice compared with the CoreTg/ApoAIITg mice (37). The results thus indicate that HCV core protein inhibits VLDL assembly and secretion via an inhibition of MTP function, leading to intracytoplasmic storage of triglycerides and steatosis (37).

The findings of impaired MTP functioning as a result of HCV core protein expression have more recently been extended to humans. In the transgenic model described above, there was no impact of core protein expression on MTP expression at the RNA or the protein level; however, as noted earlier, expression of core protein (1b) has been found to decrease MTP mRNA expression in cell lines (27). In a study of 58 patients infected with various HCV genotypes, a highly significant (P=0.0017), inverse correlation was found for liver MTP mRNA levels and the degree of hepatic steatosis that was independent of genotype (Fig. 6), suggesting an important role for MTP in steatosis (38). Interestingly, on comparing genotype 3 with non-genotype 3 patients, a significantly lower hepatic MTP activity (but not MTP mRNA level) was observed (Table 2), and this was correlated with significantly lower Apo-B, TC, HDL and LDL levels than non-genotype 3 patients (Table 2). Additional evidence demonstrated that these results were not simply because of liver failure and resultant defects in protein synthesis, and the results suggest differences among HCV genotypes in their mechanisms of reducing MTP activity (38). Taken together, these findings suggest an important role for MTP activity in HCV-induced steatosis and provide a mechanism to explain, at least in part, the differences among HCV genotypes in inducing lipid abnormalities.

Figure 6.

 Correlation of microsomal triglyceride transfer protein (MTP) mRNA levels according to the degree of steatosis (grade 0=no steatosis). Levels were assessed by real-time polymerase chain reaction using GAPDH as an internal control. Expression is reported relative to the HepG2 MTP/GAPDH ratio. The inverse correlation between liver MTP and steatosis grade was significant (P=0.0017) (38). GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Table 2.   Microsomal triglyceride transfer protein gene expression and specific activity, apolipoprotein levels and lipid parameters in patients with non-genotype 3 hepatitis C virus and genotype 3 hepatitis C virus
 HCV non-3
(n=45)
HCV-3
(n=13)
P-value
  1. Adapted with permission from Mirandola et al. (38). Copyright © 2006 American Gastroenterological Association Institute.

  2. Apo, apolipoprotein; HCV, hepatitis C virus; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MTP, microsomal triglyceride transfer protein; NS, not significant.

MTP mRNA levels23.424.6NS
MTP-specific activity140011850.004
Apo-B (g/L)0.930.780.04
ApoAI (g/L)1.61.4NS
Total cholesterol (mg/dl)184.7148.50.0007
Triglycerides (mg/dl)107.983.3NS
HDL (mg/dl)58.545.40.009
LDL (mg/dl)103.585.50.05

Reactive oxygen species

Several lines of evidence indicate an important role for ROS in the pathogenesis of chronic HCV (39). Hepatic oxidative stresses during HCV infection can be mediated by multiple mechanisms, including aberrant mitochondrial functioning, stresses on the endoplasmic reticulum and immune cell-mediated damage (39). While there is ample evidence of both production and oxidative stresses with HCV infection, it remains unclear exactly how these processes contribute to the pathogenesis of the disease. Certainly, the modulation of mitochondrial function and redox status could be an important mechanism of viral evasion of host defences, while induced by HCV can be important contributors to hepatocellular damage, fibrosis and carcinogenesis (39). Expression of the core protein in stably transfected cell lines led to significantly increased production (3.4-fold; P<0.001) and increased lipid peroxidation products (two-fold; P<0.05) (40). Furthermore, mitochondria appeared to be the source of increased, as an inhibitor of mitochondrial electron transport abrogated the core protein-induced increase, and a fraction of core protein was associated with mitochondria in the perinuclear region (40). Similar observations were made when core protein was expressed in transgenic mice. There was a significant two-fold increase in hepatic lipid peroxidation products upon exposure to a carbon tetrachloride insult, which was not observed in non-transgenic mice, indicating an abnormal vulnerability to oxidant stress with core protein expression (40).

In another study of transgenic mice, it has been shown that core protein expression caused an ‘oxidative mitochondrial phenotype’ that was characterized by oxidation of mitochondrial glutathione and pyridine nucleotide pools, a defect in mitochondrial complex I-mediated electron transport and increased production (41). This study also demonstrated a direct association of core protein with the mitochondrial outer membrane, and that core protein caused an increase in calcium (Ca2+) influx, increased production, glutathione oxidation and reduced activity of complex I when incubated with normal mitochondria (41). More recently, a mechanism of core protein-induced Ca2+ influx into mitochondria involving the mitochondrial Ca2+ uniporter has been described (42). Core protein effects on the uniporter directly increase the ability of mitochondria to sequester Ca2+ in response to Ca2+ release from the endoplasmic reticulum, the net result being increased mitochondrial and a mitochondrial permeability transition.

There is evidence that may be an important mediator of mitochondrial injury and DNA damage in hepatocytes. Infection with HCV has been shown to cause a significant increase in and a decrease in mitochondrial membrane potential (43). In this model, this damage was mediated by both as well as nitric oxide (NO), as inhibitors of NO could abrogate these effects. The results of this study also showed that production leads to DNA damage (double-stranded breaks), increased production of lipid peroxides and the induction of proliferative transcription factors such as signal transducers and activators of transcription-3. These may be important in the generation of mutations leading to hepatocellular carcinoma, steatosis and increased proliferation in hepatocytes respectively (43). Evidence for a more direct role of oxidative stress in steatosis comes from the findings that the increased phosphorylation of SREBP1 and 2 and the resultant transactivation of their target genes (e.g. fatty acid synthase) in response to HCV infection, as described earlier, can be effectively inhibited with the anti-oxidant compound PTDC (33).

Molecular basis of core protein effects

The molecular basis for HCV core protein effects on lipid accumulation have recently been investigated in at least two studies that have analysed specific amino acid residues in the protein that are associated with steatosis. Jhaveri et al. (44) have utilized a well-characterized repository of samples from genotype 3a-infected patients who either had significant or no detectable steatosis in an effort to find core protein polymorphisms that correlated with steatosis. Using this approach, these investigators identified amino acid substitutions occurring at positions 182 and 186 of the HCV-3a core protein that caused increased lipid accumulation compared with non-steatosis clones when expressed in cell lines. The results of their pilot study thus suggest that this region of the core protein plays an important role in the regulation of lipid metabolism and trafficking, and this could explain in part the differences among HCV-3a isolates in their ability to cause steatosis (44).

In a separate study, Hourioux et al. (45) used sequence comparison to identify a position in domain II of the protein that is required for lipid accumulation. Domain II (amino acids 120–179) is thought to be important for the association of the core protein with the endoplasmic reticulum and with lipid droplets. In a sequence analysis utilizing all six genotypes and 25 subtypes, a phenylalanine residue (F) at position 164 was found to be the only residue specific to genotype 3a core proteins, being present in 24 of 33 genotype 3 sequences (73%), while a tyrosine residue (Y) was present at position 164 in all other genotypes (45). Mutation of this Y residue to F in the 1a genotype core protein (Y164F) demonstrated that, whereas both the wild-type 1a core protein as well as the Y164F mutant accumulated lipid when expressed in cells, the cumulative lipid droplet area was significantly larger with expression of the Y164F mutant, and this difference was not related to differences in expression or proteolytic cleavage (45). These authors suggest that the F residue may have a higher affinity for lipids than the Y residue at position 164, and these findings may in part explain the differences in the degree of steatosis induced by HCV-3a vs other genotypes. Interestingly, mutation of residue 164 has also been shown to be important in the activation of the fatty acid synthase promoter by the HCV core protein (26). When the reverse mutation of the 3a core protein was constructed (F164Y), the degree of transcriptional activation on the fatty acid synthase promoted was significantly diminished compared with the wild-type 3a core protein (P=0.00057), suggesting the importance of this residue not only in lipid accumulation but also in the transcriptional effects of the 3a core protein.

Another recent study has investigated in vitro lipid interactions with full-length core proteins isolated from patients with genotype 3a infection, with and without steatosis, and genotype 1b patients who were steatosis free (46). No motifs or amino acids related to the presence or absence of steatosis were found in this study; however, core protein was found to be co-localized with lipid droplets when expressed in cells, regardless of the HCV genotype, and this interaction was independent of the presence or absence of steatosis in core protein derived from 3a-infected patients (46). Interestingly, cells expressing the 3a core proteins contained significantly larger volumes of lipid droplets than either control cells (P<0.0001) or genotype 1b core protein-expressing cells (P=0.001), but there was no significant difference between the 3a core protein-expressing cells based on the presence or absence of steatosis; similar results were obtained when the size of the lipid droplets was considered. These findings suggest that core protein co-localization with lipid droplets is not the sole mechanism of steatosis associated with genotype 3a infection (46).

Conclusions

Multiple mechanisms may account for the development of steatosis in persons with chronic HCV infection, and these include effects mediated by HCV core protein expression, as well as comorbid metabolic conditions; these interactions are summarized in Figure 7. The accumulation of fat in the liver is a significant problem in patients infected with HCV and is associated with increased metabolic and cardiovascular risk; significant changes in hepatic lipid content, including elevation in total lipid and TAG, are observed in patients with NAFLD and NASH (Table 1). In patients with HCV infection, significantly lower TC, LDL-C and HDL-C are observed (Fig. 1a), especially among patients with genotype 3a infection (Fig. 1b), who show higher degrees of hepatocellular fat and steatosis (Fig. 2). These lipid disruptions have important correlations with response to antiviral therapy (Figs 3 and 4), implying an essential role for lipid metabolism in the HCV life cycle. The effect of HCV infection is mediated in large part by the HCV core protein, which shows the ability to cause lipid accumulation in hepatocytes, and there are genotype-specific polymorphisms (e.g. F164) that may explain the greater propensity for genotype 3a to cause steatosis. Core protein also has significant effects on the transcription of genes involved in lipid metabolism (e.g. CPT1A; Fig. 5), and the effect of HCV infection on the expression of genes such as MTP is well correlated with steatosis (Fig. 6). The association of core protein with mitochondria has significant effects on steatosis, including decreased fatty acid degradation, disproportional increases in fatty acid uptake and changes in Ca2+ influx and the mitochondrial outer membrane (32, 41, 42). These effects also appear to be important in ROS generation by core protein and its effects on lipid peroxidation and hepatocyte damage (37, 40, 43). Several genes including PA28-γ (36) and PPAR-α (30, 32) have been shown to be essential for HCV-related steatosis, and the targeted modulation of these molecular pathways, in addition to established therapies for reducing liver fat (i.e. insulin sensitizers and weight loss), may be an important component of future therapies for HCV-related steatosis.

Figure 7.

 Summary of hepatitis C virus (HCV) effects on hepatic steatosis. Expression of the HCV core protein in chronic HCV infection affects multiple processes leading to hepatic lipid accumulation, with selected genotypes (e.g. genotype 3a) more strongly associated with steatosis. The degree of steatosis may also be affected by co-existent conditions such as metabolic syndrome, type 2 diabetes and obesity, and also HCV replication levels. Core protein increases reactive oxygen species (ROS) and lipid peroxidation, leading to liver damage and fibrosis. Core protein also affects gene expression, leading to upregulation of genes involved in fatty acid metabolism, such as fatty acid synthase (FAS), and reduces microsomal triglyceride transport protein (MTP) function, leading to impairment of very low-density lipoprotein (VLDL)-triglyceride (TG) and apolipoprotein B (Apo-B) secretion, and ultimately reductions in plasma total cholesterol and Apo-B concentrations (hypobetalipoproteinaemia). Gene targeting studies have demonstrated the requirement for peroxisome proliferator-activated receptor-α (PPAR-α) and the proteasome activator PA28 γ in HCV core protein-mediated steatosis. NASH, non-alcoholic steatohepatitis; NAFLD, non-alcoholic fatty liver disease.

Acknowledgements

This work was supported by the Swiss National Science Foundation (grants number 3200B0-103727/1 and 320000-116544) to F. N.

Conflicts of interest: Both authors have received honoraria for participating in an Advisory Board held in December 2007.

Ancillary