PNPLA3 genetic variants determine hepatic steatosis in non-obese chronic hepatitis C patients

The influence of patatin-like phospholipase domain-containing 3 (PNPLA3) genetic variants in the development of liver steatosis in Asian chronic hepatitis C patients remains elusive. A total of 1018 biopsy-proven chronic hepatitis C patients were enrolled for evaluation. The proportions of PNPLA3 rs738409 GG genotype carriage were 7.8% (44/563), 15.8% (58/367) and 19.3% (17/88) in patients with no (liver fat content <5%), mild (5–33%) and moderate/severe (>66%) hepatic steatosis, respectively (trend P < 0.001). Stepwise logistic regression analysis revealed that the strongest factor independently associated with steatosis was the carriage of the PNPLA3 rs738409 GG genotype (odds ratio [OR]/95% confidence intervals [CI]:2.34/1.557–3.515, P < 0.001). Among the patients with BMI < 24 kg/m2, carriage of the rs738409 GG genotype was the only factor associated with hepatic steatosis (OR/CI:3.44/1.824–6.500, P < 0.001). PNPLA3 genetic variants had minimal effects on hepatic steatosis among overweight or obese patients. Compared to patients with BMI<24 kg/m2/non-GG genotype, those with BMI>24 kg/m2/GG genotype were more likely to have hepatic steatosis (OR/CI:3.87/2.292–6.524, P < 0.001). In conclusions, both PNPLA3 genetic variants and BMI played important roles in hepatic steatosis among Asian chronic hepatitis C patients. However, the genetic effect was mainly restricted to non-obese patients.

variable attributive factors ranges between 40% and 80%, which is approximately twofold higher than in the general population [1][2][3][4] . The presence of hepatic steatosis may promote liver fibrosis progression in the natural course of CHC 5 . In addition, it may also determine the treatment efficacy of interferon based anti-viral therapy 6 . Two major mechanisms account for the high prevalence of hepatic steatosis in CHC, which include the direct cytopathic effect of the hepatitis C virus genotype 3 (HCV-3) viral protein and the indirect effect of metabolic derangement in patients with HCV non-3 infection 3,7,8 .
Beyond the issue of virological and environmental factors, host genomes also play a role in hepatic steatosis. Genome-wide association study (GWAS) has demonstrated that a single nucleotide polymorphism (SNP) of patatin-like phospholipase domain-containing 3 (PNPLA3) gene was associated with nonalcoholic fatty liver disease (NAFLD) 9 . In addition, PNPLA3 genetic variants have also been shown to be associated with hepatic steatosis in CHC patients [10][11][12] . However, the result was not consistent in different cohorts 13 . This finding raised the question of whether the association between PNPLA3 SNP and hepatic steatosis in CHC patients varies across ethnicities. Meanwhile, it has been reported that the determination of PNPLA3 SNPs in NAFLD was not universal in the same population with different metabolic profiles 14 . Taken collectively, the impact of PNPLA3 genetic variants on liver steatosis in Asian CHC patients with different characteristics has never been studied. We herein recruited a large CHC cohort with histologically proven steatosis and well-characterized demographics, and we aimed to determine the association of PNPLA3 genetic variants with hepatic steatosis in an Asian CHC population. Importantly, we also sought to determine whether the influence of this gene differs among subpopulations with different characteristics.

Methods
A total of 1,018 CHC patients who received pre-antiviral evaluation were consecutively recruited in a medical center and two core regional hospitals in Taiwan from 2001 to 2013. Anti-HCV antibodies were detected using a third-generation, commercially available enzyme-linked immunosorbent assay kit (AxSYM 3.0, Abbott Laboratories, Chicago, IL, USA). Serum HCV RNA was detected using qualitative real-time polymerase chain reaction (PCR) (COBAS AMPLICOR Hepatitis C Virus Test, ver. 2.0; Roche, Branchburg, NJ, USA, detection limit: 50 IU/ml) and quantification branched DNA assay (Versant HCV RNA 3.0, Bayer, Tarrytown, New Jersey, USA; quantification limit: 615 IU/ml) before 2011. The HCV genotypes were determined using the Okamoto method before 2011 15 . Both the HCV RNA and genotype were detected using real-time PCR assay (RealTime HCV; Abbott Molecular, Des Plaines IL, USA; detection limit: 12 IU/ml) since 2011 16 . All of the patients received liver biopsies before initiating antiviral therapy. Patients with current or past history of alcohol abuse (≥ 20 g daily) were excluded in the current cohort. The liver histology was graded and staged according to the scoring system described by Knodell and Scheuer 17 . Hepatic steatosis was evaluated with an H&E stain with the definition of no (liver fat content < 5%), mild (5-33%), moderate (33-66%) and severe (> 66%) hepatic steatosis. The diagnosis of diabetes was based on 1) laboratory tests with twice the fasting plasma glucose levels >126 mg/dL or hemoglobulin A 1 C > 6.5%, or 2) medical history of previously established diagnosis of diabetes. All patients gave written informed consent before enrollment. The study was approved by the ethics committee of Kaohsiung Medical University Hospital and was performed according to the guidelines of the International Conference on Harmonization for Good Clinical Practice. PNPLA3 genotyping. The PNPLA3 rs738409 was selected as the candidate SNP, and the genotype was determined using the methods described previously 18 .
Statistical analyses. The frequency was compared between groups using the χ 2 test with the Yates correction or Fisher's exact test. Group means, presented as the mean values and standard deviations, were compared using the analysis of variance and Student's t test or the Mann-Whitney U test. The serum HCV RNA levels were expressed after logarithmic transformation of the original values. The influence of PNPLA3 in liver steatosis was calculated using dominant (genotype CC vs. CG + GG) and recessive (genotype GG vs. CG + CC) genetic models of inheritance. A stepwise logistic regression analysis was performed to evaluate the independent factors associated with steatosis by analyzing the co-variants with P values <0.05 in the univariate analysis. The statistical analyses were performed using the SPSS 12.0 statistical package (SPSS, Chicago, IL, USA). All statistical analyses were based on two-tailed hypothesis tests with a significance level of p < 0.05.

Results
Patients. The mean age of the patients included in the study population was 51.8 of years and 56.6% of the patients were male ( Table 1). The majority of the patients weighed between 18.5 kg/m 2 and 30 kg/m 2 . Hepatic steatosis was observed in 455 (44.7%) of the CHC patients. The proportion of no, mild, moderate and severe hepatic steatosis was 55.3% (n = 563), 36.1% (n = 367), 7.8% (n = 79) and 0.9% (n = 9), respectively. PNPLA3 rs738409 CC, CG, and GG genotypes accounted for 41.5%, 46.9% and 11.7% of the population, respectively. The majority of patients were infected with HCV-1 and HCV-2, whereas only one patient in the current cohort was infected with HCV-3. Factors associated with hepatic steatosis in CHC patients. In the univariate analysis, the factors associated with fatty liver included older age, high body mass index (BMI), the presence of diabetes, a high r-glutamyltransferase (r-GT) level and carriage of the PNPLA3 rs738409 GG genotype, using the recessive model. CHC patients with hepatic steatosis had numerically higher proportions of CG/GG genotype carriage when compared to those without, as determined by using the dominant model (P = 0.08). Six-hundred and ninety four patients had available cholesterol and triglyceride (TG) data. Patients with steatosis had significantly higher TG level (109 + 64 mg/dL vs. 96 + 49 mg/dL, P = 0.004). Stepwise logistic regression analysis was performed to evaluate factors independently associated with hepatic steatosis. If the variable of TG was not taken into consideration, the strongest factor independently associated with steatosis was carriage of the PNPLA3 rs738409 GG genotype (odds ratio [OR]/95% confidence intervals [CI]:2.34/1.557-3.515, P < 0.001), followed by body mass index (BMI, OR/CI: 1.12/1.082-1.167, P < 0.001) and age (OR/CI:1.02/1.004-1.028, P = 0.007) by using the recessive model. By using the dominant model, the factors associated with steatosis included carriage of the PNPLA3 rs738409 CG/GG genotype (OR/CI:1.31/1.013-1.703, P = 0.04), BMI (OR/CI: 1.13/1.084-1.169, P < 0.001) and age (OR/ CI:1.02/1.004-1.028, P = 0.006). If the variable of TG was taken into account, the strongest factor associated with liver steatosis remains the carriage of PNPLA3 rs738409 GG genotype (OR/CI: 2.37/1.408-3.983, P = 0.001) ( Table 2). The PNPLA3 genotype distribution did not differ between patients with or without available TG level either by recessive model (GG genotype: 11.1% vs. 13.0%, P = 0.39) or dominant model (CG/GG genotype: 58.8% vs. 58.0%, P = 0.82).
Role of PNPLA3 genetic variants in determining hepatic steatosis among patients with different BMIs. Because BMI and PNPLA3 genetic variants are both important determinants of hepatic  (Table 4). For obese patients, hepatic steatosis was only associated with a higher r-GT level in the univariate analysis (Table 3), although no factors were associated with hepatic steatosis in obese CHC patients in the multivariate analysis. The impact of the PNPLA3 rs738409 genotype on fatty liver varied among patients with different BMIs (Supplementary Fig 1). We further explored the role of the SNP in patients with different degrees of hepatic steatosis. The proportion of PNPLA3 rs738409 GG genotype carriage was 7.8% (44/563), 15.8% (58/367) and 19.3% (17/88) in patients with no, mild, and moderate/severe hepatic steatosis, respectively (trend P < 0.001) ( fig. 1).

Interaction of PNPLA3 rs738409 genotype and BMI in hepatic steatosis. Both the PNPLA3
genetic variants and BMI determined hepatic steatosis. We further analyzed their interactions in contributing to fatty liver. There was a significantly increased proportion of mild and moderate/severe hepatic steatosis in patients with BMI> 24 kg/m 2 and the PNPLA3 rs738409 GG genotype compared to those with BMI< 24 kg/m 2 and/or non-GG genotype (P < 0.001) (fig. 2)  steatosis when compared to their counterparts. Both PNPLA3 genetic variants and BMI played important roles in hepatic steatosis in CHC patients. Importantly, we identified that the host genetic effect was mainly restricted to non-obese patients and not obese patients. For patients with BMI < 24 kg/m 2 , the carriage of the PNPLA3 rs738409 GG genotype increased the risk of hepatic steatosis 3.4-fold when compared to those individuals carrying the C allele. PNPLA3 participates in the restoration of lipid homeostasis upon aberrant intracellular lipid accumulation. The determination of the role of the PNPLA3 SNP in NAFLD was established in 2008 9 . Later, several reports have linked the genetic variants to other spectrums of liver disease such as HBV 20 and HCV infection 3,7,8,21 . Due to the direct steatotogenic effect of the HCV-3 protein, the determination of the host genome in hepatic steatosis has been restricted to non-HCV-3 infection 12,14 . Most of the studies have originated from the West, where the genetic effect has been fully explored in Caucasians. Nevertheless, Nakamura et al. recently reported that there was no association between the PNPLA3 rs738409 genotype and fatty changes in sonography in Japanese patients with HCV-1 and HCV-2 infections 13 . It is therefore imperative to validate the effect of PNPLA3 in another cohort with different ethnicities and patient characteristics in which the phenotype is clearly defined by liver biopsy. Approximately half of the   CHC patients had hepatic steatosis in the current cohort, which was similar to the prevalence in some Western reports 1-3 . However, the mean BMI was only 24.9 kg/m 2 and less than one tenth of the patients had hepatic steatosis > 33% in the current population. We confirmed that the PNPLA3 genetic variants consistently play a role in hepatic steatosis in Asian patients with HCV-1 and HCV-2 infections. The determined power was similar to the reports from the West where patients homozygous for the risk G allele had an approximately 2-fold higher risk for hepatic steatosis 12,21 As HCV infection increases the risk for liver steatosis and metabolic derangement, the relationship between HBV infection and hepatic steatosis remains conflicting 22,23 . We identified that lean CHC patients with HBV co-infection had a lower proportion of liver steatosis compared to those with HCV mono-infection. The finding was in agreement with some reports that HBV infection protects against fatty liver rather than promoting it 23 . This finding may be attributed to a lower frequency of metabolic disorders in HBV carriers 23 . In addition, hepatic steatosis may accelerate hepatitis B surface antigen clearance 24,25 . Whether the mechanism supports the inverse relationship between HBV infection and steatosis awaits further confirmation. Hepatic steatosis may promote liver fibrosis progression, although we did not observe this association in the current study. Because fibrotic tissue may replace liver fat content as the disease progresses, the linkage might be masked from cross-sectional observation 18,26 .
Interestingly, we observed that the role of the PNPLA3 genetic variants in hepatic steatosis was particularly enhanced in non-obese patients. The odds ratios further increased compared with those in the whole population, and the PNPLA3 genotype was the strongest predictor for hepatic steatosis in non-obese subjects. The reason why some lean CHC patients developed fatty liver was not fully understood. The current study in part provided a clue from the perspective of the host genetic profile. . In contrast, the PNPLA3 SNP did not independently determine hepatic steatosis in overweight or obese patients. The higher the BMI, the less the genetic composition impacted the degree of hepatic steatosis.   The current study focusing on CHC patients may echo the report that the PNPLA3 rs738409 GG genotype increases the risk of NAFLD in the general population without metabolic disorder 14 .
The current study was limited by the absence of metabolic profiles, which may interfere with the final results. Nevertheless, we demonstrated the associations between the PNPLA3 SNP and other simple demographic characteristics with hepatic steatosis. These findings should facilitate more direct clinical relevance in the interpretation of the association.
In conclusion, we demonstrated that the PNPLA3 genetic variants determined the risk of development of hepatic steatosis in Asian CHC patients. However, the effect was not universal and was mainly restricted to non-obese patients. Whether the latter finding is generalizable to other ethnicities awaits further validation.