Patatin-Like phospholipase domain-containing 3 I148M polymorphism, steatosis, and liver damage in chronic hepatitis C

Authors

  • Luca Valenti,

    Corresponding author
    1. Department of Internal Medicine, Università degli Studi, Fondazione IRCCS Ospedale Maggiore Policlinico “Ca' Granda” IRCCS, Milan, Italy
    • Luca Valenti, Centro Malattie Metabolichedel Fegato, Department of Internal Medicine, Universitàdegli Studi, Ospedale Maggiore Policlinico “Ca'Granda” IRCCS, Milano, Via F Sforza 35, 20122 Milano, Italy===

      Silvia Fargion, CentroMalattieMetabolichedelFegato, Department of Internal Medicine,UniversitàdegliStudi,OspedaleMaggiorePoliclinico“Ca'Granda” IRCCS, Milano , Via F Sforza 35, 20122 Milano, Italy===

    Search for more papers by this author
  • MariaGrazia Rumi,

    1. A.M. Migliavacca Center for Liver Disease, First Division of Gastroenterology, IRCCS, Milan, Italy
    Search for more papers by this author
  • Enrico Galmozzi,

    1. Department of Internal Medicine, Università degli Studi, Fondazione IRCCS Ospedale Maggiore Policlinico “Ca' Granda” IRCCS, Milan, Italy
    Search for more papers by this author
  • Alessio Aghemo,

    1. A.M. Migliavacca Center for Liver Disease, First Division of Gastroenterology, IRCCS, Milan, Italy
    Search for more papers by this author
  • Benedetta Del Menico,

    1. Department of Internal Medicine, Università degli Studi, Fondazione IRCCS Ospedale Maggiore Policlinico “Ca' Granda” IRCCS, Milan, Italy
    Search for more papers by this author
  • Stella De Nicola,

    1. A.M. Migliavacca Center for Liver Disease, First Division of Gastroenterology, IRCCS, Milan, Italy
    Search for more papers by this author
  • Paola Dongiovanni,

    1. Department of Internal Medicine, Università degli Studi, Fondazione IRCCS Ospedale Maggiore Policlinico “Ca' Granda” IRCCS, Milan, Italy
    Search for more papers by this author
  • Marco Maggioni,

    1. Pathology, Università degli Studi, Fondazione IRCCS Ospedale Maggiore Policlinico “Ca' Granda” IRCCS, Milan, Italy
    Search for more papers by this author
  • Anna Ludovica Fracanzani,

    1. Department of Internal Medicine, Università degli Studi, Fondazione IRCCS Ospedale Maggiore Policlinico “Ca' Granda” IRCCS, Milan, Italy
    Search for more papers by this author
  • Raffaela Rametta,

    1. Department of Internal Medicine, Università degli Studi, Fondazione IRCCS Ospedale Maggiore Policlinico “Ca' Granda” IRCCS, Milan, Italy
    Search for more papers by this author
  • Massimo Colombo,

    1. A.M. Migliavacca Center for Liver Disease, First Division of Gastroenterology, IRCCS, Milan, Italy
    Search for more papers by this author
  • Silvia Fargion

    Corresponding author
    1. Department of Internal Medicine, Università degli Studi, Fondazione IRCCS Ospedale Maggiore Policlinico “Ca' Granda” IRCCS, Milan, Italy
    • Luca Valenti, Centro Malattie Metabolichedel Fegato, Department of Internal Medicine, Universitàdegli Studi, Ospedale Maggiore Policlinico “Ca'Granda” IRCCS, Milano, Via F Sforza 35, 20122 Milano, Italy===

      Silvia Fargion, CentroMalattieMetabolichedelFegato, Department of Internal Medicine,UniversitàdegliStudi,OspedaleMaggiorePoliclinico“Ca'Granda” IRCCS, Milano , Via F Sforza 35, 20122 Milano, Italy===

    Search for more papers by this author
    • fax: +390250320296


  • Potential conflict of interest: Massimo Colombo: Grant and Research support: Schering-Plough, Roche, Novartis, Bristol Myers Squibb; Advisory committees: Schering-Plough, Roche, Novartis, Vertex, Bristol Myers Squibb, Gilead Sciences, Bayer; Speaking and Teaching: Schering-plough, Roche, Novartis, Vertex, Gilead Sciences, Bristol Myers Squibb, Bayer; Maria Grazia Rumi: Speaking and Teaching: Roche; Alessio Aghemo: Speaking and Teaching: Roche.

Abstract

Steatosis has been reported to negatively influence the natural history of chronic hepatitis C (CHC), but controversy remains over its causal role due to the confounding effect of adiposity, insulin resistance, and diabetes. The rs738409 C>G patatin-like phospholipase domain-containing 3 (PNPLA3) single nucleotide polymorphism (SNP), encoding for the I148M protein variant, influences liver fat without affecting insulin resistance and body composition. The aim of this study was to evaluate the effect of the rs738409 CG genotype on liver fat and fibrosis in CHC patients. We also explored the possible effect of PNPLA3 genotype on other steatosis-related complications, namely, treatment failure and hepatocellular carcinoma (HCC) development. To this end we considered two independent series of 325 and 494 CHC patients with available DNA and liver biopsy followed at tertiary referral centers in northern Italy. The rs738409 genotype was determined by a Taqman assay. The rs738409 GG genotype, observed in 10% of patients, was associated with steatosis independently of age, sex, body mass index (BMI), diabetes, alcohol intake, and viral genotype (odds ratio [OR] 1.90, 95% confidence interval [CI] 1.4-2.7; P < 0.001). The association with rs738409 genotype was confirmed for severe steatosis, was independent of alanine aminotransferase (ALT) and gamma-glutamyl transferase (GGT) values, and was observed in all viral genotypes but the 3. The rs738409 GG genotype was associated with fibrosis stage and cirrhosis (OR 1.47, 95% CI 1.2-1.9; P = 0.002), treatment response (n = 470; OR 0.63, 95% CI 0.4-0.8; P = 0.006), and HCC occurrence (n = 325; OR 2.16, 95% CI 1.3-3.6; P = 0.002), independently of confounders. Conclusion: The rs738409 PNPLA3 genotype influences steatosis development in CHC and is independently associated with cirrhosis and other steatosis-related clinical outcomes, such as lack of response to antiviral treatment and possibly HCC. (HEPATOLOGY 2011)

Chronic hepatitis C infection (CHC) affects more than 170 million individuals worldwide, representing a leading cause of liver-related mortality. CHC encompasses a wide spectrum of diseases, ranging from minimal disease to active hepatitis, which frequently progresses to cirrhosis and hepatocarcinoma.1 Due to the global epidemic of obesity, whose hepatic manifestation is nonalcoholic fatty liver disease (NAFLD), and to the direct effect of CHC on liver fat metabolism, liver steatosis has gained increasing attention as a modifier of CHC progression.

Indeed, steatosis occurs in more than half of hepatitis C virus (HCV)-infected patients, and has been associated with more aggressive histological features, faster progression of fibrosis, and poorer response to therapy.2-5 Both viral and host factors are believed to contribute to CHC-related steatosis.4, 6, 7

A direct cytopathic effect of HCV has been proposed based on the higher prevalence of steatosis in CHC than in other liver diseases, a specific association with genotype 3 infection,4, 8 and the induction of steatosis by viral proteins.9 Metabolic abnormalities including overweight and type 2 diabetes have also been linked to steatosis and fibrosis progression, in particular in nongenotype-3 patients.10, 11

Although most lines of evidence support a strong association between steatosis and fibrosis severity in CHC,12 doubts have been cast as to whether steatosis is the causative factor driving accelerated hepatic fibrogenesis, or rather is a simple marker associated with increased fat stores and insulin resistance, which would represent the culprits underlying disease progression.13, 14

Genetic host factors have been hypothesized to influence steatosis development and insulin resistance in CHC.15, 16 Recently, the adiponutrin/patatin-like phospholipase domain-containing 3 (PNPLA3) rs738409 C>G single nucleotide polymorphism (SNP), encoding for the I148M protein variant, has been recognized as a genetic determinant of liver fat content,17, 18 and to influence fibrosis severity in patients with NAFLD.19, 20 The mechanism whereby rs738409 influences liver fat is independent of body composition and insulin resistance,17, 21 but likely involves a decreased ability of the 148M PNPLA3 variant to regulate hepatic lipid metabolism.22 However, it is not known whether the rs738409 SNP influences the steatogenic effect of HCV and the progression of CHC.

The aim of this study was to assess the effect of rs738409 genotype on histologically assessed liver fat content and the severity of liver fibrosis in two cross-sectional series of Italian patients with CHC. To explore the possible clinical relevance of these findings, we also assessed the effect of PNPLA3 genotype on clinical outcomes previously associated with steatosis in CHC, such as treatment failure and hepatocellular carcinoma (HCC) development.

Abbreviations

ALT, alanine transaminase; AST, aspartate transaminase; BMI, body mass index; CHC, chronic hepatitis C; GGT, gamma-glutamyl transferase; HCC, hepatocellular carcinoma; HOMA-IR, homeostasis model assessment insulin resistance index; IL28B, interleukin 28B; NAFLD, nonalcoholic fatty liver disease; PNPLA3, patatin-like phospholipase domain-containing 3; SNP, single nucleotide polymorphism; SVR, sustained virological response.

Patients and Methods

Patients.

We considered 543 unrelated patients with CHC from the Metabolic Liver Disease Unit, Department of Internal Medicine (discovery series), diagnosed between January 1999 and January 2008, whose DNA samples were available. Fifty patients with coexistent causes of liver disease were excluded, including excessive alcohol intake (>60/40 g/day for more than 5 years for males/females, respectively), hepatitis B surface antigen (HBsAg) positivity, human immunodeficiency virus (HIV) infection, autoimmune hepatitis, hereditary hemochromatosis, and alpha1-antitrypsin deficiency. In all, 168 patients were excluded because of lack of or inadequate histological evaluation, whereas 325 patients were included in the study; the clinical features of these two groups were not significantly different.

Results were replicated in an additional series of 494 CHC consecutive patients who attended the A.M. Migliavacca Center for Liver Disease, at the First Division of Gastroenterology, who met the inclusion criteria and were prospectively enrolled from September 2009 to January 2010 (validation series). During the study period a total of 1,191 CHC patients attended the A.M. Migliavacca Center for Liver Disease but 697 patients could not be included as they lacked a liver biopsy, the histological sample was not adequate, or had coexistent liver diseases. The clinical features of excluded patients were not significantly different from those of patients included. Demographic and clinical features available for all patients included are shown in Table 1.

Table 1. Demographic and Clinical Features of 819 Italian Patients with CHC (325 from the Internal Medicine, Discovery Series, and 494 from the Gastroenterology Department, Validation Series), 261 Italian Patients with Biopsy Proven NAFLD, and 179 Italian Healthy Controls Without Steatosis19
 Discovery SeriesValidation SeriesNAFLD PatientsHealthy Controls
  • *

    P < 0.05 vs. controls. Parentheses indicate percent values for frequencies, SD for continuous variables. IGT: impaired glucose tolerance, IFG: impaired fasting glucose, NAFLD: nonalcoholic fatty liver disease, HCC: hepatocellular carcinoma.

  • Based on fatty liver index score36; na: not applicable.

Number325494261179
Sex F139 (42.7)*218 (44.1)*77 (29)38 (21)
Age years56.8 (14)*57.9 (12)*46.4 (11)48.4 (13)
BMI kg/m225.3 (4.3)24.4 (3.0)30.5 (7.8)*25.1 (2.7)
HDL cholesterol mg/dL51.0 (19)*55.1 (17)44.9 (12)*55.2 (13)
Triglycerides mg/dL105.1 (50)103.0 (47)155.2 (84)*90.1 (44)
Glucose mg/dL94.2 (27)*90.4 (22)*98.3 (27)*89.0 (10)
IGT-IFG or diabetes60 (18.5)*41 (8.3)*63 (24)*0
ALT UI/mL67.4 (68)*70.5 (71)*55.5 (41)*21.8 (7)
GGT UI/mL58.1 (66)*65.4 (101)*86.2 (108)*23.7 (16)
HCV genotype 1/2/3/4/5198/85/18/24/0 (61/26/6/8/0)270/137/37/49/1 (55/28/7/10/0)00
Steatosis222 (68.3)*339 (68.6)*261 (100)*0
Ishak stage 0/1/2/3/4/5/65/43/89/61/20/21/86 (2/13/27/19/7/6/26)5/86/135/99/61/45/63 (1/18/27/20/12/9/13)nana
Cirrhosis107 (32.9)*108 (21.9)*10 (4)0
HCC50 (15.4)6 (1.2)00

Insulin levels and homeostasis model assessment insulin resistance (HOMA-IR) index were available for 198 patients of the discovery series. All patients were Caucasians from Italy, except for 40 patients from northern Africa (all affected by HCV genotype 4). The prevalence of the rs738409 was not significantly different between these and the Italian patients. Additional adjustment of the analysis performed for ethnic background or exclusion of these patients did not modify the results.

The Italian healthy control group and the positive control group of 261 consecutive patients with NAFLD were previously described.23

The frequency distribution of the rs738409 SNP and clinical features of 261 patients with NAFLD and 179 healthy controls without liver enzymes and metabolic abnormalities, all from Milan,19 are also shown for comparison.

The study protocol was approved by the Institutional Review Board of the Ospedale Policlinico “Ca' Granda” IRCCS, Milan, Italy. Informed written consent was obtained from each patient and control subject and the study conforms to the ethical guidelines of the 1975 Declaration of Helsinki.

Histological Assessment.

Tissue sections were stained with hematoxylin and eosin, impregnated with silver for reticulin framework, and stained with trichrome for collagen. One expert pathologist unaware of clinical and genetic data reviewed all biopsies for steatosis grade and fibrosis stage according to Ishak stage.24 The minimum biopsy size was 1.7 cm and the number of portal areas 10. Steatosis was graded as 0: absent or <5%; 1: 5-33%; 2: 34-66%; 3: >66% of hepatocytes affected.

Clinical Endpoints.

Cirrhosis was defined as Ishak stage 5-6. Treatment success was retrospectively evaluated and defined as sustained virological response (SVR) to any therapeutic regimen containing interferons or pegylated interferons, whereas treatment failure was defined as lack of SVR to regimens based on pegylated interferons and ribavirin.

All cirrhotic patients from the discovery series (n = 107) underwent HCC surveillance. HCC was considered present when it was detected at diagnosis or at any time during follow-up (censored at November 1 2009) according to the Barcelona criteria.25

Genetic Analysis.

The rs738409 C>G SNP, encoding I148M, was genotyped by a 5′ nuclease Taqman assay (assay on demand for rs738409, Applied Biosystems, Foster City, CA) by personnel unaware of patient and control clinical status, as described.19 We genotyped 100% of patients evaluated. The IL28B rs12979860 genotype26 was determined by sequence allele specific polymerase chain reaction27 in 163 patients with available DNA samples included in the discovery series who completed a treatment course with peg-interferon and ribavirin.

Statistical Analysis.

Our sample had a >90% power of detecting an odds ratio (OR) of 2.0 for steatosis and cirrhosis, representing the primary objective of the study, in patients with CHC, according to genotype frequencies, with a significance of 5% (presupposing a recessive mode of inheritance). The results are expressed as means ± standard deviation and considered significant when P < 0.05 (two-tailed). Mean values were compared by analysis of variance (ANOVA), and frequencies by chi-square test.

The association between the PNPLA3 SNP and steatosis, cirrhosis, treatment outcome, and SVR was evaluated by logistic regression analysis adjusted for confounding variables, which included those selected a priori for their biological relevance plus those that were found associated with the outcome of interest at univariate analysis (specified in the Results section), under a recessive inheritance model, which was chosen based on previous results on the association between PNPLA3 genotype and fibrosis in NAFLD patients.19 The recessive one represented also the genetic model of inheritance for which minimal P-values were observed for these analyses, although the additive model28, 29 gave similar results. The association between PNPLA3 genotype and Ishak stage was evaluated by ordinal logistic regression analysis. Analyses were carried out with JMP 6.0 statistical analysis software (SAS Institute, Cary, NC).

Results

PNPLA3 rs738409 Genotype and Hepatic Steatosis.

The frequency distribution of the rs738409 PNPLA3 SNP was in Hardy-Weinberg equilibrium in all groups tested (Table 2), and was not significantly different between CHC patients without steatosis and controls. However, homozygosity for the rs738409 G allele was significantly more represented in CHC patients with than in those without steatosis, and the association was independent of age, sex, body mass index (BMI), daily alcohol intake, and the presence of diabetes and of HCV genotype 3 (Table 2).

Table 2. Frequency Distribution of the rs738409 C>G adiponutrin/PNPLA3 SNP in 819 Italian Patients with CHC Subdivided According to the Presence of Steatosis, 179 Italian Healthy Controls Without Steatosis (Reference for Patients with HCV Without Steatosis), and 261 Italian Patients with Biopsy-Proven NAFLD (Reference for Patients with HCV with Steatosis)
 PNPLA3 rs738409 C>G Genotype (Encoding for I148M)
 CC(148I/I)CG(148I/M)GG(148M/M)
  • P = 0.0002 for the frequency distribution of adiponutrin/PNPLA3 rs738409 genotype between patients with and without steatosis in the overall series of CHC patients.

  • *

    Adjusted for age (years), sex, BMI (Kg/m2), alcohol intake (> or < 30 g/day), presence of diabetes, and viral genotype (genotype 3 vs. other genotypes). Parentheses indicate percent values. OR: odds ratio, CI: confidence interval. P < 0.005 for the frequency distribution of the rs738409 C>G PNPLA3 SNP between CHC patients with and without steatosis in both the discovery and validation series, respectively.

Discovery series (n=325)163 (50.1)127 (39.1)35 (10.8)
 Steatosis (n=222)100 (45.0)91 (41.0)31 (14.0)
 No steatosis (n=103)63 (61.2)36 (34.9)4 (3.9)
 Adjusted OR (95% C.I.)*Reference0.81 (0.5-1.3)2.17 (1.1-5.1)
Validation series (n=494)261 (52.8)183 (37.1)50 (10.1)
 Steatosis (n=339)169 (49.8)128 (37.8)42 (12.4)
 No steatosis (n=155)92 (59.4)55 (35.5)8 (5.1)
 Adjusted OR (95% C.I.)*Reference0.81 (0.5-1.3)1.92 (1.2-3.5)
CHC overall (n=819)424 (51.8)310 (37.8)85 (10.4)
 Steatosis (n=561)269 (48.0)219 (39.0)73 (13.0)
 No steatosis (n=258)155 (60.0)91 (35.3)12 (4.7)
 Adjusted OR (95% C.I.)*Reference0.80 (0.6-1.1)2.13 (1.4-3.4)
NAFLD (n=261)108 (41.4)116 (44.4)37 (14.2)
Healthy controls (n=179)118 (65.9)56 (31.3)5 (2.8)

As expected, in CHC patients the rs738409 genotype was not significantly associated with age, sex distribution, viral genotype, BMI, diabetes, HOMA-IR index, excessive alcohol intake, cholesterol, alanine transaminase (ALT), aspartate transaminase (AST), and gamma-glutamyl transferase (GGT) levels.

The prevalence of steatosis according to rs738409 and viral genotypes is shown in Fig. 1A,B. The rs738409 G allele was significantly associated with steatosis in the overall series and in patients affected by CHC genotype 1, 2, and 4 (for steatosis grade >1), but not in those affected by CHC genotype 3. Variables associated with the presence of steatosis at multivariate analysis are shown in Table 3, upper panel. Besides the rs738409 GG genotype, steatosis was significantly associated with older age, male sex, higher BMI, presence of diabetes, and viral genotype 3. Interestingly, the rs738409 genotype was also strongly associated with moderate/severe steatosis (grade >1; Table 3, bottom panel).

Figure 1.

Prevalence of steatosis (A) and of steatosis >33% (B) according to the rs738409 C>G PNPLA3 SNP and viral genotype in 819 Italian patients with CHC. Percent values are shown on the Y axis. P-values for the frequency distribution of steatosis across PNPLA3 genotypes are shown (chi-square test).

Table 3. Clinical and Genetic Factors Independently Associated (Logistic Regression Analysis) with the Presence of Steatosis (Upper Panel), and Steatosis Moderate/Severe (Grade > 1; Bottom Panel) in 819 Italian Patients with CHC
VariablesDiscovery Series (n=325)Validation Series (n=494)Overall Series (n=819)
OR (95% CI)P valueOR (95% CI)P valueOR (95% CI)P value
  1. OR: odds ratio, CI: confidence interval (multivariate analysis results are presented), SNP: single nucleotide polymorphism.

Steatosis Present (Any Grade)
 Age (years)1.02 (1.00-1.04)0.0151.01 (1.00-1.03)0.0561.02 (1.01-1.02)0.003
 Sex (F)0.71 (0.54-0.92)0.0110.92 (0.74-1.13)0.430.82 (0.69-0.96)0.015
 BMI (Kg/m2)1.08 (1.03-1.15)0.0061.10 (1.05-1.15)<0.0011.08 (1.04-1.12)<0.001
 Alcohol intake >30/20 g/day M/F0.85 (0.61-1.18)0.331.24 (0.78-2.13)0.380.97 (0.75-1.23)0.79
 Diabetes1.29 (0.92-1.83)0.142.08 (1.26-4.01)0.0111.47 (1.13-1.97)0.006
 Viral genotype 36.04 (2.06-31.6)0.0093.15 (1.75-7.11)0.0013.77 (2.19-7.71)<0.001
 rs738409 SNP GG vs. CG + CC2.07 (1.23-4.03)0.0141.73 (1.17-2.71)0.0101.90 (1.39-2.73)<0.001
Steatosis Grade > 1
 Age (years)0.98 (0.95-1.01)0.251.01 (0.99-1.03)0.341.00 (0.99-1.02)0.81
 Sex (F)2.00 (1.25-3.31)0.0050.89 (0.69-1.15)0.370.98 (0.79-1.19)0.81
 BMI (Kg/m2)1.00 (0.91-1.09)0.981.14 (1.05-1.24)0.0011.06 (1.01-1.12)0.023
 Alcohol intake >30/20 g/day M/F0.82 (0.71-2.10)0.421.34 (0.84-2.01)0.190.98 (0.71-1.32)0.92
 Diabetes0.82 (0.39-1.45)0.551.51 (1.04-2.19)0.0281.13 (0.84-1.50)0.40
 Viral genotype 32.62 (1.37-4.96)0.0032.67 (1.82-3.94)<0.0012.53 (1.86-3.45)<0.001
 rs738409 SNP GG vs. CG + CC2.14 (1.22-3.64)0.0052.36 (1.70-3.28)<0.0012.09 (1.62-2.67)<0.001

The rs738409 GG genotype was independently associated with the presence of steatosis (OR 1.75, 95% confidence interval [CI] 1.28-2.49; P = 0.001) and with moderate/severe steatosis (OR 2.16, 95% CI 1.67-2.78; P < 0.001) also when ALT and GGT levels were added to the model. After the exclusion of patients with obesity, excessive alcohol intake, or genotype 3 infection (628 patients fulfilled these criteria), at logistic regression model considering as independent variables age, sex, BMI, diabetes, and rs738409 genotype, the rs738409 GG genotype was still independently associated with steatosis (OR 3.18, 95% CI 1.65-6.69; P = 0.001), and with steatosis >1 (OR 2.13, 95% CI 1.59-2.82; P < 0.001).

PNPLA3 rs738409 Genotype and Hepatic Fibrosis.

There was no significant association between rs738409 genotype and histological necroinflammatory activity (histological grade according to Ishak24). However, the PNPLA3 rs738409 genotype was significantly associated with Ishak fibrosis stage (P = 0.0087, estimate 0.22 per G allele) at ordinal logistic regression analysis adjusted for age, sex, BMI, presence of diabetes, and excessive alcohol intake (Table 4).

Table 4. Independent Predictors of Ishak Fibrosis Stage at Multivariate Ordinal Logistic Regression Analysis In 819 Patients with CHC
VariablesEstimate (95% CI)P value
  1. CI: confidence interval.

Age (years)0.05 (0.04-0.06)<0.0001
Sex (F)-0.09 (-0.21-0.04)0.189
BMI0.02 (0.01-0.08)0.017
Diabetes0.10 (0.08-0.46)0.0054
Alcohol intake>30/20 g/day M/F0.19 (-0.01-0.39)0.057
rs738409 SNP per G allele0.26 (0.08-0.44)0.0057

The prevalence of cirrhosis according to the rs738409 SNP and referral center is shown in Fig. 2A. The frequency distribution of rs738409 genotype was significantly different between patients with and without cirrhosis (P = 0.003). Independent predictors of cirrhosis are shown in Table 5A: the rs738409 GG genotype was associated with cirrhosis independently of confounding factors (OR 1.47, 95% CI 1.15-1.87; P = 0.002). Homozygosity for the rs738409 G allele was still associated with cirrhosis when the presence of histological steatosis was added to the model (OR 1.37, 95% CI 1.1-1.8; P = 0.01).

Figure 2.

Prevalence of cirrhosis (A), sustained virological response (B), according to the rs738409 C>G PNPLA3 SNP and referral center in 819 Italian patients with CHC, and cumulative incidence of HCC in 325 Italian patients with HCC according to the rs738409 C>G PNPLA3 SNP (C). Percent values are shown on the Y axis. P-values for the frequency distribution of cirrhosis, treatment response, and HCC across PNPLA3 genotypes are shown (chi-square test).

Table 5. Clinical and Genetic Factors Independently Associated (Logistic Regression Analysis) with the Presence of Cirrhosis (n=819; A), SVR (n=470; B), and HCC (n=325; C) in Italian Patients with CHC
(A) Cirrhosis
VariablesDiscovery Series (n=325)Validation Series (n=494)Overall Series (n=819)
OR (95% CI)P valueOR (95% CI)P valueOR (95% CI)P value
Age (years)1.04 (1.03-1.06)<0.0011.03 (1.01-1.05)0.0011.03 (1.02-1.05)<0.001
Sex (F)0.92 (0.71-1.20)0.550.83 (0.81-1.30)0.830.98 (0.82-1.16)0.98
BMI (Kg/m2)0.90 (0.86-0.94)<0.0010.93 (0.89-0.97)0.0020.92 (0.89-0.95)<0.001
Alcohol intake >30/20 g/day M/F1.00 (0.73-1.37)0.982.37 (1.58-3.59)<0.0011.50 (1.18-1.90)0.001
Diabetes1.34 (0.99-1.81)0.0571.49 (1.04-2.11)0.0251.53 (1.22-1.90)<0.001
rs738409 SNP GG vs. CG + CC1.63 (1.10-2.41)0.0141.37 (1.01-1.87)0.041.47 (1.15-1.87)0.002
(B) SVR
VariablesDiscovery Series (n=192)Validation Series (n=278)Overall Series (n=470)
OR (95% CI)P valueOR (95% CI)P valueOR (95% CI)P value
Age (years)0.97 (0.95-0.99)0.0011.01 (0.99-1.02)0.540.99 (0.97-1.00)0.054
BMI (Kg/m2)1.00 (0.96-1.05)0.820.97 (0.91-1.02)0.190.99 (0.96-1.03)0.76
Cirrhosis0.54 (0.39-0.74)0.0020.93 (0.68-1.25)0.630.73 (0.59-0.89)0.003
Viral genotype 1 + 4 vs. 2 + 30.67 (0.51-0.87)0.00320.33 (0.24-0.43)<0.00010.45 (0.37-0.54)<0.001
rs738409 SNP GG vs. CG + CC0.51 (0.28-0.83)0.0120.64 (0.39-1.01)0.060.63 (0.44-0.86)0.006
(C) HCC
VariablesDiscovery Series (n=325)    
OR (95% CI)P value    
  1. OR: odds ratio, CI: confidence interval, SNP: single nucleotide polymorphism, SVR: sustained virological response, HCC: hepatocellular carcinoma.

Age (years)1.05 (1.02-1.09)0.005    
Sex (F)0.80 (0.55-1.16)0.23    
Diabetes1.60 (1.07-2.40)0.022    
Alcohol intake>30/20 g/day M/F0.98 (0.62-1.55)0.98    
Cirrhosis4.77 (3.15-7.72)<0.001    
SVR0.56 (0.30-0.95)0.042    
rs738409 SNP GG vs. CG + CC2.16 (1.33-3.59)0.002    

In patients younger than 60 years (median value), the prevalence of GG genotype was 22% in patients with versus 8% in those without cirrhosis (P = 0.01), whereas in patients older than 60 years the prevalence of GG genotype was 14% in patients with versus 10% in those without cirrhosis (P = 0.24).

PNPLA3 rs738409 Genotype and Treatment Response.

Information on treatment response was available for 192 patients from the discovery series (59% of the whole series; 107 SVR, 56%; 18, 9% had undergone a previous treatment), and for 278 patients from replication series (59%; 124 SVR, 45%; 123, 44% had undergone a previous treatment). Of the 231 patients who achieved SVR, 26 (11.3%) were treated with interferon-based regimens, whereas 205 (88.7%) with peg-interferon and ribavirin.

The prevalence of SVR according to rs738409 genotype and referral center is shown in Fig. 2B. There was a nonsignificantly lower SVR rate in patients carrying the rs738409 GG genotype than in those with the CG or CC genotypes (P = 0.09). In genotype 1 and 4 patients (n = 286), the rs738409 GG genotype was significantly associated with a lower SVR rate than CG and CC genotypes (2/20, 10% versus 95/266, 36%; P = 0.02).

Independent predictors of SVR are shown in Table 5B. The rs738409 GG genotype was independently associated with SVR both in the two independent as well as in the overall series of patients (OR 0.63, 95% CI 0.44-0.86; P = 0.006).

Homozygosity for the IL28B rs12979860 C allele was associated with increased SVR rate in the 163 patients treated with peg-interferon and ribavirin with available data included in the discovery series (39/56, 67.8% versus 53/107, 49.5%, P = 0.019). Addition of rs12979860 genotype to the model presented in Table 5B (discovery series) did not affect the association between rs738409 genotype and SVR (OR 0.62, 95% CI 0.49-0.89). There was no evidence of linkage between the rs12979860 and rs738409 genotypes.

PNPLA3 rs738409 Genotype and Hepatocellular Carcinoma.

HCC was diagnosed in 50 (15.4%) patients included in the discovery series. The cumulative incidence of HCC was 17/163 (10.4%) in patients carrying the CC, 21/127 (16.6%) in patients carrying the CG, and 12/35 (34.3%) in patients carrying the GG rs738409 genotype (Fig. 2C; P = 0.002).

Because, due to the study design, it was not possible to conduct a prospective analysis, we conducted an exploratory analysis evaluating factors associated with the occurrence of HCC at any time between diagnosis and the end of follow-up. Independent predictors of HCC are shown in Table 5C. The rs738409 GG genotype was associated with HCC independently of confounders (OR 2.16, 95% CI 1.33-3.59; P = 0.002).

Interestingly, in 6/50 cases (12%) HCC occurred in patients without a previous diagnosis of cirrhosis. Homozygosity for the rs738409 GG genotype was associated with HCC both in patients in follow-up for cirrhosis (10/15, 66.7% versus 34/92, 40.0%; P = 0.046) and in patients without a previous diagnosis of cirrhosis (2/20, 10% versus 4/198, 2%; P = 0.038). It should be noted that histological cirrhosis could not be excluded in all these patients at the time of HCC development.

The association between the rs738409 G allele, steatosis, and liver fibrosis was further confirmed in a case-control analysis in patients matched for age, sex, viral genotype, diabetes, obesity, and excessive alcohol intake (presented in Supporting Table 1).

Discussion

In this study we evaluated in a large series of Italian CHC patients the rs738409 PNPLA3 SNP, previously reported to influence liver fat without affecting body composition and insulin resistance,17, 30 and observed that this genetic variant (1) affects steatosis development, (2) is independently associated with fibrosis and cirrhosis, and (3) may influence treatment response and HCC occurrence.

Importantly, the association between rs738409 genotype and steatosis was independent of known risk factors, such as age, sex, BMI, diabetes, alcohol intake, and biochemical markers such as ALT and GGT levels. As previously reported for NAFLD,19, 31 homozygosity for the minor rs738409 G allele conferred the strongest risk of steatosis.

In line with the prediction that, if steatosis causes fibrosis progression in CHC, then the rs738409 SNP should also be associated with advanced fibrosis, the rs738409 G allele was independently associated with Ishak stage, and the 10% of CHC patients carrying the rs738409 GG genotype had a roughly 50% higher risk of cirrhosis, a very important clinical outcome, than CG and CC subjects. This association was present in the overall series, and in the two independent patient sets, and was independent of known confounding factors. The selective association between PNPLA3 and cirrhosis in younger patients is consistent with an expected relatively greater effect of inherited factors on fibrosis progression in patients with a shorter duration of disease. Interestingly, as observed in Italian and UK patients with NAFLD,19 the association between rs738409 genotype and the severity of fibrosis was also independent of the measurable effect of steatosis, suggesting that besides neutral triglycerides accumulation, which in experimental models does not affect the progression of liver damage,32 PNPLA3 genotype affects inflammation and fibrogenesis, or opposite that in the presence of severe liver fibrosis, when steatosis usually disappears, PNPLA3 genotype reflects an impairment in hepatic fat metabolism better than histology. However, unlike than in adult patients with NAFLD,19 we did not observe any association between PNPLA3 genotype, liver enzymes, and necroinflammatory grade in CHC patients, although it is difficult to objectively evaluate histological alterations typical of NASH in the context of HCV-induced inflammation, and these hallmarks were not specifically searched for in this study. Moreover, PNPLA3 genotype was associated with steatosis severity, NASH features, and fibrosis independently of ALT levels also in Italian children with NAFLD.20

Because the rs738409 is an inherited nonmodifiable factor, which is selectively associated with steatosis by interfering with hepatic lipid metabolism,22 the present results suggest that, despite recent conflicting reports,13, 14 nonviral genetic steatosis has a causal role in determining fibrosis progression in CHC.

Despite the cross-sectional design of the present study, which does not allow drawing definitive conclusions, our data suggest also that the rs738409 SNP may influence important clinical outcomes such as treatment response and HCC development. Indeed, in line with the previously reported negative influence of steatosis on treatment outcome,3, 33 rs738409 host genotype was associated with a reduced rate of SVR achievement independently of viral genotype and of cirrhosis. Interestingly, rs738409 genotype has been reported to partially explain the increased susceptibility to NAFLD of Latinos,17 who have reduced rates of response to antiviral therapy.34 Whether evaluation of rs738409 genotype together with other host genetic factors, such as IL28B SNPs,26 may help tailor future HCV treatments should be further evaluated, especially because in the discovery series rs738409 genotype was associated with treatment outcome independently of IL28B rs12979860 genotype.

However, we do acknowledge that the two study cohorts of patients might not be considered ideal to assess the influence of rs738409 genotype on SVR. Indeed, the enrolment of many treatment-experienced patients, as well as the different therapeutic regimens and HCV genotypes, coupled with the fact that the study was limited to patients who had undergone a liver biopsy, might have accounted for a selection bias; this calls for external validation of our data. Moreover, this study was conducted in a selected group of patients from Italy followed in tertiary care centers, and therefore it is not known whether these findings also apply to individuals in the general community with CHC and to patients of different ethnicity.

Due to the inclusion criteria and the evaluation of patients with severe metabolic comorbidities, and possibly different socioeconomic background, patients who developed HCC were present almost only in the discovery series, and the analysis was thus limited to this group. Furthermore, because of the study design it was not possible to conduct a prospective analysis in this series. Notwithstanding, our analysis was able to confirm the association between HCC occurrence and known risk factors, and showed a more than 2-fold higher risk of HCC in subjects homozygous for the rs738409 GG genotype independently of confounders, in line with the emerging evidence that HCC risk is influenced by the presence of steatosis in CHC.35-37 If confirmed in larger longitudinal studies, the association between rs738409 genotype and HCC may help identify patients at increased risk among nonresponders to viral treatment and patients with cirrhosis who, even after HCV eradication, are still at a consistent risk of HCC development, especially in the presence of steatosis.37

Thus, besides contributing to clarify the pathogenesis of liver damage progression in CHC, the association between PNPLA3 rs738409 GG genotype and steatosis-related clinical outcomes might identify a subgroup of CHC patients with an adverse prognosis. This is particularly relevant in view of the increasing prevalence of this genotype with increasing severity of the disease (e.g., 4% in CHC without steatosis, 13% in CHC with steatosis, 24% in HCC patients). The evaluation of the effect of drugs, such as glitazones,38 acting on hepatic fat accumulation on the natural history of CHC in rs738409 GG subjects might represent another possible implication of our results. However, large prospective studies enrolling unselected series of patients are required to precisely define the relevance of rs738409 genotype on CHC natural history, i.e., the progression of liver fibrosis and HCC development, and treatment response.

In conclusion, PNPLA3 genotype influences steatosis development in CHC, and, likely because of the induction of NASH-like biological alterations in the liver parenchyma, is independently associated with cirrhosis and other steatosis-related clinical outcomes, such as lack of response to antiviral treatment and HCC. Whether evaluation of rs738409 genotype can help tailor treatment and follow-up of CHC patients should be assessed in prospective studies.

Acknowledgements

We thank Roberta D'Ambrosio for help in reevaluating histological data, Cristina Bertelli and Erika Fatta for clinical assistance, and Serena Rovida and Paolo Maggioni for help in collecting clinical data.

Ancillary