SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Inherited factors play a major role in the predisposition to nonalcoholic fatty liver disease (NAFLD), and the rs738409 C→G polymorphism of PNPLA3/adiponutrin, encoding for the isoleucine-to-methionine substitution at residue 148 (I148M) protein variant, has recently been recognized as a major determinant of liver fat content. However, the effect of the rs738409 polymorphism on the severity of liver fibrosis in patients with NAFLD is still unknown. In this study, we considered 253 Italian patients, 179 healthy controls, and 71 family trios with an affected child with NAFLD. Analyses were replicated in 321 patients from the United Kingdom. The rs738409 polymorphism was determined by TaqMan assays. Liver histology was scored according to Kleiner et al. Hepatic expression of genes regulating liver damage was assessed by real-time polymerase chain reaction in 52 patients. The rs738409 GG genotype was more prevalent in patients than in controls (14% versus 3%, adjusted odds ratio [OR] = 3.29, 95% confidence interval [CI] = 1.8-6.9), and in the family study, the G allele was overtransmitted to affected children (P = 0.001). In Italian and United Kingdom patients, adiponutrin genotype influenced alanine aminotransferase levels and the severity of steatosis. Adiponutrin genotype was associated with the expression of genes involved in the steatosis-related liver damage, including the proapoptotic molecule Fas ligand. In the whole series combined, adiponutrin genotype was associated with steatosis grade >1 (OR = 1.35, 95% CI = 1.04-1.76), nonalcoholic steatohepatitis (OR = 1.5, 95% CI = 1.12-2.04), and fibrosis stage >1 (OR = 1.5, 95% CI = 1.09-2.12), independent of age, body mass index, and diabetes. Adiponutrin genotype demonstrated a dose effect with heterozygote risk intermediate between CC and GG homozygotes. Conclusion: In patients with NAFLD, adiponutrin rs738409 C→G genotype, encoding for I148M, is associated with the severity of steatosis and fibrosis and the presence of nonalcoholic steatohepatitis. (Hepatology 2010;51:1209–1217)

Nonalcoholic fatty liver disease (NAFLD), affecting 20%-34% of the U.S. and European populations,1, 2 is currently considered a hepatic manifestation of the metabolic syndrome3 and is associated with increased mortality due to cardiovascular and liver diseases.4, 5

Hepatocellular fat accumulation and insulin resistance are the key factors in NAFLD pathophysiology and are deeply entangled with the progression of liver disease. Insulin resistance in the adipose tissue results in an excess of free fatty acid (FFA) supply to the liver, which may induce lipotoxicity, oxidative stress, and apoptosis, thus leading to nonalcoholic steatohepatitis (NASH).6 NASH is considered the progressive form of NAFLD and is characterized by necroinflammatory changes, ballooning degeneration, and/or fibrosis. Insulin resistance per se may be involved in the onset of liver damage, and the presence of the metabolic syndrome predicts the severity of liver disease.4, 7

Inherited factors play a major role in susceptibility to the metabolic syndrome and NASH. Predisposition to NASH appears to cluster with the heritability of metabolic risk factors within families, and ethnic differences have been reported.8 Single-nucleotide polymorphisms (SNPs) in genes involved in inflammation, oxidative stress, and fibrogenesis have been associated with the severity of liver damage in NAFLD,9-11 but data are contradicting and explain only a minority of fibrosis variations.

Recently, genomewide association studies identified an SNP in adiponutrin/patatin-like phospholipase-3 (PNPLA3), rs738409 C→G, encoding the I148M (isoleucine-to-methionine substitution at residue 148) protein variant, as the strongest genetic determinant of liver fat and alanine aminotransferase (ALT) levels.12-14 The 148M variant was more common in Hispanics, who are known to be at increased risk for NASH and cryptogenic cirrhosis.12 Importantly, the I148M SNP influenced liver fat independently of body mass, dyslipidemia, and insulin resistance. In addition, PNPLA3 variants have also been associated with alcoholic liver disease and alcohol-related cirrhosis.15

Adiponutrin/PNPLA3, which belongs to the patatin-like phospholipase family, is expressed in the liver and adipose tissue and possesses acyl hydrolase activity.16 Adiponutrin expression is increased by carbohydrate feeding and a Western-type diet,17, 18 and has lipase activity against triglycerides and acylglycerol transacylase activity, thus being likely involved in energy mobilization and storage in lipid droplets.19 Recently, it was reported that the 148M adiponutrin allele is a loss-of-function variant that predisposes an individual to steatosis by decreasing triglyceride hydrolysis in hepatocytes.20 However, although the adiponutrin genotype has previously been reported to be associated with steatosis severity and NASH in a very small series of biopsied patients,21, 22 the relationship between adiponutrin expression and liver damage, as well as the effect of the I148M SNP on liver fibrosis, have not been evaluated in adequately powered samples.

Because the degree of steatosis has previously been reported to influence the severity of liver damage in NAFLD,23 the aim of this study was to evaluate the relationship between the adiponutrin I148M polymorphism and the severity of liver fibrosis.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Patients.

We analyzed DNA from 574 unrelated patients with biopsy-proven NAFLD diagnosed between January 1999 and January 2008. These included 190 Italian patients who underwent liver biopsy because of persistently abnormal liver enzymes/serum ferritin or a long-lasting history of steatosis associated with severe metabolic abnormalities, 63 Italian severely obese patients who were found to be affected by NAFLD at routine percutaneous liver biopsy performed during bariatric surgery, and 321 adult patients from Newcastle, UK. Other causes of liver disease were excluded, including increased alcohol intake (>30/>20 g/day for males/females, respectively), as confirmed by at least one family member or friend and carboxydesialylated transferrin determination, viral and autoimmune hepatitis, hereditary hemochromatosis, and alpha1-antitrypsin deficiency. Part of this group had previously been described.9 The available demographic and clinical features are shown in Table 1. All patients included in the study had a fatty liver index >50.24

Table 1. Demographic and Clinical Features of Patients with 179 Italian Healthy Control Subjects with Normal Liver Enzymes and Metabolic Parameters and 253 Patients with NAFLD, Including 190 Patients Who Underwent Liver Biopsy Because of Persistently Abnormal Liver Enzymes and/or Metabolic Abnormalities, and 63 Patients Found to Have NAFLD at Routine Liver Biopsy During Bariatric Surgery
FeatureItalian ControlsItalian NAFLD PatientsUK NAFLD Patients
  1. Values in parentheses are % values. IFG, impaired fasting glucose; IGT, impaired glucose tolerance; NASH, nonalcoholic steatohepatitis.

Number179253321
Female sex38 (21)76 (30)123 (38)
Age (years)48.4 ± 1346.4 ± 1149.5 ± 12
BMI (kg/m2)25.1 ± 2.730.5 ± 7.834.0 ± 5.3
LDL cholesterol (mg/dL)118.7 ± 29132.5 ± 43127.0 ± 43
HDL cholesterol (mg/dL)55.2 ± 1344.9 ± 1247.0 ± 13
Triglycerides (mg/dL)90.1 ± 44155.2 ± 84248.3 ± 178
Glucose (mg/dL)89.0 ± 1098.3 ± 27121.4 ± 54
HOMA-IR2.7 ± 1.64.3 ± 3.46.4 ± 6.2
IGT-IFG or diabetes063 (25)120 (37)
ALT (IU/mL)21.8 ± 755.5 ± 4180.8 ± 64
GGT (IU/mL)23.7 ± 1686.2 ± 108130.0 ± 184
NASH (%)0124 (49)184 (57)
Fibrosis stage F0/F1/F2/F3/F4 n (%)123/85/25/10/10 (49/33/10/4/4)124/58/53/51/35 (39/18/16/16/11)

In addition, we tested 71 Italian family trios with both parents alive and an index child with biopsy-proven fibrotic NAFLD. We used transmission disequilibrium testing to look for preferential transmission of rs738409 C/G alleles to the affected children. This approach is not subject to the potential confounding effects inherent in case-control studies and is significantly more powerful at detecting true associations.10, 25

Controls.

The Italian control group included 179 geographically-matched, age-matched, and sex-matched Italian subjects out of a larger series of 482 blood donors from Northern Italy who were selected because of the lack of clinical and biochemical evidence of liver and metabolic disease and no alcohol abuse (<30/20 g/day in males/females). We excluded subjects with ALT >35/30 IU/mL in males/females, gamma-glutamyl transferase >35 IU/mL, body mass index (BMI) >28, abdominal circumference >100 cm, glucose levels ≥100 mg/dL, triglycerides ≥150 mg/dL, high-density lipoprotein (HDL) ≤45/55 in males/females or a fatty liver index >35, a value with high specificity to rule out NAFLD in the general population.24

The study protocol was approved by the Institutional Review Board of the Ospedale Policlinico MaRE IRCCS, Milan, Italy, and by Local Research Ethics Committee in Newcastle, UK. 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 visualizing reticulin framework, and stained with trichrome for visualizing collagen. One expert pathologist unaware of clinical and genetic data reviewed all biopsies for fibrosis stage at each clinical center. The severity of steatosis, features of NASH, and fibrosis was assessed according to Kleiner et al.26 NASH was considered to be present when steatosis, ballooning, and lobular inflammation were present. The minimum biopsy size was 1.7 cm and the number of portal areas was 10.

Genetic Analysis.

DNA was extracted from peripheral blood collected at the time of liver biopsy by the phenol-chloroform method. Success rate in extracting DNA was 100% for each study group. The adiponutrin/PNPLA3 rs738409 C→G SNP, encoding I148M, was genotyped by a TaqMan assay (assay on demand for rs738409; Applied Biosystems, Foster City, CA) by personnel unaware of clinical status of patients and controls. Post–polymerase chain reaction (PCR) allelic discrimination was carried out measuring allele-specific fluorescence on the Opticon2 detection system (MJ Research, Waltham, MA). Random samples were confirmed by direct genotyping that provided concordant results in all cases. Samples from both patients with NAFLD and controls were included in all batches analyzed, and quality controls were performed to verify the reproducibility of the results. Valid genotypic data were obtained for more than 99% of subjects analyzed.

Determination of Messenger RNA Levels of Adiponutrin and Genes Regulating Liver Metabolism and Damage.

In 52 patients who underwent bariatric surgery for whom adequate liver biopsy samples collected at the time of surgery were available (comprising 42 subjects with NAFLD included in the genotype frequency study, and 10 subjects with steatosis <5% and no histological abnormalities), RNA was isolated by the Trizol reagent (Invitrogen, Carlsbad, CA), digested with deoxyribonuclease I, and quality evaluated by measuring the absorbance ratio at 260/280 nm (≥1.8) and by electrophoresis. First-strand complementary DNA was synthesized using equal amounts (0.5 μg) of total RNA, with the SuperScript VILO complementary DNA synthesis kit (Invitrogen, Carlsbad, CA). The messenger RNA (mRNA) levels were analyzed by quantitative real-time PCR (qRT-PCR) with SYBR Green chemistry (Fluocycle II SYBR green mix; Euroclone, Pero, Italy). All the reactions were performed in triplicate with the Opticon2 qRT-PCR system (MJ Research, Waltham, MA). In patients with and without NAFLD, we evaluated the PNPLA3/adiponutrin mRNA levels, and we correlated the PNPLA3 I148M genotype with the expression of insulin receptor (INSR), which regulates insulin signaling and is down-regulated in the metabolic syndrome,27 steroid regulatory element binding protein 1c (SREBP1c), which regulates lipogenesis,28 peroxisome proliferator-activated receptor-α (PPAR-α), which regulates lipolysis and is decreased in NASH,29 and Fas ligand (FASL), which is a proapoptotic molecule induced by insulin resistance.30, 31 Primers are shown in Table 2. Results were normalized for β-actin and 18S RNA, which were chosen as a control because of stable expression among different samples.

Table 2. Primers Used for Evaluation of Hepatic Gene Expression Analysis
GenePrimers
  1. FASL, Fas ligand; Fw, forward primer; INSR, insulin receptor; PPAR-α, peroxisome proliferator-activated receptor-α; Rv, reverse primer; SREBP1c, steroid regulatory element binding protein 1c.

Adiponutrin (PNPLA3)Fw 5′-TTTACAGTGGCCTTATCCCT-3′
Rv 5′-GAAAGTTCGTGGACTTGACT-3′
β-actinFw 5′-GGCATCCTCACCCTGAAGTA-3′
Rv 5′-GGGGTGTTGAAGGTCTCAAA-3′
FASLFw 5′-GCACTTTGGGATTCTTTCCA-3′
Rv 5′-CCTCCATTTGTCTGGCTCAT-3′
INSRFw 5′-TCAACGGGCAGTTTGTCGAA-3′
Rv 5′-GCTGAAGTTCACACAGCGCC-3′
PPAR-αFw 5′-ATGGCATCCAGAACAAGGAG-3′
Rv 5′-TCCCGTCTTTGTTCATCACA-3′

Statistical Analysis.

Results are expressed as means ± standard deviation. Mean values were compared by analysis of variance (ANOVA) and post-hoc analysis or Wilcoxon test, when appropriate, and frequencies by Fisher's exact test, and chi-squared test for trend, when appropriate. Variables were correlated by the Spearman's rho test.

The association between the I148M variant and the presence of metabolic abnormalities, NASH, and fibrosis was evaluated by multivariate logistic regression analysis. In analyzing the association between I148M and steatosis and fibrosis, we compared allele and genotype frequencies in those with none or mild steatosis/fibrosis (0 and 1/3 for steatosis; 0 and 1/4 for fibrosis) versus those with moderate/severe steatosis/fibrosis (2 and 3/3 for steatosis; 2, 3, and 4/4 for fibrosis). Analyses were carried out with JMP 6.0 statistical analysis software (SAS Institute Inc, Cary, NC).

Results

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Adiponutrin Genotype Influences the Risk of NAFLD.

The frequency distribution of the rs738409 C→G adiponutrin genotype, which was in Hardy-Weinberg equilibrium in Italian and UK patients and in controls, is shown in Table 3. The frequency distribution of the mutant G allele was significantly higher in Italian patients with NAFLD than in geographically-matched, age-matched, and sex-matched healthy controls with normal liver enzymes and metabolic parameters and a normal fatty liver index (P < 0.0001; Table 2). In Italian adult subjects, the adiponutrin GG genotype was associated with an increased risk of NAFLD compared to the CC genotype independently of age, sex, and BMI (odds ratio = 3.29, 95% confidence interval = 1.8-6.9; Table 3).

Table 3. Frequency Distribution of the rs738409 C→G adiponutrin/PNPLA3 Genotype, Encoding for the I148M Protein Variant, in 253 Italian Patients with NAFLD and 197 Italian Healthy Subjects with Normal Liver Enzymes and Metabolic Parameters (P < 0.0001 for the Frequency Distribution of the G Allele Between Italian Patients and Controls), and in 321 NAFLD Patients from the United Kingdom
 PNPLA3 rs738409 C→G genotype (encoding for I148M)
CC (148I/I)CG (148I/M)GG (148M/M)
  • *

    For Italian patients versus Italian controls. Adjusted for age, sex, and BMI. CI, confidence interval; OR, odds ratio.

Italian patients103 (40.7)114 (45.1)36 (14.2)
Italian controls118 (65.9)56 (31.3)5 (2.8)
Adjusted OR*Reference0.803.29
95% CI0.5-1.21.8-6.9
UK patients142 (44.2)140 (43.6)39 (12.1)

The frequency distribution of the rs738409 C→G adiponutrin genotype, as well as the prevalence of the GG genotype, was very similar between Italian and UK patients with NAFLD (P = nonsignificant; Table 3).

In the family study conducted in 71 Italian trios, 56 families were “informative” in that one or both parents were heterozygous for the SNP (a total of 76 parents were heterozygous for the rs738409 C→G SNP). In these families, the variant allele was transmitted on 52 of 76 (68%) possible occasions, whereas the wild-type C allele was transmitted on only 24 of 76 (32%) occasions (P = 0.001).

Influence of Adiponutrin Genotype on Metabolic and Biochemical Parameters.

The effect of the rs738409 C→G adiponutrin genotype, encoding for the I148M protein variant, on clinical features and metabolic and biochemical parameters in Italian patients is shown in Table 4. In patients with NAFLD, the GG genotype was associated with higher LDL, and both the CG and GG genotypes were associated with lower HDL cholesterol, although it was not associated with age, sex distribution, and BMI. The GG genotype was associated with significantly higher fasting insulin levels and homeostasis model assessment of insulin resistance (HOMA-IR) score, but not with hyperglycemia or type 2 diabetes. In addition, the G allele (148M variant) was associated with higher ALT in a dose-dependent manner (P = 0.0007).

Table 4. Association of the rs738409 C→G adiponutrin/PNPLA3 SNP, Encoding for the I148M Protein Variant, with Metabolic Features and Liver Damage in 253 Italian Patients with NAFLD
CharacteristicI148M PNPLA3 genotypeP
CC (148I/I)CG (148I/M)GG (148M/M)
  • Values in parentheses are % values. M, male; F, female; NAS, nonalcoholic steatohepatitis activity score.

  • P at ANOVA,

  • *

    P < 0.05 versus the I/I PNAPLA3 genotype,

  • P < 0.05 versus the I/M genotype.

Number10311436 
Age (years)47 ± 1146 ± 1246 ± 100.59
Sex (F)32 (31)33 (28.9)11 (30.6)0.93
BMI (kg/m2)30.7 ± 830.7 ± 829.8 ± 70.10
LDL (mg/dL)129 ± 41129 ± 46152 ± 35*0.02
HDL (mg/dL)48 ± 1342 ± 12*43 ± 8*0.004
Triglycerides (mg/dL)145.6 ± 74163.5 ± 94158.5 ± 780.30
Glucose (mg/dL)96 ± 1999.9 ± 3598 ± 150.71
Fasting insulin (IU/mL)15.7 ± 916.7 ± 1023.1 ± 18*0.01
HOMA-IR3.9 ± 3.54 ± 2.75.2 ± 4.7*0.07
IFG-IGT / diabetes n (%)16 (15)/9(8)11 (10)/17(15)7 (19)/3(8)0.31
Hypertension43 (41.7)49 (43.0)17 (47.2)0.81
ALT (IU/mL)47 ± 3056 ± 44*77 ± 56*0.0007
GGT (IU/mL)91 ± 10783 ± 11973 ± 760.68

Broadly similar data were obtained in the UK cohort (Supporting Information Table 1). Patients carrying the GG genotype had significantly higher ALT levels compared to those carrying the CC or GG genotypes (76.0 ± 54 IU/L CC, 74.5 ± 56 CG, 123.0 ± 103 GG; P < 0.0001 at ANOVA, P < 0.05 for GG versus either CC or CG genotypes). In contrast to the data in Italian patients, the rs738409 genotype was not associated with abnormal HDL or LDL levels.

Association Between Adiponutrin Genotype and Liver Damage.

At histology, and as expected from previous studies, using an additive model for adiponutrin genotype, carriage of the G allele was an independent predictor for the presence of grade 2/3 steatosis together with BMI, sex, and diabetes in the combined cohort. In the UK patients alone, both genotype and age were predictors (Table 5A). More importantly, the adiponutrin genotype was strongly associated with the presence of NASH (Table 5B; P = 0.007 in the combined cohort, P = 0.03 in the UK cohort, and P = 0.07 in the Italian cohort) and stage >1 fibrosis (Table 5C; P = 0.01 in the combined cohort, P = 0.06 in the UK cohort, and P = 0.04 in the Italian cohort) independent of BMI, diabetes and steatosis (for NASH), and age, BMI, diabetes, steatosis, and NASH (for fibrosis). Furthermore, there was a dose effect of adiponutrin genotype, in that CG heterozygotes had a prevalence of NASH and fibrosis intermediate between CC and GG homozygotes (Fig. 1A,B).

Table 5. Association of the adiponutrin/PNPLA3 rs738409 Genotype (Encoding for the 148 I/M Protein Variants) and Liver Damage, as Evaluated by Multivariate Logistic Regression Analysis, in 574 Adult Italian and UK Patients with NAFLD
VariablesUK (n = 321)Italy (n = 253)Combined (n = 574)
OR (95% CI)P ValueOR (95% CI)P ValueOR (95% CI)P Value
  • CI, confidence interval; OR, odds ratio; NS, not significant.

  • *

    Based on an additive model for rs738409 genotype.

A. Steatosis (0+1 versus 2+3)
Genotype*1.6 (1.10-2.37)0.011.3 (0.88-1.9)0.11.35 (1.04-1.76)0.02
Age (years)0.97 (0.95-0.99)0.020.99 (0.97-1.02)0.90.99 (0.97-1.1)0.3
Sex0.65 (0.37-1.14)0.10.51 (0.24-2.02)0.060.59 (0.38-90.89)0.01
BMI (kg/m2)1.04 (0.99-1.1)0.071.1 (1.05-1.15)0.000011.1 (1.07-1.14)5.03 × 10−10
Diabetes0.82 (0.47-1.39)0.472.79 (1.45-5.55)0.0021.82 (1.2-2.78)0.005
B. NASH (presence versus absence)
Genotype*1.5 (1.04-2.2)0.031.7 (0.95-3.14)0.071.5 (1.12-2.04)0.007
Age (years)1.0 (1.0-1.02)0.021.0 (0.96-1.05)0.21.01 (1.00-1.02)0.006
Sex0.76 (0.37-1.5)0.40.44 (0.13-1.33)0.090.68 (0.4-1.17)0.1
BMI (kg/m2)1.04 (0.99-1.09)0.081.04 (0.98-1.12)0.21.03 (1.00-1.07)0.02
Diabetes2.38 (1.43-4.17)0.0014.54 (1.82-12.5)0.0012.78 (1.79-4.55)0.000009
Steatosis2.56 (1.8-3.72)0.000000332.1 (13.2-95.2)1.4 × 10−124.01 (3.02-5.42)2 × 10−16
C. Fibrosis (0-1 versus 2-4)
Genotype*1.54 (0.98-2.45)0.061.7 (1.02-3.1)0.041.5 (1.09-2.12)0.01
Age (years)1.04 (1.01-1.06)0.0031.02 (0.98-1.06)0.11.04 (1.02-1.06)0.0001
Sex1.28 (0.6-2.4)0.480.57 (0.2-1.54)0.21.1 (0.65-1.8)0.7
BMI (kg/m2)1.07 (1.00-1.14)0.020.99 (0.93-1.04)0.81.03 (0.99-1.07)0.08
Diabetes2.22 (1.19-4.35)0.014.2 (1.85-9.9)0.00072.56 (1.59-4.17)0.0001
Steatosis1.13 (0.71-1.8)0.590.68 (0.39-1.17)0.160.99 (0.71-1.37)0.9
NASH22.7 (11.0-51.5)1.03 × 10−1515.4 (4.8-60.4)0.0000116.4 (8.96-32.1)2 × 10−16
thumbnail image

Figure 1. Prevalence of (A) NASH and (B) fibrosis in 591 biopsied patients with NAFLD subdivided according to the adiponutrin rs738409 C→G genotype (encoding for I148M) and the referral center. (A) P = 0.0002 at chi-squared for trend for the prevalence of NASH across the adiponutrin genotypes in the combined series. (B) P < 0.0001 at chi-squared test for trend for the prevalence of fibrosis >1 across the adiponutrin genotypes in the combined series. *P < 0.05, **P < 0.005, ***P < 0.0001 versus CC genotype.

Download figure to PowerPoint

Correlation Between Adiponutrin Expression, Metabolic Features, and Liver Damage.

Hepatic adiponutrin mRNA levels, measured in 52 obese patients, were not significantly affected by the presence of the I148M variant (Fig. 2). Adiponutrin mRNA levels were significantly higher in females than in males (0.95 ± 0.39 versus 0.66 ± 0.24 arbitrary units, P = 0.002), and lower in patients with fibrosis >1 (0.64 ± 0.2 versus 0.94 ± 0.4 arbitrary units, P = 0.01). Adiponutrin mRNA levels correlated with serum triglycerides (rho = 0.63, P < 0.0001), but were not significantly associated with age, BMI, diabetes and insulin resistance, steatosis percentage, and the severity of necroinflammation.

thumbnail image

Figure 2. Expression of adiponutrin and genes involved in insulin signaling, lipogenesis, apoptosis, and fibrogenesis according to the adiponutrin I148M genotype in 52 obese patients with NAFLD. Data are shown as mean ± standard error. FASL, Fas ligand; INSR, insulin receptor; PPAR-alpha, peroxisome proliferator-activated receptor-α; SREBP1c, steroid regulatory element binding protein 1c. P < 0.05 for the expression of INSR, SREBP1c, PPAR-alpha and FASL across adiponutrin genotypes. *P < 0.05 versus I/I genotype.

Download figure to PowerPoint

Association Between Adiponutrin Genotype and Expression of Genes Associated with Steatosis-Related Liver Damage.

Expression of INSR, SREBP1c, PPAR-α, and FASL was significantly different among adiponutrin genotypes (P < 0.05; Fig. 2). Homozygosity for the 148M variant was associated with higher expression of the proapoptotic molecule FASL, and lower mRNA levels of INSR, SREBP1c, and PPAR-α compared to the I/I 148 variant (P < 0.05; Fig. 2). The expression of INSR and PPAR-α were negatively correlated with percentage of steatosis and NAFLD activity score (NAS) (rho = −0.38, P = 0.02 and rho = −0.33, P = 0.03, respectively for the correlation of INSR with steatosis percentage and NAS; rho = −0.61, P = 0.005 and rho = −0.41, P = 0.01, respectively, for the correlation of PPAR-α with steatosis and NAS). SREBP1c was positively associated with steatosis percentage (rho = 0.38, P = 0.02). FASL expression correlated with fibrosis stage (rho = 0.36, P = 0.02).

Discussion

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The Adiponutrin/PNPLA3 genotype has recently been identified by genomewide association studies as a major inherited determinant of liver triglyceride accumulation and ALT levels at the population level, and the I148M variant as the most prevalent SNP predisposing to steatosis in Caucasians.12, 13 These results confirm that the frequency of the 148M variant is significantly higher in patients with NAFLD compared to controls with normal liver enzymes and metabolic parameters.32 The risk of NAFLD was about 3.3-fold higher in subjects carrying the rs738409 GG genotype, and was detected in about 3% of healthy subjects but in 14% of patients, compared to the CC genotype. Because blood donors are often a healthy subset of the population, and we carefully further excluded subjects with any metabolic and liver enzyme abnormality, we cannot exclude that the prevalence of the G allele is higher in the general population, 20%-30% of whom have NAFLD, than in our controls.

In line with previous data,12 the rs738409 SNP was not associated with BMI and serum triglycerides, but the 148M allele was associated with ALT levels in a dose-dependent manner. The altered HDL cholesterol levels observed in patients carrying the minor allele are consistent with recent data indicating that adiponutrin genotype influences lipoprotein metabolism.33 However, because this association has not been observed in other populations12 and was not confirmed in patients from the UK, additional studies are required to define whether an interaction exists between liver steatosis and the rs738409 SNP in determining lipoprotein metabolism.

In this study, we have confirmed in a large series of biopsied patients that the adiponutrin rs738409 genotype influences both the presence of NASH and the severity of fibrosis in patients with NAFLD independent of the degree of obesity, the presence of diabetes, and the previously demonstrated effect of genotype on steatosis. This effect was observed in two independent cohorts and confirmed in transmission disequilibrium testing of family trios with fibrotic NAFLD. The association between adiponutrin genotype, steatosis, ALT levels, and NASH is not a novel finding, although histological damage has been evaluated so far only in small series of patients.21, 22, 34 In contrast, the association with fibrosis severity is completely novel. The observation that the association between genotype and NASH/fibrosis is independent of steatosis provides further evidence that steatosis may not be an a priori “first hit” in the development of progressive damage35 and suggests that adiponutrin, possibly by regulating FFA fluxes, may be involved in pathogenic mechanisms of NASH and fibrosis that are independent of its effects on the amount of hepatic triglyceride accumulation. An effect of adiponutrin genotype on FFA flux, endoplasmic reticulum, and oxidative- or cytokine-mediated stress would independently influence both steatosis, necroinflammation, and fibrosis, and would reconcile the clinical observation that steatohepatitis and fibrosis severity correlate with the degree of steatosis.19 However, we cannot exclude that the loss of liver fat, the presence of which is known to be associated with the development of severe fibrosis, may have negatively affected the association among adiponutrin genotype, steatosis, and fibrosis severity in this cross-sectional study.

Whether patients with the CG/GG genotype have different response to therapy and outcome, or may benefit from a tailored follow-up, needs to be evaluated in further studies.

In line with a functional effect of the I148M polymorphism on adiponutrin activity leading to steatosis,20 the rs738409 GG genotype, coding for the 148M/M variant, did not affect hepatic adiponutrin mRNA, but influenced the expression of molecules associated with steatosis-related liver damage, such as INSR, PPAR-α, and the membrane receptor FASL. In particular, FASL sensitizes hepatocytes to apoptosis and has been implicated in the pathogenesis of NASH,36 and indeed FASL expression was correlated with fibrosis stage in the patients evaluated. Decreased adiponutrin mRNA expression in patients with more severe liver fibrosis suggests also that decreased activity of the enzyme may predispose individuals to progressive liver disease. However, because reduced mRNA levels do not always translate to reduced protein levels, and gene expression analysis was restricted to a subset of patients subjected to bariatric surgery, these data should be interpreted with caution. Additional mechanistic studies are required to clarify these issues.

The rationale for evaluating gene expression in bariatric patients was related to the availability at our center of larger samples of hepatic tissue from needle biopsies performed at the time of surgery, and of the full spectrum of liver pathology samples ranging from normal liver without steatosis to fibrotic steatohepatitis. Moreover, even if the high BMI of the evaluated subject may limit generalizability of the results, this population has the advantage of eliminating the confounding effect of peripheral adiposity (which was present in all subjects) on liver fat and gene expression.

A limitation of this study is the lack of ultrasonographic exclusion of liver steatosis in control subjects. However, these subjects were carefully selected among a much larger pool of healthy blood donors because of normal metabolic parameters, liver enzymes, and a very low fatty liver index.24 Another limitation is related to the relatively small number of patients, all affected by severe obesity, who were subjected to gene expression evaluation. However, the strengths of this study are: (1) it is by some considerable margin the largest association study examining the influence of the adiponutrin I148M genotype on NAFLD histological severity; (2) the associations were present in two independent cohorts of patients and confirmed in a family study; and (3) this is the largest association study for which hepatic gene expression data in relation to the adiponutrin genotype are available.

In conclusion, the adiponutrin I148M genotype is associated with the risk of developing NAFLD and its progression to necroinflammation and fibrosis. Further studies on the function of adiponutrin are awaited with interest, but these data suggest that this is unlikely to be restricted to a role in the accumulation of hepatic triglyceride.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

We thank Erika Fatta, Cristina Bertelli, Daniela Bignamini, Andrea Pietrobattista, Marco Antonio Zappa, and Ezio Lattuada for clinical assistance.

References

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.