Patatin-like phospholipase domain containing 3: A case in point linking genetic susceptibility for alcoholic and nonalcoholic liver disease


  • Potential conflict of interest: Nothing to report.

Tian C, Stokowski RP, Kershenobich D, Ballinger DG, Hind DA. Variant in PNPLA3 is associated with alcoholic liver disease. Nat Genet 2009;42:21-23. Available at (Reprinted with permission.)

Two genome-wide association studies (GWAS) have described associations of variants in PNPLA3 with nonalcoholic fatty liver and plasma liver enzyme levels. We investigated the contributions of these variants to liver disease in Mestizo subjects with a history of alcohol dependence. We found that rs738409 in PNPLA3 is strongly associated with alcoholic liver disease and clinically evident alcoholic cirrhosis (unadjusted OR = 2.25, P = 1.7 × 10−10; ancestry-adjusted OR = 1.79, P = 1.9 × 10−5).


Alcoholic liver disease (ALD) includes a spectrum of histopathological injury ranging from simple fatty liver or steatosis through alcoholic hepatitis, hepatic fibrosis and cirrhosis, to end-stage liver disease. Fatty liver develops in more than 90% of drinkers whereas only 8%-15% of heavy drinkers (>40 g/day) develop cirrhosis. A strong association exists between the amount of alcohol intake and consumption patterns and ALD. In addition, factors such as sex, genetic background, and environmental factors (e.g., nutrition) appear to play an important role in determining which heavy drinker develops cirrhosis. Evidence supporting a role for genetic factors comes from twin studies1 and from the observation that the death rate from ALD is subject to wide interethnic variation that is not entirely explained by variations in the prevalence of alcohol abuse.2, 3 For example, Hispanics appear to be at particularly high risk. In the United States, the death rate (per 100,000) from ALD among men was reported to be the highest in Hispanic whites (12.6) followed by non-Hispanic African Americans (7.4), non-Hispanic whites (5.2), and Hispanic African Americans (1.8).4

Patients with all stages of ALD often have coexistent risk factors for nonalcoholic fatty liver disease (NAFLD) including obesity and hyperglycemia.5 NAFLD also follows similar histopathological sequelae to ALD from fatty liver, through inflammation to fibrosis and cirrhosis, and appears to share many pathogenic mechanisms with ALD including oxidative and endoplasmic reticulum–mediated stress and endotoxin-mediated cytokine release.6 Furthermore, NAFLD shows similar interethnic variation to ALD, with Hispanics showing the highest prevalence of disease and African Americans the lowest.7 Accordingly, it seems likely that genetic factors predisposing an individual to NAFLD may also play a role in determining the risk of ALD. This article from Tian and colleagues provides the first definitive evidence that this is indeed the case, by showing an association between ALD and a variant allele in the gene encoding PNPLA3 (patatin-like phospholipase domain containing 3) recently associated with NAFLD and/or raised serum aminotransferases in two recent genome-wide association studies (GWAS).8, 9

Romeo et al. demonstrated that in NAFLD, the PNPLA3 allele rs738409 (C→G) showed a strong association with increased hepatic fat, as determined by proton magnetic resonance spectroscopy (1H-MRS) (P = 5.9 × 10−10), and plasma alanine aminotransferase (ALT) and aspartate aminotransferase (AST) in Hispanics.9 The G allele, which is more prevalent in Hispanics (49%) than in African Americans (17%) and European Americans (23%) with NAFLD, was associated with a two-fold higher hepatic fat content in individuals homozygous for the G allele compared to CC homozygotes. In African Americans, possession of another PNPLA3 allele (rs6006460 [T]) was associated with an 18% lower hepatic fat content compared to GG homozygotes (P = 6.0 × 10−4), and this allele was extremely rare in European Americans (0.3%) and Hispanics (0.8%) compared to African Americans (10.4%). The rs738409 variant has also been associated with 1H-MRS–determined liver fat content in a Finnish study,8 and ALT and AST levels in an independent GWAS10 and in another case–control study in European White and Indian Asian populations.11

PNPLA3 encodes for adiponutrin, a transmembrane protein highly expressed in the liver and adipose tissues. Adiponutrin is primarily found in the brown and white adipose tissues in animals.12 In humans, its expression in subcutaneous and intra-abdominal adipose tissue strongly correlates with obesity,13 whereas hepatic messenger RNA expression correlates with obesity and liver fat content.8 The biochemical function of adiponutrin is uncertain, but is considered to have lipogenic transacetylase activity likely facilitating energy mobilization and lipid storage in adipose tissue and liver.13, 14 The variant rs238409[G] substitution changes codon 148 from a highly conserved isoleucine residue in vertebrates to methionine. The mutation is located in a putative splicing silencer domain, and thus it may have a gene regulatory role.15 The functional significance of this mutation and the uncertainty of how it might influence the loss or gain of protein function with relevance to liver disease are other important unknowns. It can be envisaged that this mutation may act as a “gain of function”, enhancing lipid accumulation in the liver and as a result may be an indicator of mild hepatocyte injury/inflammation due to its association with liver aminotransferases.10

Hind's group set out to examine whether variations in the PNPLA3 gene associated with NAFLD also play a role in determining liver disease susceptibility in Mestizo (mixed European and native American ancestry) individuals from Mexico City with a history of heavy drinking. They studied three groups of drinkers: those with clinically normal liver function (control, n = 305), those with abnormal liver function (“intermediate” ALD, n = 434), and those with clinically evident cirrhosis (n = 482). Genotyping was performed for the two nonsynonymous variants rs738409 and rs6006460 identified in the NAFLD GWAS study,9 as well as 15 common tagging single-nucleotide polymorphisms (SNPs) from the PNPLA3 region, 291 SNPs for assessing global ancestry, 16 ancestry-informative markers flanking the PNPLA3 region for assessing local ancestry, and 7 SNPs previously reported to be associated with cirrhosis in patients with hepatitis C. Analysis used likelihood ratio tests from logistic regression adjusted for age, alcohol intake, and duration and their interactions, and controlled for global ancestry with principle component analysis estimation. For the rs738409 SNP, comparing patients with cirrhosis with controls gave an unadjusted odds ratio of 2.25 per G allele (P = 1.7 × 10−10) and an ancestry-adjusted odds ratio of 1.79 per G allele (P = 1.9 × 10−5). All common haplotypes containing the G allele were more prevalent in individuals with cirrhosis compared to control drinkers. There was no difference in genotype distributions between intermediate patients with ALD and controls after adjusting for global and local ancestry. Further analysis suggested that the rs738409 variant accounts for 49% of the observed ancestry-related difference in cirrhosis susceptibility. Moreover, the G allele also showed a trend toward an association with increasing disease severity (frequencies of 0.70, 0.75, and 0.77 in Child-Pugh class A, B, and C, respectively), suggesting that this allele may also play a role in determining prognosis and clinical outcome. The rare variant rs6006460(T) reported by Romeo et al.9 was detected in both cirrhotic and control groups, but the very low frequency (0.0002) meant the study had insufficient power to detect any association with disease. No association was observed with any SNPs previously reported to be associated with hepatitis C–related cirrhosis. The authors conclude that their study, taken together with evidence from other studies, supports the view that rs738409 is an independent risk factor for liver dysfunction in fatty liver diseases.

Indeed, recent data from Day's group, using a real-time fluorescent allele-specific system (K-Biosciences, Essex, UK), replicate these findings in a UK cohort to provide further evidence of rs738409(G) association with ALD (Table 1; C.P. Day et al., unpublished data).

Table 1. Association of PNPLA3 Allele rs738409[G] with Alcoholic Cirrhosis
GenotypeHeavy Drinkers with Normal Liver Blood Tests (n = 182)Alcoholic Cirrhosis (n = 266)Odds Ratio
  1. Chi-squared2 for trend P < 0.0001.

  2. OR = 2.2 (95% CI = 1.53-3.18; P = 0.00002) for possession of G allele.

  3. OR = 5.57 (95% CI = 1.68-18.43; P = 0.0012) for GG genotype.

CC119 (65%)173 (47.2%)1
CG60 (33%)170 (46.5%)1.95 (1.34-2.84)
GG3 (1.6%)23 (6.3%)7.34 (2.19-24.52)

From what is known of the function of adiponutrin, these data strongly suggest that altered lipid processing plays a key role in the pathogenesis of progressive liver disease and provide further support for common pathogenic pathways in ALD and NAFLD. Clearly understanding the function of adiponutrin and its role in the pathogenesis of advanced fatty liver diseases is now the focus of considerable attention because this may lead to therapeutic advances for these common liver diseases for which there are currently no effective treatments available. For now, PNPLA3 genotyping offers the potential to identify individuals at increased risk of developing ALD and NAFLD providing the opportunity for targeted interventions.