Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

TM6SF2 rs58542926 influences hepatic fibrosis progression in patients with non-alcoholic fatty liver disease

TM6SF2 rs58542926 influences hepatic fibrosis progression in patients with non-alcoholic fatty... ARTICLE Received 2 Apr 2014 | Accepted 5 Jun 2014 | Published 30 Jun 2014 DOI: 10.1038/ncomms5309 OPEN TM6SF2 rs58542926 influences hepatic fibrosis progression in patients with non-alcoholic fatty liver disease 1 2 1,w 1 1 1 Yang-Lin Liu , Helen L. Reeves , Alastair D. Burt , Dina Tiniakos , Stuart McPherson , Julian B.S. Leathart , 3 3 4 1,5 1 Michael E.D. Allison , Graeme J. Alexander , Anne-Christine Piguet , Rodolphe Anty , Peter Donaldson , 6 7 7 8 8 4 Guruprasad P. Aithal , Sven Francque , Luc Van Gaal , Karine Clement , Vlad Ratziu , Jean-Francois Dufour , 1 1, 1, Christopher P. Day , Ann K. Daly * & Quentin M. Anstee * Non-alcoholic fatty liver disease (NAFLD) is an increasingly common condition, strongly associated with the metabolic syndrome, that can lead to progressive hepatic fibrosis, cirrhosis and hepatic failure. Subtle inter-patient genetic variation and environmental factors combine to determine variation in disease progression. A common non-synonymous poly- morphism in TM6SF2 (rs58542926 c.449 C4T, p.Glu167Lys) was recently associated with increased hepatic triglyceride content, but whether this variant promotes clinically relevant hepatic fibrosis is unknown. Here we confirm that TM6SF2 minor allele carriage is associated with NAFLD and is causally related to a previously reported chromosome 19 GWAS signal that was ascribed to the gene NCAN. Furthermore, using two histologically characterized cohorts encompassing steatosis, steatohepatitis, fibrosis and cirrhosis (combined n ¼ 1,074), we demonstrate a new association, independent of potential confounding factors (age, BMI, type 2 diabetes mellitus and PNPLA3 rs738409 genotype), with advanced hepatic fibrosis/cirrhosis. These findings establish new and important clinical relevance to TM6SF2 in NAFLD. 1 2 Institute of Cellular Medicine, The Medical School, Newcastle University, Newcastle upon Tyne NE2 4HH, UK. Northern Institute for Cancer Research, The Medical School, Newcastle University, Newcastle upon Tyne NE2 4HH, UK. Liver Unit, Department of Medicine, Addenbrooke’s Hospital, Cambridge 4 5 CB2 0QQ, UK. University Clinic of Visceral Surgery and Medicine, Inselspital Bern, 3010 Bern, Switzerland. Institut National de la Sante et de la Recherche Medicale (INSERM), U1065, Team 8, Nice F-06204, Cedex 3, France. NIHR Nottingham Digestive Diseases Biomedical Research Unit, Nottingham University Hospitals NHS Trust and University of Nottingham, Nottingham NG7 2UH, UK. Department of Endocrinology, Diabetology and Metabolism, Antwerp University Hospital, Wilrijkstraat 10, B-2650 Edegem, Antwerp, Belgium. Institute of Cardiometabolism and Nutrition, Pitie´-Salpeˆtrie`re Hospital, 75013 Paris, France. * These authors jointly supervised this work. w Present address: School of Medicine, University of Adelaide, Eleanor Harrald Building, Frome Road, Adelaide, South Australia 5005, Australia. Correspondence and requests for materials should be addressed to Q.M.A. (email: [email protected]). NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications 1 & 2014 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5309 on-alcoholic fatty liver disease (NAFLD) represents a relevant co-morbidities and factors that have previously been spectrum of progressive liver disease characterized by linked with disease progression (age, gender, body mass index Nincreased hepatic triglyceride content (HTGC) in the (BMI), presence of T2DM and PNPLA3 rs738409 genotype), and absence of excess alcohol consumption . NAFLD includes simple replicate our findings in a separate histologically characterized steatosis, non-alcoholic steatohepatitis, fibrosis and ultimately European Caucasian ‘validation’ cohort. To discover whether the cirrhosis, and is strongly associated with features of the metabolic TM6SF2 rs58542926 variant also confers an increased risk of syndrome (obesity, insulin resistance/type 2 diabetes mellitus NAFLD-related HCC, we perform a secondary case–control (T2DM) and dyslipidaemia) . Reflecting the increasing analysis comparing the overall ‘combined’ cohort of NAFLD prevalence of these conditions, NAFLD is estimated to affect patients to a cohort of NAFLD–HCC patients. approximately one-third of the population in many developed countries. Simple steatosis is generally considered to have a Results benign course and therefore to be of limited prognostic 2,3 Increased TM6SF2 rs58542926 C4T minor allele carriage relevance . However, some NAFLD patients exhibit in NAFLD. In the NAFLD discovery cohort, the TM6SF2 progressive steatohepatitis leading to cirrhosis and/or rs58542926 genotypes were confirmed to be in Hardy–Weinberg hepatocellular carcinoma (HCC), conditions that confer equilibrium with a minor allele frequency of 0.12, significantly 1,4 increased morbidity and mortality . Despite its high higher than that observed in a reference Northern European prevalence, only a minority of NAFLD patients progress to population sample (MAF 0.07, http://browser.1000genomes.org) significant fibrosis and experience the associated morbidity . or a cohort of 265 Caucasian self-reported ‘healthy workers’ Thus, similar to other common diseases (for example, obesity, recruited from offices and factories locally in the North East of T2DM and cardiovascular disease), NAFLD is best considered as England (MAF 0.07) and so supportive of an association between a complex trait in which disease phenotype results from these variants and NAFLD. Indeed, a gene-dosage effect was environmental exposures acting on a susceptible polygenic observed for both variants in the discovery cohort with the 5–7 background that comprises multiple independent modifiers . incidence of NAFLD increasing with the number of minor alleles 8–10 Genome-wide association studies (GWAS) and candidate- possessed (X for trend, P ¼ 0.0008; Supplementary Table 1). 11–15 gene studies have contributed greatly to our understanding A similar association was also confirmed for PNPLA3 rs738409, of the genetic contribution to NAFLD pathogenesis and P ¼ 0.0001 (Supplementary Table 2). Specific histological com- variability of prognosis (reviewed in ref. 7). Among the loci ponents of the NAFLD disease phenotype were next assessed identified, the non-synonymous single-nucleotide polymorphism individually. (SNP) in PNPLA3 (rs738409 c.444 C4G, p.Ile148Met), patatin- like phospholipase domain containing 3, has been validated across 8,9,12,16 multiple patient cohorts . Importantly, carriage of this SNP TM6SF2 and degree of histological steatosis. As a positive has been robustly associated not only with steatosis but also with control, and consistent with our previously reported analysis , clinically relevant factors, including severity of hepatic fibrosis/ carriage of the PNPLA3 rs738409 minor allele was significantly 12,17,18 cirrhosis and development of NAFLD-related HCC . associated with degree of steatosis in multivariate analysis Recently, Kozlitina et al. showed that a non-synonymous adopting an additive model adjusted for gender, age at biopsy, SNP in TM6SF2 (rs58542926 c.449 C4T, p.Glu167Lys), BMI and presence of T2DM (b ¼ 0.192 0.056, 95% confidence transmembrane 6 superfamily member 2, a gene of unknown  4 interval (CI) 0.082–0.301; P ¼ 6.74  10 ). However, in contrast function on chromosome 19, was associated with proton 19 to the report by Kozlitina et al. , neither TM6SF2 rs58542926 magnetic resonance spectroscopy ( H-MRS) quantified HTGC (b ¼ 0.087 0.083, 95%CI  0.076 to 0.250; P ¼ 0.296) nor based on genotyping with a genome-wide exome chip . This NCAN rs2228603 (b ¼ 0.050 0.085, 95%CI  0.116 to 0.216; variant has also been associated with dyslipidaemia and P ¼ 0.554) were found to be significantly associated with degree of cardiovascular risk . The TM6SF2 rs58542926 SNP lies within histologically determined steatosis in the 349-patient discovery 50 kb of an NCAN gene variant (rs2228603 c.274 C4T, cohort. This was also the case in the 725-patient validation cohort p.Pro92Ser) that has previously been associated with HTGC in (P ¼ 0.17). However, a trend towards significance was observed 9,21 another GWAS . Both SNPs are in strong linkage when the two cohorts were combined (b ¼ 0.111 0.059, 95%CI 0 2 disequilibrium (D ¼ 0.926, r ¼ 0.798). Conditioning on the 0.0041 to 0.2268; P ¼ 0.053), suggesting that an underlying TM6SF2 variant abrogated the effect of the NCAN variant while effect on degree of steatosis may be present but of relatively small the reverse did not occur, suggesting that TM6SF2 rs58542926 is size. An effect became apparent when the multivariate analysis in more strongly associated with the HTGC phenotype. the combined cohort was repeated after subdividing the cohort Homozygote TM6SF2 rs58542926 minor (T) allele carriage was into those with mild steatosis (S0–1) and pronounced steatosis shown to be associated with a modest but statistically significant (S2–3). Here, carriage of each copy of the TM6SF2 rs58542926 increase in H-MRS measured HTGC from 5.86 0.25% in CC C4T minor allele was associated with increased risk of greater homozygotes to 15.04 2.23% in TT homozygotes . In vitro and steatosis (odds ratio (OR) 1.379, 95%CI 1.019–1.865; P ¼ 0.037), in vivo functional studies also supported this conclusion but were although with a marginal level of significance. unable to determine whether the effect of TM6SF2 was limited to steatosis or had broader clinical relevance, as has already been shown for PNPLA3 (ref. 19). TM6SF2 and severity of histological steatohepatitis. Next, the The aim of the current study was, first, to determine whether association with steatohepatitis activity was tested using a com- the association with NAFLD reported by Kozlitina et al. could posite score incorporating severity of necroinflammation and be independently validated; and, second, to establish whether ballooning hepatocyte degeneration. TM6SF2 rs58542926, but not the TM6SF2 rs58542926 variant was associated with clinically NCAN rs2228603, was associated with severity of steatohepatitis important disease end points that have prognostic relevance (in in the discovery cohort by multivariate analysis adopting an particular stage of hepatic fibrosis or development of NAFLD- additive model adjusted for gender, age at biopsy, BMI, T2DM related HCC). To address this, we perform a quantitative analysis and PNPLA3 rs738409 genotype (b ¼ 0.288 0.139, 95%CI within a well-characterized European Caucasian ‘discovery’ 0.015–0.561; P ¼ 0.039). However, this effect was not replicated in cohort with histologically characterized NAFLD, controlling for the validation or combined cohorts. 2 NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications & 2014 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5309 ARTICLE TM6SF2 and stage of histological fibrosis. Finally, the associa- analysis incorporating known risk factors including age, gender, tion with NAFLD fibrosis stage was tested. In the discovery BMI, T2DM and presence of cirrhosis (P ¼ 0.42). cohort multivariate analyses adopting an additive model adjusted for gender, age at biopsy, BMI, T2DM and PNPLA3 rs738409 genotype found that TM6SF2 rs58542926 (b ¼ 0.549 0.135, Discussion 95%CI 0.285–0.813; P ¼ 5.57  10 ) and NCAN rs2228603 The region on chromosome 19 (19p13) flanking TM6SF2 has 9,19,21 (b ¼ 0.419 0.138, 95%CI 0.148–0.689; P ¼ 0.0026) were both been reported to be associated with NAFLD as well as significantly associated with stage of fibrosis. The association variations in plasma cholesterol, triglyceride and low-density 20,24,25 between TM6SF2 rs58542926 and fibrosis stage persisted lipoprotein levels in several previous studies. In particular, when analysis included both the NCAN rs2228603 and the a variant within the NCAN gene (rs2228603 C4T) that is in 0 2 PNPLA3 rs738409 SNPs as covariates (b ¼ 0.552 0.205, 95%CI strong linkage disequilibrium (D ¼ 0.926, r ¼ 0.798) with 0.151–0.953; P ¼ 0.0074). However, the association with NCAN TM6SF2 rs58542926 was reported to be associated with rs2228603 was lost when the analysis was conditioned on radiologically and histologically characterized NAFLD in both 9,21 rs58542926. Thus, the association is driven by the TM6SF2 GWAS and candidate-gene studies . Before the recent rs58542926 variant, and carriage of its minor allele confers sig- publication by Kozlitina et al. , examination of linkage nificantly greater NAFLD-related hepatic fibrosis independent of disequilibrium patterns across the region had already brought gender, age at biopsy, BMI, T2DM and PNPLA3 rs738409 that association into question . It was, however, the use of a genotype. genome-wide exome-chip genotyping approach, combined with This strong association between TM6SF2 rs58542926 and detailed association analysis conditioning on previously published fibrosis stage was replicated independently in the validation variants across the 19p13 region, that determined that the cohort (b ¼ 0.238 0.097, 95%CI 0.047–0.428; P ¼ 0.014) causative variant affecting HTGC was TM6SF2 rs58542926 and also clearly demonstrated in the combined cohort (ref. 19). When considered alongside a separate study by (b ¼ 0.357 0.079, 95%CI 0.203–0.511; P ¼ 6.36  10 )by Holmen et al. , which demonstrated an association with using an additive model adjusted for gender, age at biopsy, cardiovascular disease and circulating triglyceride/total BMI, T2DM and PNPLA3 rs738409 genotype in both cases. To cholesterol levels, it appears that TM6SF2 rs58542926 C-allele illustrate the potential clinical relevance of this finding, when the carriage increases circulating triglyceride/total cholesterol while multivariate analysis was repeated subdividing the NAFLD cohort T-allele carriage promotes hepatic triglyceride/cholesterol into those with mild fibrosis (F0–1) and advanced fibrosis (F2–4), retention. In clinical practice, simple steatosis is generally carriage of each copy of the TM6SF2 rs58542926 C4T minor considered to have a benign course and so degree of HTGC is 1,3,26 allele was associated consistently with a significant increased risk of limited prognostic relevance . In contrast, progressive of advanced fibrosis, independent of gender, age at biopsy, BMI, hepatic fibrosis leading to cirrhosis is the principal common 2,3 T2DM and PNPLA3 rs738409 genotype across each cohort pathway to hepatic failure and a liver-related death . Using studied (Table 1). two large, well-characterized European Caucasian cohorts with biopsy-proven NAFLD, we demonstrate that carriage of the TM6SF2 and risk of HCC. There is increasing evidence that TM6SF2 rs58542926 variant is strongly associated with the NAFLD predisposes to an increased risk of HCC , an effect presence of NAFLD and, in particular, with a significantly greater influenced by PNPLA3 rs738409 genotype independent of the risk of developing advanced hepatic fibrosis/cirrhosis. presence of cirrhosis . We therefore sought to determine Evidence to support a modifier effect of the TM6SF2 whether TM6SF2 rs58542926 had a similar effect. A cohort of rs58542926 variant on histologically determined HTGC 99 consecutive Northern European Caucasian patients with (steatosis), seen only when the 1,074-patient strong combined primary NAFLD-related HCC was identified according to the cohort was studied, is arguably more modest than might be joint European Association for the Study of the Liver and expected. Our findings do support the previously reported 9,19,21 European Association for the Research and Treatment of Cancer association , although differences in sensitivity to subtle (EASL-EORTC) guidelines . TM6SF2 rs58542926 allele and changes in HTGC between radiological and histological genotype frequencies in this cohort were compared with the modalities may have reduced the power to detect this effect . combined NAFLD cohort described above (n ¼ 1,074). In Kozlitina et al. reported that the maximal effect of the TM6SF2 univariate analysis, homozygote carriage of the TM6SF2 variant in European Caucasians was only a mean 9.2% increase in rs58542926 minor allele was associated with an increased risk H-MRS quantified HTGC in TT homozygotes above the B5.9% of NAFLD–HCC with respect to CC (OR 1.922, 95%CI 1.31–2.81; observed in CC homozygotes . Histological assessment of P ¼ 6.81  10 ); however, significance was lost in multivariate hepatic steatosis uses broad microscopic categories reflecting Table 1 | Multivariate analysis of association between TM6SF2 rs58542926 genotype and fibrosis stage F0–1 (mild) versus F2–4 (advanced). Variables Discovery cohort (n ¼ 349) Validation cohort (n ¼ 725) Combined cohort (n ¼ 1,074) OR (95%CI) P-value OR (95%CI) P-value OR (95%CI) P-value 5  5 TM6SF2 genotype 2.94 (1.76–4.89) 3.44  10 1.46 (1.03–2.09) 0.0362 1.88 (1.41–2.5) 1.63  10 PNPLA3 genotype 1.57 (1.21–2.19) 0.0086 1.32 (1.05–1.66) 0.0183 1.40 (1.16–1.69) 4.84  10 Age 1.03 (1.01–1.06) 0.0045 1.02 (1.01–1.04) 0.0041 1.03 (1.01–1.04) 1.57  10 Gender (female) 0.94 (0.57–1.56) 0.8297 1.81 (1.30–2.50) 4.50  10 1.43 (1.09–1.89) 0.0096 4  5 BMI 1.05 (1.00–1.10) 0.0368 1.03 (1.01–1.05) 9.80  10 1.04 (1.02–1.05) 3.78  10 8  11 T2DM 2.39 (1.49–3.84) 0.0003 2.73 (1.93–3.88) 1.68  10 2.57 (1.95–3.39) 1.78  10 BMI, body mass index; CI, confidence interval; OR, odds ratio; T2DM, type 2 diabetes mellitus. Additive model including age, gender, BMI, T2DM and PNPLA3 rs738409 genotype as covariates. Discovery/validation/combined cohorts: stage F0–1 (mild) n ¼ 198/439/637, stage F2–4 (advanced) n ¼ 151/286/437. NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications 3 & 2014 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5309 the proportion of hepatocytes that are visibly steatotic (S0 o5%, Conditional analysis undertaken as part of the present study S1 5–33%, S2 33–66% and S3 466%) . Based on data from adds further weight to the assertion that the 19p13 signal is previous comparative modality analysis , the modest gene effect causally related to TM6SF2 and not NCAN, not only for HTGC as 19 19 size reported by Kozlitina et al. (less than a threefold increase in was previously reported but now also for stage of hepatic HTGC above normal) would likely be encompassed within the fibrosis. These findings therefore establish a new and important histological S1 bracket, and therefore may not be apparent clinical relevance to the recently described association between histologically. Combined with the relatively low minor allele TM6SF2 and NAFLD, and suggest that TM6SF2 should be 9,29,30 frequency in the background population, smaller cohorts may considered alongside PNPLA3 (refs 8,12) and GCKR , as one therefore have insufficient statistical power for an association to of a handful of genes so far identified that are associated not only become evident. with variations in hepatic triglyceride accumulation but also with The modifier effect of the TM6SF2 variant on grade of fibrogenesis . It is noteworthy that across all the cohorts studied, steatohepatitis (disease activity) was apparent in the initial the OR for advanced fibrosis conferred by each copy of the discovery cohort analysis; however, statistical significance was TM6SF2 variant carried was consistently of similar or up to not reached in the subsequent validation analysis. The validation twofold greater magnitude than that which was observed, or has 12,31 cohort comprised a mixture of patients recruited from previously been reported , for the widely studied PNPLA3 both hepatology and bariatric services and, although the rs738409 variant (Table 1). cohorts appear well matched histologically, the validation Carriage of the TM6SF2 variant was associated with increased cohort exhibited higher mean BMI levels (38.5 9.1 versus risk of progression to NAFLD–HCC in univariate analysis. In ± ± 34.5 5.7 kg m , Po0.0001), a younger mean age (47.6 12.4 contrast to PNPLA3 (ref. 18), this effect was not sustained when versus 51.5 12.0 years, Po0.0001), a greater female preponder- confounding factors including age, T2DM and presence of ance (56.1 versus 32.1%, Po0.0001) and a lower prevalence of underlying cirrhosis were included in the model. Carriage of T2DM (32.4 versus 46.1%, Po0.0001) than the discovery cohort the TM6SF2 variant therefore does not appear to further increase (Table 2). These factors may have impacted on our ability to HCC risk independent of its effect on fibrosis stage. It should, replicate the initial association with steatohepatitis in a multi- however, be noted that the NAFLD–HCC cohort contained only variate analysis. Further study of the variant in other patient 99 patients and so, combined with a relatively modest TM6SF2 cohorts and exploration of the functional effects of TM6SF2 on rs58542926 minor allele frequency, the current study had B70% inflammatory response will be needed to address this point. power to detect an effect if an additive genetic model and risk The key finding of the current study is that carriage of the similar to that seen for fibrosis is assumed (a ¼ 0.05) .An TM6SF2 rs58542926 C4T minor allele is unequivocally asso- association cannot therefore be completely excluded, but would ciated with an increased risk of advanced NAFLD-associated seem unlikely. Studies using larger cohorts of NAFLD–HCC hepatic fibrosis. This highly significant effect was consistently patients than are presently available will be required to provide demonstrated across all the cohorts studied and was independent sufficient power to study this further. of potentially confounding factors including gender, age at time The TM6SF2 rs58542926 c.449 C4T variant is a non- of biopsy, BMI, T2DM and PNPLA3 rs738409 genotype. synonymous change producing a glutamate to lysine amino-acid Table 2 | Demographic characteristics of patient cohorts. Discovery cohort Validation cohort Combined cohort NAFLD–HCC cohort Number 349 725 1,074 99 Ethnicity European Caucasian European Caucasian European Caucasian European Caucasian Gender (female) 147 (42.1%) 407 (56.1%) 554 (51.6%) 19 (19.2%) ± ± ± ± Age, years 51.5 12.0 47.6 12.4 48.9 12.4 70.5 8.0 ± ± ± ± BMI, kg m 34.5 5.7 38.5 9.1 37.2 8.3 31.9 6.7 T2DM (yes) 161 (46.1%) 235 (32.4%) 396 (36.9%) 68 (68.7%) Steatosis score* S0 5 (1.4%) 60 (8.3%) 65 (6.1%) — S1 99 (28.4%) 206 (28.4%) 305 (28.4%) — S2 166 (47.6%) 247 (34.1%) 413 (38.5%) — S3 79 (22.6%) 204 (28.1%) 283 (26.4%) — Activity score (composite hepatocyte ballooning and necroinflammation scores)* A0 81 (23.2%) 132 (18.2%) 213 (19.8%) — A1 65 (18.6%) 133 (18.3%) 198 (18.4%) — A2 101 (28.9%) 214 (29.5%) 315 (29.3%) — A3 64 (18.3%) 149 (20.6%) 213 (19.8%) — A4 31 (8.9%) 89 (12.3%) 120 (11.2%) — Fibrosis score F0 108 (30.9%) 277 (38.2%) 385 (35.8%) F1 90 (25.8%) 162 (22.3%) 252 (23.5%) Non-cirrhotic: 32 (32.3%) F2 55 (15.8%) 161 (22.2%) 216 (20.1%) F3 66 (18.9%) 75 (10.3%) 141 (13.1%) F4 (cirrhosis) 30 (8.6%) 50 (6.9%) 80 (7.4%) Cirrhotic: 67 (67.7%) BMI, body mass index; HCC, hepatocellular carcinoma; NAFLD, non-alcoholic fatty liver disease; T2DM, type 2 diabetes mellitus. *Steatosis and activity score data incomplete in 8 (0.7%) and 15 (1.3%) of samples, respectively. 4 NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications & 2014 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5309 ARTICLE substitution at residue 167 (Glu167Lys), which is highly of disease. These were unrelated patients with histologically characterized NAFLD, derived from a patient population originally identified as having conserved across mammals . First identified as part of a ultrasonographically detected bright liver and abnormal biochemical tests large-scale sequencing project, little is currently known about (alanine transaminase and/or gamma-glutamyl transferase). the biological function of the TM6SF2 protein product . A validation cohort of 725 consecutive European Caucasian patients Adenovirus-mediated short hairpin RNA knockdown of Tm6sf2 from centres in UK and mainland Europe with histologically characterized NAFLD of different stages of disease. Patients in this cohort were unrelated in mice has been shown to increase HTGC and reduce very low- patients with histologically characterized NAFLD, derived from a patient density lipoprotein (VLDL) secretion, suggesting that TM6SF2 population originally identified as having ultrasonographically detected bright activity is necessary for normal VLDL secretion, and that liver and abnormal biochemical tests (alanine transaminase and/or gamma- impaired TM6SF2 function causally contributes to NAFLD . glutamyl transferase) or identified as having evidence of NAFLD at the time of bariatric surgery. However, these in vivo studies were of too short a duration to adequately address the effects on steatohepatitis or fibrogenesis. Together, these comprised the combined cohort of 1,074 patients with Furthermore, previous experimental evidence has shown that histologically characterized NAFLD. Demographic and histological details are hepatic triglyceride accumulation may not itself be directly shown in Table 2. A description of the ‘healthy workers’ cohort recruited in the hepatotoxic. This was elegantly demonstrated in mice by silencing North East of the United Kingdom has previously been published . A separate cohort of 99 consecutive Northern European Caucasian patients hepatic gene expression of diacylglycerol O-acyltransferase 2 with primary HCC arising on a background of NAFLD was identified (Dgat2), a key enzyme mediating the conversion of free fatty acids (NAFLD–HCC cohort). The diagnosis of HCC was established histologically or to triglyceride . Rather than ameliorating steatohepatitis, the through non-invasive assessment according to the EASL–EORTC clinical practice consequent reduction in hepatocyte triglyceride synthesis was guidelines . The presence of NAFLD was determined through histological assessment of non-tumour liver tissue or, when biopsy was not clinically associated with increased fatty acid oxidation, particularly appropriate, through radiological evidence of hepatic steatosis. through Cyp2e1, leading to greater oxidative stress, cellular damage and higher serum transaminase levels . It is therefore tempting to speculate that the function of TM6SF2 and the Liver biopsy. Liver biopsy was performed under radiological guidance. Specimens (at least 1.6 cm length and 1.5 mm thick) were fixed in 10% neutral formalin mechanism through which TM6SF2 drives NAFLD-associated for evaluation and embedded in paraffin for histological examination. Tissue sec- hepatic fibrosis may be other than through increased triglyceride tions were stained with haematoxylin and eosin, impregnated with silver for accumulation. visualizing reticulin framework and stained with Sirius Red Fast Green for visua- In conclusion, the current study confirms that TM6SF2 is lizing collagen. Liver biopsies were reviewed by a single expert liver pathologist at associated with histologically defined NAFLD, and is the first each participating centre, unaware of clinical or genetic data. The degree of stea- tosis (S0–3), activity of steatohepatitis (A0–4) and stage of fibrosis (F0–4) were demonstration that this gene serves as a powerful modifier of scored according to the validated semi-quantitative SAF score (Supplementary hepatic fibrogenesis. That this gene is also associated with disturbed 28,37 Table 3) . In 25 HCC patients, the diagnosis of HCC was confirmed cholesterol metabolism and so may modify risk of cardiovascular histologically and graded according to Edmondson and Steiner , adapted for events including myocardial infarction suggests that TM6SF2 is needle biopsy specimens. an important determinant of clinical outcome across several facets of metabolic syndrome-related end-organ damage. In light of DNA preparation. Venous blood was collected from each patient and DNA was evidence that cholesterol accumulation in hepatic stellate cells prepared from peripheral blood lymphocytes using a perchlorate–chloroform promotes NAFLD fibrosis , it is tempting to speculate that isolation method . In brief, 35 ml lysis buffer (10 mM Tris–HCl (pH 8.0), 320 mM sucrose, 5 mM magnesium chloride and 1% Triton X-100) was added to 5 ml TM6SF2 may act as a ‘switch’ with TM6SF2 rs58542926 T-allele- venous blood in a 50-ml polypropylene centrifuge tube. After mixing, the tube was mediated hepatic retention of triglyceride and cholesterol centrifuged at 3,000 g for 10 min. The supernatant was discarded and the cell pellet predisposing to NAFLD fibrosis while C-allele carriage promotes was re-suspended in 2 ml of solution B (400 mM Tris–HCl (pH 8.0) 6 0 mM EDTA, VLDL excretion, protecting the liver at the expense of increased risk 150 mM NaCl and 1% SDS). A quantity of 500ml of sodium perchlorate (5 M) was of cardiovascular disease. These data mandate further mechanistic added and the sample was mixed at room temperature for 15 min before incubating in a preheated hot block at 65 C for 30 min. Next, 2 ml of chloroform was added study to determine the physiological and pathophysiological role of and the sample was mixed for 10 min at room temperature. The tube was then this gene in various tissues and cell types as a modifier of centrifuged at 1,400g for 10 min, and the upper, clear DNA-containing phase was fibrogenesis and a putative therapeutic target. transferred to a new 15 ml polypropylene tube. Two volumes of cold ethanol were added to the aqueous phase, and the tube was gently inverted until the DNA precipitated. The DNA was spooled using a soft plastic sterile loop and allowed to Methods air dry for 20 min. DNA was then re-suspended by incubation in 200ml water at Patients. Patients were recruited from hepatology clinics at several European 60 C. Samples are quantitated and quality assessed by absorbance measurements specialist centres: the Freeman Hospital, Newcastle upon Tyne, UK; Addenbrooke’s at 260 and 280 nm. Genotyping was performed by personnel unaware of clinical Hospital, Cambridge, UK; Nottingham University Hospitals NHS Trust, status or histology of patients. Nottingham, UK; Inselspital Hospital, Bern, Switzerland; Antwerp University Hospital, Belgium; and Pitie´-Salpeˆtrie`re Hospital, Paris, France. The study had all the necessary ethical approvals (UK: Newcastle and North Tyneside 1 REC TM6SF2 rs58542926, NCAN rs2228603 and PNPLA3 rs738409 genotyping. (10/H0906/41), Norfolk REC (06/Q0106/70)] and Nottingham 2 REC TM6SF2 rs58542926, NCAN rs2228603 and PNPLA3 rs738409 genotypes were (GM010201); Switzerland: Inselspital Bern Local Ethics Committee; Belgium: determined by allelic discrimination using TaqMan reagents (Applied Biosystems Antwerp University Hospital Ethics Committee; France: CPP (Comite´ de Inc., USA) according to the manufacturer’s protocol. Control samples of known Protection des Personnes) Paris VI IDF Pitie´—Salpeˆtrie`re Hospital). All genotype were also included in every 96-well plate (blank, homozygous wild-type, participants gave informed consent. In all cases, alternative diagnoses were homozygous mutant and heterozygous). excluded, including excess alcohol intake (alcohol intakeo20 g per day for women; and o30 g per day for men), chronic viral hepatitis (hepatitis B and hepatitis C), autoimmune liver diseases, hereditary hemochromatosis, a1-antitrypsin deficiency, Statistical analysis. Statistical analyses were performed using SPSS v19.0 (IBM, Wilson’s disease and drug-induced liver injury. Clinical and laboratory data were USA) to collate and analyse cohort phenotype data. Continuous variables were collected at the time of diagnosis including basic anthropometrics so that BMI tested using Student’s t-test/one-way analysis of variance and categorical variables could be calculated, and relevant co-morbidity including the presence of T2DM 2 by w -squared test unless otherwise stated. PLINK v1.07 (ref. 40) (via the gPLINK 1  1 (fasting glucose Z7.1 mmol l (Z128 mg dl ) or treatment with anti-diabetic v2.050 GUI) was used to conduct the genetic analysis. An initial univariate drugs) and evidence of underlying cirrhosis was recorded. The degree of steatosis 2 w -squared analysis was performed. Subsequently, multivariate logistic regression (S0–3), activity of steatohepatitis (A0–4) and stage of fibrosis (F0–4) were analysis was conducted incorporating biologically relevant covariates that were scored according to the validated semi-quantitative SAF score . The main study associated with risk of NAFLD progression (age, gender, BMI, presence of cohorts were: T2DM and PNPLA3 rs738409 genotype) to test the genetic association. An additive genetic model best fitted the data and was reported. Results are expressed as beta An initial discovery cohort of 349 consecutive European Caucasian patients from ± b s.e.m. or OR with 95%CI as appropriate. Significance was taken as Po0.05 the United Kingdom with histologically characterized NAFLD of different stages throughout. NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications 5 & 2014 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5309 References 28. Bedossa, P. et al. Utility and appropriateness of the FLIP algorithm and SAF 1. Anstee, Q. M., Targher, G. & Day, C. P. Progression of NAFLD to diabetes score in the evaluation of biopsies of nonalcoholic fatty liver disease. mellitus, cardiovascular disease or cirrhosis. Nat. Rev. Gastroenterol. Hepatol. Hepatology. doi:10.1002/hep.27173 (2014). 10, 330–344 (2013). 29. Anstee, Q. M. et al. A candidate-gene approach to validation of genetic modifier 2. Musso, G., Gambino, R., Cassader, M. & Pagano, G. Meta-analysis: associations using a large cohort with histologically characterised non-alcoholic natural history of non-alcoholic fatty liver disease (NAFLD) and diagnostic fatty liver disease. J. Hepatol. 58, S46–S46 (2013). accuracy of non-invasive tests for liver disease severity. Ann. Med. 43, 617–649 30. Petta, S. et al. Glucokinase regulatory protein gene polymorphism affects liver (2011). fibrosis in non-alcoholic fatty liver disease. PLoS ONE 9, e87523 (2014). 3. Ekstedt, M. et al. Long-term follow-up of patients with NAFLD and elevated 31. Valenti, L. et al. I148M patatin-like phospholipase domain-containing 3 gene liver enzymes. Hepatology 44, 865–873 (2006). variant and severity of pediatric nonalcoholic fatty liver disease. Hepatology 52, 4. Loomba, R. & Sanyal, A. J. The global NAFLD epidemic. Nat. Rev. 1274–1280 (2010). Gastroenterol. Hepatol. 10, 686–690 (2013). 32. Purcell, S., Cherny, S. S. & Sham, P. C. Genetic Power Calculator: design of 5. Hirschhorn, J. N. & Gajdos, Z. K. Genome-wide association studies: results linkage and association genetic mapping studies of complex traits. from the first few years and potential implications for clinical medicine. Annu. Bioinformatics 19, 149–150 (2003). Rev. Med. 62, 11–24 (2011). 33. Carim-Todd, L., Escarceller, M., Estivill, X. & Sumoy, L. Cloning of the novel 6. Altshuler, D., Daly, M. J. & Lander, E. S. Genetic mapping in human disease. gene TM6SF1 reveals conservation of clusters of paralogous genes between Science 322, 881–888 (2008). human chromosomes 15q24--4q26 and 19p13.3--4p12. Cytogenet. Cell 7. Anstee, Q. M. & Day, C. P. The genetics of NAFLD. Nat. Rev. Gastroenterol. Genet. 90, 255–260 (2000). Hepatol. 10, 645–655 (2013). 34. Yamaguchi, K. et al. Inhibiting triglyceride synthesis improves hepatic steatosis 8. Romeo, S. et al. Genetic variation in PNPLA3 confers susceptibility to but exacerbates liver damage and fibrosis in obese mice with nonalcoholic nonalcoholic fatty liver disease. Nat. Genet. 40, 1461–1465 (2008). steatohepatitis. Hepatology 45, 1366–1374 (2007). 9. Speliotes, E. K. et al. Genome-wide association analysis identifies variants 35. Tomita, K. et al. Free cholesterol accumulation in hepatic stellate cells: associated with nonalcoholic fatty liver disease that have distinct effects on mechanism of liver fibrosis aggravation in nonalcoholic steatohepatitis in mice. metabolic traits. PLoS Genet. 7, e1001324 (2011). Hepatology 59, 154–169 (2014). 10. Chambers, J. C. et al. Genome-wide association study identifies loci 36. Velaga, M. R. et al. The codon 620 tryptophan allele of the lymphoid tyrosine influencing concentrations of liver enzymes in plasma. Nat. Genet. 43, phosphatase (LYP) gene is a major determinant of Graves’ disease. J. Clin. 1131–1138 (2011). Endocrinol. Metab. 89, 5862–5865 (2004). 11. Al-Serri, A. et al. The SOD2 C47T polymorphism influences NAFLD fibrosis 37. Dyson, J. K., McPherson, S. & Anstee, Q. M. Non-alcoholic fatty liver disease: severity: evidence from case-control and intra-familial allele association studies. non-invasive investigation and risk stratification. J. Clin. Pathol. 66, 1033–1045 J. Hepatol. 56, 448–454 (2012). (2013). 12. Valenti, L. et al. Homozygosity for the patatin-like phospholipase-3/ 38. Edmondson, H. A. & Steiner, P. E. Primary carcinoma of the liver: a study of adiponutrin I148M polymorphism influences liver fibrosis in patients with 100 cases among 48,900 necropsies. Cancer 7, 462–503 (1954). nonalcoholic fatty liver disease. Hepatology 51, 1209–1217 (2010). 39. Daly, A. K., Steen, V. M., Fairbrother, K. S. & Idle, J. R. CYP2D6 multiallelism. 13. Tan, H. L. et al. Association of glucokinase regulatory gene polymorphisms Methods. Enzymol. 272, 199–210 (1996). with risk and severity of non-alcoholic fatty liver disease: an interaction study 40. Purcell, S. et al. PLINK: a tool set for whole-genome association and with adiponutrin gene. J. Gastroenterol. 49, 1056–1064 (2013). population-based linkage analyses. Am. J. Hum. Genet. 81, 559–575 (2007). 14. Dongiovanni, P. et al. Genetic variants regulating insulin receptor signalling are associated with the severity of liver damage in patients with non-alcoholic fatty Acknowledgements liver disease. Gut 59, 267–273 (2010). Q.M.A. is the recipient of a Clinical Senior Lectureship Award from the Higher 15. Aravinthan, A. et al. Gene polymorphisms of cellular senescence marker p21 Education Funding Council for England (HEFCE). Aspects of this work have been and disease progression in non-alcohol-related fatty liver disease. Cell Cycle 13, supported by the facilities of the Newcastle National Institute of Health Research (NIHR) 1489–1494 (2014). Biomedical Research Centre and by the ‘Fatty Liver Inhibition of Progression’ (FLIP) 16. Kawaguchi, T. et al. Genetic polymorphisms of the human PNPLA3 gene are project funded by the European Union Seventh Framework Programme (FP7/2007- strongly associated with severity of non-alcoholic fatty liver disease in Japanese. 2013) under grant agreement Health-F2-2009-241762. PLoS ONE 7, e38322 (2012). 17. Trepo, E. et al. Association between the PNPLA3 (rs738409 C4G) variant and hepatocellular carcinoma: evidence from a meta-analysis of individual Author contributions participant data. Hepatology 59, 2170–2177 (2013). Q.M.A., A.K.D. and C.P.D. conceived the research. Clinical phenotype data collation and 18. Liu, Y. L. et al. Carriage of the PNPLA3 rs738409 C4G polymorphism confers sample acquisition/DNA preparation was performed by J.B.S.L., Q.M.A., C.P.D., S.M., an increased risk of non-alcoholic fatty liver disease associated hepatocellular H.L.R., M.E.D.A., G.J.A., A.-C.P., P.D., G.P.A., S.F., L.V.G., K.C., V.R. and J.-F.D. carcinoma. J. Hepatol. 61, 75–81 (2014). Genotyping was performed by Y.-L.L., and assisted by J.B.S.L. and RA. Histological 19. Kozlitina, J. et al. Exome-wide association study identifies a TM6SF2 variant analysis of tissues and scoring was conducted by A.D.B. and D.T. Statistical analysis and that confers susceptibility to nonalcoholic fatty liver disease. Nat. Genet. 46, interpretation of results was performed by Y.-L.L., A.K.D. and Q.M.A. The manuscript 352–356 (2014). was written and revised by Q.M.A., A.K.D. and Y.-L.L. All authors critically reviewed the 20. Holmen, O. L. et al. Systematic evaluation of coding variation identifies a manuscript for important intellectual content and approved the final submitted candidate causal variant in TM6SF2 influencing total cholesterol and manuscript. myocardial infarction risk. Nat. Genet. 46, 345–351 (2014). 21. Gorden, A. et al. Genetic variation at NCAN locus is associated with inflammation and fibrosis in non-alcoholic fatty liver disease in morbid obesity. Additional information Hum. Hered. 75, 34–43 (2013). Supplementary Information accompanies this paper at http://www.nature.com/ 22. Dyson, J. et al. Hepatocellular cancer: the impact of obesity, type 2 diabetes and naturecommunications a multidisciplinary team. J. Hepatol. 60, 110–117 (2014). Competing financial interests: The authors declare no competing financial interests. 23. European Association For The Study Of The Liver; European Organisation For Research And Treatment Of Cancer. EASL-EORTC clinical practice Reprints and permission information is available online at http://npg.nature.com/ guidelines: management of hepatocellular carcinoma. J. Hepatol. 56, 908–943 reprintsandpermissions/ (2012). 24. Kathiresan, S. et al. Six new loci associated with blood low-density lipoprotein How to cite this article: Liu, Y.-L. et al. TM6SF2 rs58542926 influences hepatic fibrosis cholesterol, high-density lipoprotein cholesterol or triglycerides in humans. progression in patients with non-alcoholic fatty liver disease. Nat. Commun. 5:4309 Nat. Genet. 40, 189–197 (2008). doi: 10.1038/ncomms5309 (2014). 25. Teslovich, T. M. et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature 466, 707–713 (2010). This work is licensed under a Creative Commons Attribution 4.0 26. Anstee, Q. M., McPherson, S. & Day, C. P. How big a problem is non-alcoholic International License. The images or other third party material in this fatty liver disease? BMJ 343, d3897 (2011). article are included in the article’s Creative Commons license, unless indicated otherwise 27. McPherson, S. et al. Magnetic resonance imaging and spectroscopy accurately in the credit line; if the material is not included under the Creative Commons license, estimate the severity of steatosis provided the stage of fibrosis is considered. users will need to obtain permission from the license holder to reproduce the material. J. Hepatol. 51, 389–397 (2009). To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ 6 NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications & 2014 Macmillan Publishers Limited. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals

TM6SF2 rs58542926 influences hepatic fibrosis progression in patients with non-alcoholic fatty liver disease

Download PDF
6 pages

Loading next page...
 
/lp/springer-journals/tm6sf2-rs58542926-influences-hepatic-fibrosis-progression-in-patients-kxIh0Z4Omj

References (90)

Publisher
Springer Journals
Copyright
Copyright © 2014 by The Author(s)
Subject
Science, Humanities and Social Sciences, multidisciplinary; Science, Humanities and Social Sciences, multidisciplinary; Science, multidisciplinary
eISSN
2041-1723
DOI
10.1038/ncomms5309
Publisher site
See Article on Publisher Site

Abstract

ARTICLE Received 2 Apr 2014 | Accepted 5 Jun 2014 | Published 30 Jun 2014 DOI: 10.1038/ncomms5309 OPEN TM6SF2 rs58542926 influences hepatic fibrosis progression in patients with non-alcoholic fatty liver disease 1 2 1,w 1 1 1 Yang-Lin Liu , Helen L. Reeves , Alastair D. Burt , Dina Tiniakos , Stuart McPherson , Julian B.S. Leathart , 3 3 4 1,5 1 Michael E.D. Allison , Graeme J. Alexander , Anne-Christine Piguet , Rodolphe Anty , Peter Donaldson , 6 7 7 8 8 4 Guruprasad P. Aithal , Sven Francque , Luc Van Gaal , Karine Clement , Vlad Ratziu , Jean-Francois Dufour , 1 1, 1, Christopher P. Day , Ann K. Daly * & Quentin M. Anstee * Non-alcoholic fatty liver disease (NAFLD) is an increasingly common condition, strongly associated with the metabolic syndrome, that can lead to progressive hepatic fibrosis, cirrhosis and hepatic failure. Subtle inter-patient genetic variation and environmental factors combine to determine variation in disease progression. A common non-synonymous poly- morphism in TM6SF2 (rs58542926 c.449 C4T, p.Glu167Lys) was recently associated with increased hepatic triglyceride content, but whether this variant promotes clinically relevant hepatic fibrosis is unknown. Here we confirm that TM6SF2 minor allele carriage is associated with NAFLD and is causally related to a previously reported chromosome 19 GWAS signal that was ascribed to the gene NCAN. Furthermore, using two histologically characterized cohorts encompassing steatosis, steatohepatitis, fibrosis and cirrhosis (combined n ¼ 1,074), we demonstrate a new association, independent of potential confounding factors (age, BMI, type 2 diabetes mellitus and PNPLA3 rs738409 genotype), with advanced hepatic fibrosis/cirrhosis. These findings establish new and important clinical relevance to TM6SF2 in NAFLD. 1 2 Institute of Cellular Medicine, The Medical School, Newcastle University, Newcastle upon Tyne NE2 4HH, UK. Northern Institute for Cancer Research, The Medical School, Newcastle University, Newcastle upon Tyne NE2 4HH, UK. Liver Unit, Department of Medicine, Addenbrooke’s Hospital, Cambridge 4 5 CB2 0QQ, UK. University Clinic of Visceral Surgery and Medicine, Inselspital Bern, 3010 Bern, Switzerland. Institut National de la Sante et de la Recherche Medicale (INSERM), U1065, Team 8, Nice F-06204, Cedex 3, France. NIHR Nottingham Digestive Diseases Biomedical Research Unit, Nottingham University Hospitals NHS Trust and University of Nottingham, Nottingham NG7 2UH, UK. Department of Endocrinology, Diabetology and Metabolism, Antwerp University Hospital, Wilrijkstraat 10, B-2650 Edegem, Antwerp, Belgium. Institute of Cardiometabolism and Nutrition, Pitie´-Salpeˆtrie`re Hospital, 75013 Paris, France. * These authors jointly supervised this work. w Present address: School of Medicine, University of Adelaide, Eleanor Harrald Building, Frome Road, Adelaide, South Australia 5005, Australia. Correspondence and requests for materials should be addressed to Q.M.A. (email: [email protected]). NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications 1 & 2014 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5309 on-alcoholic fatty liver disease (NAFLD) represents a relevant co-morbidities and factors that have previously been spectrum of progressive liver disease characterized by linked with disease progression (age, gender, body mass index Nincreased hepatic triglyceride content (HTGC) in the (BMI), presence of T2DM and PNPLA3 rs738409 genotype), and absence of excess alcohol consumption . NAFLD includes simple replicate our findings in a separate histologically characterized steatosis, non-alcoholic steatohepatitis, fibrosis and ultimately European Caucasian ‘validation’ cohort. To discover whether the cirrhosis, and is strongly associated with features of the metabolic TM6SF2 rs58542926 variant also confers an increased risk of syndrome (obesity, insulin resistance/type 2 diabetes mellitus NAFLD-related HCC, we perform a secondary case–control (T2DM) and dyslipidaemia) . Reflecting the increasing analysis comparing the overall ‘combined’ cohort of NAFLD prevalence of these conditions, NAFLD is estimated to affect patients to a cohort of NAFLD–HCC patients. approximately one-third of the population in many developed countries. Simple steatosis is generally considered to have a Results benign course and therefore to be of limited prognostic 2,3 Increased TM6SF2 rs58542926 C4T minor allele carriage relevance . However, some NAFLD patients exhibit in NAFLD. In the NAFLD discovery cohort, the TM6SF2 progressive steatohepatitis leading to cirrhosis and/or rs58542926 genotypes were confirmed to be in Hardy–Weinberg hepatocellular carcinoma (HCC), conditions that confer equilibrium with a minor allele frequency of 0.12, significantly 1,4 increased morbidity and mortality . Despite its high higher than that observed in a reference Northern European prevalence, only a minority of NAFLD patients progress to population sample (MAF 0.07, http://browser.1000genomes.org) significant fibrosis and experience the associated morbidity . or a cohort of 265 Caucasian self-reported ‘healthy workers’ Thus, similar to other common diseases (for example, obesity, recruited from offices and factories locally in the North East of T2DM and cardiovascular disease), NAFLD is best considered as England (MAF 0.07) and so supportive of an association between a complex trait in which disease phenotype results from these variants and NAFLD. Indeed, a gene-dosage effect was environmental exposures acting on a susceptible polygenic observed for both variants in the discovery cohort with the 5–7 background that comprises multiple independent modifiers . incidence of NAFLD increasing with the number of minor alleles 8–10 Genome-wide association studies (GWAS) and candidate- possessed (X for trend, P ¼ 0.0008; Supplementary Table 1). 11–15 gene studies have contributed greatly to our understanding A similar association was also confirmed for PNPLA3 rs738409, of the genetic contribution to NAFLD pathogenesis and P ¼ 0.0001 (Supplementary Table 2). Specific histological com- variability of prognosis (reviewed in ref. 7). Among the loci ponents of the NAFLD disease phenotype were next assessed identified, the non-synonymous single-nucleotide polymorphism individually. (SNP) in PNPLA3 (rs738409 c.444 C4G, p.Ile148Met), patatin- like phospholipase domain containing 3, has been validated across 8,9,12,16 multiple patient cohorts . Importantly, carriage of this SNP TM6SF2 and degree of histological steatosis. As a positive has been robustly associated not only with steatosis but also with control, and consistent with our previously reported analysis , clinically relevant factors, including severity of hepatic fibrosis/ carriage of the PNPLA3 rs738409 minor allele was significantly 12,17,18 cirrhosis and development of NAFLD-related HCC . associated with degree of steatosis in multivariate analysis Recently, Kozlitina et al. showed that a non-synonymous adopting an additive model adjusted for gender, age at biopsy, SNP in TM6SF2 (rs58542926 c.449 C4T, p.Glu167Lys), BMI and presence of T2DM (b ¼ 0.192 0.056, 95% confidence transmembrane 6 superfamily member 2, a gene of unknown  4 interval (CI) 0.082–0.301; P ¼ 6.74  10 ). However, in contrast function on chromosome 19, was associated with proton 19 to the report by Kozlitina et al. , neither TM6SF2 rs58542926 magnetic resonance spectroscopy ( H-MRS) quantified HTGC (b ¼ 0.087 0.083, 95%CI  0.076 to 0.250; P ¼ 0.296) nor based on genotyping with a genome-wide exome chip . This NCAN rs2228603 (b ¼ 0.050 0.085, 95%CI  0.116 to 0.216; variant has also been associated with dyslipidaemia and P ¼ 0.554) were found to be significantly associated with degree of cardiovascular risk . The TM6SF2 rs58542926 SNP lies within histologically determined steatosis in the 349-patient discovery 50 kb of an NCAN gene variant (rs2228603 c.274 C4T, cohort. This was also the case in the 725-patient validation cohort p.Pro92Ser) that has previously been associated with HTGC in (P ¼ 0.17). However, a trend towards significance was observed 9,21 another GWAS . Both SNPs are in strong linkage when the two cohorts were combined (b ¼ 0.111 0.059, 95%CI 0 2 disequilibrium (D ¼ 0.926, r ¼ 0.798). Conditioning on the 0.0041 to 0.2268; P ¼ 0.053), suggesting that an underlying TM6SF2 variant abrogated the effect of the NCAN variant while effect on degree of steatosis may be present but of relatively small the reverse did not occur, suggesting that TM6SF2 rs58542926 is size. An effect became apparent when the multivariate analysis in more strongly associated with the HTGC phenotype. the combined cohort was repeated after subdividing the cohort Homozygote TM6SF2 rs58542926 minor (T) allele carriage was into those with mild steatosis (S0–1) and pronounced steatosis shown to be associated with a modest but statistically significant (S2–3). Here, carriage of each copy of the TM6SF2 rs58542926 increase in H-MRS measured HTGC from 5.86 0.25% in CC C4T minor allele was associated with increased risk of greater homozygotes to 15.04 2.23% in TT homozygotes . In vitro and steatosis (odds ratio (OR) 1.379, 95%CI 1.019–1.865; P ¼ 0.037), in vivo functional studies also supported this conclusion but were although with a marginal level of significance. unable to determine whether the effect of TM6SF2 was limited to steatosis or had broader clinical relevance, as has already been shown for PNPLA3 (ref. 19). TM6SF2 and severity of histological steatohepatitis. Next, the The aim of the current study was, first, to determine whether association with steatohepatitis activity was tested using a com- the association with NAFLD reported by Kozlitina et al. could posite score incorporating severity of necroinflammation and be independently validated; and, second, to establish whether ballooning hepatocyte degeneration. TM6SF2 rs58542926, but not the TM6SF2 rs58542926 variant was associated with clinically NCAN rs2228603, was associated with severity of steatohepatitis important disease end points that have prognostic relevance (in in the discovery cohort by multivariate analysis adopting an particular stage of hepatic fibrosis or development of NAFLD- additive model adjusted for gender, age at biopsy, BMI, T2DM related HCC). To address this, we perform a quantitative analysis and PNPLA3 rs738409 genotype (b ¼ 0.288 0.139, 95%CI within a well-characterized European Caucasian ‘discovery’ 0.015–0.561; P ¼ 0.039). However, this effect was not replicated in cohort with histologically characterized NAFLD, controlling for the validation or combined cohorts. 2 NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications & 2014 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5309 ARTICLE TM6SF2 and stage of histological fibrosis. Finally, the associa- analysis incorporating known risk factors including age, gender, tion with NAFLD fibrosis stage was tested. In the discovery BMI, T2DM and presence of cirrhosis (P ¼ 0.42). cohort multivariate analyses adopting an additive model adjusted for gender, age at biopsy, BMI, T2DM and PNPLA3 rs738409 genotype found that TM6SF2 rs58542926 (b ¼ 0.549 0.135, Discussion 95%CI 0.285–0.813; P ¼ 5.57  10 ) and NCAN rs2228603 The region on chromosome 19 (19p13) flanking TM6SF2 has 9,19,21 (b ¼ 0.419 0.138, 95%CI 0.148–0.689; P ¼ 0.0026) were both been reported to be associated with NAFLD as well as significantly associated with stage of fibrosis. The association variations in plasma cholesterol, triglyceride and low-density 20,24,25 between TM6SF2 rs58542926 and fibrosis stage persisted lipoprotein levels in several previous studies. In particular, when analysis included both the NCAN rs2228603 and the a variant within the NCAN gene (rs2228603 C4T) that is in 0 2 PNPLA3 rs738409 SNPs as covariates (b ¼ 0.552 0.205, 95%CI strong linkage disequilibrium (D ¼ 0.926, r ¼ 0.798) with 0.151–0.953; P ¼ 0.0074). However, the association with NCAN TM6SF2 rs58542926 was reported to be associated with rs2228603 was lost when the analysis was conditioned on radiologically and histologically characterized NAFLD in both 9,21 rs58542926. Thus, the association is driven by the TM6SF2 GWAS and candidate-gene studies . Before the recent rs58542926 variant, and carriage of its minor allele confers sig- publication by Kozlitina et al. , examination of linkage nificantly greater NAFLD-related hepatic fibrosis independent of disequilibrium patterns across the region had already brought gender, age at biopsy, BMI, T2DM and PNPLA3 rs738409 that association into question . It was, however, the use of a genotype. genome-wide exome-chip genotyping approach, combined with This strong association between TM6SF2 rs58542926 and detailed association analysis conditioning on previously published fibrosis stage was replicated independently in the validation variants across the 19p13 region, that determined that the cohort (b ¼ 0.238 0.097, 95%CI 0.047–0.428; P ¼ 0.014) causative variant affecting HTGC was TM6SF2 rs58542926 and also clearly demonstrated in the combined cohort (ref. 19). When considered alongside a separate study by (b ¼ 0.357 0.079, 95%CI 0.203–0.511; P ¼ 6.36  10 )by Holmen et al. , which demonstrated an association with using an additive model adjusted for gender, age at biopsy, cardiovascular disease and circulating triglyceride/total BMI, T2DM and PNPLA3 rs738409 genotype in both cases. To cholesterol levels, it appears that TM6SF2 rs58542926 C-allele illustrate the potential clinical relevance of this finding, when the carriage increases circulating triglyceride/total cholesterol while multivariate analysis was repeated subdividing the NAFLD cohort T-allele carriage promotes hepatic triglyceride/cholesterol into those with mild fibrosis (F0–1) and advanced fibrosis (F2–4), retention. In clinical practice, simple steatosis is generally carriage of each copy of the TM6SF2 rs58542926 C4T minor considered to have a benign course and so degree of HTGC is 1,3,26 allele was associated consistently with a significant increased risk of limited prognostic relevance . In contrast, progressive of advanced fibrosis, independent of gender, age at biopsy, BMI, hepatic fibrosis leading to cirrhosis is the principal common 2,3 T2DM and PNPLA3 rs738409 genotype across each cohort pathway to hepatic failure and a liver-related death . Using studied (Table 1). two large, well-characterized European Caucasian cohorts with biopsy-proven NAFLD, we demonstrate that carriage of the TM6SF2 and risk of HCC. There is increasing evidence that TM6SF2 rs58542926 variant is strongly associated with the NAFLD predisposes to an increased risk of HCC , an effect presence of NAFLD and, in particular, with a significantly greater influenced by PNPLA3 rs738409 genotype independent of the risk of developing advanced hepatic fibrosis/cirrhosis. presence of cirrhosis . We therefore sought to determine Evidence to support a modifier effect of the TM6SF2 whether TM6SF2 rs58542926 had a similar effect. A cohort of rs58542926 variant on histologically determined HTGC 99 consecutive Northern European Caucasian patients with (steatosis), seen only when the 1,074-patient strong combined primary NAFLD-related HCC was identified according to the cohort was studied, is arguably more modest than might be joint European Association for the Study of the Liver and expected. Our findings do support the previously reported 9,19,21 European Association for the Research and Treatment of Cancer association , although differences in sensitivity to subtle (EASL-EORTC) guidelines . TM6SF2 rs58542926 allele and changes in HTGC between radiological and histological genotype frequencies in this cohort were compared with the modalities may have reduced the power to detect this effect . combined NAFLD cohort described above (n ¼ 1,074). In Kozlitina et al. reported that the maximal effect of the TM6SF2 univariate analysis, homozygote carriage of the TM6SF2 variant in European Caucasians was only a mean 9.2% increase in rs58542926 minor allele was associated with an increased risk H-MRS quantified HTGC in TT homozygotes above the B5.9% of NAFLD–HCC with respect to CC (OR 1.922, 95%CI 1.31–2.81; observed in CC homozygotes . Histological assessment of P ¼ 6.81  10 ); however, significance was lost in multivariate hepatic steatosis uses broad microscopic categories reflecting Table 1 | Multivariate analysis of association between TM6SF2 rs58542926 genotype and fibrosis stage F0–1 (mild) versus F2–4 (advanced). Variables Discovery cohort (n ¼ 349) Validation cohort (n ¼ 725) Combined cohort (n ¼ 1,074) OR (95%CI) P-value OR (95%CI) P-value OR (95%CI) P-value 5  5 TM6SF2 genotype 2.94 (1.76–4.89) 3.44  10 1.46 (1.03–2.09) 0.0362 1.88 (1.41–2.5) 1.63  10 PNPLA3 genotype 1.57 (1.21–2.19) 0.0086 1.32 (1.05–1.66) 0.0183 1.40 (1.16–1.69) 4.84  10 Age 1.03 (1.01–1.06) 0.0045 1.02 (1.01–1.04) 0.0041 1.03 (1.01–1.04) 1.57  10 Gender (female) 0.94 (0.57–1.56) 0.8297 1.81 (1.30–2.50) 4.50  10 1.43 (1.09–1.89) 0.0096 4  5 BMI 1.05 (1.00–1.10) 0.0368 1.03 (1.01–1.05) 9.80  10 1.04 (1.02–1.05) 3.78  10 8  11 T2DM 2.39 (1.49–3.84) 0.0003 2.73 (1.93–3.88) 1.68  10 2.57 (1.95–3.39) 1.78  10 BMI, body mass index; CI, confidence interval; OR, odds ratio; T2DM, type 2 diabetes mellitus. Additive model including age, gender, BMI, T2DM and PNPLA3 rs738409 genotype as covariates. Discovery/validation/combined cohorts: stage F0–1 (mild) n ¼ 198/439/637, stage F2–4 (advanced) n ¼ 151/286/437. NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications 3 & 2014 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5309 the proportion of hepatocytes that are visibly steatotic (S0 o5%, Conditional analysis undertaken as part of the present study S1 5–33%, S2 33–66% and S3 466%) . Based on data from adds further weight to the assertion that the 19p13 signal is previous comparative modality analysis , the modest gene effect causally related to TM6SF2 and not NCAN, not only for HTGC as 19 19 size reported by Kozlitina et al. (less than a threefold increase in was previously reported but now also for stage of hepatic HTGC above normal) would likely be encompassed within the fibrosis. These findings therefore establish a new and important histological S1 bracket, and therefore may not be apparent clinical relevance to the recently described association between histologically. Combined with the relatively low minor allele TM6SF2 and NAFLD, and suggest that TM6SF2 should be 9,29,30 frequency in the background population, smaller cohorts may considered alongside PNPLA3 (refs 8,12) and GCKR , as one therefore have insufficient statistical power for an association to of a handful of genes so far identified that are associated not only become evident. with variations in hepatic triglyceride accumulation but also with The modifier effect of the TM6SF2 variant on grade of fibrogenesis . It is noteworthy that across all the cohorts studied, steatohepatitis (disease activity) was apparent in the initial the OR for advanced fibrosis conferred by each copy of the discovery cohort analysis; however, statistical significance was TM6SF2 variant carried was consistently of similar or up to not reached in the subsequent validation analysis. The validation twofold greater magnitude than that which was observed, or has 12,31 cohort comprised a mixture of patients recruited from previously been reported , for the widely studied PNPLA3 both hepatology and bariatric services and, although the rs738409 variant (Table 1). cohorts appear well matched histologically, the validation Carriage of the TM6SF2 variant was associated with increased cohort exhibited higher mean BMI levels (38.5 9.1 versus risk of progression to NAFLD–HCC in univariate analysis. In ± ± 34.5 5.7 kg m , Po0.0001), a younger mean age (47.6 12.4 contrast to PNPLA3 (ref. 18), this effect was not sustained when versus 51.5 12.0 years, Po0.0001), a greater female preponder- confounding factors including age, T2DM and presence of ance (56.1 versus 32.1%, Po0.0001) and a lower prevalence of underlying cirrhosis were included in the model. Carriage of T2DM (32.4 versus 46.1%, Po0.0001) than the discovery cohort the TM6SF2 variant therefore does not appear to further increase (Table 2). These factors may have impacted on our ability to HCC risk independent of its effect on fibrosis stage. It should, replicate the initial association with steatohepatitis in a multi- however, be noted that the NAFLD–HCC cohort contained only variate analysis. Further study of the variant in other patient 99 patients and so, combined with a relatively modest TM6SF2 cohorts and exploration of the functional effects of TM6SF2 on rs58542926 minor allele frequency, the current study had B70% inflammatory response will be needed to address this point. power to detect an effect if an additive genetic model and risk The key finding of the current study is that carriage of the similar to that seen for fibrosis is assumed (a ¼ 0.05) .An TM6SF2 rs58542926 C4T minor allele is unequivocally asso- association cannot therefore be completely excluded, but would ciated with an increased risk of advanced NAFLD-associated seem unlikely. Studies using larger cohorts of NAFLD–HCC hepatic fibrosis. This highly significant effect was consistently patients than are presently available will be required to provide demonstrated across all the cohorts studied and was independent sufficient power to study this further. of potentially confounding factors including gender, age at time The TM6SF2 rs58542926 c.449 C4T variant is a non- of biopsy, BMI, T2DM and PNPLA3 rs738409 genotype. synonymous change producing a glutamate to lysine amino-acid Table 2 | Demographic characteristics of patient cohorts. Discovery cohort Validation cohort Combined cohort NAFLD–HCC cohort Number 349 725 1,074 99 Ethnicity European Caucasian European Caucasian European Caucasian European Caucasian Gender (female) 147 (42.1%) 407 (56.1%) 554 (51.6%) 19 (19.2%) ± ± ± ± Age, years 51.5 12.0 47.6 12.4 48.9 12.4 70.5 8.0 ± ± ± ± BMI, kg m 34.5 5.7 38.5 9.1 37.2 8.3 31.9 6.7 T2DM (yes) 161 (46.1%) 235 (32.4%) 396 (36.9%) 68 (68.7%) Steatosis score* S0 5 (1.4%) 60 (8.3%) 65 (6.1%) — S1 99 (28.4%) 206 (28.4%) 305 (28.4%) — S2 166 (47.6%) 247 (34.1%) 413 (38.5%) — S3 79 (22.6%) 204 (28.1%) 283 (26.4%) — Activity score (composite hepatocyte ballooning and necroinflammation scores)* A0 81 (23.2%) 132 (18.2%) 213 (19.8%) — A1 65 (18.6%) 133 (18.3%) 198 (18.4%) — A2 101 (28.9%) 214 (29.5%) 315 (29.3%) — A3 64 (18.3%) 149 (20.6%) 213 (19.8%) — A4 31 (8.9%) 89 (12.3%) 120 (11.2%) — Fibrosis score F0 108 (30.9%) 277 (38.2%) 385 (35.8%) F1 90 (25.8%) 162 (22.3%) 252 (23.5%) Non-cirrhotic: 32 (32.3%) F2 55 (15.8%) 161 (22.2%) 216 (20.1%) F3 66 (18.9%) 75 (10.3%) 141 (13.1%) F4 (cirrhosis) 30 (8.6%) 50 (6.9%) 80 (7.4%) Cirrhotic: 67 (67.7%) BMI, body mass index; HCC, hepatocellular carcinoma; NAFLD, non-alcoholic fatty liver disease; T2DM, type 2 diabetes mellitus. *Steatosis and activity score data incomplete in 8 (0.7%) and 15 (1.3%) of samples, respectively. 4 NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications & 2014 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5309 ARTICLE substitution at residue 167 (Glu167Lys), which is highly of disease. These were unrelated patients with histologically characterized NAFLD, derived from a patient population originally identified as having conserved across mammals . First identified as part of a ultrasonographically detected bright liver and abnormal biochemical tests large-scale sequencing project, little is currently known about (alanine transaminase and/or gamma-glutamyl transferase). the biological function of the TM6SF2 protein product . A validation cohort of 725 consecutive European Caucasian patients Adenovirus-mediated short hairpin RNA knockdown of Tm6sf2 from centres in UK and mainland Europe with histologically characterized NAFLD of different stages of disease. Patients in this cohort were unrelated in mice has been shown to increase HTGC and reduce very low- patients with histologically characterized NAFLD, derived from a patient density lipoprotein (VLDL) secretion, suggesting that TM6SF2 population originally identified as having ultrasonographically detected bright activity is necessary for normal VLDL secretion, and that liver and abnormal biochemical tests (alanine transaminase and/or gamma- impaired TM6SF2 function causally contributes to NAFLD . glutamyl transferase) or identified as having evidence of NAFLD at the time of bariatric surgery. However, these in vivo studies were of too short a duration to adequately address the effects on steatohepatitis or fibrogenesis. Together, these comprised the combined cohort of 1,074 patients with Furthermore, previous experimental evidence has shown that histologically characterized NAFLD. Demographic and histological details are hepatic triglyceride accumulation may not itself be directly shown in Table 2. A description of the ‘healthy workers’ cohort recruited in the hepatotoxic. This was elegantly demonstrated in mice by silencing North East of the United Kingdom has previously been published . A separate cohort of 99 consecutive Northern European Caucasian patients hepatic gene expression of diacylglycerol O-acyltransferase 2 with primary HCC arising on a background of NAFLD was identified (Dgat2), a key enzyme mediating the conversion of free fatty acids (NAFLD–HCC cohort). The diagnosis of HCC was established histologically or to triglyceride . Rather than ameliorating steatohepatitis, the through non-invasive assessment according to the EASL–EORTC clinical practice consequent reduction in hepatocyte triglyceride synthesis was guidelines . The presence of NAFLD was determined through histological assessment of non-tumour liver tissue or, when biopsy was not clinically associated with increased fatty acid oxidation, particularly appropriate, through radiological evidence of hepatic steatosis. through Cyp2e1, leading to greater oxidative stress, cellular damage and higher serum transaminase levels . It is therefore tempting to speculate that the function of TM6SF2 and the Liver biopsy. Liver biopsy was performed under radiological guidance. Specimens (at least 1.6 cm length and 1.5 mm thick) were fixed in 10% neutral formalin mechanism through which TM6SF2 drives NAFLD-associated for evaluation and embedded in paraffin for histological examination. Tissue sec- hepatic fibrosis may be other than through increased triglyceride tions were stained with haematoxylin and eosin, impregnated with silver for accumulation. visualizing reticulin framework and stained with Sirius Red Fast Green for visua- In conclusion, the current study confirms that TM6SF2 is lizing collagen. Liver biopsies were reviewed by a single expert liver pathologist at associated with histologically defined NAFLD, and is the first each participating centre, unaware of clinical or genetic data. The degree of stea- tosis (S0–3), activity of steatohepatitis (A0–4) and stage of fibrosis (F0–4) were demonstration that this gene serves as a powerful modifier of scored according to the validated semi-quantitative SAF score (Supplementary hepatic fibrogenesis. That this gene is also associated with disturbed 28,37 Table 3) . In 25 HCC patients, the diagnosis of HCC was confirmed cholesterol metabolism and so may modify risk of cardiovascular histologically and graded according to Edmondson and Steiner , adapted for events including myocardial infarction suggests that TM6SF2 is needle biopsy specimens. an important determinant of clinical outcome across several facets of metabolic syndrome-related end-organ damage. In light of DNA preparation. Venous blood was collected from each patient and DNA was evidence that cholesterol accumulation in hepatic stellate cells prepared from peripheral blood lymphocytes using a perchlorate–chloroform promotes NAFLD fibrosis , it is tempting to speculate that isolation method . In brief, 35 ml lysis buffer (10 mM Tris–HCl (pH 8.0), 320 mM sucrose, 5 mM magnesium chloride and 1% Triton X-100) was added to 5 ml TM6SF2 may act as a ‘switch’ with TM6SF2 rs58542926 T-allele- venous blood in a 50-ml polypropylene centrifuge tube. After mixing, the tube was mediated hepatic retention of triglyceride and cholesterol centrifuged at 3,000 g for 10 min. The supernatant was discarded and the cell pellet predisposing to NAFLD fibrosis while C-allele carriage promotes was re-suspended in 2 ml of solution B (400 mM Tris–HCl (pH 8.0) 6 0 mM EDTA, VLDL excretion, protecting the liver at the expense of increased risk 150 mM NaCl and 1% SDS). A quantity of 500ml of sodium perchlorate (5 M) was of cardiovascular disease. These data mandate further mechanistic added and the sample was mixed at room temperature for 15 min before incubating in a preheated hot block at 65 C for 30 min. Next, 2 ml of chloroform was added study to determine the physiological and pathophysiological role of and the sample was mixed for 10 min at room temperature. The tube was then this gene in various tissues and cell types as a modifier of centrifuged at 1,400g for 10 min, and the upper, clear DNA-containing phase was fibrogenesis and a putative therapeutic target. transferred to a new 15 ml polypropylene tube. Two volumes of cold ethanol were added to the aqueous phase, and the tube was gently inverted until the DNA precipitated. The DNA was spooled using a soft plastic sterile loop and allowed to Methods air dry for 20 min. DNA was then re-suspended by incubation in 200ml water at Patients. Patients were recruited from hepatology clinics at several European 60 C. Samples are quantitated and quality assessed by absorbance measurements specialist centres: the Freeman Hospital, Newcastle upon Tyne, UK; Addenbrooke’s at 260 and 280 nm. Genotyping was performed by personnel unaware of clinical Hospital, Cambridge, UK; Nottingham University Hospitals NHS Trust, status or histology of patients. Nottingham, UK; Inselspital Hospital, Bern, Switzerland; Antwerp University Hospital, Belgium; and Pitie´-Salpeˆtrie`re Hospital, Paris, France. The study had all the necessary ethical approvals (UK: Newcastle and North Tyneside 1 REC TM6SF2 rs58542926, NCAN rs2228603 and PNPLA3 rs738409 genotyping. (10/H0906/41), Norfolk REC (06/Q0106/70)] and Nottingham 2 REC TM6SF2 rs58542926, NCAN rs2228603 and PNPLA3 rs738409 genotypes were (GM010201); Switzerland: Inselspital Bern Local Ethics Committee; Belgium: determined by allelic discrimination using TaqMan reagents (Applied Biosystems Antwerp University Hospital Ethics Committee; France: CPP (Comite´ de Inc., USA) according to the manufacturer’s protocol. Control samples of known Protection des Personnes) Paris VI IDF Pitie´—Salpeˆtrie`re Hospital). All genotype were also included in every 96-well plate (blank, homozygous wild-type, participants gave informed consent. In all cases, alternative diagnoses were homozygous mutant and heterozygous). excluded, including excess alcohol intake (alcohol intakeo20 g per day for women; and o30 g per day for men), chronic viral hepatitis (hepatitis B and hepatitis C), autoimmune liver diseases, hereditary hemochromatosis, a1-antitrypsin deficiency, Statistical analysis. Statistical analyses were performed using SPSS v19.0 (IBM, Wilson’s disease and drug-induced liver injury. Clinical and laboratory data were USA) to collate and analyse cohort phenotype data. Continuous variables were collected at the time of diagnosis including basic anthropometrics so that BMI tested using Student’s t-test/one-way analysis of variance and categorical variables could be calculated, and relevant co-morbidity including the presence of T2DM 2 by w -squared test unless otherwise stated. PLINK v1.07 (ref. 40) (via the gPLINK 1  1 (fasting glucose Z7.1 mmol l (Z128 mg dl ) or treatment with anti-diabetic v2.050 GUI) was used to conduct the genetic analysis. An initial univariate drugs) and evidence of underlying cirrhosis was recorded. The degree of steatosis 2 w -squared analysis was performed. Subsequently, multivariate logistic regression (S0–3), activity of steatohepatitis (A0–4) and stage of fibrosis (F0–4) were analysis was conducted incorporating biologically relevant covariates that were scored according to the validated semi-quantitative SAF score . The main study associated with risk of NAFLD progression (age, gender, BMI, presence of cohorts were: T2DM and PNPLA3 rs738409 genotype) to test the genetic association. An additive genetic model best fitted the data and was reported. Results are expressed as beta An initial discovery cohort of 349 consecutive European Caucasian patients from ± b s.e.m. or OR with 95%CI as appropriate. Significance was taken as Po0.05 the United Kingdom with histologically characterized NAFLD of different stages throughout. NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications 5 & 2014 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5309 References 28. Bedossa, P. et al. Utility and appropriateness of the FLIP algorithm and SAF 1. Anstee, Q. M., Targher, G. & Day, C. P. Progression of NAFLD to diabetes score in the evaluation of biopsies of nonalcoholic fatty liver disease. mellitus, cardiovascular disease or cirrhosis. Nat. Rev. Gastroenterol. Hepatol. Hepatology. doi:10.1002/hep.27173 (2014). 10, 330–344 (2013). 29. Anstee, Q. M. et al. A candidate-gene approach to validation of genetic modifier 2. Musso, G., Gambino, R., Cassader, M. & Pagano, G. Meta-analysis: associations using a large cohort with histologically characterised non-alcoholic natural history of non-alcoholic fatty liver disease (NAFLD) and diagnostic fatty liver disease. J. Hepatol. 58, S46–S46 (2013). accuracy of non-invasive tests for liver disease severity. Ann. Med. 43, 617–649 30. Petta, S. et al. Glucokinase regulatory protein gene polymorphism affects liver (2011). fibrosis in non-alcoholic fatty liver disease. PLoS ONE 9, e87523 (2014). 3. Ekstedt, M. et al. Long-term follow-up of patients with NAFLD and elevated 31. Valenti, L. et al. I148M patatin-like phospholipase domain-containing 3 gene liver enzymes. Hepatology 44, 865–873 (2006). variant and severity of pediatric nonalcoholic fatty liver disease. Hepatology 52, 4. Loomba, R. & Sanyal, A. J. The global NAFLD epidemic. Nat. Rev. 1274–1280 (2010). Gastroenterol. Hepatol. 10, 686–690 (2013). 32. Purcell, S., Cherny, S. S. & Sham, P. C. Genetic Power Calculator: design of 5. Hirschhorn, J. N. & Gajdos, Z. K. Genome-wide association studies: results linkage and association genetic mapping studies of complex traits. from the first few years and potential implications for clinical medicine. Annu. Bioinformatics 19, 149–150 (2003). Rev. Med. 62, 11–24 (2011). 33. Carim-Todd, L., Escarceller, M., Estivill, X. & Sumoy, L. Cloning of the novel 6. Altshuler, D., Daly, M. J. & Lander, E. S. Genetic mapping in human disease. gene TM6SF1 reveals conservation of clusters of paralogous genes between Science 322, 881–888 (2008). human chromosomes 15q24--4q26 and 19p13.3--4p12. Cytogenet. Cell 7. Anstee, Q. M. & Day, C. P. The genetics of NAFLD. Nat. Rev. Gastroenterol. Genet. 90, 255–260 (2000). Hepatol. 10, 645–655 (2013). 34. Yamaguchi, K. et al. Inhibiting triglyceride synthesis improves hepatic steatosis 8. Romeo, S. et al. Genetic variation in PNPLA3 confers susceptibility to but exacerbates liver damage and fibrosis in obese mice with nonalcoholic nonalcoholic fatty liver disease. Nat. Genet. 40, 1461–1465 (2008). steatohepatitis. Hepatology 45, 1366–1374 (2007). 9. Speliotes, E. K. et al. Genome-wide association analysis identifies variants 35. Tomita, K. et al. Free cholesterol accumulation in hepatic stellate cells: associated with nonalcoholic fatty liver disease that have distinct effects on mechanism of liver fibrosis aggravation in nonalcoholic steatohepatitis in mice. metabolic traits. PLoS Genet. 7, e1001324 (2011). Hepatology 59, 154–169 (2014). 10. Chambers, J. C. et al. Genome-wide association study identifies loci 36. Velaga, M. R. et al. The codon 620 tryptophan allele of the lymphoid tyrosine influencing concentrations of liver enzymes in plasma. Nat. Genet. 43, phosphatase (LYP) gene is a major determinant of Graves’ disease. J. Clin. 1131–1138 (2011). Endocrinol. Metab. 89, 5862–5865 (2004). 11. Al-Serri, A. et al. The SOD2 C47T polymorphism influences NAFLD fibrosis 37. Dyson, J. K., McPherson, S. & Anstee, Q. M. Non-alcoholic fatty liver disease: severity: evidence from case-control and intra-familial allele association studies. non-invasive investigation and risk stratification. J. Clin. Pathol. 66, 1033–1045 J. Hepatol. 56, 448–454 (2012). (2013). 12. Valenti, L. et al. Homozygosity for the patatin-like phospholipase-3/ 38. Edmondson, H. A. & Steiner, P. E. Primary carcinoma of the liver: a study of adiponutrin I148M polymorphism influences liver fibrosis in patients with 100 cases among 48,900 necropsies. Cancer 7, 462–503 (1954). nonalcoholic fatty liver disease. Hepatology 51, 1209–1217 (2010). 39. Daly, A. K., Steen, V. M., Fairbrother, K. S. & Idle, J. R. CYP2D6 multiallelism. 13. Tan, H. L. et al. Association of glucokinase regulatory gene polymorphisms Methods. Enzymol. 272, 199–210 (1996). with risk and severity of non-alcoholic fatty liver disease: an interaction study 40. Purcell, S. et al. PLINK: a tool set for whole-genome association and with adiponutrin gene. J. Gastroenterol. 49, 1056–1064 (2013). population-based linkage analyses. Am. J. Hum. Genet. 81, 559–575 (2007). 14. Dongiovanni, P. et al. Genetic variants regulating insulin receptor signalling are associated with the severity of liver damage in patients with non-alcoholic fatty Acknowledgements liver disease. Gut 59, 267–273 (2010). Q.M.A. is the recipient of a Clinical Senior Lectureship Award from the Higher 15. Aravinthan, A. et al. Gene polymorphisms of cellular senescence marker p21 Education Funding Council for England (HEFCE). Aspects of this work have been and disease progression in non-alcohol-related fatty liver disease. Cell Cycle 13, supported by the facilities of the Newcastle National Institute of Health Research (NIHR) 1489–1494 (2014). Biomedical Research Centre and by the ‘Fatty Liver Inhibition of Progression’ (FLIP) 16. Kawaguchi, T. et al. Genetic polymorphisms of the human PNPLA3 gene are project funded by the European Union Seventh Framework Programme (FP7/2007- strongly associated with severity of non-alcoholic fatty liver disease in Japanese. 2013) under grant agreement Health-F2-2009-241762. PLoS ONE 7, e38322 (2012). 17. Trepo, E. et al. Association between the PNPLA3 (rs738409 C4G) variant and hepatocellular carcinoma: evidence from a meta-analysis of individual Author contributions participant data. Hepatology 59, 2170–2177 (2013). Q.M.A., A.K.D. and C.P.D. conceived the research. Clinical phenotype data collation and 18. Liu, Y. L. et al. Carriage of the PNPLA3 rs738409 C4G polymorphism confers sample acquisition/DNA preparation was performed by J.B.S.L., Q.M.A., C.P.D., S.M., an increased risk of non-alcoholic fatty liver disease associated hepatocellular H.L.R., M.E.D.A., G.J.A., A.-C.P., P.D., G.P.A., S.F., L.V.G., K.C., V.R. and J.-F.D. carcinoma. J. Hepatol. 61, 75–81 (2014). Genotyping was performed by Y.-L.L., and assisted by J.B.S.L. and RA. Histological 19. Kozlitina, J. et al. Exome-wide association study identifies a TM6SF2 variant analysis of tissues and scoring was conducted by A.D.B. and D.T. Statistical analysis and that confers susceptibility to nonalcoholic fatty liver disease. Nat. Genet. 46, interpretation of results was performed by Y.-L.L., A.K.D. and Q.M.A. The manuscript 352–356 (2014). was written and revised by Q.M.A., A.K.D. and Y.-L.L. All authors critically reviewed the 20. Holmen, O. L. et al. Systematic evaluation of coding variation identifies a manuscript for important intellectual content and approved the final submitted candidate causal variant in TM6SF2 influencing total cholesterol and manuscript. myocardial infarction risk. Nat. Genet. 46, 345–351 (2014). 21. Gorden, A. et al. Genetic variation at NCAN locus is associated with inflammation and fibrosis in non-alcoholic fatty liver disease in morbid obesity. Additional information Hum. Hered. 75, 34–43 (2013). Supplementary Information accompanies this paper at http://www.nature.com/ 22. Dyson, J. et al. Hepatocellular cancer: the impact of obesity, type 2 diabetes and naturecommunications a multidisciplinary team. J. Hepatol. 60, 110–117 (2014). Competing financial interests: The authors declare no competing financial interests. 23. European Association For The Study Of The Liver; European Organisation For Research And Treatment Of Cancer. EASL-EORTC clinical practice Reprints and permission information is available online at http://npg.nature.com/ guidelines: management of hepatocellular carcinoma. J. Hepatol. 56, 908–943 reprintsandpermissions/ (2012). 24. Kathiresan, S. et al. Six new loci associated with blood low-density lipoprotein How to cite this article: Liu, Y.-L. et al. TM6SF2 rs58542926 influences hepatic fibrosis cholesterol, high-density lipoprotein cholesterol or triglycerides in humans. progression in patients with non-alcoholic fatty liver disease. Nat. Commun. 5:4309 Nat. Genet. 40, 189–197 (2008). doi: 10.1038/ncomms5309 (2014). 25. Teslovich, T. M. et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature 466, 707–713 (2010). This work is licensed under a Creative Commons Attribution 4.0 26. Anstee, Q. M., McPherson, S. & Day, C. P. How big a problem is non-alcoholic International License. The images or other third party material in this fatty liver disease? BMJ 343, d3897 (2011). article are included in the article’s Creative Commons license, unless indicated otherwise 27. McPherson, S. et al. Magnetic resonance imaging and spectroscopy accurately in the credit line; if the material is not included under the Creative Commons license, estimate the severity of steatosis provided the stage of fibrosis is considered. users will need to obtain permission from the license holder to reproduce the material. J. Hepatol. 51, 389–397 (2009). To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ 6 NATURE COMMUNICATIONS | 5:4309 | DOI: 10.1038/ncomms5309 | www.nature.com/naturecommunications & 2014 Macmillan Publishers Limited. All rights reserved.

Journal

Nature Communications – Springer Journals

Published: Jun 30, 2014

There are no references for this article.