Alimentary Pharmacology and Therapeutics

Systematic review with meta-analysis: do interferon lambda 3 polymorphisms predict the outcome of interferon-therapy in hepatitis B infection? † & P. Lampertico* E. Galmozzi*, M. Vigano

*1st Division of Gastroenterology, “A.M. e A. Migliavacca” Center for the Study of Liver Disease, Fondazione IRCCS Ca Granda Ospedale Maggiore Policlinico, Universita degli Studi di Milano, Milan, Italy. † Hepatology Division, Ospedale San Giuseppe, Universita degli Studi di Milano, Milan, Italy.

Correspondence to: Dr P. Lampertico, 1st Division of Gastroenterology, Fondazione IRCCS Ca’ Granda - Ospedale Maggiore Policlinico, Universita degli Studi di Milano, Via F. Sforza 35, 20122 Milan, Italy. E-mail: [email protected]

Publication data Submitted 11 September 2013 First decision 1 October 2013 Resubmitted 17 December 2013 Accepted 2 January 2014 EV Pub Online 26 January 2014 This commissioned systematic review was subject to full peer-review and the authors received an honorarium from Wiley, on behalf of AP&T.

SUMMARY Background Interferon lambda 3 (IFN-k3) polymorphisms are the strongest genetic predictor of outcome of hepatitis C virus infection and of response to Pegylated interferon (PegIFN)-based therapy. Whether this holds true for hepatitis B virus (HBV) infection is matter of controversy.

Aim To review the association between host genomics and spontaneous or interferoninduced clearance of HBV with specific reference to the recently identified interleukin 28B gene now renamed IFN-k3.

Methods A literature search was performed on MEDLINE, EMBASE and Web of Science for English articles and abstracts using free text words and combinations of the following terms ‘IL28B’, ‘IFN lambda’, ‘genomics’, ‘hepatitis B virus’, ‘interferon’ ‘GWAS’, ‘treatment’, ‘SNPs’, ‘HLA’, ‘polymorphisms’.

Results Genome-wide association studies convincingly demonstrated an association between SNPs in the HLA locus and spontaneous resolution of HBV infection in subgroups of Asian patients, yet no information is available for Caucasians. The preliminary observations of an association between IFN-k3 SNP and virological and serological responses to IFN in both HBeAg-positive and -negative patients could not be replicated by subsequent studies. Yet, majority of studies performed so far suffer several limitations in terms of sample size, selection of the patients, endpoints of therapy, treatment strategies and duration of follow-up.

Conclusions While host genetics is associated with an increased likelihood of spontaneous clearance of HBV among genotype B/C patients, the relationship between IFN-k3 polymorphisms and response to IFN has not been confirmed. Further studies in large cohorts of homogeneous patients are required, before this genetic test can be recommended in clinical practice.

Aliment Pharmacol Ther 2014; 39: 569–578

ª 2014 John Wiley & Sons Ltd doi:10.1111/apt.12631

569

E. Galmozzi et al. INTRODUCTION Chronic infection with the hepatitis B virus (HBV) is a major health problem worldwide, as documented by the 400 million people who are chronically infected by the virus causing more than 600 thousand deaths each year from cirrhosis and hepatocellular carcinoma (HCC).1–3 Anti-viral therapy proved to improve quality of life and survival of hepatitis B patients by preventing progression of liver damage to cirrhosis, hepatic decompensation, HCC and death.4–6 This goal, however, can only be achieved in a subset of patients, i.e. those achieving robust suppression of HBV replication in a sustained or maintained pattern following either a short-term ‘curative’ treatment with pegylated interferon (PegIFN) or long-term ‘suppressive’ therapy with third-generation nucleos(t)ide analogues (NUCs) like entecavir and tenofovir.7 In the face of several constrains like the need for weekly injections, flu-like symptoms and limited patient access, treatment with PegIFN has several advantages over NUCs including sparing virus resistance coupled with significant immunomodulatory effects halting virus replication and enhancing host’s immune control of HBV. Owing to such a synergistic effect, IFN therapy may result in a quarter of patients achieving a sustained remission of hepatitis with chances of subsequent HBsAg seroclearance, which represents the ideal endpoint of any anti-HBV treatment.8–12 This, however, comes at the expenses of the need for parenteral injections and a burden of clinical and laboratory monitoring, whereas patient’s quality of life is often compromised by the frequent onset of flu-like side effects and the disappointment due to the lack of effectiveness in a majority of patients. The latter has therefore fostered the search for pre-treatment predictors of a response to IFN with the obvious goal of optimising cost-effectiveness of IFN regimens through pre-treatment counselling. Ideal candidates for IFN therapy are those with low baseline HBV DNA and high alanine aminotransferase (ALT) levels, younger age, female gender and carriers of HBV genotype A and B.7 Unfortunately, the fluctuating pattern of viraemia and ALT values tends to make the prediction of a response by these variables at individual level cumbersome. At present, early on-treatment HBsAg kinetics is the only parameter that can reliably predict a nonresponse to PegIFN, while all the other treatment-related variables investigated so far have a disappointingly low positive predictive power.4 More recently encouraging data stem from studies of a baseline predictor of response using a genome-wide association (GWAS) analysis of SNPs rs12979860 and rs8099917 mapping in the genomic 570

region 3 kb upstream of the gene codifying for IL28B on chromosome 19, now renamed IFN lambda. These polymorphisms, in fact, were shown to correlate with the response to PegIFN in patients with chronic hepatitis B suggesting an analogy with hepatitis C where these polymorphisms are associated with a response to Peg IFN plus Ribavirin (Rbv) in patients with the difficult to cure genotype 1 of hepatitis C virus (HCV).13–17 The fact that the variability in this genomic region in hepatitis C patients carrying different responses to IFN treatment was confirmed in subjects of different ethnicity was taken as a proof to explain, at least partially, the differences in virological response rates registered across different ethnicities. All this has boosted the interest also in the setting of hepatitis B with the obvious aim of identifying IFN-sensitive patients with a high probability of success. The biological plausibility of the association between IFN lambda, host genetic background and response to IFN lays in the critical role that immunity plays in the pathogenesis of HBV disease, thereby accounting for the clinical heterogeneity of hepatitis B. Although it is well established that the clinical course of acute infection with HBV is conditioned by the immune system’s response to HBV, chronic infection with HBV is considered a multifactorial disorder driven by various variables such as the host genetic background, virus biology and environment factors.18 The immune response is orchestrated by the human leucocyte antigen (HLA) class I and class II molecules in charge of presenting viral antigens to CD8+ cytotoxic T cells and CD4+ helper T cells respectively. Here, we review both studies describing the association between different IFN lambda SNPs and response to IFN-based therapy and on HLA alleles’ polymorphism relating to spontaneous clearance of HBV.

LITERATURE SEARCH A literature search was performed on MEDLINE, EMBASE and Web of Science for English articles and abstracts. Search terms included free text words and combinations of the following terms ‘IL28B’, ‘IFN lambda’, ‘genomics’, ‘hepatitis B virus’, ‘interferon’ ‘GWAS’, ‘treatment’, ‘SNPs’, ‘HLA’, ‘polymorphisms’. HOST GENETICS AND SPONTANEOUS CLEARANCE OF HBV GWAS studies Four GWAS comparing host genetics of HBV carriers with healthy controls and patients spontaneously clearing HBV infection, and patients with and without progressive Aliment Pharmacol Ther 2014; 39: 569-578 ª 2014 John Wiley & Sons Ltd

Systematic review with meta-analysis: host genomics and hepatitis B virus HBV-related liver disease, have been published between 2009 and 2012 (Table 1). In the first study, a significant association between chronic hepatitis B and 11 SNPs located within or around the HLA-DPA1 and HLA-DPB1 region was identified comparing 786 Japanese patients and 2201 controls (discovery phase).19 The two more significant SNPs identified in the discovery-phase analysis from each HLA-DP locus, namely HLA-DPA1rs3077 and HLA-DPB1rs9277535, were validated in three independent cohorts of patients including two sets of Japanese case–control samples (overall 992 case and 1554 controls) and a Thailand group (308 Patients and 546 controls). An internal meta-analysis confirmed the strong association of the two SNPs (combined P = 2.31 9 10 38 for rs3077 and P = 6.34 9 10 39 for

rs9277535) (Table 1). Furthermore, by case–control analysis based on direct sequencing of exon 2 of HLA-DPA1 and HLA-DPB1 genes, four haplotypes were identified: DPA1*0103-DPB1*0402 and DPA1*0103-DPB1*0401 showed protective effects (OR 0.52; 95% CI 0.35–0.75, p = 6.00 9 10-8 and OR 0.57; 95% CI 0.33–0.96, p = 0.002 respectively), whereas DPA1*0202-DPB1*0501 and DPA1*0202-DPB1*0301 were associated with susceptibility to chronic hepatitis B (OR 1.45; 95% CI 1.16– 1.81, P = 5.79 9 10 6 and OR 2.31; 95% CI 1.39–3.84, P = 0.002 respectively). A second GWAS study, which included 438 Japanese patients with persistent HBV infection and 2056 controls, not only confirmed the significant association between HLA-DP locus and HBV but also demonstrated another significant relationship with the SNPs rs2856718 and

Table 1 | GWAS and natural history of HBV in East Asian populations First author, Journal (Ref)

Reported gene

Ancestry

Sample size

Kamatani Y, NatGenet19

Japanese and Thai

2086 CHB cases vs. 4301 healthy controls

HLA-DPA1 HLA-DPB1

Mbarek H, Hum Mol Genet20

Japanese

2667 CHB cases vs. 6496 healthy controls

HLA-DPA1 HLA-DPB1 HLA-DQB1 HLA-DQB2

Nishida N, PLoSOne21

Japanese and Korean

HLA-DPA1 HLA-DPB1 HLA-DPB1 HLA-DPA1

Liu L, Viral Immunol22

Chinese

781 HBV carriers vs. 571 healthy controls 781 HBV carriers vs. 441 spontaneously HBV-resolved individuals 1944 HBV carriers with progressive liver disease vs. 854 asymptomatic HBV carriers

GRIN2A

Strongest SNP (Allele) [MAF] rs3077(A/G) [0.45] rs9277535(A/G) [0.44] rs3077(A/G) [0.40] rs9277535(A/G) [0.43] rs2856718(A/G) [0.48] rs7453920(A/G) [0.17] rs3077(A/G) [0.44] rs9277542(A/G) [0.45] rs9277542(A/G) [0.45] rs3077(A/G) [0.44] rs11866328(G/T) [NA]

OR or beta-coefficient [95% CI]

P value* 2.31 9 10 6.34 9 10

2 3 4 6

9 9 9 9

10 10 10 10

§ §

38

39

§ § 37 § 28 §

61

54

0.56 [0.51–0.61]† 0.57 [0.52–0.62]†

1.87 [1.73–2.01]‡ 1.77 [1.65–1.91]‡ 1.56 [1.45–1.67]‡ 1.81 [1.62–2.01]‡

4.40 9 10 19§ 1.28 9 10 15§ 1.60 9 10 4¶ 6.00 9 10 7¶

0.46 [0.39–0.54]† 0.50 [0.43–0.60]† 0.51 [0.36–0.72]† 0.55 [0.43–0.69]†

2 9 10

1.68 [1.40–2.02]

8

**

MAF, minor allele frequency; NA, not available. * Combined P value derived from internal meta-analysis except for ref.

22

† Odds ratio of minor allele from two-by-two allele frequency table. ‡ OR and CI are calculated using the nonsusceptible allele as reference. § Association with CHB. ¶ Association with HBV clearance. ** Progression of HBV. Aliment Pharmacol Ther 2014; 39: 569-578 ª 2014 John Wiley & Sons Ltd

571

E. Galmozzi et al. rs7453920 within the HLA-DQ locus. Using a replication set of three independent Japanese cohorts (2209 patients and 4440 controls), these two new HLA-DQ SNPs were shown to independently predict susceptibility to chronic HBV infection even after stratification with HLA-DPA1 rs3077 and HLA-DPB1rs9277535 (Table 1). Two haplotypes, DQA1*0102-DQB1*0604 and DQA1*0101DQB1*0501, were shown to be protective (OR 0.16; 95% CI 0.09–0.29 and OR 0.39; 95% CI 0.24–0.65 respectively), whereas two others, DQA1*0102-DQB1*0303 and DQA1*0301-DQB1*0601, were linked to chronic HBV infection (OR 19.03; 95% CI 2.53–143.39 and OR 5.02; 95% CI 1.87–13.45 respectively), indicating that variants in antigen-binding regions of HLA-DP and HLA-DQ may contribute to the risk of persistent HBV. In the third GWAS study,21 which was conducted in 181 HBV carriers, 185 HBV-resolved individuals and 185 healthy controls from Japan, the significant association between HLA-DPA1 and HLA-DPB1 loci, with protective effects against chronic hepatitis B, was confirmed. Moreover, using replication analysis in two independent sets of Japanese and Korean individuals (600 HBV carriers, 256 individuals with resolved infection and 387 healthy controls), the authors found that rs3077 and rs9277542 SNPs were associated with HBV clearance, although the rs9277535 showed a strong linkage disequilibrium with rs9277542 (r2 = 0.955) (Table 1). The fourth and last GWAS was carried out by Liu and colleagues22 with the aim to identify loci associated with the progression of hepatitis B. DNA pooling from two populations, 628 asymptomatic HBV carriers and 1729 carriers with progressive hepatitis B recruited from Hubei Province in south China, and 226 asymptomatic HBV carriers and 215 progressors recruited from Shandong Province in north China, were studied. The variant rs11866328 (G/T), located in the glutamate receptor ionotropic N-methyl D-aspartate 2A (GRIN2A) gene, showed a significant association with disease progression (Table 1).

GWAS validation and implementation The above-mentioned genetic associations were validated by three other studies performed in Asian patients.23–25 In addition, a recent study from Taiwan asked the question whether the rs3077 and rs9277535 SNPs play a role not only in spontaneous recovery of acute HBV infection but also on spontaneous HBsAg clearance in long-term follow-up patients with HBV.26 By comparing 100 HBsAg-positive HBeAg-negative carriers with 100 HBeAg-negative patients who spontaneously cleared 572

HBsAg over 6 years of follow-up, patients with rs9277535 non-GG genotypes were found to have higher chances to clear HBsAg than AA or AG patients (OR 1.83; 95% CI 1.04–3.21, P = 0.034). Moreover, more carriers of the rs3077 and rs9277535 GA haplotype were found to clear HBsAg than carriers of the GG haplotype (OR 2.17; 95% CI 1.14–4.16, P = 0.030). In two additional studies, the predictive power of HLASNPs and other polymorphisms, such as IL28B and TNF, was evaluated. In 203 patients achieving spontaneous HBsAg seroclearance and 203 age- and sex-matched persistently infected patients, Seto et al.27 demonstrated that spontaneous HBsAg seroclearance was linked to both the haplotype block GAT of rs3077/rs9277378/ rs3128917 (OR 2.17; 95% CI 1.06–4.45, P = 0.034) and the IL28B haplotype block CT of rs12979860/rs8099917 (OR 10.5; 95% CI 1.33–82.5, P = 0.026). In the second study, the association between HLA polymorphisms (rs361525, rs1800629, rs1799724, rs1800630) and TNF rs1799964 genotype was analysed in a South Indian population28 including 150 individuals who spontaneously cleared HBsAg and 137 patients with chronic HBV infection. The rs1800630 genotype was associated with HBV outcome in both codominant and dominant models after adjusting for age and sex. Similarly, the rs1799964 genotype was associated with HBV outcome in codominant (OR 1.57; 95% CI 1.09–2.27, P = 0.01) and dominant (OR 2.21; 95% CI 1.27–3.83, P < 0.01) models. Moreover, haplotype analysis (rs1799964/rs1800630/ rs1799724/rs1800629/rs361525) revealed that the CACGG haplotype was strongly associated with CHB infection (P = 0.0004) suggesting that inheritance of HLA and TNF polymorphisms might influence the outcome of HBV infection in the South Indian population. Overall, the association between SNPs in different locus (HLA, TNF, IL28B, GRIN2A) with HBV clearance and HBV progression has been confirmed in several countries in East Asia, including China, Thailand, Japan, Korea and South India. However, no data are so far available for Caucasian and African populations.

HOST GENETICS AND RESPONSE TO IFN THERAPY HBeAg-positive patients The study of genetic control of IFN response in these patients may be difficult owing to the interference that both HBV genotype and modality of infection exert over the patient immune response. A Taiwanese study conducted in 115 patients (63% genotype B and 37% genotype C) mostly treated with PegIFNa-2a for 6 months Aliment Pharmacol Ther 2014; 39: 569-578 ª 2014 John Wiley & Sons Ltd

Systematic review with meta-analysis: host genomics and hepatitis B virus evaluated the potential influence on treatment response of pre-core (PC) stop codon, basal core promoter (BCP), IL28B (rs8099917) and of two genetic variations of HLA-DPA1 (rs3077) and HLA-DPB1 (rs9277535).29 At baseline, 90% had rs8099917 TT genotype, 59% had rs3077 GG genotype and 53% had rs9277535 GG. The 6-month off-therapy HBeAg seroconversion and combined response rates, defined as HBeAg seroconversion plus HBV DNA

Systematic review with meta-analysis: do interferon lambda 3 polymorphisms predict the outcome of interferon-therapy in hepatitis B infection?

Interferon lambda 3 (IFN-λ3) polymorphisms are the strongest genetic predictor of outcome of hepatitis C virus infection and of response to Pegylated ...
221KB Sizes 0 Downloads 0 Views