Understanding the Host Genetics of Chronic Hepatitis B and C

[Guest guest]

Understanding the Host Genetics of Chronic Hepatitis B and C

Abstract

The outcome of hepatitis B virus (HBV) and hepatitis C virus (HCV) infections are heterogeneous, ranging from an asymptomatic self-limiting infection to cirrhosis and hepatocellular carcinoma. Several viral environmental and demographic variables have been identified as determinants of disease outcome, but these fail to explain a large proportion of the variability. Evidence from twin studies suggests that the host genetic background is an important contributor to disease outcome. Identification of genes that influence the outcome of infection has been attempted using a wide spectrum of approaches including candidate gene disease association studies, genome-wide scanning in affected sibling pairs and most recently genome-wide association studies. We summarize the main findings from a large number of studies in this review. Many studies have focused on the MHC loci from which several reproducible disease associations have been identified. More

recently, genome-wide association studies have identified an important locus within the IL-28 - Il-29 region on chromosome 29, which appears to be a major determinant of the treatment response in patients infected with HCV and also a determinant of spontaneous resolution of infection. Translation of the genetic architecture of chronic viral hepatitis into therapeutic opportunities has been slow to proceed. One clinical trial and one drug development program have been based on genetic discoveries. The use of IL-28B genotyping to predict the response to pegylated interferon and ribavirin may also find its way into clinical practice. Indeed, stratification of clinical trial populations based on IL-28B genotype is already considered mandatory.

Introduction

The outcome of both hepatitis B virus (HBV) infection and hepatitis C virus (HCV) infection are markedly heterogeneous, varying from acute asymptomatic self-limiting infection to fulminant hepatic failure to decompensated cirrhosis and hepatocellular carcinoma. Diversity in viral sequence and genotype, environmental variables, and age at the time of infection all contribute to the heterogeneity of outcome, but cannot account for all of the variability. It has been recognized for many years that the genetic background of the host influences the outcome of a viral hepatitis infection. Although this statement is generally accepted, the evidence supporting it is not as strong as assumed. Support for the theory that the host's genetic background influences the outcome of infection arises from studies showing clustering of similar outcomes (e.g., chronicity of infection) in the same family. The strongest evidence for a host genetic effect is based on twin

studies from Taiwan in HBV infection, which showed higher levels of concordance for HBsAg carriage in monozygotic twins compared with dizygotic twins[1] Although it is a somewhat circular argument, the discovery of reproducible genetic associations is also now accepted as evidence for the genetic impact of the host.[2]

There are many aims behind the study of genetic susceptibility in viral hepatitis; these include elucidation of novel mechanisms of disease pathogenesis, development of biomarkers for disease prognosis or treatment outcome, and identification of potential therapeutic targets. These aims should be born in mind when assessing the utility of the genetic discoveries outlined below.

Investigations of specific genetic determinants that influence the outcome of infection have used three or four standard approaches. Genetic association studies are usually based on a single (occasionally a cluster) candidate gene whose candidacy is based on our prior knowledge of the disease process or occasionally on variation in the level of gene expression in suitable tissues or cell lines. In these studies, genetic variants are identified through database searches or sequencing and the allele or genotype frequencies compared between cases and suitable controls. Although this type of study is statistically powerful and effective it has been argued that the approach fails to deliver novel insight into the disease process as the selection of candidate genes is based on existing knowledge about the gene or gene product. Genome-wide approaches to causal gene identification may be based on either genetic linkage in families or genetic association in

populations of unrelated cases. Genome-wide linkage studies look for chromosomal regions inherited more often than expected from Mendelian genetics in affected pairs of first-degree relatives affected by the disease outcome of interest. Genome-wide association studies (GWAS) examine the association of large numbers of genetic variants with a specific disease phenotype without any prior hypothesis. Advances in genotyping technology and large databases of genetic variants make it possible to search for new genetic influences across the entire human genome.

Candidate Gene Discoveries

MHC Class II

Hepatitis B Virus Spontaneous clearance of an HBV infection is associated with vigorous polyclonal and multispecific CD4 + T-helper cell responses in contrast to the weak responses seen in persistent infection.[3,4] As the human leucocyte antigen (HLA) molecules encoded by MHC class II are responsible for presenting peptide epitopes to CD4 + T helper cells, it is reasonable to postulate that polymorphism in the MHC class II region may explain the variation in outcome. Furthermore, failure of HBV vaccination, defined by an anti-HBs titer < 10 IU/L after three doses of HBsAg vaccination, is thought to arise through a lack of T cell help for anti-HBs-producing B cells. Therefore, failure to develop anti-HBV immunity following HBV vaccination may also be influenced by MHC class II polymorphism. Vaccine nonresponse is consistently associated

with HLA-DRB1*0301 and *0401 in many populations.[5–9]

MHC class II association studies in persistent HBV infection are summarized in Table 1. The HLA DRB locus alleles DRB1*1301/2 are consistently associated with spontaneous resolution of infection in sub-Saharan African, Asian, and Caucasian populations.[10–12] The associations of HLA-DR7 (DRB1*07-) and HLA-DR3 (DRB1*0301) with persistent infection[13,14] are interesting because DRB1*0701 and DRB1*0301 have also been associated with failure to respond to HBsAg-based vaccine.[15] Understanding the mechanisms underlying these associations is important because it will

define whether (1) T- and B-cell response to HBsAg is critical to the development of persistent infection, and (2) individuals who fail to respond to the vaccine may be more susceptible to persistent infection. DRB1*0901, DQA1*0301, DQA1*0501 and DQB1*0301 are consistently associated with persistent HBV infection in different ethnic populations.[12,16,17]

It is not yet clear what mechanism underlies the association of HLA-DRB1*1302 with spontaneous viral clearance. It is possible that the quality or magnitude of the T-cell response induced by HLA-DRB1*1302 is superior to other alleles. This is supported by work in mice indicating that different H2 backgrounds influence the character and magnitude of the T-cells response to HBV antigens.[18] Furthermore, CD4 + T cell responses in patients who have recovered from HBV are greater in those who carry the HLA-DRB1*1302 allele.[19] Alternatively, HLA-DRB1*1302 may present a wider range of epitopes than other alleles. Elution and binding studies suggest that HLA-DRB1*1302 is a relatively promiscuous peptide binder in comparison to other HLA class II

molecules.[20,21] The ability to bind and present a wide range of epitopes is likely to generate a polyclonal and multispecific T-helper cell response, as seen in individuals who spontaneously eliminate the infection. Furthermore, a broad range of potential T cell epitopes would reduce the opportunity for the virus to evade recognition through sequence variation. The importance of this ability to present a wide range of epitopes is further underlined by the finding that individuals who are heterozygous at MHC class II loci are less likely to develop persistent HBV infection than homozygous individuals.[22]

Hepatitis C Virus Consistent associations have been observed between MHC alleles and HCV outcomes.[23,24] Of these the most interesting finding has been the association of the class II allele DQB1*0301 and self-limiting HCV. Several studies in several populations have reported an association between DQB1*0301 with viral clearance, although one study reported an opposite effect.[25] The data are compelling because the DQB1*0301 allele, along with different extended haplotypes, has been associated with self-limiting HCV in different populations. A French study found associations between the DRB1*1101-DQA1*0501-DQB1*0301 haplotype and HCV clearance.[26] A UK study found the DRB1*0401-DQA1*03-DQB1*0301 haplotype associated with clearance.[27] In Italy, two studies reported an association for DRB1*1104*DQB1*0301 and clearance.[28,29] The association of the DQB1*0301 allele with self-limiting HCV in the presence of different DRB1 alleles, suggests that this particular allele, rather than the corresponding haplotypes, may play an important role in the natural clearance of HCV viremia.[30]

Another allele that is associated with HCV clearance is DRB1*1101.[23,26,31] An Italian study found a protective role for the DR5 serogroup against HCV infection.[32] The molecularly defined DRB1*1100 and DRB1*1200 group of alleles are part of the serologically defined DR5 group. The DRB1*1101 allele is also associated with susceptibility to vertically transmitted HCV infection.[33]

In a meta-analysis of the effects of DQB1*0301 and DRB1*11 employing molecularly genotyped studies conducted among Caucasians, the DQB1*0301 had a relatively strong correlation with self-limiting HCV infection (summary estimates of 3.0 (95% CI: 1.8-4.8) and 2.5 (95% CI: 1.7-3.7), for DQB1*0301 and DRB1*11, respectively).[34]

Compared with the association of MHC alleles with self-limiting infection association studies in treatment responsiveness have yielded conflicting results.[34] In a study by Kikuchi, the HLA-DR4 group was associated with nonresponse,[35] whereas another study observed the DRB1*0404 allele (a subset of the DR4 serogroup) to be associated with response.[36] One major explanation for the observed differences between these two studies is that the populations were very different with respect to race. Additionally, the study by Sim did not take differences in viral genotypes into account.

Studies of the MHC and the progression or severity of HCV have largely been inconsistent.[34] However, there is the suggestion that there might be a trend with DRB1*11 alleles and "less severe" liver disease. Three studies reported associations with normal ALT: DRB1*1104 (OR = 4.82),[37] DRB1*11 (OR = 2.36),[38] and DRB1*1101 (OR = 0.3).[39] Haruna observed the DRB1*1101 allele to be associated with less piecemeal necrosis, and Hüe observed DRB1*11 to be associated with lower Knodell scores (OR = 0.35).[40,41] Yasunami reported an association between DRB1*1101 and lower histological activity index (HAI) scores.[42] Tillman also observed DRB*11 to be lower among patients with cirrhosis (RR = 0.29).[43] Haruna reported an association between DRB1*1201 with less severe liver disease as well. It is interesting to note that the DRB1*1100 and DRB1*1200 alleles form the DR5 serogroup. Tokushige observed the combination of DRB1*0901 and the TNF-β B1 homozygous genotype associated with inactive HCV, whereas DRB1*0405 along with the TNF-β B2 homozygous genotype associated with active HCV.[44] The functional significance of these alleles is not currently known. The definition of "less severe disease" in these studies also varies greatly, ranging from biochemical (i.e., ALT levels) to histological definitions. Future studies are needed to refine these observations. Most studies reviewed did not adjust for potential environmental confounders, such as alcohol use.

HLA Class I

Although MHC class II is important for CD4 T cell responses, MHC (HLA) class I controls both CD8 T cell and natural killer (NK) cell function. Both of these cell types are able to interact directly with infected hepatocytes. CD8 T-cells respond to viral peptides of 8 to 10 amino acids presented by HLA class I molecules. NK cells are thought to respond to downregulation of HLA class I, but recent work has shown that they are also sensitive to changes in HLA class I bound peptides.[45] Viral infections are cleared by CD8-specific T cells. These cells recognize virus-derived peptide fragments presented in the context of HLA class I. However, HLA class I, it is the most polymorphic gene family within the mammalian genome and multiple alleles have similar peptide-binding specificities, thus immunogenetic studies of HLA class I often show weak and

population-specific associations. This is evident in HBV infection in which relatively few associations of HBV clearance with specific HLA class I alleles have been found, despite the well-described role for an HLA class I restricted cytotoxic T-lymphocyte (CTL) response.[46] HLA-A*0301 was associated with clearance and HLA-B*08 and HLA-B*44 were associated with persistence in multivariate model considering all alleles associated with outcome by univariate analysis.[47]

Resolution of HCV infection is also associated with a multispecific T-cell response; however, in general it is weaker than in HBV infection. Several studies have identified HLA class I alleles as being associated with resolution of HCV infection, including HLA-A, HLA-B, and HLA-C alleles. CTLs appear to be directed most strongly against peptides restricted by HLA-B allotypes[48] and HLA-B is the most diverse of the classical HLA class I loci. Resolution of HCV was originally associated with HLA-A*1101, HLA-B*57, and HLA-Cw*0102, whereas in this study persistence was associated with HLA-A*2301 and HLA-Cw*04.[49] Population diversity and diversity of the infecting inoculums may to some extent account for the difficulties in finding genetic

commonalities. These issues are to some extent addressed by the study of single-source outbreaks. In a cohort of an Irish infected form of a single source of rhesus anti-D immunoglobulin, it was found that the alleles HLA-A*03, -B*07, -B*27, and Cw*01 were associated with resolution of infection.[50] Interestingly, HLA-B*27 presents the viral epitope ARMILMTH, which generates a strong CTL response. Furthermore, mutation of this epitope has a high fitness cost to HCV, and hence may explain the importance of this HLA allele.[51] Additionally, this epitope is different in genotype 3 HCV (VRMVMMTHF), these individuals do not make a strong CTL response to this peptide variant and the frequency of HLA-B27 is higher in individuals with a chronic genotype 3a

infection as compared with those with chronic genotype 1 infection.[52] A similar correlation between genetics and function has been found for HLA-B*57. In these studies, this allele was associated with resolution of HCV infection and the authors were able to associate chronicity of HCV infection with mutations in this epitope.[53] Several of these findings have been confirmed in a large multiracial cohort of American women typed to the molecular level.[54] Consistent with the work of Kim et al,[53] this study found that HLA-B*5702 and HLA -B*5703 were

protective alleles, and also demonstrated that HLA-Cw*0102 was protective as originally shown by Thio et al,[49] and which also had a positive odds ratio in the study by Kim et al.[53]

HLA Class I and KIR

In addition to engaging CD8 + T-cells HLA class I also engages NK cells via their killer cell immunoglobulin-like receptors (KIR). These are a multigene family on chromosome 19, which like their HLA class I ligands, exhibit substantial population diversity. This diversity can be at the level of the locus as different individuals have different genes or at the allelic level. The KIR genes encode activating or inhibitory NK cell receptors. These engage groups of MHC class I alleles or "supertypes." For instance, HLA-C alleles can be divided into two different types: group 1 and group 2 based on whether amino acid 80 of the MHC class I heavy chain is either asparagine or lysine, respectively. Group I HLA-C alleles bind one group or KIR and group another. Surprisingly, it is the inhibitory receptors that appear to be most important in resolving HCV infection. The inhibitory receptor KIR2DL3 in combination with its group 1 HLA-C ligands is protective in

HCV infection in several scenarios.[55–57] These include spontaneous resolution, treatment-induced resolution, and individuals exposed to HCV through multiple episodes of high-risk behavior, who remain anti-HCV and HCV RNA negative: the so-called exposed uninfected individuals. It is thought that the interaction of KIR2DL3 with its group 1 HLA-C ligand is relatively weak, compared with other KIR:HLA-C interactions and therefore can be more easily perturbed, thus reducing the inhibitory signals to NK cells and allowing activating signals to dominate.[58] Of note is that this protection has been seen in intravenous drug users only when the transmitted viral load is thought to be lower than in those acquiring HCV through infected blood products. This

appears to be consistent with NK cells as a first line of defense, i.e., those whose innate immune systems are overwhelmed by large inocula of virions require a more vigorous CTL response; hence, specific combinations of NK cell receptors may not be protective. Furthermore, the allele HLA-Cw*0102, which is protective in HLA association studies of HCV infection, is one of the group 1 HLA-C alleles. In addition to their antiviral function, NK cells also have antitumor activity and the activating receptor KIR3DS1 in combination with its HLA-BBw4 ligand is protective against HCC in chronic HCV infection. Interestingly, a similar protective association of KIR2DL3 and group 1 HLA-C, and susceptibility for KIR2DL1 and group 2 HLA-C has recently been described for HBV infection.[59]

Cytokines

Critical to developing a successful CD8- and NK-cell response are the secretion of TH1-type cytokines from immune cells. These include TNFα, IFNγ, and IL-12. Conversely, secretion of TH2 type cytokines, such as IL-10, leads to a weaker CD8 T cell response. Polymorphisms within cytokine genes have been variable associated with the outcome of both HBV and HCV infection (Table 2).

Type I Cytokines

Hepatitis B Virus Tumour necrosis factor α (TNFα) is a proinflammatory cytokine secreted mainly by macrophages that signal through a cell surface receptor to activate NF-kB. A model of HBV replication has been established in mice where a dimer of the HBV genome was inserted as a transgene into mice. In this mouse, HBV replication is suppressed or terminated by adoptive transfer of HBV-specific cytotoxic T lymphocytes (CTL).[60] The numbers of CTL required to suppress HBV replication was found to be small relative to the number of infected hepatocytes (100% in the transgenic model). Accordingly, noncytolytic mechanisms were believed to be responsible for controlling viral replication. Antibodies to TNFα were found to prevent the effects of CTL on HBV replication, whereas infusion of TNFα mimicked the effect of the CTL.[61] Therefore, it is likely that TNFα is important in both generating immune and inflammatory responses and directly activating hepatocyte antiviral processes.

Höhler et al reported that possession of the TNFα -238A allele was associated with persistent infection,[62] although we have been unable to replicate these findings among Caucasians. However, larger studies that are more comprehensive with respect to the number of polymorphisms studied are needed before conclusions about TNFα polymorphisms may be drawn.

Among Koreans, alleles that are associated with high levels of TNFα in plasma or increased transcriptional efficiency confer resistance to persistent infection: TNFα-308A allele and HBV persistence (OR = 0.57, P = .01), and TNFα-863A with HBV clearance (OR = 1.52, P = .003).[63] These findings are consistent with the role of TNFα in the noncytolytic control of HBV replication in vivo.

Hepatitis C Virus Elevated levels of TNF have been found in the serum as well as the liver of individuals infected with HCV, and individuals with response to interferon have reportedly lower pretreatment levels of TNF mRNA.[64] Despite the controversies concerning the functional consequences of the polymorphic variants in the TNF gene, the biologic significance of TNF in the human response to infectious agents is well established.

With respect to self-limiting infection, Höhler found an association between the TNF-238A allele and chronicity in an European population.[65] Other studies conducted in Caucasian populations have failed to reproduce this association.[66,67] In contrast, a study conducted in both white and black individuals found the -863A allele associated with self-limiting HCV infection in blacks only.[68] Among whites, none of the TNF alleles studied were associated with self-limiting infection.

Several studies, conducted among predominantly Caucasian populations, demonstrated a lack of association between TNF variants and the response to anti-HCV therapy.[64–70] An additional study conducted in a Taiwanese population also found no association with respect to the response to interferon therapy.[71]

TNF variants have also been studied with respect to the progression of HCV-related liver disease and the results have been discrepant.[65,72] The functionality of TNF variants remains controversial, and these apparent discrepancies may help explain, at least in part, the varied findings arising from studies of TNF variants.[73]

IFNγ is another type 1 cytokine liberated by both T cells and NK cells. In addition to generating a strong cytolytic T cell response it may also have direct antiviral activity.[74] A polymorphism in the promoter region at position -764 has been associated with both spontaneous and treatment-induced recovery from HCV infection.[75] Interestingly, the protective allele -764G had a higher binding affinity for the transcription factor NFκβ, leading to increased activity of the promoter, implying it is associated with higher levels of IFNγ.

Another cytokine associated with a strong CTL response is IL-12. This cytokine is liberated by dendritic cells and is composed of two subunits. A polymorphism in the p40 subunit of IL-12 has been associated with spontaneous clearance of HCV and is also overrepresented in a UK cohort of multiply exposed uninfected individuals.[76–78] However, this association was not confirmed in a German population[79]

Mosbruger et al genotyped 1536 SNPs in 112 selected immune response genes in a U.S. cohort of 343 individuals that spontaneously cleared HCV infection and 547 individuals with persistent HCV infection.[80] They found that polymorphisms in the IL-18 binding protein were associated with outcome of HCV infection. This protein binds both IL-18 and IFNγ; hence, it can regulate the immune response. Consistent with this, regulatory polymorphisms in IL-18 had previously been associated with clearance of HCV infection.[81]

Type 2 Cytokines

IL-10 is one of the key cytokines that determines T-cell polarization. In addition to generating TH2-type responses, it may also be involved in the generation of a subpopulation of regulatory T cells. Indeed, elevated levels of this cytokine have been described in chronic HCV infection. The promoter region of the IL10 gene contains several polymorphisms, which have been linked to the outcome of anti-HCV therapy. In particular, the IL10 -592A or -819A alleles and the corresponding -1082A + -819T + -592A haplotype are associated with the initial response to interferon therapy.[82] This region can be further defined as an extended haplotype, the -1082A + -819T + -592A haplotype. The IL10 108bp10.R + -3575T + -2763C + -1082A + -819T + -592A extended haplotype has been correlated with sustained

response to interferon + ribavirin therapy.[69] Homozygosity for this extended haplotype was correlated with a sustained response as was the IL10 -1082G/G genotype, and the IL10 -1082G + -819C + -592C haplotype,[83] although this has not been a universal finding.[84]

Associations are much weaker for spontaneous resolution of HCV. Knapp and colleagues found correlations between homozygosity for -592A and self-limiting infection and for the -1082G genotype with persistent infection,[83] which was also reported by Vidigal.[84] Interestingly, the -1082G allele may produce higher levels of IL-10, leading to skewing of the immune response to HCV toward a TH2-type response, and away from the generation of protective CTLs.[82,85] Conversely, high levels of IL-10 may be beneficial for liver fibrosis, as evidenced by a trial of recombinant IL-10 in HCV-related cirrhosis.[86] This is consistent with a genetic association between the low IL-10-producing genotypes and fast fibrosis progression.[83]

Interestingly, IL10 has also been implicated in the outcome of HBV infection, although through polymorphisms in its receptor rather than the cytokine gene itself.[87] This is more complex than first appears as the IL-10 receptor consists of two subunits: IL-10R1 and IL-10R2 (IL-10RB). The IL10-R1 subunit is specific for the IL-10 ligand, whereas the IL-10R2 participates in ligand binding and signal transduction of IL-10, IL-22, IL-26, IL-28A, IL-28B, and IL-29.[88] As IL-10 suppresses the generation of TH1 T-helper cells and secretion of proinflammatory cytokines IL-1, IL-6, TNFα, and interferon-γ, polymorphisms in the IL-10 gene may lead to a weaker immune response to HBV, thus resulting in persistent infection. TGFβ is also

associated with a TH2-type response, and in general has antiproliferative and immunosuppressive effects on T cells and NK cells, being secreted by classical regulatory T cells (Tregs). Again, promoter polymorphisms in this gene have been associated with the outcome of HCV infection. In particular, the C allele of the SNP at -509 has been associated with spontaneous resolution of HCV infection. Consistent with this, a SNP associated with lower TGFβ levels has also been associated with recovery from HCV infection.[89] Intriguingly, regulatory T cells were also been implicated by the association TNFSF18 (GITRL), which is expressed on these cells, with the outcome of HCV infection.[80]

Type I Interferons

Consistent with a central role for type I interferon signaling in the outcome of HBV infection IFN receptor IFNAR2 was also associated with the outcome of HBV infection. IFNAR2-F8S polymorphism was overrepresented in individuals with chronic HBV infection.[87] Again, this polymorphism appears to result in a gain, rather than loss, of function implying that the molecular mechanism underpinning this genetic association is once again complex rather than simple.

Stimulation of cells by type I interferons results in expression of interferon-stimulated genes (ISGs). These include proteins such as Mx1 (MxA), PKR, and OAS, which have important antiviral properties. HCV in particular induces activation of the type I IFN pathway and has several mechanisms to dampen this response.[90] Conversely, HBV as a "stealth" virus does not seem to induce a strong type I IFN response, but instead induces IL-10, which suppresses antiviral activity.[91] Consistent with these different responses to these viruses, genetic studies of HCV have defined polymorphisms associated with clearance of HCV infection. The 84-OAS-G/G genotype was overrepresented in individuals with persistent HCV infection.[92]

The PKR gene also contains polymorphisms that have been associated with a beneficial response. These include SNPs at positions -180 and -168 of the PKR gene and a short tandem trinucleotide repeat (CGG) in the 5′-UTR. The heterozygous -168-PKR-C/T genotype is associated with self-limiting infection (OR = 2.75; 95% CI: 1.45-5.24; P = .002), but not with treatment outcomes.[92] Interestingly, the effect observed was particularly strong in females (OR = 2.98; 95% CI: 1.26–7.11; P = .006). Self-limited infection was also associated with with shorter numbers of trinucleotide repeats in the 5′-UTR than those with more than nine. Conversely, two long alleles were found more frequently among individuals with a sustained response to interferon therapy (OR = 3.29; 95% CI: 1.10-10.52; P =

..017).[92] This was associated with male, but not female gender.

Polymorphisms in the MxA gene have also been studied in HCV infection.[93] In a treatment study from Japan, homozygosity for the -88MxA-G genotype was strongly associated with nonresponse to IFN therapy (P = .0009). This effect was independent of HCV viral genotype. The SNP at position -123 of the MxA gene is in strong linkage disequilibrium with the -88 SNP. Homozygosity for -123MxA-C was associated with interferon response, and in in-vitro assays the -123-MxA-A + -88-MxA-T sequence had an approximate fourfold higher activity than the -123-MxA-C + -88-MxA-G haplotype.[93] The detrimental effect of -88-MxA-G/G genotype was confirmed in a study showing

that it was associated with HCV persistence; the -88-MxA-T/T genotype was associated with self-limited infection, and its effect in interferon therapy was also confirmed.[92]

TRAF family member-associated NFKB activator (TANK) is a protein associated with type I IFN production as it can interact with several proteins that are implicated in its induction including RIG-I. Mosbruger et al's study also identified SNPs in this gene to be associated with clearance of HCV infection. Thus, overall polymorphisms related to the type I IFNs have been strongly implicated in the outcome of HCV infection, but to a much lesser extent in HBV infection.[80]

Genes Identified Through Genome-wide Approaches

Treatment Outcome and Chronicity in HCV

In 2009 a series of reports were published based on GWAS in HCV infection providing strong evidence that a genetic variation in the region of the IL-28B gene influenced the outcome of treatment with pegylated interferon and ribavirin.[94–97] Furthermore, the same variants were found to influence whether HCV infection resolves spontaneously or develops into a persistent infection.

Ge studied over 1100 patients with hepatitis C genotype 1 infection, who had participated in a large randomized treatment trial using pegylated interferon α2a or α2b with ribavirin.[94] Patients were selected based on strict adherence criteria; those who achieved a sustained viral response (defined as undetectable viremia 24 weeks after completion of therapy) were compared with nonresponders. The SNP rs12979860, located on chromosome 19 was found to be associated with sustained viral response (SVR) with an overall P-value of 10−28. Rs12979860 is located 3 kb away from the IL-28B gene, which encodes one of the interferon λ genes. The odds ratio for SVR in patients with the CC genotype (compared with the CT or TT genotypes) at this SNP varies slightly between ethnic groups: 5.6 in Hispanics, 6.1 in African Americans, and 7.3

in Caucasians.

It has been recognized for some time that patients of African origin are less likely to achieve SVR during treatment for HCV. The favorable C allele at rs12979860 is found less frequently among people of African origin and this was found to explain ~50% of the difference between treatment responses in African Americans compared with Caucasians.

In a parallel studies, Tanaka[97] (in Japanese patients), Suppiah,[96] and Rauch[95] (in Caucasian patients) conducted GWAS in relatively small numbers of cases using replication cohorts to validate their findings. Tanaka, Suppiah, and Rauch reported strong associations of SVR with a SNP, rs809917, which lies in close proximity to rs12979860. The effect of both SNPs in the IL-28B region appears to be independent of ethnicity, gender, baseline viral load, degree of liver fibrosis, and viral genotype, which are accepted determinants of treatment response.

Rauch also conducted a GWAS on spontaneous resolution of HCV infection, which naturally occurs in 25–30% of people exposed to the virus. Again, the IL-28B locus was found to be the major determinant of spontaneous resolution. [98] investigated the significance of the rs12979860 SNP in spontaneous resolution of infection and showed that the favorable genotype in treatment responses was also favorable in spontaneous resolution.

Although it is clear that the two SNPs in the IL-28B gene region are important determinants of both treatment outcome and spontaneous resolution of infection, it is not clear how they exert their effect. The two SNPs are 3kb upstream from the IL-28B gene and are not necessarily involved in regulation of gene expression. Indeed, there is already a lack of consistency in the reported experimental data concerning the impact of the polymorphisms on the expression of IL-28B. Ge found no association between the genotypes at rs12979860 and IL-28B expression in peripheral blood mononuclear cells whereas Suppiah reported higher levels of IL28 in individuals with the rs8099917 G-risk allele. More recently, Honda et al have studied gene expression of IL-28B in liver tissue and found that it did not vary according to genotype.[99] However, the Honda study did

find that the favorable rs8099917 genotype TT was associated with lower levels of interferon-stimulated gene (ISG) expression in the liver. Previous studies have shown that high levels of pretreatment ISG expression are associated with a poor response to treatment with pegylated interferon and ribavirin. It would therefore appear that the IL-28B genotype may be a marker for intrahepatic ISG expression, but there is no clear mechanistic link between the two.

Chronicity in Hepatitis B Virus

A genome-wide linkage study in affected sibling pairs with chronic HBV infection was conducted in a West African population in Gambia.[87] A region of linkage on chromosome 21 was identified. Analysis of multiple SNPs in this region revealed evidence of association with SNPs lying within a cluster of type II cytokine receptor genes. A haplotype spanning the interferon-α receptor II (IFNAR2) and interleukin-10 receptor II (IL-10RB) was found to be linked to chronic HBV infection. The SNP in the IFNAR2 gene resulted in an amino acid change (phenylalanine to serine at position 8) in the signaling region of the protein, which resulted in variation in the cell surface expression of the receptor and alteration in interferon signaling. The SNP in IL-10RB resulted in an amino acid change from lysine to aspartic acid at position 47. The allele associated

with viral clearance was found to be a more effective transducer of the IL-10 signal, which is clearly counterintuitive given that IL-10 is an immunoregulatory cytokine. However, IL-10RB is also a component of the interferon λ receptor when paired with the IL-28R. Subsequent investigation confirmed that the favorable IL-10RB allele was a more effective signal transducer for interferon λ.

A genome-wide association study conducted in 786 Japanese patients with chronic HBV infection and 2201 controls revealed strong association with 11 SNPs in the HLA-DPA and HLA-DPB loci in the MHC class II region.[100] Replication studies in Japanese and Taiwanese patients confirmed the association of the most promising two SNPs with combined P-values of 6.34 × 10−39 and 2.31 × 10−38. Further analyses showed that the risk haplotypes were HLA-DPA1*0202-DPB1*0501 (OR = 1.45) and HLA-DPA1*0202-DPB1*0301 (OR = 2.31) and protective haplotypes were HLA-DPA1*0103-DPB1*0402 (OR = 0.52) and HLA-DPA1*0103-DPB1*0401 (OR = 0.57). These HLA-DP associations have recently been confirmed in a Chinese population.

It has often been argued that the major histocompatability complex (MHC) / HLA disease associations may arise through linkage disequilibrium with other loci within the MHC region. However, analysis of the linkage disequilibrium patterns across the MHC class II region reveal that the HLA-DP SNPs associated with HBV chronicity are within independent linkage blocks and that associations with SNPs in adjacent linkage blocks do not reach statistical significance. These findings effectively exclude the possibility of causal variants lying outside the HLA-DP genes.

The mechanism underlying this association has now been explored using gene expression analysis. The expression of the HLA-DPA and DPB mRNA is significantly higher in the genotypes associated with spontaneous resolution of infection. This suggests that the genetic variants in the HLA-DP genes determine the level of HLA-DP protein expression and therefore the magnitude of antigen presentation. In turn, higher levels of antigenic peptide presentation may induce more potent CD4 + T helper cell responses as has been observed in patients with spontaneous resolution of infection compared with patients with chronic HBV.

Progression to Liver Cancer in HBV

To date only a single GWAS has been published in liver cancer.[101] A Chinese study investigated over 400,000 SNPs in 355 chronic HBV carriers with liver cancer and 360 controls with chronic HBV. A region on chromosome 1 was highly associated with liver cancer and replication in 1962 cases and 1430 controls confirmed an association with rs17401966 (OR = 0.61, P = 1.7 × 10−18) in the KIF1B gene. This region of chromosome 1 has previously been of interest in liver cancer as loss of heterozygosity has been reported in tumor tissue. KIF1B encodes a kinesin gene involved in trafficking between cellular organelles.

Clinical Utility of Genetic Determinants

Biomarkers for Treatment Response

In viral hepatitis, we are familiar with the use of biomarkers in the clinical management of chronic viral hepatitis. The best characterized is the use of HCV genotype, which is used to determine the duration of therapy and the dose of ribavirin, as well as to advise the patient on the likelihood of achieving an SVR. To date no host genetic marker has been used in clinical practice to guide HBV or HCV management.

Genetic markers for advanced liver disease were published by Huang et al[102] in 2007. In this study, a set of genetic markers spread over seven genes were used to identify patients with advanced fibrosis among a cohort infected with HCV. In a subsequent study, this seven-ene signature was combined with demographic variables to form a cirrhosis risk score (CRS). Patients with early liver fibrosis and a higher CRS were found to have higher rates of fibrosis progression than patients with low CRS scores. However, the predictive value of CRS is relatively poor and the effect of a high CRS was found in males but not in females. Consequently, these genetic markers have not progressed to mainstream clinical utility.

There is clearly some promise in the strong association between IL-28B genotype and the outcome of HCV treatment with pegylated interferon and ribavirin. It could certainly be envisaged that IL-28B genotype could be used alongside accepted variables such as viral genotype, viral load, degree of fibrosis, age, body mass index, and gender to provide a more accurate pretreatment assessment of potential SVR rate. However, there is still insufficient data available to determine whether IL-28B genotype will influence treatment duration or dose of either pegylated interferon or ribavirin.

Therapeutic Targets

Although understanding disease pathogenesis and generating novel biomarkers are worthy outcomes from the investigation of genetic susceptibility in viral hepatitis, the ultimate goal is to identify novel therapeutic targets. To date, there have been no new treatments for viral hepatitis based on host targets identified through the study of genetic susceptibility. However, there are some promising leads in development.

The identification of an association between thrombophilic variants in coagulation system proteins and the rate of progression of liver fibrosis provides a potential therapeutic option in patients where the primary etiology of the disease cannot be removed.[104] In mice, liver fibrosis induced by carbon tetrachloride is exacerbated by the Factor V Leiden mutation, which confirms the biologic relevance of the association of this mutation with rapid disease progression in HCV and HBV infection.[105] However, treatment of mice with warfarin anticoagulant during CCl4 therapy ameliorated the fibrosis. Warfarin is commonly used in patients with thromboembolic disease and atrial fibrillation and has an acceptable side-effect profile. The therapeutic

potential of warfarin as an antifibrotic in patients transplanted for HCV infection is now being evaluated in a multicenter randomized trial.

Variants associated with reduced function in the OAS gene predispose the patient to persistent HCV infection; in rodent models, more severe mutations confer susceptibility to lethal flavivirus infection.[106] OAS is activated by viral RNA and induced by interferon. The enzyme produces 2′5′ oligoadenylates (25OAs), which activate the latent RNAse known as RNAseL. RNAseL hydrolyses viral messenger and ribosomal RNA. 25OAs are rapidly degraded by phosphodiesterase 12. Amplification of RNAseL activity using 25OA analogues or inhibitors of the phosphodiesterase 12 is currently being explored as a therapeutic option in chronic HCV infection.

The importance of the λ-interferon system is clearly highlighted by the impact of variants in both the ligand and the receptor in the outcome of HCV and HBV infection. The variant in IL-10RB associated with elimination of HBV appears to a better interferon λ transducer, suggesting that more interferon λ signaling might help to eliminate the virus in patients with chronic infection. Surprisingly, the IL-28B genotype does not appear to be a determinant of the IL-28 gene expression, but the IL-28B genotypes associated with treatment failure are strongly associated with high levels of ISG expression.[99] Trials of interferon-λ — based treatment are already underway and early studies suggest that this form of interferon may be associated with fewer side effects.

Conclusions

It is clear that polymorphisms in several genes contribute to the outcomes of HBV and HCV infection. Many of the findings arising from the study of genetic variability in this area have contributed usefully to our understanding of the biology of disease. However, we have yet to see any therapeutic benefits directly attributable to this field of research. Perhaps the most promising application is the potential use of the IL-28B genotype as a biomarker for treatment response in patients with chronic HCV infection.

http://www.medscape.com/viewarticle/744087?src=rsshttp://www.medscape.com/viewarticle/744087?src=rss