Interferon-λ3 Gene Polymorphic Variants, rs4803217 and rs12980275, Responsiveness to HBV Vaccine and Outcome of HBV and HCV Exposure in Hemodialyzed Patients

1. Background

Exploration of factors contributing to the outcomes of hepatotropic viral infections and affecting prophylactic vaccination results are essential for the general population, especially subjects with altered immune competence. The latter group includes patients with end-stage renal disease (ESRD), mainly those receiving hemodialysis (HD).

Polymorphic variants of interferon (IFN)-λ genetic region (IFNL) are continuously explored for their associations with hepatitis C virus (HCV) clearance or hepatitis B virus (HBV) limitation. The IFNL4 polymorphisms, rs12979860 and rs8099917, formerly IFNL3 rs12979860 (1) and IFNL3 rs8099917 (2), became known as the predictors of spontaneous (3) or INF-based regimen-induced HCV clearance (4-6). Less consistent are the results of the association between IFNL variants and HBV limitation (7, 8). Variants rs8099917 and rs12979860 show no association with the generation of antibodies to HBV surface antigen (anti-HBs) after HBV vaccination or exposition (9). In non-uremic populations, rs4803217 in the IFNL3 messenger RNA 3’ untranslated region (UTR) or rs12980275 downstream of IFNL3 was associated with spontaneous or therapeutic clearance of HCV (10-12) and HBV (13, 14), and rs12980275 was correlated with plasma IFN-λ3 levels (15). Moreover, rs12980275 was connected with the sustained virological response following antiviral therapy of chronic hepatitis C in HD patients (16). Both IFNL3 and IFNL4 are located on chromosome 19q13.2 at a distance of 1.3 kb (Appendix 11).

Besides, IFN-λ3, formerly interleukin-28B, is the protein product attributed to IFNL3. In a group containing chronic HBV-infected patients, self-limited HBV infection individuals, and healthy subjects, circulating IFN-λ3 levels were positively associated with major homozygosity in rs8099917, rs12979860, and rs12980275, located within IFNL on chromosome 19 (15). In HD patients treated with low permeable dialysis membranes, individuals showing rs8099917 major homozygosity (TT) presented higher IFN-λ3 levels than those harboring the G allele. Such a correlation was not revealed for rs12979860 (17).

In Chinese non-uremic individuals, higher levels of circulating IFN-λ3 were associated with HCV clearance (18) and self-limited HBV infection (15). However, chronic hepatitis C subjects from Japan demonstrated higher IFN-λ3 than healthy individuals before treatment (2). Our data on HD patients revealed that plasma IFN-λ3 was positively correlated with anti-HBs titers generated after vaccination or exposure to HBV (17, 19, 20). Plasma IFN-λ3 levels were positive predictors of survival in prevalent HD patients (20). The variant homozygosity (GG) of rs8099917 predicted a lower probability of survival in a prospective study of HD patients (21). The IFNL3 polymorphic variants were not analyzed concerning the survival of HD subjects.

2. Objectives

This study aimed to investigate whether IFNL3 rs4803217 and rs12980275 downstream of IFNL3 are associated with the anti-HBs generation, self-limitation/spontaneous resolution of HBV/HCV infection, plasma IFN-λ3, and survival of HD patients. Additionally, we analyzed haplotypes of IFNL3 and IFNL4 concerning HCV spontaneous clearance.

3. Methods

3.1. Patients

The HD population (n = 1337), previously described concerning IFNL4 polymorphic variants and anti-HBs generation (22), was qualified for genotyping IFNL3 polymorphic variants. To be included, all patients not exposed to HBV had to undergo a completed vaccination series as proposed by the Advisory Committee on Immunization Practices Centers for Disease Control and Prevention and have the status of "responders" or "non-responders" to HBV vaccination (23). All patients were Caucasians.

The HD group included 942 patients with no evidence of exposition to HBV (having repeatedly negative results for total antibodies against HBV core antigen - anti-HBc) or HCV (having repeatedly negative results of HCV RNA and antibodies to HCV - anti-HCV), 241 patients with serological markers of exposition to HBV, 80 patients exposed to HCV, and 74 patients exposed to both of these viruses. Subjects with serological markers of HBV infection were included in the HBV-infected group (n = 315), and those exposed to HCV were assigned to the HCV-infected group (n = 154). Individuals showing the acute phase of both infections were excluded. The HBV-exposed patients (n = 315) included HBsAg-positive subjects (n = 31) and individuals who self-limited infection (HBsAg-negative, anti-HBc positive, n = 284). Patients who self-limited HBV infection included subjects generated anti-HBs (n = 252), and patients presenting isolated anti-HBc positivity (n = 32). Three HBsAg-positive patients were also able to generate anti-HBs.

The HBV vaccinated group was composed of 1,022 individuals. Vaccines containing the recombinant HBV surface antigen (HBsAg) (Engerix B, GlaxoSmithKline Biologicals, Belgium; Hepavax-Gene, BIOMED SA, Poland; and Euvax B, LG Chemical, South Korea) were administered for immunization. The HD subjects were diagnosed as being able to generate protective amounts of anti-HBs if their circulating anti-HBs titers gained at least once a value of ≥ 10 IU/L after HBV vaccination or exposure. The highest individual anti-HBs titers shown in the course of HD were analyzed for associations with tested polymorphisms. Anti-HBs titers exceeding 1000 IU/L were referred to as indicating overreacted response (24). The entire HD group included 154 HCV-exposed patients (11.5% of 1337). Eighty subjects were in the vaccinated group and 74 in the HBV-exposed group. The HCV RNA positivity was detected in 93 patients. Enrolled patients exposed to HBV or HCV were never treated with antiviral medications designated for infections with these viruses.

The plasma IFN-λ3 concentration was determined in 188 HD subjects in fixed clinical conditions. Blood probes for IFN-λ3 and standard chemistry were taken on an empty stomach before the HD procedure was performed in the middle of the week.

3.2. Survival Studies

All patients successfully genotyped for IFNL3 single nucleotide polymorphisms (SNPs) were included in the retrospective longitudinal survival study. The tested period started at the beginning of renal replacement therapy (RRT) of each patient and ended in November 2019. Additionally, 402 patients tested for rs4803217 and 391 patients tested for rs12980275, who were prospectively observed since January 20, 2009 (22), were analyzed concerning survival in November 2019 (the period of the prospective study was 10.81 years).

3.3. Laboratory Determinations

The commercial ELISA kit specific for IFN-λ3 (Human Interleukin 28B (IL-28B) ELISA Kit, Sunred Biological Technology Co., Ltd., Shanghai, China) was applied for measuring the plasma IFN-λ3 level at 450 nm in an ELISA plate reader (Infinite F50, Tecan Group Ltd., Männedorf, Switzerland). The sensitivity of this kit was 0.65 ng/L. The intra-assay and the inter-assay coefficients of variation were below 10 and 12%, respectively. Microparticle enzyme immunoassay (MEIA) technology (ABBOTT, Germany) or chemiluminescent microparticle immunoassay (CMIA) method (ABBOTT, Ireland) was applied to estimate anti-HBs titers. Values of anti-HBs titers were presented numerically up to 1000 IU/L; values ≥ 1000 IU/L were shown as 1000 IU/L in result descriptions and statistical calculations.

3.4. Genotyping

Patients’ blood probes for genotyping were obtained from December 30, 2008, to January 15, 2018. The DNA material was frozen. The IFNL3 SNPs (rs4803217, rs12980275) were genotyped using high-resolution melting (HRM) curve analysis. Besides, IFNL4 rs8099917 and IFNL4 rs12979860 SNPs, previously tested in the examined patients, were also genotyped by HRM (9, 17, 22). Appendix 1 presents HRM conditions applied for the identification of tested SNPs. Successful genotyping was conducted in 1,327 patients for rs4803217 and 1,311 individuals for rs12980275. For quality control, about 10% of the probes were genotyped again using the same method; the concordance rate was 100%.

3.5. Statistics

We used the Shapiro-Wilk test to determine how continuous variables were distributed. The percentages for dichotomous variables and medians with ranges for continuous parameters are shown. The characteristics of HD patients were analyzed concerning IFNL3 SNPs using dominant, recessive, and additive inheritance models. The Mann-Whitney test was used to compare continuous parameters. The Fisher's exact test was applied to check differences in dichotomous data. The chi-square test was used to evaluate the agreement with Hardy-Weinberg Equilibrium (HWE). The SNP frequencies were computed using Fisher's exact test or Cochran-Armitage test (P_trend) to show tendencies in associations. Distributions of genotypes were evaluated by the chi-square test (P_genotype).

Odds ratios (ORs) and their 95% confidence intervals (CIs) were computed using Fisher’s exact test. Logistic regression was applied to demonstrate whether tested SNPs were significant predictors of spontaneous HCV clearance among demographic and clinical variables. Patient characteristics chosen for this analysis included age, RRT duration, and chronic glomerulonephritis because they were significantly different between groups with spontaneous HCV clearance and persistent HCV infection. All-cause, cardiovascular, infection-related, and neoplasm-related reasons for death were analyzed in dominant, recessive, and additive inheritance models. Additionally, survival analyses were performed separately in patients exposed to HCV or HBV. The Kaplan-Meier method with the log-rank test was applied to show dissimilarities in survival. The P values below 0.05 were accepted as presenting significance. The Bonferroni correction was applied in examinations of SNP results. All analyses mentioned above were computed using Statistica version 12 (Stat Soft, Inc., Tulsa, OK, USA).

As IFNL3 and IFNL4 show close positions on chromosome 19 (Appendix 11), pair-wise linkage disequilibrium (LD) between their polymorphisms was calculated. D’ and r² were obtained using Haploview 4.2 software (http://www.broad.mit.edu/mpg/haploview/). We calculated haplotype frequencies with the same software. The 1000-fold permutation test was used to show the significance of the data.

3.6. Compliance with Ethical Standards

The Institutional Review Board of Poznan University of Medical Sciences, Poland, accepted the concept of our investigations. All patients or their parents, as appropriate, provided written informed consent for research.

4. Results

4.1. HD Patients’ Data

Table 1 shows the characteristics of the examined patients. In the entire HD group, the distribution of IFNL3 polymorphic variants followed HWE (P = 0.9 for rs4803217 and P = 0.7 for rs12980275). In analyses performed in inheritance models, patients’ data did not show significant differences regarding IFNL3 rs4803217 and rs12980275 variants if the Bonferroni correction was applied (Supplementary Tables 2 and 3, respectively). However, uncorrected data of all examined HD patients indicated that bearers of the rs4803217 variant allele (T) showed a higher susceptibility for HCV RNA positivity (OR: 2.024, 95% CI: 1.267 - 3.234, P = 0.003) (Appendix 2).

4.2. IFNL3 Polymorphisms and Anti-HBs

In the HBV-vaccinated patients and HBV-exposed group, we tested the capacity to produce anti-HBs, titers of anti-HBs, the prevalence of anti-HBs titers ≥ 1000 IU/L, and anti-HBs titers when patients who did not generate protective levels of anti-HBs were excluded. Any differences in the frequency of ability to produce protective anti-HBV titers and the quantity of immunization after HBV infection were presented concerning IFNL3 rs4803217 (Appendix 5) and IFNL3 rs12980275 (Appendix 7) polymorphic variants. However, in HBV-exposed subjects, higher anti-HBs titers were shown in individuals bearing two G alleles of IFNL3 rs4803217 (P = 0.02 without Bonferroni correction, Appendix 6).

4.3. IFNL3 Polymorphic Variants and Self-Limitation of HBV Infection

The HD patients who remained HBsAg positive after HBV exposition did not differ from HBsAg negative patients in the distribution of IFNL3 polymorphic variants (Appendix 6). We also analyzed differences in the frequency of IFNL3 variant distribution between HbsAg-positive individuals not generating anti-HBs and HBsAg negative subjects generating anti-HBs (Appendix 8) and patients with isolated anti-HBc positivity (Appendix 9). All of these analyses yielded no significant results.

4.4. IFNL3 Polymorphisms and Spontaneous HCV Resolution

In HCV-exposed HD patients, spontaneous HCV elimination was correlated with tested IFNL3 polymorphic variants (Table 2). Individuals homozygous for the IFNL3 rs4803217 major allele (G) showed over three-fold higher probability for spontaneous HCV clearance. There was also a trend for association between the homozygosity of the rs12980275 major allele (A) and spontaneous HCV elimination.

As expected, HCV RNA-positive patients showed higher serum activities of liver enzymes than HCV RNA-negative individuals (Table 3). Additionally, the group with HCV RNA positivity was younger but longer on RRT and revealed chronic glomerulonephritis as a more frequent cause of ESRD than HCV RNA negative individuals. Therefore, we analyzed whether the GG homozygosity of rs4803217 was associated with spontaneous HCV clearance if age, RRT duration, and chronic glomerulonephritis were simultaneously used as possible prognostic factors for spontaneous resolution of HCV infection. A similar multivariate analysis was performed for the AA homozygosity of rs12980275. The rs4803217 GG homozygosity (OR: 2.91, 95% CI: 1.42 - 5.97, P = 0.003) and rs12980275 AA homozygosity (OR: 2.92, 95% CI: 1.42 - 6.00, P = 0.003) were positive predictors of spontaneous HCV clearance, whereas longer RRT duration was a negative predictor (OR: 0.94, 95% CI: 0.88 - 0.99, P = 0.03 together with rs4803217 and OR: 0.94, 95% CI: 0.88 – 1.00, P = 0.038 in the model with rs12980275).

4.5. Linkage Disequilibrium

We examined LD between rs12980275, rs4803217, rs12979860, and rs8099917 variants. The strongest LD was between rs4803217 and rs12979860 (r² = 0.81), and the weakest was shown between rs4803217 and rs8099917 (r² = 0.34) (Appendix 11).

4.6. IFNL3 and IFNL4 Haplotype Analysis

We analyzed haplotypes of IFNL3 and IFNL4 concerning the outcome of HCV infection (Table 4). Spontaneous HCV clearance was connected with haplotypes comprising two or more frequent alleles but without variant alleles of the examined IFNL SNPs. The most substantial relationship with spontaneous HCV clearance was demonstrated for haplotype rs12979860-C_rs8099917-T.

4.7. Plasma IFN-λ3

Circulating IFN-λ3 levels did not differ regarding tested IFNL3 SNPs. An analysis in subgroups separated by HD modality (low-flux HD, high-flux HD, and hemodiafiltration) also did not reveal significant differences in plasma IFN-λ3 concentrations (Appendix 10). The exclusion of HCV RNA positive and HBsAg positive subjects from the analysis did not show any association between plasma IFN-λ3 and tested polymorphisms.

The HCV RNA positive (IFN-λ3 74.0, 4.9 - 240 ng/L) and HBsAg positive (IFN-λ3 17.0, 9.0 - 57.7 ng/L) patients showed significantly lower plasma IFN-λ3 concentrations than patients who composed a group of subjects not exposed to HBV/HCV infections, those who had self-limited HBV infection, and those who spontaneously resolved HCV infection (IFN-λ3 92.0, 0.24 - 240 ng/L). The latter three subgroups did not differ in plasma IFN-λ3 concentrations (Kruskal-Wallis test P-value = 0.2). The HBsAg positive patients (including five HBV DNA positive patients and one subject not currently tested for HBV DNA) had lower IFN-λ3 levels than HCV RNA positive subjects (Figure 1).

Plasma IFN-λ3 concentrations in HD patients concerning HBV/HCV status [Kruskal-Wallis test P-value = 0.00005; groups: 1, HBV/HCV-not exposed subjects and those who self-limited/spontaneously resolved these infections (n = 147); 2, HCV RNA-positive subjects (n = 35); 3, HbsAg-positive patients (n = 6)].

Figure 1.

Plasma IFN-λ3 concentrations in HD patients concerning HBV/HCV status [Kruskal-Wallis test P-value = 0.00005; groups: 1, HBV/HCV-not exposed subjects and those who self-limited/spontaneously resolved these infections (n = 147); 2, HCV RNA-positive subjects (n = 35); 3, HbsAg-positive patients (n = 6)].

There were positive correlations between circulating IFN-λ3 concentrations and anti-HBs titers in all HD patients (r = 0.537, P = 4.15E-16), as well as HBV-vaccinated (r = 0.539, P = 3.22E-11) and HBV-infected (r = 0.553, P = 8.02E-06) HD subjects.

4.8. Survival and IFNL3 Polymorphisms

In the cross-sectional study, patients showing the IFNL3 rs4803217 GG genotype had higher infection-related mortality (log-rank P = 0.021), but it was not observed in either the prospective study or Cox regression analysis, including age and gender. There were no other associations between tested IFNL3 variants and survival, irrespective of study design, examined groups, or causes of death. Survival curves are shown in online resource 1.

5. Discussion

In the examined HD patients, the capacity to generate anti-HBs titers at the protective level in response to HBV inoculation or infection did not correlate significantly with IFNL3 rs4803217 and rs12980275 SNPs. The highest anti-HBs titers or prevalence of overreacted responses appeared not to depend on the examined variants after Bonferroni correction. Neither IFNL3 variants in the current study nor IFNL4 variants in the already published research (9) were associated with anti-HBs generation.

In this study, IFN-λ3, a protein product of IFNL3, was not significantly related to tested IFNL3 SNPs. Besides, IFN-λ3 showed higher circulating levels in HD patients than in healthy subjects by the same method, possibly due to the upregulation of production and/or lower degradation rate of IFN-λ3 in the uremic milieu (14). Human blood dendritic cell antigen 3 (BDCA3) (+) dendritic cells are robust IFN-λ3 generators in response to HCV stimulation (25). However, in chronic HBV or HCV infections, functional alterations of dendritic cell subsets were noticed (25, 26), so IFN-λ3 production could also be impaired. However, the determination of circulating IFN-λ3 in non-uremic subjects demonstrated not consistent results concerning differences in IFN-λ3 levels in patients resolving HCV infection or showing persistent HCV RNA positivity (2, 15, 18). Our data on HD patients showed that IFN-λ3 levels were lower in persistent HBV/HCV infections than in HD subjects not exposed to HBV/HCV or those who self-limited/spontaneously resolved these infections, which is in agreement with earlier data in non-uremic subjects (15, 18).

Circulating IFN-λ3 levels positively and strongly correlated with anti-HBs titers. The HBsAg positive patients, who did not generate anti-HBs, showed low IFN-λ3 levels. This finding confirms an earlier report by Li et al. (15). A lack of expected associations between IFNL3 polymorphic variants and IFN-λ3 and anti-HBs possibly depends on intrinsic and extrinsic influences on transcription and translation processes. Determination of the IFNL3 transcript in future studies could help approach this problem.

An association of IFNL3 rs4803217 with spontaneous or therapeutic HCV clearance was shown in non-uremic patients (10, 11). Our study revealed the positive impact of major homozygosity (GG) of IFNL3 rs4803217 on spontaneous HCV clearance in HD patients. In contrast, bearers of the rs4803217 variant allele (T) showed about two-fold higher susceptibility for persistent HCV RNA positivity. Significance was obtained at a P-value of 0.001, which was significant also after Bonferroni correction. As rs4803217 is in the strong LD with rs12979860 (0.81 in this study), and IFNL3 rs12979860 was already associated with HCV clearance in GWAS analyses (4), the level of significance shown in our research in a specific smaller group of patients should be sufficient for the recognition of a true association between IFNL3 rs4803217 and spontaneous HCV clearance. Logistic regression analysis, which included rs4803217, rs12980275, and clinical variables as possible predictors of spontaneous HCV clearance in HD patients, revealed similar predictivity of both SNPs concerning HCV resolution. Longer RRT duration diminishes the chance for spontaneous HCV clearance.

McFarland et al. (27) found that the rs4803217 T allele caused twice lower stability of the mRNA transcript acting through the AU-rich element (ARE)-mediated decomposition of IFNL3 mRNA. AU-rich elements are connected with post-transcriptional gene regulation (28). The rs4803217 G allele mRNA sets up a well-defined 3'UTR form while the T allele mRNA is more active, which controls the effectiveness of mRNA translation. Non-translating mRNA levels were higher in the case of the rs4803217 T allele mRNA (29). Therefore, the unfavorable (T) allele bearers could show worse controlling of HCV infection. In our study, HD patients with persistent HCV infection revealed a higher frequency of the unfavorable rs4803217 allele and lower levels of plasma IFN-λ3 than subjects free from HCV RNA. However, although IFN-λ3 generation is putatively lower in the rs4803217 T allele possessors due to lower transcript stability (27, 29), our study did not demonstrate the association of rs4803217 with circulating IFN-λ3, even if it was analyzed in subgroups categorized by HD modality. High permeable HD membranes may influence plasma IFN-λ3 levels (17), causing a dialysate loss of IFN-λ3, which has a molecular weight of about 22 kDa (30) and can cross HD membranes to the dialysate.

Data reported by McFarland et al. (27) and Lu et al. (29) suggest that a functional SNP, rs4803217, has a causative role in HCV resolution. Simultaneously, other SNPs of the IFNL region may correlate with HCV resolution only because of LD with SNP rs4803217. The LD between rs4803217 and rs12979860 variants is very strong (r² = 0.92) (31). In Polish HD patients, the LD between these variants seems to be weaker (r² = 0.81). In non-uremic patients with chronic hepatitis C living in the same region as the examined HD patients, the LD between rs4803217 and rs12979860 SNPs was even lower (r² = 0.75) (11). The rs4803217 and rs8099917 SNPs show weak LD [r² = 0.34 (this study), r² = 0.36 (11)]. The LD between two brands depends on intrinsic cellular components and extrinsic aspects of the human past (32). On the other hand, the association of haplotype rs12979860-C_rs8099917-T with spontaneous HCV clearance, the strongest one among other significant haplotypes concerning HCV resolution, suggests that the major rs4803217 allele may not be necessary for spontaneous HCV clearance. Moreover, the haplotype rs368234815-∆G_ rs4803217-G, which includes the favorable rs4803217 G allele and unfavorable rs368234815 ∆G allele discovered by Prokunina-Olsson et al. (33), was associated with a poorer response to anti-HCV treatment compared to haplotype rs368234815-∆G_ rs4803217-T, which is composed of the unfavorable rs4803217 T allele (31). These data suggest that the association between SNP rs4803217 and HCV clearance may occur due to its strong LD (r² = 0.90) with SNP rs368234815 in subjects of European ancestry (31). Therefore, although the major homozygosity (GG) of rs4803217 predicted spontaneous HCV clearance in HD patients in our study, the strength of this predictability may be weakened in possessors of IFNL4 rs368234815 ∆G.

The associations of tested IFNL3 SNPs with survival were not convincing in the examined HD patients. Neither all subjects nor subgroups categorized by HBV/HCV infection showed an influence of IFNL3 SNPs on the death rate, possibly due to a lack of association with plasma IFN-λ3 levels (20, 21).

5.1. Limitations

A limitation of our study was the use of recombinant hepatitis B vaccines provided by three manufacturers. Engerix B and Euvax B contained HBsAg variant adw2, while Hepavax-Gene was composed of HBsAg variant adr. We analyzed vaccinated HD patients as one group because Euvax B and Hepavax-Gene vaccines were shown as clinically identical (equally effective) to the Engerix B reference vaccine (34, 35). Moreover, as all the three vaccines were recommended uniformly powerful, some HD patients during their long-term RRT were receiving booster vaccine doses different from those used for the primary immunization. Therefore, differences in anti-HBs generation resulting from the use of the above-mentioned vaccines should be minimal or do not exist.

Similarly, anti-HBs titers were measured using two methods: MEIA and CMIA. By the manufacturer (Abbott Ireland), quantitative anti-HBs values obtained by such assays may not be equivalent due to the heterogenicity of anti-HBs. Each HD facility used one method of anti-HBs determination (MEIA or CMIA). All patients dialyzed in the same center, independently of their anti-HBs results, had performed the anti-HBs titration with this one method. How discrepancies in anti-HBs measurements contributed to a lack of significant differences between the studied groups (Tables 1 and 3) remains unclear.

5.2. Conclusion

In HD patients, natural HCV clearance is connected with the IFNL3 rs4803217 GG genotype and haplotypes comprising at least two more frequent alleles but without any variant allele of tested IFNL3/IFNL4 SNPs. Therefore, genotyping IFNL3 rs4803217 may be used for practical purposes in the prognosis of natural HCV clearance. Nevertheless, it does not forecast the self-limitation of HBV exposure, responsiveness to HBV vaccine, or patient mortality. Circulating IFN-λ3 levels are not straightly influenced by IFNL3 rs4803217 and rs12980275, but show differences dependently on HBV/HCV seromarkers, as well as firmly correlate with anti-HBs titers. Both tested polymorphisms are not associated with the survival of HD patients.