A functional polymorphism in the IL-10 promoter influences the response after vaccination with HBsAg and hepatitis A

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

The immune response to hepatitis B surface antigen (HBsAg) is mostly genetically determined. Interleukin 10 (IL-10) is a central immunoregulatory cytokine with important effects on B-cells. We have studied the influence of IL-10 promoter polymorphisms on the immune response to HBsAg and hepatitis A vaccination. We vaccinated 202 twin pairs in an open prospective study with a combined recombinant HBsAg/inactivated hepatitis A vaccine. IL-10 promoter polymorphisms were investigated in all individuals and their influence on anti-HBs, and anti-HAV responsiveness was studied. In the multiple regression analysis accounting for smoking, gender, body mass index and age, the ACC haplotype (−1082, −819 and −592) had a strong influence on anti-HBs production. Individuals carrying the ACC haplotype had anti-HBs titres almost twice as high as individuals without this haplotype. In contrast, anti-HAV production was suppressed by the presence of the −1082A allele in comparison with individuals homozygous for the −1082G allele. The contribution of the shared IL-10 promoter haplotype accounted for 27% of the genetic influence on anti-HBs antibody response. In conclusion, genetic variability in the IL-10 promoter is an important modulator of the immune response against hepatitis viral antigens. (HEPATOLOGY 2005.)

Hepatitis B virus (HBV) vaccination is highly effective in the primary prevention of acute and chronic HBV. Nationwide vaccination programs have succeeded in reducing HBV related morbidity and mortality considerably. Hepatitis B vaccine responsiveness shows large interindividual variability. Results of a twin study have shown a strong genetic determination of the HBsAg specific immune response with 60% of the observed variability explained by genetic factors. Although there is a strong contribution of genes encoded within the major histocompatibility complex (MHC) most of the genetic effect originates from genes outside the MHC.1 Possible candidates for immunoregulatory functions are cytokine genes and proteins involved in intracellular signaling pathways in lymphocytes.

Interleukin (IL-10) is an important immunoregulatory cytokine that is produced by monocytes and lymphocytes.2 It inhibits formation of proinflammatory cytokines like TNF-α in T cells and monocytes3 and downregulates MHC class II expression in monocytes.4 Contrary to its inhibitory function in T cells and macrophages, in B cells IL-10 stimulates the production of immunoglobulins and the expression of MHC class II antigens.5 We and others have recently shown that approximately 50% of the observed interindividual variability in IL-10 production can be explained by genetic factors.6, 7 Some of the observed variability in IL-10 production is determined by polymorphisms in the 5′-flanking region of the IL-10 gene on chromosome 1 (1q31). The proximal promoter contains three biallelic single nucleotide polymorphisms (SNPs) at positions −1082 (G/A), −819 (C/T) and −592 (C/A) bp from the transcription start site. These SNPs are closely linked and only 3 of 4 possible haplotypes have been reported: GCC, ACC, and ATA. In particular a G to A transition at position −1082 within an ETS consensus binding motif enhances binding of the PU.1 transcription factor and leads to a decreased transcriptional activity of −1082A haplotypes.7 In addition, 2 microsatellite polymorphisms lying 1.1 kb and 4.0 kb upstream of the transcription start as well as a number of SNPs have been correlated with high and low IL-10-production after in vitro stimulation6, 8 and have been associated with several diseases (reviewed by Bidwell et al.9).

In a prospective twin vaccination study we have investigated the contribution of IL-10 promoter variability to the anti-HBs (antibody to hepatitis B surface antigen) and anti-hepatitis A virus (HAV) immune response after a combined HAV and HBsAg vaccination. Twin studies are a valuable model to analyze genetic and environmental components of a certain phenotype. Monozygotic (MZ) twins are genetically identical and dizygotic (DZ) twins share on average 50% of their genes. Under the assumption of a similar environment in MZ and DZ twins, additional phenotypic differences in DZ twins are assumed to be due to genetic variability between them.

Abbreviations

HBV, hepatitis B virus; MHC, major histocompatibility complex; IL, interleukin; SNP, single nucleotide polymorphisms; anti-HBs, antibody to hepatitis B surface antigen; HAV, hepatitis A virus; HBsAg, hepatitis B surface antigen; MZ, monozygotic; DZ, dizygotic; BMI, body mass index; GMT, geometric mean titer.

Patients and Methods

We performed a prospective population-based twin study to assess the heritability of the immune response to vaccination with HBsAg and HAV.1 Subjects at least 18 years old and younger than 65 years were eligible for participation. The study was performed in the Wiesbaden/Mainz area in Germany. All subjects included were Caucasians. Demographic details are given in Table 1. Criteria for exclusion were previous or active HBV infection, hepatitis C virus infection, HIV infection, chronic liver or kidney disease, autoimmune disease, cancer, pregnancy or lactation, and allergy to vaccine components. In addition, individuals with a history of alcohol intake (>40g/d), intravenous substance abuse or on regular immunosuppressant medication (corticosteroids, azathioprin, cyclosporin) were not included.

Table 1. General Characteristics of Mono- and Dizygotic Twins and Anti-HAV and Anti-HBs Titers After the Third Vaccination
 Monozygotic n = 192Dizygotic n = 190
  • *

    Mean (standard deviation)

Mean age (years)*32.8 (±11.6)*36.6 (±13.8)
Female/male96/9671/119
BMI (kg/m2)*24.1 (±4.1)*24.3 (±4.2)
Current smokers131120

A total of 232 twin pairs, aged 18 to 65 years, were considered for enrolment in this study (ascertained from the general population by a media campaign and direct contacting). Thirty twin pairs were excluded because at least one co-twin was positive for anti-HBs or HBsAg at the initial testing. Two hundred and two twin pairs were included in the study of which 96 DZ and 95 MZ twin pairs received the full vaccination course. In nine twin pairs at least one twin was lost to follow-up. In one pair, vaccination was stopped because of pregnancy, and in a second pair, because of a rash after the second vaccination. Nine individuals were HBsAg nonresponders (anti-HBs <10 U/L) after the third vaccination (nonresponse rate 2.4%). The study was conducted with written informed consent of all probands involved and was performed according to the Declaration of Helsinki and ICH Guidelines of Good Clinical Practice. Local ethics committee approval was obtained.

Vaccination Protocol.

Three doses of a combined HAV/HBsAg vaccine (Twinrix, GlaxoSmithKline, Rixensart, Belgium) were injected intramuscularly in the deltoid region at 0.1 and 6 months. The vaccine consisted in a 20 μg yeast derived recombinant nonglycosylated major surface protein of HBV and ≥720 ELISA units of inactivated HAV particles adsorbed onto aluminium salts. Anti-HBs and anti-HAV were tested 4 weeks after the last vaccination to assess vaccine response. Antibody testing was done with the AUSAB Axsym and HAVAB2.0 tests (Abbott, Wiesbaden, Germany).

Zygosity was assessed by microsatellite typing using the GenePrint Powerplex 16 System (Promega, Madison, WI). With this test 15 highly polymorphic short tandem repeat loci (STR) as well as the sex specific amelogenin locus were PCR-amplified and the fluorescence labeled fragments were separated by size and color using ABI Prism 310 capillary electrophoresis equipment (Applied Biosystems, Weiterstadt, Germany). Same-sex twins identical at 15 STR-loci were considered to be MZ. Information on medication use and demographic variables was obtained by standardized questionnaire.

DNA Extraction.

DNA was isolated from previously frozen EDTA anti-coagulated blood samples by digestion with proteinase K in a non-ionic lysing buffer (50 mmol/L KCl, 10 mmol/L Tris-HCl pH 8.3, 2.5 mmol/L MgCl2, 0.45% Tween 20, 0.45 % NP-40). After heat-inactivation of proteinase K an aliquot of the DNA-solution was directly applied to PCR.

SNP Detection.

The SNPs at position −1082 and −592 were detected by specific PCR amplification and subsequent restriction enzyme digestion as described elsewhere.10, 11

Statistical Analysis.

We log transformed the anti-HBs and anti-HAV titers for the statistical analysis to obtain approximately normally distributed measures. Linear models were fit to assess the impact of body mass index (BMI), age, gender, smoking, alcohol and haplotypes/genotypes on anti-HBs and anti-HAV. To adequately adjust for the dependence of the two measurements in one pair, generalized estimation equations (proc genmod in SAS) with unstructured correlation matrices for the different types of twin pairs were used.12 Models were compared by likelihood ratio statistic.

To assess the amount of additive genetic variance which might be due to the IL-10 promoter region, we compared twice the difference of the intraclass correlation of the MZ twins and the intraclass correlation of the haplotype identical DZ twins with twice the difference of the intraclass correlation coefficient of the haplotype identical DZ twins and the intraclass correlation of all DZ twins in the 88 MZ and the 48 same-sex DZ twins who were successfully genotyped for the IL-10 promoter haplotype.13

Results

Patient characteristics have been described elsewhere.1 Distribution of IL-10 promoter haplotypes, SNP genotypes and the corresponding geometric mean titres are shown in Tables 2 and 3. We had previously reported a strong influence of gender on anti-HAV and anti-HBs responses. Both were significantly higher in females. In addition, older individuals and probands with higher BMIs showed weaker responses for anti-HBs. Smoking had a negative effect on the anti-HAV response. The influence of IL-10 SNP and promoter haplotypes was studied by using generalized estimation equations with the inclusion of these variables. Individuals carrying the ACC haplotype (geometric mean titer [GMT] 6474 U/L, 95% CI 4907–8542 U/L) had anti-HBs titres almost twice as high as individuals without this haplotype (GMT 2627 U/L, 95% CI 1919–3595 U/L, P < .003) (Table 2). In contrast, anti-HAV production was suppressed by the presence of the −1082A allele (GMT 6050 U/L, 95% CI 4738–7727 U/L) in comparison with individuals homozygous for the −1082G allele (GMT 11125 U/L, 95% CI 8160–15166 U/L, P < .012) (Table 3). No significant influence of any IL-10 promoter haplotype (ACC, ATA, or GCC) on anti-HAV response was detected (Table 2).

Table 2. Influence of IL-10 Promoter-Haplotype on Anti-HBs and Anti-HAV Titers
 nAnti-HBs (GMT, 95% CI)nAnti-HAV (GMT, 95% CI)
  1. NOTE. The proximal promoter contains three biallelic single nucleotide polymorphisms (SNP) at positions −1082 (G/A), −819 (C/T) and −592 (C/A) bp from the transcription start site. These SNPs are closely linked and only 3 of 4 possible haplotypes have been reported: GCC, ACC, and ATA. Numbers for anti-HAV are lower because individuals positive for anti-HAV before vaccination were excluded.

ACC/ACC406163.4 (3909.7–9716.4)335958.7 (3597.2–9870.2)
GCC/GCC672920.2 (1796.8–4746.2)5510650.9 (7522.9–15079.7)
ATA/ATA185483.4 (2150.8–13979.6)178498.2 (4529.8–15943.5)
ACC/GCC768362.0 (5603.2–12479.0)606677.5 (4755.4—9376.4)
ACC/ATA644927.1 (2814.2–8626.3)547257.7 (4895.9–10758.9)
GCC/ATA892089.8 (1315.9–3318.8)755680.0 (3920.0–8230.2)
ACC present1806474.28 (4906.9–8542.3)1476711.2 (5359.9–8403.2)
no ACC present1742626.6 (1919.0–3595.2)1477529.0 (5918.3–9578.2)
Table 3. Influence of IL-10 −1082 Promoter SNP on Anti-HBs and Anti-HAV Titers
−1082nanti-HBs (GMT, 95% CI)nanti-HAV (GMT, 95% CI)
  1. Abbreviation: GMT: geometric mean titer.

AA1334241.3 (2946.1–6105.9)1136866.1 (5282.0–8925.2)
AG1733730.65 (2724.18–5109.0)1416050.4 (4737.7–7726.6)
GG753229.7 (2074.3–5028.6)6211124.6 (8160.3–15165.8)

When two models for the anti-HBs response were compared, the first including only sex, age and BMI, the second including sex, age, BMI, and the IL-10 ACC haplotype, the second model showed a significantly better fit (likelihood ratio 23.7, 1 degree of freedom [DF], P < .0001 [Table 4A]).

Table 4. A. Factors With a Significant Contribution to Anti-HBs (log) Immune Response
Predictive VariableHBV—Full Model*95% CIP
  • NOTE. Table 4A: In the final model only the significantly associated IL-10 ACC promoter haplotype was considered. Table 4B. In the final model only the significantly −1082 single nucleotide polymorphism was considered.

  • *

    Estimated beta coefficient for the predictive variable.

  • The monzygote ACC and the heterozygote ACC/nonACC were computed to nonACC individuals.

  • ††

    GG compared to AG and AA combined.

Male sex−0.91−1.31–−0.50<.0001
BMI−0.0980−0.15–−0.05.0002
Age−0.04−0.06–−0.02<.0001
ACC/ACC0.970.48–1.45<.0001
ACC/non ACC††0.730.29–1.18.001
B. Factors With a Significant Contribution to Anti-HAV (log) Immune Response
Male sex−0.48−0.82–−0.15.005
Smoking−0.40−0.74–−0.05.024
Age−0.03−0.05–−0.003.025
−1082GG0.51450.12–0.91.01

When two models for the anti-HAV response were compared, the first including sex and smoking only, the second including sex, smoking and the IL-101082A polymorphism, the model including the −1082A polymorphism showed a better fit than the model without the polymorphism (likelihood ratio 16.0, 1DF, P < .001 [Table 4B]).

Because anti-HBs response showed a substantial heritable component, the contribution of shared IL-10 haplotypes was estimated by comparing intraclass correlations of DZ twin pairs who had an identical IL-10 haplotype with the DZ twin pairs who had different haplotypes (Table 5). Intraclass correlations (r values) of IL-10 haplotype identical DZ pairs (r = 0.49; 95% CI 0.20-0.70) were higher than the values for the DZ pairs with a different genotype (r = 0.37; 95% CI 0.07–0.61) but considerably lower than for MZ twins (r = 0.65; 95% CI 0.54–0.73. The contribution of the shared IL-10 promoter haplotype accounted for 27% of the total heritability. However, these results can only serve as a rough estimate since the number of MHC identical DZ pairs was small and 95% confidence intervals for the intraclass correlations are overlapping.

Table 5. Effect of Shared IL-10 Promoter: Intraclass Correlations of Anti-HBs Responses in IL-10 Promoter Identical DZ Twin Pairs Compared With All Dizygotic Twins and Monozygotic Twin Pairs
 Anti-HBs
Twin Pairs, nr (95% CI)
  1. Abbreviation: r, intraclass correlation.

Dizygotic twins, haplotype different250.37 (0.07–0.61)
Dizygotic twins, IL-10 haplotype identical230.49 (0.20–0.70)
Monozygotic880.65 (0.54–0.73)

Discussion

Numerous papers have studied the contribution of cytokine polymorphisms to the course of viral hepatitis.14–17 However, due to the heterogeneity of the studied populations, small sample sizes and ethnic differences results so far have been conflicting. Therefore, we have chosen to perform a prospective vaccination study to investigate the influence of IL-10 promoter polymorphisms on the immune response to HBsAg. Because gender, age, body mass index, and smoking significantly influence HBsAg responsiveness, this prospective approach allows the simultaneous assessment of these known confounders and the investigated IL-10 polymorphisms on HBsAg immune responsiveness. Our data show that the anti-HBs response is strongly influenced by polymorphisms in the IL-10 promoter. Individuals carrying the ACC haplotype had GMTs twice as high as individuals without this haplotype. The effect of the haplotype on the anti-HBs response was stronger than that of the −1082A polymorphism alone. We have previously shown by transfection of IL-10 reporter gene constructs into a monocytic cell line that the IL-10 ACC promoter haplotype shows a 20% lower transcriptional activity compared to the GCC haplotype.7, 18 The −1082G/A polymorphism is located within an Ets binding site. The −1082A allele confers higher binding affinity for the transcription factor PU.1 which inhibits gene expression and leads to decreased IL-10 expression in individuals carrying this haplotype.7 The influence of the IL-10 ACC haplotype appears to be as strong as that of age, gender, or BMI (Table 4A). There is increasing evidence that polymorphisms in the distal IL-10 promoter influence IL-10 gene transcription,19 thus, promoter polymorphisms in linkage disequilibrium with those investigated in this study or polymorphisms in nearby genes may contribute to the observed associations. Thus, genetically determined low IL-10 production appears to favor a strong humoral immune response to HBsAg. Responders to HBsAg vaccination show induction of IFN-γ but not IL-10–secreting T-helper cells.20In vitro the addition of IFN-γ and not TH2 cytokines resulted in strong increases of anti-HBs secreting B cells in vaccine recipients and chronic HBV carriers.21

IL-10 promoter haplotypes had the opposite effect on anti-HAV responsiveness. Individuals homozygous for the GCC haplotype had anti-HAV titres almost twice as high as individuals carrying −1082A haplotypes. Thus, higher IL-10 production leads to a stronger anti-HAV response. A high ratio of IL-10/IFN-γ production has been associated with high anti-HAV titres in a previous study.22 Differences in anti-HBs responses are most likely determined by the nature of the antigen. Twinrix contains the whole heat inactivated HAV to immunize against hepatitis A in contrast to the recombinant HBsAg used for HBV immunization. The combination vaccine has the potential advantage of bystander activation, for instance, costimulation of HBsAg-specific T-cells by cytokines secreted by HAV stimulated lymphocytes. Twinrix has been shown to elicit similar or stronger immune responses than the corresponding monovalent vaccines.

Comparison of DZ IL-10 haplotype identical and haplotype different DZ twins allows the estimation of the relative contribution of the IL-10 ACC haplotype to the overall heritability of HBsAg responsiveness. We had previously shown that more than 60% of the antibody response to HBsAg is determined by genetic variability.1 Approximately 25% of heritability appears to be determined by variability in the IL-10 promoter, shown by higher intraclass correlations of IL-10 haplotype identical in comparison with IL-10 haplotype different DZ twins. However, these results can only serve as a rough estimate because the number of MHC identical DZ pairs was small and 95% CIs for the intraclass correlations are overlapping. The effect of the IL-10 promoter variability on HBsAg responsiveness is weaker than that of the MHC, which had been estimated to account for 40% of the genetic determination of vaccine response. It is very likely that a considerable number of other genes contribute to the strong genetic influence on HBsAg responsiveness.

In conclusion, although the vaccination model is a simplification of the complex immune response during viral infections, our data underline the importance of IL-10 promoter variability for humoral immune responses to HBsAg and HAV. They provide an estimate of the relative contribution of the contribution of cytokine promoter haplotypes to the overall heritability of immune responsiveness.

Ancillary