Genetics of immune response to Epstein-Barr virus: prospects for multiple sclerosis pathogenesis

Abstract

Epstein-Barr virus (EBV) infection has been advocated as a prerequisite for developing multiple sclerosis (MS) and possibly the propagation of the disease. However, the precise mechanisms for such influences are still unclear. A large-scale study investigating the host genetics of EBV serology and related clinical manifestations, such as infectious mononucleosis (IM), may help us better understand the role of EBV in MS pathogenesis. This study evaluates the host genetic factors that influence serological response against EBV and history of IM and cross-evaluates them with MS risk and genetic susceptibility in the Swedish population. Plasma IgG antibody levels against EBV nuclear antigen-1 [EBNA-1, truncated = amino acids (aa) (325-641), peptide = aa(385-420)] and viral capsid antigen p18 (VCAp18) were measured using bead-based multiplex serology for 8744 MS cases and 7229 population-matched control subjects. The MS risk association for high/low EBV antibody levels and history of IM was compared to relevant clinical measures along with sex, age at sampling, and associated HLA allele variants. Genome-wide and HLA allele association analyses were also performed to identify genetic risk factors for EBV antibody response and IM history. Higher antibody levels against VCAp18 [odds ratio (OR) = 1.74, 95% confidence interval (CI) = 1.60-1.88] and EBNA-1, particularly the peptide (OR = 3.13, 95% CI = 2.93-3.35), were associated with an increased risk for MS. The risk increased with higher anti-EBNA-1 IgG levels up to 12× the reference risk. We also identified several independent HLA haplotypes associated with EBV serology overlapping with known MS risk alleles (e.g. DRB1*15:01). Although there were several candidates, no variants outside the HLA region reached genome-wide significance. Cumulative HLA risk for anti-EBNA-1 IgG levels, particularly the peptide fragment, was strongly associated with MS. In contrast, the genetic risk for high anti-VCAp18 IgG levels was not as strongly associated with MS risk. IM history was not associated with class II HLA genes but negatively associated with A*02:01, which is protective against MS. Our findings emphasize that the risk association between anti-EBNA-1 IgG levels and MS may be partly due to overlapping HLA associations. Additionally, the increasing MS risk with increasing anti-EBNA-1 levels would be consistent with a pathogenic role of the EBNA-1 immune response, perhaps through molecular mimicry. Given that high anti-EBNA-1 antibodies may reflect a poorly controlled T-cell defence against the virus, our findings would be consistent with DRB1*15:01 being a poor class II antigen in the immune defence against EBV. Last, the difference in genetic control of IM supports the independent roles of EBNA-1 and IM in MS susceptibility.

Keywords: DRB1; EBNA1; EBV; GWAS; HLA.

Figure 1

Higher anti-EBV IgG levels associated with increasing risk for multiple sclerosis (MS). The left-side plot illustrates the proportion of MS cases (bottom) compared to controls (top) with increasing strata of Epstein-Barr virus (EBV) antibody levels. The plots following are continuous association curves (CAC), which were determined by comparing increasing and overlapping strata (5% quantile window, +0.5% step) with the lowest 5% quantile window as reference. Line-plots for (middle) odds ratios along with 95% confidence intervals (CI, dotted lines) and (right) P-values are plotted for each antibody measure. The predicted effect is illustrated by the smoothed red curved. Predicted cut-off for biological effect was determined by inflection point (green line) and shown in comparison with median antibody level among controls (blue line).

Figure 2

HLA haplotypes associated with anti-EBV IgG levels. The figure summarizes all HLA alleles associated to anti-EBV IgG levels with significance of P < 10^–4, organized by corresponding haplotype. Heat map gradients were used to illustrate both significance level (darker = more significant) and type of effect: risk (red) and protective (blue). HLA-DRB4*01 (yellow) and its subtypes 01:01 and 01:03 are associated with the two haplotypes, DRB1*04:01 and DRB1*07:01. Values showing the beta (β) and significance (P) among all subjects adjusted for sex, age at sampling, multiple sclerosis affection status and six principal components analysis (PCA) vectors are provided in the Supplementary material.

Figure 3

Cross-comparing the distribution of weighted genetic risk scores (wGRS) for MS and anti-EBV IgG levels. Asymmetric bean plots illustrate the distribution of weighted risk scores for either (i) multiple sclerosis (MS) between high/low antibody response [continuous association curves (CAC) cut-off, A–C]; or (ii) anti-EBV IgG levels between MS cases and controls (D). See Supplementary Tables 9–11 for additional details. Analyses of weighted genetic risk scores (wGRS) for MS (A) were also conducted separately for (B) HLA and (C) non-HLA genetic risks. For A–C, left- and right-side distributions represent low and high antibody response, respectively, based on the inflection-based method. Red and grey distributions correspond to MS cases and controls, respectively. The median is displayed for each distribution and a receiver operating characteristic (ROC) curve of each comparison is illustrated on the right.