Serologic Response to the Epstein-Barr Virus Peptidome and the Risk for Multiple Sclerosis

Abstract

Importance: It remains unclear why only a small proportion of individuals infected with the Epstein-Barr virus (EBV) develop multiple sclerosis (MS) and what the underlying mechanisms are.

Objective: To assess the serologic response to all EBV peptides before the first symptoms of MS occur, determine whether the disease is associated with a distinct immune response to EBV, and evaluate whether specific EBV epitopes drive this response.

Design, setting, and participants: In this prospective, nested case-control study, individuals were selected among US military personnel with serum samples stored in the US Department of Defense Serum Repository. Individuals with MS had serum collected at a median 1 year before onset (reported to the military in 2000-2011) and were matched to controls for age, sex, race and ethnicity, blood collection, and military branch. No individuals were excluded. The data were analyzed between September 1, 2022, and August 31, 2023.

Exposure: Antibodies (enrichment z scores) to the human virome measured using VirScan (phage-displayed immunoprecipitation and sequencing).

Main outcome and measure: Rate ratios (RRs) for MS for antibodies to 2263 EBV peptides (the EBV peptidome) were estimated using conditional logistic regression, adjusting for total anti-EBV nuclear antigen 1 (EBNA-1) antibodies, which have consistently been associated with a higher MS risk. The role of antibodies against other viral peptides was also explored.

Results: A total of 30 individuals with MS were matched with 30 controls. Mean (SD) age at sample collection was 27.8 (6.5) years; 46 of 60 participants (76.7%) were male. The antibody response to the EBV peptidome was stronger in individuals with MS, but without a discernible pattern. The antibody responses to 66 EBV peptides, the majority mapping to EBNA antigens, were significantly higher in preonset sera from individuals with MS (RR of highest vs lowest tertile of antibody enrichment, 33.4; 95% CI, 2.5-448.4; P for trend = .008). Higher total anti-EBNA-1 antibodies were also associated with an elevated MS risk (top vs bottom tertile: RR, 27.6; 95% CI, 2.3-327.6; P for trend = .008). After adjusting for total anti-EBNA-1 antibodies, risk estimates from most EBV peptides analyses were attenuated, with 4 remaining significantly associated with MS, the strongest within EBNA-6/EBNA-3C, while the association between total anti-EBNA-1 antibodies and MS persisted.

Conclusion and relevance: These findings suggest that antibody response to EBNA-1 may be the strongest serologic risk factor for MS. No single EBV peptide stood out as being selectively targeted in individuals with MS but not controls. Larger investigations are needed to explore possible heterogeneity of anti-EBV humoral immunity in MS.

Figure 1.. Magnitude of Antibody Response to Epstein-Barr Virus (EBV) Peptides in Cases With Multiple Sclerosis (MS) Compared With Matched Controls

Antibody response to the 30 EBV peptides in preonset serum samples ranked by statistical significance (r1-r30) and indicating the EBV protein, UniProt ID, and amino acid (aa) sequence. A, Derived as the mean z score in identical replicates, with higher levels indicating a stronger response. A z score greater than or equal to 3.5 is an epitope hit and indicates presence of antibodies to a peptide in VirScan. Matched pairs of cases and controls are shown in the same order. Compare with eFigure 1 in Supplement 1, which shows the peptide grouped by EBV protein. B, Ordered according to strength of the immune response to peptides in the cases. EBNA indicates Epstein-Barr virus nuclear antigen.

Figure 2.. Serologic Profiles for the Epstein-Barr Virus (EBV) Peptidome in Cases With Multiple Sclerosis (MS) and Matched Controls and Their Discriminatory Potential

A, Antibody response profiles to all 2263 EBV peptides are shown. B, Amino acid (aa) sequences for full-length EBV nuclear antigen 1 (EBNA-1), UniProt ID P03211 are shown. C, Includes the antibody response to all 2263 EBV peptides (permutation test with 1000 permutations, P = .04 for the separation of cases and controls). The EBV peptides contributing most to the separation according to variable importance in the projection scores are shown in eFigure 3 in Supplement 1.

Figure 3.. Antibody Response to Epstein-Barr Virus (EBV) Peptidome and Risk for Multiple Sclerosis (MS)

Rate ratios (RRs) for MS from conditional logistic regression models comparing top vs bottom tertile of antibody response to the 625 EBV peptides with at least 1 epitope hit (z score ≥3.5 in technical replicates) in 1 individual and P value for trend across tertiles. The models in panel B were additionally continuously adjusted for standardized total anti–EBV nuclear antigen 1 (EBNA-1) antibody level (total z score to EBNA-1 peptides). Statistically significant results are plotted above the dotted horizontal line (P for trend <.05). The EBV peptides with RRs less than 0.01 or greater than 100 (all with P >.05) are not shown to facilitate plotting. Of the 10 excluded peptides, the antibody response to 4 of these was associated with MS in unconditional logistic regression models adjusted for sex, age, race and ethnicity, and total anti-EBNA-1 antibodies (comparing top vs bottom tertiles, Q07286 aa57-112: odds ratio [OR], 25.12 [95% CI, 3.23-388.77; P for trend = .006]; Q1EH47 aa18-73: OR, 18.47 [95% CI, 2.50-230.59; P for trend = .01]; P03213 aa85-140: OR, 7.63 [95% CI, 1.41-55.08; P for trend = .03]; Q3KSU8 aa673-728: OR, 0.12 [95% CI, 0.01-0.65; P for trend = .02]).

Figure 4.. Antibody Response to Epstein-Barr Virus (EBV) Peptides Associated With Multiple Sclerosis (MS) Before and After Adjusting for Total Anti–EBV Nuclear Antigen 1 (EBNA-1) Antibodies

Rate ratios (RRs) for MS from conditional logistic regression models comparing the top to the bottom tertile of antibody response to EBV peptides with at least 1 epitope hit (z score ≥3.5 in technical replicates) in 1 individual. All peptides that were statistically significantly associated with MS before adjusting for total anti-EBNA-1 are shown (62 of 71 EBV peptides described in the Results with both P <.05 comparing the top vs bottom tertile and P for trend across tertiles <.05). ^aRemained statistically significant (P for trend <.05) after adjusting continuously for standardized total anti-EBNA-1 antibody level (total z score to EBNA-1 peptides).