A new study claims that Epstein-Barr virus (EBV) infection is the cause of multiple sclerosis, suggesting that people who do not acquire the virus will not develop MS (Bjornevik et al. Science 2022;375:296-301).
The study examined blood samples from the U.S. Department of Defense serum repository for the period 1993-2013. The repository routinely collects blood from military personnel for HIV screening at the start of military service and every two years thereafter. Samples were tested for EBV antibody positivity; the methodology was not reported. Incident MS cases (n=801) were matched with non-MS controls (n=1566).
Overall, only 1 of 801 MS cases (0.12%) occurred in an EBV-negative individual (hazard ratio 26.1 for EBV-positive vs. -negative). Among MS subjects who were EBV-negative at baseline, 34 of 35 seroconverted prior to MS onset; 1 remained EBV-negative. The median time from the first EBV-positive sample to MS onset was 5 years (range 0-10 years). The hazard ratio for MS was 32.4 with seroconversion compared to remaining seronegative. The rate of acquisition of another herpesvirus, cytomegalovirus (CMV), was similar in subjects who later developed MS and controls, suggesting that CMV infection is unrelated to the etiology of MS.
These results were similar to those reported by the same group a decade ago (Levin et al. Ann Neurol 2010;67: 824-830). That study also used DoD repository samples and compared 305 individuals who developed MS and 610 matched controls. All 10 MS patients who were EBV-negative at baseline seroconverted prior to MS onset. That study concluded that EBV infection was an MS risk factor rather than a cause of MS because the findings could not rule out possible confounds or reverse causation, i.e. that the MS disease process increased the risk of EBV infection.
The same limitation applied to the present study, so the authors added other analyses in an attempt to show causality. The group examined levels of serum neurofilament-light (NfL), a marker of neuroaxonal damage, before, during and after EBV infection. The same group previously used DoD blood samples to show that NfL levels were significantly elevated about 6 years prior to MS onset compared to controls (Bjornevik et al. JAMA Neurol 2020;77:58-64).
The present study used a smaller analysis of samples obtained from 25 MS patients and 79 controls. sNFL levels in cases and controls were similar before EBV infection and at initial acquisition. However, sNFL increased significantly more after EBV infection in MS patients compared to controls. For the MS group, the within-person change was 60%, although it should be noted that absolute change in sNfL was small (from 5 pg/mL to about 8-9 pg/mL) and the 95% confidence interval was large. The authors stated that there were no signs of neuroaxonal degeneration prior to EBV seroconversion, so the increase in sNfL indicated that the MS pathological process only began after EBV infection. The conclusion was that EBV is a cause and not a consequence of MS.
This may overstate the results and the study has a number of limitations. The demographics of the military personnel studied differs substantially from that of the MS population. The sample size of the sNfL analysis was also small; the group included data from its 2010 study, and it is unclear why the same was not done for the sNfL analysis from 2020. One MS patient in this study was EBV-negative, which may have been due to a laboratory error. EBV-negative CIS/MS has been reported although cases were extremely rare in the International CIS study (1 in 1047 cases) (Dobson et al. Neurol Neuroimmunol Neuroinflamm 2017;4:e318). It would be useful to confirm causality in other populations, such as pediatric-onset MS patients, and to demonstrate EBV infection prior to the development of radiologically isolated syndrome (RIS).
However, the data do suggest that EBV infection is necessary, if not sufficient, for the development of MS. Following the primary infection, other genetic, environmental and immune factors likely determine the onset and subsequent development of MS.
If EBV were the cause of MS, EBV vaccination could eradicate MS provided the vaccine induced sterilizing immunity. In January, Moderna began a U.S. phase I trial of mRNA-1189, an mRNA vaccine candidate that encodes EBV envelope glycoproteins (gH, gL, gp42, gp220) involved in cell entry. Antiviral drugs have already shown some benefit in MS but have been ineffective in clearing latent EBV (Lycke et al. J Neurol 1996;243:214-224. Bech et al. Neurology 2002;58:31-36).
There is also the potential for EBV-specific T cell therapies to target EBV-infected autoreactive B cells presenting EBV antigens (e.g. EBV nuclear antigen 1 [EBNA-1], latent membrane protein [LMP]-1 and -2A). Interestingly, a lower EBNA-1 titre has been shown to be associated with an improved response to treatments such as natalizumab (Dominguez-Mozo et al. Sci Rep 2020;10:14244). At least one phase I trial of EBV-specific CD8+ treatment has already been completed and indicated some clinical benefit (Pender et al. JCI Insight 2018;3:e124714).
However, the window of opportunity for intervention might be narrow. While EBV infection may be the initiating event, epitope spreading, and downstream pathological processes may make EBV a less important factor later in the clinical course. This could explain why some studies have reported no association of serum EBV antibodies with clinical or radiological disease activity (Giess et al. PLoS One 2017;12:e017527).
An intriguing implication is that at least some of the efficacy associated with B cell-directed therapies and autologous stem cell transplantation is due to depletion of EBV-infected B cells. It remains to be determined if B cell-directed therapies can be fine-tined to target only EBV-infected cells, or if the efficacy of anti-CD20 agents could be enhanced or sustained with concomitant use of antiviral agents, as some authors have recently discussed (Bar-Or et al. Trends Mol Med 2020;26:296-310).