High EBV-reactive IgG levels in serum are linked to typical inflammatory cerebrospinal fluid features in multiple sclerosis

Background Epstein-Barr virus (EBV) infection is strongly associated with multiple sclerosis (MS). However, it is ill-defined which of the MS-typical inflammatory alterations of basic cerebrospinal fluid (CSF) parameters, i.e. intrathecal cellular inflammation versus intrathecal total production of antibodies, are primarily linked to EBV-mediated humoral immune signatures in serum and/or CSF. Methods We analyzed paired serum and cerebrospinal fluid (CSF) samples from 316 MS patients from Switzerland and Sweden. Serum IgG levels (sIgG) against EBV nuclear antigen 1 (EBNA1) and EBV capsid antigen (EBV-CA) were quantified by ELISA, and high levels (IgG hi ) were defined as the highest quartile (Q4) of serum EBNA1- or EBV-CA-specific IgG levels (sEBNA1-IgG or sEBV-CA-IgG), respectively. Intrathecal EBV-specific antibody production was defined as CSF/serum antigen-specific antibody index (CAI) ≥1.5. We assessed associations of EBV-specific measures with CSF leukocyte count, presence of pleocytosis, presence of blood-CSF barrier dysfunction, presence of CSF-specific oligoclonal bands (OCB) and intrathecal immunoglobulin indices/fractions using correction for multiple comparisons, as well as age- and sex-adjusted models and continuous dose–response analyses. Results sEBNA1-IgG hi levels were associated with higher CSF leukocyte counts (median 6.0 vs 3.0 cells/µL; q=0.022), more frequent pleocytosis (55.7% vs 37.6%; q=0.022), and higher rate of CSF-specific OCB (93.7% vs 79.3%; q=0.029) compared to lower sEBNA1-IgG levels. sEBV-CA-IgG hi levels were not associated with cellular CSF parameters, but with a higher total IgG index (0.78 vs 0.65; q=0.016). In age- and sex-adjusted models, a 10-fold increase in sEBNA1-IgG levels was associated with a 2.54-fold higher CSF leukocyte count (β=0.405, 95% CI 0.217–0.592, p=2.86×10 –5 ) and 2.73-fold higher odds for pleocytosis (OR = 2.73, 95% CI 1.59–4.71, p=2.83×10 –4 ), whereas sEBV-CA-IgG showed no significant correlation. Dose-response models showed monotonic associations between sEBNA1-IgG levels and both CSF leukocyte counts and general intrathecal humoral activation, whereas sEBV-CA-IgG levels were associated mainly with overall intrathecal humoral activity. Intrathecal EBV-specific antibody production was uncommon and not associated with basic CSF inflammatory or barrier parameters. Conclusion These findings are consistent with an outside-in model, in which peripheral, but not intrathecal activation of EBV-specific humoral immunity, and rather sEBNA1-IgG hi - than sEBV-CA-IgG hi -reactivity links to general intrathecal cellular and humoral immune activation in MS.