Immune profiling of COVID-19 vaccine responses in people with multiple sclerosis on B cell-depleting therapy (preprint)

ABSTRACT

Background and ObjectivePeople with multiple sclerosis (pwMS) receiving B cell-depleting therapies have impaired antibody responses to vaccination. In a proportion of individuals, repeat vaccination against COVID-19 leads to seroconversion. We sought to describe the immune phenotype of pwMS on ocrelizumab, and identify clinical and immunological determinants of an effective vaccine response. MethodsThis was a single-centre, prospective cohort study. Peripheral blood samples were collected from pwMS receiving ocrelizumab (n = 38) pre and post administration of a third dose of mRNA COVID-19 vaccine. Immunogenicity was measured by T cell IFN{gamma} ELISpot, antibody titres, and live virus neutralisation. Humoral immunity was benchmarked against pwMS receiving natalizumab (n = 15), and against a correlate of real-world protection (50% reduction in incidence of infection) from SARS-CoV-2 ancestral and omicron BA.5 variants. The peripheral immune phenotype was comprehensively assessed by flow cytometry, and potential clinical and phenotypic determinants of response to vaccination identified. ResultsImmune cell populations relevant to disease and vaccine response were altered in pwMS receiving ocrelizumab versus natalizumab treatment, including depleted CD20-expressing B cell, T cell and NK cell populations, and elevated CD27+CD38+ T cell and NK8 cell frequencies. Following a third vaccine dose, 51% of pwMS on ocrelizumab were seropositive for SARS-CoV-2 receptor-binding-domain IgG, and 25% and 14% met the threshold for effective neutralisation of live SARS-CoV-2 ancestral and omicron BA.5 virus, respectively. B cell frequency at the time of vaccination, but not time since ocrelizumab infusion, was positively correlated with antibody response, while a strong negative correlation was observed between CD56bright NK cell frequency and antibody response in the ocrelizumab group. In this exploratory cohort, CD3-CD20+ B cells (% of lymphocytes; OR=3.92) and CD56bright NK cells (% of NK cells; OR=0.94) were predictive of an effective neutralising antibody response in second dose non-responders (AUC 0.98). DiscussionOcrelizumab treatment was associated with an altered immune phenotype, including recently described T cell and NK populations with potential roles in disease pathogenesis. However, seroconversion was severely impaired by ocrelizumab, and less than half of those who seroconverted following a third vaccine dose demonstrated effective immunity against SARS-CoV-2 ancestral or omicron BA.5. B cell frequency was associated with an effective antibody response, while immunomodulatory CD56bright NK cells were identified as a potential negative determinant of response in those with inadequate B cell numbers. Immune phenotype rather than time since ocrelizumab infusion may help to stratify individuals for prophylaxis.