key: cord-0256218-8tarj62m authors: Stefanski, A.-L.; Rincon-Arevalo, H.; Schrezenmeier, E.; Karberg, K.; Szelinski, F.; Ritter, J.; Jahrsdoerfer, B.; Schrezenmeier, H.; Ludwig, C.; Sattler, A.; Kotsch, K.; Chen, Y.; Claussnitzer, A.; Haibel, H.; Proft, F.; Guerra, G. M.; Durek, P.; Heinrich, F.; Gomes, M. F.; Burmester, G. R.; Radbruch, A.; Mashreghi, M.-F.; Lino, A. C.; Doerner, T. title: B cell numbers predict humoral and cellular response upon SARS-CoV-2 vaccination among patients treated with rituximab date: 2021-07-22 journal: nan DOI: 10.1101/2021.07.19.21260803 sha: 2aaea1b2b0481c81fdcc858874a57f9fecbc7f77 doc_id: 256218 cord_uid: 8tarj62m Objectives: Patients with autoimmune inflammatory rheumatic diseases receiving rituximab (RTX) therapy show substantially impaired anti-SARS-CoV-2 vaccine humoral but partly inducible cellular immune responses. However, the complex relationship between antigen-specific B and T cells and the level of B cell repopulation necessary to achieve anti-vaccine responses remain largely unknown. Methods: Antibody responses to SARS-CoV-2 vaccines and induction of antigen-specific B and CD4/CD8 T cell subsets were studied in 19 rheumatoid arthritis (RA) and ANCA-associated vasculitis (AAV) patients receiving RTX, 12 RA patients on other therapies and 30 healthy controls after SARS-CoV-2 vaccination with either mRNA or vector based vaccines. Results: A minimum of 10 B cells/uL in the peripheral circulation was necessary in RTX patients to mount seroconversion to anti-S1 IgG upon SARS-CoV-2 vaccination. RTX patients lacking IgG seroconversion showed reduced antigen-specific B cells, lower frequency of TfH-like cells as well as less activated CD4 and CD8 T cells compared to IgG seroconverted RTX patients. Functionally relevant B cell depletion resulted in impaired IFNgamma secretion by spike-specific CD4 T cells. In contrast, antigen-specific CD8 T cells were reduced in patients independently of IgG formation. Conclusions: Patients receiving rituximab with B cell numbers above 10 B cells/ul were able to mount humoral and more robust cellular responses after SARS-CoV-2 vaccination that may permit optimization of vaccination in these patients. Mechanistically, the data emphasize the crucial role of co-stimulatory B cell functions for the proper induction of CD4 responses propagating vaccine-specific B and plasma cell differentiation. Infectious diseases and associated complications comprise an important cause of morbidity and mortality in patients with autoimmune inflammatory rheumatic diseases (AIIRDs) 1 . Increased susceptibility to infectious diseases in these patients is most likely due to an immunosuppressive effect of the disease itself and/or related to immunosuppressive treatment 2 . COVID-19, caused by the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) requires particular considerations in AIIRD patients by rheumatologists. Rituximab (RTX), an anti-CD20 monoclonal antibody leading to B cell depletion and used in AIIRDs like rheumatoid arthritis (RA) and ANCA-associated vasculitis (AAV), has been found as risk factor for poor COVID-19 associated outcomes. 3 4 . Since a suitable treatment for COVID-19 has not been developed yet, vaccination is of crucial importance to protect these vulnerable patients. Meanwhile, various phase III clinical trials have demonstrated the efficacy and safety of mRNAbased vaccines (BNT162b2 5 6 , mRNA-1273 7 ) and viral vector-based vaccines (ChAdOx1 8 , Ad26.COV2.S 9 ) to prevent severe COVID-19 disease or death. In AIIRD patients, vaccination is generally regarded safe and efficacious 10 . However, in particular under B cell depleting therapy with RTX, hampered humoral and cellular responses following influenza, pneumococcal and hepatitis B vaccination have been reported [11] [12] [13] [14] [15] [16] . Data available about SARS-CoV-2 vaccine response in rituximab treated AIIRD patients reveal substantially impaired humoral [17] [18] [19] but partly inducible cellular immune responses 20 . However, little is known about the complex mechanisms between T, B and plasma cells, as well as the level of B cell repopulation necessary for proper vaccine response among RTX patients. In this study, we investigated the characteristics of humoral and cellular antigen-specific CD4/CD8 and B cell immune response upon SARS-CoV-2 vaccines in patients treated with RTX compared with HC and RA patients on other therapies. This study recruited 19 patients receiving rituximab (16 RA and 3 AAV patients, RTX group), 30 healthy controls (HC group) and 12 RA patients on other therapies as additional control group (RA group). Most study participants were vaccinated twice with the mRNA vaccine BNT162b2, one RTX patient received 2x mRNA-1273. There were 3 HC, one RA and one RTX vaccinated twice with the viral vector vaccine ChAdOx1 (indicated in green throughout the figures). According to national recommendations 21 , 3 RTX patients and 3 HC received 1x ChAdOx1 followed by a heterologous boost with 1x BNT162b2 (indicated in blue throughout All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 3 [2 -6] All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Antibody responses to SARS-CoV-2 vaccines were assessed in all individuals, 6±3 days after boost. All HC became positive for anti-S1 IgG and IgA and showed more than 90% SARS-CoV-2 neutralisation. Noteworthy, IgA and IgG anti-vaccine titres were markedly diminished 6±3 days after boost in the RA control group and especially in RTX patients compared to HC (Fig. 1A) . Anti-S1 IgG were detected in 8/12 (66.7%) of the RA group and 8/19 (42.1%) of the RTX group. Simultaneously, 5/12 (41.7%) and 9/19 (47.4%) of the RA and RTX group, respectively developed anti-S1 IgA antibodies. Virus neutralizing antibodies were found in 8/12 (66.7%) among RA control patients and 9/19 (47.4%) RTX group (Fig. 1A) . Two RTX patients with unknown prior infection (identified as anti-nucleocapsid protein positive Fig. S1 , red quadrates), developed high titers of anti-S1 IgG, IgA and neutralizing antibodies comparable with HC. As previously reported 22 , AIIRD patients may show a delayed humoral immune response after vaccination. To address this question, we collected additional samples from RA and RTX patients 3-4 weeks after boost (Fig. 1B) . Two RA and five RTX patients developed positive IgG antibodies, IgA was found in two RA and one RTX patients 3-4 weeks after boost. Neutralizing antibodies were detected in two RA and six RTX at this later time point. Among the RTX patients, who did not show seroconversion at day 7 after boost, there was a significant increase in IgG and neutralizing antibody formation at the later time point (Fig. 1C) . Noteworthy, IgG titers correlated with neutralizing antibodies (Fig. 1D) . Thus, and considering the delayed vaccine response, 10/12 (83.3%) of RA patients and 13/19 (68.4%) of the RTX patients showed IgG seroconversion with neutralizing antibodies after SARS-CoV-2 vaccination, even though in lower titers compared with HC. Interestingly, the data suggested that patients under RTX exhibited a potential dichotomous response: 13/19 RTX patients seroconverted to IgG (RTX IgG+), while 6/19 did not (RTX IgG-). To identify potential factors resulting in IgG seroconversion among RTX patients, further study addressed potential differences between the two groups. We analyzed the B cell compartment among the different groups (gating strategy shown in Fig. 2A ). RTX patients presented with significantly lower relative and absolute B cell numbers compared to HC and RA control group (Fig. 2B) . Notably, a significant difference in the frequency and absolute B cell number was also found between RTX IgG+ and nonseroconverted RTX patients (Fig 2C) . In our RTX cohort, 10 B cells/µl in peripheral circulation All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2021. ; https://doi.org/10.1101/2021.07.19.21260803 doi: medRxiv preprint (or 0.4% of lymphocytes accordingly) were identified as the minimum to mount seroconversion to anti-S1 IgG among RTX treated patients (Fig 2C) . In the RTX group, B cell numbers correlated with humoral anti-vaccine responses since the absolute number of B cells (Fig. 2D) and B cell frequency (Fig. S2A ) correlated with anti-S1 IgG titers and even more pronounced with neutralizing antibodies. This clearly suggests that humoral protection elicited by vaccination is dependent on the critical availability of B cells in RTX treated patients. In the RA and HC groups we did not find a significant correlation between B cell numbers and serologic response (data not shown), suggesting that the correlation between B cell numbers and IgG response is restricted to patients with B cell counts below the lower limits of normal. Next, we studied SARS-CoV-2 specific B cell responses in RTX treated patients, using flow cytometry to quantify RBD-specific B cells in peripheral blood 23 (gating strategy shown in Fig 2A) . While no significant difference was seen between HC, RA and the RTX group ( for HC 23 , RA control and RTX patients were able to generate IgG+ plasmablasts upon vaccination. We found no significant difference between the groups regarding RBD+ B cell subset distribution or immunoglobulin isotypes ( Fig. 2H and S2B, C). Simultaneously, we wondered how dynamics of CD4/8 T cell subsets interrelate with induction of vaccine-specific B cells and IgG. Contrary to B cells, there was no difference regarding the frequency, absolute numbers or memory formation in CD4 (Fig. S3A , C) and CD8 T cells (Fig. S3B , D) between HC, RA and RTX patients. Subsequent analysis addressed the differences between vaccine-responders and non-responders among RTX patients (representative gates shown in Fig. 3A ). Interestingly, patients who lacked anti-vaccine IgG antibodies showed lower frequencies of circulating TfH-like CD4 T cells, defined as CD4+CXCR5+PD1+, as well as of activated CD4/8 T cells co-expressing CD38+HLA-DR+ (Fig. 3B ). Activated CD4 T cells correlated with absolute B cell numbers (Fig. 3C ). This data suggests an impaired bidirectional All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. vaccination-induced humoral immunity. The overall occurrence of spike-specific CD4 T cells (representative gates shown in 3D) compared with unstimulated controls was found similar in all groups: 86.7% (13/15) in HC, 83% (10/12) in RA and 73.7% (14/19) in RTX patients ( Fig 3E) . This was also consistent with a comparable magnitude of the response between the groups (Fig. 3F ) as well as similar memory subset distribution (Fig. S4C) . A more detailed study of the RTX group showed that the majority of RTX IgG+ patients (10/13, 76.9%) versus 50% of RTX IgG-(3/6) patients showed an appropriate antigen-specific CD4 T cell increase upon stimulation. With regard to functional analyses of cytokine secretion by spike-specific CD4 T cells, non-seroconverted RTX patients showed a significantly reduced TNFα ( Fig 3G) and IFNγ (Fig. 3H ) production compared to RTX IgG+ responders (representative gates shown in Fig. S4A ). Since most patients in the non-seroconverted RTX group had very low circulating B cell counts, we wondered if there is a relation between reduced B cells and impaired cytokine production by antigen-specific CD4+ T cells. Indeed, IFNγ but not TNFα production showed a correlation with absolute B cell number, suggesting the importance of B cell co-stimulatory functions for the proper and interactive induction of CD4 responses. Compared with unstimulated controls, 93.3% of HC (14/15) but only 58.3% of RA control (7/12) and 57.9% of RTX patients (11/19) showed an increase of spike-specific CD8 T cells coexpressing CD137 and IFNγ (as shown in Fig. 3J and Fig. S4B ) upon stimulation (Fig. 3K) . To assess the degranulation function of CD8 T cells, we analyzed the co-expression of CD107a and IFNγ. The responder rate for CD8 T cells co-expressing CD107a and IFNγ after stimulation was overall low with 60% in HC (9/15), 41.6% (5/12) in the RA control and 42.1% (8/19) in the RTX group (data not shown). Regarding the amplitude of CD8 responses, spike-specific CD8 T cells, co-expressing CD137 and IFNγ (Fig. 3L ) and CD107a and IFNγ (Fig. S4C) , as well as their memory subset distribution were comparable between all groups (Fig. S4D-E) . To identify further predictive factors for IgG seroconversion in RTX patients, we performed a correlation matrix (Fig. 4) including antigen-specific T and B cell subsets as well as demographic data. IgG titers and neutralizing antibodies correlated with RBD+ plasmablasts and memory compartments, as we previously showed 23 . Furthermore, neutralizing antibodies All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2021. ; https://doi.org/10.1101/2021.07.19.21260803 doi: medRxiv preprint correlated with the frequency of activated CD38+HLA-DR+ CD4/8 T cells as well as with IFNγ producing antigen-specific CD4 T cells. Activated CD38+HLA-DR+ CD4/8 T cells correlated also with RBD+ plasmablasts, while circulating TfH-like CD4 T cells with total RBD+ B cells. Notably, there was a correlation between TNFα and IFNγ producing antigen-specific CD4 T cells with RBD+ plasmablasts and switch-memory B cells. There was a correlation between IgG titer and interval of time since the last RTX infusion, but no significant correlation with age or with DAS28. Interestingly, induced antigen-specific CD8 responses upon stimulation did not correlate with humoral immunity, neither with B nor CD4 T cell subsets, suggesting an independent, more direct antigen-driven cellular immunity compared with CD4/CD19-interaction required for IgG formation. To further investigate the specific differences during SARS-CoV-2 vaccination in RTX treated patients, we sorted CD27++CD38++ plasmablasts, CD27+ memory B cells and HLA- (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Therefore, our study aimed to investigate humoral and cellular responses in RTX treated patients versus controls. Consistent with previous data 17- 19 22 , serologic IgG conversion with formation of neutralizing antibodies was significantly lower and delayed in both, RA and even more pronounced in the RTX cohort compared to HC. This finding was closely linked to the availability of peripheral B cells, activated CD4/8 T cells as well as circulating TFH-like cells. Ongoing antigen exposure through mRNA vaccines seems to permit prolonged GC maturation 28 , which might be an explanation for the further increase in antibody titers during an additional period in some patients. Besides IgG non-responders among RTX patients, two patients in the RA group with normal B cell numbers did not develop anti-S1 IgG antibodies. After completing the analysis, the underlying cause may be most probably related to impaired T cell responses: in one patient due to inhibition of co-stimulation by abatacept, consistent with a prior report 18 . The other patient treated with a JAK inhibitor lacked cytokine production by antigen-specific T cells after ChAdOx1 booster. Even though significantly lower IgG responses were reported upon ChAdOx1 booster in healthy vaccinees 29 , it remains to be delineated whether the treatment and/or selected vaccine may account for this finding. Induction of vaccine-specific IgG in individuals upon ChAdOx1/BNT162b2 was comparable with twice BNT162b2 vaccinations. Interestingly, IgA formation was comparable across all groups, although the protective potency of IgA remains to be determined. Of utmost importance, our RTX cohort showed a correlation between IgG seroconversion, neutralizing antibodies and absolute B cell number. Here, a minimum of 10 B cells/µl in the peripheral circulation appeared to be a threshold to signify an appropriate cellular and humoral vaccination response. Patients with B cell numbers below this threshold presented not only with lower antigen-specific B cells, but also they showed substantially diminished circulating TfH-like CD4 T cells, reduced activated CD4/8 T cells co-expressing CD38 and HLA-DR, as well as impaired IFNγ secretion of spike-specific CD4 T cells. The frequency of IFNγ secreting antigen-specific CD4 T cells also correlated with the absolute number of B cells, suggesting that these cells interact to achieve proper anti-S1 responses. Mechanistically, the current data suggest the critical role of available co-stimulatory B cell functions for the induction of proper CD4 Th response. This is consistent with observations of previously described impaired B-T cell crosstalk in rituximab treated patients 30-32 , leading to reduced frequencies of activated T cells 32 , downregulation of CD40L in CD4 T cells 30 31 , and reduced antigen-specific CD8 T cells after influenza vaccination 14 . All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. With regard to induction of antigen-specific CD8 T cells upon stimulation, the RTX and RA groups showed a reduced responder rate compared with HC. However, other than for antigenspecific CD4 T cells, neither B cell depletion nor IgG formation correlated with spike-specific CD8 T cells, suggesting that their induction occurred independently upon SARS-CoV-2 vaccination. It is not clear how these vaccine-specific CD8 T cells provide antiviral protection on clinical grounds. The debate about what correlates with protection after vaccination against SARS-CoV-2 is ongoing, while it is widely accepted, that neutralizing antibodies are considered to be a reliable surrogate of protection against virus variants 33 34 . The threshold for protective SARS-CoV-2 IgG-titer is still unknown, although non-human primate studies suggest that it is likely very effective already at low titers 35 . Our study provides evidence that detection of RBD-specific B and spike-specific CD4 T cells may provide cellular correlates of this response, while the CD8 response occurred in an independent way. The role of these two lines of vaccine response needs to be further delineated. Limitations of the study are the small number of RA and RTX patients and the heterogeneity of the groups (including different DMARD regimes, different vaccination strategies). However, co-medication with prednisolone or methotrexate do not appear to bias our results, since there was no difference in humoral response in our RTX cohort upon vaccination (Fig. S7A, B) . Here, we present a first study investigating humoral as well as antigen-specific T and B cellular (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2021.  Future treatment developments targeting T cells may take into account the directive role of critical B cell counts, All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. ChAdOx1 followed by 1x BNT162b2, indicated in blue. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 22, 2021. ; https://doi.org/10.1101/2021.07.19.21260803 doi: medRxiv preprint Comparing the burdens of opportunistic infections among patients with systemic rheumatic diseases: a nationally representative cohort study Incidence and prevalence of vaccine preventable infections in adult patients with autoimmune inflammatory rheumatic diseases (AIIRD): a systemic literature review informing the 2019 update of the EULAR recommendations for vaccination in adult patients with AIIRD No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity Factors associated with COVID-19-related death in people with rheumatic diseases: results from the COVID-19 Global Rheumatology Alliance physician-reported registry COVID-19 Outcomes in Patients Undergoing B Cell Depletion Therapy and Those with Humoral Immunodeficiency States: A Scoping Review Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine BNT162b2 mRNA Covid-19 Vaccine in a Nationwide Mass Vaccination Setting Efficacy and Safety of the mRNA-1273 SARS-CoV-2 Vaccine Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK Safety and Efficacy of Single-Dose Ad26.COV2.S Vaccine against Covid-19 2019 update of EULAR recommendations for vaccination in adult patients with autoimmune inflammatory rheumatic diseases Effect of methotrexate, anti-tumor necrosis factor α, and rituximab on the immune response to influenza and pneumococcal vaccines in patients with rheumatoid arthritis: a systematic review and meta-analysis Is rheumatoid arthritis associated with reduced immunogenicity of the influenza vaccination? A systematic review and meta-analysis The effect of rituximab on vaccine responses in patients with immune thrombocytopenia Evaluation of the immune response to hepatitis B vaccine in patients on biological therapy: results of the RIER cohort study SARS-CoV-2 vaccination in rituximab-treated patients: evidence for impaired humoral but inducible cellular immune response LB0003 IMMUNOGENICITY AND SAFETY OF THE BNT162b2 mRNA COVID-19 VACCINE IN ADULT PATIENTS WITH AUTOIMMUNE INFLAMMATORY RHEUMATIC DISEASES AND GENERAL POPULATION: A MULTICENTER STUDY Glucocorticoids and B Cell Depleting Agents Substantially Impair Immunogenicity of mRNA Vaccines to SARS-CoV-2. medRxiv Correspondence on 'SARS-CoV-2 vaccination in rituximab-treated patients: evidence for impaired humoral but inducible cellular immune response SARS-CoV-2 vaccination responses in untreated, conventionally treated and anticytokine-treated patients with immune-mediated inflammatory diseases Impaired humoral immunity to SARS-CoV-2 BNT162b2 vaccine in kidney transplant recipients and dialysis patients Sigmoidal model with 95% confidence bands We kindly thank all study participants for taking part and all team members of the rheumatology outpatient clinic Steglitz (Dr. Kirsten Karberg and Dr. Henning Christian Brandt), Berlin, as well as of AGZ rheumatology Charité Mitte Berlin (Dr. Anne Claußnitzer), Germany for organization of patient schedules. The concept of the study was developed by ALS, HRA, ES, ACL, and TD.