Dynamics of SARS-CoV-2 immunity after vaccination and breakthrough infection in rituximab-treated rheumatoid arthritis patients: a prospective cohort study

Abstract

Background: SARS-CoV-2 vaccination in rheumatoid arthritis (RA) patients treated with B cell-depleting drugs induced limited seroconversion but robust cellular response. We aimed to document specific T and B cell immunity in response to vaccine booster doses and breakthrough infection (BTI).

Methods: We included 76 RA patients treated with rituximab who received up to four SARS-CoV-2 vaccine doses or three doses plus BTI, in addition to vaccinated healthy donors (HD) and control patients treated with tumor necrosis factor inhibitor (TNFi). We quantified anti-SARS-CoV-2 receptor-binding domain (RBD) Spike IgG, anti-nucleocapsid (NC) IgG, 92 circulating inflammatory proteins, Spike-binding B cells, and Spike-specific T cells along with comprehensive high-dimensional phenotyping and functional assays.

Findings: The time since the last rituximab infusion, persistent inflammation, and age were associated with the anti-SARS-CoV-2 RBD IgG seroconversion. The vaccine-elicited serological response was accompanied by an incomplete induction of peripheral Spike-specific memory B cells but occurred independently of T cell responses. Vaccine- and BTI-elicited cellular immunity was similar between RA and HD ex vivo in terms of frequency or phenotype of Spike-specific cytotoxic T cells and in vitro in terms of the functionality and differentiation profile of Spike-specific T cells.

Interpretation: SARS-CoV-2 vaccination in RA can induce persistent effector T-cell responses that are reactivated by BTI. Paused rituximab medication allowed serological responses after a booster dose (D4), especially in RA with lower inflammation, enabling efficient humoral and cellular immunity after BTI, and contributed overall to the development of potential durable immunity.

Keywords: ACPA; B cell; COVID-19; T cell; breakthrough infection; mRNA vaccination; rheumatoid arthritis; rituximab.

Figure 1

Serological responses in rituximab-treated rheumatoid arthritis (RA) patients. (A) Serological response post-vaccination. Longitudinal IgG anti-RBD (BAU/mL) 1 month (m1) after dose 2 (D2) to dose 4 (D4) in patients (n = 65, 63, 55) with at least two samples. Median titer for RA D4 was 586 BAU/mL [IQR: 1–7,390]. Wilcoxon matched-pairs signed-rank test; P-values are indicated. The P-values are indicated as *for p < 0.05 and ****p < 0.0001. (B) Serological response post-D4 of vaccine and rituximab therapy. IgG anti-RBD (BAU/mL) responses after D4 of the vaccine according to the time interval since the last rituximab infusion with a median of 288 days [IQR: 155–361] corresponding to 10 months (within the last 10 months in red or more than 10 months in blue) before vaccination. IQR of 74–20,726 BAU/mL for RA with more than 10 months and IQR of 1–2,814 BAU/mL for RA with recent therapy. (C) Serological response post-vaccination and anti-citrullinated protein auto-antibodies (ACPA). IgG anti-RBD (BAU/mL) in ACPA-negative [IQR: 1–4,911] and ACPA-positive RA [IQR: 86–10,079] were measured at D4. (D) Serological responses before and after breakthrough infections. The biplots show IgG anti-RBD (BAU/mL) vs. IgG anti-nucleocapsid (AU/mL) in patients and healthy donors (HD); the samples are sequential after dose 3 (D3) or D4. The IgG anti-nucleocapsid titer has an IQR of 7.5–22 and n = 12 for breakthrough infection (BTI) RA and an IQR of 10.7–22.5 and n = 54 for BTI HD. The P-values are indicated as *for p < 0.05, **for p < 0.01, and ****for p < 0.0001 (see also Supplementary Figure S1 for further analysis of serological responses and decay of specific IgG).

Figure 2

Humoral response in rituximab-treated rheumatoid arthritis (RA) patients after vaccination and infection. (A) Identification of Spike- and RBD-binding B cells. The representative dot plots show B cells that bind both wild-type (WT) Spike and WT RBD (upper region, Spike⁺ RBD⁺) or WT Spike but not WT RBD (lower region, Spike⁺ RBD^-). Examples are shown for two serological responder RA and two healthy donors (HD) 1 month after D2 and D3. (B) Quantification of Spike⁺ RBD⁺ vs. Spike⁺ RBD^–binding B cells. Left: frequency of Spike-binding B cells Spike⁺ RBD⁺ or Spike⁺ RBD^- 1 month after the vaccine in HD (D2-3) vs. rituximab-treated patients (D2-4). Right: relative distribution of Spike⁺ RBD⁺ vs. Spike⁺ RBD^- binding B cells 1 month after vaccination in HD (D2-D3) and RA patients (D2-D4). Mann–Whitney test; two-tailed P-values <0.01 (**) and <0.001 (***). (C) Phenotype of Spike-binding B cells. Frequency of Spike-binding and total B cells expressing CD27, CD71, Blimp-1, or IRF-4 1 month after D2 and D3. Wilcoxon matched-pairs signed-rank test, P-value with * for p < 0.05 and ** for p < 0.01. (D) WT or Omicron Spike- and RBD-binding B cells after breakthrough infection (BTI). Left panel: detection of cross-reactive CD19⁺ B cells that bind both Omicron (BA.1.1) Spike and WT Spike, upper region vs. only WT Spike (lower region, WT Spike⁺ BA.1.1 Spike^-). Right panel: detection of CD19⁺ B cells that bind both Omicron and WT RBD (WT RBD⁺ BA1.1⁺), Omicron RBD only (WT RBD^- BA1.1⁺), or WT RBD only (WT RBD⁺ BA1.1^-). (E) Quantification of WT or Omicron Spike- and RBD-binding B cells after BTI. Distribution of Spike and RBD-binding B cells as shown in (D). For HD and RA, left, IQR [0.091%–0.16%] for WT Spike, IQR [0%–0.22%] for cross-reactive WT/BA.1.1 Spike; IQR [0.17%–0.34%] for RBD WT only, IQR [0%–0.074%] for RBD WT/BA.1.1, and IQR [0.41%–1.2%] for RBD BA1.1-only. Percentage of nucleocapsid-binding B cells, right, with IQR [1.4%–1.7%]. Mann–Whitney test; P-value non-significant. (F) Phenotype of SARS-CoV-2-specific B cells after BTI. Spike- and RBD-binding B cells with specificities as in (D). For Omicron BA1.1-only, cross-reactive (WT/BA.1.1) and WT-only Spike/RBD, or nucleocapsid. The heat plots show differentiation markers as indicated, clustered into a function: antibody-secreting cells (ASC), memory, and naïve (see also Supplementary Figure S3 for an in-depth analysis of B cell phenotypes and specificities).

Figure 3

Cellular response in rituximab RA patients after vaccination and infection. (A) Functionality of SARS-CoV-2 specific T cells and serological response. Frequency of Spike-specific T cells after D2 in serological responder and non-responder RA. The frequency of response from unstimulated cells was subtracted as a non-specific background [see Supplementary Methods and (4, 21)]. Gated CD4 (left) and CD8 (right) T cell activation in response to Spike peptides in patients (after D2) vs. HD after D2 and D3. IQR [0.02%–0.20%] for Spike-specific CD4 T cells and IQR [0.09%–1%] for Spike-specific CD8 T cells in responder RA. Mann–Whitney two-tailed P-value with ** for p < 0.01 and *** for p < 0.001. (B) Functionality of SARS-CoV-2 specific T cells after D4-induced seroconversion. Frequency of Spike-specific T cells before (D3m1) and after (D4m1) serological response in RA. CD4 (left) and CD8 (right) T cell activation in response to Spike peptides in non-responder (NR) vs. responder (R) patients. IQR of Spike-specific CD4 T cell response for D3m1 [0.005%–0.095%] and IQR for D4m1 [0.0075%–0.15%]) for CD4 and IQR of Spike-specific CD8 T cell response for D3m1 [0.005%–0.095%] vs. D4m1 [0.005%–0.14%]. Wilcoxon matched-pairs signed-rank test; P-value non-significant. (C) Detection and quantification of Spike-specific CD8 T cells in RA; the dot plot shows pre-/post-vaccination examples. MHC-Class I restricted dextramers and tetramers were used to identify ex vivo Spike-, CMV-, EBV-, and FLU-specific CD8 T cells. Left flow plots: example of the patient before vaccination and 1 month after D2, CMV: HLA-binding T cells (upper left region) vs. Spike: HLA-binding T cells (lower right region). Right scatterplot: frequency of T cells binding peptide: HLA after D2 in patients vs. controls. IQR [0.115%–0.81%], n = 61 for RA and IQR [0.08%–0.95%], n = 52 for HD. Mann–Whitney test; *** indicates p < 0.001. (D) Functionality of SARS-CoV-2 specific helper T cells after breakthrough infection (BTI) in patients. Frequencies of Spike-and non-Spike (M, N, O)-specific CD4 T cells are shown: red symbols, patients with PCR-documented BTI; blue dots, vaccinated only. (E) Quantification of Spike- and non-Spike-specific CD8 T cells in vaccinated RA after BTI. Peptide: HLA class I dextramers and tetramers were used to identify ex vivo Spike-, non-Spike, CMV-, EBV-, and FLU-specific CD8 T cells. Mann–Whitney test, P-value non-significant (see also Supplementary Figures S4 , S5 ).