Impact of BNT162b2 mRNA anti-SARS-CoV-2 vaccine on interferon-alpha production by plasmacytoid dendritic cells and autoreactive T cells in patients with systemic lupus erythematosus: The COVALUS project

Abstract

Objective

To evaluate the specific response of SLE patients to BNT162b2 vaccination and its impact on autoimmunity defined as in vivo production of interferon-alpha (IFNα) by plasmacytoid dendritic cells (pDCs) and autoreactive immune responses.

Methods

Our prospective study included SLE patients and healthy volunteers (HV) who received 2 doses of BNT162b2 vaccine 4 weeks apart. Subjects under immunosuppressive drugs or with evidence of prior COVID-19 were excluded. IgG anti-Spike SARS-CoV-2 (anti-S) antibodies, anti-S specific-B cells, anti-S specific T cells, in vivo INF-α production by pDCs, activation marker expression by pDCs and autoreactive anti-nuclear T cells were quantified before first injection, before second injection, and 3 and 6 months after first injection.

Results

Vaccinated SLE patients produced significantly lower IgG antibodies and specific B cells against SARS–CoV-2 as compared to HV. In contrast, anti-S T cell response did not significantly differ between SLE patients and HV. Following vaccination, the surface expression of HLA-DR and CD86 and the in vivo production of IFNα by pDCs significantly increased in SLE patients. The boosted expression of HLA-DR on pDCs induced by BNT162b2 vaccine correlated with the overall immune responses against SARS-CoV-2 (anti-S antibodies: r = 0.27 [0.05–0.46], p = 0.02; anti-S B cells: r = 0.19 [-0.03-0.39], p = 0.09); anti-S T cells: r = 0.28 [0.05–0.47], p = 0.016). Eventually, anti-SARS-CoV-2 vaccination was associated with an overall decrease of autoreactive T cells (slope = - 0.00067, p = 0.015).

Conclusion

BNT162b2 vaccine induces a transient in vivo activation of pDCs in SLE that contributes to the immune responses against SARS-CoV-2. Unexpectedly BNT162b2 vaccine also dampens the pool of circulating autoreactive T cells, suggesting that vaccination may have a beneficial impact on SLE disease.

1. Introduction

Anti-SARS-CoV-2 vaccination is a major issue in patients with systemic lupus erythematosus (SLE) because of their comorbidities, immunosuppressive treatment and the severity of COVID-19 in this population [1,2]. International boards have however provided conflicting guidelines regarding mRNA vaccination in autoimmune disorders [3,4]. Vaccination coverage has been challenged by the common belief that vaccination could induce flare of the disease [5,6]. Patients with autoimmune diseases were excluded at first from SARS-CoV-2 vaccine clinical development programs. Although data from the large COVAX registry were reassuring [7], both patients and physicians still fear that vaccination could induce SLE flares by boosting innate immunity [8].

Plasmacytoid dendritic cells (pDCs) - the most potent producers of type I IFN (IFN-I) [13] - is a key cell involved in SLE pathogenesis [9,10]. Since type I interferons (IFN) play a role in both SLE pathophysiology and COVID-19 immune response [[8], [9], [10], [11], [12]], a specific risk associated with anti-SARS-CoV-2 vaccine could be postulated. BNT162b2 mRNA anti-SARS-CoV-2 vaccine may activate pDCs in SLE patients and increase the risk of lupus flares.

In this study, we tested the ability of BNT162b2 vaccine to induce interferon-alpha (IFNα) production by pDCs in vivo and analysed its putative impact on both autoreactive T cells and B cell-driven immune protection against SARS-CoV-2 in patients with SLE.

2. Methods

2.1. Population

We conducted a prospective observational study that included SLE patients who received a first dose of BNT162b2 mRNA anti-SARS-CoV-2 vaccine between March and May 2021 at our national center for autoimmune diseases [5]. To be included, patients had to give their written consent and fulfilled these criteria: being 18 years old or older, having a SLE defined by the 2019 EULAR/ACR classification criteria [12], having a non-clinically active disease, and being eligible to anti-SARS-CoV-2 vaccination according to French national guidelines [13]. Patients were secondarily excluded from the analysis if they were under immunosuppressive treatment (including cyclophosphamide, mycophenolate mofetil, azathioprine or rituximab) or had a prior history of COVID-19 defined by a positive SARS-CoV-2 PCR infection test or by a positive SARS-CoV-2 serology (anti-S or anti-N) before vaccination. As controls, 11 healthy volunteers (HV) among healthcare workers were included and received their first vaccine dose during the same time period.

2.2. Study scheme

All were vaccinated with BNT162b2 mRNA vaccine according to French national guidelines [13]. Briefly, SLE patient and HV received 2 doses of BNT162b2 vaccine 4 weeks apart and were screened at baseline (before first dose), at second dose, and at 3 and 6 months after the first dose. At inclusion and/or at follow-up visits, data on demographics, SLE history, treatment history, SLE disease activity (SLEDAI-2k), COVID-19 infection and/or side effects of the vaccine were recovered. Blood and urine samples were collected at each visit, in order to measure anti-dsDNA auto-antibodies titers, C3 and C4 levels, and anti-S and anti-N SARS-CoV-2 IgG antibodies titers, and proteinuria. Ficoll-Paque isolated peripheral blood mononuclear cells (PBMCs) were divided in two samples: one was freshly used for the pDC stimulation procedure and one was stored at – 80 °C for further analysis.

2.3. Ethical statement

This study involves human participants and was approved by Institutional Review Board (IRB) 00006477 of HUPNVS, Université Paris Cité, AP-HP, authorization number CER-2020-114. Participants gave informed consent to participate in the study before taking part.

2.4. Protective immune response to BNT162b2 mRNA vaccine

Vaccine efficacy was evaluated through different methods. First, we looked at the B-cell response with a standard quantitative IgG anti-Spike SARS-CoV-2 (anti-S) serology (Alinity-Abott). According to French guidelines [14], we chose a cut-off value of 260 BAU/mL in anti-S IgG serum level to define “responders” and “non-responders”. We also quantified blood anti-S-specific-B cells induced by the vaccine with tetramers made of biotinylated-Spike protein (Recombinant SARS-CoV-2 Spike-Prot (HEK)-Biotin, Miltenyi-Biotec) and fluorescent streptavidin (PE and PE-Cy7 streptavidin, BioLegend) in a flow cytometry assay (according to manufacturer instructions [15], Fig. S1). Then, we looked at the T-cell response using a 6-h stimulation with a pool of peptides covering the full sequence of the spike protein. T cell vaccine response was assessed by using the percentages of CD154⁺CD4⁺ T cells among IFN-γ-secreting cells [21] after stimulation with a pool of peptides covering the whole SARS–CoV-2 wild-type spike protein (PepTivator® SARS-CoV-2 Prot_S, Miltenyi Biotec). Results were normalized by subtracting at each time point, and for each patient, the percentage of responses observed at time 0.

2.5. Interferon-α2b production by plasmacytoid dendritic cells

Freshly isolated PBMCs were counted and resuspended at 10⁶/mL in complete medium (RPMI 1640 GlutaMax [Gibco] supplemented with 10% fetal calf serum, 10 mM HEPES buffer, 1% sodium pyruvate, 1% minimum essential medium non-essential amino acids (all from ThermoFisher scientific), and 37.5 μM of β-mercaptoethanol (Sigma-Aldrich). One million cells (1 mL) were then added to a 48-well plate supplemented with 2 μg/mL brefeldin A (BFA) in a 37 °C incubator with 5% CO2 for 6 h. At the end of the incubation, cells were harvested, washed with phosphate buffered saline (PBS), and stained with a viability marker (Ghost Dye BV510; Tonbo biosciences) for 20 min at 4 °C. After washing with FACS buffer (PBS, 1% bovine serum albumin, 1 mM EDTA, and 0.05% sodium azide), unspecific antibody-binding sites were saturated with an Ig blocking buffer containing 100 μg/mL of polyclonal human IgG, polyclonal rat IgG and polyclonal mouse IgG (Innovative Research, Inc., Novi, MI, USA). Then, PBMCs were stained with a panel off fluorophore-conjugated antibodies (Ab) targeting surface markers: CD303-Alexa Fluor (AF) 488, CD141-PerCP-Cy5.5, CRTH2-AF647, CD1c-AF700, FcεR1α-APC/Fire750, CCR3-Brilliant Violet (BV) 605, HLA-DR-BV785, CD123-PE-Dz594 (all from BioLegend) and CD3/CD19/CD56/CD14 Brilliant Ultra Violet (BUV) 395 (T, B, NK, classical monocyte lineages, all from BD Biosciences) in FACS buffer for 20 min. After a PBS wash, cells were fixed for 20 min using a fixation buffer according to the guidelines of the manufacturer (BioLegend). PBMCs were then permeabilized and stained with IFNα2b-PE (from BD Biosciences) in Permeabilization and wash (Perm/wash) buffer (Biolegend) supplemented with 20 μl of Ig blocking buffer for 20 min at 4 °C. Finally, PBMCs were washed in Perm/wash buffer and resuspended in FACS buffer until acquisition. We also monitored the activation markers CD86 and HLA-DR expression levels on pDCs with a flow cytometry assay performed on whole blood cells, after ACK lysis. For assessment of the CD86 and HLA-DR expression levels, the ratio of the geometric mean fluorescence intensity (gMFI) of CD86 and HLA-DR to the corresponding fluorophore-conjugated isotype control gMFI was calculated and expressed in arbitrary units (A.U.). Flow cytometry data acquisition was realized using a Becton Dickinson 5-lasers LSR Fortessa X-20 and data analysis using Flowjo vX (BD Biosciences, Franklin Lakes, NJ, USA). Plasmacytoid dendritic cells (pDCs) were defined among living cells as (CD3/CD19/CD56/CD14)⁻CD123⁺FcεR1a^loCD303⁺HLA-DR⁺ cells (gating strategy presented in ESM, Fig. S2).

2.6. Autoreactive T cells

Quantification of specific autoreactive T cells was realized by defining the proportion (%) of activated (double positive CD154⁺/CD69⁺) cells among non-naïve CD45⁻RA^-CD4⁺ T cells after stimulation with a pool of nuclear antigens [16] containing an equimolar concentration of Sm/RNP, SS-A, SS-B and histones (all from Immunovision). Briefly, all the PBMCs collected at the different follow-uptimes for a given individual were thawed and put in complete medium. They were then stimulated in a 96-well U-bottom plate at 2 × 10⁶ per mL with either 1) only a co-stimulating pool containing anti-CD-40 Ab (1 μg/mL, clone HB14, BioLegend) and anti-CD28 Ab (1 μg/mL, clone CD28.2, BioXCell-InVivoMab) in complete medium (non-stimulated condition) or 2) the costimulating pool and SARS-CoV-2 Prot_S PepTivator® (0.6 nmol/mL) or 3) the costimulating pool and the nuclear antigens pool (5 μg/mL) for 6 h at 37 °C in a 5% CO2 incubator. BFA (2 μg/mL) and Monensin (10 μmol/L) were added 2 h after the beginning of the incubation. At the end of the incubation, cells were harvested, washed with phosphate buffered saline (PBS), and stained with a viability marker (Ghost Dye 510) for 20 min at 4 °C. After washing with FACS buffer, unspecific antibody-binding sites were saturated with Ig blocking buffer. Then, PBMCs were stained with a panel of fluoropohore-conjugated Ab targeting surface markers: CD3-AF488, CD69-PerCP-Cy5.5, CD8a-BV605, CD4-BV785 and PE/Cy7-CD45-RA (all from BioLegend) in FACS buffer for 20 min. After a PBS wash, cells were fixed for 20 min using a fixation buffer according to manufacturer's instructions (BioLegend). PBMCs were then permeabilized and unspecific antibody-binding sites were saturated with blocking solution in Perm/Wash buffer and cells were stained with CD154-APC/Fire750 (from BioLegend) in Perm/Wash buffer for 20 min at 4 °C. Finally, PBMCs were washed in Perm/Wash buffer and resuspended in FACS buffer until acquisition. For each analysis the percentage of the non-stimulated condition was subtracted from the one of the stimulated conditions to take into account the background signal.

2.7. Statistical analyses

Data are expressed as median with inter-quartile range (IQR) for quantitative variables or frequency (percentage) for categorical variables. Comparison of quantitative variables between two different time points were made using a paired Wilcoxon test. To establish the significance of the trend observed during follow-up for the autoimmune activity, we used a linear mixed effect model (package lme4) defining time as both a random and a fixed effect (more details in ESM). Two-sided P values of <0.05 were considered to indicate statistical significance. Statistical analyses were performed with R version 4.1.0 and Graphpad v9 softwares.

3. Results

3.1. Population

Between March and May 2021, 57 SLE patients and 11 healthy volunteers (HV) vaccinated with BNT162b2 mRNA vaccine were included in the study. SLE patients were mostly female (86.0%) with a median age at inclusion of 44.0 [38.1, 50.8] years. All patients were under hydroxychloroquine (n = 51, 89.5%) and/or corticosteroids (n = 35, 61.4%) at a median dose of 7 [[5], [6], [7], [8], [9]] mg/day. Among them, 10 (17.5%) patients were excluded from the analysis because they were receiving azathioprine, mycophenolate mofetil or rituximab. Eleven (21.0%) patients were also excluded because of a previous history of COVID-19 infection - assessed by the detection in serum of anti-N SARS-CoV-2 antibodies at screening. Eventually, 36 SLE patients with SLEDAI-2k at 0 [0–2] at inclusion, who were not receiving immunosuppressive drugs, had no previous history of COVID-19 and had received 2 doses of BNT162b2 vaccine 4 weeks apart were considered for the analysis (Table 1 ). The immune response following vaccine were analysed at baseline before the first dose (T0), at second dose (M1), and 3 (M3) and 6 (M6) months after the first dose.

Table 1.

Characteristics of SLE patients at baseline.

	SLE patients n = 36
Male gender, n (%)	5 (13.9)
Age, in years (median [IQR])	43.7 [36.3, 49.7]
Duration of SLE disease, years (median [IQR])	9.3 [4.1, 13.2]
SLE characteristics, n (%)
Class III/IV nephritis	9 (25.0)
Skin lesions,	22 (61.1)
Arthritis	26 (72.2)
Pleural/pericardial effusion	10 (27.8)
Immune cytopenia	12 (33.3)
CNS involvement	4 (11.1)
Anti dsDNA at baseline
<1 N	20 (55.6)
1–3 N	9 (25.0)
>3 N	7 (19.4)
Antiphospholipid biology	13 (36.1)
Associated disease, n (%)
Antiphospholipid syndrome	6 (16.7)
Sjögren syndrome	6 (16.7)
Anti-IFNα Ab detected at baseline (%)	6 (16.7)
SLEDAI-2k at baseline (median [IQR])	0 [0–2]
Treatment at baseline, n (%)
Corticosteroids	17 (47.2)
daily dose, mg/d (median [IQR])	7.0 [5.0, 9.0]
Hydroxychloroquine	31 (86.1)
[HCQ] <500 μg/L	10 (27.8)
[HCQ] 500–1500 μg/L	21 (58.3)
[HCQ] >1500 μg/L	4 (11.1)
Belimumab	1 (2.8)

CNS, central nervous system; [HCQ], hydroxychloroquine concentration measured in whole blood.

3.2. Humoral response to SARS-CoV-2 vaccine in SLE patients

SLE patients and HV had no detectable anti-S and anti-N at baseline. At M1, M3 and M6, 9.1% (n = 3/33), 63.3% (n = 19/30), and 8.0% (n = 2/25) of SLE patients had specific anti-S titers >260 BAU/ml, respectively. One month after the last vaccine shot (M3), the percentage of SLE patients able to mount a humoral response against SARS-CoV-2 were lower as compared to HV (63.3% vs 100%, p = 0.038). Accordingly, anti-S antibodies titers were also lower in SLE patients as compared to HV at M1 (37.1 [8.9–88.9] BAU/mL in SLE vs 120.4 [64.3–251.9] BAU/mL in HV, p = 0.005), M3 (350 [130.7–769.4] vs 877.3 [531.7–1334.2], p < 0.001) and M6 (59.1 [28.9–124.4] vs 219.8 [140.4–521.6], p < 0.0001) (Fig. 1 A). Those data confirmed the poor immune protection conferred by 2 doses of BNT162b2 vaccines in SLE patients, even though they were not treated with immunosuppressive drugs [17],

Fig. 1.

SARS-CoV-2-specific immune responses following vaccination

Anti-Spike IgG blood titer measured in BAU/mL by ELISA in SLE patients (in blue) compared to HV (in red). Points represents the mean and error bars the standard error of the mean (A). Blood concentrations of anti-S specific B cell measured by flow cytometry using tetramers of biotinylated spike-protein associated to fluorescent streptavidin. Points represented the mean and error bars the standard error of the mean (B). Correlation between anti-spike IgG blood level and anti-S specific B cells in SLE and HV at all the time points. Pearson's coefficient is provided (C). T cell response to SARS-CoV-2 vaccine measured by flow cytometry as the percentage of CD154+/IFNγ+ among CD4 cells after stimulation with a pool of peptides covering the sequence of Spike protein. For each time Ti, we represented here the delta Ti- T0 to take into account background and basal signal (D). Points represented the mean and error bars the standard error of the mean* p < 0.05 **p < 0.001 ***p < 0.001.

3.3. Cellular immune response to SARS-CoV-2 vaccine in SLE patients

We next analysed the cellular immune response induced by BNT162b2 vaccine. Among SLE patients, the absolute count of circulating anti-Spike specific B cells measured by flow cytometry increased overtime and appeared maximal 6 months after the first vaccine shot (5 [[4], [5], [6], [7], [8], [9], [10]].10^–5/mL anti-S specific B cells at T0 vs 16 [6–29] at M6, p = 0.009). Consistent with the poor humoral response, the absolute count of anti-S specific B cells was significantly lower in SLE patients as compared to HV at M1 (5 [[4], [5], [6], [7], [8], [9], [10]].10^–5/mL anti-S specific B cells in SLE vs 14 [[11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23]].10–5/mL in HV; p = 0.004), M3 (11 [[6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19]].10^–5/mL vs 30 [18–53].10^–5/mL, p = 0.008) and M6 (16 [6–29].10^–5/mL vs 27 [25–82].10^–5/mL, p = 0.02) (Fig. 1B). Interestingly, anti-S antibodies titers correlated with anti-S specific B cell counts in blood (Fig. 1C) suggesting that the humoral response reflects the overall B cell response to SARS-CoV-2 vaccine in SLE.

The anti-SARS-CoV-2 T cell response to vaccine was then assessed by measuring the frequency of CD154⁺CD4⁺T cells producing IFN-γ after in vitro stimulation with a pool of peptides covering the full sequence of the SARS-CoV-2 spike protein. Anti-SARS-CoV-2 CD154⁺IFN-γ⁺CD4⁺ T increased overtime after vaccination in SLE patients. However, in contrast to the B cell responses to vaccine, anti-SARS-CoV-2 T cell response did not significantly differ between SLE patients and HV (Fig. 1D). Such findings confirmed that despite evidence for impaired humoral response, SARS-CoV-2 vaccine still induces a robust T-cell response in patients with autoimmune diseases [18,19].

3.4. Impact of anti-SARS-CoV-2 vaccination on IFN-alpha production by plasmacytoid dendritic cells and specific autoreactive T cells

We next analysed the impact of BNT162b2 mRNA anti-SARS-CoV-2 vaccine on pDCs in SLE patients. At T0, the ability for unstimulated pDCs to spontaneously produce IFNα in vivo was stronger in SLE patients as compared to HV (0.64 [0.27–1.09] % of pDCs⁺IFNα⁺ in the SLE versus 0.32 [0.25–0.47] % of pDCs⁺IFNα⁺ in the HV, p = 0.04) (Fig. 2 A). Following anti-SARS-CoV-2 vaccine, the frequency of pDCs⁺IFNα⁺ raised at M1 in both groups (1.27% [0.6–2.6] in SLE, p < 0.001; 1.5% [0.95–2.12] in HV, p = 0.03) but remained high at M3 in SLE patient only (1.25% [0.85–1.83], p = 0.004) to next return to baseline at M6. Consistent with the activation of pDCs induced by anti-SARS-CoV-2 vaccine, the amount of both CD86 (2.9 [2.6–3.4] MFI ratio versus 2.3 [1.8–30] at T0, p = 0.002) and HLA-DR (41.9 [34.2–61.3] versus 24.4 [16.4–32.8] at T0, p < 0.001) markers were increased on pDCs at M3 in SLE (Fig. 2 B, C). Interestingly, anti-S antibodies titers, (r = 0.27 [0.05–0.46], p = 0.02), anti-S specific B cells (r = 0.19 [−0.03-0.39], p = 0.09) and anti-S specific T cells (r = 0.28 [0.05–0.47], p = 0.016) correlated with the enhanced expression of HLA-DR on pDCs boosted by BNT162b2 mRNA vaccine at M1 and M3 (Fig. 2 D, E, F).

Fig. 2.

Impact of BNT162b2 vaccine on ex vivo IFNα production by pDCs and pDCs activation.

Percentage of IFNα2b production by circulating pDCs during follow-up. In blue, SLE patients, in red HV (A). Activation markers on circulating pDCs during study follow-up measured by mean fluorescence intensity ratio of HLA-DR (B) and CD86 (C) in flow cytometry. Correlation between pDCs activation measured by HLA-DR gMFI on pDCs at M1 and M3 in SLE and HV and anti-S IgG blood level (D), anti-S specific B cells (E) and anti-S specific T cells (F) *p < 0.05 **p < 0.001 ***p < 0.001.

We next sought to determine whether BNT162b2 mRNA vaccine had an impact on the pool of autoreactive T cells in SLE. Specific autoreactive T cells - defined as the percentages of CD154⁺ CD69⁺ CD45⁻ RA⁻ CD4⁺ T cells identified after ex vivo stimulation with a pool of nuclear antigens including Sm/RNP, SS-A, SS-B and histones were measured at T0, M1, M3 and M6. As expected, the amount of autoreactive T cells at T0 were higher in SLE patients than in HV in whom those cells were barely detected (0.125% [0.042–0.38] vs 0.05 [0–0.26], p = 0.03). In line with the activation of pDCs 4 weeks after the first vaccine shot, we observed a 1.6-fold increase – but not significant - of specific autoreactive T cells at M1 as compared to T0 (0.200% [0.07–0.41] vs 0.125% [0.042–0.38], p = 0.50) that eventually dampened from M1 to M6. Overall, anti-SARS-CoV-2 vaccination was associated with a decrease of specific autoreactive T cells in SLE overtime (β for fixed-effect of time in the linear mixed-effect model = - 0.00067, p = 0.015) (Fig. 3 A). We found no association between the activation of pDCs and anti-dsDNA titers, serum C3/C4 levels and autoreactive T cell counts (Fig. S3).

Fig. 3.

Autoreactive-specific immune responses following vaccination

Anti-dsDNA IgG blood levels among SLE patients during follow-up. Box plot represents the median and the interquartile range of patients with detectable anti-dsDNA Ab at baseline [left-panel]. Points and lines represent the individual levels of each included patients at each visit [right-panel] (A). Complement fraction C3 and C4 blood level among SLE patients during follow-up. Points represents the mean and error bars the standard deviation (B). SLE disease activity measured by SLEDAI-2k during follow-up (C). Autoreactive anti-nuclear specific T cell activity among SLE patients and HV during follow-up defined as the percentages of activated (double positive CD154+/CD69+) cells among non-naïve CD4 T cells after stimulation with a pool of nuclear antigens. Time-associated slope is significantly negative in a linear mixed-effect model in the SLE group: (β for fixed-effect of time in the linear mixed-effect model = -0.00067, p=0.015). Points represented the mean and error bars the standard error of the mean (D).

3.5. Clinical outcome

During the 6-months study time, one SLE patient (2.8%) experienced PCR-confirmed COVID-19 that did not require admission to hospital. Two (5.5%) SLE patients experienced lupus flare: one had a relapse of immune thrombocytopenia between T0 and M1 and one had a relapse of a class III lupus nephritis between M3 and M6. In both cases, SLE flares occurred after corticosteroids were tapered by the clinician. Except for these 2 patients, we did not observe any modification of SLEDAI-2k scores, serum levels of anti-dsDNA IgG and serum levels of complement during follow-up (Fig. 3B, C, D).

4. Discussion

In the present study, we first confirmed the results of previous studies showing a B cell defect that impairs immune protection following anti-SARS-CoV-2 vaccine in SLE patients [17]. Second, we showed that BNT162b2 mRNA induces a short-term in vivo activation of the pDCs that contributes to the immune response against SARS-CoV-2 vaccine. Third, and most strikingly, BNT162b2 mRNA appears to dampen the pool of circulating autoreactive T cells, suggesting that vaccine may actually protect SLE patients from lupus flare.

In order to avoid treatment-related bias, we selected SLE patients who were not receiving immunosuppressive drugs. The observed B cell defect preventing a strong immune response to BNT162b2 mRNA could not be ascribed to drug-induced immunosuppression. Such poor immune response to vaccine is probably related to SLE – as our group previously reported in the setting of anti-pneumococcal vaccination [[20], [21], [22]]. Because SLE entails a risk for severe COVID-19 [1,2], SLE patients are obvious candidates for alternative vaccination strategies [23].

Besides concern for poor immune responses to vaccine, vaccination also appears to increase disease activity. RNA vaccines against SARS-CoV-2 may trigger Toll-like receptors-9 and -7 constitutively expressed on pDCs to induce further production of type I IFN, known to play a key role in SLE pathogenesis [10]. Indeed in our study, RNA vaccines activated pDCs in vivo which contributed to the overall protective response against SARS-CoV-2, in line with a previous study in mice [24]. Unexpectedly, specific auto-reactive anti-nuclear T cells blood levels decreased following BNT162b2 mRNA, leaving the possibility that the vaccine exert an anti-inflammatory and protective effect on lupus disease. Concordant with this intriguing result, a previous study has shown that vaccination (i.e. anti-pneumococcal vaccination) improves lupus disease in MRL/lpr mice probably by modifying the T follicular helper cells/T follicular regulatory cells balance in secondary lymphoid organs [25]. Overall, there was no change in SLEDAI score pre- and post-vaccination in all but 2 patients, supporting the safety of the vaccination in SLE patients [26]. These data should hopefully reduce vaccine hesitancy and encourage both patients and physicians to support vaccination.

Our study has several limitations. The findings were based on a small number of patients. There was vaccine hesitancy among SLE patients in our cohort [5] and the patients in this study may not be fully representative. Last, our study was not designed to determine the in-depth immune process that mechanistically linked the vaccine to the decrease of specific auto-reactive T cells.

This study has several strengths. First, our design selected homogeneous SLE patients who were not receiving immunosuppressive drugs, had no disease high activity at vaccination and no prior SARS-CoV-2 infection. Second, the study assessed BNT162b2 mRNA vaccine’ effects on lupus-specific disease activity pre- and post-vaccination by using several markers including diseases score activity, anti-dsDNA, complement level but also pDCs activation in vivo and autoreactive T cells quantification ex vivo. Third, the protective immune response to vaccine was assessed by analysing not only anti-SARS–CoV-2 antibodies but also SARS-CoV-2 specific B and T cells.

In conclusion, BNT162b2 mRNA vaccine induces IFN-alpha production by pDCs that contributes to the immune protective response against SARS-CoV-2 and appears to have an unexpected beneficial protective impact on autoimmunity in SLE.

Funding

Arthur Mageau was supported by a PhD fellowship provided by Fondation pour la Recherche Médicale (FDM202106013488). This work was supported by the Agence Nationale de la Recherche (n°ANR-21-COVR-0034 COVALUS), the Agence Nationale de la Recherche sur le SIDA et les Maladies Infectieuses Emergentes (ANRS MIE), AC43 Medical Virology and Emergen Program, the University Paris Cité and the Assistance Publique Hôpitaux de Paris.

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Patient consent for publication

Obtained.

Author contributions

KS and NC directed the project. AM, TG, NC and KS designed the study. AM, JT, PNR, and VMF conducted analysis. DD, ND, CF, CM and TP were involved in the project development. AM and KS wrote the manuscript. All authors reviewed and approved of the final manuscript.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: