Multiple effects of TNFα inhibitors on the development of the adaptive immune response after SARS-CoV-2 vaccination

Patient recruitment and biosampling

The study was approved by the ethics committee of the Christian-Albrecht University Kiel (D409/21). There was no patient and public involvement in conducting the study. Recruitment of patients and repetitive biosampling was performed as previously described.(2) The SAVE-CID cohort consists of 47 healthcare workers and other risk groups who received their first SARS-CoV-2 vaccination in January 2021 followed by a second vaccination 5 or 3 weeks later. Samples taken 7 days after the third vaccination were also examined in 12 patients. All patients received BNT162b2 (Comirnaty, Pfizer/BioNTech) or mRNA-1273 (Spikevax, Moderna). Biosampling and data acquisition as well as data on antibody concentration, surrogate neutralisation data and clinical characterisation of this cohort has already been published.(2, 3) The patient groups were age and gender matched resulting in mean ages of 43 (TNFα inhibitor, median: 42.5), 41.25 (Healthy Control, median: 39) and 41.63 (oDMARDs, median: 46).

Production of SARS-CoV-2 S1 proteins

S1 domain of the S protein(GenBank: MN908947) with different tags were baculovirus-free produced in High Five cells (Thermo Fisher Scientific) by transient transfection as previously described.(15, 16) Protein purification was performed depending 1 or 5 mL column on Äkta go (Cytiva), Äkta Pure (Cytiva), or Profina System (BIO-RAD). HiTrap Fibro PrismA (Cytiva) was used as resins for Protein A purification (Fc-tagged proteins). For His-tag purification of insect cell supernatant HisTrapexcel column (Cytiva) was used. All purifications were performed according to the manufacturer’s manual. S1-HIS was further purified by SEC by a 16/600 Superdex 200 kDa pg (Cytiva).

Isolation of peripheral blood mononuclear cells (PBMC) and Flow Cytometry

PBMCs from EDTA blood were isolated within three hours of blood collection by density gradient centrifugation (Biocoll, Bio&SELL GmbH, Feucht, Germany). Afterwards, 4×10⁶ PBMCs were incubated with his-tagged S1 protein (own protein or Euroimmun, Lübeck, Germany).(9) PBMCs were then stained with pre-mixed antibodies (CD19-PerCP-Vio-700 (REA657, Miltenyi Biotec, Bergisch-Gladbach, Germany), CD20-PE-Cio770 (REA780, Miltenyi Biotec), CD3-Pacific Blue (OKT3, Biolegend, San Diego, USA), CD14 Pacific Blue (M5E2, Biolegend), CD27-APC (M-T271, Biolegend), anti-HIS-PE (JO95-G46, Biolegend), Biolegend), HLA-DR-VioGreen (REA805, Miltenyi Biotec), CD138-BV605 MI15, Biolegend)) and analysed using a MacsQuant 16 Cytometer (Miltenyi Biotec, Bergisch-Gladbach, Germany). Secondary staining using AF-700 coupled S1-fc protein were used as gating and staining control to exclude false positive events (see figure S1 and S2 for more information). For the calculation of immune cells per blood volume, 50µL of whole blood was stained, (CD3-Pacific Blue (Biolegend), CD14-FITC (REA599, Miltenyi Biotec), CD4-PE (Vit4, Miltenyi Biotec), CD19-PerCP-Vio700 and CD45-APC-Vio770 (H130, Biolegend)), lysed, (Red blood lysis, BD) and measured on a MACSQuant 16 (Miltenyi Biotec). The cell counts per 50µL blood for each sample were used to calculate all other cell counts from the PBMC staining. Staining and measurement were performed in PBS containing 1% BSA, 0.5% EDTA and 0.1% sodium azide.

Measurement of Antibody Secreting Cells by Three-Colour Fluorospot

PBMCs were isolated as described above and different T cell numbers were incubated for three hours in a 96-well 0.45µM PVDF Immobilon-FL membrane plate (Merck Millipore, Burlington, MS, USA) which was pre-coated with SARS-CoV-2 protein or FAB2-fragments against Immunoglobulins A, M (both Southern Biotech, Birmingham, USA), G (Jackson Immunoresearch, Cambridge, UK) and blocked with 15%. FCS in PBS. After 3 washing steps with deionised water, the wells were stained using IgA-FITC, IgG-FC-AF555 and IgM-AF647 (all Southern Biotech). Measurement was performed on a Bioreader 6000Fb equipped with Eazyreader software (Bio-SYS, Karben, Germany)

Antigen-reactive T cell enrichment (ARTE)

The ARTE was performed as previously described (17-20). In brief, 0.5-1×10⁷ PBMCs were stimulated for 7 hours with 0.5 µg/peptide/ml SARS-CoV-2 S peptide pool (JPT, Berlin, Germany) in presence of 1 µg/ml CD40 and 1 µg/ml CD28 pure antibody (both Miltenyi Biotec, Bergisch Gladbach, Germany). 1 µg/ml Brefeldin A (Sigma Aldrich) was added for the last two hours. Cells were magnetically isolated using the CD154 MicroBead Kit (Miltenyi Biotec). After surface staining with CD4-APC-Vio770 (M-T466), CD8-VioGreen (REA734), CD14-VioGreen (REA599), CD20-VioGreen (LT20) (all Miltenyi Biotec), CD45RA-PE-Cy5 (HI100), PD-1 Brilliant Violet 605 (EH12.2H7), CCR7-Brilliant-Violet-785 (G043H7) (all BioLegend), cells were fixed, permeabilized and stained intracellular with CD154-FITC (REA238), IL-21-PE (REA1039) (both Miltenyi Biotec), IFN-γ-PerCP-Cy5.5 (4S.B3), TNFα-Brilliant-Violet-650 (MAb11), IL-10-PE-Dazzle (JES3-9D7) (all BioLegend), IL-2-BV711 (5344.111), Ki-67-Alexa Fluor 700 (B56) (both BD Biosciences). Viobility 405/520 Fixable Dye (Miltenyi Biotec) was used to exclude dead cells. Data were acquired on a LSR Fortessa (BD Bioscience, San Jose, CA, USA). CD154⁺ background cells enriched from the non-stimulated control were subtracted and frequencies of antigen-specific T cells were determined based on CD154⁺ T cells after enrichment, normalised to the total number of CD4⁺ T cells applied on the column.

Binding strength (avidity) of SARS-CoV-2 IgG antibodies

The IgG avidity was assessed with the recomLine SARS-CoV-2 IgG assay on a Dynablot Plus system together with a BLOTrix reader and the recomScan software (all from Mikrogen GmbH, Neuried, Germany) as reported previously.(21) This immunoblot consists of a nitrocellulose strip separately carrying recombinant nucleocapsid protein (NP) and the S1-and RBD-subunits of the S protein. Binding of IgG to these SARS-CoV-2 antigens in presence or absence of avidity reagent was automatically measured and assigned to four categories: no avidity detectable (=0), low avidity (=1), intermediate avidity (=2) and high (=3) avidity.(22)

Measurement of neutralising antibodies against a pre-VOC strain and a BA.2 strain

Sera were tested in triplicate using a Vero cell-based live virus neutralisation test (cVNT) in 96-well format under biosafety level 3 conditions, as previously reported.(10, 22) In brief, sera were diluted 1:10 to 1:1280 in cell culture medium free of fetal calf serum. As antigens for the cVNT, we used either 50 plaque-forming units per well of a B.1 strain (pre-VOC of 2020) or an Omicron BA.2 strain, which we had previously isolated (23) and characterised by whole-genome sequencing.(10, 23) After four (pre-VOC) or six (BA.2) days of incubation, cells were fixed by addition of paraformaldehyde and stained with an aqueous crystal violet methanol solution. Serum dilutions (titres) > 1:10 that prevented the formation of a cytopathic effect in ≥2 wells were considered to contain neutralising antibodies (NA); if no exact titre could be given, the geometric mean of the two adjacent titres was calculated.(10, 22)

Data analysis and statistics

Flow cytometry data was analysed using FlowJo v10 (BD Bioscience). Statistical analyses and graphs were prepared using RStudio (version 2022.02.0+443) and Prism 8 (GraphPad Software, LLC). For analysis of differences between the groups, Kruskal-Wallis Test and non-parametric pairwise comparisons were performed.

Anti-S-IgG-antibody concentrations, IgG avidity and neutralisation efficacy decrease 6 months after vaccination

Under TNFα inhibitor therapy, anti-S-IgG levels after the second immunisation were significantly lower than in patients receiving oDMARDs or in healthy controls (figure 1a). None of the subjects showed anti-NP IgG reactivity, making infection breakthrough unlikely (data not shown). The IgG avidity and neutralisation capacity against the pre-VOC strain were high in all tested groups 14 days after second vaccination (avidity: median avidity index (MAI) = 3; NA: geometric mean titres (GMT)=1:98–1:234)(figure 1B, C). Six months after second vaccination, IgG avidity and pre-VOC NA titres significantly decreased in TNFα-inhibitor treated vaccinees (n=8; avidity: MAI=1.25; NA: GMT=1:2) compared to patients receiving oDMARDs (n=7; avidity: MAI=3, p=0.0012; NA: GMT=1:38, p=0.0293) and healthy controls (n=12; avidity: MAI=3, p ≤0.0001; NA: GMT=1:25, p=0.0347) (figure 1A). Relative to the pre-VOC strain, anti-BA.2 NA titres were significantly lower than against the pre-VOC strain in all three groups at day 14 after the second vaccination (GMT=1:2-1:4 vs. GMT=1:98-1:234; p=0.0001-0.0072) and were not detected after six months (GMT<1:10, figure 1D). At 7 days after the third vaccination, anti-BA.2 NAs were detectable in all subjects except patients taking a TNFα blocker (n=4 per group; GMT=1:62-1:95 vs. GMT=1:3). At this time point, IgG avidity was high in all subjects except one patient receiving anti-TNFα treatment.

• Download figure
• Open in new tab

Figure 1:

A Patients using TNFα blockers show reduced antibody serum levels, avidiy and neutralisation at different timepoints. Serum IgG antibody against SARS-CoV-2 Spike 1 subunit at different timepoints as single values per patient. Lines indicate the median. B IgG avidity indices of anti-S IgGs of SARS-CoV-2. C Neutralising antibodies against the wild-type (wt) and D BA.2 variant of SARS-CoV-2 at 14 days and six months after the second vaccination. For BA.2 and IgG avidity, a time point 7 days after the third vaccination was added (n=4 per group); values are given as individual values and median. Statistical differences: Kruskal-Wallis test with Dunn’s post hoc test, significant differences are indicated as *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001

Plasma cell populations are altered in patients using TNFα blockers

Patients under TNFα inhibiting therapy showed higher numbers of plasmablasts in the peripheral blood 7 days after the second vaccination (median: 9.153cells/µL) compared to patients receiving oDMARDs (median: 2.205cells/µL, p=0.0254) (figure 2A) and healthy controls (2.657 cells/µL, p=0.0344).

• Download figure
• Open in new tab

Figure 2:

Patients using TNFα blockers show higher numbers of overall plasmablasts and SARS-CoV-2 specific plasmablasts 7 days after the second vaccination. A Flow cytometry analysis of B cell subsets at different timepoints after a vaccination against SARS-CoV-2. Values are shown as cells per microlitre of blood. Boxes and whiskers indicate median and 95%CI. B SARS-CoV-2 specific plasmablasts. IgA⁺ plasmablasts C and SARS-CoV-2 specific D IgA⁺ positive plasmablasts. E and F: CD138⁺ unspecific and SARS-CoV-2 specific plasmablasts which resemble mature plasma cells. If not indicated otherwise, cell counts per microlitre blood are shown as single values with median. Statistical differences: Kruskal-Wallis test with Dunn’s post hoc test, significant differences are indicated as *P<0.05, **P<0.01, ****P<0.0001

When comparing SARS-CoV-2 specific plasma cells at the same timepoint, no differences between patients on TNFα blockade (median 0.295 cells/µL) and healthy controls (median 0.333 cells/µL) were detected, while patients treated with oDMARDs other than TNFα inhibitors had significantly lower cell numbers compared to healthy controls (median: 0.204 cells/µL; p=0.0015) (figure 2B).

Anti-TNFα treated patients generally displayed more peripheral blood IgA plasma cells than controls and patients receiving oDMARDs (figure 2C). This trend became significant on day 14 after vaccination for patients under oDMARDs (median 0.404 cells/µL) and TNFα blockers (median 1.818 cells/µL; p=0.0397).

With regards to SARS-CoV-2-specific IgA-plasma cells, counts were lower in the anti-TNFα group on day 7 after the second vaccination (median: 0.016 cells/µL) compared to healthy controls (median: 0.052 cells/µL; p=0.0203) and patients treated with oDMARDs (median:0.035 cells/µL; p=0.0299) (figure 1D).

Overall numbers of mature circulating CD138⁺ plasma cells were higher in patients on TNFα blockade (median: 0.101 cells/µL) than in healthy controls (median: 0.031 cells/µL; p=0.0055) and patients treated with oDMARDs (median: 0.019 cells/µL; p=0.0015) on day 14 after second vaccination (figure 2E). Differences at time points between the first and the second vaccination were not significant. We were not able to detect differences in the number of SARS-CoV-2 specific plasma cells. However, on day 7 after second vaccination, healthy controls had higher numbers of circulating CD138⁺ plasma cells in the peripheral blood (median: 0.004 cells/µL) than patients using TNFα inhibitors (median: 0.002 cells/µL; p=0.0055) or receiving oDMARDs (median: 0.001 cells/µL; p=0.0026) (figure 2F).

Anti-TNFα treatment does not change the number of SARS-CoV-2 specific antibody secreting cells

Relative to the other two sample groups, anti-TNFα patients showed a pronounced increase in SARS-CoV-2 specific antibody secreting cells (ASCs) of the IgM isotype at day 7 post second vaccination, while total numbers of ASC of all other isotypes remained comparable between groups (figure 3A). Analysing spot size in our Fluorospot assays as a surrogate for the amount of secreted antibody per cell, both patient groups generally displayed larger spot sizes suggesting increased antibody secretion per ASC (figure 3B). These differences were not significant except for IgA (p=0.0026 for TNF vs. Healthy Control and p=0.001 for oDMARDs vs. Healthy Control). No differences were detected between the two patient groups. The number of ASCs correlated well with SARS-CoV-2 serum IgG levels and the number of SARS-CoV-2 positive plasmablasts at the same timepoint (figure 3C-D). No antigen specific plasma cells were detected in the blood of any participant before the first vaccination (data not shown).

• Download figure
• Open in new tab

Figure 3:

Antibody secreting cells (ASC) measured in 3-colour-Fluorospot 7 days after the second vaccination. See figure S3 for representative image. A Number of total and SARS-CoV-2 specific ASCs per µL blood. B Spot size distribution. Single points resemble the mean spot size per spot of one donor. Correlation (Pearson) of the number of SARS-Cov-2 specific ASCs C against SARS-CoV-2 specific plasma blasts in flow cytometry and D against SARS-CoV-2 serum IgG. Statistical differences: Kruskal-Wallis test with Dunn’s post hoc test, significant differences are indicated as *P<0.05, ***P<0.001

SARS-CoV-2 specific T cells show signs of delayed activation

SARS-CoV-2 S-specific CD4+ T cells could be detected at similar frequencies in all groups after second vaccination (figure 4A). We also observed no differences in cytokine production (TNFα, IFNγ, IL-2, IL-21 or IL-10) by the S-specific T cells between TNFα patients, oDMARDs and controls (figure 4B). In contrast, TNFα patients showed significantly lower expression of PD-1 and a trend towards increased levels of Ki-67, suggesting a delayed or still ongoing activation of these cells (figure 4C).

• Download figure
• Open in new tab

Figure 4:

SARS-CoV-2 specific memory T cells in patients receiving TNFα blockers show a delayed activation status. Antigen-reactive T cell enrichment was performed two months after the second vaccination. A Frequencies of spike-reactive CD154+ memory cells within CD4+ T cells. B Percentage of cytokine production within spike-reactive CD154+ memory cells. C Percentage of PD-1 and Ki-67 positive cells within CD154+ memory cells. Each dot represents one donor, lines indicate mean values. Statistical differences: Kruskal-Wallis test with Dunn’s post hoc test, significant differences are indicated as **P<0.01