Primary ChAdOx1 vaccination does not reactivate pre-existing, cross-reactive immunity

Abstract

Currently available COVID-19 vaccines include inactivated virus, live attenuated virus, mRNA-based, viral vectored and adjuvanted protein-subunit-based vaccines. All of them contain the spike glycoprotein as the main immunogen and result in reduced disease severity upon SARS-CoV-2 infection. While we and others have shown that mRNA-based vaccination reactivates pre-existing, cross-reactive immunity, the effect of vector vaccines in this regard is unknown. Here, we studied cellular and humoral responses in heterologous adenovirus-vector-based ChAdOx1 nCOV-19 (AZ; Vaxzeria, AstraZeneca) and mRNA-based BNT162b2 (BNT; Comirnaty, BioNTech/Pfizer) vaccination and compared it to a homologous BNT vaccination regimen. AZ primary vaccination did not lead to measurable reactivation of cross-reactive cellular and humoral immunity compared to BNT primary vaccination. Moreover, humoral immunity induced by primary vaccination with AZ displayed differences in linear spike peptide epitope coverage and a lack of anti-S2 IgG antibodies. Contrary to primary AZ vaccination, secondary vaccination with BNT reactivated pre-existing, cross-reactive immunity, comparable to homologous primary and secondary mRNA vaccination. While induced anti-S1 IgG antibody titers were higher after heterologous vaccination, induced CD4⁺ T cell responses were highest in homologous vaccinated. However, the overall TCR repertoire breadth was comparable between heterologous AZ-BNT-vaccinated and homologous BNT-BNT-vaccinated individuals, matching TCR repertoire breadths after SARS-CoV-2 infection, too. The reasons why AZ and BNT primary vaccination elicits different immune response patterns to essentially the same antigen, and the associated benefits and risks, need further investigation to inform vaccine and vaccination schedule development.

Keywords: SARS-CoV-2; antigen-specific T-cells; cross-reactivity; heterologous vaccination; humoral response.

Figure 1

Kinetics of adaptive immune responses upon heterologous vaccination. Immune response kinetics following heterologous SARS-CoV-2 vaccination of unexposed young healthy donors (n=17) vaccinated with AZ at d0, then BNT 85(+/-3) days after first dose and 142(+/-7) days after second dose. (A) Ex vivo stimulation of PBMCs with S-I and S-II peptide pools. The percentage of CD40L+4-1BB+ within CD4+ T cells among stimulated PBMCs was divided by the percentage of these cells in the unstimulated control, resulting in the stimulation index (SI). Dotted lines indicate a SI of 1.5 and 3, separating non-responders from responders with uncertainty and definite responders. (B) Frequencies of CD3lo cells among S-I- or S-II-reactive CD40L+4-1BB+ CD4+ T cells of T cell responses with a SI ≥ 1.5.  (C) Frequencies of HLA-DR+CD38+ among CD40L+4-1BB+ CD4+ T cells. (D) Direct comparison of the SI, frequencies of CD3lo and HLA-DR+CD38+ cells of CD40L+4-1BB+ CD4+ T cells at indicated time points upon stimulation with S-I or S-II peptide pools. (E) Serum anti-SARS-CoV-2 S1 IgA and IgG titer ratios. Only significant differences are shown with *p < 0.05, **p<0.01, ***p<0.001 (A-C, E-F: Wilcoxon matched-pairs signed-rank test between consecutive days, D: Mann-Whitney test). SI below 1 were excluded from further analysis, as they are below the lower limit of detection. Black line indicates the median.

Figure 2

Heterologous vaccination results in slower induction of CD4+ T cell responses, but higher IgG responses compared to homologous vaccination. (A) Ex vivo stimulation of PBMCs from donors receiving heterologous AZ-BNT and homologous BNT-BNT vaccination with S-I and S-II peptide pools at indicated time points. SI of antigen-specific CD40L+4-1BB+ CD4+ T cells is shown. (B) Ex vivo stimulation of PBMCs from AZ or BNT primary vaccinated donors with S-I and S-II peptide pools early and at peak of immune response. SI of antigen-specific CD40L+4-1BB+ CD4+ T cells is shown. (C) Frequencies of CD3lo cells among S-I- or S-II-reactive CD40L+4-1BB+ CD4+ T cells of T cell responses with a SI ≥ 1.5. (D) Ex vivo stimulation of PBMCs from AZ or BNT primary vaccinated donors with S-I and S-II peptide pools early and at peak of immune response. Frequencies of CD3lo antigen-specific CD40L+4-1BB+ CD4+ T cells are shown. (E) Serum anti-SARS-CoV-2 S-1 IgG and IgA antibody levels (OD) were determined at indicated time points. Upper and lower levels of detection were set at 1 and 13 (IgG)/ 10 (IgA), respectively, indicated by dotted lines. (F) Anti-SARS-CoV-2 B.1.1.7 (Alpha) and B1.1.529 (Omicron) subtype BA.5 variant spike neutralizing capacity at d14 post primary, d14 post secondary and 12 weeks post booster vaccination. Positivity thresholds: >10 ID50 for spike neutralization. Serum ID50 values less than the lowest serum dilution tested (1:10) were assigned a value of 5 for plotting the graph and for statistical analysis. Only significant differences are shown with *P < 0.05, **P < 0.01, ***P < 0.001, Mann-Whitney test. SI below 1 were excluded from further analysis, as they are below the lower limit of detection. Black line indicates the median.

Figure 3

Cross-reactive cellular responses are not effectively induced by AZ primary vaccination. (A) Ex vivo stimulation of PBMCs with S-II peptide pool. Donors were separated into cross-reactive responders according to an SI of >3 at d0 and non-cross-reactive responders (baseline SI <3). (B) Stimulation index of antigen specific CD40L⁺4-1BB⁺ CD4⁺ T cells, frequencies of CD3^lo cells among iCope-reactive CD40L⁺4-1BB⁺ CD4⁺ T cells of T cell responses with a SI ≥ 1.5, and HLA-DR⁺CD38⁺ among iCope-reactive CD4⁺ T cells are shown. (C) Comparison of SI, CD3^lo and HLA-DR⁺CD38⁺ between heterologous (AZ-BNT-BNT) and homologous (BNT-BNT-BNT) vaccinated donors.(D) Foldchange of the SI of iCope-specific T cells from d0 to d7 and d0 to d14. Only significant differences are shown with *P < 0.05, **P < 0.01, ***P < 0.001. A, C-D: Mann-Whitney test B: Wilcoxon matched-pairs signed-rank test. SI below 1 were excluded from further analysis, as they are below the lower limit of detection. Black line indicates the median.

Figure 4

Cross-reactive humoral responses are not effectively induced by AZ primary vaccination. (A) Signal from sample incubation on peptide microarrays for selected peptides following AZ-BNT (n=16), BNT-BNT (n=15) vaccination or infection (convalescents (CS), n=17). (B) Levels of anti-S1 or anti-S2 IgG binding antibody intensity units in indicated cohorts. Dotted lines indicate lower cut-off at 18 for values classified as positive and upper cut-off at 172. BL=baseline, FU=follow-up. ns=not significant, *P < 0.05, **P < 0.01, ***P < 0.001. Mann-Whitney test. Black line indicates the median.

Figure 5

Three months post last antigen contact, single cell RNA sequencing of S-I-reactive CD4+ T cells reveals comparable TCR repertoire in heterologous (AZ-BNT-BNT) vaccinated, homologous (BNT-BNT-BNT) vaccinated and homologous vaccinated and infected (BNT-BNT-INF) individuals. (A) D50 index indicates the number of clones occupying 50 % of the repertoire. (B) Inverse Simpson Index indicates the TCRαβ repertoire diversity. High values represent a more even distribution of clonotypes, whereas low values indicate enrichment of certain clonotypes. (C) Rare clonal proportion shows the summary proportion of clonotypes with specific counts. (D) Venn diagram displaying the repertoire overlap between groups. (E) Circos plot giving the numbers of shared clones between samples of the different cohort. (F) UMAP clustering of seven lymphocyte subsets based on marker genes and distributions of groups across them. (A, B) Mann-Whitney test revealed no significant differences between groups (P > 0.05). Black line indicates the median.