COVID-19 mRNA booster vaccine induces transient CD8+ T effector cell responses while conserving the memory pool for subsequent reactivation

Abstract

Immunization with two mRNA vaccine doses elicits robust spike-specific CD8⁺ T cell responses, but reports of waning immunity after COVID-19 vaccination prompt the introduction of booster vaccination campaigns. However, the effect of mRNA booster vaccination on the spike-specific CD8⁺ T cell response remains unclear. Here we show that spike-specific CD8⁺ T cells are activated and expanded in all analyzed individuals receiving the 3^rd and 4^th mRNA vaccine shots. This CD8⁺ T cell boost response is followed by a contraction phase and lasts only for about 30-60 days. The spike-specific CD8⁺ T memory stem cell pool is not affected by the 3^rd vaccination. Both 4^th vaccination and breakthrough infections with Delta and Omicron rapidly reactivate CD8⁺ T memory cells. In contrast, neutralizing antibody responses display little boost effect towards Omicron. Thus, COVID-19 mRNA booster vaccination elicits a transient T effector cell response while long-term spike-specific CD8⁺ T cell immunity is conserved to mount robust memory recall targeting emerging variants of concern.

Fig. 1. Cellular and humoral effector response after the 3^rd vaccine dose.

Calculated ex vivo frequencies of spike-specific CD8⁺ T cells throughout 1^st, 2^nd and 3^rd vaccination (a). Ki-67,and CD38 expression within spike-specific non-naïve CD8⁺ T cells (b). t-SNE representation of flow cytometry data comparing A*01/S₈₆₅- and A*02/S₂₆₉-specific CD8⁺ T cells at peak response after 1^st, 2^nd and 3^rd vaccination. Expression levels of CD38, T-BET, BCL-2, PD-1 and TOX are depicted (c). Statistical significance was determined by two-way ANOVA with main model (a, b) comparing the effects of targeted epitopes (p_e) and of the time course (p_t). Source data are provided as a Source Data file.

Fig. 2. Humoral response after the 3^rd vaccine dose.

Antibody neutralization activity as 50% plaque reduction neutralization tests (PRNT50) for SARS-CoV-2 variants B.1, B.1.617.2 and B.1.1.529 after 3^rd vaccination. Median values are depicted with 95% confidence interval error bars with n = 28 individual tested (a). Quantification of serum anti-SARS-CoV-2 spike IgG levels following 3^rd vaccination with n = 26 individuals tested (b). Calculated ex vivo frequencies of A*01/S₈₆₅- (left) and A*02/S₂₆₉-specific (right) CD8⁺ T cells, with antibody neutralization activity after 4^th vaccination (c). Median values are depicted with 95% confidence interval error bars. Statistical significance was calculated by Kruskal-Wallis test (a) comparing the effects of time course (p_t < 1×10⁻¹⁴). Source data are provided as a Source Data file.

Fig. 3. Spike-specific CD8⁺ T memory cells before and after the 3^rd vaccine dose.

Composition of spike-specific CD8⁺ T cell memory subsets at 3 (n = 7) and 9 (n = 11) months after 2^nd, and 3 (n = 11) months after 3^rd vaccination (a). Molecule expression of spike-specific non-naïve CD8⁺ T cells at 3 (n = 8) and 9 (n = 12) months after 2^nd, and 3 (n = 11) months after 3^rd vaccination normalized to bulk naïve CD8⁺ T cells (b). Calculated ex vivo frequencies of BCL-2^hi (c) and T_SCM (d) spike-specific CD8⁺ T cells. t-SNE representation of BCL-2^hi (e) and T_SCM (f) spike-specific CD8⁺ T cells at 3 and 9 months after 2^nd, and 3 months after 3^rd vaccination. Expression levels of CCR7, TCF-1 and T-BET are depicted for BCL-2^hi, and CD127, CD11a, CXCR3 and CD27 are depicted for T_SCM spike-specific CD8⁺ T cells. Statistical significance was calculated by two-way ANOVA with full model and Tukey’s test for multiple comparison (a, b) to examine the effect of sampling time points (p_t) on memory subsets and marker expression, and two-way ANOVA with main model (c, d) to compare the effects of targeted epitopes (p_e) and time course (p_t). Source data are provided as a Source Data file.

Fig. 4. Reactivation capacity of spike-specific CD8⁺ T cells before and after the 3^rd vaccine dose.

Calculated ex vivo frequencies of TCF-1⁺ non-naive spike-specific CD8⁺ T cells (a). Expansion capacity of spike-specific CD8⁺ T cells over 14 days of in vitro expansion throughout 1^st, 2^nd and 3^rd vaccination (b). Percentage of IFNγ and TNF (c), and CD107a (d) producing CD8⁺ T cells upon peptide stimulation related to the percentage of spike-specific CD8⁺ T cells from all CD8⁺ T cells at 3 (n = 10) and 9 (n = 12 for c, n = 11 for d) months post 2^nd, and 3 (n = 11) months post 3^rd vaccination after in vitro expansion. Median values are depicted with 95% confidence interval error bars. Expression of Granzyme B (e) and Perforin (f) of spike-specific CD8⁺ T cells at 3 (n = 6) and 9 (n = 12) months post 2^nd, and 3 months (n = 10 for e, n = 11 for f) post 3^rd vaccination after in vitro expansion with representative histograms. Median values are depicted with 95% confidence interval error bars. Statistical significance was calculated by two-way ANOVA with main model (a, b) to compare the effects of targeted epitopes (p_e) and time course (p_t), two-way ANOVA with full model and Tukey’s multiple comparison test (c), and Kruskal-Wallis test (d, e, f) to examine the effect of sampling time on functionality (p_t). Source data are provided as a Source Data file.

Fig. 5. CD8⁺ T effector cell response after breakthrough infection and the 4^th vaccine dose.

Calculated ex vivo frequencies of spike-specific CD8⁺ T cells before and after breakthrough infection and 4^th vaccination (a). Proportion of CD38⁺, Ki-67⁺ and T-BET^hi within non-naive spike-specific CD8⁺ T cells at peak expansion after breakthrough infection and 4^th vaccination with n = 11 for Omicron and n = 2 for Delta breakthrough infections, and n = 4 individuals receiving a 4^th vaccination (b). Statistical significance was calculated by two-way ANOVA with main model to compare the effects of antigen triggers (p_a) on epitope-specific T cell frequencies (a) and two-way ANOVA with full model and Tukey’s test for multiple comparison (b) to examine the effects of antigen triggers on activation marker expression. Source data are provided as a Source Data file.

Fig. 6. Characterization of CD8⁺ T memory cells after breakthrough infection and the 4^th vaccine dose.

Reactivation capacity of spike-specific CD8⁺ T cells 1 month after symptom onset and 4^th vaccination. Median values are depicted with 95% confidence interval error bars with n = 6 for Omicron and n = 2 for Delta breakthrough infections, and n = 4 individuals receiving a 4^th vaccination (a). t-SNE representation of spike-specific CD8⁺ T cells 1 month after symptom onset and 4^th vaccination. Expression levels of CD38, CCR7, TCF- and BCL-2 are depicted (b). Calculated ex vivo frequencies of BCL-2^hi non-naïve spike-specific CD8⁺ T cells before and after breakthrough infection and 4^th vaccination (c). Antibody neutralization activity as 50% plaque reduction neutralization tests (PRNT50) for SARS-CoV-2 variants B.1, B.1.617.2 and B.1.1.529 after breakthrough infection and 4^th vaccination with n = 11 for Omicron and n = 1 for Delta breakthrough infection, and n = 2 individuals receiving a 4^th vaccination (d). Statistical significance was calculated by Kruskal-Wallis test (a) to compare the effect of the antigen triggers on expansion (p_exp) and interferon production (p_inf) and two-way ANOVA with main model (c, d) to compare the effects of antigen triggers (p_a) on epitope-specific T cell frequencies (c) or the effects of VOCs (p_v) and time course (p_t). Source data are provided as a Source Data file.