mRNA vaccination boosts S-specific T cell memory and promotes expansion of CD45RAint TEMRA-like CD8+ T cells in COVID-19 recovered individuals

Abstract

SARS-CoV-2 infection and mRNA vaccination both elicit spike (S)-specific T cell responses. To analyze how T cell memory from prior infection influences T cell responses to vaccination, we evaluated functional T cell responses in naive and previously infected vaccine recipients. Pre-vaccine S-specific responses are predictive of subsequent CD8⁺ T cell vaccine-response magnitudes. Comparing baseline with post-vaccination TCRβ repertoires, we observed large clonotypic expansions correlated with the frequency of spike-specific T cells. Epitope mapping the largest CD8⁺ T cell responses confirms that an HLA-A∗03:01 epitope was highly immunodominant. Peptide-MHC tetramer staining together with mass cytometry and single-cell sequencing permit detailed phenotyping and clonotypic tracking of these S-specific CD8⁺ T cells. Our results demonstrate that infection-induced S-specific CD8⁺ T cell memory plays a significant role in shaping the magnitude and clonal composition of the circulating T cell repertoire after vaccination, with mRNA vaccination promoting CD8⁺ memory T cells to a T_EMRA-like phenotype.

Keywords: CD4 T cells; CD8 T cells; CyTOF; SARS-CoV-2; vaccination.

Graphical abstract

Figure 1

CD4⁺ and CD8⁺ T cell response to vaccination in naive and COVID-19 recovered individuals (A–D) (A) Percentage of CD4⁺ cells expressing IFN-γ or IL-2 and (B) percentage of CD8⁺ T cells expressing IFN-γ and/or IL-2 in response to peptide pools spanning the SARS-CoV-2 spike (S) protein. Phenotypes were determined by expression of CD45RA and CCR7 cell-surface markers (see STAR Methods), and intracellular cytokine staining in naive and COVID-19 recovered individuals was quantified by (C) percentage of CD4⁺ or (D) percentage of CD8⁺ T cells secreting IFN-γ and/or IL-2 in response to peptide pools spanning SARS-CoV-2 spike protein. Plots compare naive (blue, n = 55) versus individuals with previous SARS-CoV-2 infection (black, n = 98), with p values indicating results of a two-sided Wilcoxon rank-sum test (conditional on sufficient sample size for comparison). (E) At the post-second-dose time point, the frequency and proportion of CD8⁺ expressing IFN-γ and/or IL-2 after stimulation with SARS-CoV-2 S peptides classified as naive, T_CM, T_EM, T_EMRA in each individual donor. Participants without (left) and with (right) a previous SARS-CoV-2 infection are arranged on the x axis in ascending order based on their overall S-reactive CD8 T cell response.

Figure 2

Correlations between immunological measures in COVID-19 recovered participants (A) Pairwise Spearman rank correlations between key immunological measurements in pre-vaccine and post-first dose samples (n = 22). (B) Correlation between percent of pre-vaccine reactive CD4⁺ T cells (IFN-γ and/or IL-2) after stimulation with S1 peptides pools versus percent of reactive CD8⁺ T cells after one dose. (C) Autocorrelation between pre-vaccine and post-first dose % CD8⁺ T cells expressing IFN-γ and/or IL-2 after stimulation with S1 peptides. (D) Correlation between percent S1 reactive CD4⁺ T cells pre-vaccine and the post-first dose titers of anti-RBD IgG. (E) Correlation between pre-vaccine S1 CD8⁺ T cells and post-first dose anti-RBD IgG. R refers to Spearman rank correlation.

Figure 3

S-specific T cell responses after a first and second mRNA vaccine dose in COVID-19 recovered participants (A–E) Study participants who received two doses of mRNA vaccine (BNT162b2 or mRNA-1273) were selected for analysis if blood draws were conducted within 6 to 30 days following administration of the first mRNA vaccine dose and 6 to 30 days following the second mRNA vaccine dose. PBMCs were stimulated with S1 (A and C) and S2 (B and D) peptide pools. The p values were computed with a two-sided paired signed-rank test to assess whether the percentage of IFN-γ or IL-2-positive CD4⁺ and CD8⁺ T cells changed following administration of the second vaccine dose. The number out of 15 indicates the number of participants who were assessed as responders based on MIMOSA response call (see STAR Methods). We noted that five of six of the highest post-second dose CD8⁺ T cell responses following S1 stimulation were from patients expressing the HLA-A∗03:01 allele (indicated on the plots by red dashes). These sample reflect vaccination at least 100 days following infection with the timing of infection and post-mRNA sample collection relative to first vaccination shown in (E).

Figure 4

Associations between increasing cytokine (IFN-γ or IL-2) response in CD4⁺ and CD8⁺ T cells and detectable post-vaccination T cell clonal expansions measured at the TCRβ locus PBMCs were analyzed by ICS following administration of first (1) and second (2) mRNA vaccine doses in 15 COVID-19 recovered individuals. (A) Percent of T cells secreting IFN-γ or IL-2 after stimulation with combined S1 and S2 results. (B) Percent of repertoire composed of expanded TCRβ clones (2× fold change relative to the pre-vaccination time point and false discovery rate-adjusted q value < 0.05, Fisher’s exact test). (C) Clonal breadth of expanded clones—the number of unique expanded clones divided by the total number of unique productive clones sequenced. (D) Association between the percent of the repertoire composed of expanded clones and percent of CD4⁺ and CD8⁺ T cells expressing IFN-γ or IL-2. (E) Association between breadth of expanded clones and percent of CD4⁺ and CD8⁺ T cells expressing IFN-γ or IL-2. Line connects each participant’s post-first and post-second dose time point, which is designated with a filled circle. R refers to Spearman rank correlation.

Figure 5

Epitope mapping and mass cytometry-based phenotypic profiling using the A∗03 KCYGVSPTK tetramer (A) Mini-pool epitope mapping of PBMCs, collected 16–19 days after the second vaccine dose, from four A∗03-expressing participants by IFN-γ ELISpot. (B) 15-mer peptide-level mapping within S345:411 mini-pool. (C) 15-mer peptide-level mapping within the S573:643 mini-pool. (D) Mapping of tetramer-positive T cells to phenotypic space after multiple antigen exposures following infection and vaccination. Timing of sample collection in days post symptoms (dps) and after last vaccination (+x days) are indicated in each panel. (E) Distribution of phenotypes in A∗03 KCYGVSPTK tetramer-positive cells. (F and G) (F) Heatmap depicting mean marker abundance measured by mass cytometry in each cluster (G) UMAP projection of high-dimensional CD8⁺ T cell phenotypes from five participants. CD8⁺ T cells were down sampled to have the same number of events (10,000 cells per participant).

Figure 6

Ex vivo phenotypes of CD8⁺ and SARS-CoV-2 HLA-A∗03 KCY S-specific T cells from participant P1 after recovery from COVID-19 and three subsequent mRNA immunizations (A) UMAP projection of single-cell gene expression (scRNA-seq) and surface protein expression measured with antibody-derived tags (scADT). Each point is a single cell colored based on 20 phenotypes derived from Leiden clustering of normalized scRNA-seq and scADT-seq data across four longitudinal samples (resolution = 1.0). (B) Proportion of single cells manifesting each CD8⁺ T cell phenotype at each sampling time point. (C) Phenotypes of tetramer-positive single cells stained with HLA-A∗03 KCY tetramer and confirmed to match tetramer-positive CDR3β or CDR3α recovered from independently tetramer-sorted bulk repertoires. Gray points are tetramer negative CD8⁺ cells in longitudinal samples. (D) Proportion of tetramer-positive single cells manifesting each CD8⁺ T cell phenotype. (E) Proportion of total CD8⁺ T cells of each (long-lived effector) LLE-T_EMRA, T_EMRA, (less-differentiated) LD-T_EMRA, and activated CD8⁺ T cells. (F) Genes and ADTs most differentially expressed between LLE-T_EMRA and classical T_EMRA CD8⁺ T cells after the post-second vaccine time point. (G and H) (G) T_EMRA-associated gene markers similar between LLE-T_EMRA and classical T_EMRA cells, and (H) exhaustion-associated gene markers.