Age and Latent Cytomegalovirus Infection Do Not Affect the Magnitude of De Novo SARS-CoV-2-Specific CD8+ T Cell Responses

Abstract

Immunosenescence, age-related immune dysregulation, reduces immunity upon vaccinations and infections. Cytomegalovirus (CMV) infection results in declining naïve (T_naïve) and increasing terminally differentiated (T_emra) T cell populations, further aggravating immune aging. Both immunosenescence and CMV have been speculated to hamper the formation of protective T-cell immunity against novel or emerging pathogens. The SARS-CoV-2 pandemic presented a unique opportunity to examine the impact of age and/or CMV on the generation of de novo SARS-CoV-2-specific CD8⁺ T cell responses in 40 younger (22-40 years) and 37 older (50-66 years) convalescent individuals. Heterotetramer combinatorial coding combined with phenotypic markers were used to study 35 SARS-CoV-2 epitope-specific CD8⁺ T cell populations directly ex vivo. Neither age nor CMV affected SARS-CoV-2-specific CD8⁺ T cell frequencies, despite reduced total CD8⁺ T_naïve cells in older CMV^- and CMV⁺ individuals. Robust SARS-CoV-2-specific central memory CD8⁺ T (T_cm) responses were detected in younger and older adults regardless of CMV status. Our data demonstrate that immune aging and CMV status did not impact the SARS-CoV-2-specific CD8⁺ T cell response. However, SARS-CoV-2-specific CD8⁺ T cells of older CMV^- individuals displayed the lowest stem cell memory (T_scm), highest T_emra and PD1⁺ populations, suggesting that age, not CMV, may impact long-term SARS-CoV-2 immunity.

Keywords: CMV; COVID‐19; aging; antigen‐specific CD8+ T cells; immunosenescence.

FIGURE 1

Convalescent COVID‐19 cohort and HLA‐I profiles. (A) Overview of convalescent donor cohort. Distribution of age (B), days post onset symptoms (C), sex (D), and disease severity (E). (C) Donors with unspecified days post‐onset symptoms were not included (younger CMV⁻ n = 4 and younger CMV⁺ n = 2 adults). (F) Number of donors with HLA‐I allele of interest per donor group, donors were selected based on the expression of at least one of the four prominent HLA‐I allotypes (bold). (G) Number co‐expressed HLA‐I allotypes of interest among donors in the cohort. (H) Protein distribution of 35 SARS‐CoV‐2 epitopes per HLA‐I allotype. (I) Number of SARS‐CoV‐2‐derived epitopes tested per donor group. (B, C, I) Each dot represents an individual donor (younger CMV⁻ n = 19, younger CMV⁺ n = 21, older CMV⁻ n = 19, and older CMV⁺ n = 18 adults). (B) Statistical analysis was performed using a one‐way ANOVA followed by the Tukey HSD test, horizontal bar indicates the mean. (C, I) Statistical analysis was performed using a Wilcoxon rank‐sum test with Bonferroni‐Holm's multiple comparison correction, horizontal bar indicates median. Significant p‐values are provided above the graph. (A) Created with BioRender.com.

FIGURE 2

Robust SARS‐CoV‐2 CD8⁺ T cell frequencies independent of age and CMV status. (A) CD8⁺ T cell frequencies, each dot represents an individual donor. (B) Overview of the heterotetramer combinatorial coding approach, adjusted from [50]. (C) Representative flow cytometry plots created by Boolean gating CD8⁺ T cell populations expressing tetramer fluorophores (full gating strategy in Figure S2). Frequency of SARS‐CoV‐2 tetramer⁺CD8⁺ T cells per donor group (D) or subdivided per HLA‐I allotype (E), each dot represents an epitope‐specific population of an individual donor (younger CMV⁻ n = 19, younger CMV⁺ n = 21, older CMV⁻ n = 19 and older CMV⁺ n = 18 adults). (D, E) Open symbols indicate tetramer⁺CD8⁺ T cell populations of 3–8 detected cells, these were excluded from phenotypic analysis. Frequency of tetramer⁺CD8⁺ T cells is shifted by 10⁻⁴ to allow for visibility on logarithmic y axes (i.e., no detected tetramer⁺ events displayed as 10⁻⁴). (E) The number of HLA‐I‐restricted SARS‐CoV‐2 tetramer⁺CD8⁺ T cell populations are indicated at the top of the graph, not determined (ND) indicates HLAs for which no donors were available. The four prominent HLA‐I allotypes donors were selected for are indicated in bold. (A) Statistical analysis was performed using a one‐way ANOVA followed by the Tukey HSD test, horizontal lines indicate mean. (D, E). Statistical analysis was performed using a Wilcoxon rank‐sum test including Bonferroni‐Holm's multiple comparison correction, horizontal lines indicate median. No significant differences were identified. (B) Created with BioRender.com.

FIGURE 3

SARS‐CoV‐2‐specific CD8⁺ T cells display a robust T_cm phenotype regardless of age or CMV status. (A) Representative FACS gating strategy used to characterize the phenotype profile of SARS‐CoV‐2 epitope‐specific CD8⁺ T cells. CD27, CD45RA, CD95, and PD1 were used to identify T_cm (CD27⁺CD45RA⁻), T_em (CD27⁻CD45RA⁻), T_emra (CD27^dim/‐CD45RA⁺), and T_{naïve‐like} (CD27^highCD45RA⁺) cells, the latter can be further subdivided based on CD95 expression into T_naïve (CD27^highCD45RA⁺CD95⁻) and T_scm (CD27^highCD45RA⁺CD95⁺). Gates were defined based on the total CD8⁺ T cell population (top panel). The bottom panel displays overlays of the dual‐tetramer positive cells (red dots) on the entire CD8⁺ T cell population (grey dots). (B) Radar plots of median phenotype frequencies of the total CD8⁺ T cell per donor group (younger CMV⁻ n = 19, younger CMV⁺ n = 21, older CMV⁻ n = 19, and older CMV⁺ n = 18). (C) Radar plots of median tetramer⁺CD8⁺ T cell populations per donor group (younger CMV⁻ n = 19, younger CMV⁺ n = 18, older CMV⁻ n = 17, and older CMV⁺ n = 18). (D) Frequency of total T_naïve, T_scm, T_cm, T_em, T_emra, and PD1⁺ tetramer⁺CD8⁺ T cell population per donor group, the horizontal bar indicates median (T_naïve, T_scm and PD1 populations: younger CMV⁻ n = 8, younger CMV⁺ n = 10, older CMV⁻ n = 9 and older CMV⁺ n = 9 adults; T_cm, T_em, and T_emra similar numbers as in B). Statistical analysis was performed using a Wilcoxon rank‐sum test including Bonferroni‐Holm's multiple comparison correction. Significant p‐values are displayed next to the radar plots (B, C) or above the graphs (B–D).