NK- and T-cell granzyme B and K expression correlates with age, CMV infection and influenza vaccine-induced antibody titres in older adults

Abstract

Granzymes are a family of serine-proteases that act as critical mediators in the cytolytic and immunomodulatory activities of immune cells such as CD8⁺ T-cells and natural killer (NK) cells. Previous work indicates that both granzyme B (GZB) and K (GZK) are increased with age in CD8⁺ T-cells, and in the case of GZB, contribute to dysfunctional immune processes observed in older adults. Here, we sought to determine how GZB and GZK expression in NK-cells, and CD4⁺, CD8⁺, and gamma-delta T-cells, quantified in terms of positive cell frequency and mean fluorescence intensity (MFI), differed with age, age-related health-traits and the antibody response to high-dose influenza vaccine. We found that the frequency and MFI of GZB-expressing NK-cells, and CD8⁺ and Vδ1+ T-cells, and GZK-expressing CD8⁺ T-cells was significantly higher in older (66-97 years old; n = 75) vs. younger (24-37 years old; n = 10) adults by up to 5-fold. There were no significant associations of GZB/GZK expression with sex, frailty or plasma levels of TNF or IL-6 in older adults, but those who were seropositive for cytomegalovirus (CMV) exhibited significantly higher frequencies of GZB+ NK-cells, and CD4⁺, CD8⁺ and Vδ1+ T-cells, and GZK+ CD8⁺ T-cells (Cohen's d = .5-1.5). Pre-vaccination frequencies of GZB+ NK-cells were positively correlated with vaccine antibody responses against A/H3N2 (d = .17), while the frequencies of GZK+ NK and CD8⁺ T-cells were inversely associated with A/H1N1 (d = -0.18 to -0.20). Interestingly, GZK+ NK-cell frequency was inversely correlated with pre-vaccination A/H1N1 antibody titres, as well as those measured over the previous 4 years, further supporting a role for this subset in influencing vaccine antibody-responses. These findings further our understanding of how granzyme expression in different lymphoid cell-types may change with age, while suggesting that they influence vaccine responsiveness in older adults.

Keywords: CMV; NK-cells; aging; granzyme B; granzyme K; influenza; t-cells; vaccination.

FIGURE 1

Gating strategy depicting the identification of GZB- and GZK-expressing NK-cells and CD4⁺, CD8⁺ and γδ⁺T-cells.

FIGURE 2

A comparison of granzyme B (GZB) and K (GZK) expression in young (YA) and older (OA) adults. The geometric mean and 95% confidence interval for (A) the frequency of GZB, GZK or GZB/GZK-expressing cells, and (B) the mean fluorescence intensity (MFI) of GZB or GZK expression within those populations are shown, with the significance of age-related differences determined by t-test on log-transformed values. ***, p < .001; **, p < .01; *, p < .05.

FIGURE 3

Factors associated with GZB and GZK expression in older adults. (A) The log frequency of GZB or GZK-expressing cells and (B) the log mean fluorescence intensity (MFI) of each population was regressed on age, sex, frailty index, CMV serostatus, and plasma TNF and IL-6, while adjusting for site of enrollment. For sex and CMV status, contrasts are males vs. females and positive vs. negative participants, respectively. For all others, coefficients are relative to the magnitude of increase shown in brackets. No overlap of the 95% CI with the red dotted line indicates that the association is significant at the nominal level, while the ‘x’ and ‘#’ indicate that the respective associations for GZB or GZK maintained significance at an FDR adjusted p < .05.

FIGURE 4

Granzyme-positive cell frequency is associated with vaccine responsiveness and antibody titres in older adults receiving the high-dose vaccine. (A) To model vaccine responsiveness, log-transformed titres for influenza A/H1N1, A/H3N2 and B at 4-week post-vaccination were regressed on the log-transformed frequencies of GZB- or GZK-expressing cells, adjusting for age, sex, enrollment site and the log pre-vaccination titres. (B) Associations between A/H1N1 titre levels at baseline (i.e., 0), and 4-, 10-, 20-week post-vaccination with the frequency of GZK+ NK-cells or CD8⁺ T cells, omitting adjustment for pre-vaccination titres. No overlap of the 95% CI with the red dotted line indicates significance at the nominal level, while the ‘#’ indicates that significance was maintainted at an FDR adjusted p < .05. No significant differences beyond the nominal level were observed in (A).

FIGURE 5

GZK-expressing NK- and CD8 T-cell frequency is inversely associated with prior A/H1N1 antibody titres in older adults. (A) Participants enrolled in the current cohort had also been enrolled in multiple years of a previous randomized vaccine trial, in which A/H1N1 antibody titres against the vaccine strain were measured. (B) Associations for standardized log A/H1N1 antibody titres measured at baseline (i.e., 0) and 4-, 10-, and 20-week post-vaccination, obtained for each year of the randomized trial or all years combined, with standardized log GZK+ NK-cell or CD8⁺ T-cell frequencies obtained in 2018/19. Models for individual years were adjusted for age, sex, vaccine dose, and site of enrollment, whereas models for all years combined were adjusted for age, sex, dose and random intercepts for site, participant and year of study. No overlap of the 95% CI with the red dotted line indicates significance at the nominal level, while the ‘#’ indicates that significance was maintainted at an FDR adjusted p < .05.