Aging and cytomegalovirus infection differentially and jointly affect distinct circulating T cell subsets in humans

Abstract

The impact of intrinsic aging upon human peripheral blood T cell subsets remains incompletely quantified and understood. This impact must be distinguished from the influence of latent persistent microorganisms, particularly CMV, which has been associated with age-related changes in the T cell pool. In a cross-sectional cohort of 152 CMV-negative individuals, aged 21-101 y, we found that aging correlated strictly to an absolute loss of naive CD8, but not CD4, T cells but, contrary to many reports, did not lead to an increase in memory T cell numbers. The loss of naive CD8 T cells was not altered by CMV in 239 subjects (range 21-96 y), but the decline in CD4(+) naive cells showed significance in CMV(+) individuals. These individuals also exhibited an absolute increase in the effector/effector memory CD4(+) and CD8(+) cells with age. That increase was seen mainly, if not exclusively, in older subjects with elevated anti-CMV Ab titers, suggesting that efficacy of viral control over time may determine the magnitude of CMV impact upon T cell memory, and perhaps upon immune defense. These findings provide important new insights into the age-related changes in the peripheral blood pool of older adults, demonstrating that aging and CMV exert both distinct and joint influence upon blood T cell homeostasis in humans.

Figure 1. Age-related proportional alterations in CD8 but not CD4 T cell subsets in CMV− subjects

Fresh PBMC were phenotyped by multi-color FCM (see methods). Events (≥ 3×10⁵) were gated on lymphocytes using forward and side scatter. T cells were defined as CD3+ CD4+ or CD3+CD8β+ and within each, N, CM and EM T cells were defined as CD28^int95^low, CD28^hi95^hi and CD28^low95^hi, respectively. Data shows changes in relative representation of total (A), naïve (B), CM (C) and EM (D) CD8 (gray closed circles) and CD4 (black open circles) cells with aging. Data originate from n = 152 individuals, ranging in age from 21–101. Each dot represents an individual. The lines drawn in the figure and p values are based upon the regression parameters as detailed in the methods and illustrated within supplemental Table, part B (See Table ST1).

Figure 2. Absolute numbers of CD8 T cells and CD8 naïve cells decline with age in CMV− subjects

Absolute cell numbers for each subset were obtained from fresh whole blood counts and from FCM analysis of fresh PBMC, as detailed in Methods. Subject cohort, staining, gating, phenotypic definition and statistical treatment were as in Fig. 1. Absolute numbers of total (A), naïve (B), CM (C) and EM(D) CD8 (gray closed circles) and CD4 (black open circles) cells are plotted against age. The lines drawn in the figure and p values are based upon the regression parameters as detailed in the methods and illustrated within supplemental Table, part A (ST1).

Figure 3. Inflation of Effector Memory CD8 and CD4 cells in CMV+ subjects with age

Methodology, analysis and panels are identical to Figure 2, except that n = 239 individuals ranging in age from 21– 96, and seropositive for CMV were analyzed. The lines drawn in the figure and p values are based upon the regression parameters as detailed in the methods and illustrated within supplemental Table (ST1).

Figure 4. Cohort size considerations

PBMC from the entire cohort (n=391) (239 CMV+, grey closed squares; 152 CMV−, black open squares, A.) or its subsets of 80 (40 CMV+ and 40 CMV− donors, B) or 120 (80 + another 40 weighed; total 73 CMV+ and 47 CMV−, C) individuals were analyzed for the possible impact of CMV upon the decline of naive T cells plotted against age. Note the changes in significance (p value by regression analysis) as a consequence of cohort size.

Figure 5. Multi-color FCM gating strategy (A, B)

Representative flow cytometric analysis of one donor demonstrating gating to identify T cell subset populations (naïve (N), central memory (CM), and effector memory(EM)) and their respective overlaps between phenotypic gating strategy (CD28/CD95 vs. CD45ra/CCR7). Color coding as follows: green = naïve cells, orange = central memory cells, blue (dark and lt. blue) = effector memory and terminal effector memory respectively.

Figure 6. Impact of phenotyping marker selection

Multiple surface marker definitions confirm fluctuations of N and EM T cells with age and CMV, respectively. Frozen PBMC from a weighted subset of our cohort, consisting of a total of 116 individuals, comprised of CMV− (n=45) and CMV+(n=71) donors were thawed, rested overnight and phenotyped by multi-color FCM as in methods.≥ 3 × 10⁵ events were analyzed, gated via the lymphocyte forward/size scatter gate, then on live cells, then with the main T cell subsets defined as CD3+CD4+ or CD3+CD8β+ and the naïve cells defined as CD28^int95^lo or CD45RA^hiCCR7^hi. Correlation between absolute numbers for Naive, left panels A (CD4), C (CD8) and Effector Memory (right panel B (CD4), D(CD8) cells defined as CD28^int95^low (phenotype 1, closed circles) and CD45RA+CCR7+ (phenotype 2, open circles) for the naïve and CD28^hi95^hi (phenotype 1, closed circles) and CD45RA+ or - CCR7− (phenotype 2, open circles) for the effector memory is shown. p values of R² from regression concordance are all highly significant. E, F = Corroboration of EM changes using CD57. The same cohort was analyzed by gating on CD3+CD4+ (open circles) or CD3+CD8+ (gray closed circles) cells to enumerate CD95+, CD45RA+or- CD28− CD57+ EM cell. The analysis shows that this cell type accumulates solely in the presence of CMV infection (F) but not its absence (E).

Figure 7. CD8 EM expansion is dependent upon CMV Titer and Age

Results shown in Figs. 2 & 3 were plotted against anti-CMV Ab titers by separating the cohort (n=391) into upper and lower anti-CMV Ab titer half (lower<349>upper); multiple regression model with age and CMV status as covariates was used to calculate significance. CD8 EM numbers are shown for anti-CMV titer hi, lo and CMV(−): A, for individuals 40y and younger; B, individuals 50–64yrs of age; and C, for individuals 65yrs of age and older. D. Slope of absolute naïve CD8 cells count over age groups. E. Slope of absolute EM T cell inflation over age groups. Data from CMV−, CMVAb^lo and CMVAb^hi subjects are shown in solid line/squares, dashed line/triangels and dotted line/circles, respectively. Significance ** p=0.01 to p=0.001, ***p>0.001 was determined by two-way ANOVA followed by Bonferroni correction for multiple comparisons. NS= not significant.

Figure 8. Variance in Absolute Count is Significantly Increased in the Inflated Effector Memory Pool of CMV(+) Donors

Absolute T cell inflation within age bins from CMV− (open squares) and CMV+ (closed squares) subjects. Plots show the standard deviation of absolute naïve (A, D, G) and EM (B, E, H) T cell counts. Overall N=391; CMV(+) = 239. CMV(−) =152, and as noted within axis of each panel. Panels C, F, and I illustrate that the CMV− and CMV+ cohorts are age matched within each age bin. Significance as noted within the panels.

Figure 9. Variance in Absolute Count of Naïve Pool Declines with Advancing Age Independent of CMV Status; Variance Increases Significantly in aged CMV+ donor subset

Plots show the standard deviation of absolute naïve (A) and EM (B) T cell counts. N=391; CMV(+) = 239. CMV(−) =152. Absolute T cell inflation over age from CMV− (dark line, open squares) and CMV+ (gray line, closed squares) subjects.