Telomere length dynamics in human memory T cells specific for viruses causing acute or latent infections

Abstract

Background: Declining telomere length (TL) is associated with T cell senescence. While TL in naïve and memory T cells declines with increasing age, there is limited data on TL dynamics in virus-specific memory CD4+ T cells in healthy adults. We combined BrdU-labeling of virus-stimulated T cells followed with flow cytometry-fluorescent in situ hybridization for TL determination. We analyzed TL in T cells specific for several virus infections: non-recurring acute (vaccinia virus, VACV), recurring-acute (influenza A virus, IAV), and reactivating viruses (varicella-zoster virus, VZV, and cytomegalovirus, CMV) in 10 healthy subjects. Additionally, five subjects provided multiple blood samples separated by up to 10 years.

Results: VACV- and CMV-specific T cells had longer average TL than IAV-specific CD4+ T cells. Although most virus-specific cells were CD45RA-, we observed a minor population of BrdU+ CD45RA+ T cells characterized by long telomeres. Longitudinal analysis demonstrated a slow decline in average TL in virus-specific T cells. However, in one subject, VZV reactivation led to an increase in average TL in VZV-specific memory T cells, suggesting a conversion of longer TL cells from the naïve T cell repertoire.

Conclusions: TLs in memory CD4+ T cells in otherwise healthy adults are heterogeneous and follow distinct virus-specific kinetics. These findings suggests that the distribution of TL and the creation and maintenance of long TL memory T cells could be important for the persistence of long-lived T cell memory.

Figure 1

Telomere length (TL) measurement using flowFISH on proliferating T lymphocytes depends on fixation-permeabilization and RNA nuclease treatment. PBMC from healthy adults were either stimulated with immobilized anti-CD3 and anti-CD28 for 4 days or were cryopreserved on day 0 and thawed for processing on day 4; CD3⁺ T cells were magnetically sorted from both samples on day 4. Each sample was divided for flowFISH TL analysis and telomere restriction fragment (TRF) Southern blotting. (A) FlowFISH analysis using 3 different pre-hybridization conditions: without fixation-permeabilization prior to probe hybridization (no fix/perm), with fixation-permeabilization (Fix/Perm) prior to hybridization, and with fixation-permeabilization followed by RNase treatment prior to hybridization (Fix/Perm +RNase). Statistical comparisons were done using unpaired t-test, *** p<0.001, and ns=not significantly different. (B) TRF Southern blot analysis. DNA was extracted from CD3+ T cells isolated ex vivo or after stimulation with anti CD3+CD28 for four days and subject to Southern Blot Analysis. TRF lengths are shown at the bottom of both lanes and are the average of three separate 20 pixel-wide analyses using MatLab software running the MaTelo macro (see Methods).

Figure 2

BrdU-flowFISH allows for TL measurement in proliferating CD4⁺ and CD8⁺ T lymphocytes. (A) Flow cytometry gating strategy for TL measurement from probe mean fluorescence intensity (MFI) in BrdU⁺ CD4⁺ and CD8⁺ cells and in BrdU-negative naïve (CD45RA⁺) T cells. (B) Representative proliferative responses (BrdU+ FSC-A^high population) of CD4+ and CD8+ T cells to viral antigen. Values shown are frequencies of CD4⁺ or CD8⁺ T cells that were BrdU+. (C) Histograms of TL distribution for virus-specific CD4⁺ T cells defined as in B. Mean and median fluorescence intensity values, in arbitrary units (AU), are shown for each plot. Coefficient of variation (CV) is also shown.

Figure 3

TL and CD45RA⁺ frequencies in proliferating CD4⁺ T cells. (A) The frequency of BrdU+ CD4⁺ T cells on the indicated days after in vitro stimulation with viral antigens. BrdU was added to culture wells 3 days before harvest in all cases. (B) Histograms of TL in BrdU+ cell populations from A. Values indicate the frequencies of long telomere cells. (C) Distribution of TL and CD45RA staining in BrdU+ cells from A. Gate frequencies indicate the percentages of CD45RA^- and CD45RA⁺ cells in each sample. Data are representative of two independent experiments.

Figure 4

TL in CMV- and VACV-specific CD4⁺ T cells are longer than TL in IAV-specific CD4⁺ T cells. (A) Frequencies of BrdU+CD4⁺ T cells in ten healthy adults. (B) Mean TL measured in CD4⁺ T cells grouped by subject in units of molecules of equivalent soluble fluorescence (MESF). (C) Absolute TL in CD4⁺ T cells grouped by virus. (D) TL in CD4⁺ T cells that proliferated to viral antigen normalized to TL in naïve CD4⁺ T cells (BrdU^negative CD45RA⁺) (TL/TL_naïve) in the same subject. Statistical analyses: ** p < 0.01, * p < 0.05 by Wilcoxon paired, signed rank test. (E) Linear regression analyses for correlations between virus-specific proliferation frequencies (% BrdU+) and TL/TL_naive. P values are from Pearson correlation and linear regression testing. For CMV, n=9; all others, n=10.

Figure 5

VACV-specific memory CD4⁺ T cells have a higher frequency of CD45RA⁺ cells with long telomeres. (A) CD4⁺ T cell proliferative responses (BrdU⁺) gated by CD45RA^+/−from 1 subject. Histograms of TL in CD45RA^- and CD45RA⁺ cells are overlaid, and mean and median fluorescence intensities for each population are shown. (B) Mean and median telomere probe fluorescence intensity values for all subjects’ grouped by virus. Box and whisker plots indicate minimum to maximum values with the lines inside the boxes depicting the median values for each set of measurements. (C) Frequencies of BrdU⁺ CD45RA⁺ CD4⁺ T cells with long telomeres in one subject. (D) Frequency of CD45RA⁺ CD4⁺ T cells with long telomere within the BrdU+ population for 10 subjects. * p < 0.05 by Wilcoxon paired, signed rank test. Fluorescence is shown in arbitrary units (AU).

Figure 6

Longitudinal analysis of TL in CD4⁺ virus-specific memory T cells. (A) TL was measured in CD4⁺ BrdU^- CD45RA⁺ (dotted lines) and CD4⁺ BrdU+ T cells (solid lines) from five healthy subjects. PBMC samples were obtained 8 to 10 years apart. (B) The data from A were grouped according to different virus-specific T cell populations for all 5 subjects. Dotted lines are derived from the average slopes and y-intercepts. (C) Average TL kinetic of each virus-specific T cell population and naïve T cell average TL line from each plot in B is presented as single plot to allow comparisons of the average T cell TL kinetics in this cohort of healthy adults.

Figure 7

Reactivation associated with increased proliferative responses and restored TL in VZV-specific CD4⁺ and CD8⁺ T cells. Dot plots show virus-specific (A) CD4⁺ and (B) CD8⁺ proliferative responses across three time points in the same subject. BrdU+ frequencies (percent) are shown in the gate. PBMC were collected approximately two years prior (top row), two weeks after (middle row) or fourteen months after (bottom row) VZV reactivation. Graphs depict mean TL in (C) CD4⁺ and (D) CD8⁺ T cells along with TL in naïve T cells (from media-only culture BrdU^-CD45RA⁺) (dashed line) across the three time points. (E) VZV-specific CD4⁺ and CD8⁺ T cell from A and B are further delineated by CD45RA expression. (F) Mean TL for VZV-specific T cells by CD45RA expression shown in 4E. Error bars are standard errors from triplicate hybridizations of the same sample. By unpaired t-test, *** p < 0.001, ** p < 0.01, ns = not different.