.
ABSTRACT
Aging leads to a progressive functional decline of the immune system, which renders the elderly increasingly susceptible to disease and infection. The degree to which immune cell senescence contributes to this functional decline, however, remains unclear since methods to accurately identify and isolate senescent immune cells are missing. By measuring senescence-associated ß-galactosidase activity, a hallmark of senescent cells, we demonstrate here that healthy humans develop senescent T lymphocytes in peripheral blood with advancing age. Particularly senescent CD8+ T cells increased in abundance with age, ranging from 30% of the total CD8+ T cell population in donors in their 20s and reaching levels of 64% in donors in their 60s. Senescent CD8+ T cell populations displayed features of telomere dysfunction-induced senescence as well as p16-mediated senescence, developed in various T cell differentiation states and established gene expression signatures consistent with the senescence state observed in other cell types. On the basis of our results we propose that cellular senescence of T lymphocytes is a major contributing factor to the observed decline of immune cell function with advancing age and that immune cell senescence, therefore, plays a significant role in the increased susceptibility of the elderly to age-associated diseases and infection.
INTRODUCTION
Cellular senescence is a stable proliferative arrest that is encountered by mammalian cells in response to a variety of signals and stresses (1). In mammals, this response functions to suppress cancer development and plays also important roles during tissue repair, wound healing and embryonic development (2). Although senescent cells (SCs) can be cleared from tissue by adaptive and innate immune responses under these circumstances, not all SCs are cleared and consequently accumulate progressively in various tissues during aging (3). In mouse models, this age-associated accumulation of SCs has a substantial negative impact on fitness and health, as it shortens lifespan and contributes to the development of numerous age-associated diseases (4, 5). While it is currently unclear why some SCs evade immune cell clearance and accumulate in tissues, it is possible that immune cells themselves progressively become senescent with age, thereby increasingly weakening immune responses that would otherwise clear SCs from aged tissue and suppress infections from viruses and other pathogens (3). This hypothesis, however, has proven challenging to test, primarily due to the lack of a suitable marker that can identify senescent mammalian immune cells accurately and efficiently (6, 7).
In mammalian cells, at the least two pathways activate the senescence program. One pathway, mediated by the tumor suppressor p53 and cyclin-dependent kinase (CDK) inhibitor p21, primarily becomes activated in response to a persistent DNA damage response (DDR), such as the one caused due to telomere dysfunction. In normal somatic human cells, or cells that lack detectable telomerase activity, dysfunctional telomeres are generated not only due to repeated cell division cycles that cause progressive telomere erosion, but also as a consequence of genotoxic stresses that cause double stranded DNA breaks (DSBs) in telomeric repeats (8, 9). A second senescence pathway is activated due to upregulation of the CDK4/6 inhibitor p16INK4a, which results in a stable pRb-dependent cell cycle arrest. While genotoxic stresses can activate this senescence pathway in certain cell types, p16INK4a is also upregulated in the absence of telomere dysfunction or a persistent DDR (10, 11). Although the molecular triggers of this DDR independent senescence response are still largely unclear, p16INK4a mediated-senescence clearly has important physiological consequences, as it significantly contributes to aging and the development of age-related disorders in mammals (5).
Despite differences in pathway activation, SCs share a number of features. One characteristic that is common to all SCs is that they secrete numerous NF-kB and p38-MAPK regulated pro-inflammatory cytokines and other molecules, collectively called the senescence associated secretory phenotype or SASP (12). Although the SASP differs in composition depending on cell type, senescence-inducing signal, and time elapsed following senescence induction, a primary function of the SASP is to generate a pro-inflammatory environment that stimulates an immune response(3). Another feature common to SCs is that they up-regulate certain heterochromatin proteins, such as macroH2A, leading to a stable repression of cell proliferation genes (13, 14). In addition, SCs develop a greater abundance of lysosomal content and a reduction in lysosomal pH (15), resulting in increased expression of lysosomal ß-galactosidase. This hallmark of SCs in particular allows their detection in cultures and in tissue, regardless of the senescence-inducing signal or senescence pathway activated (16).
Our current understanding of cellular senescence stems primarily from studies conducted using mammalian fibroblast cultures. Senescence pathways in other cell types, including those of circulating peripheral blood mononuclear cells (PBMCs) CD4+ T cells, CD8+ T cells, monocytes, B cells natural killer (NK) cells, and plasmacytoid dendritic cells (pDC’s), are still incompletely understood (17). Although subsets of PBMCs, such as cytotoxic CD8+ T cells undergo replicative senescence in culture, whether and to what degree they do so also in vivo remains unclear (7, 18). A primary reason for this uncertainty is that specific markers used to identify senescent CD8+ T cells in the past (19, 20), such as a loss of the cell surface receptors CD28 and CD27 and a gain of expression of CD45RA, CD57, TIGIT and/or KLRG1 do not accurately characterize all T cells that have permanently lost the ability to proliferate due to acquisition of macromolecular damage, upregulation of cyclin-dependent kinase (CDK) inhibitors, and development of senescence associated ß-Galactosidase (SA-ßGal), criteria that define the state of cellular senescence (16). In fact, CD8+ T cells that have lost expression of CD28 and/or that display varying levels of CD45RA, CD57, TIGIT or KLRG1 maintain the ability to proliferate following appropriate stimulation, which is incompatible with a classical senescence response (18, 19).
Here, we describe an accurate and efficient method to quantify, isolate, and characterize live senescent immune cells from peripheral blood of human donors. We reveal the identity of PBMC subsets that increasingly undergo cellular senescence in healthy humans with age, uncover causes for cellular senescence in circulating CD8+ T cells, and characterize the pathways activated in senescent CD8+ T cells at levels that may provide insights into therapeutic opportunities to modulate T cell senescence in disease, infection, and advanced age.
RESULTS
A method to isolate live SCs for subsequent analysis
One feature that is common to SCs is that they develop increased expression levels SA-ßGal (16). This hallmark of SCs allows their detection using chromogenic (X-Gal) (21) or fluorogenic (FDG) (22) ßGal substrates, albeit with limitations due to poor cell permeability or lack of cellular retention, respectively. To mitigate these limitations, we tested a cell permeable and self-immobilizing fluorogenic SA-ßGal substrate (fSA-ßGal) for its ability to label live SCs for prolonged periods, so that they can be accurately analyzed, quantified, and isolated by flow cytometry. As anticipated, incubation with fSA-ßGal caused GM21 fibroblasts in oncogene-induced senescence (OIS), either alone or mixed at a 3:1 ratio with non-senescent GM21 fibroblasts, to develop significantly increased fluorescence signal intensities compared to proliferating non-senescent control fibroblasts (Fig. S1A). Live cells in OIS could be accurately sorted and isolated from non-SCs by FACS, allowing us to conduct cell proliferation assays, immunofluorescence analysis, and gene expression profiling. We show that FACS isolated cells with high SA-ßGal activity (fSA-ßGal high) were significantly impaired in their ability to proliferate, suppressed expression of cyclin A, and displayed increased expression levels of senescence genes p21, p16INK4a, IL1B, and IL8 compared to cells with low activity (fSA-ßGal low; Fig. S1B-E). Similar results were obtained when comparing etoposide-treated fibroblasts undergoing DNA damage-induced senescence with proliferating non-senescent fibroblasts (Fig. S1F-H). Collectively, these results demonstrate that fSA-ßGal is specific to labeling live human SCs and efficient for isolating them from mixed cell populations for subsequent analysis.
Subsets of human PBMCs increasingly develop high SA-ßGal activity with advancing age
To determine whether subsets of human peripheral blood mononuclear cells (PBMCs) undergo cellular senescence as a function of age in circulation, we collected and analyzed blood from healthy human donors in two age groups: 1) “young”, which included donors between the ages 23 and 30 years (average age 25 years; 20s), and 2) “old” which included donors between the ages 57 and 67 years (average age 64 years; 60s), around the age the human immune system is found to exhibit age-associated deficits. Young and old donors were always recruited in pairs, allowing us to process and analyze freshly isolated PBMCs from donor pairs in parallel (Fig. 1A and Fig. S2A). Cord blood, together with blood from healthy donors in their 20s, was also collected and analyzed for the presence of SCs. Surprisingly, age-associated increases of mean fluorescence intensities (MFIs) of SA-ßGal signals were observed in all subsets analyzed for some donor pairs, including T lymphocytes, plasmacytoid dendritic cells (pDC’s), natural killer (NK) cells, monocytes, and B cells (Fig. 1B-C and Fig. S2B). Our data therefore suggest that all subsets of human PBMCs can undergo cellular senescence in vivo. Most consistent age-associated increases in mean fluorescence intensities (MFIs), however, were observed only for T cells (Fig. S2B-C). In order to determine which PBMC subsets increasingly develop high SA-ßGal activity with advancing age, we quantified the percentages of cells that display increased fSA-ßGal signal intensities by taking advantage of the fact that young donors, and often also old donors, had distinct populations of cells with lower and higher fSA-ßGal fluorescence. This allowed us to set gates at intersections where the two populations met, designating the population with lower signal intensity as “fSA-ßGal low” and the population with higher signal intensity as “fSA-ßGal high” (Fig. 1B). Quantitative analysis of fSA-ßGal high cells revealed statistically significant age-associated increases in human T lymphocytes, particularly in CD8+ T cells. For this subset, we discovered a striking age-associated increase of cells with high SA-ßGal activity, ranging from 30 ± 3 % in blood from donors in their 20s to 64 ± 4 % in donors in their 60s (Fig. 2C; p<0.0001). CD4+ T cells also displayed a significant age associated increase of cells with high fSA-ßGal activity, albeit to a lesser degree compared to CD8+ T cells (15 ± 3 % in young and 31 ± 6 % in old donors, p=0.02). No gender differences in the percent of fSA-ßGal high cells were observed in either T cell population (not shown). Not surprisingly, percentages of cells with high SA-ßGal activity were the lowest in cord blood, including in T lymphocytes, which contained fewer that 5% of this SA-ßGal expressing population (Fig. 1C and Fig. S2D-E). Our data therefore demonstrate that an increasing fraction of T lymphocytes, in particular CD8+ T cells, develop high SA-ßGal activity with age.
Fig. I Humans display increased percentages of T lymphocytes with high fSA-ßGal signal intensities in advanced age. (A) Experimental strategy to quantify, isolate, and characterize senescent subsets of PBMCs from donors in their 20s (young) and 60s (old). IF: immunofluorescence analysis. (B) Representative fSA-ßGal intensity profiles and gates used to quantify fSA-ßGal high cells for indicated PBMC subsets from a young (blue) and old (red) donor. © Quantification of the percentages of fSA-ßGal high cells in cord blood (CB), young (blue) and old (red) donors for indicated PBMC subsets. Whiskers indicate mean +/- S.E.M and are indicated for each subset. Statistical significance was determined by an unpaired, two-tailed Student’s t test.*** p<0.0001; ** p = 0.0005; * p < 0.05; NS: not significant.
CD8+ T cells with high levels of SA-ßGal activity display features of telomere dysfunction-induced senescence (TDIS) and p16INK4a-mediated senescence
To test whether CD8+T cells with high SA-ßGal activity are senescent, we sorted and collected them by FACS based on low, intermediate, or high fSA-ßGal signal intensities (Fig. 2A) and measured cell proliferation, senescence gene expression, and development of features specific to SCs. Significantly, the ability of CD8+T cells to proliferate following activation was inversely proportional to fSA-ßGal signal intensities. While 91 ± 1% of cells with low SA-ßGal activity were able to undergo more than one cell division during a 5 day period of stimulation, the fractions of proliferating cells with intermediate and high fSA-ßGal signal intensities were reduced to 71% ± 6% and 33 ± 6%, respectively, under the same conditions (Fig. 2B). No significant differences in proliferation capacities were observed between T cells collected from young or old donors at respective fSA-ßGal signal intensities (Fig. 3A). Gene expression analysis by RT-qPCR demonstrated significantly increased expression levels of senescence genes p16INK4a and p21, in cells with high SA-ßGal activity compared to cells with intermediate or low activities (Fig. 2C), which is consistent with a senescence response. Levels of IL6 mRNA, however, decreased with SA-ßGal activity, which is unlike senescent human fibroblasts (12). In addition, immunofluorescence analysis revealed that protein levels of p16INK4a and the senescence marker macroH2A were the lowest in CD8+ T cells with low SA-ßGal activity and increased proportionally in cells with intermediate and high SA-ßGal activities (Fig. 2D and Fig. S3B-C). Similarly, cells that displayed multiple foci of 53BP1, a DDR factor that localizes to sites of DSBs, were less frequently detected in cells with low SA-ßGal activity, compared to cells with intermediate or high activities, irrespective of donor age (Fig. 2D-E and Fig. S3D-F). Approximately 40% of DDR foci analyzed were localized at telomeric repeats, revealing that a substantial fraction of DSBs in CD8+ T cells are caused as a result of telomere dysfunction (Fig. 2F and Fig. S3G). Overall, the mean number of telomere dysfunction-induced DNA damage foci (TIF) per cell increased proportionally with increasing SA-ßGal activity, irrespective of donor age, demonstrating a direct correlation between the senescence state and presence of dysfunctional telomeres in CD8+ T cells (Fig. 2G and Fig. S3G-H). Our data therefore demonstrate that humans increasingly develop senescent CD8+ T cells with features of TDIS and p16INK4a-mediated senescence with age.
Fig. 2 CD8+ T cells with high fSA-ßGal signal intensities are senescent. (A) Dot plot illustrating gates used to sort and isolate CD8+ T cells by FACS from young and old donors, as indicated, based on low, inter(med)iate, and high fSA-ßGal signal intensities. (B) Representative Cell Trace Violet histograms of CD8+ T cells, sorted as in (a), following anti-CD3 and anti-CD28 stimulation for 5 days from a young and old donor. Bar graph, quantification of more than one cell divisions of CD8+ T cells from 6 healthy donors (3 young, 3 old) following polyclonal stimulation for 5 days. n = 3; * p = 0.0002; ** p < 0.0001. © RT-qPCR expression profiles for indicated senescence-associated genes in CD8+ T cells at indicated fSA-ßGal levels in young (blue) and old donors (red). Average and S.E.M. of five (CDKN1A and CDKN2A) and three (IL6) independent experiments. Statistical significance was determined by a two-tailed unpaired t-test, with Y-low as the reference value. * p < 0.05; ** p< 0.001. (D) Immunofluorescence analysis of sorted CD8+ T cells from a young and old donor, as indicated, using antibodies against 53BP1 (green) and p16INK4a (red). Outline of cell nuclei is indicated with white border. Scale bar: 10 □m (E) Quantification of the percentage of CD8+ T cells positive for 53BP1 foci at each fSA-ßGal level in young and old donors combined. Young: n = 8, Old: n = 8. Whiskers depict mean +/- S.E.M. Statistical significance was determined by a one-way ANOVA. ** p = 0.001. (F) Sorted CD8+ T cells with high fSA-ßGal signal intensities were simultaneously immunostained using antibodies against 53BP1 (green) and analyzed by FISH to detect telomeres (red). Blue: DAPI. Enlarged versions of the numbered DNA damage foci showing colocalization with telomeres are shown in the right micrographs. Scale bar: 10 □m (G) Quantification of mean TIF per cell in sorted CD8+ T cells, as indicated, in young and old donors combined. Young: n = 5, Old: n = 5. Whiskers depict mean +/- S.E.M. Statistical significance was calculated by a one-way ANOVA. * P = 0.0280.
.../...
.
