.
F U L L T E X T S O U R C E : Aging Cell
Abstract
The susceptibility of human CD4+ and CD8+ T cells to senesce differs, with CD8+ T cells acquiring an immunosenescent phenotype faster than the CD4+ T cell compartment. We show here that it is the inherent difference in mitochondrial content that drives this phenotype, with senescent human CD4+ T cells displaying a higher mitochondrial mass. The loss of mitochondria in the senescent human CD8+ T cells has knock‐on consequences for nutrient usage, metabolism and function. Senescent CD4+ T cells uptake more lipid and glucose than their CD8+ counterparts, leading to a greater metabolic versatility engaging either an oxidative or a glycolytic metabolism. The enhanced metabolic advantage of senescent CD4+ T cells allows for more proliferation and migration than observed in the senescent CD8+ subset. Mitochondrial dysfunction has been linked to both cellular senescence and aging; however, it is still unclear whether mitochondria play a causal role in senescence. Our data show that reducing mitochondrial function in human CD4+ T cells, through the addition of low‐dose rotenone, causes the generation of a CD4+ T cell with a CD8+‐like phenotype. Therefore, we wish to propose that it is the inherent metabolic stability that governs the susceptibility to an immunosenescent phenotype.
1 INTRODUCTION
The human immune system functionality declines with age in a process referred to as immunosenescence. The functional outcomes of this process include the compromised ability of older individuals to mount protective immune responses against both previously encountered and new pathogens (Akbar, Henson, & Lanna, 2016). Additionally, there is a marked decrease in vaccine efficacy in these populations. While these age‐associated alterations arise from defects in different leucocyte populations, the dysfunction is most profound in T cell subsets (Akbar et al., 2016). Furthermore, aging is associated with a chronic low‐grade inflammatory state, termed inflammaging (Franceschi et al., 2000), and mediates an important role in a range of age‐related degenerative pathologies (Baker et al., 2011). The source of this inflammation has yet to be defined. Senescent T cells are found to accumulate with age and represent a likely contributor to this inflammatory state that is observed during aging (Akbar et al., 2016).
Primary human senescent T cells are a highly differentiated subset of cells found within the CD27−CD28− population (Parish, Wu, & Effros, 2010). This subset can be further characterized on the basis of CD45RA expression, with highly differentiated T cells that re‐express CD45RA identified as the senescent T cell population (EMRA; effector memory CD45RA re‐expressing T cells). They display multiple characteristics of senescence including a low proliferative activity, high levels of DNA damage and loss of telomerase activity (Henson et al., 2014). However, the response patterns of CD4+ and CD8+ T cells to aging differ, with CD8+ T cells being more susceptible to both phenotypic and functional changes during aging (Czesnikiewicz‐Guzik et al., 2008). The CD8+ EMRA T cell subset accumulates in higher proportions with age and is more prevalent following in vitro culture than the CD4+ EMRAs (Czesnikiewicz‐Guzik et al., 2008). The cause of this difference has been suggested to be due to the differing homeostatic mechanisms and an increased gene expression instability of regulatory cell surface molecules in the CD8+ EMRA subset (Czesnikiewicz‐Guzik et al., 2008). We would like to postulate an alternate view that the metabolic versatility seen in CD4+ T cells confers a metabolic advantage to the CD4 EMRA subset allowing these T cells to better withstand the intrinsic or extrinsic effects governing differentiation.
Metabolic examination of CD4+ and CD8+ T cells suggests that their metabolic programming allows differential immunological functions to be performed. We demonstrate here that CD4+ T cells have a greater mitochondrial mass and are consistently more oxidative than CD8+ T cells, allowing them to sustain effector function. Whereas the metabolic programs that prioritize rapid biosynthesis such as glycolysis are favoured by CD8+ T cells, allowing for faster growth and proliferative rates (Cao, Rathmell, & Macintyre, 2014). We have previously shown that CD8+ EMRA T cells display impaired mitochondrial function (Henson et al., 2014) but are still unclear as to whether CD4+ EMRA T cells also exhibit mitochondrial dysfunction. We provide evidence that this is not the case, and CD4+ EMRA T cells have fitter, healthier mitochondria that are better able to meet the energy requirements of the CD4+ EMRA subset. Therefore, we propose that it is the inherent metabolic stability that governs the susceptibility to an immunosenescent phenotype.
2 RESULTS
2.1 Human CD4+ EMRA T cell development at a slower rate due to their higher mitochondrial content
Human T cells can be subdivided into four populations on the basis of their relative surface expression of CD45RA and CD27 molecules (Figure S1a). The four subsets are defined as naïve (N; CD45RA+CD27+), central memory (CM; CD45RA−CD27+), effector memory (EM; CD45RA−CD27−) and effector memory T cells that re‐express CD45RA (EMRA; CD45RA+CD27−). We and others have demonstrated that the EMRA population exhibit numerous characteristics of senescence (Appay, Lier, Sallusto, & Roederer, 2008; Di Mitri et al., 2011; Henson et al., 2014); indeed, it has also been known for some time that CD4+ EMRA T cells senesce at a slow rate than their CD8+ counterparts (Figure 1a). It was thought that the difference in EMRA accumulation was due to their differing cytokine stabilities (Czesnikiewicz‐Guzik et al., 2008); however, we now demonstrate that it is a difference in mitochondrial mass between CD4+ and CD8+ EMRAs that governs the rate at which they develop.
Figure 1. Human CD4+ EMRA T cells are acquired at a slower rate owing to a higher degree of mitochondrial content. (a) The accumulation of senescent CD4+ and CD8+ T cells with age defined by the markers CD45RA and CD27. (b) Representative flow cytometry plots from middle‐aged donors and cumulative graphs of MitoTracker Green staining in CD4+ and CD8+ EMRA T cells analysed directly ex vivo. Data expressed as mean ± SEM of six donors. © Electron microscope images of CD4+ and CD8+ EMRA T cells imaged directly ex vivo from middle‐aged donors. Yellow arrows mark mitochondria. Graph shows the percentage by cell volume of mitochondria in senescent T cell subsets determined by a point‐counting grid method from 20 different electron microscope images. (d) PGC1α expression in CD45RA/CD27‐defined EMRA T cell subsets from middle‐aged donors. Data expressed as mean ± SEM of nine donors. p‐values were calculated using a t test. ** p < .01
Using MitoTracker Green, a mitochondrial‐specific dye that binds the mitochondrial membranes independently of mitochondrial membrane potential (MMP), we found the CD4+ EMRA subset isolated from middle‐aged donors (av. age 41 years ± 5) to have a significantly higher mitochondrial mass than CD8+ EMRAs, nearly double the amount of mitochondrial content (Figure 1b). The CD4+ EMRA subset retains their mitochondrial content compared to earlier less differentiated subsets (Figure S2a), whereas the CD8+ EMRAs do not (Henson et al., 2014). This was also borne out when the EMRA subsets were examined ex vivo by electron microscopy. We observed significantly fewer mitochondrial in the CD8+ EMRA compartment when compared to the CD4+ EMRA fraction using a point‐counting method (Figure 1c). Furthermore, when we investigated the expression of PGC1α (peroxisome proliferator‐activated receptor gamma coactivator 1‐alpha), the key regulator of mitochondrial biogenesis, the CD4+ EMRA subset showed significantly higher ex vivo levels of this marker than the CD8+ EMRAs (Figure 1d). This phenomenon was found to be independent of chronological age, as the mitochondrial content of CD4+ and CD8+ EMRA T cells isolated from older individuals (av. age 71 ± 3) was the same as that of younger individuals (Figure S2b,c). Collectively, our results demonstrate that senescent CD4+ T cells have increased mitochondrial mass in comparison with their CD8+ counterparts.
2.2 Distinct mitochondrial functions in CD4+ and CD8+ EMRA subsets
The increased mitochondrial mass seen in the CD4+ EMRA subsets suggests they may exhibit distinct mitochondrial functions compared to the CD8+ EMRAs. Indeed, using TMRE, which measures mitochondrial transmembrane potential, we found the CD4+ EMRAs had a higher proportion of hyperpolarized mitochondria than the CD8+ EMRA subset, which displayed a hypopolarized phenotype (Figures 2a and S3a). The mitochondrial membrane potential provides the charge gradient required for Ca2+ sequestration and the regulation of reactive oxygen species (ROS) production. Cell stress causes a dysregulation in the mitochondrial membrane potential, with hyperpolarization resulting in the production of excess ROS leading to oxidative stress. While a state of hypopolarization is also harmful, as low amounts of ROS cause reductive stress, which is as detrimental to homeostasis as oxidative stress (Zorova et al., 2018).
Figure 2. Mitochondrial dysfunction is observed in CD8+ but not CD4+ EMRA T cell subsets. (a) Representative flow cytometry plots and cumulative graphs of TMRE staining from middle‐aged donors showing membrane potential in CD45RA/CD27 T cell subsets directly ex vivo defined showing the percentage of cortactin‐positive (a) CD4+ and (b) CD8+ T cells analysed directly ex vivo. Data expressed as mean ± SEM of six donors. (b) Mitochondrial ROS measured using MitoSOX by flow cytometry in CD4+ and CD8+ EMRA T cells from middle‐aged donors. Data expressed as mean ± SEM of six donors. © Mitochondrial ROS production expressed as a ratio of mitochondrial mass. Calculated from data shown in Figures 1b and 2. (d) γH2AX expression as determined by flow cytometry in CD45RA/CD27‐defined T cell subsets directly ex vivo from middle‐aged donors; the graph shows the mean ± SEM for five donors. (e) Oxygen consumption rates (OCR) of the EMRA CD4+ and CD8+ T cell subsets from middle‐aged donors were measured following a 15‐min stimulation with 0.5 µg/ml anti‐CD3 and 5 ng/ml IL‐2; the cells were then subjected to a metabolic stress test using the indicated mitochondrial inhibitors. Data are representative of four independent experiments. (f) The basal OCR, extracellular acidification rate (ECAR) and spare respiratory capacity were measured following a 15‐min stimulation with 0.5 µg/ml anti‐CD3 and 5 ng/ml IL‐2. Graphs show the mean ± SEM for four donors. (g) ATP concentration in EMRA T cell subsets from middle‐aged donors, graphs show the mean ± SEM for five donors. p‐values were calculated using a t test. *p < .05, **p < .01, and ***p < .005
As hyperpolarized mitochondria can be a source of ROS that can potentiate senescence (Cui, Kong, & Zhang, 2012), we next measured mitochondrial ROS production using MitoSOX. We found the amount of ROS to be significantly higher in CD4+ EMRAs (Figures 2b and S3b); again, the observed differences were found to be independent of donor age (Figure S3c,d). However, both EMRA subsets produced significantly more ROS than their less differentiated counterparts (data not shown). However, the increased ROS production seen in CD4+ EMRAs was neutralized owing to the higher mitochondrial mass, meaning that CD8+ EMRA T cells produced more ROS per mitochondria that can potentially enhance their senescent phenotype (Figure 2c). Furthermore, increased ROS can also cause DNA damage and the activation of the DNA damage response, elevated during senescence. The examination of phosphorylated H2AX (γH2AX), a member of the histone H2A family that is part of the DNA damage response, in EMRAs revealed that the CD8+ EMRA subset displayed a higher level of this marker compared to CD4+ EMRAs (Figure 2d). However, both EMRA subsets express the highest amount of DNA damage compared to their less differentiated subsets (data not shown). We suggest that a loss of mitochondrial mass in the CD8+ EMRA subsets is a key mediator in generating an enhanced senescent state.
We then examined differences in mitochondrial respiration between the CD4+ and CD8+ EMRA subsets. Differences were found in both the baseline respiration and respiration following injection of oligomycin, FCCP and rotenone and antimycin A (Figure 2e). The CD4+ EMRA population retain their ability to respond to challenge akin to the other CD4+ memory subsets (Figure S3e), while we have shown this not to be the case for the CD8+ EMRA compartment (Henson et al., 2014). The oxygen consumption rate (OCR), together with the spare respiratory capacity, the potential amount of stored energy a cell has to respond to challenge were both upregulated in CD4+ EMRAs, further implying a difference in mitochondrial content. While the extracellular acidification rate (ECAR), a marker of lactic acid production and glycolysis, was only marginally increased compared to the CD8+ EMRA subset (Figure 2f), furthermore, the amount of ATP made by CD4+ EMRAs was also greater than that of the CD8+s (Figure 2g). These results suggest that the CD4+ EMRA subset has enhanced mitochondrial fitness that allows for a greater flexibility in the type of metabolism they can engage.
2.3 CD8+ EMRA T cells display impaired nutrient uptake
T cells utilize a variety of energy sources including glucose and lipids; however, their metabolic preferences are governed not only by their differentiation status but also by mitochondria fitness (Cui et al., 2012). Indeed, a lack of regulatory control over nutrient usage is a recurrent theme accompanying senescence and aging (Brewer, Gibbs, & Smith, 2016). We therefore sort to determine whether there were differences in glucose and fatty acid uptake in CD4+ and CD8+ EMRA T cell subsets. CD4+ EMRAs were found to take up more of the fluorescent glucose analogue 2‐NBDG from their extracellular environment than their CD8+ EMRA counterparts (Figure 3a). This was found to be independent of the age of the donor (Figure S4a). Indeed, CD4+ EMRAs showed higher expression of glut1, the major glucose transporter in T cells using an RNA‐labelled probe (Figure 3b). Analysis of microarray data revealed high expression of alternate glut family members (Callender et al., 2018). Interestingly, CD8+ EMRAs displayed a higher expression of the class III glucose transporters glut8 and glut10 (Figure 3b). Both these transporters are found intracellularly and are thought to transport glucose or galactose across intracellular organelle membranes (Mueckler & Thorens, 2013). The uptake of fluorescently labelled palmitate, BODIPY C16, a long‐chain fatty acid, was also quantified. Similar to our observations for glucose uptake, CD4+ EMRA T cells also utilize significantly more palmitate than their CD8+ counterparts (Figure 3c), again independent of the chronological age of the donor (Figure S4b). Furthermore, CD4+ EMRA T cells express higher levels of both the fatty acid translocase CD36 and the fatty acid transporters FATP2 and FATP3 (Figure 3d). Taken together, these results suggest that the increased mitochondrial fitness of CD4+ EMRA T cells enables these cells to better utilize glucose and lipids, which may limit the impact of senescence.
Figure 3. Impaired nutrient uptake by CD8+ EMRA T cells. (a) Glucose uptake was assessed using the fluorescent glucose analogue 2‐NBDG in CD4+ and CD8+ T CD45RA/CD27‐defined EMRA T cells from middle‐aged donors by flow cytometry following a 15‐min incubation. Data expressed as mean ± SEM of seven donors. (b) Examples and data showing expression of the glucose transporters glut1, glut8 and glut10 in senescent T cell subsets directly ex vivo from middle‐aged donors. Graphs show the mean ± SEM for four donors. © Lipid uptake was measured using fluorescently labelled palmitate, BODIPY C16 by flow cytometry following a 15‐min incubation in CD4+ and CD8+ EMRA T cells from middle‐aged donors. Data expressed as mean ± SEM of seven donors. (d) Examples and graphs showing the fatty acid translocase CD36 and FATP2 and −3 directly ex vivo from middle‐aged donors. Data expressed as mean ± SEM of six donors. p‐values were calculated using a t test. *p < .05, **p < .01, and ***p < .005
2.4 Impaired proliferation and migration of CD8+ EMRA T cells
The end result of the DNA damage response is the activation of p53. p53 regulates cell cycle arrest limiting cell growth and proliferation, as well as playing a crucial role in limiting cell motility, a critical process for optimal T cell function (Muller, Vousden, & Norman, 2011). In line with the theory that the acquisition of the CD4+ EMRA T cell subset occurs at a slower rate than their CD8+ counterpart, we find that the expression of p‐p53 is higher in the CD8+ EMRAs compared to the CD4+s (Figure 4a), although the expression of p‐p53 in the CD4+ EMRAs is the highest of all the CD4+ memory subsets (data not shown). Furthermore, the proliferative defect is more pronounced in the CD8+ EMRA subset, measured using ki67 (Figure 4b) and migration impaired (Figure 4c). Transwell chemotactic assays were used to assess migration; HUVECs were activated using 20% autologous donor sera, in order to create a more appropriate ex vivo environment, and were found to be no different to activation with IFNγ (Figure S4c). Migration was assessed in response to CXCL10 and CXCL12, chemokines promoting the migration of memory T cells or 20% autologous serum, and the data expressed as a percentage of the total CD4+ or CD8+ T cells found to have migrated. CD4+ EMRA T cells were less able to respond to CXCL10 and CXCL12 than autologous serum, presumably due to the loss of CXCR3 and CXCR4 from the CD4+ EMRA T cells (Brainard et al., 2007; Hess et al., 2004). CD8+ EMRAs on the other hand retain expression of CXCR3 and CXCR4 and migrate in response to both chemokines and autologous serum, all be it to a lesser extent than their CD4+ counterparts (Figure 4c). This was once again not dependent on the chronological age of the donors (Figure S4d).
.../...
