Orai, STIM, and PMCA contribute to reduced calcium signal generation in CD8+ T cells of elderly mice

Ca2+ is a crucial second messenger for proper T cell function. Considering the relevance of Ca2+ signals for T cell functionality it is surprising that no mechanistic insights into T cell Ca2+ signals from elderly individuals are reported. The main Ca2+ entry mechanism in T cells are STIM-activated Orai channels. Their role during lymphocyte aging is completely unknown. Here, we report not only reduced Ca2+ signals in untouched and stimulated, but also in central and effector memory CD8+ T cells from elderly (18-24 months) compared to adult (3-6 months) mice. Two mechanisms contribute to the overall reduction in Ca2+ signals of CD8+ T cells of elderly mice: 1) Reduced Ca2+ currents through Orai channels due to decreased expressions of STIMs and Orais. 2) A faster extrusion of Ca2+ owing to an increased expression of PMCA4. The reduced Ca2+ signals correlated with a resistance of the cytotoxic efficiency of CD8+ T cells to varying free [Ca2+]ext with age. In summary, reduced STIM/Orai expression and increased Ca2+ clearing rates following enhanced PMCA4 expression contribute to reduced Ca2+ signals in CD8+ T cells of elderly mice. These changes are apparently relevant to immune function as they reduce the Ca2+ dependency of CTL cytotoxicity.


INTRODUCTION
Aging of the immune system is a cumulative phenomenon and contributes to morbidity and mortality in man due to the greater incidence of infection, as well as autoimmune phenomena and cancer in elderly individuals [1][2][3]. Multiple changes in T cell populations are considered to be critical contributors to age-associated immune dysfunction [4]. A rather unexplored field of this dysfunction in aging is the Ca 2+ homeostasis, which especially in T cells, is tightly regulated by an orchestra of channels, pumps, transporters, and receptors [5,6]. The increase in cyto-plasmic free [Ca 2+ ] following T cell receptor activation is an essential element of various activation pathways [7] and necessary for the induction of gene expression in CD8 + T cells [8][9][10][11]. In addition, Ca 2+ orchestrates the cytotoxicity of CD8 + T cells as the main function of this cell type [12]. For murine and human immune cells, store operated Ca 2+ entry (SOCE) is the main pathway for Ca 2+ to enter T cells [6,13,14]. Highly Ca 2+selective Orai channels and STIM proteins, as Ca 2+ sensors, are the key players in SOCE and with their distinct characteristic properties [15,16] they shape the Ca 2+ pattern in T cells [17]. In contrast to the human SOCE components, four different Orai (Orai1, Orai2S, AGING Orai2L, Orai3) proteins [18] and four STIM (STIM1, STIM1L, STIM2.1 and STIM2.2) proteins [19][20][21] have been identified in the mouse so far, but their role during aging in T cells is completely unknown.
While STIM/Orai channels are the likely candidates to mediate Ca 2+ entry in aging T cells, other mechanisms could in principle contribute to store-operated Ca 2+ signals: K + channels (like Kv1.3 and KCa3.1) and TRPM4 may control the membrane potential and thereby the net Ca 2+ entry through Orai channels [22,23]. Furthermore, Ca 2+ ATPases and mitochondria export Ca 2+ from the cytosol and thereby shape Oraidependent Ca 2+ signals [24].
Following T cell activation in mice, several groups have reported a decline in Ca 2+ signals with age [25,26]. However, since the molecular mechanisms of these decreased Ca 2+ signals in the context of aging are still elusive, we performed comprehensive assessments of the molecular repertoire governing Ca 2+ signals in CD8 + T cells and investigated the influence of varying [Ca 2+ ]ext on the main function of these cells, the lysis of target cells.

Elderly mice show an increased proportion of CD8 + T cells and a shift in subpopulations
First, we examined the health status of the mice used for experiments. For this, body weight and spleen weight were measured (Table 1). Splenomegaly is a frequently observed sign of non-specific necropsy [27], therefore we excluded those mice from our cohort. A shift in the distribution of T cell subtypes to less naïve and more memory cells is considered as a marker of immunosenescence [26,28,29]. To assess the CD8 + T cell subtype distribution of the cells we stained them with surface markers CD44 and CD62L to differentiate between naïve (N: CD62L high CD44 low ); effector memory (EM: CD62L low CD44 high ), and central memory CD8 + T cells (CM: CD62L high CD44 high ) (Supplementary Figure  1A, 1D). The untouched cells of adult mice display a pronounced population of CD62L high CD44 low naϊve cells, a lesser relative amount of CD62L high CD44 high central memory cells and only a small proportion of CD62L low CD44 high effector memory cells (Supplementary Figure 1A Figure 1C, 1F). Furthermore, the population of central memory cells showed a significant increase for the CD8 + T cells of adult and a significant decrease for the CD8 + T cells of elderly mice after stimulation. In summary, the CD8 + T cells of adult mice mainly consist of naϊve cells and shifts primarily to central and secondarily to effector memory cells, whereas the CD8 + T cell pool of elderly mice mostly consists of memory T cells and shifts to more effector memory cells with T cell activation.

Untouched CD8 + T cells from elderly mice show reduced SOCE
Ca 2+ influx is an essential step in T cell activation and regulation of diverse cellular functions and the main pathway of Ca 2+ in immune cells is through STIM-gated Orai channels [5]. We first investigated store-operated Ca 2+ responses of elderly and adult untouched CD8 + T cells activated by thapsigargin (TG), an irreversible inhibitor of the sarco/endoplasmic reticulum Ca 2+ -ATPase (SERCA) pumps. We measured Ca 2+ mobilization with two different approaches. The first approach combined the measurement of Ca 2+ release and entry (combined protocol, Figure 1A) in the presence of 0.5 mM [Ca 2+ ]ext, whereas the second one separates them ( Figure 1D) by carrying out the classical Ca 2+ re-addition protocol. In the latter protocol, cells are initially treated with TG in Ca 2+ -free Ringer's solution to deplete the endoplasmic reticulum (ER) Ca 2+ stores, and Ca 2+ influx is assured by subsequent addition of Ca 2+ . CD8 + T cells from elderly and adult mice differ in their ability to increase the internal calcium concentration ([Ca 2+ ]int) after TG-triggered stimulation in both approaches ( Figure 1A, 1D). Resting [Ca 2+ ]int in CD8 + T cells from both mice was not significantly changed in combined and re-addition protocol ( Figure  1A, 1D, Supplementary Tables 1, 2). Treatment of cells with TG in the absence of [Ca 2+ ]ext resulted in a transient rise in [Ca 2+ ]int revealing no significant change in the size of TG-releasable peak ( Figure 1D, Supplementary Table 2). Upon re-addition of [Ca 2+ ]ext but also with the combined Ca 2+ protocol, [Ca 2+ ]int mobilization was reduced in elderly CD8 + T cells by about 25 % (Figure 1A, 1B, 1D, 1E, Supplementary  Tables 1, 2). The rate of Ca 2+ entering CD8 + T cells is an indirect read-out for Orai channel function. For the re-addition protocol, the Ca 2+ entry rate was significantly slower in cells from elderly compared to adult mice ( Figure 1F); a similar tendency was observed in the combined protocol ( Figure 1C). Similar results were obtained in case [Ca 2+ ]ext was increased to 2 mM (Supplementary Figure 2, Supplementary  Tables 3, 4). Table 1. Spleen to body weight ratio of adult (n = 91) and elderly (n = 71) mice ± SEM.

Reduced ICRAC in untouched CD8 + T cells from elderly mice
Since mainly Ca 2+ release-activated Ca 2+ currents (ICRAC) through Orai channels are responsible for Ca 2+ influx in T cells, we assessed ICRAC in untouched CD8 + T cells from adult and elderly mice to determine agerelated changes ( Figure 2). We performed whole-cell patch-clamp experiments by perfusing cells with a high concentration of the Ca 2+ chelator BAPTA, and IP3 to deplete the Ca 2+ stores. Both adult and elderly CD8 + T cells immediately developed a small, inwardly rectifying current (ICRAC) with a reversal potential above +40 mV. Elderly CD8 + T cells showed reduced currents by 46 % (-1.89 ± 0.37 compared with -3.51 ± 0.36 (pA/pF) at 120 sec, p < 0.0001) (Figure 2A) with the typical inwardly rectifying current-voltage relationship (IV, Figure 2B).

Reduced store-operated Ca 2+ entry in in vitro stimulated CD8 + T cells from elderly mice
Since the negatively isolated and untouched cells do not have a long life span (24 hours) and, just like virusinfected or cancer cells, stimulation leads to activated CD8 + T cells, we decided to validate Ca 2+ signals in CD8 + T cells after in vitro stimulation. We therefore stimulated the CD8 + T cells with anti-CD3/CD28 stimulation beads and examined SOCE on day 3 after stimulation. The overall Ca 2+ signals analyzed in combined and re-addition protocols were reduced in stimulated CD8 + T cells between 60 to 64 % compared Ca 2+ Imaging with 1 µM TG as stimulus applied in the presence of 0.5 mM [Ca 2+ ]ext of CD8 + T cells (combined Ca 2+ protocol) from adult (black, n = 5) and elderly (red, n = 7) mice. The scatter dot plot in (B) displays the corresponding statistics of Ca 2+ influx peak and Ca 2+ plateau and in (C) the corresponding influx rates. (D) Ca 2+ Imaging with 1 µM TG applied in the absence of [Ca 2+ ]ext before re-addition of 0.5 mM Ca 2+ (readdition protocol) of CD8 + T cells from adult (black, n = 5) and elderly (red, n = 7) mice. The scatter dot plot in (E) displays the corresponding statistics of Ca 2+ influx peak and Ca 2+ plateau and (F) the corresponding influx rates. Ca 2+ signalling curves are presented as mean ± SEM. Scatter dot plots are presented as mean ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001. . This suggested that the molecular composition of the CRAC channel and STIM sensors may change during T cell stimulation. Still, TGinduced SOCE, measured as a peak of the Ca 2+ response was significantly reduced in stimulated elderly CD8 + T cells compared to adult as control ( Figure 3B, 3E). Besides the peak, also the Ca 2+ plateau, as an important determinant of Ca 2+ dependent cellular responses, was reduced in elderly CD8 + T cells ( Figure 3B, 3E). For the re-addition protocol, the Ca 2+ entry rate was significantly slower in cells from elderly compared to adult mice ( Figure 3F); a similar tendency was observed in the combined protocol ( Figure 3C). In contrast to untouched CD8 + T cells, the application of 2 mM [Ca 2+ ]ext was able to rescue the impaired Ca 2+ signal in the elderly CD8 + T cells at least to some extend (Supplementary Figure 3). Measurements of ICRAC in  ]ext before re-addition of 0.5 mM Ca 2+ (readdition protocol) of CD8 + T cells from adult (black, n = 4) and elderly (red, n = 4) mice. The scatter dot plot in (E) displays the corresponding statistics of Ca 2+ influx peak and Ca 2+ plateau and (F) the corresponding influx rates. Ca 2+ data are presented as mean ± SEM. Scatter dot plots are presented as mean ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.
AGING CD3/CD28 bead-stimulated CD8 + T cells were not successful due to their already overall small whole-cell currents that were presumably even more reduced in the T cells from elderly mice.

CD8 + T cells from elderly mice show reduced Ca 2+ signals after T cell receptor stimulation and are less affected in their cytotoxic function by varying free external Ca 2+ concentrations
To test for a functional relevance of reduced [Ca 2+ ]int we investigated SOCE in response to a more physiological stimulus. Antibody binding to the CD3/T-cell receptor complex activates T cells and evokes Ca 2+ signals [30].
To explore the differences in TCR-induced [Ca 2+ ]int mobilization between adult and elderly CD8 + T cells we activated the TCR by application of a soluble anti-CD3ε antibody. Figure 4 shows that TCR activation leads to increased Ca 2+ influx in untouched ( Figure 4A) and stimulated ( Figure 4B) CD8 + T cells but could not reach the levels seen in TG-experiments ( Figure 1A, 3A). Mean [Ca 2+ ]int mobilization of the untouched cells was faster and reached overall a higher plateau compared to the stimulated counterparts. As in TG-induced SOCE, CD8 + T cells isolated from elderly mice show less efficient TCR-induced [Ca 2+ ]int mobilization compared to adult mice.
The characteristic function of CD8 + T cells is to eliminate virus-infected and degenerated targets cells [31]. In our previous paper, we reported how changes in [Ca 2+ ]ext influence human CD8 + T cell cytotoxicity [12]. Therefore, we investigated, whether different concentrations of [Ca 2+ ]ext have an impact on the killing efficiency of CD8 + T cells in the context of aging.
Either EGTA or CaCl2 were added to AIM V, a medium optimized for the cultivation of lymphocytes, to decrease or increase free [Ca 2+ ]ext. Quantification of cancer cell killing revealed that free [Ca 2+ ]ext higher or lower than the value of 798 μm of AIM V reduced the cytotoxicity of CD8 + T cells of adult mice ( Figure 4C) similarly as for human CD8 + T cells [12]. In contrast, CD8 + T cells of elderly mice where largely unaffected by fluctuations in free [Ca 2+ ]ext ( Figure 4C). Averaging all experiments showed that elderly CD8 + T cells showed similar cytotoxic efficiency against target cells at a relatively wide range of free [Ca 2+ ]ext between 74 up to 1715 µM compared to their adult counterpart. Thus, Ca 2+ dependent regulation of cytotoxicity is reduced in CD8 + cells from elderly mice.

STIM and Orai molecules are differently expressed in untouched and stimulated CD8 + T cells from elderly mice
The Ca 2+ influx profile in lymphocytes is strictly determined by the composition of Orai channels and STIM Ca 2+ sensors [5,32,33]. The ratio of Orai to STIM determines the characteristics and properties of ICRAC [34,35]. Therefore, we performed a detailed analysis of the expression levels of Orais (Orai1, 2, and 3) and STIMs (STIM1 and 2) in untouched and stimulated CD8 + T cells from adult and elderly mice by quantitative real-time PCR and western blot analysis. As expected, mRNA of both STIMs and all three Orais are abundantly expressed in untouched ( Figure 5A) and stimulated ( Figure 5C) CD8 + T cells from both age groups. To facilitate comparison, expression levels of Orai and STIM genes from elderly mice were normalized to reference genes and shown as relative AGING fold change to the adult group. The untouched CD8 + T cells from elderly mice showed a significant reduction in mRNA transcript levels for Orai2 and STIM1 ( Figure  5A), which could also be confirmed at protein levels after densitometry analysis ( Figure 5B, Supplementary Figure 4C). Surprisingly, as shown in Figure 5A, mRNA levels of STIM2 were not changed in untouched CD8 + T cells of elderly mice but the statistical analysis of the western blots revealed a significant reduction on protein levels ( Figure 5B). The mRNA reduction of Orai1 ( Figure 5A) could not be confirmed at the protein level ( Figure 5B). In addition, a significant downregulation for STIM1 and STIM2 was seen in stimulated CD8 + T cells from elderly mice at mRNA transcript and protein levels ( Figure 5C, 5D). The level of Orai3 mRNA remained unchanged in untouched and stimulated cells ( Figure 5A, 5C). Protein analysis of Orai2 or Orai3 was hampered by limited sensitivity of commercially available antibodies. Since the patchclamp attempts with CD3/CD28 bead-stimulated CD8 + T cells from adult and elderly mice were unsuccessful, we were wondering if this is due to the reduced levels of Orai and STIM after stimulation. Indeed, the stimulated CD8 + T cells showed a significant downregulation of all STIM and Orai mRNA transcript levels compared to the untouched CD8 + T cells for both age groups (Supplementary Figure 4A, 4B). In conclusion, the overall expression levels of STIM and Orai correlate well with the reduction of Ca 2+ entry in CD8 + cells from elderly compared to adult cells and also with the reduction of Ca 2+ signals in stimulated compared to untouched cells.

Enhanced Ca 2+ clearance rate in CD8 + T cells from elderly mice
In addition to the influx mechanisms, [Ca 2+ ]int of course depends on Ca 2+ efflux that could contribute to the difference in Ca 2+ signals of CD8 + T cells from adult and elderly mice. We therefore compared the clearance rates of both untouched ( Figure 6A) and stimulated ( Figure 6D) Figure 6B) and stimulated ( Figure 6E) CD8 + T cells isolated from adult and elderly mice were plotted against the respective Ca 2+ plateaus. Both the untouched and the stimulated CD8 + T cells from the elderly mice show significantly faster Ca 2+ Normalised relative mRNA expressions of Orai1, 2 and 3 and STIM1 and 2 of untouched CD8 + T cells from adult (grey, n = 12) and elderly (red, n = 11) mice. (C) Normalised relative mRNA expressions of SOCE components of stimulated CD8 + T cells from adult (grey, n = 12) and elderly (red, n = 11) mice. Protein quantification after normalization to GAPDH of SOCE components from (B) untouched (n = 7 -9) and (D) stimulated (n = 5 -7) CD8 + T cells lysates isolated from adult (black) and elderly (red) mice. Data obtained are presented as mean ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.
AGING extrusion rates compared to the iso-cells from the adult mice ( Figures 6B, 6E, Supplementary Table 5). PMCA1 and PMCA4 are the two, out of four, known PMCA isoforms that are ubiquitously distributed in all tissues and they are also the major Ca 2+ extrusion pathway in T cells [37]. PMCA4b is highly expressed in T cells and important to shape [Ca 2+ ]int in T cells [37]. It may therefore contribute to the higher Ca 2+ clearance rate seen in elderly CD8 + T cells. We tested the mRNA expression levels of PMCA1 and 4 in untouched and stimulated CD8 + T cells from adult and elderly mice ( Figure 6C, 6F). In agreement with the Ca 2+ clearance rate analysis, PMCA4b expression was increased in CD8 + cells from elderly compared to adult mice; the values for the PMCA1 isoform remained constant ( Figure 6C, 6F). In conclusion, PMCA4b upregulation in CD8 + T cells from elderly mice might contribute to the observed phenotype of reduced [Ca 2+ ]int.

Altered SOCE in central and effector memory CD8 + T cells from elderly mice
While changes in the (relative) frequency of distinct T cell subsets have been described, nothing is known about Ca 2+ signal differences between subsets and if these explain the differences in the total population of CD8 + T cells from elderly and adult mice. In order to characterize the [Ca 2+ ]int mobilization of the most abundant subtypes of stimulated CD8 + T cells, we performed FACS sorting to separate the subtypes of adult and elderly mice and measured their Ca 2+ signals. The normalized data of the adult CD8 + T cells subpopulations is summarized in Figure 7. Overall the [Ca 2+ ]int reduction of CD8 + T cells from elderly compared to adult mice is more prominent in the central memory (CM) than the effector memory (EM) population ( Figure 7). The CM from elderly mice showed a more drastic Ca 2+ influx peak reduction (20 -25 %) ( Figure 7A-7C, 7E) than the EM (~ 10 %) ( Figure 7G-7I, 7K). The same applies for the Ca 2+ plateaus of CM (~ 25 %) ( Figure 7A-7C, 7E) and EM (~ 10 %) ( Figure 7G-7I, 7K). The influx rates of both CM and EM showed a significant reduction for the more physiological combined protocol (CM: ~ 35 %; EM: ~ 23 %) ( Figure 7D, 7F, 7H, 7J). In conclusion, from the two predominant subtypes of stimulated CD8 + T cells, the CM from elderly mice revealed a more distinctly reduced overall Ca 2+ response compared to those of adult animals ( Figure 7A, 7B). For the EM this age-related reduction was less pronounced ( Figure 7G, 7H). Considering the percentage distribution of the subpopulations in both age cohorts (Supplementary Figure 1), the CM/EM ratios reflect the Ca 2+ signal differences seen in the total population of CD8 + T cells ( Figure 3A, 3D).

DISCUSSION
Proper Ca 2+ homeostasis is essential for the immune system to act fast and specific to eliminate tumor cells AGING at early stages. Altered Ca 2+ fluctuations have already been associated with numerous age-related diseases, such as neurodegenerative [39][40][41], autoimmune and inflammatory disorders [42]. Ca 2+ responses are regulated negatively and positively by several mechanisms involving channels, pumps and sensors [43]. During aging, a continuous and steady decrease of all immune function takes places [2,[44][45][46] and especially for T cells alterations in Ca 2+ homeostasis have been reported [25,26].
Although Ca 2+ is an important second messenger the underlying cause or mechanism of Ca 2+ defects at molecular level during aging are not well understood and remain elusive. In this study, we perform for the first time a detailed analysis of Ca 2+ signals in CD8 + T cells from elderly mice and link the aging-related reduction in Ca 2+ signals to reductions of the main key players in the Ca 2+ signaling pathway, further leading to coherent, concurrent functional alterations in Ca 2+ homeostasis. Specifically, we found: 1) reduction of the expression of STIM and Orai proteins leads to reduced Ca 2+ entry; 2) upregulation of PMCA4 additionally contributes to faster Ca 2+ extrusion 3) less efficient TCR-induced [Ca 2+ ]int mobilization and 4) increased insensitivity to Ca 2+ fluctuations during cytotoxic activity. These findings likely account for the observed reduced Ca 2+ signals in CD8 + T cells from elderly mice and contribute to the declined T cell responses.
Generally, one would rather expect an overall reduction in expression of the key proteins involved in the Ca 2+ homeostasis during aging. However, this is not what we observed in CD8 + T cells: whereas Orai1 and STIM1/2 proteins are downregulated, PMCA4b is in contrast upregulated. Both expression changes, however, result in reduced [Ca 2+ ]int. Thus, it can be speculated that there is a deliberate change or adjustment in CD8 + T cells to reduced Ca 2+ signals during the aging process. The agerelated reductions in CD8 + T cell Ca 2+ signals observed in our elderly mice may not imperatively implicate flawed cellular pathways and functions. Hence, one of the fundamental questions is how CD8 + T cells may benefit from reduced Ca 2+ entry in age?
For the lysis of target cells, CD8 + cytotoxic T Lymphocytes (CTL) require a sequence of programmed steps, including target cell binding (conjugate formation), delivery of the lethal hit, target cell lysis, and killer cell recycling [47][48][49]. Several of these steps in the CTL killing machinery require or are modulated by Ca 2+ itself [13]. It is well conceivable that Ca 2+ fluctuations may greatly influence target cell killing by CTL and NK cells, and CRAC channels are well-suited to modulate killing because their dissociation constant KD for Ca 2+ permeation is in the range of 0.84 and 1.17 mM [50,51].
Recently, we analyzed the cytotoxicity of human CTL and NK cells against cancer cells [12]. CTLs showed Ca 2+ dependent cytotoxicity with an optimum for cancer cell elimination at rather low free [Ca 2+ ] concentrations. Downregulation of Orai1 in CTLs led to decreased Ca 2+ signals and increased efficiency to eliminate cancer cells [12]. Alterations of the STIM:Orai stoichiometry might regulate the killing efficiency of CD8 + T cells by changing the cells Ca 2+ signals to be closer to or further away from the Ca 2+ optimum for target cell killing. In the context of aging, we were able to detect changes at the mRNA levels in untouched and stimulated CD8 + T cells but not all Orai and STIM are affected in the same way. While Orai3 has hardly changed, decreased mRNA levels for Orai1, STIM1 and STIM2 could also be confirmed at the protein levels ( Figure 5, Supplementary Figure 4). Surprisingly, the data for STIM2 showed no tendency at mRNA levels for the untouched CD8 + T cells ( Figure 5A) but a significant decrease at protein levels ( Figure 5B). Interestingly, also for Orai1 we observed a slight but not significant increase on protein level ( Figure 5B) while the mRNA levels were significantly decreased ( Figure 5A). This discrepancy may indicate a change in the turnover rate [52], translational regulation [53] or posttranslational modification [54] of these proteins. The stoichiometry of the STIM:Orai ratio determines the current size and inactivation properties [34,35,55] and thus Ca 2+ signals and cell-specific cellular responses. Therefore, small variations in external Ca 2+ could significantly alter Ca 2+ signals and Ca 2+ -dependent target cell killing. Surprisingly, CD8 + T cells of elderly mice were mostly unaffected by fluctuations in free [Ca 2+ ]ext during killing compared to the CD8 + T cells from adult mice ( Figure  4C). This independence from external Ca 2+ fluctuations could be an advantage for the per se impaired CD8 + T cells from elderly mice. This insusceptibility to the wide range in free [Ca 2+ ]ext may offer an excellent adaptation to constantly changing surroundings such as the tumor microenvironment [56].
Additionally, there is no doubt about the importance of Ca 2+ for cell proliferation not only for immune cells but also in the context of cancer cells [57][58][59][60]. Increased Ca 2+ levels lead to an increased proliferation rate and SOCE plays a central role in its regulation [61]. However, the role of Orai channels in proliferation seems to be more complex considering the different results for different mouse models used for the investigation on this matter [62,63]. Deletion of Orai [62] as well as STIM1 in T cells did not alter the proliferation, while T cells lacking both STIM1 and STIM2 proliferate to a much lesser extent [64]. This data support the hypothesis that the threshold of [Ca 2+ ]int necessary for T cell proliferation seems to be low and proliferation does not require high [Ca 2+ ]int levels [65]. However, we did observe declined AGING proliferative capacity in the first three days during stimulation which has been lifted from day four. Similar observations of declined proliferation have already been reported for CD4 + and CD8 + T cells from elderly mice and humans [4,26,61,[65][66][67]. Maybe, this decelerated proliferation benefits the cytotoxic capacity of CD8 + T cells and represents an elaborate adaptation during aging process.
Apoptotic deletion of activated T cells is an essential physiological process to terminate the immune response and the control of the overall number of immunocompetent cells [68,69]. Ca 2+ signaling plays a critical role for the initiation and effectuation of cell death [70][71][72][73]. Studies about apoptosis in aging are controversial and have shown increased apoptosis in T cells while others have observed a decrease or no effect [74][75][76][77][78]. Beside STIM-and Orai-mediated Ca 2+ signals being crucial for T cell cytotoxicity and proliferation, excessive or prolonged Ca 2+ entry can also lead to cell death [62]. Since it has been shown that Orai1 deficiency renders T cells resistant to death upon longterm exposure to anti-CD3 and anti-CD28 [62], the decreased Ca 2+ influx in CD8 + T cells from elderly individuals could provide potential protection against increased apoptosis.
Although the signaling machinery in T cells is extremely complicated and many steps remain to be clarified, age-related changes in Ca 2+ entry may be one important cause of cell-mediated immune response decline with aging. In this study we were able to demonstrate the contribution of Orai, STIM and PMCA in this multifaceted network of channels and pumps in Ca 2+ homeostasis at molecular levels. Additional studies are required to clarify the influence of the reduced Ca 2+ signaling in the context of CD8 + T cells cytotoxicity as one of the main killers in our body.

Mice
C57BL6/J mice were bread in our own colony from stock purchased from Charles River Laboratories. Only female mice were used between 12 and 24 weeks (adult mice) or between 78 and 102 weeks (elderly mice). Mice with splenomegaly (spleen to body weight ratio above 0.6) or macroscopically visible tumors were excluded from the studies. Mice were housed under specific pathogen-free conditions. At the designated times, animals were sacrificed by cervical dislocation and their spleens were harvested. Splenocytes were isolated by pushing the spleen through a 70 µm cell strainer (Corning®) and depleting the erythrocytes by a hypoosmolar solution. All animals used for this study were sacrificed and their organs harvested in compliance with the German Animal Protection Law (Tierschutzgesetz, §11, Abs.1 Nr.1).

T cell culture and stimulation
CD8 + T cells were negatively isolated from murine splenocytes by using the Dynabeads™ Untouched™ Mouse CD8 Cells Kit (ThermoFisher). Murine CD8 + T cells were cultured in AIM V medium, supplemented with 10 % FCS, 100 U/ml IL-2, and 50 µM β-Mercaptoethanol, without stimulus. For the stimulition, the above mentioned medium was supplemented with Dynabeads™ Mouse T-Activator CD3/CD28 for T-Cell Expansion and Activation (ThermoFisher). The stimulated CD8 + T cells were kept in the medium with a 5:4 cell-to-bead ratio at 37 °C and 5 % CO2 for up to 4 days.

Flow cytometry and cell sorting
Splenocytes and CD8 + T cells of adult and elderly mice were stained with the antibodies listed in 'Abs and reagents' and incubated for 20 min in the dark at room temperature. Stained samples were acquired using a FACSVerse™ (BD Biosciences) flow cytometer, and acquired data were analyzed with FlowJo software (FlowJo, LLC). For the sorted subtypes, stained 72 h stimulated CD8 + T cells were sorted on a FACSAria™III (BD Biosciences) sorter and left to rest from the sorting procedure for at least 4 h.

Quantitative real-time PCR
Total RNA from CD8 + T cells was extracted by phenol-chloroform extraction.

Fluorescence-based Ca 2+ -imaging
Ca 2+ -Imaging was performed according to Alansary, Kilch et al. paper [79]. Briefly, murine CD8 + T cells were loaded in AIM V medium with 1 µM Fura2-AM for 30 min at room temperature and allowed to attach to polyornithine-coated glass coverslips for 15 min. All experiments were carried out at room temperature in self-built perfusion chambers with low volume and high solution exchange rate. The external Ca 2+ Ringer solution contained (in mM): 155 NaCl, 2 MgCl2, 10 glucose, 5 HEPES and 0.5 or 2 CaCl2 (0.5 / 2 Ca 2+ Ringer) or no CaCl2, but 1 EGTA and 3 MgCl2 instead (0 Ca 2+ Ringer). The pH was adjusted to 7.4 with NaOH. Images were analyzed with VisiView software (Visitron). Quantification of the trace shows the ratio (340 nm/380 nm) corresponding to Ca 2+ influx peak and plateau and Ca 2+ influx and efflux rate as ratio (340 nm/380 nm)s -1 . Parameters analyzed were the influx peak, as maximal Ca 2+ signal reached after TG application after (re-)adding of Ca 2+ , and the average plateau. The plateau mirrors the balance between influx and efflux and was analyzed before the application of 0 mM Ca 2+ (Ca 2+ removal).

Patch-clamp measurements
Patch-clamp experiments were performed in the tightseal whole-cell configuration at RT. Voltage ramps of 50 ms duration spanning a range of -100 to +100 mV were delivered from a holding potential of 0 mV at a rate of 0.5 Hz. All voltages were corrected for a liquid junction potential of 10 mV. Currents were filtered at 2.9 kHz and digitized at 100 µs intervals. Capacitive currents were determined and corrected before each ramp.

Real-time killing assay
The real-time killing assays were carried out as described in Kummerow et al. 2014 [80]. Briefly, P815 mastocytoma cells were loaded with 500 nM calcein-AM in AIM V medium containing 10 mM HEPES at room temperature for 15 minutes. The loaded P815 cells were then settled at 2.5×10 4 cells per well into black 96well plates with clear-bottoms (353219, Corning, Amsterdam, Netherlands). CD8 + T cells were pulsed with 2 μg/ml anti-CD3ε antibody and subsequently added onto the P815 cells at a 20:1 effector to target ratio. Target lysis was measured either in an M200 Infinite plate reader (Tecan, Crailsheim, Germany) or a Genios Pro (Tecan) using bottom reading function at 37°C. The quantification of free Ca 2+ concentration in AIM V medium supplemented with different amounts of Ca 2+ or EGTA was done as previously described in Zhou et al. 2018 [12].

Statistical analysis
All values are given as mean ± SEM or SD. Data were analyzed using VisiView (Visitron), Microsoft Excel (Microsoft), Igor Pro (Wavemetrics), Image Lab™ (Bio-Rad) and GraphPad (GraphPad Software Inc.). Rate constants (k values) of iso-cells were calculated for each trace by exponential decay analysis after removal of external Ca 2+ . Significances of data were calculated with an unpaired two-sided Student's t-test if Gaussian distribution was given. If no Gaussian distribution was given, data were analyzed with the nonparametric Mann-Whitney test. For multi-parameter analysis data were analyzed with ANOVA. Degrees of significance AGING were set at * p < 0.05, ** p < 0.01, *** p < 0.001 and **** p < 0.0001.  and stimulated (light red, n = 13) CD8 + T cells from elderly mice. Representative Western blots of SOCE components from untouched (C) and stimulated (D) CD8 + T cells. Scatter dot plots are presented as mean ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Supplementary Tables
Supplementary Table 1