Beneficial effect of resveratrol on T cell oxidative metabolism and anti-tumor function is conditioned by prior in vivo T cell history

Abstract

Despite the clinical success of redirected T cells in the setting of cancer adoptive cell immunotherapy, patients may exhibit resistance to treatment, resulting in uncontrolled disease and relapses. This phenomenon partly relies on impaired ex vivo-produced T cell metabolic fitness, including a decreased respiratory reserve, as well as a greater sensitivity to tumor-mediated metabolic stress. To improve the respiratory capacity of cultured T cells, we sought to target the nicotinamide adenine dinucleotide/sirtuine-1/reactive oxygen species (ROS) axis through supplementation of culture medium with resveratrol. Resveratrol-treated T cells display broader respiratory capacities, along with sustained ROS control ability. Strikingly, we reveal that the effect of resveratrol on T cells is restricted to cytomegalovirus (CMV)-exposed donors, a virus known to promote immune aging. Herein, CMV prior infection is associated with the influence of terminally differentiated T cells on the fate of companion T cell subsets. Moreover, beyond resveratrol’s effect on redirected T cell metabolic features, it provides a functional anti-tumor advantage to these CMV-seropositive donor-derived T cells, in a third-generation CD123-specific chimeric antigen receptor-T cell in vitro model. This highlights the necessity to consider patient’s intrinsic attributes, especially immune aging-related ones, when assessing new T cell production processes to improve clinical efficacy, pushing the limits of personalized medicine.

Graphical abstract

Mercier-Letondal and colleagues demonstrate that immunotherapeutic T cells treated with resveratrol display enhanced respiratory and ROS control capacities, relying on increased mitophagic flux and SIRT-1 activity. This metabolic improvement is associated with T cell donor’s immune aging-related attributes and with enhanced CAR-T cell anti-tumor function.

Introduction

For several decades, we have witnessed the gradual advent of the T cell-based cancer immunotherapy era. The first successes associated with chimeric antigen receptor-bearing T (CAR-T) cell treatments have resulted in their clinical use in the setting of malignant hemopathies. To date, seven CD19- or BCMA-targeting CAR-T specialties have received marketing authorization since 2017 from regulatory agencies worldwide. The last one was approved by the US Food and Drug Administration only in late 2024.¹ After a myriad of clinical trials in the field of solid tumors, displaying mitigated success, the first transgenic T cell receptor-bearing T cell immunotherapeutic product, directed against MAGE-A4, has been authorized in 2024 to treat metastatic or unresectable synovial sarcoma.²

All these significant clinical results might not, however, make us forget the remaining hurdles regarding primary and secondary resistance to T cell-based treatments. Resistance not only implies tumor-mediated mechanisms but also involves administrated T cell fitness.^3,4,5,6 Indeed, an ex vivo culture phase is required to expand and genetically modify immunotherapeutic T cells and may lead to phenotypic and functional T cell distortions, involving a lack of long-term in vivo physical and/or functional persistence and thus affecting their anti-tumor efficiency.⁷

This lack of long-term persistence of immunotherapeutic T cells implies both their likeliness to precociously die after injection and their likeliness to dysfunction once the metabolically inhospitable tumor micro-environment (TME) is reached.^8,9,10 Therefore, the T cell culture process needs to be refined in order to allow for the generation of suitably armed for persistence anti-tumor T cells. In this aim, favoring a high respiratory capacity of cultured T cells appears to be a strategy of choice. Early memory T cells, such as stem central memory T cells (T_SCM) and central memory T cells (T_CM), represent desirable subsets to ensure a better in vitro and in vivo anti-tumor function in the setting of adoptive immunotherapy.^8,11,12,13 Those early memory T cells are characterized by a high fatty acid-related oxidative (FAO) catabolism and an efficient mitochondrial metabolism. These two parameters are constitutive of a high spare respiratory capacity (SRC).^14,15,16,17 Moreover, an elevated SRC is correlated with a more sustainable motility in vitro¹⁸ and is prone to increase resistance to TME-mediated nutrient deprivation.^19,20

To date, several studies reported encouraging results regarding oxidative phosphorylation (OXPHOS) promotion in cultured T cells. For instance, replacing interleukin (IL)-2 with IL-15 or IL-21, replacing CD28 with CD137 for co-stimulation, or engineering CAR-T with an inhibition resistant peroxisome proliferator-activated receptor-γ coactivator-1α (PGC1α) gene exhibited positive effect on SRC and FAO-derived OXPHOS.^{21,22,23,24,25}

It has already been demonstrated that mitochondrial fitness, and consequently, SRC, is associated with nicotinamide adenine dinucleotide (NAD) metabolism.²⁶ NAD is the co-factor of critical metabolic enzymes involved in oxido-reductive activity, either regarding glycolysis or the Krebs cycle, and OXPHOS-associated pathways. Moreover, NAD is the substrate of several families of enzymes, including sirtuine (SIRT)’s family. As such, appropriate NAD content and balance between its oxidized (NAD⁺) and its reduced (NADH) form are required to fine-tune the cellular metabolism. SIRT-1 is a deacetylase belonging to the SIRT family and is described to have various substrates. Among them, SIRT-1 activates the transcription factor PGC-1α.^27,28,29 PGC-1α activity is positively associated with mitochondrial biogenesis, quality maintenance, and structural organization, as well as FAO promotion and oxidative stress-fostering reactive oxygen species (ROS) control.^{30,31,32,33,34,35,36} Thus, high cellular NAD availability is an indicator of metabolic fitness, whereas weak NAD cellular rate and immune aging mutually contribute to each other.^37,38,39,40 The anti-oxidant molecule resveratrol (RSV) has been reported to promote SIRT-1 deacetylase activity through elevated NAD⁺/NADH ratio sensing.⁴¹ It has already been demonstrated that regarding human or murine CD4⁺ T cells or thymocytes, RSV culture medium supplementation is likely to increase the quantity⁴² and quality⁴³ of mitochondria along with OXPHOS pathway activity, ATP production,⁴³ and SIRT-1 activation.⁴⁴ In addition, at such a similar dose, RSV does not display any prejudicial influence on cell proliferation or death.⁴⁵

Despite these previous results indicating that RSV influences mitochondrial metabolism, including that of T cells, the specific impact of RSV as a medium supplement in a setting of ex vivo immunotherapeutic T cell production remains elusive. Herein, in a context of classical ex vivo immunotherapeutic T cell production, we evaluated the effects of culture medium supplementation with RSV, prone to target the NAD/SIRT-1/ROS axis, on the metabolic and respiratory capacities of T cells and on their inherent related-oxidant stress control. Knowing that SIRT-1 activity depends on immune aging-related NAD concentration, we further questioned the impact of immune aging-associated biological parameters on metabolic RSV effect. Last, we compared the anti-tumor function of RSV-treated or untreated CAR-T cells.

Results

RSV-treated T cell metabolic capacities and proliferative and differentiation status

Activated and cultured T cells, treated with or without RSV, were evaluated throughout the culture period for proliferation and on day 10 for metabolic capacities and differentiation phenotype (Figure 1A). RSV-cultured T cells display significantly greater OXPHOS resort than control T cells, as assessed by baseline oxygen consumption rate (OCR) values (Figure 1B). Similarly, RSV treatment allows for a significant increase in T cell-associated respiratory capacity, as demonstrated by the SRC values (Figure 1C). Interestingly, even if the evaluated donors do not individually benefit from RSV treatment regarding SRC, greater metabolic flexibility is observed after exposure to RSV. Therefore, a more flexible and countervailing resort to oxidizable substrates is enabled by RSV treatment in this context. Blockade of fatty acids, glutamine, or pyruvate catabolism by Etomoxir, BPTES, or UK5099, respectively, leads to a relative increase of T cell SRC after RSV exposure compared with controls. In contrast, glucose commitment in the glycolysis pathway blockade by 2-DG rather dramatically impairs T cell oxidation capacities, independent of previous RSV exposure and of pyruvate availability (Figure 1D). Similarly, and as expected, T cell culture supplemented with Everolimus decreases mechanistic target of rapamycin (mTOR)c1 and glycolytic activity and negatively influences T cell oxidative functions to the extent that they exceed a certain basal threshold (Figure S1A). Taken together, these results show that the sustained mitochondrial oxidation process requires optimal glycolysis activation to be functional and that this requirement does not only rely on pyruvate as an oxidizable substrate.

Figure 1.

T cell metabolic capacity and proliferation and differentiation status

(A) Assay design. Figure created with BioRender.com. (B) T cell OCR values at baseline, with the mean (SD)-representing bar. Data from 30 independent donors. Paired two-tailed Student’s t test. (C) T cell OCR values in the standard oxidation test [mean (SD)]. Data from 26 independent donors. Paired two-tailed Student’s t test on SRC-calculated values. (D) (Left) Glutaminolysis, fatty acid oxidation, mitochondrial pyruvate transport, and glycolysis inhibition pathways. Figure created with BioRender.com. (Right) T cell SRC after the pathway inhibition oxidation tests and the corresponding mean (SD). Data from 5–8 independent donors. Paired two-tailed Student’s t test. (E) T cell ECAR values at baseline [mean (SD)]. Data from 25 independent donors. Paired two-tailed Student’s t test. (F) T cell ECAR values in the standard glycolysis test [(mean (SD)]. Data from six independent donors. Paired two-tailed Student’s t test on glycolytic capacity calculated values. (G) T cell total ATP and glyco- and mito-associated APRs [(mean (SD)]. Data from 23 independent donors. Paired two-tailed Student’s t test. (H) Proliferation index (cell count relative to day 0) throughout the cell culture of T cells [(mean (SD)]. Data from 28 independent donors. Two-tailed Wilcoxon matched-pairs signed rank test on day 10-obtained values. (I) T cell differentiation profiles [FCM panel differentiation (1)]. Percentages [(mean (SD)] of naive T cells (white bar), T_SCM (light gray bar), T_CM (middle gray bar), T_EM cells (gray bar), and T_EMRA (dark gray bar). Data from 15 independent donors. Paired two-tailed Student’s t test on each T cell subset.

Further, T cell basal and maximal glycolytic capacities are significantly greater after RSV exposure (Figures 1E and 1F). Overall, RSV-derived glycolytic flux enables T cells to produce more ATP molecules (total APR), either through proper glycolysis (glycolysis-derived APR [glyco-APR]) or pyruvate and alternative substrates oxidation (mitochondrial-derived APR [mito-APR]) (Figure 1G).

Despite an expected increase in glycolytic flux-associated biosynthesis ability, the in vitro proliferative capacity of RSV-exposed T cells is impaired (Figure 1H). This proliferative defect similarly affects CD8⁺ and CD4⁺ T cells, as demonstrated by a comparable CD4/CD8 ratio (Figure S1B) and is more pronounced during the first 6 days of the culture period (Figure S1C). The cell size and the translation process are comparable between T cells cultured with or without RSV, as indicated by the intensity of forward scatter and tRNA-linked puromycin staining, respectively (Figures S1D and S1E). Collectively, these results suggest that glycolysis is directed toward ATP production rather than toward biosynthetic features. As metabolic capacities support and are in turn supported by T cell differentiation, we further studied the T cell differentiation profile. The results reveal that RSV affects the proportions of both CD8⁺ or CD4⁺ T cell memory subsets. This impact implies a decrease of T_SCM along with an increase of T_CM populations. The overall proportions of these early memory T cells nevertheless remain comparable. Moreover, an interesting decrease regarding effector memory expressing CD45RA T cells (T_EMRA) proportion is observed after RSV supplementation. As expected, regardless of the culture conditions, naive T cells are no longer detectable at the end of the propagation period (Figure 1I).

To sum up, RSV-treated T cells display greater and more flexible respiratory capacities than control T cells, partly because of the reliance on glycolytic flux. Moreover, glycolysis is enforced in RSV-exposed T cells. Nevertheless, increased glycolysis does not seem to be associated with intense biosynthesis but rather with ATP production. Finally, an unexpected reduction in the T_SCM rate is observed, while compensated by a T_CM rate concomitant increase, and accompanied by a decrease in the T_EMRA proportion.

RSV treatment impact on ROS management and related metabolic consequences

Activated and cultured T cells in the presence or absence of RSV were analyzed on day 10 regarding ROS quantity through H2DCFDA staining, ROS accumulation management, and ROS scavenging power (Figure 2A). First, the ROS content of CD8⁺ and CD4⁺ T cells is similar under both RSV and control conditions (Figure 2B). This finding could be surprising since RSV-induced more sustained oxidative metabolism is prone to increase ROS mitochondrial production. Thus, we further evaluated T cell ROS accumulation management capacity via H₂O₂ medium supplementation followed by ROS content quantification. Compared with their untreated counterparts, RSV-exposed CD8⁺ and CD4⁺ T cells are more likely to deeply scavenge ROS (Figure 2C). This more efficient ability of RSV-treated T cell to eliminate ROS is associated with higher respiratory rates. Indeed, H₂O₂-exposed T cells exhibit an increased SRC after RSV treatment compared with the control condition (Figure 2D). Surprisingly, glycolysis-derived reduced NAD phosphate (NADPH)-based cellular ROS scavenging systems do not demonstrate any increased activity in RSV-treated T cells. Thus, neither the NADPH-based total anti-oxidant capacity (targeting both glutathione and thioredoxin ROS scavenging systems) nor the specific glutathione system effectors display any differential activity between RSV and control conditions (Figures 2E–2H). To summarize, RSV-treated T cells are better able to eliminate excessive ROS and further maintain their respiratory capacity, even after the triggering of an ROS burst. In this setting, NADPH-based scavenging systems-increased activity is not involved in this phenomenon, suggesting an alternative ROS-buffering pathway.

Figure 2.

RSV treatment impacts ROS management and associated-metabolic consequences

(A) Assay design. Figure created with BioRender.com. (B) RSV-treated to medium control CD8⁺ and CD4⁺ T cell H2DCFDA MFI fold change [mean (SD)]. Data from 24 independent donors. (C) H₂O₂-exposed to medium control T cell H2DCFDA MFI fold change [mean (SD)]. Data from 15 independent donors. Paired two-tailed Student’s t test. (D) H₂O₂-exposed to medium control T cell SRC fold change [mean (SD)]. Data from 11 independent donors. Paired two-tailed Student’s t test. (E) T cell NADPH-based total anti-oxidant capacity [mean (SD)]. Data from three independent donors. Paired two-tailed Student’s t test. (F) T cell oxidized (dark color) or reduced (light color) glutathione content [mean (SD)]. Data from three independent donors. Paired two-tailed Student’s t test. (G) T cell glutathione reductase activity [mean (SD)]. Data from three independent donors. Paired two-tailed Student’s t test. (H) T cell glutathione peroxidase activity [mean (SD)]. Data from three independent donors. Paired two-tailed Student’s t test.

RSV-related mitochondrial fitness and turnover

Respiratory capacities and ROS-induced oxidative stress levels are closely related to mitochondrial function. We thus assessed both mitochondria quantity and polarization status of RSV-treated or untreated T cells, along with their mitophagic activity (Figure 3A). Mitochondrial quantification, monitored through Mito Tracker Green dye, does not reveal any difference between RSV and control culture condition, regarding neither CD8⁺ nor CD4⁺ T cells (Figure 3B). Mitochondria polarization level, evaluated via TMRM intensity measurement, trends to decrease in CD8⁺ T cells after RSV treatment, whereas no difference is detected in CD4⁺ T cells (Figure 3C). However, the level of mitochondrial polarization of RSV-treated or untreated CD8⁺ T cells remains on the same order of magnitude, compared with mitochondrial uncoupler carbonyl cyanide p-trifluoro methoxyphenylhydrazone-exposed T cells (Figure S2). Taken together, these snapshot analyses do not allow for detecting any differences in mitochondrial dynamics or turnover. We further unravel mitophagic activity through mitophagy-related protein expression levels. Under RSV-treated conditions, we observed an increased expression of PINK1 and BNIP3L/Nix, which are involved in depolarized mitochondria labeling and addressing these labeled mitochondria to the autophagosome, respectively. Conversely, the expression of Parkin, which is known to be activated by and to cooperate with PINK1 in the context of mitophagy, does not differ between RSV and control conditions. These results do not allow us to speculate on its possible activity. Moreover, LC3B is equally expressed under both conditions, suggesting that the global autophagic process is not involved, but rather relies on mitophagy specifically (Figures 3D and 3E). Taken together, these observations underlie an RSV-associated increase in mitophagic flux and, with respect to the previously reported absence of increase in mitochondrial quantity (Figure 3B), an increase in turnover. Altogether, these data show that, despite the unclear impact of RSV on snapshot mitochondrial quality analyses, RSV positively influences mitophagic flux and mitochondrial turnover. These results indicate increased mitochondrial fitness, and this phenomenon provides a plausible explanation for the RSV-mediated increase in respiratory and ROS accumulation management capacities.

Figure 3.

Mitochondrial fitness and turnover

(A) Assay design. Figure created with BioRender.com. (B) CD8⁺ (left) and CD4⁺ (right) T cell MitoTracker Green MFI [mean (SD)]. Data from 11 independent donors. Paired two-tailed Student’s t test. (C) CD8⁺ and CD4⁺ TMRM fluorescence intensity values. Data from 11 independent donors. Paired two-tailed Student’s t test. (D and E) Western blot analysis of mitophagy-related protein expression. (D) PINK1, Parkin, BNIP3L/Nix, and LC3B expression-related raw data. Data from one of three representative donors. (E) T cell relative expression [mean (SD)] of PINK1, Parkin, BNIP3L/Nix, and LC3B. Data from three independent donors. Paired two-tailed Student’s t test.

SIRT-1 involvement in the RSV-mediated effect on cultured T cells

To assess the role of RSV in increasing oxidative metabolism and managing ROS excess, we explored RSV-induced SIRT-1 activation. To do so, we first investigated the relative expression of SIRT-1 at both the mRNA and protein levels (Figure 4A). Surprisingly, relative SIRT-1 mRNA transcription is lower after T cell RSV treatment (Figure 4B). Nevertheless, SIRT-1 protein expression is equivalent between the RSV-exposed and medium only conditions (Figures 4C and 4D). We then evaluated the enzymatic activity of SIRT-1, which is likely involved in RSV-mediated effects. First, the SIRT-1 activity-related NADH/NAD⁺ ratio tends to decrease in RSV-treated T cells, despite a similar overall NAD⁺ content (Figures 4E and 4F). Next, we investigated the influence of EX-527, a SIRT-1-specific inhibitor, on RSV-treated T cells, especially regarding SRC, as well as ROS accumulation handling and subsequent managed oxidative stress. The EX-527-supplemented condition results in an intermediate level of respiratory metabolism-associated SRC compared with that in the RSV-alone and medium conditions (Figure 4G). EX-527 exposure prevents the RSV-induced increase of ROS management capacity (Figure 4H). Similarly, following H₂O₂ burst triggering, the beneficial influence of RSV on T cell SRC is abrogated by EX-527 treatment (Figure 4I). Taken together, these findings suggest that the effects of RSV on oxidative metabolism and ROS management are linked, at least partly, to SIRT-1 activity.

Figure 4.

SIRT-1 is involved in RSV-mediated effects

(A) Assay design. Figure created with BioRender.com. (B) T cell relative SIRT-1 mRNA expression [mean (SD)]. Data from three independent donors. Paired two-tailed t test. (C and D) T cell relative SIRT-1 protein expression as determined by western blotting analysis. (C) Raw SIRT-1 expression data. Data from one of three representative donors. (D) Relative SIRT-1 expression [mean (SD)]. Data from three independent donors. Paired two-tailed Student’s t test. (E and F) T cell NAD content biochemical analysis. (E) NAD⁺ cellular content [mean (SD)]. Data from eight independent donors. Paired two-tailed Student’s t test. (F) NADH/ NAD⁺ ratio [mean (SD)]. Data from five independent donors. Paired two-tailed Student’s t test. (G–I) SIRT-1 activity was evaluated via EX-527. (G) T cell OCR values at baseline and in the standard oxidation test [mean (SD)]. Data from three independent donors. Paired two-tailed Student’s t test on SRC calculated values. (H) H₂O₂ exposed to medium control T cell H2DCFDA MFI fold change; CD8⁺ (left) or CD4⁺ (right) T cells [mean (SD)]. Data from 9–15 independent donors. Paired two-tailed Student’s t test. (I) H₂O₂ treated to medium control T cell SRC fold change [mean (SD)]. Data from 7–11 independent donors. Paired two-tailed Student’s t test (medium vs. RSV) or two-tailed Wilcoxon matched-pairs signed rank test (medium vs. RSV+EX-527 and RSV vs. RSV+EX-527).

As RSV has been described as an mTOR and AMP-dependent kinase (AMPK) modulator,⁴⁶ we assessed the activity levels of these two kinase complexes. At the RSV concentration of 5 μM, neither p70s6 nor AMPK phosphorylation status were modulated (Figures S3A and S3B). Collectively, these results suggest that at a concentration of 5 μM, the impact of RSV on T cell oxidative functions is at least partly mediated by the modulation of SIRT-1 activity, whereas mTORc1 and AMPK do not seem to be involved in this process.

Influence of individual immune aging-related parameters on the RSV-mediated impact on cultured T cells

Inter-donor variability in the SRC response to RSV treatment is observed, as evidenced by an increased gap between the SD and the SEM values compared with those in the control medium (Figure 5A). Contrary to sex, age and previous cytomegalovirus (CMV) infection are clearly identified to be related to immune aging.⁴⁷ We hypothesized that these two parameters may be partly responsible for the observed variability of SRC after RSV exposure of T cells. We thus evaluated the influence of age, CMV status, and sex of the donor on T cell respiratory capacity increase, inducing RSV treatment. As expected, sex does not seem to influence RSV-mediated SRC benefit. In contrast, this benefit is superior for donors over 50 years of age than for younger people and even more superior for CMV-seropositive donors than for CMV-seronegative donors (Figure 5B). We thus reanalyzed the data from Figures 1C, 2C, and 2D relative to oxidative metabolism and ROS excess handling according to CMV status. Thus, we determine that the RSV-mediated beneficial impact on T cell SRC is restricted to CMV-seropositive donors (Figure 5C). Similarly, the RSV-induced increase in ROS management capacity is also associated with prior CMV infection (Figure 5D). Ultimately, the positive impact of RSV on the SRC after an H₂O₂ shot is only observed in the context of CMV seropositivity. Taken together, this information allows us to assume that the effect of RSV on T cell respiratory and ROS control capacities is potentiated by immune aging-influencing parameters, and this is especially true regarding CMV exposure.

Figure 5.

Influence of individual-related attributes on RSV-mediated impact

(A) T cell SEM and SD comparative values obtained from the SRC set of values. The change (Δ) value is the result of subtracting the SEM value from the SD value. Data were calculated from a normal distribution of 26 values. (B) Percentages of donors who benefit from RSV treatment (black bars) or not (gray bars) according to sex (left), age (center), and CMV status (right). RSV-mediated SRC benefit is defined by an RSV-to-control SRC ratio of >1.2. Data from 26 healthy donors. Chi-square contingency test. (C) T cell OCR values at baseline in the standard oxidation test [(mean (SD)]. Data from 26 independent donors segregated according to CMV status. Paired two-tailed Student’s t test on SRC calculated values for each donor category. (D) H₂O₂-exposed to medium control T cell H2DCFDA MFI fold change; CD8⁺ (left) or CD4⁺ (right) T cells [mean (SD)]. Data from 14 independent donors were segregated according to CMV status. Paired two-tailed Student’s t test for each donors category. (E) H₂O₂ exposed to medium control T cell SRC fold change [mean (SD)]. Data from 11 independent donors were segregated according to CMV status. Paired two-tailed Student’s t test for each donor category.

Influence of CMV on RSV-mediated effects

CMV infection is known to impact the T cell differentiation status. Indeed, CMV-specific CD8⁺ T cells account for a greater proportion of T_EMRA than other β herpes viruses, such as Epstein-Barr- or herpes simplex virus-specific CD8⁺ T cells.^48,49,50,51 Moreover, through its likeliness to lead to immuno-senescence, CMV is prone to drive a global imbalance between naive T cells and T_EMRA, in favor of the latter.⁴⁹ On the one hand, even if often described as accounting for 1%–2% of the T cell compartment,^52,53 total CMV-specific T cells may in fact represent from 10% to 40% of the compartment,⁵⁴ contributing to T_EMRA accumulation. On the other hand, the detection of CMV-associated T cell markers goes beyond the simple T cell CMV specificity.^53,55,56 Taken together, these elements point out that CMV infection triggers T_EMRA cell accumulation in the donor’s peripheral blood. Herein, we demonstrate similar T_EMRA rates in peripheral blood mononuclear cell (PBMC)-derived CD8⁺ and CD4⁺ T cells from CMV-seropositive donors compared with those reported by Scheer et al.⁵⁶ (Figure S4A). To assess the effect of RSV on T_EMRA frequency, we first magnetically sorted T_EMRA cells from PBMC of CMV-positive healthy donors. T_EMRA-enriched (T_EMRAE) and T_EMRA-depleted (T_EMRAD) fractions were cultured with or without RSV for 10 days, and their proliferation potential and viability were assessed every 3–4 days (Figure 6A). Regardless of RSV treatment, the T_EMRAE fractions exhibit a drastically impaired proliferative capacity (Figure 6B). In addition, starting from day 3 of culture, especially after RSV treatment, the T_EMRAE fractions display a trend to a lower viability than the T_EMRAD fractions do (Figure 6C).

Figure 6.

CMV exposure-related influence on T cells and RSV-mediated effects

(A) Assay design. Figure created with BioRender.com. (B) T cell proliferation index throughout cell culture [mean (SD)]. Data from five independent donors. Paired two-tailed Student’s t test. (C) viable T cell percentages throughout cell culture [mean (SD)]. Data from five independent donors. Paired two-tailed Student’s t test. (D) Assay design. Figure created with BioRender.com. (E and F) T_EMRA differentiation index [FCM panel Differentiation(2); T_EMRA percentage ratio calculated as follows: T_EMRAD+T_EMRAE/T_EMRAD, gated on T_EMRAD-derived T cells]; CD8⁺ (left) or CD4⁺ (right) T cells [mean (SD)]. (E) On day 10 of culture. Data from five independent donors. Paired two-tailed Student’s t test. (F) On day 3 of culture. Data from five independent donors. Paired two-tailed Student’s t test.

Thereafter, we aimed to establish the impact of T_EMRA presence on the fate of other T cell subsets. Thus, the T_EMRAE fraction was labeled with Cell Trace Yellow dye and mixed back or not with the T_EMRAD fraction in a similar proportion to that initially observed for each donor. Subsequent cell activation and culture with or without RSV were performed (Figure 6D). The phenotypic evolution of T_EMRAD cells mixed back or not with T_EMRAE-labelled cells, either CD8⁺ or CD4⁺ T cells, was analyzed during the culture process until day 10 (Figure S4B). Focusing on T_EMRA cell differentiation from T_EMRAD cells during the culture process, we note a decrease in the number of RSV-treated and T_EMRAD-derived T_EMRA CD8⁺ and CD4⁺ T cells in the presence of T_EMRAE (Figure 6E). This result is consistent with previously reported data regarding T_EMRA rates on day 10, as shown in Figure 1I. Notably, this decrease is preceded by an early increase, detected at day 3 of culture, in RSV-treated and T_EMRAD-derived T_EMRA CD8⁺ (trend toward significance) and CD4⁺ T cells in the T_EMRAE presence of T_EMRAE (Figure 6F).

Altogether, our findings demonstrate that, after CMV infection, T_EMRAE cells are less able to expand ex vivo and tend to die more easily after day 3 of culture, especially in the presence of RSV. These RSV-treated and previously CMV-imprinted T_EMRAE cells specifically did not lead to T_EMRAD terminal differentiation, and this following a prior increase in T_EMRAD-derived T_EMRA differentiation. CMV-imprinted T_EMRA cells allow for purging RSV-induced undesirable terminally differentiable and differentiated T cells.

Impact of RSV treatment on CAR-T cell cytotoxic functions

To extend our previous data regarding RSV impact from cultured T cells to anti-tumor engineered CAR-T cells, we assessed the respiratory capacities of CMV-positive donor-derived and RSV-exposed CD123-specific CAR-T cells (Figure 7A). We confirm that CAR-T cells tend to behave like untransduced T cells (UTCs) regarding RSV-induced SRC increase (Figure 7B). Notably, the transduction efficiency is comparable between CAR-T cells treated or not with RSV (Figure S5A), and neither CAR-T cells nor UTCs not display significant differences regarding proliferation and differentiation profiles after RSV exposure (Figures S5B and S5C).

Figure 7.

RSV-exposed or not CAR-T cell anti-tumor efficiency

(A) Assay design. Figure created with BioRender.com. (B) CAR-T cell and UTC OCR values in the standard oxidation test [mean (SD)]. Data from eight independent donors. Paired two-tailed Student’s t test (RSV-untreated CAR-T cells vs. UTCs, RSV-treated or UTCs and CAR-T cells) or Wilcoxon matched-pairs signed rank test (RSV-treated CAR-T cells vs. UTCs) on SRC calculated values. (C–E) Acute cytotoxicity model. (C) Percentage of remaining alive CAL-1 after a 24-h co-culture with CAR-T cells or UTCs [mean (SD)]. Data from four independent donors. Paired two-tailed Student’s t test with (RSV-treated or not CAR-T cells; RSV-treated CAR-T cells vs. UTCs) or without Welch’s correction (RSV-treated CAR-T cells or UTCs; untreated CAR-T cells vs. UTCs). (D) Percentage of specific target lysis of CAL-1 by CAR-T at 1:1, 1:5, 1:10, and 1:20 effector to target ratios and after 24 h of co-culture [mean (SD)]. Data from four independent donors. (E) Total integrated intensity of dead cell-associated fluorescence monitoring from 24 h to 56 h post co-culture at a 1:5 effector:target ratio [mean (SD)]. Data from four independent donors. Paired two-tailed Student’s t test. (F–H) Chronic cytotoxicity assay; CAR-T cells in vitro tumor control capacity (assessed through T cell percentage), evaluated after repeated stimulations. (F) After RSV withdrawal [mean (SD)]. Values from three independent donors. Paired two-tailed Student’s t test regarding CAR-T percentages at the onset time of the defect in control of tumor growth. (G and H) After maintaining exposure to RSV (G). CAR-T percentages after iterative stimulations. Values from three independent donors. (H) CAR-T percentages at the onset time of the defect in control of tumor growth. Values from three independent donors. Paired two-tailed Student’s t test.

Further, we sought to characterize in vitro the CD123-specific CAR-T anti-tumor functions, in both acute and chronic stimulation-mimicking settings (Figure 7A). In the acute cytotoxicity model, we show that CAR-T cells, regardless of prior RSV treatment, are able to specifically eliminate CAL-1 target cells at a 1:1 ratio after 24 h (Figure 7C). Further, we do not find evidence of any difference in cytotoxic properties between RSV-treated or untreated CAR-T cells, whatever the effector:target ratio is at a 24 h co-culture time point (Figure 7D) or according to a continuous monitoring from 24 to 56 h (Figure 7E). In the chronic stimulation-mimicking cytotoxicity assay, RSV treatment was maintained (Figures 7G and 7H) or not (Figure 7F) during the cytotoxicity assay. CAR-T cells treated or not with RSV similarly control CAL-1 target cell growth in case of RSV deprivation. Their failure to control CAL-1 growth occurs according to a comparable delay, reflecting a similar chronic CAR stimulation-mediated functional impairment (Figure 7F). Conversely, maintaining exposure to RSV throughout the cytotoxicity assay specifically displays a trend to a more sustained capacity of RSV-treated CAR-T cells to control the target tumor cell line growth compared with untreated ones, even if the calculated p value remains statistically non-significant (p = 0.0754) (Figures 7G, 7H, S5D, and S5E).

Taken together, these data show that supplementation of culture medium with RSV displays a positive effect on the oxidative metabolism and the anti-tumor function in a setting of chronic antigenic stimulation of CAR-T cells, provided that RSV exposure of CAR-T cells remains continuous.

Discussion

In this study, we sought to explore the impact of RSV culture medium supplementation on T cells in the context of ex vivo immunotherapeutic T cell production.

We first establish that RSV increases oxidative metabolism under basal conditions as well as under artificially enforced conditions, revealing an RSV-boosted increase in SRC. This increase of OXPHOS requires sustained glycolytic flux, is independent of pyruvate production capacity, and is accompanied by an increased flexibility in terms of oxidizable substrate resort. These data are comparable with those previously published by Wenes et al.,⁵⁷ who demonstrated an OXPHOS dependency on glycolysis and an increased fatty acid oxidation potential in the case of mitochondrial pyruvate transport blockade. This observed decoupling between OXPHOS-directed pyruvate availability and glycolysis is prone to be associated with oxidant stress management. Indeed, the latter is triggered by an imbalance between OXPHOS-mediated ROS production and glycolytic byproducts-related scavenging. Glucose deprivation mediates aldolase-induced AMPK and subsequent OXPHOS activation and increases ROS production.⁵⁸ Concomitantly, glucose deprivation unleashes the pentose phosphate pathway and subsequent NADPH production, thus impairing the ROS detoxification potential.^59,60 Therefore, both phenomena may contribute to OXPHOS collapse.

We further demonstrate a decrease in the proliferation capacity of RSV-treated T cells and uncover a greater albeit unexploited glycolysis-associated biosynthetic potential. We could elaborate on the involvement of RSV-induced genotoxic stress and the subsequent decrease in cycling, as reported by Craveiro et al.⁴³ In addition to this proliferation defect, increased RSV-related T cell ATP production occurs, even regarding glycolysis-derived part, suggesting that glycolysis is directed toward energy synthesis rather than toward biosynthesis. The enzymes pyruvate dehydrogenase (PDH) and PDH kinase, which are both involved in pyruvate fate, might be at least partially responsible for this differential commitment. Indeed, Frisch and colleagues⁶¹ described an increase of T cell SRC after inhibition of PDH kinase 1.

At first glance, the RSV-mediated T cell differentiation profile is surprising. Indeed, it has been previously reported that early memory T cells, representing a desirable T cell subset in a setting of adoptive immunotherapy,^8,11,12,13 display a more oxidative and less glycolytic metabolic profile than more differentiated T cells.^14,15,16,17 It is also known that unbalancing glycolysis and OXPHOS resort to the detriment of glycolysis is likely to induce T_SCM accumulation during the ex vivo culture period.^{21,62,63,64,65,66,67,68,69} In our study, RSV medium supplementation led to a decrease in T_SCM rates to the benefit of the T_CM subset, along with glycolytic flux enforcement. Our results suggest that glycolysis vs. OXPHOS imbalance should mainly involve a glycolysis decrease to be related to T_SCM differentiation. An RSV-induced increase of SRC and associated ROS control capacity in T cells are not particularly related to elevated T_SCM rates, impaling a differentiation-independent mitochondrial fate. This highlights a non-systematic coupling between differentiation status and metabolic support. Comparable results were obtained by Gross et al.⁷⁰ in the context of supplementation of CAR-T culture medium with galactose; the OXPHOS and mitochondrial function of these CAR-T cells were increased without significant modification of their differentiation level. An RSV-induced T cell differentiation profile along with metabolic and proliferative attributes points out the importance of the balance between catabolism and anabolism regarding T cell fate. These two metabolic orientations drive opposite functional issues, and OXPHOS and glycolysis definitely do not recapitulate them respectively. Furthermore, RSV treatment induces a decrease in T_EMRA cell rates. The T_EMRA subset includes senescent cells, which are reported to undergo metabolic dysfunction.^49,71 This point might at least partially explain the effect of RSV on T cell metabolic behavior, independent of glycolysis mitigation.

The increased OXPHOS capacity of RSV-induced T cells does not result in an accumulation of intracellular ROS, indicating an efficient ROS scavenging potential. This increased ROS control capacity does not specifically involve NADPH-related oxidative stress control systems. This scavenging activity is obviously sufficient to control ROS excess in T cells, which is induced by the high respiration rate associated with RSV. The explanation might rely on the involvement of an additional ROS-scavenging system, such as the catalase enzyme. Indeed, the glutathione system is the most abundant ROS-reducing system in cells, but it is not the most efficient system for eliminating H₂O₂, with catalase being a plausible candidate for doing so.⁶⁰

Herein, we show that mitochondria quantity and snapshot polarization status are not involved in the RSV-induced increase in OXPHOS. In contrast, increased mitophagic flux is implied, as evidenced by increased PINK1 and BNIP3L/Nix expression, emphasizing sustained mitochondrial turnover. This prompter elimination flux of depolarized mitochondria is unlikely to allow for the detection of convincing differential TMRM staining, and fitter mitochondria content could explain the more efficient oxidative process. Nevertheless, given the lack of an obvious link between the mitochondrial content, the mitochondrial polarization level, and PINK1 activity in the context of short-term RSV treatment of human fibroblasts,⁷² further investigations are needed.

The impact of RSV on T cells relies at least partly on SIRT-1 activity, as EX-527 prevents the effect of RSV on respiration, control of ROS accumulation, and oxidant stress management capacities. As the SIRT-1 protein expression level is not modulated by RSV, SIRT-1 activity, as a result of its oxidized NAD-favored form, is more likely to be involved in its RSV-mediated influence. Notably, SIRT-1 activity has been described as increasing PINK1 expression,³⁹ and PGC-1-α has been reported to amplify BNIP3L/Nix expression.³¹ As such, our results support the SIRT-1 involvement hypothesis. More direct SIRT-1 activity assessment has previously been reported. First, a cleaved form of SIRT-1 has been described and has an impact on its activity.⁷³ Second, an activating liver kinase B1-dependent SIRT-1 phosphorylation has already been reported.⁷⁴ In our experimental context, we did not demonstrate any existing RSV-mediated changes in SIRT-1 cleavage and phosphorylation status (unpublished data), highlighting the possible existence of other regulatory mechanisms. RSV, supplied at a concentration of 5 μM, which represents a relatively low dose according to the literature, does not affect mTORc1 or AMPK activity. The absence of mTORc1 inhibition in cultured T cells is consistent with the high glycolytic flux maintenance. In contrast, in line with its impact on pro-mitochondrial oxidative metabolism, observing an increase in AMPK activity might have been expected. With respect to the redundant effects of SIRT-1 and AMPK on PGC-1α activity, we hypothesize that increased SIRT-1 activity is sufficient to target PGC-1α activation, and this without any substantial AMPK intervention. Nonetheless, as SIRT-1 only partially recapitulates the RSV-mediated effect on T cells, we could elaborate on the implication of other SIRTs like SIRT-3. Indeed, SIRT-3 has already been shown to be activated by RSV; it is mainly localized in the mitochondrial matrix and is involved in mitochondrial homeostasis and ROS scavenging.^75,76

We further establish that advanced age (>50 years) and CMV seropositivity positively influence RSV-induced outcomes. Age and CMV infection are both known to trigger systemic chronic inflammation, which participates in immuno-senescence and further development of T cell-specific dysfunction.⁴⁹ Therefore, age and age-related chronic inflammation are associated with a decrease in the NAD concentration^{37,38,47,71,77} and are likely to drive the breakdown of SIRT-1 activity. This could explain the more important RSV rescue action in this older population in our current setting. Similar mechanisms might be involved in CMV-driven inflammation and immuno-senescence. Notably, the beneficial effect of RSV on mitochondrial metabolism is restricted to CMV-seropositive donors. This highlights the influence of donor history and immune aging-related previous indelible imprinting on T cell fate. Prior T cell exposure to senescence-associated secretory phenotype (SASP) factors is likely to have a particular influence in this setting.^78,79,80 This highlights the critical importance of considering and reporting relevant patient attributes, such as CMV status, in the context of autologous T cell therapy development and clinical evaluation. This would lead to the proposal of safe and valuable production processes adapted to specific patients. With respect to the impact of preexisting chronic CMV infection on T cells, our results suggest a trend toward RSV-mediated T cell terminal differentiation in the early phase of cell culture, followed by T_EMRA-specific RSV-induced purging during the later culture phase. This purge is consistent with the lower T_EMRA rates observed in RSV-cultured T cells. Notably, initial T_EMRA cells are prone to become undetectable during cell culture, either because of a proliferation defect or because of an RSV-related propensity to die, which may be connected to early-stage RSV-induced T cell proliferation impairment and subsequent RSV culture-induced T_EMRA cell disappearance. Early RSV-related non-T_EMRA cell propensity for terminal differentiation can be likened to a phenomenon described by Klebanoff et al.⁸¹ called precocious differentiation. At this stage, the specific role of RSV remains to be deciphered. RSV is likely to promote anti-oxidant or pro-oxidant effects according to its concentration,⁴⁵ the pro-oxidant concentration threshold could be lower in certain types of T cells, such as T_EMRA or senescent T cells. Moreover, previous SASP exposure is known to impair NAD metabolism⁸⁰; thus, it is likely that SIRT-1 activation through RSV treatment might, at least partly, compensate for this negative effect.

Finally, in a setting of CD123-specific CAR-T cells ex vivo generation, in an acute cytotoxicity model, RSV-mediated enhanced oxidative capacities do not drive more potent anti-tumor cytotoxic functions in vitro. In this setting, RSV supplementation does not modify early memory T cells and effector memory (T_EM) T cells rates. Knowing that early memory T cells and T_EM are, respectively, prone to self-renewing and to activation-induced cell death,⁸ this result may explain the comparable behavior of RSV-treated or untreated CAR-T cells in an acute CAR-T reactivation context. Another explanation for the absence of an RSV-induced benefit relative to acute CAR-T cytotoxicity could be in the PGC1-α-mediated mitochondrial fitness prone to drive Tregs differentiation, as reported by Damasceno et al.,⁸² both in mouse and human. In contrast, in a CAR-T chronic stimulation-mimicking context, functional impairment of untreated CAR-T cells tends to occur more abruptly, relative to continuously RSV-treated counterparts. Despite a need to further confirm these first data, this might point out the anti-tumor beneficial impact of RSV and its non-definitive effect on CAR-T cells. We might assume that an increased respiratory capacity allows for greater resistance to metabolic stress in CAR-T cells, and, subsequently, a more durable capacity to control tumor growth. Hope and colleagues⁸³ recently demonstrated that high NAD metabolism positively determines CAR-T cell therapy outcomes. The effect of RSV on NAD redox balance and SIRT-1 activity might be prone to efficiently countervail an immune aging-constitutive defect of NAD cellular quantity and to efficiently rescue CAR-T effector functions.

Taken together, these results lead us to conclude that RSV culture medium supplementation in the context of an immunotherapeutic T cell production process is likely to have a beneficial effect on T cell oxidative metabolism capacity. The positive impact of RSV on T cell oxidative metabolism has therefore been proven, provided that the donor’s intrinsic and immune aging-related parameters, such as CMV status, are properly addressed and harnessed. This dramatically points out the need to consider patient characteristics when assessing new improvements in the T cell production process, pushing the limits of personalized medicine. Notably, in our model, the beneficial effect of RSV on T cell respiratory capacities and oxidative stress control mediates more durable anti-tumor CAR-T cell functions, at least in vitro. Therefore, RSV is likely to improve T cell-based immunotherapies, especially for immune-aging patients.

Materials and methods

Biological materials

PBMCs were collected from healthy donors at the Etablissement Français du Sang as apheresis kits after providing informed consent and according to the collection agreement AC-2020-4129. The CAL-1 cell line, classified as a tumoral plasmacytoid dendritic cell line and known to express CD123 at its surface, was obtained from Dr. Maeda (University of Nagasaki, Nagasaki, Japan) and cultured according to supplier’s instructions. This cell line was periodically checked for Mycoplasma contamination.

T cell isolation, activation, and culture and lentiviral transduction

Healthy donor T cells were magnetically isolated and activated via CD3/CD28 microbeads (111.31D, Fisher Scientific, Illkirch, France) according to the manufacturer’s instructions. Bead-attached T cells were cultured in RPMI-1640 medium (11544526, Fisher Scientific) supplemented with 10% human serum (internal product) in the presence of 500 IU/mL IL-2 (Proleukin, Clinigen Healthcare, Schiphol, the Netherlands) and 5 μM RSV (554325, Sigma-Aldrich, Saint-Louis, MO, USA). When required, 1 μM EX-527 (E7034, Sigma-Aldrich) was added to the RSV culture mixture. This complete medium was renewed every 2–3 days until 14 days of culture, after viable cell counting with the Trypan blue exclusion method. When required, at day 2, activated and cultured T cells were transduced (CAR-T) or not (UTC) using a CD123-specific CAR-T and ΔCD19 selection gene expressing lentiviral construct, as previously described.⁸⁴ Briefly, 0.2.10e6 T cells were spinoculated in lentiviral supernatant (multiplicity of infection = 10) for 1 h at 32°C and 800×g. At day 14, transduction efficiency was assessed through membrane staining with CD3 BV421 (BD Biosciences, La Jolla, CA, USA; 562426), and CD19 APC (Miltenyi Biotec, Bergisch Gladbach, Germany; 130-113-165) antibodies, and analyzed by flow cytometry (FCM). Alternatively, prior to CD3/CD28-mediated activation and sorting, PBMC were isolated via centrifugation on a Ficoll density gradient (CMSMSL01-01, Eurobio, Les Ulis, France) and magnetically sorted via a human Effector T cell isolation kit (130-094-485, Miltenyi Biotec)-derived reagents according to the manufacturer’s instructions. Briefly, labeling with CD8⁺ CD45RA⁺ Effector T cell Biotin Antibody Cocktail, followed by anti-APC-Biotin, and anti-Biotin MicroBeads and magnetic separation via an LD column (130-042-901, Miltenyi Biotec) was used for T_EMRA cell sorting. If any, T_EMRA-isolated cells were stained with 5 μM CellTrace Yellow (C34567, Fisher Scientific) according to the manufacturer’s instructions. The stained T_EMRA cells were mixed with other previously sorted T cell subsets in a proportion related to the initial one.

Cytotoxicity assays

After the initial 14-day culture period, RSV-treated or not CAR-T cells were co-cultured with CAL-1 cells.

Acute stimulation model

CAR-T cells were co-cultured with CAL-1 at different ratios (1:1, 1:5, 1:10, or 1:20) for 24 h before T cell and CAL-1 FCM-based quantification, through a membrane staining using CD3 BV421 and CD123 PE-Cy7 (Sony, Tokyo, Japan; RT2130050). UTCs were co-cultured with CAL-1 at a 1:1 ratio, as a control.

Alternatively, CAR-T cells or UTCs were co-cultured with CAL-1 at 1:5 ratio in the presence of Incucyte Cytotox Green Reagent (Essen Bioscience, Ann Arbor, MI, USA; 4633) and cell lysis was monitored during 56h with an Incucyte S3 Live Cell Analyser (Sartorius, Göttingen, Germany).

Chronic stimulation model

CAR-T cells were co-cultured with CAL-1 (1:15 T cells vs. CAL-1 ratio) for 1 week before T cell and CAL-1 respective FCM-based quantification, through a membrane staining using CD3 BV421 and CD19 APC. Operations were renewed every 7 days until T cells were no longer able to control CAL-1 proliferation.

Assessment of T cell metabolism

Metabolic analysis was performed via Agilent Seahorse technology (Agilent Technologies, Santa Clara, CA, USA; Seahorse XFe96), as previously described.⁸⁵ Briefly, at D_-1, the culture plate was coated with 100 μg/mL poly-D-lysine (A-003-E, Sigma-Aldrich) overnight at 4°C before being rinsed with distilled water and desiccated. Concomitantly, the cartridge sensor (Seahorse mini Fluxpak XFe96, 102601-100, Agilent Technologies) was rehydrated with 200 μL/well distilled water overnight in a 37°C non-CO₂ incubator. At D₀, the distilled water was replaced with 200 μL/well XF Calibrant (100840-100, Agilent Technologies), further incubated for 1 h in a 37°C non-CO₂ incubator. We seeded 200,000 T cells in unbuffered prewarmed XF base medium supplemented or not with 2 mM glutamine, 10 mM glucose, and 1 mM pyruvate (103681-100, Agilent Technologies) into the culture plate and allowed to adhere for 2 h in a 37°C non-CO₂ incubator. Mitochondrial and glycolytic metabolism assays were performed as described in Table S1.

Each Seahorse run step comprised three loops of 3 min of mixing, 2 min of waiting, and 3 min of measurement. Data were acquired and analyzed through Wave 2.6.1 software (Agilent Technologies). The OCR and extracellular acidification rate (ECAR) parameters were measured with a Seahorse XFe96. SRC and glycolytic capacity were calculated by subtracting the baseline values from the BAM-15-induced maximal OCR and oligomycin-induced maximal ECAR, respectively. The ATP production rate (APR), including glyco-APR and mito-APR, were calculated according to Agilent Technologies via the following formulas:

Glyco-APR = (ECAR_baseline × BF × vol × Kvol) − [(OCR_baseline − OCR_{rotenone/antimycin A}) × CCF]

Mito-APR = (OCR_baseline − OCR_oligomycin) × 2 × P/O

where, under our assay conditions, the buffer factor (BF) = 2.5, microchamber volume (vol) = 2.28, volume scaling factor (Kvol) = 1.6, CO₂ contribution factor (CCF) = 0.61, and number of molecules of ATP synthesized per atom of O reduced by an electron pair (P/O) = 2.75.

FCM-based assays and staining

Membrane staining

We stained 200,000–500,000 T cells with fixable viability dye-eFluor 780 (65-0865-18, Fisher Scientific) for 10 min at 4°C, and the concentrations of the monoclonal antibodies were adjusted for 20 additional minutes before extensive washing with 1× PBS (11530546, Fisher Scientific) followed by FCM analysis. Appropriate isotypic controls were included in all the staining designs. The panels composition is summarized in Table S2. The “Differentiation (1)” panel was acquired on an LSR Fortessa flow cytometer (BD Biosciences) and analyzed with BD FACS Diva 9.0 software. The “Differentiation (2)” panel was acquired on a Cytoflex LX (Beckman, Brea, CA, USA) and analyzed with FlowJo 10.10.1 software.

Mitochondria-associated parameters

Concomitant with CD3 BV421 and CD8 BV510 membrane staining, T cells were labeled with either 50 nM MitoTracker Green reagent (17501655, Fisher Scientific), 50 nM TMRM (15656139, Fisher Scientific), or 0.1 μM H2DCFDA (11560326, Fisher Scientific) for 30 min at 37°C before extensive washing in PBS 1× and data acquisition. Notably, H2DCFDA staining was preceded if any by a 2-h contact with 100 μM H₂O₂ and 2 additional resting hours after removing H₂O₂.

MitoTracker Green staining were acquired on a CANTO II flow cytometer (BD Biosciences) and analyzed with BD FACS Diva 9.0 software. H2DCFDA and TMRM data were acquired on either a CANTO II or a Cytoflex LX and analyzed with either BD FACS Diva 9.0 or FlowJo 10.10.0, respectively.

Biochemical analysis

Western blotting

The cells were collected in Tris-buffered saline (TBS) and resuspended in lysis buffer (50 mM Tris/HCl, pH 7.4; 150 mM NaCl; 0.1% SDS; 1% Nonidet P-40; 0.5% Na deoxycholate; and 1 mM EDTA) supplemented with protease inhibitor (4693132001, Sigma-Aldrich). After sonication, the protein concentration was determined with a Pierce BCA Protein Assay Kit (A65453, Fisher Scientific). Then, 30 μg of protein sample was subjected to SDS-PAGE and transferred onto Hybond TM-P polyvinylidene difluoride membranes (10600021, Dutscher, Bernolsheim, France). The membrane was blocked for 2 h at 4°C with 0.1% Tween 20 in TBS (TBS-Tween) supplemented with 5% nonfat milk and incubated overnight with primary antibodies. After being washed with TBS-Tween, the membrane was incubated with appropriate HRP-conjugated secondary antibodies for 1 h and visualized via chemiluminescence with Pierce ECL2 Western Blotting Substrate (32132, Fisher Scientific) via ChemiDoc XRS+ with Image Lab software (Bio-Rad, Hercules, CA, USA). Protein expression was quantified via the ImageJ tool.

All the primary antibodies used for western blotting are listed in the Table S3, available as a supplemental file.

NAD⁺ and NADH respective cellular content quantification

NAD and NADH were extracted as described by Lu et al.⁸⁶ Briefly, a mixture of 40:40:20 acetonitrile:methanol:0.1 M formic acid was added to the cell pellet, which was vortexed and laid on ice for 3 min. Fifteen percent ammonium bicarbonate was then added to neutralize the sample. The samples were vortexed, cooled on dry ice for 20 min and then centrifuged at 16,000×g, after which the supernatants were harvested for liquid chromatography mass spectrometry analysis. LC-MS analysis was performed via a method derived from the method reported by Ortmayr et al.⁸⁷ with an Atlantis T3 column and with mobile phases containing 99:1 water:acetonitrile (0.1% formic acid) and 99:1 acetonitrile:water (0.1% formic acid). Analysis was performed on a high-resolution mass spectrometer (Exploris 120; Fisher Scientific).

NADPH-based anti-oxidant cellular capacity quantification

NADPH-based quantification of anti-oxidant capacity, reduced and total glutathione levels, glutathione reductase activity, and peroxidase enzymatic activity were evaluated in T cells via the Total Antioxidant Capacity Assay Kit (ab65329, Abcam, Cambridge, UK), the Glutathione Fluorescent Detection Kit (EIAGSHF, Fisher Scientific), the Glutathione Reductase Fluorescent Activity Kit (EIAGRF, Fisher Scientific), and the Glutathione Peroxidase Activity Assay Kit (fluorimetric) (ab219926, Abcam), according to the supplier’s respective recommendations. All samples and standards were run in duplicate. The absorbance or fluorescence of the reaction product was detected at appropriate wavelengths by a CLARIOstar Plus device (BMG Labtech, Ortenberg, Germany).

Quantitative real-time PCR

Total mRNA from 3.10E6 T cells was extracted with the RNeasy Mini Kit (74106, Qiagen, Hilden, Germany) according to the manufacturer’s instructions. cDNA was subsequently synthesized via the Advantage RT-for-PCR Kit (639506, Takara-Bio, San Jose, CA, USA) and subjected to quantitative PCR analysis via the TB Green Premix Ex Taq II (Tli RNase H Plus) Kit (RR820L, Takara-Bio) and CFX96 Real-Time PCR Detection System (Bio-Rad). The data were analyzed according to the 2^−ΔΔCt method, with the reference sample being 24 h RSV-treated T cells. The sequences of primers used to link the SIRT-1 gene and the actin housekeeping gene are described in Table S4.

Statistical analysis

Appropriate statistical analyses were performed for all mean comparisons, as mentioned case by case in the associated figure captions. Overall, normal distribution and homogeneous variance demonstrating-data, assessed with a Shapiro-Wilk test and a Fisher test, respectively, were analyzed according to a Student’s t test (paired or unpaired; one tailed or two tailed). In the case of heterogeneous variance, a Wilcoxon test was preferentially realized in the setting of paired data. In the case of a non-normal distribution, a Welch’s correction is brought to the Student’s t test. Contingency-related data were analyzed according to a chi-squared test. p values of greater than 0.10, between 0.05 and 0.10, and less than 0.05 were considered as statistically unsignificant, as a trend to statistical significance, and as statistically significant, respectively.

Data availability

All supporting data, including raw files, are available upon reasonable request and with permission (contact the corresponding author).

Acknowledgments

The authors thank Dr. Brigitte Bonnin from the society Diaclone for her help regarding Clariostar Plus device use and Dr. Gwenaël Rolin for his help regarding Incucyte S3 device use. The authors thank “Ligue contre le cancer” foundation and “MiMEDI” consortium for financial support.

Author contributions

P.M.L. designed the project, performed and analyzed experimentations, and wrote the manuscript; C.M. brought a critical analysis of the manuscript and contributed to its improvement; B.R. developed and performed NAD⁺/NADH dosage; B.D. optimized and performed western blotting assays; P.P. brought a critical analysis of mitochondria-related assays; O.A., J.G., and Y.G. supervised the work.

Declaration of interests

The authors declare no competing interests.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used CURIE in order to improve written English language. After using this tool, the authors reviewed and edited the content as needed and takes full responsibility for the content of the publication.