P2RX7 Enhances Tumor Control by CD8⁺ T Cells in Adoptive Cell Therapy

The purinergic receptor P2RX7 is a ligand-gated ion channel that recognizes extracellular ATP (eATP; ref. 1). P2RX7 stimulation induces Ca²⁺ influx and K⁺ efflux, but prolonged exposure to very high eATP concentrations results in the formation of a macropore that ultimately induces cell death (1, 2). Elevated concentrations of eATP can act as a “danger signal” during infection or acute tissue damage, but eATP also accumulates in many solid tumors and is a component of the tumor microenvironment (TME; ref. 3). In myeloid cells, P2RX7 stimulation leads to the assembly of the NLPR3 inflammasome and promotes immunogenic cell death, innate immune activation, and enhanced presentation of tumor antigens (4). Accordingly, previous studies have shown that host deficiency of P2RX7 leads to accelerated growth of B16 melanoma tumors, which is accompanied by reduced lymphocyte infiltration and decreased intratumoral inflammatory cytokines (5, 6). As a result of these studies, P2RX7 is thought to promote antitumor immune responses. However, some tumors express conformationally altered forms of P2RX7 that may promote their survival in the eATP-rich TME, whereas P2RX7 expressed by tumor-infiltrating immune cells makes them vulnerable to macropore formation and death (7). Hence, the impact of P2RX7 on tumor growth and control is complex.

P2RX7 is also expressed by numerous T-cell populations and is important for the metabolic fitness and durability of memory CD8⁺ T cells following acute and chronic infections (8, 9). Parallels between CD8⁺ T-cell responses in chronic infections and tumors suggest that P2RX7 may have a similarly beneficial role in CD8⁺ T cell–mediated antitumor immunity. On the other hand, it has been proposed that the abnormally high eATP concentrations in the TME could instead result in P2RX7-mediated T-cell death (7). Indeed, using an adoptive cell therapy (ACT) approach, one study suggests that P2RX7 expression by mouse CD8⁺ T cells limits their ability to control the growth of B16 melanoma (10). However, that study used CD8⁺ T cells that were expanded with IL2 alone, and it is known that the inclusion of IL12 during CD8⁺ T-cell stimulation substantially improves tumor control in B16 melanoma models, mediated through increased production of CD8⁺ effector cytokines and decreased induction of inhibitory receptors associated of T-cell exhaustion (11, 12). Hence, it is unclear whether P2RX7 limits or enhances the immunotherapeutic potential of ACT using optimally activated CD8⁺ T cells.

T cell–based immunotherapies have already led to impressive outcomes in the clinic for patients with a variety of malignancies (13–15). However, there are still barriers hindering these therapies from reaching their full potential, and a pressing need to better understand how the TME impairs T-cell accumulation, function, and survival. Determining whether the expression and stimulation of P2RX7 on tumor-specific CD8⁺ T cells is an asset or detriment for tumor immunotherapy is important to inform strategies that could enhance the efficacy and durability of T cell ACTs. It is becoming clear that mitochondrial function is compromised in tumor-infiltrating T cells, and that strategies that enhance the levels of functional mitochondria improve ACT by CD8⁺ T cells (16–20). Our previous work shows that P2RX7 signaling in long-lived memory CD8⁺ T-cell subsets is essential for normal mitochondrial homeostasis and function, with P2RX7 deficiency leading to impaired mitochondrial maintenance, function, and ultrastructure; dysregulated oxidative phosphorylation (OXPHOS); and altered levels of intracellular ATP and AMP (8, 9). However, there is evidence that the different levels of P2RX7 stimulation may enhance or inhibit mitochondrial maintenance (1, 21), so it is currently unclear how the presence of P2RX7 affects mitochondrial function and T-cell survival in the eATP-rich TME of solid tumors (7).

In this study, we demonstrate a CD8⁺ T cell–intrinsic role for P2RX7 in promoting antitumor immune responses, using the B16 melanoma model. This effect was observed for CD8⁺ T cells activated in the presence of IL12, which substantially enhances tumor control and simultaneously leads to increased P2RX7 expression. P2RX7-deficiency resulted in impaired mitochondrial homeostasis of transferred CD8⁺ T cells, in keeping with previous studies (8, 9). Furthermore, P2rx7^–/– CD8⁺ T cells displayed elevated expression of certain exhaustion markers, increased apoptosis, reduced proliferation, and selective loss as the tumor burden increased. Additionally, we found that transient in vitro stimulation of CD8⁺ T cells with a P2RX7 agonist in place of IL12 also improved ACT and enhanced mitochondrial function while reducing the expression of some exhaustion markers. These results indicate that P2RX7 expression and stimulation can be leveraged to promote optimal CD8⁺ T cell–mediated antitumor immunity.

Supplementary Table S1 provides information on the reagents used in this study, including RRID.

Female 6- to 8-week-old adult C57BL/6 and B6.SJL mice were purchased from Charles River (via the NCI). P2rx7^–/– mice were purchased from Jackson Laboratories and backcrossed to the P14 (LCMV-gp33/D^b specific) and OT-I (OVA/K^b specific) TCR transgenic backgrounds, provided by R. Ahmed (Emory University) and K. Hogquist (University of Minnesota), respectively. Thy-1 and CD45 were introduced as congenic markers to distinguish wild-type (WT) and P2rx7^–/– cells using B6.SJL (CD45.1) or B6.PL (Thy1.1) mice purchased from either Charles River or Jackson Laboratories. Animals were maintained under specific pathogen-free conditions at the University of Minnesota. All experimental procedures were approved by the Institutional Animal Care and Use Committee at the University of Minnesota. Animals were randomly assigned to experimental groups.

B16.gp33 cells (provided by Dr. Ananda Goldrath, University of California, San Diego, in 2018) and B16.OVA cells (provided by Dr. Matt Mescher, University of Minnesota, in 2003) were cultured in RPMI-1640 (Corning, #10–040-CV) supplemented with 10% FBS (Atlas Biologicals, #FS-0500-AD), 100 U/mL penicillin/streptomycin (Gibco, #15070063), and 2 mmol/L L-glutamine (Corning, #25–005-CI). After thawing cell lines, B16 cells were expanded for 3 days and passaged once prior to transfer into mice. Cell lines did not undergo Mycoplasma testing. Authentication of OVA or gp33 expression on tumor cell lines was confirmed by in vitro or in vivo recognition by OT-I or P14 CD8⁺ T cells, respectively.

Naïve CD8⁺ T cells were isolated from the spleens of WT or P2rx7^–/– P14 or OT-I transgenic mice with the mouse-naïve CD8⁺ T-cell isolation kit (Miltenyi Biotech, #130–096–543). A total of 2.5×10⁵ isolated T cells were stimulated in flat-bottom 24-well plates with anti-CD3 (10 μg/mL; Bio X Cell, #BE0001–1), anti-CD28 (20 μg/mL; Bio X Cell, #BE0015-1), human IL2 (2.5 IU/mL; PeproTech, #200-02), and with or without murine IL12 (5 ng/mL; R&D Systems, #419-ML-050/CF) for 72 hours in RPMI-1640 medium supplemented with 10% FBS, 100 U/mL penicillin/streptomycin, and 2 mmol/L L-glutamine at 37°C/5% CO₂. For P2RX7 agonism experiments, isolated CD8⁺ T cells were similarly activated with anti-CD3/-CD28 and IL2, and then 100 μmol/L of BzATP (Sigma-Aldrich, #B6396–25MG) or dH₂O as a vehicle treatment was added every 24 hours to the cultures for the 72-hour activation. After 72 hours, 2.5×10⁵ to 5.0×10⁵ activated T cells were transferred intravenously (i.v.) into recipient mice (see In vivo tumor experiments).

Mice were injected subcutaneously with 1.5×10⁵ to 3.0×10⁵ B16.gp33 cells or B16.OVA cells in the right flank. After tumors became palpable (∼7 days), tumor-bearing mice received 5.0×10⁵ activated P14 or OT-I T cells (or 2.5×10⁵ of each for cotransfer experiments). Tumor growth and survival were monitored by measuring height and width every other day with calipers until mice reached an endpoint criterion of 100 mm² or ulceration. Where indicated in the figure legends, mice were treated with a P2RX7 antagonist A-438079 (80 mg/kg; Abcam #ab120413), or PBS as a vehicle treatment, intraperitoneally (i.p.) every other day beginning when tumors became palpable (∼7 days after tumor injection).

Mice were sacrificed at times indicated in the figure legends, and spleens, tumor-draining inguinal lymph nodes (dLN), and tumors were harvested and homogenized. In all experiments involving the isolation of lymphocytes from tumors, 50 μg of Treg-Protector (anti-ARTC2.2) nanobody (BioLegend, #149802) was injected intravenously (i.v.) 30 minutes prior to mouse sacrifice as previously described (8, 9). Tumors were removed, cut into small pieces, and digested with 3 mg/mL Collagenase type I (Worthington #LS004197) solution for 1 hour at 37°C and then dissociated via gentleMACS Dissociator (Miltenyi Biotec) twice. Spleens and dLNs were homogenized and passed through a 70-μm cell strainer. Isolated mouse cells were stained with antibodies specific for CD8α (clone 53-6.7, BD Biosciences), CD44 (clone IM7, Tonbo Biosciences), CD69 (clone H1.2F3, BioLegend), CD103 (clone M290, BD Biosciences), P2RX7 (clone 1F11, BD Biosciences), CD279 (PD-1; clone 29F.1A12, BioLegend), CD366 (Tim3; clone RMT3–23, BioLegend), CD223 (Lag3; clone C9B7W, BioLegend), CD39 (clone Duha59, BioLegend), TCF1 (clone C63D9, Cell Signaling Technologies), TOX (clone TXRX10, Thermo), Ki67 (clone SolA15, Thermo), IFNγ (clone XMG1.2, Tonbo Biosciences), granzyme B (clone GB11, Thermo Fisher Scientific). All cells were stained at antibody dilutions of 1:200, except for granzyme B (1:100), IFNγ (1:100), TCF1 (1:100), TOX (1:50), and P2RX7 (1:50). Cells stained intracellularly for granzyme B, IFNγ, Ki67, TCF1, and Tox were permeabilized using the FoxP3/Transcription Factor fixation/permeabilization kits (Tonbo Biosciences, #TNB-0607-KIT) according to the manufacturer's protocols. Cell viability was determined with Ghost Dye 780 (Tonbo Biosciences, #13-0865-T100). Enumeration of lymphocytes was achieved using CountBright Absolute Counting Beads (Invitrogen, #C36950).

For ex vivo restimulation assays, digested/homogenized tumors, dLNs, and spleens were incubated with 2 μL/mL PMA/ionomycin cell stimulation cocktail (Tonbo Biosciences, #TNB-4975-UL400) for 4 hours at 37°C in RPMI-1640 supplemented with 2% FBS. We then proceeded with extracellular/intracellular staining as described above.

For mitochondrial mass and membrane potential measurements, 1–2×10⁶ cells were incubated with 100 nmol/L MitoTracker Green (MTG; Thermo Fisher Scientific, #M7514) and 80 nmol/L tetramethylrhodamine ethyl ester (TMRE; Thermo Fisher Scientific, #T669) simultaneously for 20 minutes in RPMI-1640 supplemented with 2% FBS at 37°C prior to ex vivo staining. In order to account for experiment-to-experiment variability, geometric mean fluorescence intensities (gMFI) of either MTG or TMRE in donor WT CD8⁺ P14 cells within the spleen of each experiment were averaged, and fold change of donor WT or P2rx7^–/– CD8⁺ T cells within the tumor or tumor dLN for each experiment was calculated relative to splenic WT donor cells. Mitochondrial reactive oxygen species (ROS) were evaluated by incubating 1–2×10⁶ cells with 5 μmol/L MitoSOX reagent (Thermo Fisher, #M36008) in RPMI-1640 supplemented with 2% FBS at 37°C prior to ex vivo staining.

Apoptosis was analyzed using an Annexin V Apoptosis Detection Kit I (BD #556454) according to the manufacturer's instructions. Briefly, after extracellular staining, cells were washed 2× in PBS (without serum) and resuspended in Annexin V binding buffer. Approximately 15 to 30 minutes prior to reading the samples, Annexin V–APC and propidium iodide (PI) staining solution was added directly to the sample in Annexin V binding buffer. Apoptotic cells were discriminated from dead/necrotic cells by PI and viability Ghost Dye staining (i.e., Annexin V⁺ LD/PI^–).

Flow-cytometric analysis was performed on an LSR II or LSRFortessa (BD Biosciences), and data were analyzed using FlowJo software v10.8 (Treestar).

Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were assessed in in vitro–activated WT and P2rx7^–/– T cells (activated with or without IL12 or BzATP for 72 hours) using a 96-well XF Extracellular flux analyzer, according to the manufacturer's instructions (Seahorse Bioscience), similar to earlier studies (8, 9). Briefly, 2×10⁵ activated cells were plated onto Seahorse XFe96 (Agilent) plates coated with CellTak (Corning, 354240) solution the night before in warm Seahorse base media (Agilent, #102353-100) supplemented with 2 mmol/L L-glutamine, 1 mmol/L sodium pyruvate (Corning, #25-000-Cl), and 10 mmol/L glucose (Sigma, G8769-100ML; pH 7.4) the day of assay. Seahorse XF Mito Stress kit (Agilent #103015–100) components (1.5 μmol/L oligomycin, 1 μmol/L FCCP, 0.5 μmol/L Rotenone/Antimycin A) were added to hydrated 96-well sensor cartridge (incubated overnight at 37°C with Agilent calibrant solution) before performing the assay. Spare respiratory capacity (SRC) and OCR and ECAR values were defined as previously described (16, 22).

The killing of B16 melanoma cells was evaluated in real-time using the IncuCyte platform. Activated WT or P2rx7^–/– P14 CD8⁺ T cells were plated into 96-well flat clear-bottom tissue culture–treated microplates, previously coated with Poly-L-ornithine (Sigma-Aldrich #A-004-M) for 2 to 3 hours, along with CellTrace Far Red (Thermo Fisher Scientific, #C34564) stained (according to the manufacturer's instructions) B16.gp33 or B16.OVA cells at 1:1 and 3:1 effector:target ratios. Caspase-3/7 green dye (5 μmol/L; Sartorius #4440) was added just prior to beginning the assay to identify dying cells. The plate was then placed in an IncuCyte ZOOM platform housed inside a cell incubator at 37°C/5% CO₂. Images from 5 technical replicates were taken every 30 minutes for 48 hours using a 4× objective lens and then analyzed using IncuCyte Basic Software v2018A (Sartorius). Graphed readouts represent the percentage of live B16 melanoma targets (CellTrace Far Red⁺Caspase-3/7^–), normalized to live targets alone at the starting (0 hours) time point.

Resting CD8⁺ T cells were isolated from Ficoll-purified PBMC isolates from leukapheresis products (Memorial Blood Center) using a naïve human CD8⁺ negative selection kit (Stem Cell, #19258). Purified cells were stimulated for 72 hours with anti-CD3/-CD28beads (Gibco #11161D) with 4 ng/mL human IL2, and with or without 50 ng/mL IL12 (R&D Systems, #219-IL-005/CF). After 72 hours, activated CD8⁺ T cells continued to be cultured with concentrations of IL2 ± IL12 indicated in the figure legends until RNA expression analysis.

Human CD8⁺ T cells activated with IL2 ± IL12 were first homogenized using QIAshredder columns (QIAGEN), and RNA was extracted using RNeasy kit (Qiagen, #79654) following the manufacturer's instructions. From isolated RNA samples, cDNA was prepared with the SuperScript III First-Strand Synthesis SuperMix (Thermo Fisher Scientific, #11752050) following the manufacturers’ instructions. Gene expression was assessed (in triplicate) with an ABI 7700 sequence-detection system, and amplification was detected with SYBR Green PCR Master Mix (Applied Biosystems #4309155) as previously described (23). The sequences of primer pairs used to measure P2RX7 RNA were: forward, 5′-TGT GTC CCG AGT ATC CCA CC-3′, reverse, 5′-GGC ACT GTT CAA GAG AGC AG-3′; the primers for glucuronidase beta (GUSB; used as a control housekeeping gene; ref. 24) were: forward, 5′-GTC TGC GGC ATT TTG TCG-3′, reverse, 5′-CAC ACG ATG GCA TAG GAA TGG-3′.

Data were subject to the Kolmogorov–Smirnov test to assess the normality of samples. Statistical differences were calculated by using a paired or unpaired two-tailed Student t test or one-way ANOVA with Tukey post-test. In cases where one group in the comparisons was normalized, statistical analysis used ANOVA with Brown–Forsythe and Welch corrections. Statistical differences for survival data were assessed using a log-rank Mantel–Cox test. All statistical analyses were performed using Prism 9 (GraphPad Software). Graphical data are shown with error bars indicating the SD or SEM. P values of <0.05 (*), <0.01 (**), <0.001 (***), or <0.0001 (****) indicate significant differences between groups.

The data generated in this study are available within the article and its supplementary data files or upon request from the corresponding author Stephen C. Jameson (james024@umn.edu).

Previous studies suggested that P2RX7 expression is necessary for the host to elicit an effective antitumor immune response (5, 6). In striking contrast, however, one report has indicated that P2RX7-deficient CD8⁺ T cells exhibit enhanced tumor control (10). To investigate this further, we adoptively transferred either WT or P2rx7^–/– tumor-specific transgenic CD8⁺ T cells into mice with B16 melanoma tumors (the model used by Romagnani and colleagues; ref. 10). Briefly, B16 melanoma cells expressing gp33 (B16.gp33), the epitope recognized by P14 CD8⁺ T cells, were injected subcutaneously into WT hosts. Once tumors became palpable, either WT or P2rx7^–/– P14 CD8⁺ T cells were administered i.v. after 72 hours of in vitro activation (Fig. 1A). WT and P2rx7^–/– P14 were activated with IL12 as a tertiary signal (in addition to anti-CD3/CD28 and IL2) to optimize antitumor immune responses, as previously described (11, 12). Survival of tumor-bearing mice was significantly improved by adoptive transfer of WT P14 cells, compared with transfer of P2rx7^–/– P14 cells (Fig. 1B). Tumor burden was also substantially reduced in mice that received WT P14 cells, with some mice having fully controlled tumors for >40 days (lines indicated by # in Fig. 1C; Supplementary Fig. S1a). In contrast, tumor control was markedly impaired in mice transferred with P2rx7^–/– P14 cells (Fig. 1B and C). The effect of P2RX7 deficiency was confirmed using OT-I CD8⁺ T cells, which were transferred into mice with B16 tumors that express ovalbumin (B16.OVA; Supplementary Fig. S1b), paralleling the system used by Romagnani and colleagues (10). Despite these differences in tumor control between WT and P2rx7^–/– P14 cells, there was no difference in the ability of the in vitro activated populations to produce IFNγ (Supplementary Fig. S1c) or kill B16 target cell lines (Supplementary Fig. S1d), in keeping with previous studies indicating that P2rx7^–/– CD8⁺ T cells differentiate normally into effector cells (8, 9).

A potentially important difference between our studies and those of Romagnani and colleagues (10) was their omission of IL12 in the CD8⁺ T-cell activation protocol. Indeed, in parallel studies we found that the advantage we observed for WT P14 cells in controlling tumors (Fig. 1B and C) was lost if IL12 was not included during CD8⁺ T-cell stimulation, with the transfer of either WT or P2rx7^–/– P14 cells only modestly improving survival of tumor-bearing mice compared with mice that received no CD8⁺ T-cell transfer (Fig. 1D; Supplementary Fig. S2a and S2b). Similar outcomes were observed in the OT-I/B16-OVA model (Supplementary Fig. S2c). Consistent with Romagnani and colleagues (10), P2rx7^–/– P14 cells stimulated without IL12 conferred slightly improved survival relative to WT P14 cells (Fig. 1D), although this was not statistically significant in our hands. IL12 exposure has been reported to increase P2RX7 expression on CD8⁺ T cells (25), and we confirmed this on P14 cells assessed at the end of in vitro activation (Supplementary Fig. S2d). Similarly, P2RX7 gene expression by human CD8⁺ T cells was enhanced by culture with IL12 (Supplementary Fig. S2e). This is consistent with the hypothesis that the P2RX7-mediated contribution to tumor control may be amplified by IL12-induced expression of P2RX7.

To determine how P2RX7 deficiency affects donor CD8⁺ T cells transferred into tumor-bearing mice, we harvested B16.gp33 tumors, dLNs, and spleens from mice 7 days after transfer of IL12-activated WT or P2rx7^–/– P14 cells. At this time point, tumor burden was only slightly greater in mice that received P2rx7^–/– P14 cells as compared with recipients of WT P14 cells (Supplementary Fig. S3a), providing an opportunity to identify phenotypic differences between WT and P2rx7^–/– P14 cells, which may correspond with the failure of P2rx7^–/– P14 cells to effectively control tumors. This analysis revealed fewer P2rx7^–/– P14 cells per gram of tumor and in the dLN and spleen, compared with WT P14 cells (Fig. 2A), despite no significant differences in the percentage of total tumor-infiltrating CD8⁺ T cells (Supplementary Fig. S3b). Although tumor control by P2rx7^–/– CD8⁺ T cells was impaired, WT and P2rx7^–/– P14 cells did not differ significantly in the expression of exhaustion markers Tim3 and PD-1 within the tumor. However, in the dLN, the frequency of Tim3-expressing cells and PD-1 expression levels were elevated in the P2rx7^–/– P14 population relative to WT P14 cells (Fig. 2B and C; Supplementary S3d and S3e). This finding is important in light of studies showing that cells in the dLN may constitute a reservoir for sustaining the population of antigen-specific CD8⁺ tumor-infiltrating lymphocytes (26–28). Elevated expression of exhaustion markers may suggest a recent antigen encounter or, perhaps, approaching T-cell dysfunction/exhaustion.

We found that P2rx7^–/– P14 cells within the dLN exhibit effector T-cell qualities, as evidenced by an increased frequency of cells expressing granzyme B (GzmB), despite no difference in GzmB expression among the cells within the tumor (Fig. 2D). Moreover, tumor-infiltrating P2rx7^–/– P14 cells were functional, efficiently producing IFNγ after ex vivo restimulation with PMA/ionomycin (Supplementary Fig. S3c). Together, these data suggest that the P2rx7^–/– P14 cells within the tumor are slightly reduced in number but do not show overt signs of exhaustion. However, P2rx7^–/– P14 cells within the dLN, which presumably continue to seed the tumor over time, exhibit effector-like characteristics but are beginning to increase the expression of markers associated with T-cell dysfunction, which may ultimately compromise the ability of P2rx7^–/– P14 cells to sustain an effective antitumor response.

We have previously shown that P2rx7^–/– CD8⁺ T cells have dysregulated metabolism in an acute infection model (8). Given the correlation between T-cell metabolic fitness and efficacy of antitumor responses (29), we were interested in whether P2rx7^–/– P14 cells have altered mitochondrial dynamics in response to metabolic stressors in the TME. We measured mitochondrial mass and membrane potential of WT and P2rx7^–/– P14 cells from spleens, dLNs, and tumors of B16.gp33-engrafted mice by staining for MTG and TMRE, respectively (Fig. 3A–C). The TMRE/MTG ratio, indicating mitochondrial activity efficiency, was decreased in P2rx7^–/– P14 cells within the tumor, dLN, and spleen, relative to WT P14 cells (Fig. 3C). This suggests that, despite both WT and P2rx7^–/– P14 cells increasing mitochondrial mass in response to cues in the TME, the mitochondria in tumor-infiltrating WT P14 cells are more functional than the mitochondria in tumor-infiltrating P2rx7^–/– P14 cells.

Additionally, we observed a population of P14 cells that had increased mitochondrial mass (MTG^hi) but very low mitochondrial membrane potential (TMRE^lo; Fig. 3D). This population (MTG^hiTMRE^lo) likely represents cells that accumulate “depolarized” and nonfunctional mitochondria, and this is known to occur at a higher frequency among CD8⁺ T cells within the TME (19). Consistent with this, we found that both WT and P2rx7^–/– P14 cells from tumors of some B16.gp33 engrafted mice had a higher percentage of MTG^hiTMRE^lo cells compared with P14 cells from the dLNs and spleens. Despite high variability among samples, this trended higher for tumor-infiltrating P2rx7^–/– P14 cells compared with WT (Fig. 3D). Together, these results suggest that tumor-infiltrating P14 cells that lack P2RX7 have dysfunctional mitochondria, which has been linked to impaired CD8⁺ T-cell antitumor responses.

Differences in initial tumor control by WT and P2rx7^–/– CD8⁺ T cells could potentially affect the phenotype or function of these cells even at early time points, due to an altered TME. Hence, we next investigated whether differences seen for WT and P2rx7^–/– CD8⁺ T cells would be recapitulated when both populations were present in the same TME. Equal numbers of WT and P2rx7^–/– OT-I cells were cotransferred into mice with palpable B16.OVA tumors. This approach permitted the evaluation of cell-intrinsic effects of P2RX7 deficiency, whereas the cotransferred WT cells are expected to limit tumor growth, allowing for the evaluation at later time points after ACT. Seven and 18 days later, the ratio and phenotype of cells infiltrating the tumor, dLN, and spleens of B16.OVA engrafted mice were analyzed. Between these time points, the tumor size increased considerably, as expected (Fig. 4A). At day 7, WT and P2rx7^–/– OT-I cells were present at equal frequencies within the tumors and secondary lymphoid organs (SLO; Fig. 4B; Supplementary Fig. S4a). However, by day 18, we observed a preferential loss of P2rx7^–/– OT-I cells within the tumor, dLN, and spleen (Fig. 4B; Supplementary Fig. S4a). Furthermore, although the frequency of cells expressing exhaustion markers was greater among the WT donor population at day 7, by day18 we observed increased expression of PD-1 and Lag3 by the P2rx7^–/– OT-I population (Fig. 4C) and the P2rx7^–/– populations showed an impaired ex vivo production of IFNγ (Fig. 4D). The frequency of the TCF1⁺ “stem-like” population within the tumor initially favored the P2rx7^–/– population, but this advantage was lost by day 18 (Supplementary Fig. S4b). At both time points, we observed significantly increased frequencies of Annexin V⁺ (LiveDead/PI^–) cells among the P2rx7^–/– OT-I cells within the tumor (Fig. 4E), suggesting these cells are preferentially undergoing apoptotic death, potentially contributing to the loss of P2rx7^–/– cells by day 18. P2rx7^–/– OT-I in the tumor and dLN showed reduced proliferation (measured by Ki67 staining) at day 7, and this defect was sustained in the dLN at day 18 (Fig. 4F). Finally, tumor-infiltrating P2rx7^–/– displayed reduced expression of the “residency marker” CD103 within the tumor (Fig. 4G), which may suggest impaired tumor residency. Taken together, these data suggest a progressive loss in function and maintenance among P2rx7^–/– CD8⁺ T cells as time and the tumor burden increase, resulting in increased cell death and reduced representation of P2rx7^–/– CD8⁺ among tumor-infiltrating lymphocytes.

We also assessed the mitochondrial fitness of cotransferred WT and P2rx7^–/– OT-I cells at both time points. At day 18, we observed a significant increase in the frequency of P2rx7^–/– OT-I cells with depolarized mitochondria within the tumor (Fig. 5A). Additionally, consistent with our findings from the single-transfer model, we observed that WT OT-I cells within the tumor and dLN had higher mitochondrial mass and improved membrane potential and activity per mitochondrion at both time points (Fig. 5B–D), suggesting the impaired mitochondrial fitness observed for P2rx7^–/– CD8⁺ T cells was cell-intrinsic. Additionally, P2rx7^–/– CD8⁺ T cells within the tumor at day 18 had significantly lower levels of mitochondrial ROS, consistent with the mitochondrial function being reduced (Fig. 5E). These results demonstrate that the P2rx7^–/– CD8⁺ T-cell population is susceptible to varied mitochondrial defects even when the TME is normalized through the use of a cotransfer approach, indicating that this dysfunction is cell-intrinsic.

It was unclear whether the mitochondrial defects observed in tumor-infiltrating P2rx7^–/– CD8⁺ T cells were established during in vitro activation or were a result of cells encountering antigen in the hypoxic, nutrient-depleted TME. Furthermore, the impact of IL12 on the metabolic activity of cultured CD8⁺ T cells has not been clearly defined. We initially assessed oxidative phosphorylation and glycolysis in WT and P2rx7^–/– P14 cells at 72 hours of stimulation, using extracellular flux “Seahorse” assays. IL12 culture significantly increased the baseline and maximum OCR of WT P14 cells, but not P2rx7^–/– P14 cells (Fig. 6A–C), and it showed a trend toward increased SRC (Fig. 6D). Consistent with the metabolic defects in P2rx7^–/– CD8⁺ T cells being primarily related to mitochondrial function, we observed no differences in the glycolytic capacity of WT and P2rx7^–/– CD8⁺ T cells, as measured by ECAR (Supplementary Fig. S5a–b).

We next determined how this related to mitochondrial homeostasis of in vitro–cultured WT and P2rx7^–/– P14 T cells. IL12 increased mitochondrial mass in both populations, and the effect was significantly more potent for P2rx7^–/– cells, yet the mitochondrial membrane population was not elevated proportionally, such that IL12 exaggerated the differences in relative membrane potential (TMRE/MTG ratio) between WT and P2rx7^–/– T cells (Fig. 6E–G). These results illustrate the ability of IL12 to boost the levels of functional mitochondria in WT CD8⁺ T cells but not to the same extent in P2rx7^–/– cells. These results mirror mitochondrial defects observed in vivo and suggest that P2rx7^–/– P14 are disadvantaged metabolically prior to arrival in the TME, which may compromise the durability of the antitumor immune responses.

Our findings indicated that there may be an opportunity to boost ACT antitumor efficacy by stimulation of P2RX7 in vitro prior to cell transfer. We previously showed that stimulation of CD8⁺ T cells with the ATP analogue BzATP, a P2RX7 agonist, leads to increased cell proliferation and Ca²⁺ influx, without compromising viability (8). Here, we found that BzATP treatment of CD8⁺ T cells, in the absence of IL12, enhanced mitochondrial function, leading to increased OCR and SRC in WT but not P2rx7^–/– CD8⁺ T cells (Supplementary Fig. S6), but did not significantly change the cell-surface expression of P2RX7 (Supplementary Fig. S2d). To test whether P2RX7 agonism could be used to enhance immunotherapy by CD8⁺ T cells cultured in the absence of IL12, we activated P14 CD8⁺ T cells in vitro for 72 hours with 3 doses of 100 μmol/L BzATP (added every 24 hours) or vehicle (H₂O) treatment, and then transferred these cells into mice bearing B16.gp33 tumors, without further BzATP treatment. We found that the addition of BzATP during in vitro activation significantly improved survival and reduced tumor area in comparison with vehicle-treated P14 cells (Fig. 7A–C), albeit less effectively than IL12 cultured cells (Fig. 1). In parallel studies, we found that BzATP treatment had no effect on tumor control mediated by P2rx7^–/– P14 cells, indicating the effect requires BzATP stimulation of P2RX7 (Supplementary Fig. S7a). To determine whether continued P2RX7 activity was needed to sustain antitumor activity, some animals were also treated with the P2RX7 antagonist A438074 (A438) in vivo every other day beginning at the time of T-cell transfer. As previously described in similar studies with P2RX7 antagonists (5, 6), A438 treatment had a modest beneficial effect on the survival of mice that did not receive T cells. However, A438 treatment significantly compromised the ability of BzATP-treated donor WT P14 T cells to enhance survival (Supplementary Fig. S7b). Preliminary studies indicated that BzATP stimulation did not further enhance the efficacy of IL12 cultured WT P14 cells for tumor control (Supplementary Fig. S7c), which may indicate that these treatments operate in the same functional pathway.

To determine the effects of transient BzATP treatment on the phenotype and mitochondrial homeostasis of tumor-infiltrating lymphocytes, tumors from mice were harvested 7 days after T-cell transfer. We observed that BzATP treatment during in vitro activation led to reduced numbers of P14 cells per gram of tumor at the time of harvest (Supplementary Fig. S8a), but increased proliferation of those cells (Fig. 7D). It also reduced the frequency of P14 cells expressing the inhibitory receptors Tim3 (Fig. 7E), but not PD-1 (Supplementary Fig. S8b). Consistent with transient P2RX7 stimulation in vitro promoting mitochondrial function of tumor-infiltrating lymphocytes, BzATP-treated P14 cells in the tumor exhibited higher mitochondrial membrane potential and function per mitochondrion, despite no significant change in mitochondrial mass or change in frequency of MTG^hiTMRE^lo cells (Fig. 7F–H; Supplementary Fig. S8c). However, BzATP-treated P14 cells in lymphoid tissues showed reduced mitochondrial mass and membrane potential (Fig. 7F–H), suggesting in vitro BzATP treatment may preferentially benefit mitochondrial function in tumor-infiltrating cells.

Together, these results demonstrate that P2RX7 stimulation is beneficial for CD8⁺ T cell–mediated antitumor immunity and highlight the potential for the P2RX7 stimulation of cultured CD8⁺ T cells to promote tumor control.

Previous studies have shown that mice that lack P2RX7 display accelerated tumor growth and have fewer tumor-infiltrating lymphocytes compared with WT mice (5, 6). However, P2RX7 is expressed on many cell types, including various immune populations and many tumors (7), making the cell-intrinsic role of P2RX7 difficult to discern from such studies. Indeed, expression of P2RX7 by tumor cells (including B16 melanoma) can contribute to their growth, as indicated by studies showing that P2RX7 pharmacologic blockade limits tumor growth (5, 6). The situation is further complicated by the capacity of high levels of eATP to transform P2RX7 signaling from a nonselective cation channel into generating a lethal macropore; the high eATP concentrations often found in the TME may therefore make P2RX7 expression a liability rather than a benefit for immune cells. Indeed, although we previously showed that P2RX7 expression is beneficial for the establishment and maintenance of CD8⁺ memory T cells following responses to chronic and acute infections (8, 9), one study suggested that P2RX7 was detrimental for CD8⁺ T cell–mediated tumor immunotherapy (10). Nevertheless, it was not clear what role P2RX7 would play in adoptive immunotherapy by CD8⁺ T cells that were optimally stimulated for efficient tumor control. In particular, our studies herein built on the observation that IL12 treatment of tumor-specific CD8⁺ T cells during in vitro activation dramatically improves their persistence and ability to eradicate tumors (11). Indeed, although our data indicate that P2RX7 expression offers no benefit for suboptimally stimulated CD8⁺ T cells, the enhanced tumor control induced by IL12-cultured CD8⁺ T cells was severely blunted by P2RX7 deficiency. P2RX7 expression correlated with the production of functional mitochondria in tumor-infiltrating CD8⁺ T cells, an effect that was also observed following in vitro activation prior to adoptive transfer.

As tumors expand, their increased metabolic demand creates an environment that is extremely hypoxic and depleted of nutrients. This poses an obstacle that infiltrating immune cells must adapt to in order to meet their own metabolic needs. Mitochondrial health and homeostasis are dependent on controlled biosynthesis and autophagic degradation of mitochondria (30). These processes are often impaired in tumor-infiltrating CD8⁺ T cells, resulting in poor mitochondrial quality and the accumulation of damaged mitochondria compared with splenic CD8⁺ T cells (17, 19). Consistent with those studies, we found that mitochondria in tumor-infiltrating CD8⁺ T cells had a lower mitochondrial membrane potential and an increased frequency of depolarized mitochondria. These characteristics were significantly worsened in P2rx7^–/– CD8⁺ T cells within the tumor. P2RX7-deficient CD8⁺ T cells also showed mitochondrial impairments and reduced baseline and maximum levels of OXPHOS after in vitro activation. The glycolytic capacity of in vitro activated P2rx7^–/– CD8⁺ T cells was not impaired compared with WT CD8⁺ T cells. T cells shift their metabolism to predominately utilize aerobic glycolysis upon activation (31). P2RX7-deficient CD8⁺ T cells may be able to compensate for mitochondrial defects initially by relying on glycolysis, but likely lack the plasticity necessary to adapt to changes in nutrient and oxygen availability in the TME.

Metabolic alternations in tumor-infiltrating CD8⁺ T cells are early events that ultimately lead to T-cell exhaustion (19, 32, 33). Furthermore, manipulation of CD8⁺ T cells to optimize mitochondrial function and structure correlates with reduced expression of inhibitory receptors (34–36) and can even restore the functionality of exhausted T cells (18, 37). Consistent with this, we found that intratumoral P2rx7^–/– CD8⁺ T cells exhibited increased expression of some exhaustion markers and reduced production of IFNγ as the antitumor response progressed. These changes occurred alongside continued mitochondrial dysfunction and correlate with selective loss of P2rx7^–/– CD8⁺ T cells within the tumor and dLN. These data reinforce the strong link between mitochondrial fitness and T-cell function and survival. Our data also indicate that mitochondrial function may in some cases be an early indicator of an effective T-cell antitumor response before the onset of exhaustion.

Although our data suggest that the role of P2RX7 in promoting CD8⁺ T-cell responses and metabolic fitness is similar in the context of acute/chronic infections and tumors (8, 9), the numeric defect among tumor-infiltrating P2rx7^–/– CD8⁺ T cells was more subtle than our observations in infection models (8, 9). This may relate to the early time points of analysis required for studies involving the rapidly growing B16 model, but it is also possible that this finding relates to the dual functions of P2RX7: While P2RX7 deficiency may compromise mitochondrial maintenance, overstimulation of P2RX7, due to high expression levels or elevated concentrations of eATP in the TME may induce P2RX7 to form a macropore resulting in the death of WT T cells (1, 38, 39). In normal tissues, eATP concentrations are in the nanomolar range, but in the TME, eATP concentrations can reach hundreds of micromols (3). P2RX7 macropore formation can also be induced by the NAD-mediated ribosylation of P2RX7 by the ectoenzyme ARTC2.2 (2). For this reason, we used nanobody-mediated ARTC2.2 antagonism at the time of harvest, which we have shown improves the recovery of P2RX7-expressing CD8⁺ T cells during tissue processing where NAD might be released from necrotic cells (40). However, this does not protect P2RX7-expressing T cells from high concentrations of eATP within the tumor in vivo. For these reasons, it is likely that the TME may be especially hostile to P2RX7-expressing WT CD8⁺ T cells, yet our data indicate that this is outweighed by their improved ability to control tumor growth. Still, given the complicated nature of P2RX7 signaling, particularly in high eATP/NAD environments where the receptor could be a liability, it is not surprising that there is some discrepancy on the cell-intrinsic role of P2RX7 in the context of CD8⁺ ACT. Furthermore, P2RX7 may have a distinct role on CD4⁺ T cells, where its expression may promote the survival of regulatory T cells (1). Indeed, Romagnani and colleagues observed that P2RX7 deficiency was advantageous for CD4⁺ T-cell survival and tumor control (10). Nevertheless, our findings suggest a need for careful evaluation of how tumor-specific T cells are activated in order to address whether P2RX7 plays a beneficial versus a detrimental role.

In summary, we report that P2RX7-deficient CD8⁺ T cells exhibit increased mitochondrial dysfunction, reduced proliferation, increased levels of apoptosis, and elevated expression of exhaustion markers as tumor burden increases, leading to loss of P2rx7^–/– CD8⁺ donor cells within the tumor and in SLO. We conclude that these combined defects account for impaired tumor control when P2rx7^–/– CD8⁺ T cells are used for ACT. Strategies to boost mitochondrial health CD8⁺ tumor-infiltrating lymphocytes have shown to improve T-cell survival and function under metabolic stressors imposed by the TME (35, 41). Consistent with those studies, we report that transient P2RX7 agonism (with BzATP) increased mitochondrial mass and function in tumor-infiltrating WT CD8⁺ T cells resulting in enhanced tumor control. These data demonstrate that stimulation of the P2RX7 signaling pathway is a promising therapeutic option that could easily be optimized and incorporated into existing ACT protocols. Furthermore, the requirement for P2RX7 to generate an effective antitumor response has clinical relevance as P2RX7 loss-of-function variants are relatively common in the human population (42), and patients with those mutations may require ACT protocols that boost the mitochondrial health of transferred T cells.

K.M. Wanhainen reports grants from NIH NCI during the conduct of the study. H. Borges da Silva reports grants from NIH (NIAID) and the Cancer Research Institute during the conduct of the study. S.C. Jameson reports grants from NIH NIAID during the conduct of the study. No disclosures were reported by the other authors.

K.M. Wanhainen: Conceptualization, data curation, formal analysis, funding acquisition, validation, investigation, writing–original draft. C. Peng: Formal analysis, investigation, methodology. M.H. Zhou: Validation, investigation. B. de Gois Macedo: Validation, investigation. S. O'Flanagan: Formal analysis, investigation, methodology. T. Yang: Investigation, methodology. A. Kelekar: Supervision, methodology. B.J. Burbach: Supervision, investigation, methodology. H. Borges da Silva: Conceptualization, data curation, formal analysis, supervision, funding acquisition, validation, investigation, methodology. S.C. Jameson: Conceptualization, resources, supervision, funding acquisition, methodology, writing–original draft.

This work was supported by the NIH (R01 AI145147 to S.C. Jameson; R00 AI139381 to H. Borges da Silva; F30 CA250231 to K.M. Wanhainen), and H. Borges da Silvawas supported by a Paul C. Shiverick/CRI Irvington fellowship. We thank Ananda Goldrath (UCSD) and Matthew Mescher (UMN) for providing us with the B16.gp33 and B16.OVA cell lines, respectively, and thank Peter Hinderlie and Jeffrey Miller (UMN) for their assistance with the IncuCyte assays. We appreciate all members of the Jamequist lab for input on these studies and critical comments on the manuscript.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.