ARTC2.2/P2RX7 Signaling during Cell Isolation Distorts Function and Quantification of Tissue-Resident CD8⁺ T Cell and Invariant NKT Subsets

This article is featured in In This Issue, p.1909

During preparation of ex vivo cell suspensions or in response to microbial infections, inflammation, or tumor growth, high concentrations of the nucleotides ATP and NAD can be released from apoptotic, necroptotic, or stressed cells (1). Extracellular ATP (eATP) stimulates P2RX7, which is a nonselective ligand-gated ion channel expressed by several immune cell types. Prior research focused on myeloid cells (2, 3), but P2RX7 is also expressed by lymphocyte populations (4–7). When activated by high concentrations of eATP, P2RX7 forms reversible nonselective pores that can mediate activation signals but can ultimately lead to cell death if eATP exposure persists (2, 8). ADP-ribosylation of P2RX7 by the ectoenzyme ARTC2.2, in contrast, induces irreversible pore formation and subsequent cell death. ARTC2.2 is activated by extracellular NAD (9). Importantly, ARTC2.2 activation-induced P2RX7 pore formation occurs at much lower concentrations of NAD compared with that of eATP (10). ARTC2.2 is catalytically active even when cells are at 4°C (1). The subsequent formation of P2RX7 pores, however, only happens at temperatures above 24°C, suggesting the effects of ARTC2.2-mediated ribosylation could be manifested during tissue processing that involves incubation at room temperature or 37°C—such as the steps necessary for lymphocyte isolation from nonlymphoid tissues (11, 12). Indeed, previous studies have shown extensive cell death of T cell populations under these circumstances, especially cells expressing high levels of ARTC2.2 and P2RX7, like CD4⁺ T regulatory cells (T_reg) (1). Moreover, even cells that survive isolation steps may be compromised for in vitro functional assays (13).

To tackle this issue, ARTC2.2-specific antagonist nanobodies to block the ARTC2.2/P2RX7 signaling axis were developed (9). Previous studies successfully used this strategy to recover lymphocytes with high expression of ARTC2.2, including T_reg and invariant NKT cells (iNKT) (13). Two recent reports showed that ARTC2.2 blockade also prevents the death of liver CD8⁺ tissue-resident memory T cells (T_RM) during tissue preparation (14, 15). Overall, this indicates that T lymphocytes in nonlymphoid tissues are sensitive to death induced by activation of ARTC2.2 and P2RX7. Despite the pioneering nature of these reports, several questions remain. First, these studies focused on elevated frequency of live cells and short-term functional assays, rather than numeric comparisons of cells in the tissues. This made it hard to quantify to what extent ARTC2.2 blockade prevented loss of iNKT and T_RM. Especially in the case of T_RM, a severe underestimation of cell numbers detected by isolation and flow cytometric assays has been reported, in comparison with cell numbers calculated by immunofluorescence in situ (16), and it is unclear to what extent activation of the ARTC2.2/P2RX7 axis contributes to this.

Second, T_RM and iNKT are not homogeneous populations, with potential differences dictated both by differentiation state and tissue-specific microenvironmental signals. iNKT, for instance, are composed of functionally and transcriptionally distinct effector subsets that include T-bet⁺ PLZF^low NKT1, PLZF^high NKT2, and RORγt⁺ PLZF^intermediate NKT17 cells (17–19). Notably, the previous ARTC2.2 blockade studies focused on liver and spleen iNKT, most of which are NKT1 cells. Whether other iNKT subsets coexpress ARTC2.2 and P2RX7 and whether blockade of this pathway can rescue these cells is unexplored. As for T_RM, expression of CD69 and CD103 have been used as “signature” markers (20, 21). However, coexpression of CD69 and CD103 is not consistent among nonlymphoid tissues (21); moreover, parabiosis and in situ staining showed that some cells that do not express either of these molecules may nevertheless be bona fide T_RM (16). Whether these different subpopulations of T_RM vary in their susceptibility to ARTC2.2/P2RX7-mediated cell death is untested.

In this study, we provide an analysis of coexpression of ARTC2.2 and P2RX7 in different iNKT subsets and T_RM, in lymphoid and nonlymphoid tissues. In summary, we show that high expression of both molecules correlated with high susceptibility to ex vivo cell death during harvest procedures. Moreover, we found that blocking ARTC2.2 using nanobodies preserved the cell viability and in vitro function, favoring a more precise description of the numeric proportions of different subsets, as well as their exact functionality. Finally, we explored the possibility of using direct P2RX7 inhibition for the preservation of T cell numbers.

Six- to ten-week-old C57BL/6 (B6) and B6.SJL (expressing the CD45.1 allele) mice were purchased from Charles River Laboratories (via the National Cancer Institute). P2rx7^−/− (Mouse Genome Informatics strain designation: 2386080) mice were obtained from The Jackson Laboratory. Lymphocytic choriomeningitis virus (LCMV)/DbGP33–specific TCR transgenic P14 mice were fully backcrossed to B6 and P2rx7^−/− mice, with introduction of CD45.1 and CD45.2 congenic markers for identification. Mice were infected with LCMV Armstrong strain (2 × 10⁵ PFU, i.p.). 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.

To prevent ADP-ribosylation of P2RX7 during harvest procedures, experimental mice were injected i.v. with 50 μg of the ARTC2.2 blocking nanobody (S+16a; Treg Protector, BioLegend) diluted in 200 μl of PBS, 30 min prior to sacrifice (14). Control mice received vehicle (PBS) at the same time.

For short-term blockade of P2RX7 during harvest procedures, experimental mice were injected i.v. with 200 μl of either 35 μM (22) Brilliant Blue G (BBG; Sigma-Aldrich) or, alternatively, with 80 mg kg⁻¹ of A-438079 (0.5% DMSO in PBS; Sigma-Aldrich) 30 min prior to mice sacrifice. Control mice received vehicle injections (PBS or 0.5% DMSO in PBS) at the same time.

Lymphocytes were isolated from tissues, including spleen, skin-draining lymph nodes (LN), mesenteric LN (mLN), thymus, blood, lung, small intestine (SI) intestinal epithelium, SI lamina propria (LPL), and liver, as previously described (16, 23) with the indicated mouse pretreatments (vehicle, ARTC.2.2 nanobody or P2RX7 antagonists). Briefly, this involved digestion at 37°C for 30 min in DTE (SI intestinal epithelial lymphocyte [IEL]) or 30–45 min in Type I Collagenase (SI LPL, lung), with stirring, followed by Percoll gradient centrifugation at room temperature. For secondary lymphoid organs (SLO), processing was followed by RBC lysis with ACK lysis buffer, at room temperature. For discrimination of vascular-associated CD8⁺ T cells in nonlymphoid tissues, in vivo i.v. injection of PE-conjugated CD8α Ab was performed as described (24), 3 min prior to sacrifice. Direct ex vivo staining and intracellular cytokine staining were performed as described previously (23, 25). Briefly, cells were stimulated (as below) for 6 h with GolgiPlug added for the final 4 h. Foxp3 Fix/Perm kit (eBioscience) was used for intracellular detection. Fluorochrome-conjugated Abs were purchased from BD Biosciences, BioLegend, eBioscience, Cell Signaling Technology, Tonbo, or Thermo Fisher Scientific. iNKT were detected using CD1d tetramers loaded with PBS-57 (provided by the National Institutes of Health Tetramer Facility) and TCRβ staining, and the distinct iNKT subsets were distinguished as described previously (17, 18); briefly, the NKT1 cells were defined as PLZF^low T-bet⁺, NKT2 cells were defined as PLZF^high ROR-γt⁻ T-bet⁻, and NKT17 cells were defined as PLZF^intermediate ROR-γt⁺. Polyclonal CD8⁺ T cells were identified as TCRβ⁺ CD8α⁺ CD4⁻. To detect LCMV-specific CD8⁺ T cell responses, tetramers were prepared as described previously (26). Among LCMV-specific CD8⁺ T cells, the following markers were used to distinguish these respective populations: central memory T cells (T_CM) (CD44⁺CD62L⁺), effector memory T cells (T_EM) (CD44⁺CD62L⁻CD127⁺), SLO T_RM (in spleen, LNs: CD44⁺CD62L⁻CD69⁺), and T_RM (i.v. CD8α⁻CD69^+/−CD103^hi/int/lo). For survival assessment, cells were stained with Live/Dead dye (Tonbo Biosciences). For measurement of mitochondrial mass and membrane potential, cells were incubated with MitoTracker Green (MTG) (Thermo Fisher Scientific) and tetramethylrhodamine (TMRE) (Cell Signaling Technology) simultaneously for 15 min at 37°C prior to ex vivo staining. For assessment of proliferation upon in vitro restimulation, cells were stained with Ki-67 (eBiosciences) after fixation using the Foxp3 Kit (Tonbo Biosciences). Flow cytometric analysis was performed on an LSR II or LSR Fortessa (BD Biosciences), and data were analyzed using FlowJo software (Tree Star).

To assess P2RX7 expression kinetics after initial activation, P14 splenocytes from naive mice were stimulated for 72 h with gp33 peptide (1 μM, KAVYNFATM; New England Peptide) and IL-2 (10 ng/ml). To measure the proliferation, survival, and cytokine production of liver iNKT and spleen/SI memory (P14) CD8⁺ T cells, cells from experimental mice were isolated as described above and stimulated in vitro with vehicle (RPMI 1640), PMA (20 ng/ml) plus ionomycin (1 μM), or gp33 peptide (1 μM). Cells were cultured for 4, 24, or 72 h. For all experiments, complete RPMI media (RPMI 1640 supplemented with 10% FBS, 100 U/ml penicillin/streptomycin, 2 mM l-glutamine) was used.

Data were subjected to the Kolmogorov-Smirnov test to assess Gaussian distribution. Statistical differences were calculated by using unpaired two-tailed Student t test or one-way ANOVA with Tukey posttest, where indicated. All experiments were analyzed using Prism 5 (GraphPad Software). The p values <0.05 (*), <0.01 (**), <0.001 (***), or <0.0001 (****) indicated significant differences between groups, and nonsignificant differences were indicated with “ns.”

We first sought to evaluate the expression of P2RX7 and ARTC2.2 in subsets of iNKT from an array of lymphoid and nonlymphoid tissues. Previous studies indicate that ARTC2.2/P2RX7 activation strongly affects the viability of liver iNKT (27). Liver iNKT are predominantly NKT1, and we confirmed that liver NKT1 cells express high levels of both ARTC2.2 and P2RX7 (Fig. 1A). Similar expression was seen on NKT1 from other nonlymphoid tissues and in the spleen and mLN, whereas mature NKT1 in the thymus showed low expression of both ARTC2.2 and P2RX7 (Fig. 1B, 1C). Examination of NKT2 and NKT17 cell subsets (17), however, showed distinct expression patterns: NKT2 cells expressed high ARTC2.2 and P2RX7 in all sites, including the thymus, whereas NKT17 cells expressed low levels of both molecules, whether recovered from thymus, LN, or lung (Fig. 1B and data not shown).

We further assessed ARTC2.2 and P2RX7 expression in CD8⁺ T cells in unimmunized mice. Whereas CD8 single-positive thymocytes expressed low levels of both molecules, peripheral naive CD8⁺ T cells expressed ARTC2.2 but not P2RX7 (Fig. 1D, top). P2RX7 and ARTC2.2 were both highly expressed by most CD8⁺ T cells found in nonlymphoid tissues such as the SI at steady-state (Fig. 1D, top). As expected, these cells had a phenotype resembling Ag-experienced cells (Supplemental Fig. 1A), thus suggesting CD8⁺ T cells activated by either Ag or homeostatic cytokines express higher levels of ARTC2.2 and P2RX7. Indeed, following cognate Ag-induced activation, P2RX7 expression rose in CD8⁺ T cells (Supplemental Fig. 1B). Furthermore, we generated P14-immune chimeras, in which congenic distinct host mice were adoptively transferred with P14 CD8⁺ T cells (bearing a TCR transgene specific for an LCMV epitope) and subsequently infected with LCMV. Virtually all memory P14 populations displayed ARTC2.2, but P2RX7 expression level varied widely in distinct memory subpopulations (Fig. 1D, bottom and Fig. 1E). Whereas T_EM showed low P2RX7 expression, a substantial fraction of T_CM, essentially all resident memory cells present in SLO, and all T_RM in nonlymphoid tissues exhibited high P2RX7 expression (Fig. 1D, bottom and Fig. 1F). Hence, the majority of P14 T_RM are P2RX7⁺ ARTC2.2⁺, suggesting they might also be susceptible to NAD-induced cell death.

Considering the high expression of both ARTC2.2 and P2XR7 in NKT1 cells, especially those residing in the nonlymphoid tissues (liver and SI), we asked whether tissue harvest and processing led to loss of NKT1 cells in an ARTC2.2-dependent pathway. To test this, we used a nanobody specific to ARTC2.2, previously shown to inhibit this enzyme’s function in T_reg cells (27). In keeping with previous reports (13, 27), we observed a significantly higher yield of NKT1 cells in liver of mice treated i.v with anti-ARTC2.2 30 min prior to sacrifice (Fig. 2). The action of the ARTC2.2 blockade was not limited to the liver because we observed a notably higher yield (5-fold) of NKT1 cells in the IEL of treated mice (Fig. 2). Although not statistically significant, there was also a clear trend of higher yield of NKT1 cells in LPL by pretreatment with anti-ARTC2.2 nanobody (Fig. 2). Importantly, ARTC2.2 blockade led to increased numbers of NKT1 cells recovered, which was not specifically reported in previous studies (13, 27). In contrast, we did not observe any difference in the recovery of NKT1 cells in thymus, spleen, or mLN between mice injected with anti-ARTC2.2 nanobody or PBS (Fig. 2). It is important to note that tissue processing of liver and intestines (but not of thymus, spleen, or LNs) requires short-term (20–60 min) incubations at room temperature (25°C) or 37°C, conditions that permit ARTC2.2-mediated P2RX7 pore formation. Together, our data suggest that nonlymphoid tissue processing might result in ARTC2.2-mediated loss of NKT1 cells, and blockade of this pathway is a crucial step for the recovery of optimal numbers of this cell population.

We next evaluated whether blocking the ARTC2.2/P2RX7 signaling axis could likewise enhance recovery of CD8⁺ T_RM from different nonlymphoid tissues in LCMV-immune mice. As reported before (14, 15), we found that nanobody pretreatment led to an increased recovery of viable liver T_RM (Fig. 3A). This effect was not limited to the liver because the numbers of T_RM in other nonlymphoid tissues were also elevated in mice injected with the anti-ARTC2.2 nanobody (Fig. 3A) and were associated with decreased cell death (Supplemental Fig. 2A). We also observed an increase in numbers of CD44⁺ polyclonal CD8⁺ T cells from SI with nanobody treatment (Supplemental Fig. 2B). Interestingly, Masopust and collaborators (16) reported that processing tissues for recovery of T_RM significantly underestimates their numbers in nonlymphoid tissues, compared with in situ quantification using immunofluorescence. Our results suggest that ARTC2.2-mediated cell death during cell isolation can account for nearly all of this discrepancy in SI IEL T_RM numbers and substantially corrects T_RM numbers from other tissues tested (Fig. 3B). Both CD103⁺ and CD103⁻ SI IEL T_RM express ARTC2.2, with slightly higher expression in CD103⁻ T_RM (Supplemental Fig. 2C). Consistently, recovery of both subsets was enhanced by ARTC2.2 nanobody treatment (Fig. 3C, 3D). Importantly, CD103⁻ SI IEL T_RM have a higher proportional representation in nanobody-treated samples (Fig. 3C), which corresponds with the finding that CD103⁻ T_RM are underrepresented by conventional tissue extraction protocols (16). Our data indicate that ARTC2.2-mediated cell death plays a role in this issue and suggest that blockade of this enzyme not only induces increased recovery of T_RM numbers but also offers a more faithful representation of T_RM subsets after tissue harvests.

Furthermore, we observed that nanobody injection significantly enhanced recovery of lung T_RM (Fig. 3A). Indeed, lung showed the highest discrepancy between flow cytometry and quantitative immunofluorescence in the determination of T_RM numbers (16) (Fig. 3B). Therefore, it may not be surprising that a majority of lung T_RM underwent ARTC2.2-mediated cell death during tissue preparation. Following conventional tissue processing procedures, only 5–10% of lung virus-specific CD8⁺ T cells are identified as i.v.⁻ (i.e., likely to be in the lung parenchyma), whereas quantitative immunofluorescence estimates the frequency of the i.v.⁻ population is 25–30% (16). In lungs processed after ARTC2.2 nanobody treatment, nearly 30% of specific CD8⁺ T cells were i.v.⁻ (Fig. 3E), which is in close agreement with the quantitative immunofluorescence approach. This cements the notion that anti-ARTC2.2 blockade prior to organ harvest renders a faithful representation of T_RM subsets in several nonlymphoid tissues.

We recently reported that expression of P2RX7 is necessary for T_RM development in nonlymphoid tissues (7). Superficially, this appears to be at odds with the increased susceptibility to NAD-induced cell death for T_RM described in this article. In our previous study we used ARTC2.2 nanobody blockade prior to tissue harvest (7), based on the expectation that P2RX7 deficiency would avert cell death mediated by ARTC2.2 (13). To directly test this idea, we cotransferred equal numbers of wild-type (WT) and P2rx7^−/− P14 cells into recipient mice, which were then infected with LCMV, and 8 wk later, cells were recovered from the SI IEL with or without anti-ARTC2.2 nanobody treatment just prior to harvest. Without treatment, we saw no advantage in WT compared with P2rx7^−/− T_RM numbers (Fig. 3E, 3F). In contrast, ARTC2.2 blockade resulted in an∼5-fold increased recovery of WT relative to P2rx7^−/− P14 T_RM (Fig. 3F). Importantly, there was no difference in the numbers of P2rx7^−/− P14 T_RM in PBS- versus nanobody-treated recipient mice (Fig. 3F), confirming that P2RX7 expression dictates the susceptibility of T_RM to ARTC2.2-induced cell death.

In previous studies, ARTC2.2 blockade prior to harvest prevented shedding of CD62L in iNKT and T_reg cells (9). T_CM also express CD62L and are found primarily in SLO and blood (28). Of note, some staining protocols and tissue processing experiments with these organs involve incubations at room temperature or 37°C, such as peptide/MHC tetramer enrichment or staining of chemokine receptors (29–31). We observed that a considerable portion of T_CM express high levels of both ARTC2.2 and P2RX7 (7) (Fig. 1). Hence, we tested whether incubation of spleen cells from P14-immune chimeras at 37°C would influence the staining and numbers of CD62L⁺ T_CM detected. We observed a significant decrease in percentages of CD62L⁺ P14 cell in samples incubated at 37°C, whereas this was prevented by anti-ARTC2.2 blockade (Supplemental Fig. 3A, 3C). This is likely due to CD62L shedding rather than T_CM death because the total P14 cell numbers were not altered by nanobody treatment (Supplemental Fig. 3B).

In nonlymphoid tissues, most memory CD8⁺ T cells are resident (16), and virtually all have been characterized as CD62L⁻, based chiefly on flow cytometric analysis (20). Nevertheless, recent studies indicate CD62L⁺ memory CD8⁺ T cells can migrate into peripheral tissues (16, 32). Given that most nonlymphoid tissue processing protocols involve 37°C incubation steps, we asked whether anti-ARTC2.2 blockade could improve detection of migrating CD62L⁺ memory CD8⁺ T cells in nonlymphoid tissues. Indeed, we detected significantly higher numbers of CD62L⁺ P14 cells in the SI of nanobody-treated mice compared with their PBS-treated counterparts (Supplemental Fig. 3D). Together, our data show that, akin to iNKT and T_reg cells, anti-ARTC2.2 blockade prior to tissue harvest permits a more precise assessment of CD62L⁺ CD8⁺ T_CM in both lymphoid and nonlymphoid tissues.

Activation of the ARTC2.2/P2RX7 signaling axis also induces the loss of other cell surface molecules (9). Consistent with a previous study (27), we observed that in vivo treatment with anti-ARTC2.2 nanobody prevented loss of CD27 in both splenic and liver iNKT (Fig. 4A). Furthermore, we evaluated other surface markers that have been used for phenotyping of iNKT and CD8⁺ T cells. We observed that cell surface levels of CD69 and P2RX7 in iNKT were substantially preserved by in vivo treatment of anti-ARTC2.2 nanobody, whereas expression of ARTC2.2 itself, CD122, and CD4 was not affected (Fig. 4A). To our knowledge, the nanobody-mediated preservation of CD69 and P2RX7 expression in iNKT is a novel finding and suggests that some previous studies of iNKT may not faithfully describe their phenotype ex vivo. Whether ARTC2.2 activation alters other surface markers in iNKT is still unclear and will be a focus of future research.

Next, we asked whether blockade of the ARTC2.2/P2RX7 signaling axis would improve functional properties of iNKT after in vitro stimulation. Liver mononuclear cells from mice treated with PBS or anti-ARTC2.2 nanobody were cultured in vitro with the presence of PMA/ionomycin for 4 h. The treatment of anti-ARTC2.2 nanobody prior to tissue harvest not only led to much better recovery of iNKT after short-term in vitro stimulation (Fig. 4B) but also to a substantially higher frequency of IFN-γ production by those NKT1 cells (Fig. 4C, 4D).

Recently, there has been an increasing interest in harnessing iNKT for immunotherapy, which usually requires prolonged in vitro culture for transduction and/or expansion. Therefore, we tested the possibility that blockade with anti-ARTC2.2 nanobody might enhance the feasibility of culturing iNKT in vitro. Indeed, we observed that in vivo treatment of mice with anti-ARTC2.2 nanobody enhanced the number and viability of iNKT after short-term in vitro culture by∼5-fold (Fig. 4E, 4F). Taken together, these data show that in vivo treatment of anti-ARTC2.2 nanobody to block the ARTC2.2/P2RX7 signaling pathway prior to isolation significantly improves viability and functionality of iNKT for in vitro stimulation and cell culture.

We also tracked the effect of anti-ARTC2.2 nanobody on the expression of cell surface molecules in CD8⁺ T cell T_RM. As for iNKT, expression levels of CD69, CD44, and especially P2RX7 were increased on T_RM following transient ARTC2.2 blockade, and expression of ARTC2.2 itself was also elevated (Fig. 5A, Supplemental Fig. 4A). Of note, nanobody pretreatment caused no changes in the expression of these molecules in memory CD8⁺ T cells isolated from spleen, harvested at 4°C (data not shown). We next sought to define if ARTC2.2 blockade preserves the function of T_RM during in vitro assays (Fig. 5B). In line with our ex vivo assays (Fig. 2) and a recent report with liver T_RM (14, 15), injection of the nanobody prior to nonlymphoid tissue preparation led to a substantial increase in the recovery of viable SI IEL T_RM and a small increase in the viability of spleen memory cells (Fig. 5C, Supplemental Fig. 4C) following 4 h of in vitro restimulation. Moreover, we observed an increased frequency of IFN-γ–producing P14 T_RM, especially following pharmacological stimulation using PMA/ionomycin (as is often used to assess T cell function independent of Ag specificity) (Fig. 5D, 5E). Memory CD8⁺ T cells rapidly increased surface P2RX7 level after stimulation in vitro (Supplemental Fig. 4B), which is consistent with this receptor playing a role in memory CD8⁺ T cell reactivation (7), but at the same time, rendering these cells susceptible to death by NAD/ARTC2.2.

Prolonged in vitro stimulation and/or manipulation of T_RM is difficult because of the high cell death rate (20, 33, 34), which has made in vitro assessment of prolonged T_RM reactivation and assays on immunometabolism impractical. Strikingly, injection of anti-ARTC2.2 nanobody prior to cell collection led to a significant increase in T_RM viability after 24 h of in vitro culture, in either the presence or absence of restimulation (Fig. 5F, left and center). ARTC2.2 inhibition prior to in vitro stimulation led to a striking rescue of IFN-γ–producing T_RM, which were increased by more than 100-fold (Fig. 5F, right), and a more moderate increase in viability of restimulated splenic memory P14 cells (Supplemental Fig. 4D). This trend continued at 72 h of culture, when we observed increased viability of T_RM following in vivo nanobody treatment (Supplemental Fig. 4E, 4F), and this population showed enhanced cytokine production and proliferation compared with controls (Supplemental Fig. 4G, 4H). The major hallmarks of T_RM secondary immune responses after local Ag challenge are increased in situ cytokine production (35, 36) and proliferation (37). Thus, transient ARTC2.2 blockade prior to harvest makes reliable in vitro assessment of T_RM function practical.

NAD/ARTC2.2-induced cell death in T_RM occurs through the activation of P2RX7 (Fig. 3F). A recent report showed that P2RX7 blockade during in vitro assays prevents CD4⁺ follicular helper T cell death (38). We assessed whether in vivo P2RX7 blockade just prior to harvest could also prevent T_RM death and improve recovery. Blockade of P2RX7 with the nonspecific inhibitor BBG and the specific synthetic inhibitor A-438079 (39) by i.v. injection 30 min prior to sacrifice led to recovery of increased numbers of SI P14 T_RM (Fig. 6A), albeit to a lesser extent than that of using ARTC2.2 nanobody treatment (Fig. 3). Like ARTC2.2 blockade, P2RX7 blockade also led to an increased representation of CD103⁻ T_RM (Fig. 6B), suggesting that direct P2RX7 blockade can be a viable option for ex vivo analysis of T_RM subsets.

We previously discovered a fundamental role of P2RX7 in controlling the establishment and function of circulating and resident memory CD8⁺ T cells (7). Importantly, we demonstrated that in vitro P2RX7 blockade for as short as 6 h induced defects on CD8⁺ T cell survival, metabolism, and function like those observed in P2rx7^−/− CD8⁺ T cells (7). Therefore, we evaluated whether preharvest P2RX7 blockade would also lead to metabolic alterations in memory CD8⁺ T cells. Our data show that short-term in vivo P2RX7 blockade induced a decrease in mitochondrial mass (measured by MTG staining) in both spleen and SI P14 cells (Fig. 6C). In contrast, ARTC2.2 blockade prior to harvest did not induce alterations in mitochondrial mass or membrane potential, as measured by TMRE staining (Fig. 6D). Recently, P2RX7-specific antagonist nanobodies were developed (40), which might offer a better opportunity for directly blocking this receptor to yield a more accurate estimate of tissue-resident lymphocyte numbers. Additionally, it was reported that P2RX7 antagonists improve the functionality of P2RX7-expressing CD4⁺ T cells in vitro (38). Based on our findings, transient blockade of P2RX7 could be used to quantify T_RM subsets, but its use should be interpreted with caution for prolonged in vitro culture or functional assays of T_RM, for which transient ARTC2.2 blockade would be the preferred approach.

Early studies indicated that the activation of the ARTC2.2/P2RX7 pathway during ex vivo preparation of a single-cell suspension could lead to profound apoptosis and dysfunction in T_reg and liver iNKT (13). However, the expression of ARTC2.2 and P2RX7 in iNKT or CD8⁺ T_RM in tissues other than liver has not been extensively evaluated. Moreover, whether ARTC2.2/P2RX7 signaling might lead to disproportionate changes in cell number, phenotype, or function of specific subsets of iNKT and CD8⁺ T_RM during tissue processing had not been evaluated. In the current study, we established a detailed analysis of P2RX7 and ARTC2.2 expression in iNKT and CD8⁺ T cells and short-term blockade of this ARTC2.2/P2RX7 axis shortly before tissue harvest could significantly improve the recovery and ex vivo function of iNKT and CD8⁺ T_RM. Our results suggest that both ARTC2.2 and P2RX7 are highly expressed in peripheral but not thymic NKT1 cells. This might be due to the environmental cues present only in periphery. Retinoic acid reportedly induces P2RX7 expression in intestinal lymphocytes (41), being a potential factor that influences the expression of ARTC2.2 and P2RX7 in iNKT and CD8⁺ T cells residing in nonlymphoid tissues. Whether retinoic acid itself or similar molecules drive expression in other tissues is still unknown and will certainly be a subject of future studies. Moreover, it is worth mentioning that cell-intrinsic factors might also contribute to this phenomenon, as NKT2 cells seem to express high levels of ARTC2.2 and P2RX7 regardless in the thymus or periphery, whereas NKT17 cells do not express these receptors in either the thymus or periphery. Altogether, these results demonstrate considerable heterogeneity in the expression of P2RX7 and ARTC2.2 across various subpopulations of T cells (even among iNKT subsets), suggesting differential susceptibility to NAD-induced and ATP-induced cell death and signaling. In the future, it will be important to determine whether iNKT differentiation signals override local environmental cues for expression of P2RX7 and ARTC2.2 and whether activation of these receptors is involved in shaping the representation of distinct iNKT subsets.

The physiological role of P2RX7 in T cell function and homeostasis is still the subject of intense investigation. It has been reported that P2RX7 activation can favor mouse CD4⁺ T cell IL-2 production following restimulation and the differentiation of Th17 cells but restrains generation of T_reg and T follicular helper cells (5, 42, 43). Moreover, we recently reported that P2RX7 is crucial for establishment of long-lived memory CD8⁺ T cell populations, including T_RM (7). Interestingly, our data presented in this article suggest a “Jekyll and Hyde” function of this receptor for T_RM biology, in which high expression of P2RX7 can be beneficial in most cases (e.g., for memory establishment) but potentially detrimental if these cells encounter situations (such as highly inflammatory environments) in which ARTC2.2 activation by NAD may be anticipated. Indeed, a recent report showed P2RX7 to be detrimental for liver and SI IEL T_RM in the presence of sterile inflammation (44). It will be important to expand these studies and determine how long-term survival of various tissue-resident T cell populations, as well as their in vivo function, is affected by inflammatory challenge in a P2RX7/ARTC2.2-dependent way. Likewise, ARTC2.2 and P2RX7 are highly expressed by some subsets of iNKT, yet the impact of this expression on differentiation and function of these populations is unclear. Furthermore, it will be interesting to explore the role of P2RX7 and ARTC2.2 in mucosal-associated invariant T cells (19, 29, 45–48), another innate-like T cell population, in future studies.

Taken together, our results help cement the notion that high expression of ARTC2.2 and P2RX7 act as biomarkers for susceptibility to cell death during tissue isolation in murine models. This was confirmed by our comprehensive ARTC2.2 blockade study. Beyond the scope of previous reports, we showed in this study that inhibition of ARTC2.2-mediated cell death not only preserves viability of nonlymphoid tissue T cell populations (iNKT and T_RM) but allows a more accurate representation of different subsets among those recovered lymphocytes. Moreover, we show this blockade permits the accurate assessment of functionality of tissue-resident T cells in vitro, which opens several future possibilities—for example, metabolism assays and transduction of T cell populations isolated from nonlymphoid sites, both of which have been technically challenging. In summary, we report that short-term blockade of ARTC2.2/P2RX7 activation during cell isolation is a crucial technical step for the optimal study of the biological characteristics of iNKT and T_RM in nonlymphoid tissues and should be used when practical.

We thank the University Flow Cytometry Resource core facility (University of Minnesota) for technical support. We thank the members of the Jamequist Laboratory and Center for Immunology for insightful discussion and the National Institutes of Health Tetramer Core for provision of peptide/MHC tetramers and CD1d tetramer.

The authors have no financial conflicts of interest.