Lack of p53 Augments Antitumor Functions in Cytolytic T Cells

Adoptive transfer of tumor epitope reactive T cell in cancer patients has generated much interest due to promising control of tumor growth (1). However, susceptibility to immunosuppression and reduced survival of effector T cells in an oxidative tumor microenvironment are the key confounding factors in immunotherapy (2, 3). We have previously shown that reactive oxygen species (ROS) scavengers can inhibit repetitive T-cell receptor (TCR) stimulation-mediated activation induced cell death (AICD) of tumor-reactive T cells without interfering with cytokine production (4), a measure of CTL function, placing redox regulation at a central point for therapeutic intervention.

The altered expression of a redox active transcription factor p53 leads to uncontrolled cell proliferation, senescence, and cell death (5). However, only a handful of studies have reported the role of p53 in shaping T-cell immune response. Grayson and colleagues (6) reported slightly higher memory response in p53-KO mice compared with p53-sufficient mice, and only minor differences in proliferation, apoptosis, or maintenance of “non-self” viral antigen-specific T cells. A recent study has shown that in order to mount an effective antigen-specific proliferative response, CD4⁺ T cell kinetically downregulate the expression of tumor suppressor p53 until 72–96 hours (7). Another study showed that p53 inhibits systemic autoimmune diseases by inducing regulatory T cells (Treg; ref. 8). Because p53 is also required for TGFβ gene responses by cooperating with Smads (9), we hypothesized that T cells from p53-KO mice will be less prone to TGFβ-mediated immunosuppression in a tumor microenvironment, and with less incidence of inducible regulatory T cell (iTreg) generation a durable antitumor T-cell response could be mounted by targeting p53. Further, p53 negatively regulates glycolysis through activation of TP53-induced glycolysis regulator (TIGAR; ref. 10) and positively regulates oxidative phosphorylation (OXPHOS) through upregulation of SCO2, a member of the COX-2 assembly involved in the electron-transport chain (11). Because long-term T-cell effector and memory response is also metabolically regulated (12), we determined if differences in metabolic signature due to lack of p53 expression correlate to antitumor T-cell function.

Our study demonstrates that p53 deficient T cells exhibited enhanced effector function and proliferation while maintaining the CD62L^hiCD44^hi central memory (Tcm) phenotype. Further, p53-KO T cells are not transformed to iTregs and exhibit elevated cytolytic properties with remarkable tumor control in a mouse melanoma model. Thus, p53 could serve as target for improving ACT.

C57BL/6 (cat. # 000664) and p53-KO (cat. # 002101) mice were obtained from The Jackson Laboratory. Development of h3T TCR transgenic mouse has been described recently (13). Briefly, the class I–restricted human tyrosinase epitope (YMDTMSQV)_368-376 reactive TCR isolated from tumor-infiltrating lymphocytes of an HLA-A2⁺ metastatic melanoma patient was used to generate this transgenic mice. Animals were maintained in pathogen-free facilities and procedures approved by the Institutional Animal Care and Use Committee.

Recombinant cytokines were purchased from BioLegend. Complete IMDM (cIMDM) media containing 10% FBS, penicillin, and streptomycin were used for T-cell differentiation. On day 3 of culture, T cells were harvested and either processed for intracellular cytokine analysis, RNA preparation using TRIzol (Invitrogen) or used for adoptive cell therapy.

Mouse melanoma tumor (B16-F10) and human melanoma (624-MEL) were maintained in vitro in cIMDM. B16-F10 (0.25 × 10⁶) and 624-MEL (2.5 × 10⁶) were injected subcutaneously (s.c.) into left flank of C57BL/6 or Rag1^−/− C57BL/6 mice or NSG-A2 mice, respectively. Twenty-four hours before adoptive transfer of T cells on day10, the recipient mice were injected cyclophosphamide (4 mg/mice, i.p.).

Three days after TCR activation, transgenic T cells were restimulated for 4 hours with either cognate antigen or nonspecific antigen-loaded T2-A2 cells at a 5:1 ratio. Apoptosis was measured by staining for Annexin V according to the manufacturer's protocol, followed by flow cytometry. Data were analyzed with FlowJo software (Tree Star).

Cells were stained with fluorescent-labeled deoxy-glucose analog, 2NBDG (Cayman Chemicals) according to the manufacturer's protocol. Cells were washed and stained with other fluorochrome-conjugated antibodies and acquired by flow cytometry. All analyses were done on viable cells. Mitochondrial oxygen consumption or glycolytic flux was measured using the XF 24 analyzer (Seahorse Bioscience) as described earlier (14).

Detailed protocols for staining the cells for surface markers and intracellular cytokines have been described earlier (15). Data were analyzed with FlowJo software (Tree Star).

Total RNA was isolated using TRIzol reagent (Invitrogen). cDNA was generated from 1 μg total RNA using iScript cDNA Synthesis Kit (BioRad). Quantitative real-time PCR was performed using a SYBR Green mix (Biorad) in the CFX96 Detection System (BioRad). The fold change in expression of molecules in h3T-p53 KO T cells was calculated over h3T cells and expressed as relative fold change. The TGFβ Pathway PCR array (Qiagen) was used to monitor the expression of 84 genes, along with five housekeeping genes and control for genomic DNA contamination, RNA quality, and general PCR performance. Data analysis was performed using Qiagen's proprietary web-based analysis tool.

All data reported are the arithmetic mean from three or five independent experiments performed in triplicate ±SD unless stated otherwise. The unpaired Student t test was used to evaluate the significance of differences observed between groups, accepting P < 0.05 as a threshold of significance. Data analyses were performed using the Prism software (GraphPad). In vivo data were analyzed using Kaplan–Meier methods, and pairwise comparisons of survival distributions were done via the log-rank test. Mice that did not reach a tumor size of 400 mm³ by the end of the experiment were sacrificed and had survival time censored in the analysis.

To determine the role of p53 in tumor epitope-specific T cells, we crossbred p53 knockout (p53-KO) mice with h3T TCR transgenic mice (13). Supplementary Figure S1A shows the PCR-based genotype screening for the h3T-p53 KO. Using cell trace violet dye we noticed that upon stimulation with cognate antigen, the TCR transgenic T cells from h3T-p53 KO proliferated faster until 48 hours (left) as compared with the wild-type (wt) h3T T cells (Fig. 1A). The difference persisted even after 72 hours (right) of stimulation showing greater cell division in h3T-p53 KO–derived T cells. This increased proliferation could be attributed solely to the absence of p53, because the expression of activation-induced cell surface molecules such as CD69 or CD25 (IL2Rα) was similar in h3T-p53 KO– and h3T-derived T cells (Supplementary Fig. S1B). In keeping with the increase in proliferation, a higher number of total splenocytes and thymocytes were retrieved from h3T-p53 KO mice (Fig. 1B; Supplementary Fig. S1C). Our data show that TCR activated h3T-p53 KO–derived T cells have higher expression of Cyclin D, a key cyclin protein involved in regulating cell-cycle progression, and is repressed by p53 (16). The expression of cyclin-dependent kinase inhibitors CDKn1a, CDKn2a, and CDKn2b, which are regulated by p53, was also significantly reduced in h3T-p53 KO cells as compared with h3T T cells (Fig. 1C). In addition, higher proliferation rate could lead the T cells close to replicative senescence with increased CD62L^lo phenotype and susceptibility to cell death (3). A recent study has also shown that p53 isoform switching regulates tumor-associated replicative senescence in T cells (17). However, we observed that h3T-p53 KO T cells not only exhibit higher percentage of CD62L⁺CD44⁺ T central memory (Tcm) phenotype as compared with h3T T cells (Fig. 1D), but also showed lower expression of senescence-associated β-galactosidase and increased CD28 expression (Fig. 1E). Thus, reduced expression of p53 modulates cell-cycle progression of T cells without inducing replicative senescent phenotype.

Because Tcm phenotype is associated with higher antioxidant capacity and reduced cell death (3), we determined ROS/RNS levels and AICD levels between h3T-p53 KO versus h3T T cells. Upon TCR restimulation with cognate tyrosinase antigen, h3T-p53 KO T cells secreted less ROS (measured using DCFDA dye), and RNS (measured using DAF dye), as compared with h3T T cells (Fig. 2A). This also correlated to increased antioxidant levels as determined by cell surface thiol (c-SH; measured using melamide dye) and intracellular glutathione (iGSH; measured using monocholorobimane dye; Fig. 2B) in p53-KO T cells. Further, a quantitative real-time analysis revealed that antioxidant enzymes catalase and superoxide dismutase (SOD) levels were also elevated in activated h3T-p53 KO T cells as compared with h3T T cells. While TCR restimulation-induced ROS/RNS levels could affect downstream signaling that involves JNK and leads to T cell death (4), we observed that upon TCR restimulation, h3T-p53 KO T cells exhibit reduced JNK phosphorylation (Fig. 2D, top), and cell death, as indicated by reduced loss of mitochondrial membrane potential (measured using DiOC₆; Fig. 2D, bottom). Phosphatidyl serine upregulation (using Annexin V) among the Vβ12⁺CD8⁺ TCR transgenic T cells was also reduced (Fig. 2E). Admittedly, while the difference between the Annexin V⁺ cells in p53-KO versus WT was about 10% to 15%, the number of cells that were Annexin V^lo (between 0 and 10² on the x-axis) was appreciable (Supplementary Fig. S1D). To further confirm if ROS/RNS levels are important mediators of p53 phosphorylation, the TCR-activated wt T cells were either pretreated for 45 minutes with antioxidant compound L-NAC (10 mmol/L), or left untreated before TCR restimulation. We observed that reduced RNS accumulation (determined by DAF staining) after antioxidant L-NAC pretreatment also correlated with reduced p53 phosphorylation (Fig. 2F). Thus, the loss of p53 in T cells results in their increased antioxidant capacity, which renders them less susceptible to oxidative stress–mediated cell death.

Recent studies have shown that p53 is also involved in regulating various metabolic pathways (18), by balancing glycolysis and oxidative phosphorylation, regulating the production of ROS. While determining how p53 loss regulates T-cell metabolism, we observed that uptake of fluorescent glucose 2-NBDG was higher in TCR-activated h3T-p53 KO T cells as compared with h3T T cells (Fig. 3A, i). Next, we observed significantly higher mRNA levels of glycolytic pathway enzymes in TCR-activated p53-KO T cells as compared with the h3T T cells (Fig. 3B). Specifically, glycolysis genes, hexokinse (HKII), phosphofructokinase (Pfk), lactate dehydrogenase A (LDHA; Fig. 3B, i), and key glycolysis regulator hypoxia-inducing factor (HIF1α; Fig. 3B, ii; *, P ≤ 0.05; ref. 19) were found to be significantly upregulated. Further, the expression of TIGAR (Tp53-induced glycolysis and apoptosis regulator), a known negative regulator of glycolysis that is activated by p53 (10), was also reduced in h3T-p53 KO T cells as compared with h3T T cells (Fig. 3B, ii). Increased expression of glycolytic genes was also observed when comparing magnetic bead-sorted CD8⁺ T cells from p53-KO T cells to the wt CD8⁺ T cells (Supplementary Fig. S1E). However, expression level of peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α), a key regulator of mitochondrial biogenesis, was significantly decreased in h3T-p53 KO T cells as compared with h3T T cells (Fig. 3B, iii). Further, we also observed increased expression levels of key enzymes involved in the regulation of pentose phosphate pathway (PPP) that are required for nucleotide synthesis. In concordance with a recent study that showed p53 inhibits PPP (20), we observed that the mRNA expression of glucose-6 phosphate dehydrogenase (G6PD) was 4-fold higher, and that of ribose-5-phosphate isomerase (RPIA) was about 3-fold higher in activated h3T-p53 KO T as compared with h3T T cells (Fig. 3B, iv). Because cells with glycolytic phenotype exhibit a significantly higher extracellular acidification rate (ECAR) than those dependent upon oxidative phosphorylation, which display higher oxygen consumption rate (OCR; ref. 14), we determined the ECAR and OCR levels in real time using bioanalyzer (Seahorse). Our data show that 3-day activated h3T-p53 KO T cells exhibit higher ECAR as compared with h3T T cells (Fig. 3C). Thus, enhanced glycolysis accompanied by increased commitment to PPP could be contributing to T-cell anabolism (21). The higher degree of glycolysis also correlated to the higher usage of the mTOR pathway as observed by elevated phosphorylation levels of ribosomal protein S6 (Fig. 3D), which is reported to mediate glycolysis (22). Transgenic T cells at the basal level or stimulated with control peptide showed lower pS6 staining, indicating that the increase pS6 in h3T-p53 KO was antigen specific. Thus, these data suggest that increased antioxidant capacity and glycolytic commitment of p53-KO T cells could be due to increased expression of PPP molecules—as G6PD reduces nicotinamide adenine dinucleotide phosphate (NADP) to NADPH, and NADPH in turn maintains the level of glutathione to help protect against oxidative damage—a scenario that could be useful in maintaining persistence of tumor-reactive T cells in the oxidative tumor microenvironment.

Because loss of p53 results in increased glycolysis, a key metabolic pathway that regulates cytokine IFNγ (23), we compared expression of effector molecules between h3T-p53 KO and h3T T cells. Our data demonstrate that upon TCR stimulation with tyrosinase antigen, the fraction of T cells secreting cytokines IL2, IFNγ, and TNFα were about 2-fold more as compared with the h3T T cells (Fig. 4A). Overnight antigen stimulation also confirmed that h3T-p53 KO T cells secreted twice the amount of IFNγ as compared with h3T T cells (Fig. 4A, right). Importantly, h3T-p53 KO T cells also exhibit increased externalization of lysosomal protein CD107a (Fig. 4B), an indicator of enhanced cytotoxic granule exocytosis (24), which indicates increased cytolytic ability of h3T-p53 KO T cells over p53-sufficient h3T T cells. We also observed that the expression of signature transcription factors for type-1 cytotoxic (Tc1) cells as T-bet and IRF-4 was higher in p53-KO CD8⁺ T cells (Fig. 4C). In addition, an increased expression of genes related to key effector molecules such as GM-CSF, Granzyme B, IL1Rn, IL23R, and IL22 was noticed in h3T-p53 KO T cells than in h3T T cells (Fig. 4D). These data indicate that h3T-p53 KO T cells are highly poly-functional cells and exhibit increased effector function as compared with h3T T cells.

To determine the antitumor potential of p53-deficient T cells, B16-A2 tumor melanoma cells were established subcutaneously in HLA-A2 mice for 14 days before transferring HLA-A2–restricted tyrosinase reactive Vβ12⁺ TCR transgenic CD8⁺ splenic T cells from h3T-p53 KO or h3T mouse (schema in Fig. 5A). We observed that h3T-p53 KO T cells showed long-term tumor control than those that received h3T T cells (Fig. 5B; Supplementary Fig. S2A). At the experimental endpoint, 10-fold higher transferred T cells were tracked in the peripheral blood of the recipient group that received h3T-p53 KO splenic T cells (Fig. 5C), which exhibited CD62L⁺CD44⁺ central memory phenotype (16% in h3T vs. 44% in h3T-p53 KO; Fig. 5D). The retrieved h3T-p53 KO transgenic T cells continued to produce more effector cytokines IFNγ and TNFα than retrieved h3T T cells upon restimulation (Fig. 5E). We also noted that a fraction (4%–11%) of h3T-p53 KO–transferred T cells exhibited CD44⁻CD62L⁺ phenotype that is known to harbor stem-cell memory phenotype cells (25), which correlated with 2-to-3-fold higher expression of stemness genes Tcf7, Lef1, and PRDM1 as compared with h3T T cells (Fig. 5F). These data indicate that the quantitative and qualitative differences between the h3T and h3T-p53 KO T cells may account for differences in ability to control tumor growth in vivo. Further, to confirm the feasibility of this approach using TCR-engineered T cells, we used the splenic T cells from the C57BL/6 wild-type mice and p53-KO mice that were rendered tumor antigen specific by using retroviral transduction of tyrosinase reactive HLA-A2⁺ TIL1383I TCR (26). Adoptive transfer of the tyrosinase reactive TIL1383I TCR-transduced T cells also showed long-term control of subcutaneously established B16-A2 tumors in the HLA-A2 recipient mice (Fig. 5G). Next, we tested the efficacy of p53 inhibitors pifithrin-mu (PFT-μ) and pifithrin-alpha (PFT-α) in controlling tumor growth. For this purpose, subcutaneously established B16-F10 murine melanoma in C57BL/6 mice was treated by adoptively transferring 1 × 10⁶ melanoma epitope gp100 reactive T cells that were activated for 3 days with cognate antigen in the presence or absence of inhibitors. We observed that as compared with the activated T cells alone, the p53 inhibitor pretreated T cells resulted in a significantly improved control of tumor growth and thus survival of the tumor-bearing mouse (Fig. 5H). These data show that pharmacological inhibition of p53 could be a clinically translatable ACT approach.

Next, we addressed if the improved tumor control exhibited by h3T-p53 KO T cells is due to reduced susceptibility to immunosuppression or reduced plasticity toward iTregs conversion. Importantly, h3T-p53 KO T cells exhibited reduced expression of TGFβRI and TGFβRII as compared with the h3T-derived splenic T cells (Fig. 6A). Further, ex vivo programming conditions that use TGFβ and IL2 (Fig. 6B, i) showed that p53-KO–derived splenic CD4⁺ T cells exhibit less susceptibility to iTreg conversion (Fig. 6B, ii). The quantitatively reduced iTregs also corresponded to the reduced FoxP3 expression in the p53-KO–derived splenic CD4⁺ T cells as compared with wild-type T cells (Fig. 6B, iii). A detailed analysis of differences in the TGFβ signaling pathway was performed using the TGFβ Signaling Targets RT² Profiler PCR Array (Qiagen). Our data in Fig. 6C show that a number of signaling molecules were differentially expressed in T cells obtained from p53-KO mice, which may have contributed to the enhanced antitumor phenotype. For example, Furin, is a direct target gene of the IL12/STAT4 pathway, regulates Th1/2 cell balance by limiting conversion to Th2 phenotype, and its expression directly correlates to the stability and long-term secretion of IL2 by CD4⁺ T cells (27). Similarly, increased expression of activating transcription factor (ATF) 3 that is a positive regulator of IFNγ gene expression (28) supports our observation of increased Th1 cytokine response in p53-KO T cells. The cell division cycle 6 (Cdc6), a target of p53 that coordinates S phase and mitosis was also upregulated in p53-KO T cells. (29). Similarly, Ctnnb1, a gene that encodes β-catenin protein, is upregulated in p53-KO T cells. Ptk2, protein tyrosine kinase 2 (also known as focal adhesion kinase), which plays an important role in T cell–antigen-presenting cell conjugation, and HMOx1 encoded heme-oxygenase-1 levels, a target of p53 (30), were also increased in p53-KO T cells. The expression of inhibitor of dna binding 2 (Id2), which promotes generation of distinct CD8⁺ T-cell memory, was also increased in p53-KO T cells (31). Tumor suppressor p53 is an essential partner of Smads, affecting TGFβ signaling at various points in the pathway (32). Importantly, the expression of mitogen-activated protein kinase kinase kinase 7 (Map3K7) was reduced in p53-KO T cells. This kinase mediates the signal transduction induced by TGFβ and controls a variety of cell functions, including transcription regulation and apoptosis (33). Expression of E2F4—a transcription factor that plays a crucial role in the control of cell cycle and regulating antigen recall response in CD8⁺ T cells—was also elevated in p53-KO T cells (34). The p53-KO T cells expressed higher Gadd45b, which augments antitumor immune response by enhancing the expression of IFNγ, granzyme B, CCR5 in T cells (35), and protecting from apoptosis by p38 activation and JNK inhibition (36). WFS1, a gene encoding an endoplasmic reticulum (ER) membrane protein and involved in survival of pancreatic β-cells, was also found to be upregulated in p53-KO T cells. As expected, we also found that Myc levels were enhanced in p53-KO T cells, which also showed higher HIF1α and glycolytic commitment. Thus, targeting p53 in T cells result in modulating TGFβ-mediated signaling molecules.

To determine the translational potential of inhibiting p53, pharmacological inhibitors PFT-μ and PFT-α pretreated human tyrosinase TCR TIL1383I transduced T cells were characterized. Upon p53 inhibition, we noticed a significant increase in glucose uptake as measured by 2-NBDG (Fig. 7A), which also correlated to an increased fraction of IFNγ secreting cells upon antigen restimulation (Fig. 7B). Importantly, AICD was also reduced when PFT-μ pretreated TIL1383I TCR-transduced T cells were restimulated with the cognate tyrosinase epitope (Fig. 7C). Additionally, p53 inhibition not only downregulated ROS and RNS, but also reduced the expression of CD95, CD95L, exhaustion molecules Lag3 and PD1 on TCR-activated T cells (Supplementary Fig. S2B) Thus, pharmacologically inhibiting p53 in human TCR-transduced cells mimicked results of increased metabolic activity in form of glycolysis, with increased effector functions and lower cell death. Further, we observed that pretreatment with either PFT-μ or PFT-α results in fewer human T cells exhibiting FoxP3 expression under iTreg ex vivo programming condition (Supplementary Fig. S2C). Lastly, tracking studies using human T cells transduced with melanoma reactive TIL1383I TCR that were pretreated with a combination of p53 inhibitors and transferred into NSG-A2 mice showed increased persistence at 72 hours as compared with the untreated counterparts (Fig. 7D). Thus, we believe that the strategy to inhibit p53 expression is potentially translatable and could improve the efficacy of ACT.

p53 is regarded as the “guardian of genome integrity” due to its complex role in regulating cellular differentiation (37). More than 50% of tumors have a direct mutation of p53, which promotes invasion, metastasis, proliferation, and cell survival (38). p53 is also a central regulator of the glycolysis and TGFβ signaling pathways (9, 19). Therefore, we hypothesized that rendering properties of higher proliferation, lower cell death, and increased persistence to CTLs by lowering its p53 expression could improve adoptive T-cell immunotherapy. Our data establish that p53-deficient tumor-specific CD8⁺ T cells exhibit increased glycolytic commitment that correlates to higher IFNγ secretion, increased persistence due to high stem-cell–related gene expression, reduced susceptibility to immunosuppression and iTreg conversion due to reduced TGFβ signaling. All these features result in a robust effector phenotype that leads to improved tumor control.

A recent study showed that antigen-specific proliferative responses of CD4⁺ T cells require downregulation of tumor suppressor p53 (7), and that inhibiting p53-regulating protein Mdm2 resulted in its sustained expression and prevented proliferation. Our data show that T cells obtained from the TCR transgenic mice h3T developed on p53-KO background (i.e., h3T-p53 KO) proliferate rapidly, and maintain antigen specificity upon TCR stimulation. Given the role of p53 in negatively regulating cell-cycle progression by blocking cyclin D1 (16), we observed that h3T-p53 KO T cells exhibited higher expression of cyclin D and lower levels of cyclin inhibitors that correlate to increased proliferation leading to enlarged spleen and thymus. However, increased proliferation was not associated with shedding of CD62L molecule (3, 39), because we observed that h3T-p53 KO T cells exhibit CD62L^hi central memory (Tcm) phenotype. Importantly, h3T-p53 KO T cells also exhibited elevated expression of stem-cell–specific transcription factors such as Tcf7, Lef-1, and PRDM1. Notably, p53 has been shown to bind at the promoter region of Oct4 and Nanog, which are required for self-renewal and maintenance of embryonic stem cells in an undifferentiated state, and reduce their gene transcription (40). Given a recent report that a subset of memory T cells with stem-cell–like phenotype (referred to as Tscm) exists within the Tcm fraction (41), it is likely that the Tscm phenotype is increased in the p53-KO T cells.

p53 also regulates aerobic respiration at the glycolytic and OXPHOS steps via transcriptional regulation of its downstream genes TIGAR and SCO2 (10, 11). In line with the role of p53 as a negative regulator of glycolysis (11), our data show that p53-KO T cells exhibit higher glycolytic commitment as observed by glucose uptake and increased expression of key glycolytic genes. This increase in glycolysis corresponds to the reduced expression of TIGAR, an inhibitor of the fructose-2, 6-bisphosphate, which is normally activated by p53 to regulate glycolysis (11). Another property of p53-deficient T cells was their ability to persist longer and exhibit lower degree of AICD. It has been shown that p53 leads to upregulation of a number of prooxidant enzymes as quinone oxidoreductase, proline oxidase, BAX, and PUMA, leading to oxidative stress and consequently to apoptosis (42–44). Upon its mitochondrial translocation, p53 binds to and inhibits MnSOD, playing a direct role in promoting ROS formation and eventually in apoptosis (45). Thus, an inverse correlation between p53 and antioxidant capacity may have contributed to the increased persistence and antitumor T-cell response. Interestingly, a detailed necropsy of the tumor-bearing recipient animals showed increased inflammatory reactions, without any evidence to suggest that this was due to the transfer of p53 KO T cells (data not shown).

Another confounding factor that limits long-term tumor control by ACT is suppressive tumor microenvironment, which is abundant with suppressive cytokines such as TGFβ. Importantly, key cellular responses to TGFβ signals have been shown to rely on p53 family members (9). This study shows that p53-KO T cells display an impaired response to TGFβ signals. Additionally, Smad and p53 protein complexes converge on separate cis binding elements on a target promoter and synergistically activate TGFβ-induced transcription (6). Thus, it is likely that p53-KO T cells displayed diminished transcriptional activation of key TGFβ target genes (32). It has also been shown that p53 enhances the transcription of Treg signature transcription factor Foxp3 by binding to the promoter and the conserved noncoding DNA sequence-2 of the Foxp3 gene (8), and that fewer nTregs and iTregs are obtained from p53-KO mice. Our data also confirm this observation, which implies that impaired TGFβ signaling molecules may be responsible for reduced plasticity in p53-deficient T cells. It is also likely that increased glycolytic commitment identified in h3T-p53 KO T cells by elevated levels of ECAR values, glycolytic genes, and HIF1α metabolically downregulates Treg differentiation, because HIF1α has been shown to attenuate Treg development by binding Foxp3 and targeting it for proteasomal degradation (46). Importantly, p53 pharmacological inhibitors also improved T cell–mediated tumor control, and pifithrin-pretreated murine and human T cells exhibited increased persistence in vivo. Overall, this study shows that p53 is a central regulator of multiple pathways (such as glycolysis, ROS, TGFβ signaling), and its inhibition could be important for ACT of tumor.

No potential conflicts of interest were disclosed.

Conception and design: A. Banerjee, K. Thyagarajan, C.M. Paulos, C. Voelkel-Johnson, S. Mehrotra

Development of methodology: A. Banerjee, K. Thyagarajan, P. Chakraborty, P. Kesarwani, K. Moxley, M.P. Rubinstein, C.M. Paulos, M.I. Nishimura, S. Mehrotra

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Banerjee, K. Thyagarajan, S. Chatterjee, P. Kesarwani, M. Soloshchenko, M. Al-Hommrani, K. Andrijauskaite, K. Moxley, H. Janakiraman, M.J. Scheffel, K. Helke, M.I. Nishimura, S. Mehrotra

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A. Banerjee, K. Thyagarajan, S. Chatterjee, P. Chakraborty, P. Kesarwani, K. Helke, K. Armenson, M.P. Rubinstein, E.-G. Mayer, S. Mehrotra

Writing, review, and/or revision of the manuscript: A. Banerjee, K. Thyagarajan, S. Chatterjee, K. Helke, M.P. Rubinstein, D.J. Cole, C. Voelkel-Johnson, M.I. Nishimura, S. Mehrotra

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Thyagarajan, M. Soloshchenko, M. Al-Hommrani, K. Moxley, K. Helke, V. Palanisamy, M.P. Rubinstein, C.M. Paulos, S. Mehrotra

Study supervision: S. Mehrotra

The authors acknowledge Drs. Zihai Li, Radhika Gudi, and Ephraim Ansa-Addo in the Department of Microbiology and Immunology at MUSC for their help with this article.

The work was supported in part by NIH grants R21CA137725 and R01CA138930 to S. Mehrotra and P01CA154778 to M.I. Nishimura. The Cell Evaluation and Therapy Shared Resource is supported by P30 CA138313.

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.