Adoptive T-cell transfer therapy is an FDA- approved treatment for leukemia that relies on the ex vivo expansion and reinfusion of a patient's immune cells, which can be engineered with a chimeric antigen receptor (CAR) for more efficient tumor recognition. Type 17 T cells, controlled transcriptionally by RORγ, have been reported to mediate potent antitumor effects superior to those observed with conventionally expanded T cells. Here, we demonstrate that addition of a synthetic, small-molecule RORγ agonist during ex vivo expansion potentiates the antitumor activity of human Th17 and Tc17 cells redirected with a CAR. Likewise, ex vivo use of this agonist bolstered the antitumor properties of murine tumor-specific CD4+ and CD8+ T cells. Expansion in the presence of the RORγ agonist enhanced IL17A production without compromising IFNγ secretion in vitro. In vivo, cytokine neutralization studies revealed that IFNγ and IL17A were required to regress murine melanoma tumors. The enhanced antitumor effect of RORγ agonist treatment was associated with recovery of more donor T cells in the tumor and spleen; these cells produced elevated levels of cytokines months after infusion and expressed markers of long-lived stem and central memory cells such as Tcf7 and CD62L. Conversely, untreated cells mainly exhibited effector phenotypes in the tumor. Cured mice previously treated with agonist-primed T cells were protected from tumor rechallenge. Collectively, our work reveals that in vitro treatment with a RORγ agonist generates potent antitumor Type 17 effector cells that persist as long-lived memory cells in vivo.
Significance: RORγ agonists can be used in vitro during T-cell expansion to enhance the efficacy of adoptive cell therapy (e.g., CAR-T) and to provide long-term protection against tumors.
Graphical Abstract: http://cancerres.aacrjournals.org/content/canres/78/14/3888/F1.large.jpg. Cancer Res; 78(14); 3888–98. ©2018 AACR.
The nuclear receptor RORγt is a master transcription factor that controls the development of CD4+ and CD8+ lymphocytes that secrete IL17A, called T helper 17 (Th17) and T cytotoxic 17 cells (Tc17), respectively (1). RORγt also plays a role in the differentiation of IL17A-producing innate immune cells, such as innate lymphoid cells, NK cells and γδ T cells (2, 3). Although defined by IL17 secretion, Type 17 cells are polyfunctional effectors that can co-secrete IL22 and IFNγ upon tumor recall responses and possess durable memory properties in vivo (4, 5). The involvement of Th17 and Tc17 cells in autoimmunity, tumor immunity, and mucosal defense has been well established (6, 7). Importantly, a recent case report of a patient with colon cancer treated with the checkpoint inhibitor pembrolizumab revealed that IL17 blockade with secukinumab provided dramatic relief of immune-mediated skin toxic effects but was associated with a subsequent loss of the antitumor efficacy, suggesting that the IL17/Th17 axis plays a role in the antitumor effects of immunotherapy (8).
While IL17+T cells are abundant in the mucosal tissues and support gut-related homeostasis (9), few such cells exist in the blood of healthy individuals. However, many Th17 and Tc17 cells infiltrate tumors, especially compared with the density of these cells in the nontumor tissue of patients (10). This heightened presence of Type 17 cells in tumor tissue holds true for many types of malignancies, implying that tumors themselves produce factors that promote RORγt expression. Approximately 15% of human CD4+ T cells in tumors express RORγt and this transcription factor is induced by cytokines TGFβ and IL6, both produced at high levels in inflamed tissues and transformed cells (11). We reported that RORγ activation with a novel small molecule agonist potentiates the function of antitumor Th17 cells to a greater extent than the endogenous agonist desmosterol (11, 12). This activation is associated with enhanced cytokine production and CTL activity as well as reduced Treg formation in vitro and effective antitumor immunity in syngeneic models (11).
In the context of cellular therapy for cancer, several reports have shown that Type 17 T cells eradicate large human and murine tumors to a far greater extent than bulk CD4 T cells, Th1 or Th2 cells (4, 5, 13–15). Additional investigation revealed that tumor-specific RORγt+ Th17 cells persisted longer than T-bet expressing IL2-expanded Th0 and Th1 cells due to their stem-like memory properties. Thus, we were interested in how the addition of a synthetic RORγ agonist to the ex vivo expansion of TCR or CAR T-cell cultures would affect their function, memory phenotype, persistence, and antitumor activity when infused into mice with large murine or human tumors. To address this question, murine pmel-1 TCR transgenic CD8, TRP-1 TCR transgenic CD4 and mesothelin human CAR T-cell models were used (4, 16, 17).
Herein, we report that the RORγ agonist LYC-54143 potentiates the antitumor activity of Th17 and Tc17 cells when added to ex vivo cell expansion cultures of both CAR-expressing human T cells and tumor-specific CD4 and CD8 T cells. LYC-54143 treatment generates cells that produce elevated effector cytokines and increased markers associated with stem-like memory T cells (18, 19). When lymphocytes from mice receiving LYC-54143–treated Th17 plus Tc17 cells were analyzed after eradicating or controlling tumors, we found that the donor cells were more prevalent and were composed of diverse memory phenotypes compared with mice unfused with untreated cells. In addition, the agonist-primed cells expressed heightened Tcf7, a transcription factor downstream of the canonical Wnt pathway essential for stemness (20) and long-lasting immunity to cancer (19). Importantly, mice cured with LYC-54143–primed cells were protected from repeated tumor challenges several months later. Collectively, our work reveals that antitumor Type 17 cells generated in the presence of a small-molecule RORγ agonist have enhanced antitumor activity and persist as long-lived memory cells in vivo. These data support the utility of including these drugs as part of cell manufacturing and expansion protocols.
Materials and Methods
Mice and tumor lines
C57BL/6J (B6), TRP-1 TCR transgenic mice, pmel-1 TCR transgenic mice, and NOD/scid/γ chain knock out (NSG) mice were purchased from The Jackson Laboratory, housed, and bred in the Medical University of South Carolina Hollings Cancer Center (MUSC, Charleston, SC). NSG mice were housed under specific pathogen-free conditions in micro-isolator cages and given autoclaved food and acidified water. Housing and experiments were conducted with Institutional Animal Care and Use Committee's approval at Medical University of South Carolina. B16F10 (H-2b) melanoma was maintained in culture media (RPMI 1640 w/L-glutamine, 10% FBS, 1% penicillin–streptomycin, NEAA, and Na Pyruvate, and 0.1% BME and Hepes). M108 xenograft tumors were cultured and engrafted as described previously (15).
TRP-1 splenocytes (which contain MHC-II–restricted CD4+ T cells expressing TRP-1-recognizing transgenic TCR Vβ14) were activated using 10 Gy-irradiated B6 456 splenocytes (feeder cells) pulsed with 1 μmol/L TRP-1 peptide and polarized to a Th17 phenotype at 2 × 106 cells/2 mL of cell media in one well of a 24-well plate with the following cocktail: 100 ng/mL rhIL6 (NIH repository), 100 ng/mL rhIL21 (Shenandoah), 30 ng/mL rhTGFβ1 (Biolegend), 10 ng/mL rhIL1β (NIH Repository), 10 μg/mL each of anti-mIFNγ clone XMG1.2, anti-mIL4 clone 11B11, and anti-mIL2 clone JES6-1A12 (Bio X Cell). Th0 polarization occurred under peptide activation with irradiated feeder cells with the following cocktail: 100 IU/mL rhIL2 (NIH repository). Cultured cells were supplemented with new media containing 100 IU/mL rhIL2 (NIH repository) throughout expansion. In experiments where indicated, RORγ agonist LYC-54143 (synthesized at Lycera) was added to the cultures during peptide activation and then again 2 days later (10 μmol/L). For further description of RORγ agonist and their use in immunotherapy of cancer, see, for example, international patent application publication WO 2015/131035.
Pmel-1 T cells.
pmel-1 splenocytes (which contain MHC-I-restricted CD8+ T cells expressing gp100-recognizing transgenic TCR Vβ13) were activated using 1 μmol/L hgp100 peptide + 100 IU rhIL2/mL and primed on day 1 with type 17-polarizing cytokines with our without RORγ agonist (as above for TRP-1 cells). Cells were supplemented with culture media containing 100 IU rhIL2/mL and expanded as indicated. CD8+ T-cell cultures were in vitro activated with feeder cells and peptide 12 hours before transfer.
Human normal donor peripheral Th17/Tc17 cells
To generate mesothelin-specific T cells, sorted pan T (CD4+ or CD8+) cells were activated with CD3/CD28-coated beads and programmed to a Th17 or Tc17 phenotype and then transduced with a chimeric anti-mesothelin single-chain variable fragment (scFv) fusion protein containing the T-cell receptor ζ (TCRζ) signaling domain and 4-1BB that was generated as described previously (13). CD4 T cells or CD8+ T cells were polarized to Th17 or Tc17 phenotype as follows: 10 ng/mL rhIL1β, 10 ng/mL rhIL6, 20 ng/mL rhIL23, 10 μg/mL anti-hIL4 clone 11B11, and anti-hIFNγ clone H22 (eBioscience). Cells were either primed with RORγ agonist or not at 10 μmol/L on day 0 and 2 postbead activation. Experiments were conducted with fetal calf serum containing endogenous sources of TGFβ. Cell cultures were maintained with 100 IU/mL of rhIL2 and cells were expanded for up to two weeks.
Adoptive cell therapy
B6 mice were given 4.5 × 105 B16F10 cells subcutaneously and tumors were allowed to establish between 7 and 10 days before ACT. One day before therapy, mice received nonmyeloablative 5 Gy total body irradiation. T cells were infused via tail vein. NSG mice were given 5 × 106 M108 suspended in Matrigel subcutaneously. Tumors were allowed to establish for 35 days prior to adoptive therapy. In all experiments, mice were randomized to treatment groups and tumor burden was monitored in blinded fashion using perpendicular caliper measurements and reported as tumor area (mm2).
Tissue distribution assays
Spleens from treated mice were harvested and mechanically disrupted using the tip of a syringe plunger. Cells were filtered through a wire mesh, red blood cells lysed with RBC lysis buffer (Biolegend), and then resuspended in cell media for analysis. Tumors were sectioned, then incubated in 1 mg/mL collagenase type II (Life Technologies) at 37°C for 1 hour. Digested tissue was filtered, resuspended in cell media, and plated for assay. Before probing with antibodies, FC block (Biolegend) was applied to cells at 1 μg/100 μL. TRP-1 donor T cells were identified as CD4+Vβ14+ cells while pmel-1 donor T cells as CD8+ Vβ13+ cells.
Flow cytometry and ELISA
Flow cytometry was performed with a BD FACSverse 500 instrument. Intracellular staining of cytokines were conducted using IC fixation and permeabilization system (Thermo Fisher Scientific). Staining of transcription factors was conducted using FOXP3 fixation and permeabilization system (Thermo Fisher Scientific) per manufacturer's instructions. Antibodies used were anti-mCD3-efluor450 clone 17A2, IL17A, and IFNγ, anti-h/mRORγt-PE clone AFKJS-501 9, anti-h/mCD44-PerCPCy5.5 clone IM7, anti-hCD4-APCH7 clone RPA-T4, anti-mCD4-APC/PE clone RM4-5, and anti-mCD62L- APC clone MEL-14 (eBioscience). Anti-human CD45, anti-mouse Vβ13, anti-mouse TCF7 were from Biolegend, and anti-mouse Vβ14 was purchased from BD Biosciences. ELISA for IL17A, IL22, and IFNγ was performed using DuoSet ELISA kits (R&D Systems) per the manufacturer's instructions.
Kaplan–Meier survival curves were assessed for significance using a log-rank test between treatment groups. A P value of <0.05 was considered significant. Comparisons between two groups were analyzed using Student t tests with Welch's correction for parametric distribution or Mann–Whitney signed rank tests for nonparametric distribution. A P value of <0.05 was considered significant. For comparisons between multiple groups, a one-way ANOVA was performed followed by multiple comparisons. A P value of <0.05 was considered statistically significant.
RORγ agonist augments the function of CAR human Th17 and Tc17 cells
We reported that a series of RORγ agonists could augment the effector function of murine Type 17 T cells in vitro and improve their antitumor activity when administered as an oral therapy in mice bearing syngeneic tumors (11). To extend these results and evaluate how RORγ agonists would affect human chimeric antigen receptor (CAR) T cells, human total T (CD4+ and CD8+) cells were enriched from the peripheral blood of healthy individuals, programmed using Type 17 polarizing conditions (IL1β, IL6, IL23), activated with anti-CD3/CD28 beads and 1 day after activation, transduced with a lentiviral vector that encodes a CAR recognizing mesothelin (Fig. 1A). The redirected T cells were further expanded in the presence of IL2 and IL23. The results showed that 10 days after expansion, a representative agonist (LYC-54143) increased RORγt by approximately 20% compared with untreated Type 17 cells with similar numbers of total cells in both conditions (Fig. 1B). Consistent with the findings in mouse T cells (11), this in vitro agonist treatment induced the cells to produce 6-fold more IL17A without compromising their ability to produce IFNγ (Fig. 1C and D). Moreover, CAR transduction efficiency was not compromised in T cells treated with this agonist (Supplementary Fig. S1). Previously, we reported that Type 17 CAR T cells primed in the presence of an agonist lyse human tumors better than untreated cells (11). We next sought to test if the function of CAR Type 17 cells was heightened when reactivated against a battery of different tumors expressing mesothelin, including leukemia (K562-meso), mesothelioma (M108), ovarian cancer (Ov79), and pancreatic cancer (Panc1; ref. 21). We found that agonist-treated CAR Type 17 cells secreted more IL17A when cocultured with mesothelin-positive tumors compared with untreated CAR T cells (Fig. 1E). As an important control, IL17A production was nominal when CAR T cells were incubated with mesothelin-negative tumor lines (such as K562 or K562 lines overexpressing CD19), regardless of if they were primed with an agonist. Collectively, our data indicate that this RORγ agonist augments the functional capacity of human Type 17 cells.
Transfer of human CAR Type 17 cells regresses mesothelioma in vivo when primed in vitro with a RORγ agonist
T cells redirected with CAR constructs containing the ICOS cytoplasmic tail induce RORγ and IL17 expression and regress human tumors more effectively than those activated through CD28 (10). Moreover, activator beads coated with anti-CD3/ICOS augment the function and antitumor activity of human CAR Th17 cells more effectively than those stimulated with anti-CD3/CD28 beads (13). Consequently, we posited that a small molecule that activates RORγt in human CAR Type 17 cells stimulated with anti-CD3/CD28 beads would regress tumors in vivo more effectively than untreated cohorts. To test this idea, we co-infused Th17/Tc17 cells that had been redirected with a mesothelin CAR, primed with the RORγ agonist LYC-54143 and activated with CD3/CD28 beads into NSG mice bearing a human mesothelioma. As a control, an equal number of untreated CAR Th17/Tc17 cells were co-infused into mesothelioma-bearing mice. As in Fig. 2A, RORγ agonist–treated cells mediated the best antitumor activity in mice. While the infusion of untreated cells was initially effective, the response was less durable than that observed with infused cells treated in vitro with the RORγ agonist. Importantly, these cells must be redirected with a CAR to mediate tumor regression, as mock transduced Th17/Tc17 cells (untransduced) were only slightly (not significantly) different from untreated animals. We next asked if extended duration of antitumor activity and regression observed with agonist-primed CAR Th17/Tc17 cells might be the result of better persistence of these cells in vivo compared with untreated CAR cells. Indeed, 30 days after infusion, more than twice as many agonist-treated cells were detected in tumor than untreated control cells (Fig. 2B). Collectively, our data suggest that in vitro priming of human CAR Th17/Tc17 cells with a RORγ agonist augments their capacity to secrete IL17 and conditions the cells for better persistence in vivo, which is associated with long-term regression of mesothelin-positive tumors.
Th17 cells co-infused with Tc17 cells mediate potent anti-melanoma activity in vivo when primed in vitro with a RORγ agonist
To assess if RORγ agonist–treated Type 17 (Th17 and Tc17) T cells mediate durable antitumor responses in mice and the impact of a RORγ agonist on the antitumor activity of IL2 expanded T cells, traditionally used in adoptive cell therapy (ACT) clinical trials (22–24), we employed the MHCI-restricted transgenic TCR pmel-1 CD8+ T cell and MHCII-restricted TRP-1 CD4+ T cell tumor ACT mouse models that recognize either gp100 or TRP-1 antigens, respectively, on B16F10 melanoma. First, we determined the impact of a RORγ agonist on IL2-expanded T cells (Th0) using the TRP-1 CD4 T cell tumor model. In these experiments, TRP-1 Th0 cells were generated in vitro from the MHCII-restricted transgenic TCR mice using antigen, irradiated antigen presenting cells and IL2 without polarizing cytokines in the presence or absence of RORγ agonist LYC-54143. RORγ agonist treatment enhanced IL17A production in Th0 cells in a dose-dependent manner; however, the absolute titer was considerably lower than that observed from cytokine-polarized Th17 cells (Supplementary Fig. S2). The production of IFNγ by IL2-expanded Th0 cells was not affected by RORγ agonist treatment (Fig. 3A). Interestingly, when the in vitro–treated cells were infused into mice with established B16F10 tumors, LYC-54143 treatment improved the antitumor activity of Th0 cells (Fig. 3B; Supplementary Fig. S3A). Overall, more than half of the mice infused with agonist-treated Th0 cells had an extension of life for approximately 10 more days. However, this therapy did not promote long-term cures and most animals relapsed and did not survive for more than 1 month after ACT.
We next compared the antitumor activity of Th17 with Th0 cells using the TRP-1 system. Consistent with a previous report (4), we also found that Th17 cells provided better antitumor activity over Th0 cells (Fig. 3C vs. Fig. 3B). However, superior antitumor activity was mediated by Th17 cells treated in vitro with LYC-54143 with long-term regressions observed (Fig. 3C; Supplementary Fig. S3B) and the majority of animals surviving 2 months after ACT (Supplementary Fig. S3C). Interestingly, at the time of infusion, agonist-treated Th17 cells produced only slightly more IL17A and IL22 than untreated Th17 cells likely due to the robust cytokine cocktail used to polarize the cells in vitro (Fig. 3D). Despite the modest effect on cytokine production in vitro, the addition of the RORγ agonist to the Th17 cultures imparted an improvement in antitumor activity in vivo highlighting the role of additional, cytokine-independent pathways induced by RORγ activation. Similar to TRP-1 Th0 and Th17 CD4 T cells, pmel-1 CD8 Tc17 cells also possessed enhanced antitumor properties following in vitro expansion with RORγ agonist LYC-54143 (Fig. 3E; Supplementary Fig. S3D).
The human CAR T experiment utilized pan T cells, to assess the contributions of both Th17 and Tc17 subsets to mediate durable antitumor responses we co-infused equal numbers of TRP-1 Th17 cells and pmel-1 Tc17 cells into mice bearing established B16F10 melanoma. The coinfusion of both Th17 and Tc17 cells resulted in an effective treatment (Fig. 3F; Supplementary Fig. S3E). These data are consistent with previous publications showing that Th17 cells can augment the activation of CD8+ T cells and adoptive transfer of CD4+ T cells helps maintain the function of transferred CD8+ T cells (14, 25). Even in this setting, expansion in the presence of RORγ agonist LYC-54143 further augmented the antitumor effect with all animals achieving complete or partial regressions. These data might imply that the most effective antitumor responses are mediated by combination of Th17 and Tc17 cells and that the addition of RORγ agonists to even highly polarized effectors cells further augments their antitumor activity.
IL17 and IFNγ, not IL22, production by agonist-treated T cells mediates antitumor immunity
The plasticity of Th17 cell cytokine production and their ability to produce IFNγ have been reported to be important to their antitumor activity (5). Thus, we next set out to determine the importance of IL17, IL22, and IFNγ to the antitumor efficacy observed following coinfusion of agonist-primed Th17 and Tc17 cells. To address this question, we programmed pmel-1 Tc17 or TRP-1 Th17 cells in the presence or absence of the RORγ agonist and infused them into lympho-depleted melanoma-bearing mice. As expected, neutralizing IL17A or IFNγ in mice infused with untreated cells (Fig. 4A and B) or agonist-treated cells (Fig. 4C and D) impaired their antitumor activity. The most impaired treatment outcome in mice cytokine ablated of IL17A or IFNγ occurred in mice treated with the most effective therapy (i.e., agonist-treated cells; Fig. 4C and D). Conversely, neutralizing IL22 did not affect treatment outcome in mice, suggesting that this cytokine does not alter tumor immunity, at least in the context of adoptive T-cell transfer therapy. Collectively, our data reveal that IL17 and IFNγ but not IL22 production by agonist-treated co-infused Th17 and Tc17 cells are important for treatment outcome.
Agonist-primed Th17 and Tc17 cells persist, co-secrete elevated IL17 and IFNγ, and possess a distinct memory profile
We have previously shown that higher numbers of RORγ agonist–treated OT-1 Tc17 T cells can be found in the spleen and tumor of mice compared with untreated Tc17 cells (11). As agonist-treated Th17 plus Tc17 cells control tumor growth in mice to a greater extent than untreated cells, we sought to determine if these cells persisted in mice to a greater extent than untreated cells. To address this question, we analyzed the frequency, function, and memory profile of donor cells in mice that experienced long-term antitumor activity in vivo (>71 days after ACT). As we found that infusing 5 × 105 untreated donor cells into mice bearing 10-day established tumors was ineffective and did not protect mice long term (Fig. 3F). Therefore, for these studies, we infused more donor cells (2 × 106) into mice with smaller tumors so that both groups of mice receiving agonist-treated cells and untreated cells would survive long term for us to test this question in both treatment groups. As shown in Fig. 5A, 71 days after infusion, TRP-1 Th17 cells primed with a RORγ agonist persisted in the tumor at 2-fold higher levels than untreated Th17 cells. Likewise, pmel-1 Tc17 cells primed with RORγ were detected in the tumors at approximately 4-fold greater levels than untreated cohorts (Fig. 5A). On the day of transfer approximately 10% of cells were producing IFNγ regardless of LYC-54143 treatment during in vitro priming (Fig. 3D); however 71 days after transfer, the percentage of IFNγ+ cells is increased (Fig. 5B). Interestingly, in vitro treatment with RORγ agonist 71 days prior resulted in maintenance of more IL17+ cells (2.8% vs. 8.3%) and the development of more IL17+IFNγ+ double positive population (Fig. 5B).
Moreover, both Th17 and Tc17 cells, when in vitro primed with an agonist 71 days prior, secreted more IL17A when reactivated with their cognate antigens (Fig. 5C). Interestingly, when the memory phenotype of the transferred cells in the tumor was examined by flow cytometry, we found that the agonist-primed Th17 and Tc17 cells possessed a wider repertoire of central memory (CD44+CD62L+) and stem-like memory (CD44-CD62L+) 71 days after infusion into tumor-bearing mice (Fig. 6A and B). Conversely, untreated cells mainly consisted of central and effector memory cells (CD44+ CD62L−; Fig. 6A and B). Of note, this finding was more striking with Th17 cells than Tc17 cells (Supplementary Fig. S4). Additional investigation revealed that agonist-primed Th17 cells in the tumor expressed more Tcf7 than untreated donor cells (Fig. 6C), suggesting that agonist therapy supports the generation of T cells with durable stem memory. The memory phenotype of transferred cells is in sharp contrast to in vitro–activated cells, which predominantly are CD44+CD62L− effector cells. Collectively, our data reveal that a short in vitro exposure to RORγ agonist bolsters Th17 and Tc17 cells, which co-secrete more cytokines, have superior in vivo persistence, and develop a distinct memory phenotype during responses to tumors in vivo.
RORγ agonist treatment in vitro induces long-term T-cell memory and drives durable protection after ACT
As agonist-primed cells possessed a stem-like memory phenotype, we hypothesized that mice receiving this therapy would be protected from tumor rechallenge. To address this question, we co-infused TRP-1 Th17 and pmel-1 Tc17 cells generated in vitro in the presence or absence of the RORγ agonist LYC-54143 to mice bearing B16F10 melanoma tumors as above. Sufficient numbers of cells were infused to mediate full tumor regression in both groups (Supplementary Fig. S5A). We then rechallenged these mice that had survived long term from cellular therapy with a second subcutaneous injection of B16F10 melanoma 45 days after adoptive transfer. As a control, we gave melanoma to previously untreated (naïve) mice. As shown in Fig. 7A, we found that tumors grew rapidly in naïve mice. In contrast, mice previously infused with anti-melanoma Th17 and Tc17 cells without agonist treatment were initially protected from tumor rechallenge. However, 2 weeks after rechallenge, tumors began to grow in these animals. In contrast, mice were protected for more than 1 month after tumor rechallenge if they had originally received agonist-primed Th17 and Tc17 cells (Fig. 7A). Also, 75 days after adoptive transfer of T cells, when we rechallenged these mice a third time with B16F10 tumor cells, they remained protected from melanoma for approximately 20 days with 2 of 4 mice remaining tumor free. Conversely, if these mice were rechallenged with EL4 tumors, the malignancy grew (Fig. 7B), showing that the memory response was antigen specific. Consistent with the antitumor effect in each animal, blood levels of donor cells inversely correlated with tumor growth (Supplementary Fig. S5B). Thus, collectively RORγ agonists can dramatically potentiate the stem-like memory phenotype of Type 17 T cells and provide superior, long-term protection against tumor challenges in vivo.
Herein, we report that a novel RORγ agonist can potentiate the function, persistence, and antitumor activity of both human CAR T cells and murine Th0, Th17, and Tc17 cells in two distinct and clinically relevant mouse tumor models.
We found that the antitumor activity was improved in mice co-infused with agonist-treated T cells. Interestingly, RORγ agonist treatment not only supported survival of T cells but also helped them co-secrete IL17A, IL22, and IFNγ. It is important to appreciate that while Type 17 cells mediated the most potent antitumor properties in vivo, the IL2-expanded, unprogrammed Type 0 cells (normally used in the clinic) were also augmented by RORγ agonist treatment. These data suggest that the addition of a small-molecule RORγ agonist can be rapidly translated into ACT expansion protocols currently used in the clinic.
In mice surviving from agonist-Type 17 therapy for more than 2 months, we found that donor cells were composed of diverse populations of memory cells (including stem, central, and effector). While mice infused with untreated cells were mainly effectors with some central memory lymphocytes. Remarkably, agonist-primed cells within tumors expressed heightened Tcf7, a transcription factor downstream of the canonical Wnt pathway essential for stemness (20) and directly associated with lymphocytes with durable immunity to cancer (19). Importantly, this altered memory phenotype of cells treated in vitro with a RORγ agonist several months prior were functionally superior to untreated cells and provided protection from repeated tumor challenges. Collectively, our work reveals that antitumor Type 17 cells provide durable memory responses in vivo when treated in vitro with a RORγ agonist that bolsters, at least in part, the Wnt/TCF7 pathway.
Type 17 T cells mostly provide antitumor immunity when transferred into mice with established tumors as shown in this work and many other reports (4, 10, 14, 21); however, both pro- and antitumor activities of IL17A have been reported, with most protumor activities observed in inflammation-induced tumorigenesis in the gut (7, 26). In our ACT model, blocking IL17A at least partially reduced the antitumor activity of transferred Type 17 T cells, suggesting an antitumor role of IL17A (Fig. 4). Interestingly, a recent report indicates that treatment of anti-PD1–induced autoimmune toxicity by anti-IL17A secukinumab resulted in loss of antitumor activity in a patient with metastatic colon cancer (8). Similarly, in patients with melanoma receiving PD-1 therapy, the frequency of IL17-producing T cells was increased in responders versus nonresponders (27). Together these results suggest that RORγ and Type 17 cells play important roles in cancer immunotherapy and may mediate the antitumor activity of checkpoint inhibitors, such as anti-PD1.
There is a significant need to generate durable memory in T cells in the field of cancer immunotherapy. Persistence of transferred lymphocytes correlates with cancer regression in patients receiving adoptive cell therapy (28, 29). Many investigators are enriching CD62L+ T cells from bulk TILs or CAR T-cell cultures, as these central memory T cells are more efficacious than the effector memory CD62L− T cells normally expanded with high dose IL2 under rapid expansion protocols. Other researchers are adding distinct cytokines (such as IL15; ref. 30) or pharmacologic reagents (such as Akt inhibitors; ref 31) to cultures to generate central or stem memory T cells.
To our knowledge, our work is the first demonstration that small molecules stimulating the activity of one transcription factor—RORγt—can profoundly affect the persistence and function of T cells. We also have shown that LYC-54143, as a representative RORγ agonist, affects multiple pathways through elevating cytokines, enhancing survival of T cells and expression of costimulatory receptors, and blocking Treg generation and decreasing multiple coinhibitory receptors (11). All of these activities could also contribute to the improved antitumor immune responses observed here. Indeed, when TRP-1 and pmel-1 Type 17 T cells in the tumor were analyzed 71 days after transfer, reduced percentage of Treg cells and decreased PD-1 expression were found in agonist-primed cells (Supplementary Fig. S6). It is remarkable that a short in vitro exposure of tumor-specific T cells to a small-molecule RORγ agonist not only enhances the in vivo survival of cells, but also results in persistent effects on cytokine production and the generation of long-lived memory cells following adoptive transfer. These results suggest that RORγ agonist treatment likely induces long-lasting epigenetic changes in T cells. Future studies will assess possible epigenetic effects of RORγ agonists.
Our findings have immediate translational implications as additional investigation revealed that small-molecule RORγ agonist can also enhance the antitumor activity of IL2-expanded, nonpolarized T cells in vivo. This finding is particularly exciting, as these cells are commonly used in clinical trials for adoptive T-cell transfer therapy. Nonpolarized T cells do express RORγt albeit at lower levels than T cells polarized under Type 17 conditions. In humans, RORγt is expressed in some circulating T cells and in TILs (13). Thus, it is conceivable that a RORγ agonist could be easily added to nearly any tumor-specific T-cell culture such as TIL expansions or bulk lymphocytes redirected with antigen receptors (TCRs or CARs) to endow long-lived memory responses to tumors. In addition to the potential utility of RORγ agonists in in vitro priming of T-cell therapies, clinical studies of an oral RORγ agonist LYC-55716 in patients with cancer are currently ongoing (NCT02929862). Collectively, our findings have important implications in the field of cancer immunotherapy, including adoptive T-cell transfer therapy as well as other T-cell–based therapies such as vaccines, cytokines, and immune checkpoint inhibitors.
Disclosure of Potential Conflicts of Interest
X. Hu has ownership interest (including patents) in Lycera Corp. Y. Liu has ownership interest (including patents) in and is a consultant/advisory board member for Lycera Corp. L.L. Carter has ownership interest (including patents) in Lycera Corp. C.M. Paulos reports receiving commercial research grant from Lyceea. No potential conflicts of interest were disclosed by the other authors.
Conception and design: X. Hu, X. Liu, C. Spooner, J. Moisan, W. Zou, L.L. Carter, C.M. Paulos
Development of methodology: K. Majchrzak, L.L. Carter, C.M. Paulos
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Majchrzak, X. Liu, M.M. Wyatt
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): X. Hu, X. Liu, M.M. Wyatt, J. Moisan, L.L. Carter, C.M. Paulos
Writing, review, and/or revision of the manuscript: X. Hu, X. Liu, C. Spooner, J. Moisan, L.L. Carter, C.M. Paulos
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): X. Hu, C.M. Paulos
Study supervision: X. Hu, L.L. Carter, C.M. Paulos
We thank all chemists at Lycera Corp. for making RORγ agonist compounds. This work was supported by funding from Lycera Corp.
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.