Female C57BL/6 mice and Pmel-1 T-cell receptor (TCR)-transgenic mice (B6.Cg Thy1^a-Tg(TcraTcrb)8Rest/J) were purchased from the Jackson Laboratories. OT-I Rag2^−/− mice have been previously described.23 OT-I and Pmel-1 mice were bred in-house. All mice were age matched (7–10 weeks old) at the beginning of each experiment and maintained under specific pathogen-free conditions and housed in the Laboratory Animal Resources facility. All animal studies were conducted in accordance with and approved by the Institutional Animal Care and Use Committee at the Roswell Park Comprehensive Cancer Center.

Generation of mouse embryonic fibroblast-derived iPSC lines, 2A-4F-118 and 2A-4F-136, was previously described.24 OP9 and OP9 cells expressing a notch ligand, delta-like-1 (OP9-DL1) cells were purchased from RIKEN (Japan). The iPSC lines were maintained in Gibco Dulbecco's Modified Eagle Medium (DMEM) containing 15% fetal bovine serum (FBS) (Sigma-Aldrich), 1% non-essential amino acid (NEAA) (Gibco), 2 mM L-glutamine (Gibco), 0.5% penicillin/streptomycin (Gibco), 55 µM 2-mercaptoethanol (2-ME) (Gibco), 0.5 µM PD0325901 (mitogen-activated protein kinase inhibitor) (Stemgent), 3 µM CHIR99021 (glycogen synthase kinase-3 inhibitor) (Stemgent), and 1000 U/mL of leukemia inhibitory factor (Millipore) on mitomycin-C-treated SNL cells (Cell Biolabs) as described.25 SNL cells were cultured in DMEM containing 7% FBS (Sigma-Aldrich), 2 mM L-glutamine, and 0.5% penicillin/streptomycin. OP9 and OP9-DL1 cells were cultured on gelatin-coated dishes in OP9 medium: αMEM supplemented with 20% non-heat inactivated FBS (Biowest), 0.5% penicillin/streptomycin and 2.2 g/L of sodium bicarbonate (Sigma-Aldrich).

B16-F10 (B16) tumor cell lines were purchased from the American Type Culture Collection (ATCC), authenticated at ATCC and maintained. The AT-3, B16-OVA, and MC38 cell lines were gift from Drs Scott Abrams, Sharon Evans (Roswell Park Comprehensive Cancer Center), and Weiping Zou (University of Michigan), respectively. AT-3 tumor cells expressing green fluorescent protein (GFP) (AT-3-GFP) were described.26 B16, B16-OVA and MC38 cells were cultured in RPMI 1640 (Gibco) supplemented with 10% FBS, 1% NEAA, 2 mM L-glutamine, 0.5% penicillin/streptomycin, and 55 µM 2-ME. AT-3 and AT-3-GFP cells were cultured in DMEM (Gibco) supplemented with 10% FBS (Sigma-Aldrich), 1% NEAA, 2 mM L-glutamine, 1 mM sodium pyruvate, 15 mM 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid, 0.5% penicillin/streptomycin, and 55 µM 2-ME. These cell lines were authenticated by morphology, phenotype and growth, and routinely screened for Mycoplasma, and were maintained at 37°C in a humidified 5% (B16, B16-OVA and MC38) or 7% (AT-3, AT-3-GFP) CO₂ atmosphere.

The differentiation of mouse iPSCs to DCs is composed of three steps as previously described,12 27 but step 2 was slightly modified in our laboratory. In brief, iPSCs (1×10⁵ cells) were seeded onto OP9 cell layers in 10 cm dishes and cultured in OP9 media (step 1). After 7 days of incubation, cells were collected, suspended with OP9 media containing 100 ng/mL of mouse granulocyte macrophage colony-stimulating factor (mGM-CSF) (Peprotech), and 1×10⁶ cells were seeded onto a newly prepared OP9-DL1 cell layer (step 2). On day 14, loosely adherent cells were collected by pipetting, and 5×10⁵ cells were transferred to new dish coated by 10% poly 2-hydroxyethyl methacrylate (pHEMA) (Sigma-Aldrich) without feeder cells and cultured in RPMI 1640 containing 10% FBS, 100 ng/mL of mGM-CSF and 55 µM of 2-ME for another 14 days (step 3). On day 28, the non-adherent cells were collected and used for further analysis or in vivo study.

Mouse BM-derived DCs (BM-DCs) were generated as described.27 BM cells from flushed marrow cavities of femurs and tibiae of female C57BL/6 mice aged 6–8 weeks old were cultured in RPMI 1640 supplemented with 10% heat-inactivated FBS (Sigma-Aldrich), 2 mM L-glutamine, 55 µM 2-ME, 0.5% penicillin/streptomycin, and 20 ng/mL mGM-CSF in petri dishes with 100 mm diameter (Falcon #351029, BD Biosciences) at 2×10⁶ cells/10 mL. At day 3, 10 mL of media containing 20 ng/mL mGM-CSF were added to the plates. At days 6 and 8, half of the culture supernatant was changed with media containing 20 ng/mL mGM-CSF. Non-adherent cells were collected on day 10.

CD8⁺ cells were isolated from Pmel-1 splenocytes using EasySep Mouse CD8α Positive Selection Kit (STEMCELL Technologies). The isolated cells were labeled with CellTrace carboxyfluorescein succinimidyl ester (CFSE) (Thermo Fisher Scientific), and seeded onto 24-well plate (2×10⁶ cells/well) in media containing interleukin (IL)-2 (60 IU/mL) (Prometheus Laboratories) with H-2D^b-restricted epitope of the influenza nucleoprotein (NP) peptide, NP_366–374 (ASNENMETM; GenScript) (1 µM) or human (h)gp100_25–33 peptide (KVPRNQDWL: GenScript) (1 µM). Then, iPSC-DCs (1×10⁶) stimulated with mouse IL-4 (mIL-4) (Peprotech) (10 ng/mL), mouse tumor necrosis factor alpha (mTNFα) (Peprotech) (5 ng/mL), and lipopolysaccharides (LPS) (Sigma-Aldrich) (0.5 µg/mL) for 2 days were added to each well and co-cultured for 2 days. In vitro-activated Pmel-1 CD8⁺ T cells were evaluated for proliferation, surface markers, and cytokine production against NP_366–374 and hgp100_25–33 peptides.

AT-3 (5×10⁵) tumor cells were implanted into the left fourth mammary gland under anesthesia with isoflurane. B16 (5×10⁵) tumor cells were injected subcutaneously on the left flank. To test the establishment of immunological memory, mice that had durable complete responses to the treatment and fresh naïve C57BL/6 mice were injected subcutaneously with AT-3 (5×10⁵) and MC38 (5×10⁵) tumor cells in the right flank and on the back, respectively. To generate mice bearing bilateral tumors, AT-3 (5×10⁵) tumor cells were implanted into the left fourth mammary gland (primary tumors) under anesthesia with isoflurane, and AT-3 (5×10⁵) tumor cells were injected in the right flank (distant tumors) subcutaneously 2 days after the inoculation for primary tumors. iPSC-DCs (1×10⁶ cells) in 50 µL phosphate buffered saline (PBS) were injected intratumorally every 4 days. Irradiation of mammary and subcutaneous tumors was conducted as recently described.26 Briefly, the mice were anesthetized with isoflurane and positioned for exposure to radiation (9 Gy) under a 2 mm thick lead shield containing 1 cm² hole, limiting exposure to the tumors. Irradiation was performed with an orthovoltage X-ray machine (Philips RT250, Philips Medical Systems) at 75 kV using a 1×2 cm cone. Radiation was given 1 day after first DC injection and repeated every 6 days. Anti-PD-L1 antibody (Ab) (clone 10F.9G2, BioXCell) or rat IgG_2b Ab (clone LTF-2, BioXCell) was injected intraperitoneally at a dose of 200 µg/mouse every 3 days starting on the day first iPSC-DCs were injected. Tumor growth was measured three times a week, and the volumes were calculated by determining the length of short (l) and long (L) diameters (volume=l²×L/2). Experimental endpoints were reached when tumors exceeded 20 mm in diameter or when mice became moribund and showed signs of lateral recumbency, cachexia, lack of response to noxious stimuli, or observable weight loss.

Single-cell suspensions of blood and tumors were prepared for flow cytometric analysis. Red blood cells were lysed using ACK Lysis Buffer (Life Technologies). Cells were incubated with Abs in PBS containing 2% FBS for 20 min at room temperature after blockade by anti-CD16/CD32 (BD Biosciences). The following Abs were used: anti-CD8 (clone 53-6.7, BD Biosciences), anti-CD45 (clone 30-F11, Invitrogen), anti-CD90.1 (clone Ox-7, BioLegend), anti-programmed cell death protein 1 (PD-1) (clone 29F.1A12, BioLegend), anti-CD127 (clone A019D5, BioLegend), anti-interferon (IFN) regulatory factor-4 (IRF4) (clone IRF4.3E4, BioLegend), anti-IRF8 (clone V3GYWCH, Thermo Fisher Scientific), anti-Slamf6 (clone 330-AJ, BioLegend), anti-T cell immunoglobulin and mucin domain-containing protein 3 (TIM3) (clone B8.2C12, BioLegend), anti-IFNγ (clone XMG1.2, Thermo Fisher Scientific), anti-TNFα (clone MP6-XT22, Thermo Fisher Scientific), anti-IL-2 (clone JES6-5H4, BD Biosciences), anti-CD62L (clone MEL-14, BioLegend), anti-CD25 (clone PC61, BioLegend), anti-Ly6C (clone HK1.4, BioLegend), anti-CD11b (clone M1/70, BD Biosciences), anti-CD11c (clone HL3, BD Biosciences), anti-I-A^b (clone AF6-120.1, BioLegend), anti-XCR1 (clone ZET, BioLegend), anti-CD24 (clone M1/69, BD Biosciences), and anti-F4/80 (clone BM8, BioLegend). LIVE/DEAD Fixable Near-IR Dead Cell Stain Kit (Thermo Fisher Scientific)-stained cells were excluded from the analysis. To identify apoptotic cells, cells were stained with annexin V, Alexa Fluor 488 conjugate (#A13201, Invitrogen). For analysis of IRF4 and IRF8 expression in iPSC-DCs, iPSC-DCs (1×10⁶) were stimulated with mIL-4 (10 ng/mL), mTNFα (5 ng/mL), and LPS (0.5 µg/mL) for 2 days. Cells were harvested, permeabilized with True-Nuclear Transcription Factor Buffer Set (BioLegend) and stained with anti-IRF4 and IRF8 Abs. For intracellular staining of cytokines, in vitro-activated Pmel-1 CD8⁺ T cells were co-cultured with NP_366–374 (1 µM) or hgp100_25–33 (1 µM) and splenocytes from C57BL/6 mice in the presence of Brefeldin A (BD Biosciences) for 5 hours. Cells were harvested, permeabilized with Fixation and Permeabilization Kit (BD Biosciences) and stained with anti-IFNγ, TNFα, and IL-2 Abs. Samples were analyzed using LSRII or LSRFortessa (BD Biosciences) with FlowJo software (TreeStar).

iPSC-DCs (1×10⁶ per 50 µL) labeled with 5-(and-6)-(((4-chloromethyl) benzoyl) amino) tetramethylrhodamine (CellTracker Orange CMTMR, #C2927, Invitrogen) were injected into the established orthotopic AT-3 or AT-3-GFP tumors in the left fourth mammary gland. After 24 hours, mice were untreated or treated with RT (9 Gy). TdLNs (left inguinal LN) were harvested 24 hours after RT for flow cytometry or imaging flow cytometry. ImageStream-MKII was used for imaging flow cytometry (Luminex Corporation) which enables spatially correlated image analysis of spectrally separated cell imagery. Data were analyzed using IDEAS software (Luminex Corporation). Immunophenotypically defined cell aggregates were identified as described.26

Statistical analysis was performed using two-tailed Student’s t-test or Mann-Whitney U test (for comparisons between two groups), or the Mantel-Cox method (log-rank test) for survival analysis using GraphPad Prism V.9.02 (GraphPad Software). Data are presented as mean±SEM.

In this study, we aimed to generate DCs from previously established mouse embryonic fibroblasts-derived iPSC clones, 2A-4F-118 and 2A-4F-136.24 First, we tested whether mouse iPSCs could differentiate into functional DCs using the previously established protocol12 with some modifications based on the emerging evidence that notch signaling facilitates in vitro generation of mouse and human DCs28 29 (figure 1A). In this study, we differentiated mouse iPSCs on OP9 feeder layers for 6 days (step 1). On day 6, cells were harvested and transferred onto OP9-DL1,30 and cultured in the presence of GM-CSF for 7 days (step 2). On day 13, loosely adherent cells were recovered and transferred to a dish coated with pHEMA to create a hydrophobic surface and prevent cell adhesion. On day 26, the majority of the floating iPSC-derived cells were positive for CD11c and Major Histocompatibility Complex (MHC) class II similar to BM-derived cells differentiated in the presence of GM-CSF (figure 1B). Mouse iPSC-derived cells expressed CD24, CD11b, F4/80, CD80 and IRF4, but not B220, CD8α, XCR1 and IRF8, resembling conventional type 2 DCs (cDC2) (figure 1C,D).31 32 Mouse iPSC-DCs also expressed DEC205, which is an endocytic receptor of DCs and mediates cross-presentation of antigens onto MHC class I.33 34 Hereafter we focus on clone 2A-4F-118 for further experiments.

Next, we tested whether iPSC-DCs could activate antigen-specific CD8⁺ T cells in vitro. To this end, we labeled Pmel-1 CD8⁺ T cells that can recognize hgp100_25–33 with CFSE, and co-cultured them with iPSC-DCs in the presence of hgp100_25–33. NP epitope NP_366–374 was used as an irrelevant control peptide. Pmel-1 CD8⁺ T cells vigorously proliferated in response to hgp100_25–33 but not to NP_366–374 (figure 2A). Phenotypic analysis revealed that hgp100 but not NP peptide pulsed-iPSC-DCs downregulated L-selectin (CD62L) and upregulated CD25 and PD-1 in proliferating Pmel-1 CD8⁺ T cells (figure 2B). Furthermore, Pmel-1 CD8⁺ T cells activated by hgp100-pulsed iPSC-DCs had a capacity to produce effector cytokines, IFNγ, TNFα and IL-2 against hgp100_25–33 but not NP_366–374 (figure 2C). Collectively, these findings indicate effective activation and effector differentiation of antigen-specific CD8⁺ T cells by iPSC-DCs.

Given the capacity of iPSC-DCs to activate and expand antigen-specific CD8⁺ T cells in vitro, we next investigated in vivo antitumor efficacy of iPSC-DCs. Local irradiation to the tumor can induce immunogenic tumor cell deaths, resulting in the release of tumor-associated antigens (TAAs), which are uptaken, processed and presented to T cells by DCs.15–21 Our previous work suggests that a combination of intratumoral BM-DC injections and RT synergistically increases antitumor efficacy compared with administration of BM-DCs or RT alone in preclinical models.22 These observations prompted us to test in vivo antitumor efficacy of iPSC-DCs in combination with RT, using two syngeneic orthotopic mouse models of poorly T cell-inflamed tumors, AT-3 triple negative breast cancer and B16 melanoma, which display primary resistance to PD-1/PD-L1 blockade.26 Although each single therapy of in situ iPSC-DC injection or RT exhibited modest tumor growth delay and improved survival in AT-3 bearing mice, combination of these markedly delayed tumor growth and prolonged survival, indicating synergistic antitumor efficacy of this combination therapy (figure 3A, online supplemental figure 1A). Furthermore, we also observed the synergistic antitumor reactivity of the combined therapy against rapidly growing B16 tumors (figure 3, online supplemental figure 1B)

Higher density of CD8⁺ T cells in the TME correlates with better prognosis in various types of cancers including melanoma and breast cancer,35 and response to anti-PD-1/PD-L1 therapy.36 In contrast, high frequency of tumor-infiltrating myeloid cells and tumor-associated macrophages (TAMs) associates with poor prognosis and resistance to immune checkpoint blockade.26 37 38 Therefore, we examined by flow cytometry analysis whether combined therapy could alter the frequency of tumor-infiltrating CD8⁺ T cells and CD11b⁺ myeloid cells (online supplemental figure 2A). Despite increased antitumor efficacy, we found that the frequency of CD8⁺ tumor-infiltrating lymphocytes (TILs) and a CD8⁺ T cell to CD11b⁺ myeloid cell ratio were unchanged by the combination therapy (online supplemental figure 2B). We then evaluated the myeloid compartment of the TME (online supplemental figure 3A). The frequency of DCs was increased in mice treated with intratumoral injection of iPSC-DCs alone (online supplemental figure 3B). However, adding RT to in situ iPSC-DC administration decreased the frequency of DCs in the TME. The frequency of TAMs was decreased by the combination therapy. These findings suggest that the majority of injected iPSC-DCs remained in situ, and that RT may have facilitated their trafficking to the TdLN.

To investigate this notion, we took two complementary approaches. First, mice bearing AT-3 tumors were intratumorally injected with 1×10⁶ fluorescently labeled iPSC-DCs, and were treated with or without RT (figure 4A). Twenty-four hours later, tumors and TdLN were analyzed for flow cytometry (online supplemental figure 4). The frequency of iPSC-DCs was lower in the tumor and higher in the TdLN in RT-treated mice compared with mice which did not receive RT (figure 4B). We also examined the viability of iPSC-DCs in the tumor and TdLN by flow cytometry (online supplemental figure 4). We did not observe many apoptotic iPSC-DCs in the tumor while 40%–60% of iPSC-DCs in the TdLN express annexin V regardless of local RT (figure 4C).

To further confirm the trafficking of iPSC-DCs to the TdLN, we examined the frequency of injected iPSC-DCs in mice bearing AT-3-GFP tumors by imaging flow cytometry (online supplemental figure 5). Consistent with the finding from flow cytometric analysis, RT enhanced migratory capacity of iPSC-DCs (online supplemental figure 6A) with increased frequencies of GFP⁺ iPSC-DCs (figure 4D, online supplemental figure 6B) and CD40⁺ iPSC-DCs (figure 4E, online supplemental figure 6C) in TdLN. Furthermore, RT-treated mice exhibited increased frequencies of GFP⁺ iPSC-DCs and GFP⁺ iPSC-DC-XCR1⁺ DCs in contact with CD8⁺ T cells in TdLN (figure 4F,G and online supplemental figure 6D,E). Taken together, these results suggest that RT augments trafficking of intratumorally injected iPSC-DCs, upregulates CD40 and facilitates a cross-talk with other DC subsets and CD8⁺ T cells in TdLN.

Emerging evidence has indicated that intratumoral Slamf6⁺ PD-1 intermediate (PD-1^int) TIM3⁻ progenitor exhausted CD8⁺ T cells with stem-like properties but not Slamf6⁻ PD-1 high (PD-1^hi) TIM3⁺ terminally exhausted CD8⁺ T cells mediate cellular expansion and tumor control in response to anti-PD-1/PD-L1 therapy.39 40 Therefore, we evaluated Slamf6 and TIM3 expression in CD8⁺ TILs in mice bearing AT-3 tumors (figure 5A). We found that the frequency of Slamf6⁺ TIM3⁻ CD8⁺ TILs was higher in CD8⁺ TILs from mice treated with combined therapy of in situ injection of iPSC-DCs and RT (figure 5B). This subset expressed higher levels of CD62L and CD127 with an increased ratio of PD-1^int to PD-1^hi cells compared with Slamf6⁻ TIM3⁺ CD8⁺ TILs (figure 5C), suggesting that combined therapy facilitated the infiltration of the stem-like progenitor exhausted subpopulation.

PD-L1 can be adaptively induced in the TME by inflammatory factors secreted by tumor-specific T cells.36 Therefore, we examined whether combined therapy would increase PD-L1 expression in tumor-residing myeloid cells. PD-L1 expression was synergistically upregulated in TAMs and DCs by in situ injection of iPSC-DCs and RT (figure 5D).

Findings of increased PD-L1 expression in the TME prompted us to hypothesize that combined therapy would facilitate the priming of antigen-specific CD8⁺ T cells that infiltrate into the TME. To this end, we used an adoptive transfer system of OVA-specific TCR transgenic OT-I CD8⁺ T cells into mice bearing OVA-expressing B16 (B16-OVA) tumors. The frequency of infused OT-I CD8⁺ T cells (CD90.1) in C57BL/6 mice (CD90.2) was evaluated after treatment with intratumoral iPSC-DC administration and/or RT (figure 6A). We found that in situ injection of iPSC-DCs and RT increased the frequency of tumor-infiltrating OT-I CD8⁺ T cells (figure 6B). Increased frequency of adoptively transferred Pmel-1 CD90.1⁺ CD8⁺ T cells was also observed in C57BL/6 mice bearing B16 tumors treated with the combined therapy (online supplemental figure 7). Notably, higher frequency of Slamf6⁺ TIM3⁻ CD8⁺ T cells was identified in infiltrating OT-I T cells compared with endogenous (CD90.2) CD8⁺ T cells (figure 6C), suggesting that the combined therapy triggers an increase of progenitor exhausted tumor-specific T cells in the tumor. Of note, we evaluated function of iPSC-DCs in comparison with BM-DCs in vivo (figure 6A), and found that iPSC-DCs and BM-DCs demonstrated equivalent antitumor efficacy (figure 6D) and capacity of increasing tumor-specific Slamf6⁺ TIM3⁻ CD8⁺ TILs (figure 6E,F) with local RT.

Given the increased frequency of tumor-specific CD8⁺ T cells as well as the potent local antitumor response by combination of in situ injection of iPSC-DCs and RT, we next investigated whether this combination therapy would trigger a systemic antitumor immune response. To this end, we used a bilateral mouse tumor model, treated the primary tumor, and monitored the growth of untreated distant tumors (figure 7A). Combination therapy substantially increased the frequency of circulating effector memory CD44⁺ CD62L⁻ CD8⁺ T cells (figure 7B). While in situ iPSC-DC administration alone or RT alone had no effect on distant tumor growth, the combination of these attenuated distant tumor growth (figure 7C, online supplemental figure 8). Consistent with these findings, combined therapy synergistically increased CD8⁺ T cells (figure 7D), and a CD8⁺/CD11b⁺ cell ratio (figure 7E), in the untreated distant tumor. Synergistic increase of CD8⁺ TILs in the distant tumor was confirmed by immunohistochemistry analysis (figure 7F). Together, these results indicate that combined therapy of in situ administration of iPSC-DCs and local RT induces systemic antitumor immunity and is capable of controlling the growth of distant tumors.

Our findings of the increased stem-like progenitor exhausted CD8⁺ TILs and upregulation of PD-L1 expression in tumor-residing macrophages and DCs led us to postulate that antitumor efficacy of the combination therapy could be augmented by PD-1/PD-L1 blockade. To address this, we treated AT-3-bearing mice with combination of in situ injection of iPSC-DCs, RT, and anti-PD-L1 Ab or isotype Ab (figure 8A). Although anti-PD-L1 Ab therapy did not alter the immunogenicity of RT, it substantially improved tumor growth and survival in mice treated with a combination of in situ injection of iPSC-DCs and RT (figure 8B,C). Furthermore, mice with durable regressions rejected the AT-3 tumor rechallenge, but not unrelated MC38 colon adenocarcinoma, demonstrating the establishment of tumor-specific immunological memory (figure 8D). Taken together, these findings suggest in situ induction of iPSC-DCs and RT could overcome primary resistance to anti-PD-L1 therapy.