Neoadjuvant in situ immunomodulation enhances systemic antitumor immunity against highly metastatic tumors

Abstract

Neoadjuvant immunotherapy, given before surgical resection, is a promising approach to develop systemic antitumor immunity for the treatment of high-risk resectable disease. Here, using syngeneic and orthotopic mouse models of triple-negative breast cancer, we have tested the hypothesis that generation of tumor-specific T-cell responses by induction and activation of tumor-residing Batf3-dependent conventional type 1 dendritic cells (cDC1) before resection improves control of distant metastatic disease and survival. Mice bearing highly metastatic orthotopic tumors were treated with a combinatorial in situ immunomodulation (ISIM) regimen comprised of intratumoral administration of Flt3L, local radiotherapy, and in situ TLR3/CD40 stimulations, followed by surgical resection. Neoadjuvant ISIM generated tumor-specific CD8⁺ T cells that infiltrated into distant non-irradiated metastatic sites, which delayed the progression of lung metastases and improved survival after the resection of primary tumors. The efficacy of neoadjuvant ISIM was dependent on de novo adaptive T-cell immunity elicited by Batf3-dependent DCs and was enhanced by increasing dose and fractionation of radiotherapy, and early surgical resection after the completion of neoadjuvant ISIM. Importantly, neoadjuvant ISIM synergized with PD-L1 blockade to improve control of distant metastases and prolong survival, while removal of tumor-draining lymph nodes abrogated the antimetastatic efficacy of neoadjuvant ISIM. Our findings illustrate the therapeutic potential of neoadjuvant multimodal intralesional therapy for the treatment of resectable tumors with high risk of relapse.

Significance:

Neoadjuvant induction and activation of cDC1s in primary tumors enhances systemic antitumor immunity, suppresses metastatic progression, improves survival, and synergizes with anti-PD-L1 therapy.

Introduction

Neoadjuvant therapy is an established standard of care therapy for solid tumors with high risk of relapse such as breast cancer (1). This approach offers advantages over adjuvant therapy for several reasons: i) it can confirm therapeutic efficacy and direct the choice of adjuvant therapy; ii) it can reduce tumor size and facilitate surgical resection; iii) it provides valuable information regarding pathological response data as surrogate outcome markers for relapse-free and overall survival; and iv) it can eradicate occult distant micrometastatic disease without delay. In addition to these benefits, neoadjuvant immunotherapy may allow patients to develop tumor-specific systemic immunity and/or immunological memory while the tumors remain “in situ” during treatment, a condition which is more difficult to acquire with systemic treatment after surgical resection (2-4). Indeed, evidence from preclinical studies revealed that neoadjuvant immunotherapy is more efficacious than adjuvant immunotherapy (5,6).

Among immunotherapeutic approaches, targeting programmed cell death protein-1 (PD-1)/PD1 ligand-1 (PD-L1) immune checkpoints has transformed the treatment of various advanced and metastatic cancers (7,8). This paradigm shift has led to testing this strategy in the neoadjuvant setting, which revealed that neoadjuvant anti-PD-1/PD-L1 blockade therapy was feasible, caused few side effects, did not delay surgery, and induced a major pathological response in a significant number of patients with a variety of cancer types (2-4,9). However, it remains to be determined whether this approach stimulates systemic antitumor immunity, controls distant micrometastatic disease, and improves survival in patients.

Triple-negative breast cancer (TNBC), defined by the lack of estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2, is frequently associated with an increased risk for early recurrence and high mortality (10). However, TNBC is characterized by a high degree of mutational burden and high PD-L1 expression compared with other subtypes of breast cancer (11,12), suggesting that TNBC might be an attractive target for PD-1/PD-L1 blockade therapy. Supporting this possibility, combined chemo-immunotherapy with anti-PD-L1 antibody (Ab) plus nab-paclitaxel was found to prolong progression-free survival in patients with PD-L1 positive tumors (13), and was approved for the treatment of advanced and metastatic TNBC by the Food and Drug Administration in 2019. While neoadjuvant chemotherapy is frequently utilized for the management of early-stage TNBC, a recent clinical study has shown that anti-PD-1 Ab and chemotherapy is superior to placebo and chemotherapy in achieving a pathologic complete response (pCR) in the neoadjuvant setting (9).

Clinical outcomes in patients with TNBC are strongly influenced by the tumor immune microenvironment. The presence of tumor-infiltrating lymphocytes (TILs) associates with response not only to chemotherapy but also to immunotherapy in TNBC (14-18). However, the baseline frequency of CD8⁺ TILs is low in TNBC (18), and only a subset of TNBC patients show clinical benefit to PD-1/PD-L1 blockade therapy (15,19). Therefore, strategies that increase CD8⁺ TILs are likely to increase response to neoadjuvant immunotherapy and/or chemo-immunotherapy.

Radiotherapy (RT) has been used for the management of breast cancer for many years, and neoadjuvant radiotherapy is commonly utilized in patients with various cancers such as lung and gastrointestinal cancers. RT is known to exert direct cytotoxic effects on tumor cells; however, recent research is revealing its influence on the immunogenicity of tumors, thus affecting the overall outcome of RT (20,21). There is a growing consensus that RT triggers immunogenic cell death and local release of type I interferon (IFN) (22), enhances antigen processing and cross presentation (23), and leads to infiltration of cytotoxic T cells (24,25). While RT alone is usually insufficient to overcome the immunosuppressive tumor microenvironment (TME), strategies to boost immune-stimulating effects of RT are under intensive investigation (20,21,26).

Batf3-dependent conventional type 1 dendritic cells (cDC1s; migratory CD103⁺ and lymphoid CD8α⁺ DCs in mice, and CD141⁺ DCs in humans) are critical for priming and expansion of tumor-specific CD8⁺ T cells (27-32), facilitating their infiltration into the TME (33), and thus enhancing the efficacy of PD-1/PD-L1 blockade (34,35). Emerging evidence suggests that induction and activation of tumor-residing cDC1s greatly enhances the therapeutic efficacy and immunogenicity of RT (36,37). We have recently reported that a combinatorial in situ immunomodulation (ISIM) regimen comprised of in situ delivery of Fms-like tyrosine kinase 3 ligand (Flt3L), RT (9 Gy), and dual toll-like receptor 3 (TLR3)/CD40 stimulation 1) mobilizes cDC1s to the TME; 2) promotes maturation of cDC1s; 3) facilitates trafficking of cDC1s carrying tumor antigens to tumor-draining lymph nodes (TdLN); 4) elicits de novo adaptive T cell immunity; 5) induces an influx of stem-like Tcf1⁺ Slamf6⁺ CD8⁺ T cells in the tumor; 6) decreases intratumoral macrophages, polymorphonuclear and CX3CR1⁺ monocytic myeloid-derived suppressor cells (MDSCs) via IFN regulatory factor 8 (IRF8); and 7) renders myeloid-enriched, poorly T cell-inflamed tumors responsive to anti-PD-L1 therapy (37,38). However, it remains unknown whether ISIM-induced systemic antitumor immunity could control distant micrometastases and improve survival in the neoadjuvant setting.

In this study, we are testing the hypothesis that in situ induction and activation of cDC1s in the primary tumor before resection not only enhances systemic tumor-specific T-cell immunity, but also controls growth of distant metastases and improves survival using spontaneously metastatic TNBC mouse models. We investigate the determinants of neoadjuvant ISIM-induced systemic antitumor immunity, the role of TdLN, and potential synergy with anti-PD-L1 therapy. Our study strongly supports new clinical evaluation of the potential of induction and activation of tumor-residing cDC1s in the neoadjuvant setting for high-risk resectable cancers.

Materials and Methods

Mice

Female C57BL/6 mice and Batf3^−/− mice on C57BL/6 mice background were purchased from the Jackson Laboratories and were bred in-house. Female Balb/c-AnNCr mice were from Charles River Laboratories. All mice were age matched (7–10 wk old) at the beginning of each experiment and kept under specific pathogen-free conditions and housed in the Laboratory Animal Resources. All animal studies were conducted in accordance with and approved by the Institutional Animal Care and Use Committee (IACUC) at Roswell Park Comprehensive Cancer Center.

Cell lines

The 4T1 and E0771 tumor cell lines were purchased from the American Type Culture Collection (ATCC) and CH3 BioSystems, respectively. The AT-3 cell line was gift from Dr. Scott Abrams (Roswell Park). Tumor cells expressing luciferase (4T1-luc, E0771-luc, and AT-3-luc) were generated with infection of lentiviruses encoding luciferase (pLenti PGK V5-LUC Neo, Addgene plasmid #21471). 4T1 and E0771 cells were cultured in RPMI 1640 (Gibco) supplemented with 10% fetal bovine serum (FBS) (Sigma-Aldrich), 1% Non-Essential Amino Acid (NEAA) (Gibco), 2 mM L-glutamine (Gibco), 0.5% penicillin/streptomycin (Gibco), and 55 μM 2-mercaptoethanol (Gibco). AT-3 and AT-3-luc cells were cultured in DMEM (Gibco) supplemented with 10% FBS, 1% NEAA, 2 mM L-glutamine, 0.5% penicillin/streptomycin, and 55 μM 2-mercaptoethanol. 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% (4T1 and E0771) or 7% (AT-3) CO₂ atmosphere.

Neoadjuvant in situ immunomodulation (neo-ISIM)

4T1-luc (2 × 10⁴), E0771-luc (5 × 10⁵), or AT-3-luc (5 × 10⁵) tumor cells were orthotopically implanted into the left fourth mammary gland of female mice under anesthesia with isoflurane. Tumor-bearing mice were treated with intratumoral administration of hFlt3L (10 μg/dose; Celldex Therapeutics, Inc.) in 30 μL PBS or control PBS for 5 consecutive days. After the completion of Flt3L injection, local radiation was performed with an orthovoltage X-ray machine (Philips RT250, Philips Medical Systems) at 200 kV, 1.0 mm Cu filter, 18.4 mA using a 1×2 cm cone (37). For radiation treatment, the mice were anesthetized with isoflurane and positioned under a 2 mm thick lead shield with small apertures limiting exposure to the tumors, and received a single dose of 9 or 15 Gy, or three fractions of 3 or 9 Gy in consecutive days. Mice were then treated with injection of agonistic anti-CD40 Ab (50 μg/dose; clone FGK4.5, BioXcell) and high molecular weight poly (I:C) (50 μg/dose; InvivoGen) at the peritumoral site subcutaneously. Surgical resection of the primary tumor was performed as described before (39). In some experiments, at the time of tumor implantation, we identified and removed the inguinal lymph node, which is the TdLN for tumors implanted to the fourth mammary fat pad (40). To test impact of radiation to the TdLN, left inguinal LN region was irradiated (9 Gy) two times after resection of the tumor. To establish bilateral established tumors, AT-3 (5 × 10⁵) tumor cells were orthotopically implanted to into the left fourth mammary gland, and two days later, AT-3 (5 × 10⁵) tumor cells were injected subcutaneously to the right flank. In some experiments, AT-3-luc cells (5 × 10⁶) were injected from tail vein at the day of surgery. AT-3 (5 × 10⁵) or E0771 (5 × 10⁵) cells were injected subcutaneously in the left or right flank for tumor rechallenge study. Tumor growth was measured 3-5 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.

Generation of bone marrow-derived CD103⁺ DCs

CD103⁺ DCs were generated from bone marrow cells as described (41). In brief, bone marrow cells were suspended in DC-induction medium (RPMI 1640 containing 10% FBS, 200 ng/ml human Flt3L, 5 ng/ml GM-CSF (Peprotech), penicillin/streptomycin and 50 μM 2-mercaptoethanol) at a cell density of 5 x 10⁵ cells/ml. Cell suspension (10 ml) was then plated on a 100-mm Petri dish (FALCON, #351029) and incubated at 37°C, 5% CO₂ for 5 days. On day 5, 5 ml of freshly prepared DC-induction medium was added to the culture and incubated for further 4 days. On day 9, floating cells were harvested and re-plated at 3 x 10⁶ cells in 10 ml DC-induction medium per Petri dish to expand the cells. The cells were incubated for further 6 days and the floating cells were collected and used on day 15 for experiments. For each dendritic cell preparation, DC marker expression was analyzed using flow cytometry with the majority of cells being positive for MHC class II, CD11c, CD11b, CD24, predominantly XCR1, DEC205, CD103. DCs (1 x 10⁶ cells) were injected intratumorally.

In vivo antibody treatment

For PD-L1 blockade, anti-PD-L1 Ab (clone 10F.9G2, BioXCell) or isotype control rat IgG2b (clone LTF-2, BioXCell) were given intraperitoneally every three days from the day RT performed at a dose of 200 μg/mouse for three times (37). For in vivo depletion of lymphocytes, 200 μg of anti-CD4 (clone GK1.5, BioXCell), anti-CD8β (clone Lyt 3.2, BioXCell), or control rat IgG2b (clone LTF-2, BioXCell) were injected intraperitoneally every three days from the day when RT was given for three times (37). Depletion of each subset was confirmed on day 2 (data not shown).

Treatment with Fingolimod (FTY720)

FTY720 was given to mice to inhibit lymphocyte migration out of secondary lymphoid organs. FTY720 stock solution (10 mg/ml in water) was diluted to a 0.2 mg/ml in 3% Tween-20 directly before administration. Mice received a dose of 20 μg FTY720 or vehicle (3% Tween-20) intraperitoneally as a control (37). Therapy was initiated one day before RT and was given daily until the date of surgery.

Flow Cytometry

Single cell suspension from mouse blood and tumors were prepared for flow cytometric analysis. Lungs were digested by Collagenase/Hyaluronidase (Stem cell Technologies). Red blood cells in blood were lysed using ACK Lysis Buffer (Life Technologies). Cells were incubated with antibodies in PBS containing 2% FBS for 20 mins at room temperature after being blocked by anti-CD16/CD32 (BD Biosciences). Samples were acquired and analyzed using Fortessa (BD Biosciences) and FlowJo software (TreeStar), respectively. Antibodies used in this study are listed in Table S1.

In vivo bioluminescence imaging

Mice were injected with D-luciferin (1.5 mg/20 g body weight) intraperitoneally. In 10 minutes, images were obtained by in vivo bioluminescence imaging (IVIS Spectrum imager) with 1 min exposure (37). Quantification of bioluminescence signal was determined using the Living Image (PerkinElmer Inc.) and average radiance (Total Flux/cm²/Sr) was calculated, implementing standard region of interests (ROI) drawn over the tumor site.

Statistical analysis

Statistical analysis was performed using a two-tailed Student’s t-test or a Mann-Whitney U test for comparisons between two groups, a one-way ANOVA with Tukey’s multiple comparisons for comparisons more than 2 groups, a two-way ANOVA with Bonferroni post-hoc test, or the Mantel-Cox method (log-rank test) for survival analysis using GraphPad Prism 8.02 (GraphPad Software). P < 0.05 was considered statistically significant. Data are presented as mean ± SEM.

Results

Neoadjuvant in situ immunomodulation (neo-ISIM) enhances systemic tumor-specific T-cell immunity, and improves survival.

To test the potential of neo-ISIM to control the metastatic burden after primary tumor resection, we utilized a 4T1-luc tumor model, where mice develop spontaneous lung metastases following inoculation of tumor cells in the mammary fat pad (42). We treated 4T1-luc bearing mice with in situ administration of Flt3L, radiotherapy (RT) and dual TLR3/CD40 stimulation (neo-ISIM), resected the primary tumor 6 days after the neo-ISIM, and monitored tumor burden with in vivo bioluminescence imaging (IVIS) (Fig. 1A). In situ delivery of Flt3L markedly increased DCs including CD103⁺ DCs in the TME (Supplementary Fig. 1A, B). Upregulation of CD40 and CD86 was observed in CD103⁺ DCs at the TdLN after the completion of neo-ISIM (Supplementary Fig. 2A, B). Consistent with this, we found substantially increased frequency of activated and differentiated CD4⁺ and CD8⁺ T cells in peripheral blood (Supplementary Fig. 3A, B). A considerable reduction of tumor bioluminescence was observed in the primary tumor 1 day after neo-ISIM (Day 9) (Fig. 1B). Consistent with this, the neo-ISIM treated tumors were smaller and weighed less than PBS-injected control tumors at the time of resection on day 14 (Fig. 1C). All mice developed lung metastases by day 22; however, progression of metastases was substantially delayed in neo-ISIM treated mice compared to control mice (Fig. 1B).

Figure 1. Neoadjuvant in situ immunomodulation (neo-ISIM) enhances systemic tumor-specific T-cell immunity, and improves survival.

A, Treatment protocol of neo-ISIM for mice bearing orthotopic 4T1-luc tumors.

B, Representative bioluminescent imaging of four 4T1-luc tumor bearing mice per group treated with PBS (NT) or neo-ISIM at different timepoints are shown.

C, Picture of the resected tumors and mean tumor weight on day 14 in different treatment groups as indicated (n = 9 per group; top row, NT; bottom row, neo-ISIM). Scale bar, 10 mm.

D, Representative flow cytometric plots showing gp70 tetramer (Tet)⁺ CD8⁺ T cells and frequency of Tet⁺ cells among CD8⁺ T cells in peripheral blood and the lungs of mice treated with or without neo-ISIM (n = 9 per group). Numbers denote percent Tet⁺ cells. Lungs were harvested 6 days after TLR3/CD40 stimulation.

E, Survival curves in 4T1-luc-tumor bearing mice treated with PBS (NT) or neo-ISIM (n = 6-9 per group).

Two-tailed unpaired t-test (C, D), or log-rank (Mantel-Cox) test (E). Mean ± SEM. Data shown are representative of two or three independent experiments.

To gain insight into the mechanisms underlying the delayed progression of lung metastases by neo-ISIM, we sought to evaluate whether neo-ISIM could facilitate expansion and infiltration of tumor-specific CD8⁺ T cells into the lungs. To this end, we used a tetramer (Tet) to identify H-2L^d-restricted MuLV gp70-specific CD8⁺ T cells in peripheral blood and the lungs of mice bearing orthotopic 4T1-luc tumors, which harbor gp70 epitopes as an endogenous tumor-associated antigen (TAA) (43). The frequency of Tet⁺ CD8⁺ T cells in circulation and the lungs was increased in neo-ISIM treated mice compared with untreated mice (Fig. 1D). In agreement with improved control of metastases and increased Tet⁺ CD8⁺ T cells in peripheral blood and the lungs, overall survival was markedly prolonged in neo-ISIM treated mice (Fig. 1E). To explore the applicability of this regimen against large palpable tumors, we started the treatment on day 6, and evaluated progression of lung metastases and survival (Supplementary Fig. 4A). Surgical resection was performed on day 13 to ensure complete resection and minimize local recurrence of untreated tumors. Neo-ISIM effectively inhibited the growth of large established tumors, curtailed metastatic progression, and improved survival (Supplementary Fig. 4B-D).

To confirm the generality of the findings observed in a 4T1-luc model, we set up similar experiments using E0771-luc tumors, another spontaneous metastatic tumor model (44). Again, we observed reduced primary tumor burdens and improved overall survival by neo-ISIM (Supplementary Fig. 5A, B). Surviving mice were rechallenged with E0771 in the contralateral mammary fat pad, and injected unrelated tumors, AT-3 on back. Age-matched, untreated nontumor experienced mice were used as controls. All surviving neo-ISIM treated mice rejected E0771 tumors while we found normal growth of unrelated AT-3 tumors (Supplementary Fig. 5C), suggesting the establishment of antigen-specific long-term immunological memory. Altogether, these results indicate that neo-ISIM generates systemic tumor-specific antitumor immunity, controls growth of distant metastatic disease, improves overall survival, and establishes immunological memory.

Synergistic antimetastatic efficacy of neo-ISIM is dependent on adaptive T cell immunity mediated by cDC1s.

We next examined whether CD4⁺ and CD8⁺ T cells were involved in the control metastatic disease of neo-ISIM by using neutralizing antibodies in 4T1-luc tumor models. Antitumor efficacy of neo-ISIM was substantially decreased or abrogated with depletion of CD4⁺ T cells or CD8⁺ T cells, respectively, suggesting that the efficacy of neo-ISIM was dependent on adaptive T cell immunity (Fig. 2A).

Figure 2. Synergistic antimetastatic efficacy of neo-ISIM is dependent on adaptive T cell immunity mediated by cDC1s.

A, Survival curves in 4T1-luc tumor bearing mice treated with neo-ISIM as described in figure 1A. Anti-CD4 and anti-CD8β depleting antibody (Ab), or isotype Ab were injected intraperitoneally every third day for three times from the day when RT was given (n = 7-8 per group).

B, C, Survival curves in AT-3-luc tumor bearing Batf3^−/− or wild type (WT) C57BL/6 mice treated with PBS (NT), neo-ISIM or neo-ISIM with an intratumoral injection of bone marrow-derived CD103⁺ dendritic cells one day before RT (B, n = 6-8, C, n = 6-7 per group). AT-3-luc cells (5 × 10⁶) were injected into tail vein at the day of surgery.

D, Survival curves in 4T1-luc tumor bearing mice treated with different combinatorial treatment (n = 5-9 per group).

NS not significant,*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, log-rank (Mantel-Cox) test. Mean ± SEM.

To assess the relevance of cDC1s in antitumor efficacy of neo-ISIM, we orthotopically implanted AT-3-luc tumor cells to Batf3^−/− mice and wild type (WT) C57BL/6 mice, treated them with neo-ISIM, and injected AT-3-luc cells intravenously when the primary tumors were resected. Therapeutic efficacy of neo-ISIM observed in WT mice was abrogated in Batf3^−/− mice (Fig. 2B). To further determine the role of tumor-residing cDC1s in therapeutic efficacy of neo-ISIM, bone marrow-derived CD103⁺ DCs were injected intratumorally into AT-3 tumors a day before RT in Batf3^−/− mice. In situ delivery of cDC1s partially restored the antitumor efficacy of neo-ISIM in Batf3^−/− mice (Fig. 2C), demonstrating the critical role of intratumoral cDC1s in the therapeutic efficacy of neo-ISIM.

To examine whether all components of neo-ISIM are needed to control metastatic disease, we treated 4T1-luc tumor bearing mice with Flt3L plus RT, Flt3L plus TLR3/CD40 agonists, RT plus TLR3/CD40 agonists or combination of all. We found that all three components synergistically increased the antitumor efficacy of multimodal intralesional therapy, and improved survival (Fig. 2D). Altogether, these results suggest that synergistic antitumor efficacy of multimodal neo-ISIM is dependent on adaptive T cell immunity elicited by cDC1s.

Neo-ISIM modulates systemic antitumor immunity via the egress of T cells from the LN.

Among various DC subsets, tumor-residing cDC1s are able to transport TAA to LN, where they can cross-present TAA to CD8⁺ T cells (28,35). To examine the role of secondary lymphoid organs (SLO) in the neo-ISIM treatment, mice were treated with vehicle control or FTY720 to inhibit new T cell migration from SLO (45). Although administration of FTY720 alone did not affect the progression of lung metastases, it substantially compromised the neo-ISIM-mediated control of lung metastases (Fig. 3A). Of note, we found that FTY720 significantly decreased the numbers of Tet⁺ CD8⁺ T cells in the resected primary tumors of mice treated with neo-ISIM (Fig. 3B). In line with these findings, FTY720 administration significantly worsened the overall survival in mice treated with neo-ISIM compared to mice treated with vehicle control and neo-ISIM (Fig. 3C).

Figure 3. Neo-ISIM modulates systemic antitumor immunity via the egress of T cells from the LN.

A, Mice were treated with PBS (NT) or neo-ISIM and primary tumors were resected on day 14 as described in figure 1A. Mice received FTY720 or vehicle intraperitoneally everyday till the surgery from one day before RT (n = 7-8 per group). Representative bioluminescent imaging of three 4T1-luc tumor bearing mice per group in different treatment as indicated at different timepoints are shown.

B, The numbers (/g) of gp70 tetramer (Tet)⁺ CD8⁺ T cells in resected primary 4T1-luc tumors (n = 7-9 per group).

C, Survival curves of mice in different treatment as indicated (n = 7-8 per group).

**p < 0.01, ***p < 0.001, one-way ANOVA with Tukey’s multiple comparisons (B) or log-rank (Mantel-Cox) test (C). Mean ± SEM. Data shown are representative of two independent experiments.

Tumor-draining lymph nodes (TdLN) are critical for neo-ISIM to generate systemic antitumor immunity.

At the time of resection of the primary tumors, we observed an increased size of TdLN in neo-ISIM treated mice compared to untreated mice (Fig. 4A). This finding and results from experiments using FTY720 (Fig. 3) together with a recent report showing the importance of TdLN for the generation of RT-induced abscopal effect (46) prompted us to investigate the significance of TdLN in neo-ISIM. To this end, we surgically resected the TdLN at the time of tumor implantation, and treated mice with or without neo-ISIM followed by the resection of the primary tumor. ISIM-induced reduction of the primary tumor was markedly compromised when TdLN was resected, (Fig. 4B). Consistent with this, TdLN resection resulted in larger metastatic burden in the lungs (Fig. 4C), and abrogated therapeutic efficacy of neo-ISIM (Fig. 4D). During surgical resection of the primary tumors, we also recognized that the TdLN was located close to the tumor within the field of RT (Supplementary Fig. 6A), suggesting that radiation to the TdLN might not negatively affect the antitumor efficacy of neo-ISIM. To test impact of radiation to the TdLN, we treated palpable left primary 4T1-luc tumors with neo-ISIM and surgery followed by radiation (9 Gy x 2 fractions) to the left inguinal region where TdLN is located (Supplementary Fig. 6B). We found that radiation to the TdLN had no impact on survival (Supplementary Fig. 6C), suggesting that this regimen is effective even when the tumors is close to the TdLN. Taken together, these results indicate that the presence of TdLN is a critical determinant of the antitumor efficacy of neo-ISIM.

Figure 4. Tumor-draining lymph nodes (TdLN) are a critical site for neo-ISIM to generate systemic antitumor immunity.

A, Representative pictures of TdLN of five 4T1-luc tumor bearing mice treated with PBS (NT) or neo-ISIM at day 14 in figure 1A. Scale bar, 10 mm.

B-D, Mice bearing orthotopic 4T1-luc tumors were treated with PBS (NT) or neo-ISIM as described in figure 1A (n = 8-10 per group). In some groups, TdLN were resected at the time of tumor implantation (day 0). Weight of the resected tumors on day 14 (B), representative bioluminescent imaging of three 4T1-luc tumor bearing mice per group in different treatment as indicated at different timepoints (C), and survival curves from all treated mice (n = 8-10 per group) (D) are shown.

NS not significant, **p < 0.01, ***p < 0.001, ****p < 0.0001, one-way ANOVA with Tukey’s multiple comparisons (B) or log-rank (Mantel-Cox) test (D). Mean ± SEM. Data shown are representative of two independent experiments.

PD-L1 blockade enhances the antitumor efficacy of neo-ISIM and improves survival.

Although induction and activation of tumor-residing cDC1s overcomes primary resistance to PD-1/PD-L1 blockade in the treated tumors (37), whether this occurs in untreated distant metastatic site remains unknown. We first evaluated PD-1 expression in CD8⁺ T cells in the lungs after neo-ISIM. We observed increased PD-1 expression in CD8⁺ T cells infiltrating into lungs as well as primary tumors of neo-ISIM treated mice compared to untreated mice (Fig. 5A). Furthermore, Tet⁺ CD8⁺ T cells exhibited higher levels of PD-1 expression compared with Tet⁻ CD8⁺ T cells in both primary tumors and lungs (Fig. 5B), suggesting that neo-ISIM-induced tumor-specific T cells might be preferentially inhibited by PD-L1 signaling.

Figure 5. PD-L1 blockade therapy enhances antitumor efficacy of neo-ISIM and improves survival.

A, B, Mice bearing orthotopic 4T1-luc tumors were treated with PBS (NT) or neo-ISIM as described in figure 1A (n = 9-10 per group). Primary tumors and lungs were harvested on day 14.

Representative flow cytometric plots and percentage of PD-1⁺ cells among CD8⁺ T cells (A) and frequency of PD-1⁺ cells in gp70-specific tetramer (Tet)⁻ or Tet⁺ CD8⁺ T cells (B) in primary tumors (left) and the lungs (right) from mice treated with PBS (NT) or neo-ISIM (A) and neo-ISIM (B) are shown. Numbers denote percent PD-1 positive cells (n = 9-10 per group).

C, Treatment protocol of neo-ISIM + anti-PD-L1 Ab (αPD-L1) or isotype Ab (rat IgG2b) for mice bearing orthotopic 4T1-luc tumors.

D, Representative flow cytometric plots showing gp70 Tet⁺ CD8⁺ T cells and the numbers (/g) of Tet⁺ cells among CD8⁺ T cells in resected 4T1-luc tumors (n = 9-10 per group). Numbers denote percent Tet⁺ cells.

E, Representative bioluminescent imaging of four 4T1-luc tumor bearing mice per group in different treatment as indicated on day 26.

F, Survival curves of mice in different treatment as indicated (n = 8-10 per group).

**p < 0.01, ***p < 0.001, two-tailed unpaired t-test (A), paired t-test (B), one-way ANOVA with Tukey’s multiple comparisons (D), or log-rank (Mantel-Cox) test (F). Data shown in A, B, E and F are representative of two independent experiments. Mean ± SEM.

Given the increase of PD-1⁺ CD8⁺ T cells and higher susceptibility to PD-L1-mediated inhibition of tumor-specific T cells in the lungs, we next sought to test whether PD-L1 blockade could augment the neo-ISIM-elicited control of metastatic lesions (Fig. 5C). In agreement with our recent studies and others (5,37,38), we did not observe changes of the frequency of tumor-specific CD8⁺ T cells (Fig. 5D) or regression of the primary tumors (Supplementary Fig. 7) by anti-PD-L1 Ab alone. In contrast, anti-PD-L1 Ab and neo-ISIM synergistically increased tumor-specific CD8⁺ T cells in the primary tumors (Fig. 5D), improved the control of metastatic tumors (Fig. 5E), and prolonged survival (Fig. 5F). Collectively, these data suggest that neo-ISIM-induced systemic antitumor efficacy could be further augmented by PD-L1 blockade.

Increasing dose and fractionation of RT, and early surgical resection after neo-ISIM improve therapeutic efficacy of neo-ISIM.

We next investigated the RT dose and fractionation, and timing of surgical resection to maximize the antitumor efficacy of neo-ISIM. We first used mice bearing bilateral AT-3 tumors (37) to determine the optimal dose and fractionation of RT in multimodal intralesional therapy causing the regression of distant non-irradiated tumors. We treated primary orthotopic left mammary tumors with in situ delivery of Flt3L, RT at various doses of irradiation (0, 3, 9, or 15 Gy) in a single fraction, and TLR3/CD40 agonists, and monitored the growth of untreated secondary tumors (Supplementary Fig. 8A). We found that control of secondary as well as primary tumor growth was RT dose dependent. We next examined whether increase of RT dose and fractionation has an influence on the growth of secondary tumors. To this end, mice received a single dose of 9 Gy or three fractions of 3 or 9 Gy in consecutive days to the primary tumors (Supplementary Fig. 8B). Growth of the secondary as well as primary tumors was markedly inhibited in mice treated with 9 Gy radiation for three consecutive days compared with mice receiving total 9 Gy radiation as either a single-dose or fractionated regimen. To probe whether these results are translated into the neo-ISIM regimen, we treated mice bearing orthotopic 4T1 tumors with in situ delivery of Flt3L, a single-dose or three fractions of 9 Gy irradiation, and TLR3/CD40 agonists (Fig. 6A). Indeed, the use of three fractions of 9 Gy irradiation in the neo-ISIM regimen markedly improved survival, demonstrating that dose and fractionation of RT in neo-ISIM are a critical determinant for the control of distant metastasis.

Figure 6. Increasing dose and fractionation of RT and early surgical resection after neo-ISIM improve therapeutic efficacy of neo-ISIM.

A-C, Mice bearing orthotopic 4T1-luc tumors were treated with PBS (NT) or neo-ISIM consisting of different RT dose and fractionation schedule (A, n = 9-10 per group), different timing of surgical resection of tumors (B, n = 5-7 per group), and in combination with anti-PD-L1 Ab (αPD-L1) or isotype Ab (rat IgG2b) (C, n = 6-7 per group). *p < 0.05, **p < 0.01, ***p < 0.001, or ****p < 0.0001, log-rank (Mantel-Cox) test. Data shown are representative of two independent experiments.

We initially developed a regimen with surgical resection performed 6 days after the completion of neo-ISIM (Fig. 1A), allowing the host enough time to develop systemic immunity. However, our previous work showed that ISIM could trigger rapid regression of primary tumors (37,38), and we found that therapeutic efficacy could be observed when primary tumors were resected one day after the completion of neo-ISIM (Supplementary Fig. 4D and 6C). To evaluate the optimal timing of surgical resection in the neo-ISIM, we treated palpable 4T1-tumor bearing mice with RT only or combination of Flt3L, RT and TLR3/CD40 agonists followed by early or late surgical resection (Fig. 6B). Although we observed in situ administration of Flt3L and TLR3/CD40 agonists enhanced immunogenicity of RT against distant metastases regardless of the timing of surgery, earlier resection of tumors substantially improved survival. Lastly, we assessed whether anti-PD-L1 therapy would augment therapeutic efficacy of neo-ISIM consisting of three fractions of 9 Gy and early surgical resection after neo-ISIM (Fig. 6C). We found that antitumor efficacy of this regimen was further enhanced by anti-PD-L1 therapy.

Discussion

Surgical resection remains the mainstay of treating cancers, but many patients develop locoregional and/or distant relapse of disease following resection with microscopically cancer-free margins. In TNBC, the relapse rate is as high as 34% in patients that underwent a potentially curative resection (10). This is related, at least, in part to the lack of systemic adaptive immunity especially in patients with immunologically cold tumors characterized by the myeloid-enriched, poorly T cell-inflamed TME. The strategy we describe here enables the host with poorly T cell-inflamed tumors to develop systemic tumor-specific T-cell immunity and immunological memory before resection. We demonstrate that neoadjuvant in situ induction and activation of cDC1s in the primary tumor generates tumor-specific T cells that effectively traffic to distant metastatic sites, controls growth of non-irradiated orthotopic metastatic tumors, improves survival, and establishes systemic immunological memory. Furthermore, we show this neo-ISIM-induced systemic antitumor immunity is dependent on cDC1s, T cells, and the presence of TdLN, and can be further potentiated by anti-PD-L1 therapy.

RT-induced immunogenic death of cancer cells can trigger antigen presentation and priming of tumor-specific CD8⁺ T cells (20,21,24); however, RT alone is still usually insufficient to overcome immunologic barriers in the TME, and abscopal effects following RT are rarely seen in patients even with immune checkpoint inhibitors (47,48). Our findings of the requirement of Batf3-dependent cells for neo-ISIM-mediated systemic antitumor efficacy is in line with a previous study showing that the abscopal effects of RT and immunotherapy are lost in Batf3^−/− mice (49). Our results further demonstrated that increasing the frequency of tumor-residing cDC1s augments immunogenicity of RT in poorly T cell-inflamed tumors.

Our results revealed the pivotal role of TdLN for the development of systemic antitumor efficacy of neo-ISIM. These findings align with a recent study showing the importance of TdLN for RT-induced abscopal effect in a preclinical model (46), and our previous work demonstrating that ISIM facilitates trafficking of antigen-loaded cDC1s to the TdLN, and generates tumor-specific T cells that exhibit an effector phenotype and cytokine-producing capacity in TdLN (37). The findings described here is particularly relevant in the management of breast cancer where sentinel LN biopsy or axillary LN dissection is performed at the time of resection. These data also suggest that intralesional therapy might be well-suited for the neoadjuvant setting when patients have intact TdLN.

Although in situ induction and activation of cDC1s augments immunogenicity of RT, and mediates regression of non-irradiated tumors in preclinical models and patients (36,37), optimal RT dose, fractionation and delivery schedule to maximize the engagement of tumor-residing cDC1s remain to be determined. We initially chose a single 9-Gy dose based on our studies and others (36-38), and for potential concerns that increasing dose and fractionation of RT might cause apoptosis of cDC1s recruited in the TME. However, we have recently reported minimal apoptosis of intratumorally injected DCs after RT (50), which prompted us to increase RT dose and fractionation of neo-ISIM. Our findings suggest that abscopal effect can be enhanced by increasing dose and fractionation of RT after in situ induction of DCs even when the TdLN is located close to the tumor within the field of RT while optimal radiotherapy regimen may differ depending on the frequency of DCs prior to RT and/or type of immunotherapy to combine with.

More work is needed to understand the mechanisms underlying the resistance to neo-ISIM and anti-PD-L1 therapy at distant metastatic lesions. This might be due to the tumor antigens that are not presented to T cells after neo-ISIM of the primary tumor as we see inter-tumor or inter-metastatic heterogeneity within the same patient with various cancers including breast cancer (51,52). Although this mechanism has not been formally investigated in the current study, we have recently shown that serial ISIM could reshape repertoires of intratumoral T cells, and overcome acquired resistance to anti-PD-L1 therapy (37). This approach, however, might be challenging in the neoadjuvant setting because it causes significant delay of surgery. To circumvent this issue, we tested different dose and fractionation schedule within one cycle of neo-ISIM. RT is known to broaden the TCR repertoire of TILs (53), and it is conceivable that increasing RT dose and fractionation may have contributed to the improved survival by reshaping the TCR repertoire of TILs. Future work is necessary to investigate the correlation of dose and fractionation of RT and TCR diversity in metastatic sites. Another potential mechanism underlying the resistance might be due to tumor cells that developed resistance after initial neo-ISIM, and subsequently metastasized to the lungs. Consistent with this scenario, we found that early surgical resection substantially improved survival compared with late surgical resection after the completion of neo-ISIM. Nevertheless, immunotherapy that expeditiously elicits a broad systemic immune response would be beneficial as neoadjuvant therapy.

Although the primary tumor needs to be accessible for injections, an in situ approach has several advantages (54,55). First, this approach does not require the need for identification of patient- and tumor-specific antigens such as neoantigens, or isolation, modification, and ex vivo expansion of patient- and tumor-specific T cells such as chimeric antigen receptor or TCR-transduced T cells for generation of antitumor T cell immunity. Second, it may elicit diverse T-cell responses against heterogeneous tumor cell populations. This broad antitumor response can be further amplified by another immunotherapy targeting polyclonal T cells such as immune checkpoint inhibitors. Third, this strategy allows for the delivery of high concentration of immunomodulatory agents in the tumor, and minimizes systemic and off-target toxicities. Fourth, if successful in the neoadjuvant setting, it elicits adaptive immunity, overcomes ignorance to solid tumors, and establishes systemic immunological memory to minimize local and distant relapse after surgical resection. Many clinical studies are underway to evaluate the safety and antitumor reactivity of intratumoral immunotherapy with a wide variety of combinations (54,55), which includes Flt3L, TLR3 and CD40 agonists (NCT03789097, NCT03788083).

Finally, given the increased frequency of tumor-specific CD8⁺ T cells in primary as well as distant non-irradiated tumors by neo-ISIM, we anticipate that this strategy could aid neoadjuvant chemotherapy, where higher pretreatment TILs correlates with increased pCR rates, and improved survival (17). Therefore, although future studies are needed to confirm this, and the scope of our studies was limited to immunotherapy, incorporation of ISIM may benefit patients with TNBC undergoing neoadjuvant chemotherapy. Overall, data presented here highlight the clinical potential of the combinatorial multimodal intralesional therapy in the neoadjuvant setting to develop systemic antitumor immune response for the treatment of highly metastatic poorly T cell-inflamed tumors such as TNBC.