Undifferentiated pleomorphic sarcoma (UPS) is one of the most common and aggressive adult soft tissue sarcomas (STS). Once metastatic, UPS is rapidly fatal. Most STS, including UPS, are resistant to conventional immunotherapies as these tumours have low numbers of spontaneous tumour infiltrating lymphocytes (TILs) and are densely populated with immune suppressive macrophages. Intra-tumoural activation of the STimulator of INterferon Genes (STING) pathway is a novel immunotherapeutic strategy to recruit anti-tumour TILs into the tumour microenvironment. In a murine model of UPS, we have demonstrated that intra-tumoural injection of a murine-specific STING agonist, DMXAA, results in profound immune mediated tumour clearance. Recently, molecules capable of activating both human and mouse STING pathways have been developed. In pursuit of clinically relevant therapeutic opportunities, the purpose of this study is to evaluate the anti-tumour potential of two agonists of the human and murine STING receptors: ADU-S100 and MSA-2 as monotherapies and in combination with the immune checkpoint inhibitor, anti-PD1 in a murine model of UPS.

Immune competent mice were engrafted with murine UPS cells in the hindlimb muscle. Once palpable, mice in the monotherapy group were treated with a single intra-tumoural dose of 1) ADU-S100 or 2) MSA-2 or 3) DMXAA. In additional experimental groups, mice were treated with the different STING agonists and monoclonal anti-PD1. Tumour volume measurements and tumour bioluminescence were measured over time. To quantify dynamic changes in immune populations and in the tumour immune microenvironment, STING treated UPS tumours were evaluated using flow cytometry and mRNA quantification at various timepoints after therapy.

DMXAA monotherapy produced complete tumour eradication in 50% of mice, whereas both ADU-S100 or MSA-2 monotherapy only extended survival but did not result in complete tumour clearance. Flow cytometry and transcriptional profiling of tumours at multiple timepoints post-treatment showed similar inflammatory changes and increased TILs numbers across all STING agonists. The addition of anti-PD1 treatment to STING therapy significantly extended survival times with both ADU-S100 and MSA-2, and resulted in 14% complete tumour clearance with ADU-S100. No complete survivors were observed with MSA-2-anti-PD1 combinations therapy.

STING activation is a promising immunotherapeutic strategy for UPS. Recently developed human STING agonists are not as effective as DMXAA despite similar immunologic responses to treatment. STING and anti-PD-1 treatment were therapeutically synergistic for both human STING agonists. These results justify further research around STING activation as a therapeutic modality for STS. DMXAA may possess additional off-target therapeutic properties beyond STING activation which warrants further investigation. Elucidating these differences may be critical to further optimize STING therapy for human STS.

Soft tissue sarcomas (STS) are rare, aggressive malignancies derived from connective tissues such as muscle and fat. Undifferentiated pleomorphic sarcoma (UPS) is one of the most common STS in adults. UPS is an aggressive, highly metastatic sarcoma, and is resistant to chemotherapy. New therapies for UPS are desperately needed. STS have an immune desert tumour immune microenvironment (TIME), characterized by a paucity of tumour infiltrating lymphocytes and subsequent resistance to immunotherapies such as immune checkpoint inhibitors. Strategies capable of creating an immune-rich, inflamed TIME may improve immunotherapy efficacies for sarcoma. Activation of the STING (stimulator of interferon genes) receptor can induce potent innate and adaptive immune responses within immunogenic solid tumours. However, this approach has never been attempted in immune-inert sarcomas.

Purpose: To determine the therapeutic anti-tumour effects of STING activation in UPS tumours.

We have developed an inducible, immune-competent mouse model of UPS. We evaluated intra-tumoural injection of the murine STING receptor agonist, DMXAA, into UPS-bearing immune-competent mice. DMXAA was injected into palpable UPS tumours of the hindlimb. Tumour volume and bioluminescence imaging was recorded bi-weekly. DMXAA treated UPS tumours were also evaluated for necrosis and immune infiltration at defined time points.

UPS tumours developed necrosis and lymphocytic infiltration 72 hours after DMXAA treatment. A single intra-tumoural dose of DMXAA into UPS tumours resulted in durable cure in 50% of mice. All survivors rejected a re-challenge of the UPS tumours in both the contralateral hindlimb and lung, suggesting adaptive immunity. The therapeutic effects of DMXAA were mitigated in lymphocyte deficient Rag2 knockout mice.

STING therapy is a promising immunotherapeutic opportunity for immune-inert sarcomas. Our data warrants further preclinical investigations in other sarcoma models and in combination with other immune-based therapies.

Impaired bone healing biology secondary to soft tissue deficits and chemotherapy contribute to non-union, fracture and infection following limb salvage surgery in Osteosarcoma patients. Approved bone healing augments such as recombinant human bone morphogenetic protein-2 (rhBMP-2) have great potential to mitigate these complications. rhBMP-2 use in sarcoma surgery is limited, however, due to concerns of pro-oncogenic signalling within the tumour resection bed. To the contrary, recent pre-clinical studies demonstrate that BMP-2 may induce Osteosarcoma differentiation and limit tumour growth. Further pre-clinical studies evaluating the oncologic influences of BMP-2 in Osteosarcoma are needed. The purpose of this study is to evaluate how BMP-2 signalling affects Osteosarcoma cell proliferation and metastasis in an active tumour bed.

Two Osteosarcoma cell lines (143b and SaOS-2) were assessed for proliferative capacity and invasion. 143b and SaOS-2 cells were engineered to upregulate BMP-2. In vitro proliferation was assessed using a cell viability assay, motility was assessed with a scratch wound healing assay, and degree of osteoblastic differentiation was assessed using qRT-PCR of Osteoblastic markers (CTGF, ALP, Runx-2 and Osx). For in vivo evaluation, Osteosarcoma cells were injected into the intramedullary proximal tibia of immunocompromised (NOD-SCID) mice and local tumour growth and metastases were assessed using weekly bioluminescence imaging (BLI) and tumour volume measurements for 4–6 weeks. At the experimental end point we assessed radiographic tumour burden using ex-vivo micro-CT, as well as tibial and pulmonary gross and histologic pathology.

SaOS-2 was more differentiated than 143b, with increased expression of Runx-2 (p = 0.009), Osx (p = 0.004) and ALP (p = 0.035). BMP-2 upregulation did not stimulate an osteoblast differentiation response in 143b, but stimulated an increase in Osx expression in SaOS-2 (p = 0.002). BMP-2 upregulation in 143b cells resulted in increased proliferation in vitro (p = 0.014), faster in vitro wound healing (p = 0.03), significantly increased tumour volume (p = 0.001) with enhanced osteolysis detected on micro-CT, but did not affect rates of lung metastasis (67% vs. 71%, BMP-2 vs. Control). BMP-2 over-expression in SaOS-2 cells reduced in vitro proliferation when grown in partial osteogenic-differentiation media (p < 0.001), had no effect on in vitro wound healing (p = 0.28), reduced in vivo SaOS-2 tumour burden at 6 weeks (photon counts, p < 0.0001), decreased tumour-associated matrix deposition as assessed by trabecular thickness (p = 0.02), and did not affect rates of lung metastasis (0% vs. 0%).

Our results indicate BMP-2 signalling incites a proliferative effect on a poorly differentiated Osteosarcoma cell line (143b), but conditionally reduces proliferative capacity and induces a partial differentiation response in a moderately-differentiated Osteosarcoma cell line (SaOS-2). This dichotomous effect may be due to the inherent ability for Osteosarcoma cells to undergo BMP-2 mediated terminal differentiation. Importantly, these results do not support the clinical application of BMP-2 in Osteosarcoma limb salvage surgery due to the potential for stimulating growth of poorly differentiated Osteosarcoma cells within the tumour bed. Additional studies assessing the effects of BMP-2 in an immune-competent mouse model are ongoing.

Impaired bone healing biology secondary to soft tissue deficits and chemotherapy contribute to non-union, fracture and infection following limb salvage surgery in Osteosarcoma patients. Approved bone healing augments such as recombinant human bone morphogenetic protein-2 (rhBMP-2) have great potential to mitigate these complications. rhBMP-2 use in sarcoma surgery is limited, however, due to concerns of pro-oncogenic signalling within the tumour resection bed. To the contrary, recent pre-clinical studies demonstrate that BMP-2 may induce Osteosarcoma differentiation and limit tumour growth. Further pre-clinical studies evaluating the oncologic influences of BMP-2 in Osteosarcoma are needed. The purpose of this study is to evaluate how BMP-2 signalling affects Osteosarcoma cell proliferation and metastasis in an active tumour bed.

Two Osteosarcoma cell lines (143b and SaOS-2) were assessed for proliferative capacity and invasion. 143b and SaOS-2 cells were engineered to upregulate BMP-2. In vitro proliferation was assessed using a cell viability assay, motility was assessed with a scratch wound healing assay, and degree of osteoblastic differentiation was assessed using qRT-PCR of Osteoblastic markers (CTGF, ALP, Runx-2 and Osx). For in vivo evaluation, Osteosarcoma cells were injected into the intramedullary proximal tibia of immunocompromised (NOD-SCID) mice and local tumour growth and metastases were assessed using weekly bioluminescence imaging and tumour volume measurements for 4–6 weeks. At the experimental end point we assessed radiographic tumour burden using ex-vivo micro-CT, as well as tibial and pulmonary gross and histologic pathology.

SaOS-2 was more differentiated than 143b, with significantly increased expression of the Osteoblast markers Osx (p = 0.004) and ALP (p = 0.035). BMP-2 upregulation did not stimulate an osteoblast differentiation response in 143b, but stimulated an increase in Osx expression in SaOS-2 (p = 0.002). BMP-2 upregulation in 143b cells resulted in increased proliferation in vitro (p = 0.014), faster in vitro wound healing (p = 0.03), significantly increased tumour volume (p = 0.001) with enhanced osteolysis detected on micro-CT, but did not affect rates of lung metastasis (67% vs. 71%, BMP-2 vs. Control). BMP-2 over-expression in SaOS-2 cells reduced in vitro proliferation when grown in osteogenic-differentiation media (p < 0.001), had no effect on in vitro wound healing (p = 0.28), reduced in vivo SaOS-2 tumour burden at 6 weeks (photon counts, p < 0.0001), decreased tumour-associated matrix deposition as assessed by trabecular thickness (p = 0.02), but did not affect rates of lung metastasis (0% vs. 0%).

Our results indicate BMP-2 signalling incites a proliferative effect on a poorly differentiated Osteosarcoma cell line (143b), but conditionally reduces proliferative capacity and induces a partial differentiation response in a moderately-differentiated Osteosarcoma cell line (SaOS-2). This dichotomous effect may be due to the inherent ability for Osteosarcoma cells to undergo BMP-2 mediated terminal differentiation. Importantly, these results do not support the clinical application of BMP-2 in Osteosarcoma limb salvage surgery due to the potential for stimulating growth of poorly differentiated Osteosarcoma cells within the tumour bed. Additional studies assessing the effects of BMP-2 in an immune-competent mouse model are ongoing.

Soft tissue sarcomas (STS) have not demonstrated favourable clinical responses to emerging immunotherapies such as checkpoint inhibitors. Studies in carcinomas and melanoma have demonstrated that tumours lacking T-cell infiltrates are associated with poor responses to immunotherapies. It is postulated that STS lack tumour asscoiated lymphocytes which renders these tumours insensitive to checkpoint inhibitors. Our objective was to develop a novel syngeneic mouse model of STS and characterize the immune phenotype of these tumours. Additionally, we sought to evaluate the therapeutic responses of these sarcomas to checkpoint inhibitors and a Type I interferon agonist.

K-ras mutagenesis and p53 deletion was induced using a Lenti-Cre-recombinase injection into the hindlimb of 3 week old C57BL/6 mice. Tumours were harvested and characterized using standard histopathology techniques and whole trascriptome sequencing (RNAseq). Full body necrospy and histopathology was performed to identify metastases. Flow cytometry and immunohistochemistry was used to evaluate tumour immune phenotypes. Tumours were implanted into syngeneic C57BL/6 mice and the therapeutic responses to anti-CTLA4, anti-PD1 and DMXAA (Type I interferon agonist) were performed. Tumour responses were evaluated using bioluminescent imaging and caliper measurements.

Soft tissue sarcomas developed in mice within 2–3 months of Lenti-Cre injection with 90% penetrance. Histologic analyses of tumours was consistent with a high-grade myogenic sarcoma characterized by smooth muscle actin, Desmin and Myogenin D positive immunostaining. Using crossplatform normalization protocols, geneexpression signatures of the mouse tumours most closely correlated with human undifferentiated pleomorphic sarcoma (UPS). Collectively, gene expression signatures of this murine sarcoma correlated with all muscle-derived human sarcomas (ERMS, ARMS, Synovial sarcoma, UPS). No lung or other visceral metastases were observed in all mice who developed spontaneous tumours. Immune phenotyping demonstrated a paucity of tumour-infiltrating lymphocytes (TILs, (TAMs). 50% of identified TILs in these murine sarcomas expressed PD-1, yet tumours were not responsive to anti-PD1 therapy or anti-CTLA4 therapy. A single intra tumoural (i.t.) injection of the Type I interferon agonist, DMXAA resulted in 80–90% tumour necrosis 72 hrs post-injection, decreased tumour viability up to 2 weeks post-injection and a marked infiltration of CD8+ T-cells and anitgen presenting dendritic cells and macrophages. Additional longitudinal experiments demonstrate a sustained and progressive anti-tumour effect in 83% (5/6) mice up to 6weeks following a single i.t. injection of DMXAA. All control treated mice (6/6) reached humane endpoint within 14 days. At 3 months post-DMXAA treatment, 4/6 mice were free of disease. We re-injected UPS tumours into these mice and tumours did not grow, suggesting abscopal effects after DMXAA treatment of primary tumours.

We have characterized a new orthotopic and syngeneic mouse model of a myogenic soft tissue sarcoma. Like most human STS sub-types, these tumours have an immune inert tumour microenvironment and are not sensitive to checkpoint inhibitors. This model, syngeneic to C56BL/6 mice will enable future opportunities to investigate how various branches of the immune system can be targetted or manipulated to unearth new immunotherapeutic strategies for sarcoma. Using this model we have demonstrated that a single, intra-tumoural injection of a Type I interferon agonist can result in anti-tumour effects, recruit cytotoxic lymphocytes and antigen presenting cells with into the the tumour microenvironment. Abscopal tumour rejection after DMXAA treatement suggest adaptive T-cell responses against UPS are active in this model. Future work is needed to determine if upregulation of Type I inferferon pathways can be used as a therapeutic strategy for sarcoma or as a sensitization strategy for checkpoint inhibitors.