$340K Collaborative Grant Funds Research on the Immune Response to Sarcomas

October 21, 2015 - The Liddy Shriver Sarcoma Initiative is pleased to announce the awarding of a $340,400 research grant for a collaborative study on the immune system and the potential role of immunotherapy in sarcomas. The grant brings together investigators from Austalia, Canada, France and the United States in a two-year study of two sarcoma subtypes.

Using medicine to help a patient's immune system fight cancer is an exciting area of research today. Despite the fact that cancer immunotherapy was named Science magazine’s 2013 advance of the year, few immunotherapy options exist for people with sarcomas. This study aims to reveal the immune system's role in osteosarcoma and undifferentiated pleomorphic sarcoma and to build a foundation for the use of immunotherapy in sarcoma treatment.

Osteosarcoma and undifferentiated pleomorphic sarcoma are aggressive cancers that tend to metastasize, threatening the lives of children and adults. These types of sarcoma may be particularly vulnerable to immunotherapies.

Dr. Neilsen explains: "Both these sarcoma types (the most common soft tissue and primary bone sarcomas) have genetically complex, mutation-filled DNA, and produce large numbers of mutant proteins. These sarcomas are therefore more likely to be recognized as foreign and consequently to attract immune responses."

The investigators aim to learn how the immune system reacts to sarcomas and how the tumors manage to evade the immune system and continue growing. Dr. Blay states that the overall goal of the research is "to understand how immune cells, white blood cells, contribute to tumor progression and can be manipulated to eradicate cancer." Based on this new knowledge, the investigators can recommend treatment strategies that should help the immune system kill the tumors.

One such strategy is to use checkpoint inhibitors, drugs that block a cell's "don't kill" signal and engage the immune system. The hope is that matching the right checkpoint inhibitor to a sarcoma subtype will enable those sarcoma patients to benefit from immunotherapy.

Dr. David Thomas says that scientists are excited about newly developed checkpoint inhibitors: "These agents 'unleash' the capacity of our own immune cells, specifically T-cells, to kill cancer cells." He will be working to discover which T-cells, among the billions that humans carry, actually kill sarcomas. This information can be used in a second immunotherapy treatment strategy: the design of vaccines that are specific to individual cancer types.

The Potential of Immunotherapy and the Importance of Collaboration

According to Dr. Thomas, the most extraordinary aspect of immunotherapies in other cancers has been the duration of the patients' responses. He says, "When patients are treated with standard chemotherapy, the cancers have a tendency to come back after treatment ends. With some patients on immunotherapy, the immune system continues to eradicate cancer in an ongoing way."

Dr. Robert Maki notes that investigators have been interested in advancing immune research on sarcomas for several years, but progress can be slow in rare diseases. This International Collaborative Grant will help to ensure that sarcoma patients are not left behind as immunotherapy science advances.

Dr. Demicco explains: "Because sarcomas are so rare, it is difficult for any one institution to have adequate case numbers for a meaningful study. This granting program encourages researchers to assemble interested and enthusiastic collaborative partners with materials and resources that may not be available at their home institutions. This sharing of ideas, materials, data, and expertise creates a synergy crucial for studies of rare diseases such as sarcoma."

The Investigators

Each investigator involved in this study has an established infrastructure of equipment, specimens, and supporting personnel for sarcoma research. In addition, each member of the research team has demonstrated a track record of productivity and accomplishment in studying sarcomas. Investigators include:

• Torsten O. Nielsen (MD, PhD, FRCPC): a professor of pathology and laboratory medicine at the University of British Columbia, a clinician-scientist and musculoskeletal consultant pathologist based at the Vancouver Coastal Health Research Institute and the British Columbia Cancer Agency.
• Jean-Yves Blay (MD, PhD): a professor of medical oncology at the University Claude Bernard Lyon I, a medical oncologist and current Director General of the Comprehensive Cancer Center of Lyon, the Centre Léon Bérard.
• Elizabeth Demicco (MD, PhD): an assistant professor of pathology at the Mount Sinai Hospital in New York City.
• Robert Maki (MD, PhD, FACP): a professor of medicine, pediatrics, and orthopaedics at Mount Sinai Hospital in New York City.
• David Thomas (FRACP, PhD): the director of the Kinghorn Cancer Centre and head of the cancer division at the Garvan Institute of Medical Research in Sydney, Australia.

The Funding

This International Collaborative Grant was co-funded by the Liddy Shriver Sarcoma Initiative ($290,400), the Alan B. Slifka Foundation ($25,000) and the Wendy Walk ($25,000). These funds, in addition to $459,600 which the investigators are focusing on this research, means that $800,000 will be used to support these studies. We hope that this research will make important and useful contributions to the understanding of the use of immunotherapeutic approaches in the treatment of undifferentiated pleomorphic sarcoma and osteosarcoma.

Copyright © 2015 Liddy Shriver Sarcoma Initiative.

ImmunoSarc: Understanding the Immune Response in Sarcomas

An ESUN Article by Torsten Nielsen, MD, PhD, FRCPC; Jean-Yves Blay, MD, PhD;
Elizabeth G. Demicco, MD, PhD; Robert Maki, MD, PhD, FACP; David Thomas, FRACP, PhD

Introduction

In the past 40 years, there has been remarkable progress in the understanding of the immune system and ways to turn it on and off to treat diagnoses as diverse as autoimmune disease and cancer. Most recently, engineered T-cells and immune checkpoint inhibitors have demonstrated activity in melanoma, non-small-cell lung cancer, and other carcinomas. Given these exciting developments in immunology and immunotherapy, we will examine sarcomas and the existing immune responses against them in greater detail than ever before.

We hypothesize that the best anti-sarcoma immune responses with checkpoint inhibitors will be seen in sarcomas that (1) express the largest number of neoantigens and (2) express immune checkpoint markers that can be inhibited by new immunomodulatory drugs. By extension, given the near absence of mutations in translocation-associated sarcomas, we further hypothesize that aneuploid sarcomas with the greatest spontaneous mutational burden and inflammatory infiltrates will be the sarcomas with the best chance for a good clinical result from immune checkpoint inhibitors.

To test our hypothesis, we will focus especially on two aneuploid sarcomas: Undifferentiated Pleomorphic Sarcoma and Osteogenic Sarcoma (Osteosarcoma). To determine if these sarcoma subtypes are genuinely good targets for immune checkpoint inhibitors we will:

1. Examine undifferentiated pleomorphic sarcoma and osteosarcoma and associated immune infiltrates for immune regulatory molecules
2. Study the T cells and tumors in an osteosarcoma mouse model system for neoantigens and the T cells that recognize them
3. Characterize the neoantigens and T cells in people with osteosarcoma and undifferentiated pleomorphic sarcoma

These studies will provide the translational science underpinning for the use of checkpoint inhibitors in the management of sarcomas. The analyses conducted here may be replicated to inform use of specific immunotherapies in other sarcoma subtypes.

Background

Despite advances in our understanding of cancer biology and the development of standardized, effective treatment with surgery and radiation, sarcomas metastasize in about half of patients. Once this happens, cures are rare, even with new generations of therapies that target the underlying molecular drivers. Next generation sequencing has revealed that cancer genomes are even more complex than expected, with many subclones within each tumor. These subclones are genetically unstable and constantly generating new mutations that are selected-for in a Darwinian fashion, leading to resistance against virtually all conventional cytotoxic, hormonal, small molecule, or antibody-based targeted therapies. Even the sarcomas that affect children and adolescents, which often start out genetically simple, can adapt in this fashion.¹

This problem is ultimately unsolvable without an anti-cancer approach that evolves as rapidly as tumors do. Given the limited arsenal of existing anti-cancer drugs and given that drug development takes decades (for diseases that progress in weeks), the practical answer is to harness the body’s immune system. The generally-accepted principle has been that (1) cancers express different self-antigens, and so they are not recognized as foreign; and/or (2) cancers have developed immune escape mechanisms well before they grow to a clinically-appreciable size. Consequently, immune stimulation has long been thought to be a futile approach.

It is clear that the mutation burden of a tumor impacts the immune response against that mutated tumor. In key experiments, tumors developing in immunocompetent mice could be grown in immunodeficient mice, but tumors generated in immunodeficient mice do not grow well in immunocompetent mice of the same background strain.² More careful analysis shows a loss of immunogenic antigens in cancers transplanted from immunodeficient to immunocompetent mice, indicating a selective pressure on the cancer: cancer immunoediting.

Given these changes in perspective, cancer immunotherapy was Science magazine’s 2013 advance of the year.⁵ Passive immunotherapy has proven remarkably successful in selected cancer subtypes, such as lymphoma. Strategies to reactivate anti-tumor immune responses with drugs targeting CTLA4 and PD1/PDL1 have led to striking advances in the therapy of melanoma, some carcinomas, and hematopoietic neoplasias.

Immunotherapy strategies for people with sarcomas⁶ are comparatively less advanced than in other cancers. Evidence from mouse models⁷ and the observation that immunosuppressed patients have three times the incidence of (non-Kaposi) sarcomas⁸ support a major role for the immune system in preventing the development and progression of these cancers. Osteogenic sarcoma represents a specific sarcoma in which immunotherapy has proven effective, such that regulatory authorities have approved muramyl tripeptide for human use in primary osteogenic sarcoma. New strategies have been developed based on immune checkpoint inhibitors and tumor antigen targeting using cancer vaccines, adoptive T-cell transfer, and chimaeric antigen receptors. Thus, there is an immediate opportunity to assess the potential for such strategies to work across the spectrum of human sarcomas.

We will focus much of the work on the antigens found within sarcomas and the immune checkpoints that appear to regulate existing immune responses against neoantigens. Immune checkpoints are crucial for maintaining self-tolerance and limiting immune responses in peripheral tissues; they are also involved in allowing immune tolerance of tumors. There is growing interest in the immunoregulatory receptor PD1 and its large family of ligands as a potential mechanism of tumor immune tolerance and escape. In the inflammatory setting, stimuli such as IFNγ may upregulate PD-L1 expression in peripheral tissues and in immune cells to suppress the immune response. Many different cancers (melanoma, ovarian cancer, colorectal cancers) up-regulate PD-L1 constitutively or in response to inflammation.

Little is known to date concerning immune checkpoint mechanisms in sarcomas. So far, one study reporting the expression of PD1/PD-L1 in osteosarcoma showed that 84% of the osteosarcoma samples expressed PD-L1, among which 24% had very high levels of PD-L1 transcript. This finding suggests that PD-L1-directed therapies could be relevant for a subset of patients with osteosarcoma.

In a similar fashion, very little is known about the potential neoantigenic repertoire of aneuploid sarcomas. There are precious few undifferentiated pleomorphic sarcoma and osteosarcoma samples in The Cancer Genome Atlas (TCGA repository), and there is no data regarding T-cell receptor usage in these tumor samples.

In the ImmunoSarc program, by gathering a group of international researchers/centers with large accrual and core research structures in place, we will fill this pre-clinical knowledge gap. This new data should, at a minimum, improve clinical trial design, efficiency, and disease prioritization for the coming wave of sarcoma immunotherapy trials.

The Investigative Team

Torsten O. Nielsen (MD, PhD, FRCPC): is a Professor of Pathology and Laboratory Medicine at the University of British Columbia, a clinician-scientist and musculoskeletal consultant pathologist based at the Vancouver Coastal Health Research Institute and the British Columbia Cancer Agency. He co-chairs the Sarcoma Disease Site Committee for the NCIC-Clinical Trials Group, and serves as associate director of the MD/PhD program for his university. Dr. Nielsen has been active in sarcoma research for the past 15 years, translating genomic profiling results into new diagnostics, predictive biomarkers, experimental therapeutics and clinical trials. He also has an active research program in breast cancer, focusing on the development and validation of biomarker tests, including for immuno-oncology.

Jean-Yves Blay (MD, PhD): is a Professor of Medical Oncology at the University Claude Bernard Lyon I, Medical Oncologist, and current Director General of the Comprehensive Cancer Center of Lyon, the Centre Léon Bérard, President of the French Sarcoma Group and of the NETwork for expert centers in SARcomas, Director of the Integrated Site for Innovative Research in Oncology in Lyon (SIRIC), Past-President of the European Organization of Research and treatment of Cancer, corresponding member of  the French Academy of Medicine. His research work in the last 15 years has been focusing on basic (molecular nosology), translational (identification of novel targets), and clinical research (new drugs development and novel strategies) on sarcoma, along with translational research on the interactions of cancer cells and immune cells in breast and ovarian carcinoma models, in particular the mechanisms of immune deactivation. This work is conducted within the Unit 17 of the Cancer Research Center of Lyon, co-led with Dr. C Caux. The ImmunoSARC project merges these two aspects of the research themes.

Elizabeth Demicco (MD, PhD): is an Assistant Professor of Pathology at the Mount Sinai Hospital in New York City. Her clinical and research focus is on mesenchymal neoplasia, with particular interest in translational studies in sarcoma. Since 2010, she has authored or co-authored over 20 papers on bone and soft tissue sarcomas.

Robert Maki (MD, PhD, FACP): is a Professor of Medicine, Pediatrics, and Orthopaedics at Mount Sinai Hospital in New York, NY, USA. He has worked in sarcoma medical oncology for over 15 years, with research experience in sarcomas for over 25 years, including a PhD in immunology examining innate and adaptive immunity in sarcomas. He has interests in sarcomas of soft tissue and bone, their biology, and in clinical trials and new drug development for this diverse group of diseases. He has mentored over 25 fellows and MD/PhD students and continues to develop collaborative clinical and translational studies.

David Thomas (FRACP, PhD): is the Director of the Kinghorn Cancer Centre and Head of Cancer Division at the Garvan Institute of Medical Research in Sydney, Australia. Dr. Thomas has a particular focus on the impact of genomics on cancer medicine, and the immunobiology of sarcomas.

Aims of the Integrated ImmunoSarc Research Program

The overall objective of the ImmunoSarc program is to interpret sarcoma biology in the context of the improved understanding of the immune response in cancer. Specifically, we will identify potential antigen targets harbored by sarcomas, as well as describe and assess the meaning of immune infiltrates observed in various sarcoma subtypes, emphasizing immune checkpoint proteins and assessing the potential role of immunotherapies in these diseases.

Within this theme, our aims are:

1. To assess and validate immune biomarkers and immune system infiltrates in specific sarcoma subtypes
2. To evaluate the immune response in genetically engineered mouse models (GEMMs) and in humans with osteosarcoma
3. To explore the evolution of sarcomas in relation to immunoediting and neoantigens

Figure 1 outlines the flow of information and materials for this international collaboration. In brief, Aim 1 will use materials from New York and Canada for biomarker screening and assay development, and materials from Canada and France for validation. This will be co-led by Demicco and Nielsen, with support from pathologists affiliated with the French group (Blay). Aim 2, co-led by Thomas and Blay, will use assay methods developed in Aim 1 for immune cell subset and checkpoint biomarkers, with pathology support from Demicco and Nielsen. Aim 3, led by Maki (with support from Thomas and Blay), will analyze French specimens using the facilities available in New York.

Aim 1. Assessment and validation of immunologic biomarkers and immune system infiltration in specific sarcoma subtypes

This will be the world’s most comprehensive survey describing immune infiltrates across sarcomas, by virtue of access to large banks of patient tumor tissues from Canada, New York, and France:

• The Genetic Pathology Evaluation Centre at the University of British Columbia has collected a series of 1,809 mesenchymal tumors, already assembled into tissue microarrays and used for the assessment of immunohistochemistry biomarkers.^9,10 The original specimens and follow-up information are available for these patients.
• Mount Sinai Hospital has access to patient tissue from 700 mesenchymal tumors. Tissue microarrays have been assembled for undifferentiated pleomorphic sarcoma, osteosarcoma, and synovial sarcoma.
• Conticabase (France) includes 14,205 mesenchymal tumors from patients treated at an expert center, with regular follow-up and updated clinical information. Tissue microarrays have been constructed for over 400 patients, focusing on undifferentiated pleomorphic sarcoma and leiomyosarcoma.

Using commercial antibodies, we will perform antibody staining on the sarcoma tissues, and establish a scoring system for each biomarker. The Conticabase tissue microarrays – of special value due to their large size and detailed linked clinical data – will be used to validate our findings.

Through this aim, we will, more thoroughly than ever before:

Aim 1a. Quantify the immune infiltrates, by simple H&E assessment. This is an inexpensive, practical method to associate immune response to sarcoma subtype, which can be applied in any laboratory. Pathologists can easily identify and count overall levels of tumor-infiltrating lymphocytes. While certain sarcomas are known to be associated with tumor-infiltrating lymphocytes, a systematic survey using consistent analytical methods has not been performed. Standard methods for tumor-infiltrating lymphocyte assessment will be applied, as disseminated recently by an international group, including Nielsen.¹²

Aim 1b. Characterize the immune infiltrates by subtypes, including lymphocytes, macrophages, dendritic cells, and mast cells. Studies in carcinoma and melanoma have shown that the composition of the immune infiltrate has biologic, prognostic, and therapeutic significance. Tumor-infiltrating lymphocytes, for example, include cytotoxic, regulatory, helper, and gamma-delta T cells, natural killer (NK) cells, and B cells that cannot be distinguished by H&E staining. These (and other immune infiltrates) are instead distinguished by the immunohistochemistry antibodies (Table 1). This work will characterize the number and type of tumor-infiltrating lymphocytes, tumor-associated macrophages, dendritic cells (the most efficient antigen-presenting cells (APCs)) and mast cells present in human sarcomas. Sarcoma types will then be ranked by total, cytotoxic, and regulatory immune cell infiltrates, as a guide for which histologies are indeed the most immunogenic. The presence and types of infiltrates will also be linked to prognosis.

Aim 1c. Describe expression of immune regulatory molecules and their ligands, on immune infiltrates and sarcoma cells. Studies in melanoma and carcinoma have identified a series of important immune cell receptors and their ligands that are critical in stimulating or dampening the immune response. New therapies that target these molecules have recently become available. The cell surface receptor and ligand proteins we are interested in, along with their existing therapies, are listed in Table 2a. Briefly, these include:

• The tumor necrosis factor receptor (TNFR) superfamily, which activate anti-tumor immune responses.^15,16
• The B7-CD28 superfamily of immune checkpoints, which include CTLA-4, PD1 and PD-L1/2, BTLA, ICOS, and B7-H3/4.
• Additional lymphocyte co-regulators, including LAG3, TIM3, NKG2/2D, and the KIR family receptors.
• Cytokines and their receptors, including IL-12 and IL-23 (secreted immunostimulatory factors), IL-6 and IFNα/β (discussed in detail in Aim 2), and CSF1R (macrophage stimulating receptor).

Major histocompatibility complex (MHC) proteins expressed on sarcoma cells (principally HLA-A/B/C and -E), which can be downregulated as an immune escape mechanism in aggressive Ewing sarcoma and osteosarcoma cases.^25.26

To date, studies of these biomarkers in sarcoma are limited to unvalidated small series covering only a few subtypes. We have critically reviewed the literature and identified protocols (antibody, conditions, scoring methodology) for most of these proteins, as detailed in Table 2b.

Aim 1d. Identify tumor-associated antigens and neoantigens considered to be immunogenic and targetable with new therapies. Some tumor types – such as synovial sarcoma and myxoid liposarcoma – generally lack immune infiltrates, implying that they may be inherently immunoevasive. For these tumors, the issue is not bypassing regulatory immune checkpoints, but rather, making the immune system recognize and react to the cancer in the first place. Interestingly, these tumor types do express CT antigens, including SSX and NY-ESO-1. We will survey sarcomas for tumor cell expression of these and other immunogenic, targetable antigens.

Aim 1e. Employ expression profiling as an alternative to immunohistochemistry, which only assesses single biomarkers in a semi-quantitative manner, with somewhat subjective interpretation. Gene expression profiles produce quantitative, robust immune signatures in a highly reproducible way. NanoString is a new technology that can investigate the expression level of a specified set of genes, often as panels of related genes. It can be applied to paraffin-embedded tissue (such as the tissue microarray source blocks), and fortunately, an immune expression panel has recently become available.²⁹ Therefore, for those sarcoma subtypes that consistently demonstrate the most immune infiltrates, this technology will be applied in a pilot study of 100 cases, to more thoroughly characterize the immune infiltrate components and active pathways.

Expected Outcomes of Aim 1

This work will yield the largest comprehensive analysis of immune infiltrates and related biomarkers in sarcomas. Patients presenting with high levels of tumor-infiltrating lymphocytes, macrophages, and NK cells are thought to have "immunologically primed" tumors, which are the most likely to respond to immune checkpoint therapies. Further, tumor types lacking such responses but expressing targetable antigens represent "immunoevasive" tumors, which may be better subjected to immune stimulation strategies.

Aim 2. Evaluation of the immune response in genetically modified mouse models (GEMM) and humans with bone sarcomas.

While Aim 1 will provide a description of potential avenues for successful application of immunotherapy in sarcoma, an experimental model is needed to better understand the mechanisms by which sarcomas escape immunosurveillance and how these mechanisms may be overcome with appropriate immunotherapy. One such model has been developed for osteosarcoma by Dr. Thomas’ group.

Model System

As radiation is a known risk factor for the development of osteosarcoma,³⁰ we use a radiocarcinogen-induced osteosarcoma mouse model (Figure 2). To induce osteosarcoma, mice are injected with radioactive calcium, which generates osteosarcoma in approximately 100% of mice,^31,32 with latency around 8 months. These mice are immunocompetent, and are thus an ideal model in which to study the anti-tumor immune response. Moreover, the use of radiation to induce tumors mimics the complex genomes of human tumors.

Using Rb+/- mice as the genetic background for 45Ca-induced osteosarcoma, we have shown that a number of cytokines – including IL-6 – are important for maintaining RB1 tumor suppressor function. IL-6-deficient mice are predisposed to developing radiation-induced osteosarcomas, suggesting that IL-6 also functions as a tumor suppressor.³⁵

Behavior of individual tumors was further studied using an allograft model of tumors derived from the radiation-induced osteosarcomas. Tissue from tumors grown in wild-type and in IL-6-deficient (IL-6-/-) mice were transplanted to both wild-type and IL-6-deficient mice. Of note, tumors from IL-6-deficient tumors showed reduced growth when implanted in wild-type mice (Figure 3), demonstrating that IL-6 impacts the anti-tumor immune response.

Critical to the studies proposed, the IL-6-deficient osteosarcoma allografts ultimately develop resistance to the immune suppression within wild-type hosts and begin to grow out (Figure 9). Understanding the mechanisms of this phenomenon is the primary goal of this aim.

Aim 2a: Interrogate how osteosarcoma evades IL-6-mediated tumor suppression in vivo: Our preliminary studies show that the immune microenvironment puts pressure on osteosarcomas to restrict tumor growth. In response, some tumors evolve – through as yet poorly-understood mechanisms – to escape this immunosurveillance. These mechanisms may be either genetic or epigenetic. We propose to take advantage of syngeneic models to determine whether there are consistent differences between osteosarcomas that occur in the presence or in the absence of IL-6.

Analyses of the osteosarcoma immune microenvironment and immune checkpoint expression will be performed by immunohistochemistry. Building on knowledge developed in Aim 1, we will examine the subtypes of immune infiltrates in our mouse orthotopic osteosarcoma model. We will also evaluate T-cell activation, cytokine production (in particular IL-6, IL-23, and IL-12), and PD1/PD-L1 expression. These results will be compared to findings about the immune microenvironment in a rat osteosarcoma model and in human osteosarcoma samples (representing a wide range of clinical scenarios: primary vs. metastasis, diagnosis vs. post-treatment, primary vs. relapse).

We will next compare the whole genomes between primary osteosarcomas derived from wild-type and IL-6-deficient mice (8 each), to identify recurrent genotypic differences between tumors evolving in a more immune permissive environment versus those evolving under IL-6 tumor suppression. To compensate for tumor heterogeneity, a parallel study will be run, using at least 2 IL-6-deficient osteosarcoma cell lines, which will be implanted into wild-type and IL-6-deficient mice (see Figure 8). Tumors will be allowed to grow until spontaneous resistance emerges.  We hypothesize that at least some of the strategies for this “escape” from IL-6-mediated immunosurveillance will be identifiable using genome sequencing, and that some of the same genetic adaptations will be observed between tumors implanted in different mice. In addition, we will compare genome differences between parental cell lines and those clones that escape immunosurveillance. This strategy will provide method to distinguish passenger from driver mutations. Mouse whole genome sequencing will be performed to identify recurrent mutations. Specific attention will be paid to genes involved in cell-host signaling and on immune pathways.

Aim 2b: Identify how immune checkpoints are involved in osteosarcoma response to conventional cytotoxic chemotherapy and PI3K/Akt/mTOR-targeted therapy: Prior studies from Dr. Blay’s group showed that, in both human tumors and in a rat model, chondrosarcoma cells expressed significant PD-L1. In the rat model, therapeutic agents against the PI3K/Akt/mTOR pathway caused a decrease in PD-L1 expression. This was confirmed in vitro in different sarcoma cell lines, and will be further investigated in a study exploring everolimus treatment prior to definitive surgery.

We will extend these observations to osteosarcoma, using our in vivo mouse and rat models to investigate the impact of PD1/PD-L1 blockade on:
   1. Tumor progression, both local invasion and metastatic spread.
   2. Osteosarcoma response to conventional chemotherapy (doxorubicin and ifosfamide).
   3. Osteosarcoma response to targeted therapy against the PI3K/Akt/mTOR pathway.

In addition to tumor response, we will evaluate local and metastatic tumors treated by these various combination therapies for expression of PD1/PD-L1, alterations in tumor-infiltrating lymphocytes, and cytokine production.

If relevant, in view of the previous observations, osteosarcoma cells overexpressing PD-L1 will be generated by lentiviral transfection. After analyzing the effects of PD-L1 overexpression in vitro, the effect of PD-L1 transfection on osteosarcoma progression will be analyzed in vivo, as well as its impact on immune cell infiltrates and cytokine production.

Expected Outcomes of Aim 2

The studies in Aim 2 will establish how the osteosarcoma genome is sculpted by immune pressures in vivo, and they will specifically shed light on pathways affecting the fundamental mechanisms of immune escape. This is valuable for understanding the role of the immune system (cytokines, immune checkpoints) in eliminating or suppressing osteosarcoma, and how this relates to tumor growth and progression in an immunocompetent, syngeneic model. Combining immunotherapy with conventional chemotherapy may be rationally proposed for further pre-clinical experiments, with the aim of expanding this to the clinical setting for the development of powerful targeted therapy.

Aim 3. Investigation of immune evolution in the natural history of sarcomas

One of the ways that cancer cells are thought to become recognizable to the immune system as "foreign" (and therefore in need of destruction) is the creation of abnormal peptides (neoantigens) by driver or passenger mutations within the cancer.⁴¹ This is believed to happen more frequently in aneuploid tumors such as leiomyosarcoma and undifferentiated pleomorphic sarcoma, which have much greater mutational burdens than that seen in translocation-associated sarcomas.

The most important place to assess immune responses to tumor neoantigens is in the tumor itself.⁴² The presence of high levels of tumor-infiltrating lymphocytes in several types of sarcoma has been previously described and will be confirmed and characterized in Aim 1. Another way to better-understand the immune response to a particular tumor is to sequence the T-cell receptors (TcR) present on tumor-infiltrating lymphocytes and to identify if any are particularly common, reflecting recognition of tumor neoantigens. These clones can then be detected at a much lower level in the peripheral circulation. Information about the T-cell repertoire can be useful to guide therapy in patients treated with immune checkpoint inhibitors. In these cases, pre-treatment tumor samples alone would be enough to identify the T-cell repertoire that should be explored in the peripheral circulation.

Here, we will assess the interplay between tumor antigenicity and T-cell infiltrates in human sarcoma samples during disease progression, focusing on the most common aneuploid sarcomas, which also tend to be the most likely to metastasize: undifferentiated pleomorphic sarcoma, leiomyosarcoma, and osteosarcoma. Tumor samples (50 paired primary sarcomas and their metastases), will be obtained from the Conticabase tumor bank.

Aim 3a. To compare immune infiltrates, antigen expression, and immune checkpoint markers between paired primary tumors and metastases, using methods developed in Aim 1. Data will be tied to clinical aspects of the disease, relapse, metastasis, and sensitivity to treatment. 

To better understand neoantigen expression in these sarcomas, we will then assess a subset of tumors (paired primary and metastasis) by whole exome sequencing. Once the mutations are determined, potential neoantigen generation will be assessed via virtual translation of the sequences, focusing in particular on missense mutations. These mutated protein sequences will be used to create theoretical peptides, which will be analyzed to identify those most likely to bind strongly to MHC-I molecules (which would enable the immune system to recognize them as “foreign”). This will allow us to determine the antigen diversity among individual sarcomas, and to generate a neoantigen catalog for each tumor class. We will also compare the mutation profiles of primary and metastatic tumors to determine if there is a difference in neoantigen presence in metastatic disease.

Aim 3b. To characterize the spectrum of T-cell receptors within sarcomas, for primary tumors, metastases, and, where available, peripheral blood obtained at initial diagnosis and after disease relapse. We will use data from tumor genome sequencing to confirm anti-tumor responses to sarcoma neoantigens by tumor-infiltrating lymphocytes.

Comparisons will include:

1. TcR repertoire in pre-treatment vs. treated tumor samples
2. TcR repertoire in primary vs. metastatic disease
3. TcR repertoire in tumor samples vs. peripheral blood mononuclear cells
4. TcR repertoire in patients who have received immune checkpoint inhibitors

This data will allow us to determine the relative frequency of specific T-cell clones in the tumor and in the periphery, which may allow us to utilize TcR testing prospectively in future studies, if common antigens and TcR are identified.

Aim 3c. To understand the effects of pembrolizumab in advanced sarcomas. Pre-treatment tumor samples and post-treatment buffy coat mononuclear cells collected in the course of the SARC028 trial – an open label, multi-centre, single stage, phase II clinical trial of pembrolizumab in highly mutated sarcomas, including undifferentiated pleomorphic sarcoma and osteosarcoma — will undergo T-cell analyses, as above. Data will be tied to tumor response.  In addition, the most interesting of the validated biomarkers from Aim 1 will be assessed in sarcoma tissues from the SARC028 trial.

Our approach focuses on the role of immune recognition of antigens presented by MHC class I molecules only, and thus, it is not capable of describing the entirety of the complex immune response elicited by sarcomas. Nonetheless, we believe that the described approach will interrogate the most relevant part of the anti-tumor immune responses and yield important new insights into the immune response to sarcomas.

Expected outcomes of Aim 3

This work will help to delineate the evolution of immune cell infiltrates in both tissue and peripheral blood over the course of sarcoma progression from local to metastatic disease. We expect to identify immune infiltrates in primary tumors associated with higher (or lower) risk of progression, independent of classical predictors for relapse.  Moreover, we will describe in greater detail the differences between primary and metastatic sarcomas, focusing on osteosarcoma and undifferentiated pleomorphic sarcoma.

For specific sarcoma subtypes, we will be able to specify TcR usage and to link this to tumor neoantigens as possible pairs necessary/sufficient for an anti-sarcoma immune response. We will also determine the relative frequency of specific TcR in the tumor vs. peripheral circulation, to determine if it is realistic to use peripheral blood to assess immune responses within a tumor.

This will serve as a proof-of-principle for future studies in patients, when it becomes possible to design immune responses against patient-specific neoantigens. Members of our group include translational clinical trial oncologists, who would be perfectly positioned to act quickly on such information.

Clinical Translation

All proposed work contributes to a common objective of identifying the immune cell populations and antigens that are most important to expected responses in patients with metastatic sarcoma, treated with immune checkpoint inhibitors. The results of Aims 1-3 will define methods that can be applied to recently-opened sarcoma immuno-oncology clinical trials (such as SARC028). Our work will also identify which sarcoma types are most likely to benefit from immunomodulatory versus immunostimulatory strategies and suggest the best combinations of immunotherapy with other available treatment strategies. Team members, with executive roles in SARC, EORTC, NCIC-CTG, and the Australasian Sarcoma Study Group, will be able use the results to directly influence the design of sarcoma immuno-oncology trials.

Copyright © 2015 Liddy Shriver Sarcoma Initiative.