Report on a Study of Synovial Sarcoma

By Torsten O. Nielsen, MD, PhD; Marc Ladanyi, MD;
Scott W. Lowe, PhD; and Diederik de Bruijn, PhD

Abstract

This International Collaborative Grant was created to develop a coordinated research program in synovial sarcoma by taking advantage of the resources and research capacity of an international team of clinician investigators and scientists. The original aims of this International Collaborative Grant as presented in the companion Experimental Plan were:

1. To use shRNA screening technology to interrogate epigenetic modifiers of gene expression that are likely to include critical partners contributing to SS18-SSX oncogenesis.
2. To identify cofactors responsible for creating a permissive state for SS18-SSX oncogenesis, which might be exploited as therapeutic targets.
3. To develop Zebrafish models of synovial sarcoma for validation and preclinical therapeutic translational studies.

The synovial sarcoma International Collaborative Grant team has been working together for more than two years. The team is comprised of researchers from the Memorial Sloan Kettering Cancer Center in New York led by Drs. Marc Ladanyi and Scott Lowe, from the University of British Columbia in Vancouver led by Dr. Torsten Nielsen, and from Radboud University Medical Center in Nijmegen, Netherlands, led by Dr. Diederik de Bruijn.

This team met face-to-face in New York (August 2014), Philadelphia (April 2015), Salt Lake City (November 2015) and will meet in Lisbon (November 2016), and over the course of the ICG granting period has recruited additional expertise in the field through collaborations with Dr. Kevin Jones (Huntsman Cancer Institute, Salt Lake City) and Dr. Bertha Brodin (Karolinska Institutet, Stockholm). The research program has been able to achieve goal one, make significant progress on goal two after initial studies showed technical problems, and shifted goal three toward a more feasible murine system, while also branching into new research directions as novel findings and opportunities emerged.

Results

1. To use shRNA screening technology to interrogate epigenetic modifiers of gene expression that are likely to include critical partners contributing to SS18-SSX oncogenesis

RNA interference (RNAi) screening is a state-of-the-art technology that enables the dissection of biological processes and identification of potential therapeutic targets. To identify epigenetic vulnerabilities that are specifically generated by the SS18-SSX fusion protein we screened two cell lines using an shRNA library targeting epigenetic modulators (2400 shRNAs): a synovial sarcoma cell line (derived from a mouse model in which the human SS18-SSX2 fusion oncogene is expressed in the myogenic progenitor compartment) and C2C12 mouse myoblasts as an untransformed control. Using the scoring criterion of more than two-fold mean depletion in three independent replicates, a total of 119 shRNAs were scored in the synovial sarcoma cell line. These included three shRNAs targeting the human SS18-SSX as well as other positive controls (shRpa3 and shBcl2). Among the 119 shRNAs depleted in the sarcoma cell line, three independent shRNA targeting a histone demethylase specifically scored in sarcoma cells (Banito et al., presented at CTOS 2015 Salt Lake City, and manuscript in preparation). These results were validated in in the mouse synovial sarcoma cell line as well as in three human synovial sarcoma cell lines (HS-SY-II, YaFUSS, FUJI). Besides affecting proliferation in vitro, inhibition of this target repressed tumor growth in vivo in human synovial sarcoma xenografts. In agreement for a role for this protein in this disease, its levels are higher in a panel of 57 human synovial sarcoma samples when compared with benign soft tissue tumors. Furthermore, we used CRISPR/Cas9 technology to validate our results and find specific protein domains in our candidate that could be targeted for synovial sarcoma therapy (Banito et al, submitted to CTOS 2016 Lisbon, and manuscript in preparation). In addition to finding potential therapeutic targets, we used CRISPR/Cas9 technology to endogenously tag SS18-SSX1 in human cell lines (Figure 1). This tool will allow the study of additional molecular interactions and SS18-SSX occupancy genome-wide.

Significance: Our approach allowed the identification of specific epigenetic dependencies, uncovered a new potential drug target in synovial sarcoma and shed light into the biological mechanisms underlying this disease. Additionally, we created new and powerful tools to better understand SS18-SSX biological function, which may provide alternative strategies for the design of targeted therapies to treat this disease.

2. To identify co-factors responsible for creating a permissive state for SS18-SSX oncogenesis, which might be exploited as therapeutic targets

2.1 Proximity ligation assay development for targeted therapy investigation

The proximity ligation assay is a novel technique used to assess direct protein-protein interactions in their cellular environment. It was utilized to assess in situ (i.e. inside a living cell) interactions of SS18-SSX with its cofactor TLE-1 in human synovial sarcoma cell lines. A fluorescent marker attached to an antibody linker allowed us to visualize protein-protein interactions. Fluorescent microscopy was used to measure the number of SS18-SSX/TLE-1 proximity interactions per nucleus, with and without experimental treatments. Having established this method, not used in synovial sarcoma before, a small-molecule inhibitor library of 16 000 candidate drug compounds was screened in synovial sarcoma cell lines. The top hits, based first on their ability to stop the growth of synovial sarcoma cells, were assessed in a novel high-throughput proximity ligation assay format to see if the drug was breaking up the SS18-SSX/TLE-1 interaction, which would be a sign that it is specifically reversing the main oncogenic mechanism driving these cancers. 

SS18-SSX/TLE-1 interactions were detectable by proximity ligation assay, exclusively in human synovial sarcoma cell lines (this abnormal interaction being unique to this form of cancer). A group of drugs known as Class 1 HDAC inhibitors were found that specifically disrupts the SS18-SSX/TLE-1 interaction in situ, leading to synovial sarcoma cell death. The assay was successfully scaled up into a high-throughput screening format to look at the full library in a single experiment. A novel small molecule inhibitor capable of disrupting the interaction and reactivating normal gene transcription was uncovered from a 16 000 compound library.

Significance: The proximity ligation assay can be utilized to measure the relative interaction of the SS18-SSX/TLE-1 complex in situ in synovial sarcoma (Figure 2). This interaction can be specifically disrupted by class I HDAC inhibition as well as by a novel drug from the compound library. The assay we developed can be (and is now being) applied in a high-throughput format for use in drug discovery platforms for other translocation-associated sarcomas with known effector partners. A paper was published on this discovery by Laporte et al. in Oncotarget (2016), a publicly-accessible, open access journal.

2.2 Drug screening to uncover novel therapeutic targets

In parallel work, a different high-throughput drug screen was undertaken in a panel of synovial sarcoma cell lines to uncover novel sensitizing agents and targetable pathways. In this case, we used a 900 compound library comprised of known anticancer drugs which are ready for clinical use to screen a panel of fusion oncogene-confirmed synovial sarcoma cell line models. Compound efficacy was evaluated based on potency against synovial sarcoma cell lines and specificity for targeting this type of cancer but not other cancers or nonmalignant cells. To identify particularly effective anticancer strategies, we also tested combinations of top agents for potentially synergistic activity. Effective compounds were validated in specific assays including the proximity ligation assay described above, apoptotic induction assays, mechanism of synergy studies, and in vivo drug treatment studies in mouse models of synovial sarcoma. The cellular effects of HDAC inhibition were further assessed by transcriptome analysis and supporting functional studies.

We found that the top hits in this unbiased screen of anticancer drugs were HDAC and proteasome inhibitors. Alone, but especially in combination, these agents induce cytosolic stress, increase reactive oxygen species and activate pro-apoptotic factors, leading to synovial sarcoma cell death even at very low doses. The combination of HDAC inhibitors and proteasome inhibitors significantly inhibits tumor growth in a mouse model of synovial sarcoma. Mechanistically, HDAC inhibition prevents aggresome formation, reactivates expression of tumor suppressors, downregulates the mitotic cell cycle phenotype and initiates apoptotic marker expression.

Significance: HDAC inhibition changes the phenotype of synovial sarcoma cells from growth to death (via apoptosis), and this effect is significantly enhanced by combining with proteasome inhibitors. As others have already successfully tested the combination of these two types of anticancer drugs in multiple myeloma, toxicities are known and this combination is eminently testable in human sarcoma clinical trials, which we are working to organize.

2.3 Synovial sarcoma cell line analysis

In order to investigate a potential role for secondary mutations in synovial sarcoma oncogenesis and better understand the genetic context of the cell lines used in our screens and functional studies, seven synovial sarcoma cell lines were sequenced using the MSK-IMPACT clinical next generation sequencing assay, consisting of a 341 gene panel which covers most known cancer mutations (especially those that can be targeted by existing drugs). Very few additional mutations beyond the known SS18-SSX fusions were detected, none of which were consistent among the cell lines (Table 1), indicating no evidence for a requirement of a second mutation.

Significance: This result suggests that fusion oncoprotein SS18-SSX is necessary and sufficient for synovial sarcoma oncogenesis, as there are no other consistent additional mutations seen.

Table 1. Synovial Sarcoma Cell Lines Sequences for Additional Mutations

Cell line	SS type	Fusion Type	MSK-IMPACT 341 gene panel Number of secondary mutations
SYO-1	Biphasic	SYT-SSX2	1
FUJI	Monophasic	SYT-SSX2	2
Yamato-SS	Biphasic	SYT-SSX1	8
MoJo	Monophasic	SYT-SSX1	1
HS-SY-II	Monophasic	SYT-SSX1	1
YaFUSS	Monophasic	SYT-SSX1	10
ASKA-SS	Biphasic	SYT-SSX1	0

2.4 Inducible model development for genetic screening studies

A cell line with an inducible version of the SS18-SSX oncoprotein, known as HEK-SS18/SSX, was used to study long term effects of oncoprotein expression on cell biology. During this experiment, the growth speed and cell morphology was evaluated and samples were analyzed for SS18-SSX expression. The experiment yielded three observations:

1. Growth speed had a dip at the start, but growth picked up after some time. Over the course of 6 weeks, the difference between induced and non-induced HEK cells was smaller than expected.
2. Morphology of the HEK-SS18SSX cells changed significantly during this prolonged induction.
3. Expression of the SS18-SSX fusion protein was strong immediately after the start of induction, but diminished within a week or so and stayed lower for the rest of the time.

The data implies the possible existence of a post-transcriptional regulatory mechanism that causes decreased expression of SS18-SSX over time, at least when expressed in HEK cells. If that mechanism existed, it would be expected to occur every time the oncoprotein was induced. Experiments with pulsed induction of SS18-SSX expression were undertaken to investigate the mechanism. The effect was not found to be consistent, and was not strong enough for use in genetic screening.

Conclusions: The difference in growth speed would not be sufficient to do a CRISPR/Cas9 experiment for the identification of growth modifiers. Similar experiments have only worked in situations where the differences in growth speed were much more dramatic. The observed impact on morphology is still under investigation.

3. To develop Zebrafish models of synovial sarcoma for validation and preclinical therapeutic translational studies

A zebrafish xenograft model of synovial sarcoma was evaluated in order to streamline drug discovery studies and therapeutic development. The zebrafish model was found to be inconsistent during our experimental analyses and tumors were not effectively cultured with the cell lines available. Cell line xenograft mouse models were also investigated, demonstrating similar inconsistencies. Neither was suitable for the preclinical in vivo drug validation studies we had hoped to perform.

After this, we decided that the genetically engineered mouse model of synovial sarcoma established in the lab of Mario Capecchi (originally described in M. Haldar et al., 2007, Cancer Cell) was a more reliable model system for therapeutic validation in synovial sarcoma. In collaboration with Dr. Kevin Jones, drug evaluation studies were undertaken in updated inducible SS18-SSX mouse models of synovial sarcoma (based on TAT-Cre orthotopic injection into SS18-SSX floxed mice, yielding consistent tumors that faithfully recapitulate most aspects of human synovial sarcoma). Top results from the drug development studies described above are currently under investigation in these models and are demonstrating promising results in initial studies.

Significance: A genetically engineered mouse model of synovial sarcoma is more effective than multiple animal xenograft models for drug evaluation in therapeutic discovery experiments. We brought a new collaborator into our team, who has developed new improved mouse models of synovial sarcoma.

Summary and Discussion

In these two years, this international collaborative group has accomplished a great deal in the biological investigation and therapeutic targeting of synovial sarcoma. The importance of a novel enzyme that modifies the epigenetic histone code in the oncogenesis of synovial sarcoma has been demonstrated, advancing our understanding of sarcoma biology, with possible implications for more specific drug targets. Models of synovial sarcoma have been investigated and optimized for drug assessment in pre-clinical studies, allowing for more effective treatment studies. A cutting-edge technique for targeted drug discovery in translocation-associated sarcomas has been developed and an innovative drug combination for improved therapy has been uncovered. This novel treatment approach is currently under evaluation for clinical trial application. Together, our group has made significant progress in understanding the biology and treatment of synovial sarcoma, while establishing active collaborations for continued and future projects.

 

Acknowledgements: The principal investigators would like especially to thank Aimee Laporte (PhD student, University of British Columbia) and Ana Banito (Postdoctoral Fellow, Memorial Sloan-Kettering Cancer Center) for their consistent efforts keeping the different laboratory teams in close communication, and for their contributions to this report.

Copyright © 2016 Liddy Shriver Sarcoma Initiative.

An International Collaborative Study of Synovial Sarcoma

By ESUN Article by Scott W. Lowe, PhD; Marc Ladanyi, MD;
Diederik de Bruijn, PhD; Ana Banito, PhD and Torsten O. Nielsen, MD, PhD

Abstract

Synovial sarcoma afflicts young adults and has a high mortality rate. The driving molecular cause, a translocation producing the SS18-SSX fusion oncoprotein, is well known but is not targetable by any known therapy. Recent evidence suggests that SS18-SSX functions by epigenetic dysregulation of transcription, but only in a permissive environment that is present in only very specific cells (likely mesenchymal stem-like cells). Technological advances in si/shRNA knockdown strategies now make it possible to screen for the critical epigenetic partner genes as well as those required to create and maintain an appropriate permissive state. While SS18-SSX cannot be targeted directly, it is possible that these partners might be targetable, allowing the development of more specific and effective systemic therapies for this disease. A translational plan involving development of novel zebrafish models of synovial sarcoma is also presented that will accelerate completion of convincing preclinical studies that would support clinical trials. With their executive roles on clinical trial groups and their network of colleagues in Canada, the Netherlands and the USA, the members of this International Collaborative Grant consortium will be well positioned to initiate Phase I or II clinical trials of targeted agents.

Introduction

Synovial sarcoma constitutes up to 10% of all soft-tissue sarcomas and arises most frequently in adolescents and young adults.¹ Conventional therapies including cytotoxic drugs (doxorubicin, ifosfamide) provide limited benefit, leaving patients at risk for local recurrence and metastases. When surgery and radiation are inadequate to remove all disease, both early and in some cases very late recurrences are unfortunately common in patients who should have a long life expectancy ahead of them.²

The defining genetic event of synovial sarcoma is the translocation of the SS18 gene on chromosome 18q11 to an SSX gene (almost always SSX1 or SSX2) located on chromosome Xp11.³ The resulting SS18-SSX fusion oncogene, exclusive to and diagnostic of synovial sarcoma, is considered the main driver in the etiology of this disease.⁴

The SS18-SSX translocation generates a fusion protein that retains the transcriptional activation domain of SS18 and the transcriptional repressor domain of SSX, as depicted in Figure 1. Though it lacks a DNA binding domain, the SS18-SSX fusion protein is thought to exert its function via interaction with transcription factors and chromatin remodelers.^5,6

The presence of such a known, specific driver mutation in virtually every case of synovial sarcoma represents a possible “Achilles’ heel” in this tumor type, which may be particularly vulnerable to targeted therapy approaches, following the precedent of chronic myelogenous leukemia. The fusion oncoprotein brings about abnormal transcriptional repression of key tumor suppressor genes via recently elucidated binding partners.⁷ The SSX portion of the oncoprotein recruits the corepressor TLE1 and polycomb-group chromatin remodeling factors - key players in the epigenetic silencing of genes - while the SS18 portion of the oncoprotein interacts with the activating transcription factor ATF2. The complex is thus directed to the promoter region of genes involved in stress response, apoptosis and cell cycle control.

Recently, others have shown that SS18’s interactions with components of the human SWI/SNF complex are disrupted in the context of the SS18-SSX mutant oncoprotein, leading to other epigenetic changes that activate transcription.⁸ Thus, epigenetic dysregulation is central to the biology of synovial sarcoma. The net result, a combination of abnormal transcriptional activation and repression, is thought to ultimately deregulate developmental programs in progenitor cells, leading to transformation instead of differentiation.⁹ Given its interaction with the epigenetic machinery, it is plausible that SS18-SSX reprograms the epigenetic landscape of cells to generate a permissive environment for sarcomagenesis. In agreement with this, two studies performing transcriptional profiling and methylation analyses on cells expressing SS18-SSX have revealed deregulation in imprinted genes and other chromatin related genes.^10,11 Thus, synovial sarcoma cells driven by SS18-SSX may possess “epigenetic” dependencies – i.e. dependencies on specific chromatin modifying activities – critical for their survival.

Means to directly inhibit the SS18-SSX oncoprotein have yet to be developed; however, a search for such epigenetic dependencies may uncover therapeutic targets for treating SS18-SSX-driven synovial sarcomas that are refractory to conventional therapies. Our international collaborative group proposes to work towards this goal through a series of integrated aims to identify targetable biology, to develop efficient and appropriate preclinical models useful for the research community, and to test potential new therapies with a view to informing future clinical trials of targeted therapies.

The Synovial Sarcoma International Collaborative Grant Team

Torsten O. Nielsen, MD, PhD, FRCPC is a clinician-scientist pathologist, and a Professor of Pathology at the University of British Columbia. He is co-chair of the NCIC-Clinical Trials Group sarcoma committee and sits on the executive board of their Investigational New Drug and Correlative Sciences committees. For the past 9 years he has held grants for the study of synovial sarcoma, recently renewed by the Canadian Cancer Society in a form significantly expanded in scope. His team currently includes two graduate students, one technologist and one research associate specifically devoted to the study of synovial sarcoma, within a larger sarcoma and breast cancer translational research group directed by Dr. Nielsen. He has a track record of over 20 publications relating to the biology, diagnosis, and treatment of synovial sarcoma. Dr. Nielsen previously collaborated in a highly successful study with Dr. Ladanyi that led to the development of TLE1 as a diagnostic biomarker of synovial sarcoma,¹² which has since come into use world-wide.^13-15

Marc Ladanyi, MD is Attending Pathologist on the Molecular Diagnostics Service in the Department of Pathology and Member in the Human Oncology & Pathogenesis Program at MSKCC. Since 2010, he is also the incumbent of the endowed William Ruane Chair in Molecular Oncology at MSKCC. Over the past 20 years, his laboratory has maintained a focus on improving the understanding of sarcomas through the study of human sarcoma tissue and sarcoma cell lines, in order to raise diagnostic accuracy, refine the assessment of prognosis, and drive the development of new therapeutic strategies. In the specific area of synovial sarcoma, he has authored or co-authored approximately 20 papers, reviews, and letters touching on synovial sarcoma and has had uninterrupted NIH funding for his work on this cancer since 2000, starting with the MSKCC Sarcoma P01 CA47179 (Project 4), followed by the current MSKCC Sarcoma SPORE P50 CA140146 (Project 4). His work has also received support from the Bustany Fund for Synovial Sarcoma. In the present proposal, his group is collaborating with the Scott Lowe laboratory on the synovial sarcoma functional genomics / synthetic lethality project.

Scott Lowe, PhD is Associate Director for Basic Cancer Research at Memorial Sloan Kettering Cancer Center and an Investigator for the Howard Hughes Medical Institute. His research has made important contributions to our understanding of p53, multi-step carcinogenesis, cellular senescence, drug resistance, and cancer evolution. Dr. Lowe’s work has been recognized by several awards, including a Sydney Kimmel Foundation Scholar Award, a Rita Allen Foundation Scholar Award, the AACR Outstanding Investigator Award, AACR-NFCR Professorship in Basic Cancer Research, and the Paul Marks Prize for Cancer Research. The Lowe laboratory currently is applying mouse models, RNA interference and cancer genomics in a coordinated effort to identify cancer maintenance genes that are useful therapeutic targets relevant to specific cancer genotypes. Recently, the Lowe laboratory has brought this expertise in cancer biology to bear on the study of synovial sarcoma, under the auspices of the MSKCC Sarcoma SPORE Developmental Research Program applying novel functional genomic strategies to this cancer.

Diederik de Bruijn, PhD is staff scientist at the department of Human Genetics of the Radboud University Medical Center in Nijmegen, The Netherlands. Since 1996, he has authored or co-authored over twenty papers dealing with the functional characterization of the synovial sarcoma associated SS18 and SSX genes. Amongst others, this research has provided insight into the fundamental role of the SS18 gene in mouse embryonic development, the interacting partners of the SS18 and SSX proteins, and the role of the SS18-SSX fusion proteins in gene regulation. Previously, Dr. de Bruijn has collaborated successfully with Dr. Nielsen in a project aimed at the identification of SS18-SSX target genes.¹⁶ This research has been supported by successive grants from the Dutch cancer society.

This work will be enhanced by close collaborations with other scientists affiliated with the Memorial Sloan Kettering Cancer Center Sarcoma SPORE grant (from the NCI), the Centre for Molecular Life Sciences at Nijmegen University, the Centre for Translational and Applied Genomics and Underhill labs in Vancouver, and collaborators in the Utah sarcoma group (Figure 2).

Goals of the Integrated Synovial Sarcoma Research Program

Our goal is to bring together resources and expertise from four investigators in three countries in order to accelerate research into how the SS18-SSX oncogene causes synovial sarcoma, and what strategies might be able to block this process. Within this theme, our specific aims are:

1. To use shRNA screening technology to interrogate epigenetic modifiers of gene expression that are likely to include critical partners contributing to SS18-SSX oncogenesis.
2. To identify cofactors responsible for creating a permissive state for SS18-SSX oncogenesis, which might be exploited as therapeutic targets.
3. To develop Zebrafish models of synovial sarcoma for validation and preclinical therapeutic translational studies.

Aim 1. Defining epigenetic vulnerabilities in synovial sarcoma through pooled shRNA screens

To better understand the mechanisms of disease in synovial sarcoma and to define new therapeutic targets, our approach in this aim is to utilize RNAi technology in both existing and new experimental systems to interrogate tumor maintenance genes. In its first iteration, we will use a validated shRNA library generated in the Lowe laboratory to target genes that control chromatin alterations, and determine whether any such alterations are required specifically for the maintenance of SS18-SSX-driven tumors.

The specific aim is to perform negative selection RNAi screens to identify and characterize potential epigenetic vulnerabilities created by the SS18-SSX fusion oncogene in human synovial sarcoma cell lines. Once we gather these data, we will compare the potential vulnerabilities in each system and study those that are common to the different cells lines to identify dependencies that are specifically created by the SS18-SSX fusion oncoproteins, focusing on those that might be exploited therapeutically. Additionally, to identify SS18-SSX specific “hits”, we will compare these vulnerability profiles to the profiles of other primary immortalized human cell lines.

Our studies will determine whether sarcoma cells are dependent on specific epigenetic mechanisms that may serve to define therapeutic targets in synovial sarcomas. Following completion of the screens, we will prioritize hits based on:

1. the presence of multiple depleted shRNAs (indicating an “on target” effect),
2. the magnitude of depletion, and
3. the potential druggabilty of the gene product.

We will also pursue hits that are not obviously druggable if they suggest interesting biology, including hits relating to genes and pathways identified in previous work by Nielsen^7,17 and de Bruijn.^6,10 Of particular interest will be hits which correlate with genes or pathways implicated in the screens for genes supporting an SS18-SSX permissive state (Aim 2). Candidate shRNAs will be further validated in additional synovial sarcoma cell lines and tested for effects on other untransformed cell types. High priority hits will also be validated in established tumors in vivo, using xenografts in nude mice, and, through collaborations made possible by this ICG, in human synovial sarcoma xenografts in zebrafish (Aim 3).

1.1. Epigenetic Vulnerabilities: Preliminary Data

As a first approach and as a proof of principle for the RNAi based negative selection screens in human sarcoma cell lines, we used a mouse synovial sarcoma cell line (M5SS1) derived from a mouse model of synovial sarcoma developed in Dr. Mario Capecchi’s lab (who have an active collaboration with the Nielsen group^7,18). In this mouse model the human SS18-SSX2 fusion oncogene is expressed in the myogenic progenitor compartment (driven by the myogenic regulatory factor Myf5-Cre), leading to the generation of tumors with full penetrance that recapitulate the histological and transcriptional profiles of human synovial sarcomas.⁶

A mouse shRNA library targeting epigenetic modulators, previously used in the Lowe lab’s recently published RNAi screen in AML,¹⁹ was transduced into M5SS1 and control mouse myoblasts. Control shRNAs targeting the human SS18-SSX2 oncogene were within the top depletion hits in the synovial sarcoma cell line but were neutral in the C2C12 myoblasts, validating the screen performance. A total of 119 shRNAs were depleted in the negative screen, epigenetic factors which are currently being validated and represent candidates for targeted therapy approaches.

1.2. Epigenetic Vulnerabilities: Experimental Methods

Work will be extended to human synovial sarcoma cell lines, which can be expected to more faithfully represent human disease. Through this collaborative project, the Lowe and de Bruijn labs will gain access to a number of human tumor cell lines available from the Ladanyi and Nielsen labs. These cell lines, described in Table 1, include cell types expressing SS18-SSX1 or SS18-SSX2, and both monophasic and biphasic subtypes.

Synovial Sarcoma Cell Lines
Available for These Studies

Cell Line	Original Tumor	Fusion Type	Reference
HS-SY-II	Monophasic	SYT-SSX1	Abe20 [20]
SYO-1	Biphasic	SYT-SSX2	Beppu [21]
FUJI	Monophasic	SYT-SSX2	Nagashima [22]
YaFUSS	Monophasic	SYT-SSX1	Toguchida [23]
Aska-SS	Biphasic	SYT-SSX1	Itoh [24]
Yamoto-SS	Biphasic	SYT-SSX1	Itoh [24]
FU-SY-1	Monophasic	SYT-SSX1	Kikuchi [25]
KU-SS-I	Monophasic	SYT-SSX1	Takagi [26]

We have begun to characterize the response of the HS-SY-II and YaFUSS human sarcoma cells lines to knockdown of the SS18-SSX oncogene, and have confirmed dependency on SS18-SSX1 expression. We will proceed with the negative selection RNAi screen using a human version of the epigenetic shRNA library.¹⁹ The RNAi screen will follow the steps outlined in schematic form in Figure 3. In brief, shRNAs will be transduced as one pool into human synovial sarcoma cell lines in triplicate. GFP-positive cells will be sorted 48 hours post-transduction (t0) and following serial passage over a course of approximately 16 days (t final). Changes in library representation will be monitored using deep sequencing of shRNA guide strands amplified from genomic DNA. In order to identify SS18-SSX-specific hits in this context, we will compare the lethality signatures of the human synovial sarcoma cell lines with that of human mesenchymal stem cells or/and immortalized IMR90 human fibroblasts. In parallel, we also plan to exploit the dataset created from screening other human cell lines to identify lethality signatures that may be oncogene dependent.

Because synovial sarcoma is a mesenchymal cancer, and expression of SS18-SSX is sufficient to transform rat fibroblasts,⁵ we will generate immortalized IMR90 human fibroblasts that are engineered to inducibly express the SS18-SSX2 oncogene. We will attempt to screen the library in that context as well; because the impact of SS18-SSX fusion protein overexpression in this system may not reflect the correct physiologic context, we will prioritize the screens using synovial sarcoma cell lines. However, this system may prove to be useful as an additional setting to test the effects of shRNAs scoring in the performed screens where we have a more direct negative control (i.e. IMR90 without induction of SS18-SSX).

Following completion of the screens, we will prioritize hits based on:

1. the presence of multiple depleted shRNAs (indicating an “on target” effect),
2. the magnitude of depletion, and
3. the potential druggability of the gene product.

Candidate shRNAs will be further validated in additional synovial sarcoma cell lines (Table 1) and tested for effects on other untransformed cell types, including translocation-associated sarcomas known to present in a similar clinical setting and available in the Ladanyi or Nielsen labs. High priority hits will also be validated in established tumors in vivo, using a transplantable renal subcapsular xenograft available from the Nielsen lab, the conditional mouse model, and the Zebrafish models developed in Aim 3 (which once ready will give faster results at lower cost). Such systems will exploit recently published inducible shRNA systems developed by the Lowe group.²⁷

Aim 2. Modifiers of the Synovial Sarcoma Permissive State: Novel Therapeutic Targets

In various experimental settings, we and others have found that the efficiency of transformation by the synovial sarcoma-specific SS18-SSX oncoproteins can vary significantly. Upon SS18-SSX expression, some cells undergo cell death, while others are able to cope with this situation and evade this death response. The ability of a cell to tolerate SS18-SSX expression has been defined as the ‘permissive state’. Based on this definition, existing synovial sarcoma cell line models can be classified according to their permissiveness. Synovial sarcoma derived cell lines (Table 1) can be regarded as the descendants of permissive cells and are expected to have retained this property. In contrast, HEK293 cells do not seem to be very permissive. This is one conclusion from our previous work in which we characterized HEK293-SS18-SSX cells which were engineered to contain a tetracyclin-inducible SS18-SSX gene.¹⁰ Although these cells did not die instantly upon induction of SS18-SSX expression, they failed to proliferate and died after prolonged SS18-SSX expression.

On a cellular level, evidence suggests that this permissive state is associated with the level of differentiation. Various studies have demonstrated that (undifferentiated) pluripotent stem cells respond differently to the forced expression of SS18-SSX fusion proteins than (more differentiated) mesenchymal stem cells.^5,9,28,29 Recent molecular data suggest that this permissive state may also depend on the successful integration of SS18-SSX fusion proteins into the SWI/SNF complex, although the existing literature data on this subject appears to be somewhat contradictory. Some authors have found that induced SS18-SSX expression leads to disruption of the SWI/SNF complex,⁸ while others have demonstrated that exogenous SS18-SSX proteins can function as stable components of complete SWI/SNF complexes.³⁰ These apparently conflicting data can be reconciled, because SWI/SNF disruption was observed as an immediate early effect of SS18-SSX expression, while the stable integration was found in cells with stable SS18-SSX expression and stable growth. Since the latter was observed in surviving HEK293 subclones, cellular selection may play an important part in the adaptation of cells, suggesting the presence of modifiers that facilitate the sustained growth of SS18-SSX expressing cells. Finally, this process of cellular adaptation may be linked to the successful integration of SS18-SSX fusion proteins into the SWI/SNF complex.

We hypothesize that the permissive state of synovial sarcoma precursor cells is defined by specific modifiers that allow cells to cope with expression of the SS18-SSX oncoproteins and evade the initial death response. As these would represent critical cofactors for SS18-SSX oncogenesis and tumor cell survival, the identification of such permissive state modifiers may point the way to the identification of alternative therapeutic targets.

2.1: Identification of Permissive State Modifiers

Based on our own previous data, we regard HEK293 cells to be not very permissive to the induction of SS18-SSX expression. As such, the existing HEK293-SS18-SSX model will now be used to identify modifiers of this permissive state. As noted, we have observed growth inhibition upon tetracycline induced SS18-SSX expression, followed by cell death after prolonged exposure. Now, we will repeat the SS18-SSX induction in these cells, in combination with a positive selection RNAi screen of a genome-wide human shRNA library. The data from this screen are expected to be complementary to the data generated under Aim 1 of this proposal, which describes a negative RNAi screen with permissive (synovial sarcoma) cells.

To identify shRNA modifiers of SS18-SSX permissiveness, duplicate HEK293-SS18-SSX cultures will be transduced with ten different lentiviral shRNA pools at an infectivity of one shRNA per cell. Combined, these pools contain shRNAs targeting >15,000 human genes. Two days after transduction, SS18-SSX expression will be induced with tetracycline in the tranduced cultures and in one set of non-transduced (control) cultures, and then they will be allowed to grow for another sixteen days. Through this time, cells will be serially passaged as needed and samples will be taken at each passage. Finally, shRNAs will be rescued from the surviving cells and sequenced directly in a high throughput fashion. ShRNAs that are found to be enriched in the sequencing data of both replicates will be tested individually for their effects on SS18-SSX permissiveness of HEK293 cells. In this fashion, we expect to identify multiple individual shRNAs that affect the SS18-SSX permissive state of HEK293 cells.

2.2: Evaluation of Permissive State Modifiers as Novel Therapeutic Targets

The genes that are targeted by the shRNAs identified under specific aim 2.1 will be identified, and relevant properties of these genes, such as predicted protein functions, expression patterns and mutations found in synovial sarcomas and/or other tumors will be interrogated by bioinformatic approaches. Based on this analysis, a subset of the identified permissive state modifier genes will be further analyzed with respect to their in vivo role in synovial sarcoma development. In addition, we will address whether the proteins encoded by these genes are already known to be druggable in an experimental setting.

Based on the knowledge currently available, synovial sarcoma permissive state modifiers may influence the integration of SS18-SSX fusion proteins into the SWI/SNF complex. Therefore, it is conceivable that these will include SWI/SNF associated proteins, SWI/SNF targeting factors and/or other (as yet unknown) SWI/SNF assembly factors. Permissive state modifiers may also be linked directly to the TLE-Polycomb-HDAC complexes. This is of specific interest, given that our previous work has indicated that HDAC inhibition effectively stops synovial sarcoma cell growth. Consequently, we will cross-reference the results from these whole genome screens with the validated shRNAs for epigenetic modifiers described in Aim1.

Aim 3. Rapid in vivo screening of synovial sarcoma xenografts using Danio rerio (zebrafish)

While monolayer cultures are convenient for studying the intrinsic vulnerabilities of tumor cells, it is well established that they do not accurately reflect the phenotype of tumor growth in vivo. In the first place, enhanced proliferation and survival are usually desirable, and media are normally supplemented with high levels of growth factors, most notably IGF2 and FGF2 (major growth factor constituents of fetal bovine serum). Major differences are also introduced due to lack of physiologic signaling, including three dimensional growth, cell-cell contact, and contact with extracellular matrix. For these reasons, gene expression profiles of tumor cells grown in three dimensional spheroid culture are more representative of primary tumors,³¹ and the predictive value of drug sensitivity assays is significantly improved in 3D culture.³²

For characterization of screening hits and drug targets identified in aims 1 and 2, xenografts remain the gold standard since they also model interactions with extracellular matrix, normal host cells, and the physiological microenvironment - all critical parameters for tumor growth and survival. Animal models are also essential for evaluation of drugs for their ability to inhibit metastasis, and several models of experimental metastases have been successfully developed in zebrafish embryos. Murine models, while commonly used, are relatively labour intensive, expensive and slow to yield results, making them ill-suited for moderate to high throughput studies such as screen verifications and drug combination studies. For these investigations, zebrafish xenografts will be developed as an intermediate between in vitro monolayer screening hits and more costly murine investigations (Figure 4).

Zebrafish are a tractable model system for this application, producing upwards of 300 embryos per female at each spawn. The embryo and larval stages are essentially transparent, facilitating live cell imaging at single cell resolution. In this way human tumor cells may be conveniently labeled - by staining with dye for short-term studies (Figure 5), or by transduction fluorescent proteins, and xenografts can monitored in detail throughout the experiment. These systems allow rapid, real time evaluation of xenograft viability, angiogenesis, invasion, and metastases, and can be scaled up to high throughput screens for drug response.³³

Models of metastasis are critically important for preclinical drug development, but a satisfactory model of spontaneous metastasis in subcutaneous mouse xenografts has not been achieved in synovial sarcoma. Potentially this is due to insufficient vascularization at this site, but possibly also due to limitations of detection in the mouse. Tools for genetic manipulation of zebrafish are well developed, including over 13 000 mutant and transgenic zebrafish lines available through ZIRC, the Zebrafish International Resource Center. This will allow selection of specific mutants important for host interactions, as well as the use of reporter lines to detect activation of specific pathways. Specific host cell types can also be engineered to express fluorescent markers - in particular the zebrafish s

train Tg(vegfr2:g-rGFP), expresses green fluorescent protein (GFP) in all endothelial cells, and will be used to monitor tumor-induced angiogenesis.

3.1: Developing a zebrafish xenograft assay for screening novel chemicals and drug combinations against synovial sarcoma cell lines.

Adapting a cell tracer technique commonly used in developmental studies, synovial sarcoma cells will be labeled with the covalent membrane stain CM-Dil, which will allow detection over the short duration of embryo xenografts – alternately stable transduction of synovial sarcoma cells with red fluorescent protein will be used. For each of four synovial sarcoma cell lines, up to 1000 labeled cells will be microinjected into the yolk sac of Tg(vegfr2:g-rGFP) zebrafish embryos, at 48 hours post fertilization. In parallel studies, up to 400 cells will be injected into the perivitelline space or embryonic circulation in order to develop models of spontaneous or experimental metastasis. Cell lines will be evaluated for efficiency of engraftment, growth kinetics, induction of angiogenesis, and formation of micrometastases. Embryos will be raised in 96 well microtitre plates, and xenografts will be visualized by real time confocal fluorescent imaging.

Standard anticancer drugs used in the treatment of synovial sarcoma (doxorubicin, ifosfamide) will be evaluated as positive controls, and lead compounds from the work in aims 1 and 2 in this grant proposal will be evaluated in this system. In particular, previous work has shown synovial sarcoma to be sensitive to HDAC inhibitors, and that this class of compounds has the potential to act directly on the aberrant SS18-SSX complex responsible for gene repression. As these agents may be most effective in combination with other drugs, we are pursuing high throughput screening of chemical libraries for potential synergies with HDAC inhibitors. Positive hits from these drug screens will also be evaluated in vivo using the zebrafish model, strengthening the rationale for preclinical studies in mice.

Similarly the zebrafish xenograft system will serve to validate high confidence shRNA screening hits and drug candidates identified from monolayer systems in Aims 1 and 2. Specific pathways and host cell interactions will be manipulated in zebrafish xenografts, in order to strengthen and extend the conclusions from monolayer culture. In addition to work implanting cell line models, using our existing tumor tissue procurement systems in place at MSKCC and at UBC (where Dr. Nielsen directly handles the synovial sarcoma surgical excision specimens) small fragments of primary synovial sarcoma tissue will be labeled with red fluorescent dye CM-Dil and transplanted in zebrafish embryos as above. Viable transplants will be used to validate and reproduce the cell line data in order to confirm analogous results for known drug effects and their relative degree of signal in the assay.

3.2 Cancer Stems Cells in Zebrafish Models of Synovial Sarcoma

The fast turnaround time of the zebrafish system makes it an ideal format to test for the presence of cancer stem cells in synovial sarcoma. Cancer stem cells are an important paradigm and have been isolated in a number of human cancers, where they are classically enriched by flow cytometry within a “side population”, based on their ability to rapidly efflux the fluorescent Hoechst dye (owing to upregulation of drug transporters).³⁴ Otherwise known as tumor initiating cells, these cells are a minor subpopulation with stem-like properties, whose more differentiated progeny make up the bulk of a tumor’s mass. In other cancers chemoresistance and metastases have been attributed to this subpopulation – postulates that have not been tested in synovial sarcomas but which hold important ramifications for drug and shRNA screening studies.

The ”gold standard” assay for cancer stem cells is markedly enhanced tumorigenicity in xenografts, and the zebrafish embryo will provide an efficient and rapid assay of subpopulations at limiting dilutions. In these studies cancer stem cell isolation protocols will be assayed in zebrafish xenografts for disaggregated primary tumors and for at least four of our synovial sarcoma cell lines (Table 1). A variety of biomarkers for progenitors in mesenchymal and neural crest lineages will be evaluated in these isolates, and results from cell culture and primary tumors correlated.

Conclusion: Translation into Clinical Care

The ultimate goal of this work is to translate our insights into targetable epigenetic (Aim 1) and permissive state (Aim 2) pathways, as supported by the preclinical model work (Aim 3), into clinical trials to provide new options for patients with synovial sarcoma. The agents we identify through this work may have various timelines for drug development (i.e. if a drug repurposing strategy is identified, a phase II study might be immediately possible, whereas investigational new drugs may require preclinical toxicology and phase I studies to be completed first, and earlier stage lead compounds may well require medicinal chemistry and commercial partnerships to take forward). Our group and our affiliated members have a network of connections and experience with work at each of these levels. Dr. Nielsen is chair of the Sarcoma committee at the NCIC-Clinical Trials Group, an executive member of their Investigational New Drugs committee, and a correlative scientist providing support for North America-wide clinical trials running through the SARC consortium. However, cooperative groups can face considerable delays in trial activation due to the challenges of involving multiple centers. MSKCC has special advantages in this regard as one of the world’s largest referral centers for sarcoma therapy, making it one of only a very few centers in the world that can quickly translate leads from pre-clinical studies in synovial sarcoma into small clinical trials. Dr. Ladanyi interacts with an extremely active Phase 1 clinical trials effort at MSKCC, including a significant annual volume of patients referred for advanced treatment of synovial sarcoma. Through these connections, our ICG consortium is in an excellent position to translate our insights into clinical trials. We thank the Liddy Shriver Sarcoma Initiative and its supporters for catalyzing and supporting our international efforts to help patients with synovial sarcoma.

V11N3 ESUN Copyright © 2014 Liddy Shriver Sarcoma Initiative.

$250K International Collaborative Grant Funds Synovial Sarcoma Research

By ESUN Announcement

May 1, 2014 - The Synovial Sarcoma Research Foundation and the Liddy Shriver Sarcoma Initiative are pleased to award a $250,000 International Collaborative Grant for synovial sarcoma research. The research project, which focuses on understanding the development of the disease, is a global effort of five investigators in Canada, the Netherlands and the United States.

Synovial sarcomas comprise about 10% of soft tissue sarcomas, and they often affect adolescents and young adults. Conventional chemotherapies are of limited benefit to synovial sarcoma patients, leaving them at risk for recurrence and metastasis.

Unlike most cancers, synovial sarcomas are defined by a unique genetic mutation. Unfortunately, scientific knowledge of that mutation, the SS18-SSX fusion oncogene, has yet to lead to improved treatments. This study aims to uncover the genetic pathways that help synovial sarcoma cells with the SS18-SSX mutation develop and thrive. The research team believes that their findings could lead to effective therapeutic strategies for people with synovial sarcoma.

Dr. Torsten Nielsen, one of the study's principal investigator, explains: "In synovial sarcoma, the mutation that drives this cancer has been identified, but it isn't targeted by any existing drugs. We're trying to figure out what genes and proteins are its 'partners in crime' that allow this mutant protein to transform a cell into a dangerous cancer." Finding those partners is key, because drugs may already exist to target them and cripple synovial sarcoma cells.

Utilizing New Technology and Existing Drugs

RNA interference (RNAi) is an advanced technology that can identify genes that are important for the proliferation and survival of cancer cells. This study's investigators will use RNAi to help identify molecules that are needed to sustain the growth and survival of synovial sarcoma cells with the SS18-SSX mutation. Inhitibiting those molecules should kill synovial sarcoma cells but not normal cells.

According to Dr. Nielsen, drug companies have not been developing targeted therapies for sarcomas because each subtype is so rare. But he also believes there is hope: "It is very likely that the distinct mutational drivers that have been identified in sarcomas, such as the SS18-SSX fusion oncogene in synovial sarcoma, act via pathways that are also activated in other types of cancer, and for which inhibitory drugs do exist."

A Personal Connection

Occasionally a sarcoma researcher has a compelling personal connection to his work. In this study, Dr. Scott Lowe is motivated by the memory of a young scientist: "I had a talented postdoctoral fellow in my laboratory who was diagnosed with, and eventually died from, this disease. Henry was a fantastic person and had enormous potential as a scientist, and sadly died one week before his wife gave birth to twins. Since his death, I have been interested in finding opportunities where our work could contribute to a better understanding of the disease. The opportunity presented by this collaborative grant accomplishes this and goes further, in that it might help identify new ways to eventually treat synovial sarcoma."

The Value of Sarcoma Specific Research

According to Dr. Lowe, research into the genes that drive synovial sarcoma and other sarcomas is particularly important. He says, "These cancers differ substantially from many of the more commonly studied adult cancers that occur in organs such as the lung, breast, prostate, and colon. Owing to the shear volume of patients with the common cancers, most efforts in industry have focused on those diseases. As sarcomas are so different, little of this effort trickles down to impact sarcoma patients."

Dr. Marc Ladanyi adds, "Simply applying models and mindsets based on common adult cancers is not sufficient or effective in this group of cancers. There are numerous distinct types of sarcoma, each with different genetic alterations. A lack of appreciation for the diversity of sarcomas has in the past led to lumping biologically unrelated sarcomas in the same study, often causing important findings to be overlooked."

The members of this research team are excited to focus exclusively on synovial sarcoma, which is rare and often very difficult to treat. Dr. Lowe explains, "Collaborative research, such as this grant supported by the Liddy Shriver Sarcoma Initiative, provides the funds for researchers to sufficiently understand a disease to make drug development efforts an attractive target for biotechnology and pharma."

Working with the Initiative

The Liddy Shriver Sarcoma Initiative awards international collaborative grants like this one in order to increase the power of private donations and to accelerate research that has the potential to save and improve lives.

The Initiative uses a stringent process for grant review that Dr. Nielsen appreciates: "The peer review process for International Collaborative Grants bring in comments from twice as many reviewers as typical grant applications, ensuring that a thorough critique and consequent high-quality science will be produced from donors' funds."

Investigators welcome the opportunity to publish their plans and findings in ESUN. Dr. Ladanyi explains, "This web publication promotes a very open scientific interaction and also allows the supporters of the Liddy Shriver Sarcoma Initiative to get a more direct sense of the science they so generously fund."

The Funding

This grant is co-funded by the Synovial Sarcoma Research Foundation and the Liddy Shriver Sarcoma Initiative, each of which provided $125,000 in support. The two organizations are working together to initiate and support high-quality basic and translational research in synovial sarcoma.

The purpose of the Synovial Sarcoma Research Foundation is to initiate, stimulate, support and fund synovial sarcoma research. Romer, the son of founders Susan van Dijk and Koen Jansen, was only one year old when he was diagnosed with a synovial sarcoma in his right shoulder. He is now an energetic five-year-old. Having Romer go through the experience of general treatment when they believed something special was needed, Susan and Koen became interested in supporting research targeted at this very distinct sarcoma. To further their goal, they formed a strategic partnership with the Liddy Shriver Sarcoma Initiative to focus on synovial sarcoma research. Susan and Koen said, "The knowledge and energy that Bruce and Beverly bring to the table and their ability to have the world’s best researchers work together globally is amazing. They made it possible for the Synovial Sarcoma Research Foundation to initiate and fund this wonderful project."

The Initiative greatfully acknowledges generous donations from the Wendy Walk (in memory of Amanda Noble) and the Alan B. Slifka Foundation (in memory of Ann Davis), donations in honor of Elodie Espesset and Jason Fesel, who are currently in treatment for this disease; and donations in memory of Lucia Kramer, Artem Petrossian, Louie Pellegrini IV, Jensen Barrett, Rhonda Williams, Alison Schmeling Foerster, Wayne Coffey III, and Jennifer Colladay, who lost their lives to synovial sarcoma.

Together, we are making a difference!

V11N3 ESUN Copyright © 2014 Liddy Shriver Sarcoma Initiative.