Abstract SY15-01: Modeling the initiation of Ewing sarcoma in human neural crest stem cells

Abstract

Abstract Ewing sarcoma family tumors (ESFT) are aggressive bone and soft tissue tumors that primarily affect children and young adults. Despite intensive therapy, relapse is common and for patients who relapse or present with overt metastatic disease survival rates are only 20%. Intensification of current treatment regimens is impossible due to dose-limiting toxicities. Therefore, novel, less toxic approaches to therapy are desperately needed. For ESFT as well as other pediatric tumors, these novel therapies will ideally target tumor cells whilst sparing normal developing tissues. Attainment of this important goal demands that we better understand the molecular underpinnings of ESFT and how its development deviates from normal. ESFT can occur in any anatomic location and their cellular origin remains elusive. Genetically they are characterized by expression of chimeric fusion oncogenes, most commonly EWS-FLI1. Although it functions as an oncogene in ESFT, EWS-FLI1 is toxic to most normal cells. Recent data from our lab and others indicate that some bone marrow-derived mesenchymal (BM-MSC) and neural crest-derived (NCSC) stem cells are uniquely tolerant of EWS-FLI1. Moreover, the genetic profile of ESFT, combined with its ability to differentiate in vivo and in vitro into both neural and mesenchymal lineages implicates multipotent neuro-mesenchymal stem cells as putative cells of origin. Significantly, recent studies in model organisms have demonstrated that rare populations of NCSC persist in adult tissues and that some MSC are actually of neural crest rather than mesodermal origin. We hypothesized that these NCSC are the likely cellular ancestors of ESFT and sought to develop a human NCSC-based model of EWS-FLI1-induced transformation and ESFT initiation. NCSC are exquisitely rare in postembryonic tissues. Therefore, we devised a novel method to generate and isolate NCSC from in vitro differentiating human embryonic stem cells (hESC). Importantly, these cells, which are sorted on the basis of NCSC-markers p75 and HNK-1, retain multilineage neural crest differentiation potential in vitro and in vivo. In addition, these hNCSC can be induced to generate MSC-like cells in vitro (NC-MSC) that can in turn be differentiated along neural and glial as well as osteogenic and adipogenic lineages. Importantly, lentiviral transduction of hNCSC and their NC-MSC progeny with EWS-FLI1 is well tolerated demonstrating that, like BM-MSC, human NCSC are permissive for expression of the fusion oncoprotein. Having established that NCSC tolerate EWS-FLI1 we next evaluated genetic changes downstream of oncogene activation and also compared gene expression of primary ESFT to the three different stem cell populations (hNCSC, NC-MSC, BM-MSC) as well as 11 normal adult tissues. Significantly, whole genome expression profiling studies revealed that the molecular signature of established ESFT is more similar to hNCSC than any other tissue or stem cell population. In addition, expression of EWS-FLI1 in NC-MSC partially reactivated the more primitive NCSC genetic program and also initiated transition to a more ESFT-like state. These data suggest that EWS-FLI1-induced maintenance or reactivation of the NCSC program is integral to ESFT pathogenesis. Consistent with this hypothesis, we found that EWS-FLI1-expressing cells retained expression of NCSC markers, maintained the ability to form neurospheres in nonadherent conditions and upregulated expression of the polycomb proteins BMI-1 and EZH2. Moreover, unlike empty vector-transduced cells, EWS-FLI1+ cells avoided cellular senescence and this was dependent upon continued overexpression of BMI-1 and epigenetic repression of p16. Recent studies of epithelial tumors have discovered that polycomb target genes that are normally reversibly silenced by histone modifications in normal stem cells are often aberrantly and heritably silenced by DNA hypermethylation in cancer. We have found that EWS-FLI1 induces BMI-1 and EZH2 in hNCSC and other groups have confirmed that EZH2 is a direct transcriptional target of the fusion protein. It is our current hypothesis that EWS-FLI1-mediated deregulation of polycomb proteins in NCSC precipitates changes in the epigenome that ultimately result in DNA methylation-mediated silencing of developmental pathways, thereby contributing to maintenance of stemness and susceptibility to malignant transformation. In support of this, kinetic genome-wide evaluation of transduced NC-MSC revealed that changes in gene expression downstream of EWS-FLI1 activation are highly dynamic. In particular, with increasing passage of time, EWS-FLI1+ cells progressively downregulate transcripts involved in cellular differentiation and upregulate genes that promote cell proliferation. To begin to address whether changes in DNA methylation at the promoters of polycomb target genes might underlie these progressive changes in gene expression we used Illumina GoldenGate technology to compare ESFT tumor and cell line samples to hNCSC, BM-MSC and normal adult tissues. Significantly, we found that transcription factors that normally instruct neural differentiation are progressively silenced in EWS-FLI1-transduced NC-MSC and are also aberrantly targeted for DNA promoter methylation in established tumors. We are now performing kinetic, genome-wide studies of DNA methylation in EWS-FLI1+ NC-MSC to determine if EWS-FLI1 globally induces altered methylation of polycomb target gene promoters. Updated data from these studies will be presented at the meeting as available. In summary, we have developed a novel developmental stage and species-appropriate model in which to study the cellular origin of ESFT and the molecular mechanisms of EWS-FLI1-induced transformation. Our studies thus far indicate that ESFT is a stem cell tumor in which a single genetic mutation induces profound epigenetic changes that ultimately cooperate to induce malignant transformation. We are now investigating the nature of cooperating cell-autonomous as well as microenvironmental factors that are necessary to convert an EWS-FLI1+ NCSC into an ESFT tumor-initiating cell. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr SY15-01. doi:10.1158/1538-7445.AM2011-SY15-01