Published OnlineFirst June 23, 2014; DOI: 10.1158/1078-0432.CCR-13-0275 Clinical Cancer Molecular Pathways Research Molecular Pathways: Targeting ETS Gene Fusions in Cancer Felix Y. Feng1,2,3, J. Chad Brenner2,3,4,5, Maha Hussain3,6,7, and Arul M. Chinnaiyan2,3,4,7,8 Abstract Rearrangements, or gene fusions, involving the ETS family of transcription factors are common driving events in both prostate cancer and Ewing sarcoma. These rearrangements result in pathogenic expression of the ETS genes and trigger activation of transcriptional programs enriched for invasion and other oncogenic features. Although ETS gene fusions represent intriguing therapeutic targets, transcription factors, such as those comprising the ETS family, have been notoriously difficult to target. Recently, preclinical studies have demonstrated an association between ETS gene fusions and components of the DNA damage response pathway, such as PARP1, the catalytic subunit of DNA protein kinase (DNAPK), and histone deactylase 1 (HDAC1), and have suggested that ETS fusions may confer sensitivity to inhibitors of these DNA repair proteins. In this review, we discuss the role of ETS fusions in cancer, the preclinical rationale for targeting ETS fusions with inhibitors of PARP1, DNAPK, and HDAC1, as well as ongoing clinical trials targeting ETS gene fusions. Clin Cancer Res; 20(17); 4442–8. Ó2014 AACR. Background tion domain (from the EWS gene) to the ETS fusion and ETS transcription factors are aberrantly expressed in (ii) replacement of the N-terminus of the ETS protein by several cancers, including prostate cancer (1), the Ewing an RNA-binding domain from the EWS protein that sarcoma family of tumors (2), melanoma (3), secretory enhances posttranscriptional splicing of ETS target genes breast carcinoma (4), acute lymphoblastic leukemia (5), (10; Fig. 1). gastrointestinal stromal tumors (6), and rare cases of Both prostate cancer and Ewing sarcoma ETS genomic acute myelogenous leukemia (7). The ETS family consists rearrangements are thought to occur early in malignant TMPRSS2–ERG of 28 unique genes (reviewed in ref. 8), of which ERG, progression. For example, fusions are FLI1,andETV1 are the most frequently deregulated in observed during the transition from high-grade prostatic cancer. Prostate cancer frequently harbors rearrange- intraepithelial neoplasia lesions to invasive carcinoma ments of ETS genes, in which ERG (50% of all prostate (9, 11) and are formed at high frequency in androgen- cancers) and ETV1 (5%) are fused to the androgen-reg- stimulated cell lines under genotoxic stress (12–14). ulated promoter and 50 untranslated region of the However, mice genetically engineered to express andro- ERG ETV1 TMPRSS2 gene (1, 9). This creates an androgen-regulated gen-regulated or develop prostatic intraepithelial TMPRSS2–ETS fusion transcript that encodes a nearly neoplasia-like lesions, but do not progress to frank full-length ETS transcription factor (Fig. 1). In addition, carcinoma (9, 11, 15–17). This suggests that complete almost all Ewing sarcomas contain an ETS rearrangement, ETS-mediated transformation may require additional col- including EWS–FLI1 (90%) or EWS–ERG (5%–10%) laborating mutations. While this spectrum is only begin- gene fusions, which encode a chimeric protein notable ning to emerge (18–20), it is clear that ERG accelerates for several features, including (i) provision of an activa- prostate carcinogenesis following loss of a highly recurrent prostate cancer tumor suppressor protein called PTEN or in the context of overexpression of the androgen receptor 1Department of Radiation Oncology, University of Michigan Medical (15–17). Interestingly, TMPRSS2–ERG overexpression School, Ann Arbor, Michigan. 2Michigan Center for Translational Pathol- leads to increased self-renewal over multiple plating gen- 3 ogy, University of Michigan Medical School, Ann Arbor, Michigan. Com- hi þ prehensive Cancer Center, University of Michigan Medical School, Ann erations in Sca-1 /EpCAM basal/progenitor cells isolated Arbor, Michigan. 4Department of Pathology, University of Michigan Medical from genetically engineered mice (21), suggesting a role for School, Ann Arbor, Michigan. 5Department of Otolaryngology, University of ETS fusions in prostate cancer progenitor populations. In Michigan Medical School, Ann Arbor, Michigan. 6Department of Internal Medicine, Division of Hematology/Oncology, University of Michigan Med- contrast with prostate cancer, the cells from which Ewing ical School, Ann Arbor, Michigan. 7Department of Urology, University of sarcomas are derived are still unknown, limiting the inter- 8 Michigan Medical School, Ann Arbor, Michigan. Howard Hughes Medical pretation of genetic mouse models. Despite this impedi- Institute, University of Michigan Medical School, Ann Arbor, Michigan. ment, EWS–FLI1 overexpression has been shown to induce Corresponding Author: Arul M. Chinnaiyan, University of Michigan Med- ical School, 1400 E. Medical Center Drive, 5316 CCGC 5940, Ann Arbor, MI leukemic phenotypes when expressed in hematopoietic 48109-5940. Phone: 734-615-4062; Fax: 734-615-4055; E-mail: stem cells (22), to induce skeletal disruption when [email protected] expressed in mesenchymal progenitors using a PRX1 pro- doi: 10.1158/1078-0432.CCR-13-0275 moter (23), and to accelerate tumor formation in conjunc- Ó2014 American Association for Cancer Research. tion with TP53 deletion (23). 4442 Clin Cancer Res; 20(17) September 1, 2014 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2014 American Association for Cancer Research. Published OnlineFirst June 23, 2014; DOI: 10.1158/1078-0432.CCR-13-0275 ETS Gene Fusions in Cancer Consistent with their role in prostate cancer and Ewing Following upfront androgen deprivation therapy, many sarcoma progression, ETS transcription factors drive down- patients will relapse with castration-resistant prostate stream signaling pathways with a number of functional cancer. The restoration of androgen signaling (35) and consequences. RNAi-mediated disruption of either TMPRSS2–ERG expression (36) in castration-resistant TMPRSS2–ERG or EWS–FLI1 expression inhibits cell pro- disease provides a foundation for the hypothesis that liferation, invasion, metastasis, and xenograft growth of ETS-positive castration-resistant prostate cancer may be prostate cancer or Ewing sarcoma cell line models that preferentially responsive to next-generation antiandrogen harbor the respective fusions (24–26). Accordingly, the therapy, such as abiraterone acetate. Abiraterone blocks transcriptional program driven by overexpression of ETS androgen synthesis by inhibiting the enzyme cytochrome gene fusions is enriched for invasion and metastasis-asso- P450 17 a-hydroxysteroid dehydrogenase (37) and has ciated gene signatures (1, 27, 28). Recently, our group found improved clinical outcomes for patients with castration- that both prostate cancer and Ewing sarcoma ETS gene resistant disease in large phase III clinical trials (38, 39). fusions induce DNA double-strand breaks (25, 26). This Using patient specimens from smaller phase I/II studies of suggests that ETS gene fusions may drive a mutator pheno- metastatic patients treated with abiraterone, Attard and type and cause increased genomic instability in some cells. colleagues found that the presence of the predominant ETS Given the pathogenic roles of ETS fusions in the progres- fusion, the TMPRSS2:ERG rearrangement, in circulating sion of both prostate cancer and Ewing sarcoma, ETS fusion tumor cells (CTC) correlated with prostate-specific antigen products represent intriguing potential therapeutic targets. (PSA) response (40). In this study, 38% of patients with However, transcription factors, such as the ETS family, have ERG fusion–positive CTCs had a >90% decline in PSA level been notoriously difficult to target (29). Potential strategies with abiraterone, compared with 7% of patients with ERG for targeting ETS fusion genes include therapies directed at fusion–negative CTCs (40). In contrast, Danila and collea- the gene promoter, the RNA transcript, the fusion product gues (41) found that TMPRSS2:ERG status in CTCs was not itself, coregulators of the fusion product, other collaborat- associated with response to abiraterone. As with the castra- ing lesions, and downstream targets of the fusion. Although tion-sensitive setting, these discrepancies raise additional each of these strategies holds promise, this review focuses questions, such as whether ETS fusion status in the CTCs on agents available to patients or currently in clinical trials, accurately reflects fusion status in the metastatic lesions. leading to an emphasis on therapies directed at the andro- To address these questions, a multi-institutional random- gen-responsive promoter (in prostate cancer) or against ized phase II clinical trial (clinicaltrials.gov identifier: coregulators of the fusion product. NCT01576172) was initiated by our group at the University of Michigan with the objective of assessing several key questions, including the relationships between ETS fusion – Clinical Translational Advances status and the response to antiandrogen therapy. Speci- Targeting the promoter of ETS fusions fically, this trial, which requires biopsy of metastatic pro- The fact that the predominant ETS fusions in prostate state cancer lesions for enrollment, prospectively stratifies cancer contain an androgen-responsive promoter (1, 24, 30, patients by ETS fusion status in biopsies before random- 31) provides a strong rationale