Erythroblast transformation-specific transcription factor fusions in prostate cancer.

Abstract

224 Background: Common drivers in advanced prostate cancer (PC) involve fusions of the 3’ DNA binding domain-containing region of an Erythroblast Transformation Specific (ETS) transcription factor gene with a 5’ region of another gene, often one that is androgen-regulated. Previous work indicates that PCs harboring TMPRSS2:ERG fusions are more dependent on androgen signaling, and men with such tumors may, therefore, be more responsive to the effects of androgen deprivation therapy than men lacking this fusion. In this study, we examined the frequency of fusions involving ETS-family genes, including TMPRSS2:ERG, and determined how often such fusions would be missed if detection relied solely on DNA sequencing. Methods: To identify fusions and compare DNA-only versus DNA plus RNA sequencing, we utilized the Oncomap ExTra genomic profiling assay. This assay performs whole-exome, whole-transcriptome DNA and RNA sequencing of tumor and matched normal samples to identify somatic alterations in tumors. Single-nucleotide variants, indels, copy number alterations, alternative transcripts, and gene fusions are all detected. We specifically searched for fusions involving ETS transcription factor genes ERG and ETV1- 7. Results: A total of 512 PC patient samples assayed between April 2018 and July 2022 were included in the analysis; 226 ETS fusions were present in 223 (43.6%) patients; 3 patients carried two fusions. There were 196 ERG fusions including 185 (36.1%) patients having the familiar TMPRSS2: ERG fusion and 8 (1.6%) patients with a SLC45A3: ERG fusion, which is known to be associated with particularly poor prognosis. ETV1, 4, and 5 fusions were found in 29 (5.7%) patients and involved 12 different 5’ partner genes, the most common of which were SLC45A3 (6 fusions) and CANT1 (4 fusions). The two partner genes were on the same chromosome in 192 fusions, with 190 requiring a single deletion to create the fusion and 2 requiring a single inversion. For the 34 fusions involving genes on different chromosomes, a single translocation would be sufficient to produce the fusion gene in 23 cases; in 11 cases more than one structural rearrangement would be required. A total of 66 (29.6%) fusions were not detected in the DNA sequencing data, including 40 (21.6%) TMPRSS2: ERG fusions. In addition, 16 fusions were detected only in the DNA data, as RNA could not be sequenced, including 13 TMPRSS2: ERG fusions. Conclusions: In addition to TMPRSS2: ERG, the Oncomap ExTra assay identified several low frequency ETS fusions, all of which could be used to assist patients and physicians to select appropriate treatments. The identification of ETS fusions appears to be limited when using only DNA sequencing. Oncomap ExTra RNA analysis identified 66 additional fusions, representing almost 30% of those present, not identified by whole-exome DNA sequencing, suggesting RNA plus DNA assays detect fusions more reliably than DNA-only assays.