RNA Sequencing Identifies Transcriptionally Viable Gene Fusions in Esophageal Adenocarcinomas

Abstract

Esophageal adenocarcinoma is a deadly cancer with increasing incidence in the United States, but mechanisms underlying pathogenesis are still mostly elusive. In addressing this question, we assessed gene fusion landscapes by comprehensive RNA sequencing (RNAseq) of 55 pretreatment esophageal adenocarcinoma and 49 nonmalignant biopsy tissues from patients undergoing endoscopy for Barrett's esophagus. In this cohort, we identified 21 novel candidate esophageal adenocarcinoma-associated fusions occurring in 3.33% to 11.67% of esophageal adenocarcinomas. Two candidate fusions were selected for validation by PCR and Sanger sequencing in an independent set of pretreatment esophageal adenocarcinoma (N = 115) and nonmalignant (N = 183) biopsy tissues. In particular, we observed RPS6KB1-VMP1 gene fusion as a recurrent event occurring in approximately 10% of esophageal adenocarcinoma cases. Notably, esophageal adenocarcinoma cases harboring RPS6KB1-VMP1 fusions exhibited significantly poorer overall survival as compared with fusion-negative cases. Mechanistic investigations established that the RPS6KB1-VMP1 transcript coded for a fusion protein, which significantly enhanced the growth rate of nondysplastic Barrett's esophagus cells. Compared with the wild-type VMP1 protein, which mediates normal cellular autophagy, RPS6KB1-VMP1 fusion exhibited aberrant subcellular localization and was relatively ineffective in triggering autophagy. Overall, our findings identified RPS6KB1-VMP1 as a genetic fusion that promotes esophageal adenocarcinoma by modulating autophagy-related processes, offering new insights into the molecular pathogenesis of esophageal adenocarcinomas. Cancer Res; 76(19); 5628-33. ©2016 AACR.

Figure. 1. RPS6KB1–VMP1 fusion in EACs

(A) (Top) conserved protein domains in RPS6KB1 and VMP1, along with their length in amino acids (aa). (Middle) native gene exons included in the RPS6KB1–VMP1 fusion transcript. (Bottom) PCR amplified product spanning the E1-E7 isoform of the RPS6KB1–VMP1 fusion-junction, along with the Sanger sequencing chromatogram of the fusion-junction in representative EAC and matched SQ tissues used in the discovery RNAseq study. (B) Representative agarose gel images of RPS6KB1–VMP1 (E1-E7) fusion-junction PCR product from matched SQ and fusion-positive primary EAC tissue RNAs (N=11 pairs), included in the validation (val) cohort. Beta-2-microglobulin (B2M) was used as an endogenous RNA control. (C) Kaplan-Meier survival curves for Stage III EACs with or without the RPS6KB1–VMP1 (E1-E7) fusion.

Figure 2. Assessing the protein-coding potential of RPS6KB1–VMP1 fusion in mammalian cells

(A) Representative gel image showing the expression of full-length RPS6KB1–VMP1 fusion transcripts (E1-E7, E4-E7) in FLO-1 cell line RNA. DNA ladder (L). (B) Western blot analysis using an N-terminal (inverted Y) anti-RPS6KB1 primary antibody in FLO-1 and Cos7 cells ectopically expressing full-length RPS6KB1–VMP1 fusion transcripts or vector. (Right) radiographic image of FLO-1 and vector-carrying Cos7 cells at ~5 minute exposure (Long-exposure). (C) Western blot analysis using a negative control C-terminal (inverted Y) anti-RPS6KB1 primary antibody in Cos7 and FLO-1 cells from above showing no detection of the fusion protein.

Figure 3. Functional characterization of the RPS6KB1–VMP1 fusion

(A) (Top) effect of RPS6KB1–VMP1 fusion transcripts on CP-A cell growth. Y-axis represents the average confluence values of two independently-derived stable clones per experimental arm, with 12 replicates each, normalized to time zero (X-axis). The error bars represent the standard error of the means. (Bottom) PCR analyses of fusion transcripts in RNA derived from CP-A cells transfected with control gene (TUBB) or fusion transcripts, and Western blot analysis demonstrating fusion protein detection with the N-terminal anti-RPS6KB1 antibody. (B) (Top) localization of ectopically expressed RPS6KB1–VMP1 and VMP1 proteins in Cos7 cells, detected by anti-V5 antibody (green fluorescence), and endogenous LC3B protein, a marker of autophagosome (red fluorescence). Note the diffuse staining pattern of RPS6KB1–VMP1 fusion protein, as opposed to the specific co-localization of VMP1 with LC3B protein (scale bar, 10μm). (Bottom) Western blot analysis demonstrating the protein expression of ectopically expressed VMP1 and RPS6KB1–VMP1 fusion in Cos7 cells. (C) Immunoprecipitation (IP) and immunoblotting (IB) analysis in Cos7 cells demonstrating the lack of interaction of RPS6KB1–VMP1 (E1-E7) fusion protein with Beclin1, as compared to the RPS6KB1–VMP1 (E4-E7) fusion or VMP1 protein. Western blot analysis demonstrates equivalent amounts of endogenous Beclin1 protein present in Cos7 total cell lysates (TCL) across treatment groups. (D) (Top) Western blot analysis using anti-LC3A/B antibody in Cos7 cells, following nutrient-depletion and chloroquine treatment, showing significant induction of LC3A/B-II product in VMP1 expressing cells, as compared to Cos7 cells carrying the RPS6KB1–VMP1 (E1-E7) fusion. Also note the reduction in the protein levels of LC3A/B-I in RPS6KB1–VMP1 (E1-E7) expressing cells as compared to vector control. Western blot analysis demonstrates protein expression of ectopically expressed VMP1 and RPS6KB1–VMP1 fusion transcripts in Cos7 cells. (Bottom) Western blot analysis using anti-LC3A/B antibody in CP-A cells, stably expressing RPS6KB1–VMP1 fusion or control TUBB transcripts, after nutrient-depletion and chloroquine treatment. Note the significant induction of LC3A/B-II product in TUBB expressing CP-A cells, as compared to the CPA-cells carrying the RPS6KB1–VMP1 (E1-E7) fusion.