Identification of a RAS-activating TMEM87A-RASGRF1 Fusion in an Exceptional Responder to Sunitinib with Non-Small Cell Lung Cancer

Abstract

Purpose: We pursued genomic analysis of an exceptional responder with non-small cell lung cancer (NSCLC) through a multi-platform effort to discover novel oncogenic targets.

Experimental design: In this open-label, single-arm phase II study (NCT01829217), an enriched cohort of patients with advanced NSCLC was treated with the multi-kinase inhibitor sunitinib. The primary endpoint was objective response rate. Tissue was collected for multi-platform genomic analysis of responders, and a candidate oncogene was validated using in vitro models edited by CRISPR-Cas9.

Results: Of 13 patients enrolled, 1 patient (8%), a never smoker, had a partial response lasting 33 months. Genomic analysis of the responder identified no oncogenic variant using multi-platform DNA analysis including hotspot allelotyping, massively parallel hybrid-capture next-generation sequencing, and whole-exome sequencing. However, bulk RNA-sequencing (RNA-seq) revealed a novel fusion, TMEM87A-RASGRF1, with high overexpression of the fusion partners. RASGRF1 encodes a guanine exchange factor which activates RAS from GDP-RAS to GTP-RAS. Oncogenicity was demonstrated in NIH/3T3 models with intrinsic TMEM87A-RASGRF1 fusion. In addition, activation of MAPK was shown in PC9 models edited to express this fusion, although sensitivity to MAPK inhibition was seen without apparent sensitivity to sunitinib.

Conclusions: Sunitinib exhibited limited activity in this enriched cohort of patients with advanced NSCLC. Nonetheless, we find that RNA-seq of exceptional responders represents a potentially underutilized opportunity to identify novel oncogenic targets including oncogenic activation of RASGRF1.

©2020 American Association for Cancer Research.

Figures

Figure 1:. Many genotyping methods will miss fusions in genes not known to be cancer related.

(A) Allelotyping for a panel of key recurrent variants is usually limited to detection of specific coding (exonic) mutations in known cancer genes. (B) Targeted next-generation sequencing (NGS) can detect mutations as well as some fusions through sequencing most exons of cancer genes as well as select introns. (C) Whole exome sequencing (WES) covers all coding regions across the genome, but lack of intron coverage results in fusions going undetected. (D) Through sequencing the RNA transcripts, bulk RNA sequencing has the additional ability to detect fusions in genes not known to be cancer related.

Figure 2:. Objective response to sunitinib therapy.

(A) Waterfall plot showing best percent change from baseline in tumor size, in patients with a RET rearrangement (red) and without (black). (B) One patient had a partial response, with a best response at 33 month follow-up of 51% diameter decrease (19.2 mm to 9.5 mm, short axis).

Figure 3:. Gene expression of the novel fusion TMEM87A-RASGRF1.

(A) TMEM87A expression in the responder was higher than all 551 lung squamous cell carcinoma (LUSC) and 594 lung adenocarcinoma (LUAD) samples from the Cancer Genome Atlas (TCGA). FPKM; fragments per kilobase of exon model per million reads mapped. (B) RASGRF1 expression in the responder was higher than 99.4% of TCGA samples.

Figure 4:. Oncogenicity of TMEM87A-RASGRF1 fusion in CRISPR-edited models.

(A) RASGRF1, one of the guanine nucleotide exchange factors (GEF), activates Mitogen-activated Protein Kinase (MAPK) pathway. GAP; GTPase activating protein. (B) TMEM87A-RASGRF1 fusion gene is caused by duplication of fragments in chromosome 15. This fusion gene lacks exons 1–8 of RASGRF1, thus loses the N-terminal regulatory Pelckstrin homology 1 (PH1) domain. CC; Coiled coil domain, IQ; Isoleucine Glutamine motif, DH; Dbl homology domain, REM; Ras exchanger motif, CDC25H; CDC25 homology domain. (C) TMEM87A-RASGRF1 fusion was confirmed by Sanger sequence in bulk PC9 and NIH3T3 cell lines edited by CRISPR-Cas9. (D) Focus formation assay after 5 weeks’ culture showed foci with marked pile-up in bulk NIH/3T3TMEM87A-RASGRF1, whereas parental NIH/3T3 were inhibited to grow when they became confluent. (E) Cell viability assay after 72 hours’ treatment showed that parental PC9 cells were sensitive to EGFR tyrosine kinase inhibitor osimertinib, but single clones from PC9 with TMEM87A-RASGRF1 were resistant.

Figure 5:. MAPK activation of TMEM87A-RASGRF1 fusion in CRISPR-edited models.

(A) Phospho-receptor tyrosine kinase (RTK) array in PC9TMEM87A-RASGRF1 clone 1 showed increased expression of phospho-EGFR. No other RTKs related to sunitinib were unregulated regardless of osimertinib treatment. (B) Quantitative RT-PCR showed increased expression of RASGRF1 and KITLG, a ligand for KIT in PC9TMEM87A-RASGRF1 clone 1. No other ligands of sunitinib-related RTKs were upregulated. (C) Screening with drugs targeting sunitinib-related RTKs or MAPK pathway was performed in PC9TMEM87A-RASGRF1 clone 1. Relative viability compared with control (treated with 0.5μM osimertinib) was shown. MAPK inhibitors were effective in the presence of control osimertinib. (D) Western blot analyses of parental PC9 and PC9TMEM87A-RASGRF1 clone 1 were performed after treatment with 0.5μM osimertinib, 10nM trametinib, or 0.5μM RAF709 for 48 hours. PC9TMEM87A-RASGRF1 clone 1 maintained MAPK signals without inducing apoptosis in the presence of osimertinib, which indicated that TMEM87A-RASGRF1 activates MAPK pathways. These MAPK signals can be overcome by combination of MAPK inhibitors in the presence of osimertinib.