Identification and characterization of RET fusions in advanced colorectal cancer

Abstract

There is an unmet clinical need for molecularly directed therapies available for metastatic colorectal cancer. Comprehensive genomic profiling has the potential to identify actionable genomic alterations in colorectal cancer. Through comprehensive genomic profiling we prospectively identified 6 RET fusion kinases, including two novel fusions of CCDC6-RET and NCOA4-RET, in metastatic colorectal cancer (CRC) patients. RET fusion kinases represent a novel class of oncogenic driver in CRC and occurred at a 0.2% frequency without concurrent driver mutations, including KRAS, NRAS, BRAF, PIK3CA or other fusion tyrosine kinases. Multiple RET kinase inhibitors were cytotoxic to RET fusion kinase positive cancer cells and not RET fusion kinase negative CRC cells. The presence of a RET fusion kinase may identify a subset of metastatic CRC patients with a high response rate to RET kinase inhibition. This is the first characterization of RET fusions in CRC patients and highlights the therapeutic significance of prospective comprehensive genomic profiling in advanced CRC.

Keywords: RET fusion kinase; RET kinase inhibitor; colorectal cancer; comprehensive genomic profiling.

Figure 1. Characterization of RET fusions in CRC patients

A. Frequency of RET fusions in unselected metastatic CRC patients as detected by NGS. B. Genetic and clinicopathologic characteristics of 6 patients harboring RET fusion kinase. nd = no data and WT = wild type. C. Fusion of CCDC6 exon 11 (green) containing the coiled-coil domain to RET exon 11 (red) containing the tyrosine kinase domain to generate CCDC6-RET fusion kinase. D. Fusion of NCOAT exon 9 (orange) containing the coiled-coil domain to RET exon 12 (red) containing the tyrosine domain to generate NCOAT-RET fusion kinase.

Figure 2. Inhibition of RET fusion-positive cancer cells viability by RET kinase inhibitors

A. Relative RET mRNA levels in Lc2/ad, SW48, SW48,-KRAS^12V, SW480, and SW620 cells as measured by quantitative RT-PCR analysis and normalized to SW620 cells using primers that recognized the RET kinase domain (Kinase: +) or flanked the CCDC6-RET fusion site (Fusion: +). B.-E. Lc2/ad, SW48, SW48,-KRAS^12V, SW480, and SW620 cells were treated with indicated concentrations of regorafenib B. vandetanib C. lenvatinib D. and erlotinib E. for 72 hours and cell survival was determined relative to 0.1% DMSO-treated controls (mean ± STD; n = 3).

Figure 3. Clinical response of RET fusion-positive CRC patient to regorafinib

A. Scans of Patient 1 harboring CCDC6-RET fusion kinase with diffuse liver metastases evident on PET/CT scan on coronal (left) and transverse (upper right) sections and MRI scan on transverse section (lower right). B. Serum CEA of patient 1 treated with regorafenib 40-80 mg daily. C. Serum LDH of Patient 1 treated with regorafenib 40-80 mg daily.

Figure 4. Colorectal cancer classification based on genomic biomarkers

A. Schematic classification of colorectal cancer based on genomic biomarkers with accompanying table showing fusion partner numbers (n) and gene point mutations available from the COSMIC database. B. Schematic of a Colorectal Cancer Genomic Protocol for stage IV CRC patients with assignment of therapy based on specific genomic biomarkers.