Background: Sorafenib is an oral multikinase inhibitor that decreases tumor angiogenesis and proliferation. The antitumor efficacy and toxicity profiles of sorafenib vary among patients. Novel pathways and targets of sorafenib activity remain to be identified, and no predictive biomarkers of sorafenib activity exist to help guide clinicians. We aimed to identify novel genes associated with sorafenib activity by using an in vitro methodology based upon mouse genomics.

Methods: We profiled primary mouse embryonic fibroblasts (MEFs) from 32 inbred strains for sorafenib cytotoxicity utilizing high content imaging and simultaneous evaluation of cell health parameters. The 32 strains were genomically characterized previously (PMID: 21623374). MEFs were treated with ten concentrations (0-300 μM) of sorafenib, incubated for 72 h, and then fixed and stained. Cytochrome C was immunolabeled in the fixed cells and imaged using an automated fluorescence microscope. An image analysis software assessed sorafenib-induced cytochrome C release from mitochondria. Dose response curves were generated from data, and EC50 values for each strain were identified using a Brain-Cousens model. Genome-wide association mapping, using the EMMA algorithm, was performed on cytochrome C EC50 values to identify quantitative trait loci (QTLs) associated with sorafenib cytotoxicity. Approximately 277,000 single nucleotide polymorphisms were tested, and genomic loci with p-values ≤ 1.0×10−7 were selected for additional analyses.

Results: Interstrain EC50 variability among the 32 MEF strains was observed after 72 h (17.2-44.5 μM) sorafenib incubations. We identified one peak (chromosome 9 from 51-52 Mb; p = 1.0×10−8), which reached genome-wide significance and significantly associated with cytochrome C release. From this peak, we have identified candidate genes that may underlie variability in sorafenib-induced cytochrome C release from mitochondria. A total of nine genes expressed in MEF cells at mRNA level are present in this QTL. Of particular interest, the identified locus contains Rdx, a gene that encodes radixin. Radixin is a component of the ezrin-radixin-moesin-binding phosphoprotein-50 (EBP50) complex, and is conserved between humans and mice. In humans, radixin, as part of the EBP50 complex, has been shown to act as a tumor suppressor and to promote apoptosis in hepatocellular carcinoma by modulating β-catenin/E-cadherin (PMID: 23483729). It has also been shown to potentially enhance the metastatic potential of renal cell carcinoma (PMID: 20395446).

Conclusions: Our high-throughput cellular genetics approach has detected robust interstrain cellular differences in sorafenib activity, and identified one region, which reached genome-wide significance, that potentially associates with sorafenib-induced cytochrome C release from mitochondria. Functional validation of Rdx and other promising candidates is currently ongoing.

Citation Format: Daniel James Crona, Oscar Suzuki, O. Joseph Trask, Bethany Parks, Amber Frick, Timothy Wiltshire, Federico Innocenti. Identification of novel candidate genes associated with sorafenib cytotoxicity. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 5486. doi:10.1158/1538-7445.AM2015-5486