Abstract 5035: Predictions of therapeutic doses of a B-Raf inhibitor, GSK2118436, based on exposure-response modeling of preclinical tumor biomarker and xenograft data

Background: GSK2118436 is a potent inhibitor of B-Raf, a signal transduction kinase in the MAPK pathway that can be mutated to become an oncogenic driver in several human cancers, including melanomas (70%). GSK2118436 inhibits in a dose dependant manner phoshorylation of ERK (pERK) in cell lines encoding B-RafV600E mutation leading to inhibition of cell proliferation and tumor regression. Purpose: To predict clinical relevant doses based on exposure-response relationship from preclinical pERK inhibition tumor xenograft data and exposure observed in first-time-in-human (FTIH) study. Methods: Tumor pERK and tumor volume measurements from nude mouse xenograft models, including colon carcinoma and melanoma encoding B-RafV600E mutation, were pooled for the analysis. GSK2118436 was administered using a wide range of oral (0.1 to 300 mg/kg QD, 5 to 300 mg/kg BID) and continuous IV (3 to 100 mg/kg/day) doses. Tumor pERK inhibition was characterized as a function of GSK2118436 concentration using an inhibitory Emax model. For tumor regression, tumor volume was described using a linear growth model with tumor death rate expressed as a function of GSK2118436 concentration. The analyses were conducted using nonlinear mixed effects modeling. To extrapolate to clinical exposure, GSK2118436 concentrations in mouse were corrected for blood:plasma ratio. Results: The relationship between preclinical tumor pERK inhibition and exposure was described using a mean (90% CI) plasma IC50 of 18.5 (4.3-32.4) ng/mL (35.5nM). In preclinical models, tumor regression was observed at 30mg/kg QD orally and 3mg/kg IV and was dependent on dosing regimen. Based on exposure observed in IV studies, tumor regression was associated with >70% pERK inhibition. Tumor growth was characterized with a mean rate of 1.04 mm3/hr and was inhibited by GSK2118436 with a mean (90% CI) plasma IC50 of 28 (5.9-50.4) ng/mL (53.8nM) for tumor death rate. The model was successful in describing data observed with different modes of administration. In the FTIH study, inhibition of tumor pERK was exposure-related with >80% inhibition predicted at doses ≥100 mg BID. Consistent with target inhibition, administration of the recommended Phase 2 dose of 150 mg BID resulted in clinical activity in subjects with B-Raf-mutant melanoma. Mean exposure at trough exceeded 100 ng/mL (192nM), which is >3-fold the target exposure for tumor growth inhibition (53.8nM) based on the preclinical model. Conclusions: The exposure-respons