Background: Although combination treatment has been a useful strategy for circumventing acquired resistance (AQR) to single agent treatment, AQR also occurs to combination treatment. Nonetheless, models of AQR have thus far been restricted to single agent treatments. In this study, we tested whether models of AQR to combination treatment can be generated, and whether the determinants of AQR to combination treatment differ to those of AQR to single agent treatment.

Methods: HCT116 colorectal cancer (CRC) cells with KRAS and PIK3CA mutations were exposed continuously with combined (1 IC50 dose of each agent) or single agent (2 doses of IC25) concentrations of MEK (AZD6244) and PI3K (BKM120) inhibitors, or DMSO (parental). IC50 measurements were assessed by MTS assay after 72 hours treatment. Combination Indices (CI) was calculated by the method of Chou and Talalay, where CI>1 indicates antagonism, <1 synergy, and = 1 additivity. Phenotypic analysis was performed using standard techniques. Differentially expressed genes were determined using Affymetrix Gene 1.0ST analysis.

Results: HCT116 cells with AQR to combination treatment were obtained after 6 months (CI@fu0.5 of “HCT116CR” cells 1.8±0.1 vs parental 0.23±009, p=0.04). HCT116 cells with AQR to single agent treatment with AZD6244 (IC50 of “HCT116AR” 56.1±0.04µM vs parental 4.3±0.04µM; p=0.03) and BKM120 (IC50 of “HCT116BR” 3.1±0.21µM vs parental 1.0±0.38µM; p=0.007) were also generated. The AQR in HCT116CR cells was associated with a significant reduction in apoptosis and wound healing response after combination treatment compared to HCT116 parental, HCT116AR and HCT116BR cells. Cell cycle phase distribution analysis revealed an inability of the compounds to elicit a G1 arrest in the single agent or combination AQR cells. Basal level of phosphorylated AKT were elevated in HCT116BR and HCT116CR cells, while phosphorylated ERK levels were increased in all the three resistant lines. In addition, there was a significant decrease in phosphorylated 4EBP1 in HCT116CR cells. In cross-resistance studies, no differences in the IC50s of 5FU or oxaliplatin were observed in any of the cell lines tested, suggesting the absence of a multi-drug resistance phenotype. HCT116AR and HCT116BR cells displayed cross resistance to other MEK or PI3K inhibitors, while combination treatment with MEK and PI3K inhibitors remained synergistic in these cells. HCT116CR cells were also resistant to single agent MEK and PI3K inhibitor treatment and to an alternative combination of a MEK inhibitor (GDC0973) and PI3K inhibitor (BYL719). However, HCT116CR cells were ∼20, 11 and 8-fold more sensitive than HCT116 parental cells to MK2206 (Akt), dual PI3K/mTOR (BEZ235; PI103) and mTOR (KU0063794) inhibitors respectively. Gene expression analysis revealed significant overexpression of KRAS and AKT3 in HCT116AR and HCT116BR cells compared to HCT116 cells, consistent with cross resistance data obtained with MEK and PI3K inhibitors. In HCT116CR cells, KRAS expression remained unchanged, however an overexpression of ET-1 (endothelin 1, 33-fold) was observed that was not observed in HCT116AR and HCT116BR cells.

Conclusions: AQR to synergistic combinations can evolve in cancer cells, and the mechanisms of this combination treatment AQR may be distinctly different from those for single agent AQR. ET-1 may be a specific mediator of AQR to combination PI3K and MEK resistance in CRC.

Citation Format: Bhaskar Bhattacharya, Sarah Low, Dilys Chia, King Xin Koh, Mei Ling Chong, Touati Benoukraf, Richie Soong. Determinants of acquired resistance to combination treatment with MEK and PI3K inhibitors. [abstract]. In: Proceedings of the AACR Precision Medicine Series: Drug Sensitivity and Resistance: Improving Cancer Therapy; Jun 18-21, 2014; Orlando, FL. Philadelphia (PA): AACR; Clin Cancer Res 2015;21(4 Suppl): Abstract nr B02.