Abcg2 Expression, Function And A Snp Variant In The Nci Drug Screen Cell Lines

Proc Amer Assoc Cancer Res, Volume 47, 2006 609 ABCG2 is a member of the ATP-binding cassette family of xenobiotic transporters. Expression can be induced by exposure to cytotoxic substrates, leading to drug resistance. High expression of ABCG2 in a 'side population’ of cells –potentially representing stem cells –may characterize the early stem cell phenotype, and explain multidrug resistance in these cells. A number of non-synonymous SNPs have been found in the ABCG2 gene, including the Q141K variant which has a high frequency in normal volunteers. Different researchers have identified the Q141K variant with a loss of membrane expression, reduced drug efflux, and decreased ATPase activity. The NCI Anticancer Drug Screen (NCI-ADS) has been used successfully to identify new cancer therapies, novel cellular targets, common mechanisms of action, and mechanisms of drug resistance. For example, the 60-cell line profile of both expression and function of P-glycoprotein (Pgp) was used to identify Pgp substrates and inhibitors through correlation with cytotoxicity patterns. We pursued a similar strategy by using measured expression and function of ABCG2 in the NCI-ADS cell lines. We also explored what effect the Q141K SNP had on transporter function. ABCG2 expression was determined by RT-PCR. A wide range in expression level was found, with high levels in H460, A549, RPMI 8226, MDA-MB 231, and HT29 cells. Function was measured as the efflux of the fluorescent substrate, pheophorbide a (PhA), detected by flow cytometry with and without the ABCG2 inhibitor fumitremorgin C. High function was found in H460, A549, RPMI 8226, H23, KM12, HCC 2998, and SF-295 cells. We found a statistically significant correlation between expression and function across the cell lines (r2 = 0.67). The expression and function of cell lines containing the Q141K variant were compared to those with a wildtype ABCG2 sequence. Ten lines have the Q141K variant, with two lines also containing the 944-949 deletion variant. Plasma membrane efflux of PhA as a function of expression was calculated and categorized by SNP variant. There was a strong trend towards lower function in the lines with the Q141K SNP. When the expression and function of ABCG2 was re-plotted with just wildtype cell lines, a stronger correlation was found (r2 = 0.80). The functional data from all 60 cell lines was used as a profile to probe the NCI-ADS database using the COMPARE program, and Pearson correlation coefficients were generated. Positive correlations identified possible substrates while negative correlations identified potential inhibitors. We identified 12 agents with a positive correlation above 0.50, and 11 agents with a negative correlation above 0.60. Experiments are ongoing to determine whether these agents are substrates and inhibitors, and final results will be presented. Further preclinical studies will be necessary to determine whether ABCG2 inhibition is an effective antitumor strategy, including targeting cancer stem cells.