Our laboratory has a long-standing interest in non-Pgp mediated mechanisms of drug resistance, having established several cell line models of resistance focusing on the ABC half-transporter ABCG2. We successfully cloned ABCG2 from a mitoxantrone-resistant colon cancer cell line, S1-M1-80, that exhibited an ATP-dependent reduction in drug accumulation. Comprised of 6 transmembrane domains and a single ATP binding domain, it is thought that dimerization is required for activity, and that the gene encodes a half-transporter molecule. Overexpression of ABCG2 renders cells resistant to mitoxantrone and to the camptothecins, topotecan and SN-38 (the active metabolite of irinotecan). Both substrates and inhibitors of ABCG2 have been discovered at an accelerating pace, and the variation in inhibitors rivals that described for P-glycoprotein. There is increasing evidence supporting a role for ABCG2 in oral absorption of pharmacologic agents. We identified a naturally occurring mutation in ABCG2 (R482T; R482G) that alters substrate and inhibitor specificity; and then carried out a sequence analysis of 90 DNA samples representing a global genetic diversity and identified single nucleotide polymorphisms in ABCG2. We then established transfectants in HEK293 cells. Using the wildtype background (R482), clones expressing V12M, Q141K and D620 N were established. Clones with comparable levels of surface expression were selected, and impaired transport was observed in clones bearing the Q141K. Since gastrointestinal absorption of topotecan has been related to ABCG2 expression in the intestinal epithelium, a clear implication of this work is that the Q141K SNP could be associated with increased oral topotecan uptake. To the extent that ABCG2 is involved in drug excretion, this SNP could increase exposure to substrates such as irinotecan and topotecan. In an attempt to identify the mechanism of dimerization of ABCG2, our laboratory has studied a GXXXG dimerization motif in the transmembrane helix 1, finding that while it is critical for normal transport activity, normal cross-linking is retained in cells expressing glycine to leucine mutations at amino acids 406 and 410. Further mutational analysis is ongoing, generating mutations in residues identified by analysis of homologues or by molecular modeling, based on collaborative studies with Michael Dean and Di Xia, respectively. Our goal is to identify residues involved in the dimerization. In collaboration with Balasz Sarkadi, we have generated Sf9 insect vectors to allow transfection in the high expression insect system. This system is more tolerant of misfolded protein and has already been shown to generate functional ABCG2 molecules. This system will allow co-immunoprecipitation studies that are needed to provide confirmation of the impact of our mutations on dimerization. Quantitative mitoxantrone efflux assays used in characterizing the mutants and the SNPs were developed in part for use in clinical samples. We were able to convincingly show a linear relationship between mitoxantrone efflux and both mRNA and protein expression. Cells are incubated for 30 min at 378C in mitoxantrone with or without the ABCG2-specific inhibitor, fumitremorgin C (FTC). The cells are then washed and incubated again at 378C for 1 hr in mitoxantrone-free medium continuing with or without FTC. The cells are then washed and fluorescence quantitated by flow cytometry. The difference in the post-efflux peaks with or without FTC, the """"""""channel -shift value"""""""" is linearly related to both mRNA and surface protein expression in the cells selected for drug resistance. Since mitoxantrone is also a substrate for P-glycoprotein, we were interested in pheophorbide a when it was first described by Shinkel et al as the agent that produced phototoxicity in ABCG2 knockout mice. We reasoned that since phototoxicity had not been observed in the intensively studied Pgp knockout mice, pheophorbide a might be an ABCG2-specific substrate. This could allow more accurate clinical detection of ABCG2. We tested pheophorbide a selected cell lines and in the HEK 293 clones transfected with pcDNA vectors encoding ABCG2 with wild type, mutant and SNP sequences and found a tight correlation between cell surface expression as measured by 5D3 antibody and pheophorbide a efflux as measured by inhibition with FTC. No transport was observed in cells expressing Pgp or MRP. In addition to the studies with pheophorbide a, we have also worked to define new ABCG2 substrates and inhibitors. Several photosensitizers used in photodynamic therapy and structurally related to pheophorbide a have been found to be substrates. In collaboration with Yves Pommier, we evaluated the role of ABCG2 in mediating resistance to BN80915, a novel homocamptothecin in early clinical trial. We demonstrated that the difluoro derivative, BN80915, is a substrates for ABCG2, but that BN80915 was much more poorly transported than SN-38 (active metabolite of irinotecan) or topotecan. These studies lend support for its potential efficacy in tumors expressing ABCG2, thus circumventing ABCG2-mediated resistance. Finally, in order to evaluate clinical samples for ABCG2 expression, we have developed an immunohistochemical assay using a polyclonal antibody that we generated. With this assay, or with the flow cytometric assay with mitoxantrone or pheophorbide a, we have the ability to evaluate the potential role of ABCG2 in clinical drug resistance. Our ultimate goal is reversal of drug resistance. In collaboration with Michael Dean and the CCR Target Development Group, ABCG2-overexpressing cells have been set up to screen for inhibitors. This is an important undertaking since the potential ability to modulate oral drug absorption will be important whether or not ABCG2 proves important in oncologic drug resistance.
Showing the most recent 10 out of 33 publications
