ABCB1, ABCG2, and PTEN determine the response of glioblastoma to temozolomide and ABT-888 therapy.

PURPOSE Little is known about the optimal clinical use of ABT-888 (veliparib) for treatment of glioblastoma. ABT-888 is a PARP inhibitor undergoing extensive clinical evaluation in glioblastoma, because it may synergize with the standard-of-care temozolomide (TMZ). We have elucidated important factors controlling ABT-888 efficacy in glioblastoma. EXPERIMENTAL DESIGN We used genetically engineered spontaneous glioblastoma mouse models and allograft models that were orthotopically transplanted into wild-type (WT) and Abcb1/Abcg2-deficient (KO) recipients. RESULTS ABT-888/TMZ is not efficacious against p53;p16(Ink4a)/p19(Arf);K-Ras(v12);LucR allografts in wild-type recipients, indicating inherent resistance. Abcb1/Abcg2 mediated efflux of ABT-888 at the blood-brain barrier (BBB) causes a 5-fold reduction of ABT-888 brain penetration (P < 0.0001) that was fully reversible by elacridar. Efficacy studies in WT and KO recipients and/or concomitant elacridar demonstrate that Abcb1/Abcg2 at the BBB and in tumor cells impair TMZ/ABT-888 combination treatment efficacy. Elacridar also markedly improved TMZ/ABT-888 combination treatment in the spontaneous p53;p16(Ink4a)/p19(Arf);K-Ras(v12);LucR glioblastoma model. Importantly, ABT-888 does enhance TMZ efficacy in Pten deficient glioblastoma allografts and spontaneous tumors, even in Abcb1/Abcg2 proficient wild-type mice. Loss of PTEN occurs frequently in glioblastoma (36%) and in silico analysis on patient with glioblastoma samples revealed that it is associated with a worse overall survival (310 days vs. 620 days, n = 117). CONCLUSIONS The potential of ABT-888 in glioblastoma can best be demonstrated in patients with PTEN null tumors. Therefore, clinical trials with ABT-888 should evaluate these patients as a separate group. Importantly, inhibition of ABCB1 and ABCG2 (by elacridar) may improve the efficacy of TMZ/ABT-888 therapy in all glioblastoma patients.