Kinases are integral components of diverse signaling cascades which can mediate microenvironment-derived resistance of tumor cells to apoptosis. We thus hypothesized that tumor cell-microenvironment interactions alter the anti-tumor activity of multi-targeted kinase inhibitors and could contibute to discordant efficacy results in conventional preclinical models which do not account for these interactions vs. in clinical studies.
We evaluated 6 multi-targeted kinase inhibitors initially developed to target IGF1R (XL-228), Akt/p70S6K (EXEL-6075), B-/C-RAF(EXEL-0094), CHK1/2 (XL844), met/VEGFR2(EXEL-6323) and JAK2 (XL019). Compounds were tested (0–20 μM, 48 hrs) against luciferase-expressing human multiple myeloma (MM) (n=8), mantle cell lymphoma (n=2) and breast cancer (n=1) cell lines. For each of these treatments, tumor cells were cultured in isolation or co-cultured with immortalized human bone marrow stromal cells (BMSCs, n=2); lung fibroblasts; hepatocytes; or brain astrocytes. Tumor cell viability was quantified by tumor cell compartment-specific bioluminescence imaging (CS-BLI). For each compound and tumor cell line, the area under the dose-response curve (AUC) was calculated as % of the AUC that corresponds to 100% tumor cell viability at all doses, to provide cumulative measure of drug activity across the range of tested drug concentrations. Differences in % AUCs in presence vs. absence of accessory cells were compared by 2-tailed paired t-test. The average value and skewness of distribution for log(10)IC50 of enzymatic activity for each kinase and inhibitor tested were calculated for quantitative comparison of multi-targeted nature of these compounds.
Co-culture of most tumor cell lines with accessory cells triggered resistance for 5 of 6 compounds tested, as evidenced by statistically significant (p<0.001 for all tests) increases in %AUC with vs. without co-culture with accessory cells (average %AUC differences of 4.7%, 23.4%, 6.2%, 9.7% and 10.8% for IGF1R, Akt/p70S6K, B-/C-RAF, CHK1/2; and met/VEGFR2 inhibitors, respectively, with corresponding 95% C.I. of 2.0–7.5%, 18.2–28.6%, 2.0–10.4%, 5.3–14.1%, and 3.9–17.6%). In contrast, co-culture with accessory cells increased the response of most tumor cell lines to JAK2 inhibitor (average AUC decrease of 13.8%, 95%CI 1.9–16.4, p<0.001). Results were consistent and statistically significant when analysis was restricted to MM cell lines only. For some inhibitors, heterogeneous responses of different tumor cells to accessory cells were noted. For example, for met/VEGFR2 and CHK1/2 inhibitors, 58.0% (40/69) and 60.0% (36/60) of experimental conditions, respectively, showed co-culture-induced resistance, while 31.9% (22/69) and 18.0% (11/60) of cocultures, respectively, showed sensitization to these inhibitors. High frequency of sensitization was observed in co-cultures of a bortezomib/Dex-resistant subline of MM1S cells treated with met/VEGFR2 inhibitor (63.6%, 7/11 of co-cultures tested), and in co-cultures with BMSCs treated with CHK1/2 inhibitor (29.6% of co-cultures tested). The degree of sensitization or resistance by accessory cells (%AUC difference with vs. without co-culture) did not correlate with quantitative measures of multi-targeted nature for each inhibitor.
Accessory cells from different tissues significantly modulated tumor cell responses to diverse multi-targeted kinase inhibitors. Co-cultures enhanced the anti-tumor activity of JAK2 inhibitor, but decreased the activity of the other compounds tested. These different responses did not correlate with how multitargeted each inhibitor was. The sensitization to the JAK inhibitor may specifically reflect increased dependency on JAK signaling for tumor cells when they interact with accessory cells. Accessory cell-induced resistance to some kinase inhibitors (e.g. IGF1R, Akt, CHK1–2) may account for observed differences in their preclinical vs. clinical efficacy. These results provide a rationale for preclinical testing of compounds against large panels of tumor cell lines and clinically-relevant accessory cell types to help develop therapeutics against tumors localized in specific tissue microenvironmental niches.
McMillin:Axios Biosciences: Equity Ownership. Mitsiades:Millennium Pharmaceuticals: Honoraria; Celgene: Honoraria; Novartis Pharmaceuticals: Honoraria; Bristol-Myers Squibb: Honoraria; Merck &Co.: Honoraria; Centocor: Honoraria; Arno Therapeutics: Honoraria; Amgen: Research Funding; AVEO Pharma: Research Funding; OSI: Research Funding; EMD Serono: Research Funding; Sunesis: Research Funding; Johnson & Johnson: Research Funding; PharmaMar: Licensing royalties Other; Axios Biosciences: Uncompensated Role as advisor, Uncompensated Role as advisor Other.
Asterisk with author names denotes non-ASH members.