Tumor Cell-Dependent Differences in Stroma-Induced Changes to Bortezomib Sensitivity.

Abstract 2468

In multiple myeloma (MM) and other hematologic malignancies, bone marrow stromal cells (BMSCs) confer resistance to diverse conventional or investigational therapeutics. During the last decade, data from many groups have concurred that the in vitro anti-MM activity of the proteasome inhibitor bortezomib is very similar in the presence and absence of BMSCs, including primary and immortalized BMSCs. These well-validated observations have supported the notion that novel, more effective, therapies for the treatment of MM should ideally be, similarly to bortezomib, capable of overcoming the protective effect of BMSCs. Interestingly, however, we have observed that primary CD138+ MM tumor cells isolated from patients with clinical refractoriness to bortezomib occasionally exhibit substantial in vitro response to clinically achievable concentrations of this drug. We therefore hypothesized that, under certain previously under-explored experimental settings, BMSCs may alter the threshold of MM cell response to bortezomib-induced apoptosis. To address this hypothesis in conditions that better simulate the clinical context, we conducted compartment-specific bioluminescence imaging (CS-BLI) assays to evaluate the effect of bortezomib on tumor cells co-cultured with BMSCs for different time periods prior to bortezomib administration. We observed that prolonged tumor-stromal co-culture (48–96hrs) prior to initiation of bortezomib treatment did not affect drug sensitivity for several MM cell lines (OPM2, H929, UM9, KMS11, KMS18 and RPMI-8226) tested. Prolonged co-culture of OPM1, RPMI-8226-Dox40, OCI-My5, KMS12BM and KMS18 cells prior to bortezomib treatment enhanced its activity. Importantly, extended co-culture of MM cell lines MM.1S and MM.1R with BMSCs prior to drug treatment induced significant attenuation of their response to bortezomib, as evidenced by 2–3 fold increase of IC₅₀ values in several independent replicate experiments and a mean % area under the bortezomib dose response curve (AUC) of 5.82% vs 14.10% in the absence vs. presence of BMSCs, respectively (p=0.0079). Consistent with these in vitro results, heterotypic s.c. xenografts of Luc+ MM.1S cells mixed with Luc- BMSCs did not show statistically significant reduction in MM burden with bortezomib treatment (0.5 mg/kg s.c. twice weekly for 5 weeks) compared to vehicle-treated controls (p=0.1320), as quantified by bioluminescence imaging. In contrast, the same dose and schedule of bortezomib treatment significantly suppressed tumor burden, compared to vehicle-treated controls, of monotypic s.c. xenografts of Luc+ MM.1S cells in SCID mice (p=0.0022), as in prior experience. To evaluate the molecular mechanisms of cell non-autonomous decrease in MM cell response to bortezomib, we compared the transcriptional profiles of MM.1S cells in extended co-cultures with HS-5 BMSCs vs. MM.1S cells cultured in isolation. These studies identified a distinct transcriptional signature of stroma-induced transcripts, including several (e.g. PSMC3, ITGB7, FOS, ALDH1L2) for which transcript expression higher than the median levels for refractory MM patients correlated with shorter overall survival (p<0.02, log-rank tests) after treatment with bortezomib. These observations highlight the notion that tumor cell responses to a given agent in the presence of non-malignant stromal cells can exhibit substantial qualitative and quantitative variation, depending on the specific tumor cell type tested, as well as the particular stromal cell population and conditions of the co-culture. Our findings highlight the need to apply combinatorial high-throughput scalable platforms, such as CS-BLI, to evaluate the different permutations of interactions between tumor cells, non-malignant accessory cells of the microenvironment and administered therapeutics. This study also provides a comprehensive functional oncogenomic framework to identify prognostically relevant molecular mediators of stroma-induced resistance to therapy in MM.