Identification of Synergistic, Clinically Achievable, Combination Therapies for Osteosarcoma

Abstract

Systemic therapy has improved osteosarcoma event-free and overall survival, but 30-50% of patients originally diagnosed will have progressive or recurrent disease, which is difficult to cure. Osteosarcoma has a complex karyotype, with loss of p53 in the vast majority of cases and an absence of recurrent, targetable pathways. In this study, we explored 54 agents that are clinically approved for other oncologic indications, agents in active clinical development, and others with promising preclinical data in osteosarcoma at clinically achievable concentrations in 5 osteosarcoma cell lines. We found significant single-agent activity of multiple agents and tested 10 drugs in all permutations of two-drug combinations to define synergistic combinations by Chou and Talalay analysis. We then evaluated order of addition to choose the combinations that may be best to translate to the clinic. We conclude that the repurposing of chemotherapeutics in osteosarcoma by using an in vitro system may define novel drug combinations with significant in vivo activity. In particular, combinations of proteasome inhibitors with histone deacetylase inhibitors and ixabepilone and MK1775 demonstrated excellent activity in our assays.

Figure 1. Single-agent activities of 54 therapeutic compounds screened against 5 osteosarcoma cell lines.

(a) Percent reduction in CellTiter-Glo signal indicating decreased ATP production at 72 hours post-drug treatment (normalized to vehicle controls). (b) Relative fold change in Caspase-Glo signal indicating cell death 24 hours post-drug treatment (normalized to vehicle controls). (c) Immunocytochemistry staining of caspase-3 activity in MG63 cells after 6-hour treatment with GSK923295A (7122 ng/mL) showing caspase-3 up-regulation in response to drug treatment. (d) Top 10 drug candidates with diverse mechanisms-of-action were selected from single-agent screening results to be further evaluated in 45 drug combinations for synergistic effects against pediatric osteosarcoma. N = 3 biological replicates. Scale bar: 20 μm.

Figure 2. Combination screening results.

(a) clustering results showing top combination picks based on FA. (b) FA of the top 6 combinations evaluated in 5 osteosarcoma cell lines using CT-Glo viability assay. (c) CI of the top 6 combinations evaluated in 5 osteosarcoma cell lines using non-constant ratio CI method. (d) single and combination drug-response curves for bortezomib+panobinostat and bortezomib + romidepsin combinations. (e), isobologram analysis of carfilzomib + panobinostat, carfilzomib + romidepsin, MK1775 + ixabepilone, and MK1775 + romidepsin at ED90. *Alternative bortezomib concentrations used with MG63 (5.00, 3.70, 2.74, 2.03, 1.50 ng/mL). **Alternative romidepsin concentrations used with 143B (150, 75.0, 37.5, 18.8, 9.40 ng/mL); Saos2 (25.0, 12.5, 6.25, 3.13, 1.56 ng/mL). ***Alternative ixabepilone concentrations used with U2OS; MNNG/HOS (100, 50.0, 25.0, 12.5, 6.25 ng/mL); MG63 (25.0, 12.5, 6.25, 3.13, 1.56 ng/mL). The exact drug concentrations used are summarized in Supplemental Table S3.

Figure 3. Order-of-addition analysis of the top 6 two-drug combinations assessed in 5 osteosarcoma cell lines FA (a) and CI (b) results using CT-Glo viability assay and non-constant ratio CI method, respectively.

C indicates concurrent treatment, 1: Tx1 given 24 hours prior to T × 2, 2: T × 2 given 24 hours prior to T × 1. (c) combination indices of carfilzomib:romidepsin, carfilzomib:panobinostat, MK1775:romidepsin, MK1775:ixabepilone in panel E averaged across ED80, ED85, ED90, ED95. (d) combination drug-response curves of bortezomib:romidepsin (top) and bortezomib:panobinostat (bottom) reported by CT-Glo. (e) combination drug-response curves of carfilzomib:romidepsin (3:20), carfilzomib-panobinostat (30:14), MK1775:romidepsin (98:20), MK1775:ixabepilone (98:8) reported by CT-Glo. (f) IC50 values of bortezomib:panobinostat and bortezomib:romidepsin combinations in panel A. Statistical comparison used two-way ANOVA with post-hoc Tukey test (d,e) and 2-tail paired t-test (c,f). *P < 0.05, **P < 0.01, ***P < 0.001).

Figure 4. Cytotoxicity of top 6 combinations evaluated in primary MSCs and primary MSC-derived osteoblasts in comparison with osteosarcoma cell lines.