Multivariate cellular image-based drug profiling to identify synergistic drug combinations for the treatment of triple-negative breast cancer

摘要

Triple-negative breast cancer (TNBC) is the most aggressive form of breast cancer. Due to the lack of known receptors, these tumours are resistant to tailored therapies and the standard of care is therefore limited to combinations of chemotherapy agents. These therapies are frequently plagued with intrinsic and acquired drug resistance, highlighting the need for novel and more effective combinations. However, cost- and time-limiting aspects of assessing the large number of possible combinations hamper their identification.This work presents an innovative and effective approach for the identification of pair-wise drug combination candidates to increase the efficacy of treatment for TNBC. The proposed methodology integrates high-content screening, computational and experimental biology. Cell phenotype alterations caused by treatment of TNBC cells with 55 FDA-approved and non-approved compounds were captured after fluorescent staining. Automatic image processing was applied to delineate single cells and 211 features describing cellular phenotypes were extracted, which provided the basis to compute compound profiles using a multivariate approach. Six combination pairs were identified based on similarities and dissimilarities between these profiles and their effects on TNBC cell growth were examined.Strong synergy was confirmed for the combination of the microtubule-targeting agent Vinblastine and the KSP/Eg5 motor protein inhibitor Monastrol in three independent TNBC cell lines. A modified combination containing Ispinesib, which has the same target as Monastrol, but better bioavailability and antitumour activity in patients with advanced and recurrent breast cancer also induced synergy. Cell cycle analysis revealed that the synergistic interaction between Vinblastine and Monastrol/Ispinesib was associated with an increase in cell cycle arrest at the G2/M transition, which translated into increased apoptotic cell death. Collectively these results show that the application of image-based drug profiling is a valuable time- and cost-effective tool for the identification of synergistic interactions, which led to the discovery of a new combination effective against TNBC cell lines. Although, the presented combinations were specifically identified for TNBC, this approach is not restrained to a specific cancer type, but may be extended to other hard-to-treat cancers.