A pipeline for prioritizing and evaluating multidrug regimens for Mycobacterium abscessus

点击翻译按钮获取中文摘要

PROJECT SUMMARY Mycobacterium abscessus complex (MABSC) is a group of closely related nontuberculous mycobacteria, which primarily cause infections in persons with underlying lung disease, compromised immunity or other risk factors. MABSC organisms are intrinsically resistant to many antibiotics and treatment requires many months of multiple antibiotics, often with significant toxicity and low cure rates. One of the difficulties in optimizing the treatment of MABSC infections is the lack of a robust pipeline for the identification of multidrug regimens that may be more efficacious than the current standard of care. A strategy for prioritizing multidrug regimens is necessary, because the number of combinations to test can be overwhelming. For example, with even just 12 drugs to select from there are 220 potential 3 drug combinations. This project proposes to develop INDIGO (INferring Drug Interactions using chemo-Genomics and Orthology), a machine learning modeling approach, to identify drug combinations with strong potential synergistic activity against MABSC and to test these regimens in an animal model. The specific aims of this project are: 1) Build and refine an INDIGO-MABSC model through the collection of transcriptomic data from drug treated MABSC cultures, subsequent experimental verification of antibiotic synergy and antagonism followed by model refinement with iterative rounds of model building and testing; 2) Assess the efficacy of top predicted synergistic multidrug regimens in a zebrafish model for MABSC infection. INDIGO modeling is modular, so as new agents become available they may be incorporated into the model to make predictions for how these agents may be best combined into multidrug regimens. The overall goal of this project is to provide a streamlined pipeline for identifying regimens that can be pushed into advanced pre-clinical and clinical trials for treatment of MABSC infections.