Antimicrobial resistance challenges the sustainability of healthcare systems and results in substantial economic losses worldwide. This issue is further aggravated by paucity of new drugs and treatment options. In this study, an artificial intelligence (AI)-derived platform termed IDentif.AI is utilized to accelerate the development of effective therapeutic options for carbapenem-resistant Enterobacteriaceae (CRE). Twelve Food and Drug Administration-approved drugs are selected and the in vitro inhibitory efficacy of 155 combinations consisting of various drugs is determined at different concentrations against both Klebsiella pneumoniae and Escherichia coli. Correlating these experimental data via an AI-derived relationship, IDentif.AI rapidly determines a ranked list of drug combinations in the search space of over half a million possible combinations. Meropenem is found to strongly synergize with low doses of the anticancer drug bleomycin, showing broad-spectrum, bactericidal activity against nine isolates across three CRE species in rich and minimal media with no synergistic cytotoxicity on mammalian cells. Synergy is also detected between bleomycin and other key carbapenems in clinical use (imipenem, ertapenem). Bleomycin/carbapenem appears to be a promising combination therapy for treating various CRE infections. IDentif.AI shows profound capability of identifying pan-active drug combinations against a family of bacteria through surveying strains from different species in parallel.

中文翻译：

---

通过人工智能 (AI) 驱动的平台发现抗碳青霉烯类耐药肠杆菌 (CRE) 的广谱再利用药物组合

抗菌素耐药性挑战了医疗保健系统的可持续性，并在全球范围内造成了巨大的经济损失。由于缺乏新药和治疗选择，这个问题进一步加剧。在这项研究中，利用人工智能 (AI) 衍生平台 IDentif.AI 来加速开发针对碳青霉烯类耐药肠杆菌(CRE) 的有效治疗方案。选取12种美国食品药品监督管理局批准的药物，测定不同药物组成的155种组合在不同浓度下对肺炎克雷伯菌和大肠杆菌的体外抑制效果。IDentif.AI 通过人工智能衍生的关系将这些实验数据关联起来，在超过 50 万种可能的组合的搜索空间中快速确定药物组合的排名列表。研究发现美罗培南与低剂量的抗癌药物博来霉素具有强烈的协同作用，在丰富和基本培养基中对三种 CRE 物种的 9 种分离株显示出广谱杀菌活性，但对哺乳动物细胞没有协同细胞毒性。博来霉素和临床使用的其他关键碳青霉烯类药物（亚胺培南、厄他培南）之间也发现了协同作用。博莱霉素/碳青霉烯似乎是治疗各种 CRE 感染的一种有前景的联合疗法。IDentif.AI 通过并行调查不同物种的菌株，展现出识别针对某一细菌家族的泛活性药物组合的强大能力。