Optimization of Benzoxazinorifamycins to Minimize hPXR Activation for the Treatment of Tuberculosis and HIV Coinfection

Tuberculosis (TB) is one of the most significant world health problems, responsible for 1.5 M deaths in 2020, and yet, current treatments rely largely on 40 year old paradigms. Although the rifamycins (RIFs), best exemplified by the drug rifampin (RMP), represent a well-studied and therapeutically effective chemotype that targets the bacterial RNA polymerase (RNAP), these agents still suffer from serious drawbacks including the following: 3-9 month treatment times; cytochrome P450 (Cyp450) induction [particularly problematic for human immunodeficiency virus-Mycobacterium tuberculosis (MTB) co-infection]; and the existence of RIF-resistant (RIFR) MTB strains. We have utilized a structure-based drug design approach to synthesize and test 15 benzoxazinorifamycins (bxRIFs), congeners of the clinical candidate rifalazil, to minimize human pregnane X receptor (hPXR) activation while improving potency against MTB. We have determined the compounds' activation of the hPXR [responsible for inducing Cyp450 3A4 (CYP3A4)]. Compound IC50s have been determined against the wild-type and the most prevalent RIFR (β-S450L) mutant MTB RNAPs. We have also determined their bactericidal activity against "normal"replicating MTB and a model for non-replicating, persister MTB. We have identified a minimal substitution and have probed larger substitutions that exhibit negligible hPXR activation (1.2-fold over the dimethyl sulfoxide control), many of which are 5- to 10-fold more potent against RNAPs and MTB than RMP. Importantly, we have analogues that are essentially equipotent against replicating MTB and non-replicating persister MTB, a property that is correlated with faster kill rates and may lead to shorter treatment durations. This work provides a proof of principle that the ansamycin core remains an attractive and effective scaffold for novel and dramatically improved RIFs.