Comprehensive preclinical reevaluation of PaBQU and DBQU regimens identifies lesional liability in hard-to-treat forms of tuberculosis

Murine efficacy models inform advancement of preclinical tuberculosis treatment regimens to human clinical testing. A recent human Phase 2 trial was terminated early because investigational regimens (4-months of PaBQU or DBQU) did not meet the treatment shortening criteria outlined in the Target Regimen Profile ([≤] 3 months). We queried whether a factor leading to this early termination may have been lesional liability, meaning slow or diminished onset of effect in the caseum of complex lung lesions. We conducted a translational study, comparing the easy-to-treat BALB/c mouse, lacking complex lesions, to the hard-to-treat C3HeB/FeJ mouse that develops complex human-like lesions. We evaluated traditional and novel pharmacodynamic markers (colony forming units and RS ratio), relapse outcomes and ex vivo caseum pharmacokinetics. PaBQU and DBQU were slower to elicit bactericidal activity, RS ratio activity and prevent relapse in the C3HeB/FeJ mouse than the BALB/c mouse. A reference regimen (BPaMZ) had less lesional liability than PaBQU and DBQU. Fewer drugs in PaBQU were projected to achieve target attainment in caseum compared to BPaMZ, particularly early in treatment due to slow accumulation of bedaquiline in caseum. Here, we demonstrated application of a multi-modality pharmacokinetic-pharmacodynamic analysis that compared regimen activity in the hard-to-treat C3HeB/FeJ and easy-to-treat BALB/c mouse models, identifying lesional liability of PaBQU and DBQU. Systematic interrogation of additional diverse regimens is needed to determine the value of preclinical lesional liability as a means of predicting clinical treatment shortening activity.