A human Phase 2 trial testing 4-month PaBQU and DBQU investigational regimens was stopped early because the regimens failed to meet the Target Regimen Profile benchmark for treatment shortening (≤ 3 months). The investigators asked whether a contributing factor to the early termination could be lesional liability — that is, slow or reduced onset of drug effect within the caseum of complex pulmonary lesions that resemble human disease.
To address this, the study performed a translational preclinical comparison using two established murine models with different lesion pathologies. One model (BALB/c) lacks complex caseous lesions and is considered easy-to-treat, while the other (C3HeB/FeJ) develops complex, human-like lesions and is considered hard-to-treat. The goal was to evaluate whether regimen activity differs between models in ways that would indicate lesional liability and potentially explain clinical treatment-shortening failures.
The study contrasted regimen activity in the BALB/c and C3HeB/FeJ mouse strains. Pharmacodynamic assessment included conventional microbiologic readouts (colony forming units, CFU) and a novel molecular marker, the RS ratio, as indicators of bactericidal activity. Relapse outcomes were recorded to assess sterilizing activity and treatment durability beyond bactericidal effects.
In addition to in vivo efficacy measures, the investigators evaluated ex vivo caseum pharmacokinetics of regimen drugs to project target attainment within lesion caseum. These data were integrated in a multi-modality pharmacokinetic–pharmacodynamic (PK–PD) analysis to link drug exposure in caseum with measures of regimen activity and relapse prevention across the two animal models.
Across the comparative analyses, both PaBQU and DBQU regimens were slower to elicit bactericidal activity (as measured by CFU) and to change the RS ratio in the C3HeB/FeJ mice than in the BALB/c mice. These delays in onset of effect in the hard-to-treat model are consistent with lesional liability, where drug action is diminished or delayed within complex pulmonary lesions.
A reference regimen, BPaMZ, showed less difference in activity across the two models, indicating lower lesional liability compared with PaBQU and DBQU in these preclinical settings.
Relapse assessments paralleled the bactericidal and RS ratio findings. PaBQU and DBQU were less effective at preventing relapse in the C3HeB/FeJ mice than in BALB/c mice, consistent with slower sterilizing activity in the presence of complex lesions. The observed model-dependent differences in relapse prevention support the hypothesis that lesional drug access or activity can undermine attempts to shorten treatment duration clinically.
Ex vivo caseum pharmacokinetic analyses projected that fewer drugs comprising the PaBQU regimens would achieve target attainment within caseum compared with BPaMZ. This shortfall was particularly evident early in treatment and was partly attributed to slow accumulation of bedaquiline within caseum. These PK observations provide a mechanistic basis for the delayed bactericidal and sterilizing activity observed in the hard-to-treat model.
The authors applied an integrated pharmacokinetic–pharmacodynamic framework combining in vivo efficacy markers (CFU, RS ratio, relapse) with ex vivo caseum PK to assess regimen performance across lesion types. This multi-modality approach identified lesional liability of PaBQU and DBQU in the C3HeB/FeJ model relative to BALB/c, and highlighted differential drug penetration/accumulation in caseum as a likely contributor.
These findings suggest that standard murine efficacy models that lack complex human-like lesions may overestimate regimen potential for treatment shortening, whereas models with caseous lesions can reveal important limitations related to lesion-specific drug exposure and activity.
The study authors conclude that systematic interrogation of a broader set of regimens using diverse lesion-bearing preclinical models is needed to establish whether measures of lesional liability can reliably predict clinical outcomes for treatment shortening. Specific experimental details, numerical PK parameters, and quantitative target attainment thresholds were reported in the source but are not reproduced here; those interested in granular data should consult the full preprint.
In this translational preclinical comparison, PaBQU and DBQU exhibited delayed bactericidal and sterilizing activity and reduced caseum target attainment in the hard-to-treat C3HeB/FeJ model relative to BALB/c, consistent with lesional liability. The reference BPaMZ regimen showed less lesional liability. The work demonstrates the value of combined PK–PD and lesion-focused preclinical assessments to flag potential limitations of regimens intended to shorten tuberculosis treatment and supports further systematic evaluation of regimens across lesion-diverse models.