Accelerating drug development in infectious diseases using zebrafish disease models supported by pharmacokinetic pharmacodynamic modeling as new approach methodology (NAM)

Key results

1. ShownBedaquiline's exposure-response in M. marinum-infected larvae was described by delayed logistic bacterial growth with an Emax drug effect on the growth rate. The maximal effect was larger after 3 days post infection (dpi).

   Growth rate constant 0.0853 h^-1. Emax, expressed as a multiple of the growth rate constant, was 1.09 up to 3 dpi and 1.37 later (increase p<0.01). EC50 26.5 µg/mL; all relative standard errors below 50%. A sigmoid Emax model fitted significantly better but its estimates were too imprecise. Fit was checked with visual predictive checks (500 simulations). · Bedaquiline: exposure-response modelling; Exposure-response modelling

2. ShownBedaquiline inside the larvae rose quickly, reached a dose-dependent plateau and fell in two phases after drug removal. A one-compartment model described these data, with elimination increasing at 5 days post fertilization (dpf).

   45 medium measurements (0.5, 5, 50 µM) and 120 homogenates of 5 larvae each (0.5 to 50 µM), n=3 per timepoint. Absorption rate constant 0.000941 h^-1; elimination rate constant 0.089 h^-1, rising by 0.122 h^-1 at 5 dpf (p=0.025); all PK parameters had relative standard errors below 50%. · Bedaquiline: internal exposure; Bedaquiline: exposure-response modelling; Exposure-response modelling

3. ShownClarithromycin failed the internal-exposure criterion under waterborne dosing, yet killed bacteria when injected intravenously. It is therefore a false negative of the waterborne workflow.

   Internal amounts were under 0.1 % of the external medium, although medium concentrations were stable and washing was optimized. Injecting 0.5 or 1.0 ng (vs PBS; Kruskal-Wallis with Dunn's tests) lowered M. marinum and M. avium MAC101 burden in a dose-dependent way. Internal exposure after injection was not measured, so no exposure-response was derived. · Clarithromycin did not pass workflow criteria; Figure legends

4. ShownBedaquiline passed the potency-versus-toxicity screen. Under waterborne treatment it reduced M. marinum burden in a dose-dependent way once differences in starting inocula were accounted for, and M. avium burden also showed a significant dose-dependent effect (Figure S2).

   Minimum inhibitory concentration (MIC) 0.1 mg/L (0.18 µM), with no toxicity up to 50 µM (>25-fold margin). 60 larvae, at least n=5 per timepoint and dose (0, 0.5, 1, 5, 10, 50 µM), were imaged repeatedly. Bacterial growth slowed between the third and fourth day of infection regardless of dose. · Bedaquiline: drug efficacy

5. ClaimedThe authors state that the workflow gives in vivo efficacy and potency within 1 to 2 orders of magnitude of mammalian values, and that bedaquiline's exposure-response was close to that seen in rodents.

   This is stated in the abstract and discussion. The text shows no rodent or mammalian parameter values and no side-by-side comparison. The workflow is said to take about 2 months, with the internal-exposure stage taking three experiments (about 1 month). · Abstract; Workflow development; Results; Discussion and conclusion