Host-Mediated Bioactivation of Pyrazinamide: Implications for Efficacy, Resistance, and Therapeutic Alternatives

Abstract

Pyrazinamide has played a critical role in shortening therapy against drug-sensitive, drug-resistant, active, and latent tuberculosis (TB). Despite widespread recognition of its therapeutic importance, the sterilizing properties of this 60-year-old drug remain an enigma given its rather poor activity in vitro. Here we revisit longstanding paradigms and offer pharmacokinetic explanations for the apparent disconnect between in vitro activity and clinical impact. We show substantial host-mediated conversion of prodrug pyrazinamide (PZA) to the active form, pyrazinoic acid (POA), in TB patients and in animal models. We demonstrate favorable penetration of this pool of circulating POA from plasma into lung tissue and granulomas, where the pathogen resides. In standardized growth inhibition experiments, we show that POA exhibits superior in vitro potency compared to PZA, indicating that the vascular supply of host-derived POA may contribute to the in vivo efficacy of PZA, thereby reducing the apparent discrepancy between in vitro and in vivo activity. However, the results also raise the possibility that subinhibitory concentrations of POA generated by the host could fuel the emergence of resistance to both PZA and POA. In contrast to widespread expectations, we demonstrate good oral bioavailability and exposure in preclinical species in pharmacokinetic studies of oral POA. Baseline exposure of oral POA can be further increased by the xanthine oxidase inhibitor and approved gout drug allopurinol. These promising results pave the way for clinical investigations of oral POA as a therapeutic alternative or an add-on to overcome PZA resistance and salvage this essential TB drug.

Keywords: Mycobacterium tuberculosis; bioactivation; host metabolism; pyrazinamide; pyrazinoic acid.

Figure 1

Concentration-time profile of PZA and POA in plasma following oral administration of 1500 mg of PZA to 14 TB patients. Each subject contributed three to five plasma samples from predose to 25 h after drug administration.

Figure 2

(A) Concentration-time profile of PZA and POA following the oral administration of 150 and 300 mg/kg to uninfected rabbits. Blue diamonds, PZA 150 mg/kg; green diamonds, POA from PZA dosed at 150 mg/kg; blue circles, PZA 300 mg/kg; green circles, POA from PZA dosed at 300 mg/kg. (B) POA/PZA concentration ratio in plasma at sequential time points following oral dosing of PZA to naïve, M. tuberculosis-infected (12 weeks postinfection), and M. bovis-infected (4 weeks postinfection) rabbits. Purple bars, uninfected rabbits dosed with 150 mg/kg PZA; blue bars, M. bovis AF2122 infected rabbits dosed with 125 mg/kg PZA; green bars, uninfected rabbits dosed with 300 mg/kg PZA; orange bars, M. tuberculosis HN878 infected rabbits dosed with 250 mg/kg PZA.

Figure 3

(A) Concentration-time profile of PZA and POA metabolite following oral administration of (A) a single 150 mg/kg PZA dose to naïve C3HeB/FeJ mice, (B) a single 300 mg/kg PZA dose to naïve guinea pigs, and (C) a single 125 mg/kg PZA dose to naïve marmosets. Blue diamonds, plasma concentrations of PZA; green diamonds, plasma concentrations of POA metabolite.

Figure 4

Concentration-time profile of POA following oral administration of a single POA dose to naïve BALB/c mice (150 mg/kg, squares), guinea pigs (300 mg/kg, diamonds), and NZW rabbits (200 mg/kg of Na-POA, circles). At the last time point of 24 h, all plasma concentrations were below the limit of detection.

Figure 5

Effect of xanthine oxidase inhibition by varying allopurinol predosing schedules on the metabolism of PZA and POA in CD-1 mice and NZW rabbits. Blue bars, AUC_[0–24] of PZA; green bars, AUC_[0–24] of POA; orange bars, AUC_[0–24] of 5-hydroxy-POA.