The human proton pump inhibitors inhibit Mycobacterium tuberculosis rifampicin efflux and macrophage-induced rifampicin tolerance

Abstract

Tuberculosis treatment requires months-long combination chemotherapy with multiple drugs, with shorter treatments leading to relapses. A major impediment to shortening treatment is that Mycobacterium tuberculosis becomes tolerant to the administered drugs, starting early after infection and within days of infecting macrophages. Multiple lines of evidence suggest that macrophage-induced drug tolerance is mediated by mycobacterial drug efflux pumps. Here, using assays to directly measure drug efflux, we find that M. tuberculosis transports the first-line antitubercular drug rifampicin through a proton gradient-dependent mechanism. We show that verapamil, a known efflux pump inhibitor, which inhibits macrophage-induced rifampicin tolerance, also inhibits M.tuberculosis rifampicin efflux. As with macrophage-induced tolerance, the calcium channel-inhibiting property of verapamil is not required for its inhibition of rifampicin efflux. By testing verapamil analogs, we show that verapamil directly inhibits M. tuberculosis drug efflux pumps through its human P-glycoprotein (PGP)-like inhibitory activity. Screening commonly used drugs with incidental PGP inhibitory activity, we find many inhibit rifampicin efflux, including the proton pump inhibitors (PPIs) such as omeprazole. Like verapamil, the PPIs inhibit macrophage-induced rifampicin tolerance as well as intramacrophage growth, which has also been linked to mycobacterial efflux pump activity. Our assays provide a facile screening platform for M. tuberculosis efflux pump inhibitors that inhibit in vivo drug tolerance and growth.

Keywords: efflux pump inhibitors; mycobacterial efflux pumps; proton pump inhibitors; tuberculosis rifampicin tolerance; verapamil.

Fig. 1.

Inhibitors of macrophage-induced Mtb tolerance also inhibit efflux of EtBr and FITC-rifampicin in axenically grown Mtb. (A) Mtb mc²6206 intracellular EtBr fluorescence over time in the presence of verapamil (Ver) 0 to 100 μM. Values represent means of three technical replicates. (B) Efflux of FITC-rifampicin into supernatant over time from Mtb mc²6206 treated with Ver 0 to 100 μM. (C) Percent increase in intracellular EtBr accumulation in Mtb mc²6206 due to Ver 160 μM, piperine 350 μM, or thioridazine 5.4 μM, normalized to mean untreated value. (D) Percent inhibition of FITC-rifampicin efflux from Mtb mc²6206 due to Ver 160 μM, piperine 350 μM, or thioridazine 5.4 μM, normalized to mean untreated value. (B–D) Values represent means of three technical replicates ± SEM. Statistical analysis by one-way ANOVA with Dunnett’s multiple comparisons test. **** = P < 0.0005.

Fig. 2.

Rifampicin efflux is mediated principally by proton gradient-dependent transporters (A) Percent increase in Mtb mc²6206 intracellular EtBr accumulation due to CCCP 6.25 μM or bedaquiline (BDQ) 0.0625 μM (25% of MIC). (B) Percent inhibition of FITC-rifampicin efflux from Mtb mc²6206 due to CCCP 6.25 μM or BDQ 0.0625 μM. (C) Total Mtb mc²6206 ATP after 24 h treatment with BDQ. RLU, relative luminescence units. Values represent mean of three technical replicates ± SEM. **** = P < 0.0005. Representative of at least two experiments at similar concentrations. Statistical analysis by one-way ANOVA with Dunnett’s multiple comparisons test.

Fig. 3.

Verapamil analogs targeting PGP increase accumulation of EtBr and inhibit the efflux of FITC-rifampicin in Mtb. (A and C) Percent increase in intracellular EtBr accumulation in Mtb mc²6206 due to 25 μM drug. (B and D) Percent inhibition of FITC-rifampicin efflux from Mtb mc²6206 due to 25 μM drug. (C and D) Drugs ordered Left to Right by increasing efficacy of PGP inhibition (α_max). α_max is the maximum increase in the nuclear concentration of the PGP substrate pirarubicin in eukaryotic pirarubicin-resistant cells that can be obtained with a given compound, where α varies between 0 (no inhibitor present) and 1 (when the amount of pirarubicin in resistant cells is the same as in sensitive cells) (–46). Black dashed line indicates mean value for verapamil; blue dashed line indicates two-fold verapamil activity. (C and D) Gaussian distribution of data was confirmed by assessment of QQ plot and then outliers within each experiment were detected using Grubbs’ test (α = 0.05) and removed and remaining values were normalized to the mean untreated value. (A–D) Each symbol represents the mean normalized value for that drug for one individual experiment. (A–C) Combined data of three experiments. (D) Combined data of two (compounds 11, 17-Old, 23, 25, FC27, MC176, FRA58A) or three experiments.

Fig. 4.

Omeprazole, lansoprazole, rabeprazole, and pantoprazole are potent inhibitors of rifampicin but not EtBr efflux. (A) Percent increase in intracellular EtBr accumulation in Mtb mc²6206 due to 25 μM drug. (B) Percent inhibition of FITC-rifampicin efflux from Mtb mc²6206 due to 25 μM drug. (A and B) Dashed line indicates mean value for verapamil. Gaussian distribution of data was confirmed by assessment of QQ plot and then outliers within each experiment were detected using Grubbs’ test (α = 0.05) and removed, and then remaining values were normalized to the mean untreated value. Each symbol represents the mean normalized value for that drug for one individual experiment. Column height and error bars represent mean ± SEM of all mean values shown for that drug. (A) Combined data of three experiments. (B) Combined data of three experiments, except for itraconazole, which was tested twice.

Fig. 5.

The PPIs inhibit intramacrophage growth and rifampicin tolerance of Mtb. (A) THP-1 macrophages were infected with Mtb H37Rv and lysed for 2 h (black bars) or 96 h (white bars) after infection. The released bacteria were treated for an additional 48 h with 1 μg/mL rifampicin with or without inhibitor before enumeration of CFU. Statistical analysis by ordinary one-way ANOVA with Dunnett’s multiple comparisons test. Representative of two independent experiments. (B) Growth of Mtb H37Rv in THP-1 macrophages infected at MOI 1. 40 μM drug or vehicle was added at 2 d post-infection and growth allowed to continue a further 48 h. Macrophages were lysed in H₂O and the lysate was plated in dilution for CFU. Results representative of two or more independent experiments, except lansoprazole 40 μM, which was tested once. (A and B) Error bars represent SEM. ****P < 0.0005.