Antitubercular pharmacodynamics of phenothiazines

Abstract

Objectives: Phenothiazines have been shown to exhibit in vitro and in vivo activity against Mycobacterium tuberculosis (Mtb) and multidrug-resistant Mtb. They are predicted to target the genetically validated respiratory chain component type II NADH:quinone oxidoreductase (Ndh). Using a set of compounds containing the phenothiazine pharmacophore, we have (i) investigated whether chemical validation data support the molecular target and (ii) evaluated pharmacophore tractability for further drug development.

Methods: Recombinant Mtb Ndh was generated and its functionality confirmed by steady-state kinetics. Pharmacodynamic profiling of the phenothiazines, including antitubercular efficacy in aerobic and O2-limited conditions, time-kill assays and isobole analyses against first-line antituberculars, was performed. Potential mitochondrial toxicity was assessed in a modified HepG2 cell-line assay and against bovine cytochrome bc1.

Results: Steady-state kinetic analyses revealed a substrate preference for coenzyme Q2 and an inability to utilize NADPH. A positive correlation between recombinant Ndh inhibition and kill of aerobically cultured Mtb was observed, whilst enhanced potency was demonstrated in a hypoxic model. Time-kill studies revealed the phenothiazines to be bactericidal whilst isobolograms exposed antagonism with isoniazid, indicative of intracellular NADH/NAD(+) couple perturbation. At therapeutic levels, phenothiazine-mediated toxicity was appreciable; however, specific mitochondrial targeting was excluded.

Conclusions: Data generated support the hypothesis that Ndh is the molecular target of phenothiazines. The favourable pharmacodynamic properties of the phenothiazines are consistent with a target product profile that includes activity against dormant/persistent bacilli, rapid bactericidal activity and activity against drug-resistant Mtb by a previously unexploited mode of action. These properties warrant further medicinal chemistry to improve potency and safety.

Figure 1.

(a) Plot of coenzyme Q₂ concentration dependence of recombinant Mtb Ndh, as determined by steady-state kinetic assays. The reduction of coenzyme Q₂ was monitored at 283 nm and resultant data fitted with a rectangular hyperbolic function to determine the apparent K_m of Q₂ as 10.4 ± 3.4 μM and the V_max as 1080.9 ± 59.4 nmol/min/mg. (b) Plot of thioridazine concentration against recombinant Mtb Ndh activity for IC₅₀ determination. The IC₅₀ for thioridazine was calculated as 11.4 μM ± 1.8 using a four-parameter logistic function (Origin 8.5 software). All assays contained 10–15 mg/mL protein in 50 mM potassium phosphate, 2 mM EDTA (pH 7.4) with potassium cyanide and NADH present at final concentrations of 1 mM and 200 μM, respectively. Inhibitor was introduced prior to reaction initiation by the addition of 50 μM coenzyme Q₂.

Figure 2.

Scatterplot of the IC₅₀s of eight phenothiazine-like compounds against recombinant Mtb Ndh and aerobically cultured Mtb, as detailed in Table 2. The data demonstrate a strong positive correlation between enzyme inhibition and in vitro antitubercular activity. Data for phenothiazine are not included, as accurate IC₅₀ determinations could not be made.

Figure 3.

Time–kill curves for untreated cells (control), thioridazine, trifluoperazine, fluphenazine, flupenthixol, promazine, promethazine, chlorpromazine, perphenazine, isoniazid and ethambutol. Compounds were present at 5× IC₉₀ (established from MABA assays) and cfu/mL were determined at the appropriate timepoints (as described in the Materials and methods section). Data are the mean of two experiments.

Figure 4.

Isobolograms for thioridazine in combination with (a) rifampicin, (b) streptomycin, (c) ethambutol and (d) isoniazid. Panels (e–h) show isobolograms for trifluoperazine in combination with the aforementioned antitubercular compounds, respectively. Fixed ratios of each pair of compounds were prepared and used to determine IC₅₀ values for each compound. Subsequently, the 50% fractional inhibitory concentration (FIC) for each compound in combination was calculated (see the Materials and methods section) and plotted as an isobologram. Compounds demonstrating an additive effect have a summed FIC ≈ 1 (dashed line), whilst antagonistic and synergistic combinations have values >4 and <0.5, respectively.

Figure 4.

Isobolograms for thioridazine in combination with (a) rifampicin, (b) streptomycin, (c) ethambutol and (d) isoniazid. Panels (e–h) show isobolograms for trifluoperazine in combination with the aforementioned antitubercular compounds, respectively. Fixed ratios of each pair of compounds were prepared and used to determine IC₅₀ values for each compound. Subsequently, the 50% fractional inhibitory concentration (FIC) for each compound in combination was calculated (see the Materials and methods section) and plotted as an isobologram. Compounds demonstrating an additive effect have a summed FIC ≈ 1 (dashed line), whilst antagonistic and synergistic combinations have values >4 and <0.5, respectively.

Figure 5.

Percentage viability of HepG2 cell versus drug concentration (μM) cultured in glucose (glu) and galactose (gal) media. Panels (a) and (b) illustrate the responses elicited by a switch in media for the non-mitochondrial toxicant tamoxifen and the respiratory poison rotenone, respectively. (c) The responses for phenothiazine, promethazine and thioridazine in glucose and galactose media. In (c) the dashed vertical line represents the therapeutic maximum plasma concentration of thioridazine, whilst the dotted line indicates the concentration of thioridazine required to kill phagocytosed Mtb.^,