Permission to undertake this study and draw blood from healthy, adult human volunteers was granted by the Research Ethics Committee of the Faculty of Health Sciences, University of Pretoria in full compliance with the World Medical Association Declaration of Helsinki, 2013 (Approval number: 116/2017). Informed consent was obtained from all participants in the study prior to the blood draw [17 female:7 male, mean age ( ± SD) = 35.25 ± 12.0 years].

Bedaquiline, (1R,2S)-1-(6-bromo-2-methoxyquinolin-3-yl)-4-(dimethylamino)-2-naphthalen-1-yl-1-phenylbutan-2-ol, was purchased from AdooQ Bioscience (York, UK); wortmannin [(1R,3R,5S,9R,18S)-18-(methoxymethyl)-1,5-dimethyl-6,11,16-trioxo-13,17-dioxapentacyclo[10.6.1.0^2,10.0^5,9.0^15,19]nonadeca-2(10),12(19),14-trien-3-yl] acetate, a selective inhibitor of PI3-K, from the Sigma Chemical Co. (St Louis. MO, USA); U73122, {1-6[[(17ß)-3-methoxyestra-1,3,5(10)-trien-17-yl]amino]hexyl]-1H-pyrolle-2,5-dione}, a selective inhibitor of phospholipase C (PLC), and U46619, (Z)-7-{[1R,4S,5S,6R]-6-[(E,3S)-3-hydroxyoct-1-enyl]-2-oxabicyclo[2.2.1]heptan-5-yl}hept-5-enoic acid, a thromboxane A₂ surrogate, were purchased from Tocris Bioscience, Bristol, UK. Bedaquiline, wortmannin, U73122, and U46619 were dissolved to stock concentrations in dimethylsulfoxide (DMSO), and used at final concentrations of 0.625–10 µg/ml (equivalent to 1.13–18 µM), 1 µM and 0.625 µM and 5 µM, respectively, in the assays of platelet activation described below, which also included appropriate solvent controls.

Adenosine-5’-diphosphate (ADP) and thrombin (from human plasma) were purchased from Boehringer Mannheim Biochemical (Basel, Switzerland) and the Sigma Chemical Co. respectively and used at final maximal concentrations of 100 µM and 1 NIH unit/ml respectively. ADP is an agonist of platelet P2Y1 and P2Y12 receptors and acts as a secondary, autocrine activator of these cells, while thrombin is an agonist of the platelet proteinase-activated receptors 1 and 4 (PAR1/4). U46619 was used at a final concentration of 5 µM, as mentioned above.

Unless indicated, all other chemicals and reagents were purchased from the Sigma Chemical Co.

To prepare platelet-rich plasma (PRP), blood (anti-coagulated with five units of preservative-free heparin/ml blood) was centrifuged at 250 x g for 10 min at room temperature within 15 min of venepuncture and the essentially erythrocyte- and leukocyte-free upper layer of PRP decanted and used in the experiments described below.

PRP was diluted four-fold with phosphate-buffered saline (PBS, 0.15M, pH 6.8) containing ethylene glycol-bis(β-aminoethyl ether)-N-N-N’-N’-tetraacetic acid (EGTA, final concentration 0.6 mM) and centrifuged at 260 x g for 10 min at 20°C to deplete residual erythrocytes and leukocytes. The platelet-enriched supernates were then centrifuged at 800 x g for 20 min and the cell pellets resuspended in PBS (pH 7.4) and counted using a Sysmex XN 20000 Automated Haematology Analyzer (Sysmex Corporation, Kobe, Japan).

The low pH (6.8) of the PBS together with the addition of EGTA to deplete extracellular Ca²⁺ and preparation of the cells at a constant temperature of 20°C, were used to minimize platelet activation. We preferred this strategy to inclusion of acetylsalicyclic acid as we have previously observed that induction of synthesis of thromboxane A₂ represents a significant component of the platelet response following activation with ADP.

Platelet activation was measured by flow cytometry as the proportion (%) of CD42a⁺ platelets expressing the α-granule-derived adhesion molecule, CD62P (P-selectin), a key mediator of platelet activation and pro-thrombotic activity. For these experiments, PRP (20 µl) was added to 980 µl of Hanks’ balanced salt solution (HBSS, indicator-free, 1.25mM calcium, pH7.4) and incubated for 10 min at 37°C in the presence of bedaquiline (0.625–10 µg/ml final, equivalent to 1.13–18 µM), or wortmannin (1µM), or U73122 (0.625 µM), or an equal volume of DMSO to control systems (0.1–0.2%, final). This was followed by the addition of ADP, thrombin, U46619 or an equal volume of HBSS (background to detect spontaneous platelet activation). The tubes were then incubated for a further period of 5 min at 37°C and processed immediately thereafter for analysis by flow cytometry. The cells were stained with 5 µl each of a murine anti-human platelet CD42a-phycoerythrin (PE)-labeled monoclonal antibody (Becton Dickenson, Franklin Lakes, USA) and an anti-human CD62P-fluorescein isothiocyanate (FITC)-labeled monoclonal antibody (Beckman Coulter, Miami, FL, USA) to detect the total and activated platelet populations respectively. After 15 min of incubation in the dark, the samples were analyzed on a Gallios flow cytometer (Beckman Coulter, Miami, FL, USA) and the results expressed as the percentage of activated platelets with 50,000 cells interrogated during each measurement.

Fura-2 acetoxymethyl ester (fura-2AM) was used as the fluorescent, Ca²⁺-sensitive indicator for these experiments (18). In this series of experiments, purified platelets suspended in HBSS were incubated at 37°C for 5 min followed by addition of fura-2AM (2 µM) and a further incubation period of 25 min at 37°C after which the platelets were pelleted by centrifugation at 350 x g for 20 min. The cell pellets were resuspended in PBS. For measurement of cytosolic Ca²⁺ fluxes, 100 µl of platelet suspension was added to 2.9 ml HBSS and incubated for 10 min at 37°C in the absence or presence of either bedaquiline (5 and 10 µg/ml) or wortmannin (1 µM). The fura-2AM-loaded platelets were then transferred to disposable reaction cuvettes maintained at 37°C in a Hitachi 650 10S fluorescence spectrometer with excitation wavelengths set at 340 and 500 nm respectively. After a stable baseline was obtained, the platelets were activated by addition of either ADP (100 µM) or thrombin (1 NIH unit/ml) and alterations in fluorescence intensity measured over a 5 min time course. These results are presented as the peak Ca²⁺ concentrations calculated as described previously (18) and as typical traces from representative experiments.

For this limited series of experiments, PRP (150 µl), or an equal volume of platelet poor plasma (PPP) prepared by centrifugation of PRP at 350 x g for 5 min, was added to 1,850 µl of HBSS followed by the addition of 2 µl of bedaquiline (10 µg/ml final), wortmannin (1µM final), or DMSO to control systems and the test tubes incubated at 37°C for 20 min. Thereafter, ADP (100 µM final) or an equal volume of HBSS (background to detect spontaneous platelet aggregation), were added to the tubes, which were incubated at 37°C for 10 min with repeated inversion at 1-min intervals. For measurement of platelet aggregation, 250 µl from each tube (3 repeats) were added to a 96-well microtissue culture plate using reverse pipetting to minimize air bubbles. Light absorbance, as a surrogate of platelet aggregation, was measured spectrophotometrically at a wavelength of 650nm (19). The results are expressed as the mean percentage aggregation, which was calculated based on the absorbance of the PPP control (100% aggregation) and the PRP control (0% absorbance).

Platelets (approximately 1 x 10⁸/ml) suspended in HBSS were preincubated with bedaquiline (5 and 10 µg/ml) or an equivalent volume of DMSO for 10 min at 37°C followed by the addition of ADP (100 µM). After 3 min of incubation (20–22), the reactions were terminated and the platelets lysed by addition of an equal volume of cell extraction buffer supplemented with a 1% protease and phosphatase inhibitor cocktail (all purchased from Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA) and held on ice for 30 min. Thereafter, the cell extracts were centrifuged at 13,000 rpm for 10 min and the supernatants decanted and stored at −80°C until analysis of phosphorylated Akt1 using an Ultrasensitive solid phase sandwich ELISA system (Thermo Fisher Scientific). The results of five experiments using platelets from three separate donors are expressed as units/ml.

The results of each series of experiments are expressed as the median values with interquartile ranges, with the exception of platelet counts, contaminating cells, platelet aggregation and cytosolic Ca²⁺ fluxes, which are expressed as the mean values ± standard deviations, for a minimum of 3–4 different donors with numbers of replicate experiments indicated. Statistical analysis was performed using WinStat statistical software with the Shapiro-Wilk test used to test the distribution of data and levels of statistical significance calculated using the Mann-Whitney U-test for comparison of non-parametric data. A P value of <0.05 was considered significant.

Circulating platelet counts for all blood donors were within the normal range (150–400 x 10⁹/l), while the mean number of platelets in the concentrated, purified platelet suspensions was 558 x 10⁹/l (n=14). The numbers of contaminating neutrophils, lymphocytes and monocytes in the pure platelet suspensions were almost undetectable, the mean values with SDs being 0.98 ± 0.14. 0.23 ± 0.42 and 0.016 ± 0.02 x 10⁹, respectively.

As shown in Figure 1 , pre-treatment of PRP with bedaquiline (0.625–10 µg/ml) resulted in dose-related inhibition of ADP (100 µM)-activated upregulation of expression of CD62P, which achieved statistical significance at concentrations of 5 and 10 µg/ml bedaquiline (P<0.05). As shown in Figure 2 , however, the corresponding responses of platelets activated with thrombin (1NIH unit/ml) or U46619 (5µM) were unaffected by exposure to bedaquiline at either 5 or 10 µg/ml, demonstrating a selective inhibitory effect of the test anti-mycobacterial agent on ADP-mediated platelet activation.

As shown in Figure 3 , pre-treatment of PRP with either wortmannin (1 µM) or U73122 (0.625 µM), almost completely attenuated ADP-activated upregulation of CD62P expression, seemingly demonstrating the mechanistic involvement of PI3-K and/or PLC in this event.

The effects of wortmannin (1 µM) or U73122 (0.625 µM) on thrombin- and U46619-activated platelets are shown in Figure 4 . Pre-treatment of PRP with wortmannin did not affect the magnitude of either thrombin- or U46619-mediated upregulation of CD62P expression, while pre-treatment with U73122 almost completely attenuated upregulation of CD62P expression mediated by both thrombin and U46619. These results appear to demonstrate the key involvement of PLC, but not that of PI3-K, in upregulated expression of CD62P by platelets activated by thrombin and U46619.

As shown in Figure 5 , pre-treatment of purified platelets with bedaquiline at 10 µg/ml or wortmannin (1 µM) caused statistically significant inhibition of the immediate increase in the concentration of cytosolic Ca²⁺ following addition of ADP, while the more intense response activated by thrombin was unaffected by treatment of platelets with either of these agents. The basal cytosolic Ca²⁺ concentration of unstimulated platelets was 192 ± 21 nM, while the peak values observed for ADP-activated platelets in the absence or presence of bedaquiline or wortmannin were 420 ± 121 nM, 81 ± 74 nM (P<0.05) and 330 ± 136 nM (P<0.05), respectively. The corresponding values for thrombin-activated platelets in the absence or presence of bedaquiline were 680 ± 124 nM and 694 ± 151 nM (P=0.77), respectively (n=6–12 experiments). The slight increase in the rate of decline of fluorescence intensity in the thrombin-activated system in the presence of bedaquiline may result from attenuation of the endogenous ADP-mediated autocrine response.

These findings implicate interference with both PLC and PI3-K in bedaquiline-mediated inhibition of ADP-mediated activation of platelets.

Pre-treatment of platelets with bedaquiline (10 µg/ml) or wortmannin (1 µM) caused statistically significant inhibition of platelet aggregation. The magnitudes of ADP-activated platelet aggregation in the absence or presence of bedaquiline or wortmannin were 17, 3.7, and 0.4%, (P<0.05) (n=8–14 experiments), respectively. These findings reveal that bedaquiline inhibits several functions of ADP-activated platelets.

As shown in Figure 6 , pre-treatment of isolated platelets with bedaquiline (5 and 10 µg/ml) resulted in potent, statistically significant (P<0.05 for both), dose-related inhibition of PI3-K-mediated phosphorylation of Akt1, representing the probable mechanism by which bedaquiline attenuates ADP-induced platelet activation.