To study whether PARP may play a role during Mtb infections in human macrophages, we first evaluated intracellular Mtb survival after treatment with four clinically relevant PARPi ( Figures 1A, B ). We selected 10 µM as a relatively high PARPi concentration, which has been shown to target PARP in vitro (15–17) and which allowed us to directly compare drug efficacy of PARPi with each other and with control kinase inhibitor H-89 (18, 19). Rucaparib significantly decreased intracellular Mtb CFUs in M1 and M2 more efficiently than H-89 ( Figure 1C , n=6). Additionally, niraparib and pamiparib decreased intracellular Mtb in M1 and A-966492 decreased intracellular Mtb in M2, suggesting cell-specific effects of PARP inhibition on bacterial load for these compounds. Furthermore, we assessed the effect of PARPi on cytotoxicity using uninfected macrophages from four different donors, to ascertain that PARPi was not detrimental to healthy human macrophages. PARPi did not affect host-cell integrity and was not cytotoxic ( Figure 1D ). Next, we used a flow cytometry-based method (18) to evaluate whether the inhibition of Mtb outgrowth by PARPi was accompanied by a decrease in the percentage of cells harboring Venus-expressing Mtb. Rucaparib, but not other PARPi, significantly reduced the percentage of Mtb-infected cells both in M1 and M2 ( Figure 1E , n=8), further corroborating the Mtb-inhibiting effect of rucaparib in human macrophages. To exclude that PARPi had direct microbicidal effects, liquid Mtb cultures were treated with PARPi in the absence of macrophages in two independent experiments ( Figure 1F ). None of the PARPi limited bacterial growth directly, indicating that PARPi restrict intracellular outgrowth of Mtb by modulating host pathways.

To investigate whether PARPi-induced inhibition of intracellular bacterial growth was specific for Mtb, we evaluated whether pharmacological inhibition of PARP also restricted growth of other intracellular pathogenic bacteria, including Mycobacterium avium (Mav), Salmonella enterica serovar Typhimurium (Stm) and methicillin-resistant Staphylococcus aureus (MRSA). Interestingly, niraparib, but not other PARPi, significantly decreased Stm CFUs in M1 and M2 ( Figure 2A , n=6) without exhibiting direct microbicidal effects in two independent experiments ( Figure 2B ). Rucaparib did not significantly diminish Mav, Stm or MRSA CFUs, suggesting that the effect of rucaparib is specific for Mtb. These results imply that both rucaparib and niraparib inhibit specific species of bacteria through modeling of host signaling pathways.

Taken together, these data suggest that PARP activity is involved in Mtb survival in human macrophages, and that although there is some difference in their efficacy against Mtb-infected M1 compared to Mtb-infected M2, pharmacological inhibition of PARP is a promising strategy to target intracellular Mtb. Moreover, the effects of different PARP inhibitors are highly specific to certain pathogen species.

To evaluate the translational potential of PARPi for the development of HDT, we first tested whether PARPi displayed activity against macrophages infected with two different multi-drug resistant (MDR)-Mtb strains: an MDR-Mtb strain belonging to the Beijing genotype (strain 16319) and an MDR-Mtb Dutch outbreak strain (strain 2003-1128), both resistant to rifampicin (RIF) and isoniazid (INH). Although drug-susceptible and MDR Mtb strains are expected to respond similarly to HDT compounds, previous experiments by our laboratory suggested that in selective cases compound efficacy may differ between Mtb strains (20). Here, rucaparib significantly decreased intracellular MDR-Mtb and A-966492 showed activity against the MDR-Mtb Dutch outbreak strain ( Figures 3A, B , n=6), highlighting the potential of PARPi to control intracellular outgrowth not only of drug-sensitive but also MDR-Mtb strains.

Next, the interaction of PARPi with first-line antibiotics was studied. Macrophages from five donors were infected with Mtb and treated with PARPi in the presence or absence of a non-toxic, suboptimal dose of RIF or INH ( Figure 3C ) (20, 21). As expected, PARPi significantly decreased Mtb CFUs in M1 ( Figure 3D ). In M2, however, several HDT compounds, including rucaparib, niraparib and positive control H-89 increased Mtb CFUs in two donors, illustrating the typically high variation in the response of donors to HDT compounds. Niraparib, pamiparib and rucaparib significantly enhanced the activity of INH against intracellular Mtb in M1. The coefficient of drug interaction (CID) of niraparib, pamiparib and rucaparib with RIF was close to 1 in M1, suggesting that these PARPi had an additive effect to the antibiotic treatment and not a synergistic effect ( Supplementary Table 1 ). Furthermore, compared with other PARPi, pamiparib seemed to enhance the activity of RIF in M2 and of INH in M1 and M2, albeit not significant.

To investigate whether the effect of PARPi against Mtb correlated with macrophage activation during Mtb infection, cytokine and chemokine production was quantified in the supernatants of Mtb-infected macrophages that were treated with PARPi or vehicle control DMSO overnight. Mtb-infection induced cytokine and chemokine production in M1 and M2 and tended to induce tumor necrosis factor (TNF)-α, macrophage inflammatory protein (MIP)-1α,MIP-1β and IP-10 and fractalkine in M1 and M2 ( Figure 4A and Supplementary Table 2 , n=8). Exposure to PARPi altered the cytokine and chemokine response of M1 and M2 upon Mtb infection ( Figure 4B and Supplementary Table 2 , n=4): pamiparib and rucaparib both increased MIP-1α and MIP-1β secretion by M1 and significantly decreased IFN-α2 secretion by M2 compared to DMSO. Correlation plots between the cytokines and chemokines that were significantly modified by PARPi and Mtb CFUs did not identify a positive correlation between secretion profiles of macrophages and Mtb loads ( Supplemental Figure 1 ). Collectively, these data suggest that pamiparib and rucaparib can modify the cytokine and chemokine response to Mtb infection.

Next, we investigated the expression levels of activation markers on the surface of Mtb-infected M1 and M2 that were treated with PARPi overnight. Mtb-infection tended to increase macrophage activation markers in M1 and M2 ( Figure 4C , n=4). Surprisingly, rucaparib, but not other PARPi, tended to decrease expression of CD192 and CD209 and expression of human leucocyte antigen (HLA)-DR on the surface of Mtb-infected M1 and M2 ( Figure 4D and Supplemental Figure 2 , n=4). Furthermore, the expression of costimulatory molecule CD86 on the surface of M2 tended to be increased by rucaparib, and was significantly increased by pamiparib. These data suggest that rucaparib increases macrophage activation and possibly antigen presentation.

Given that several HDT compounds with demonstrated efficacy against intracellular Mtb in macrophages exert their activity by activating autophagy (22–25), we next assessed whether treatment with PARPi induced autophagy and lysosomal maturation. Accumulation of autophagic and lysosomal vesicles was quantified by confocal imaging on Mtb-infected and PARPi-treated macrophages from four donors that were stained with CYTO-ID or Lysotracker, respectively. Formation of autophagic or lysosomal vesicles could not be detected after 4h treatment with PARPi compared to DMSO ( Supplemental Figures 3A, B ). To examine the effect of PARPi on the autophagic flux, macrophages from three donors were treated with PARPi in the presence or absence of bafilomycin A1 (Baf) which is an autophagy inhibitor that prevents fusion between autophagosomes and lysosomes. As expected, treatment with Baf induced microtubule-associated protein light chain 3 (LC3)-II accumulation in M1 and M2 ( Supplemental Figure 3C ). However, PARPi did not increase LC3-II protein levels, regardless of the presence of Baf. Also, PARPi did not increase LAMP1 protein levels, suggesting that PARPi do not act via induction of autophagic and/or lysosomal pathways to control intracellular Mtb in human macrophages.

Collectively, our data demonstrate that PARPi modulate the immune response of Mtb-infected human macrophages via cytokine/chemokine expression and surface markers, which play a role in antigen presentation and chemotaxis. Our data also suggest that a role of induction of autophagy and phagosome maturation in PARP-induced Mtb control is not likely mechanistically involved in the mode of action.

Hexose-6-phosphate dehydrogenase (H6PD) and lactate dehydrogenase (LDH) were recently identified as additional target molecules of rucaparib (26). We therefore hypothesized that these metabolic targets might play a role in the more profound inhibitory effect of rucaparib on Mtb survival compared to other PARPi ( Figures 1C, E , 3A , 5A ). Lactate levels were significantly decreased in the supernatants of Mtb-infected M1 and M2 that had been treated with rucaparib compared to DMSO, suggesting that rucaparib indeed impaired LDH activity ( Figure 5B , n=4). Interestingly, none of the other selected PARPi significantly affected lactate levels, compatible with the fact that none of these have been reported to modulate LDH activity. To study whether inhibition of H6PD or LDH could have contributed to the effect of rucaparib on Mtb control, we treated Mtb-infected macrophages with the specific LDH-inhibitor FX-11 or the G6PD-inhibitor 6-aminonicotinamide (6-AN) and determined intracellular Mtb levels in a flow cytometry-based assay, which allows a more rapid quantification of intracellular Mtb levels compared to classical CFU assays and generally showed a greater effect window of rucaparib compared to CFU assay ( Figure 1C versus 1E ). Interestingly, FX-11 significantly decreased the percentage of Mtb-infected M1 (n=7) and M2 (n=8), whereas no effect of 6-AN on intracellular Mtb could be detected ( Figure 5C ). FX-11 did not affect the percentage of live M1 cells (n=6), but did decrease the percentage of live cells in a subset of M2 cultures (n=5). Together, these data suggest that inhibition of LDH, but not G6PD, likely contributed to the Mtb-decreasing effect of rucaparib. Corroborating this hypothesis, FX-11 treatment significantly decreased lactate levels in the supernatant of Mtb-infected M1, corresponding with decreased LDH activity ( Figure 5E , n=6). FX-11 treatment did not result in decreased lactate levels in M2, which is likely a result of cellular toxicity in M2 ( Figures 5D, E ). Collectively, these data suggest that LDH could be involved in rucaparib-induced inhibition of Mtb and identifies LDH as potential host target during Mtb infections.

Dimethyl sulfoxide (DMSO), staurosporine solution from Streptomyces sp. (STSP), H-89 dihydrochloride hydrate (H-89), bafilomycin A1 (Baf), 6-aminonicotinamide (6-AN), lactate dehydrogenase A inhibitor (FX-11), rifampicin (RIF), gentamicin, tetracycline hydrochloride, ethylenediaminetetraacetic acid (EDTA) and Hoechst 33342 were obtained from Sigma-Aldrich (Zwijndrecht, The Netherlands). PARP inhibitors (A-966492, niraparib, pamiparib and rucaparib) and isoniazid (INH) were purchased from SelleckChem.

PE anti-human CD209 (1:50, clone 9E9A8), PerCP/Cy5.5 anti-human CD192 (CCR2, 1:20, clone K036C2), PE/Cy7 anti-human CD40 (1:100, clone 5C3), Alexa Fluor^® 647 anti-human CD163 (1:50, clone CHI/61), APC/Cy7 anti-human CD206 (1:50, clone 15-2) and BV785 anti-human CD14 (1:200, clone M5E2) were purchased from Biolegend (London, United Kingdom). APC-R700 anti-human CD80 (1:200, clone L307.4), V500 anti-human HLA-DR (1:100, clone G46-6) and BV711 anti-human CD86 (1:100, clone 2331) were purchased from BD Bioscience (Vianen, The Netherlands). BV605 anti-human CD197 (CCR7, 1:200, clone G043H7) was purchased from Sony Biotechnology (Weybridge, United Kingdom). Mouse monoclonal anti-human beta actin (1:2000, clone mAbcam 8226) and rabbit polyclonal anti-human LAMP1 (1:500) were purchased from Abcam (Amsterdam, The Netherlands). Rabbit polyclonal anti-human LC3-I/II (1:250) was purchased from Cell Signaling Technology (Leiden, The Netherlands). Secondary stabilized peroxidase conjugated antibodies goat anti-rabbit IgG (H+L) (1:10,000) and goat anti-mouse IgG (H+L) (1:10,000) were purchased from Thermo Fisher Scientific (Breda, The Netherlands).

Human peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats of healthy anonymous donors (Dutch, adults) after written informed consent (Sanquin Blood Bank, Amsterdam, The Netherlands) by density gradient centrifugation over Ficoll/amidotrizoaat as described earlier (61). The use of buffy coats for research purposes was approved by the Institutional Review Board of the Leiden University Medical Center, The Netherlands. Magnetic cell sorting using anti-CD14-coated microbeads (Miltenyi Biotec, Auburn, CA) was used to isolate CD14+ monocytes. CD14+ cells were cultured for six days at 37°C/5% CO₂ in Gibco Roswell Park Memorial Institute (RPMI) 1640 medium (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS), 2 mM L-Alanyl-L-Glutamine (PAA, Linz, Austria), 100 units/ml penicillin (Thermo Fisher Scientific), 100 µg/ml streptomycin (Thermo Fisher Scientific) and either 5 ng/ml granulocyte-macrophage colony-stimulating factor (GM-CSF, Thermo Fisher Scientific) to promote M1 differentiation or 50 ng/ml macrophage colony-stimulating factor (M-CSF, R&D Systems, Abingdon, United Kingdom) to promote M2 differentiation. Cytokines were added again at day 3 of differentiation in equal concentrations. The M1 and M2 macrophage phenotypes were subsequently validated by flow cytometry based surface marker expression (M1: CD14^low, CD163^low, CD11b^high; M2: CD14^high, CD163^high, CD11b^low) and by quantifying IL-10 and IL-12p40 production by Enzyme-Linked Immuno Sorbent Assay (ELISA) following stimulation of cells in the presence or absence of 100 ng/ml lipopolysaccharide (LPS) for 24h (In vivoGen, San Diego, United States) as described before (60).

Mycobacterial strains were cultured at 37°C in Difco™ Middlebrook 7H9 Broth (BD Bioscience) containing 10% acid-albumin-dextrose-catalase (ADC, BD Bioscience), 0.5% Tween-80 (Sigma-Aldrich), 2% Glycerol (Sigma-Aldrich) and 50 µg/ml hygromycin B (Thermo Fisher Scientific) when appropriate. The following bacterial strains were used: Mtb H37Rv, Venus-expressing Mtb H37Rv (strain mc²8120), DsRed-expressing Mtb H37Rv (18), MDR-Mtb Beijing family strain 16319 (62), MDR-Mtb Dutch outbreak strain 2003-1128 (62), Mycobacterium avium subsp. avium Chester strain 101 (ATCC, Wesel, Germany), DsRed-expressing Salmonella enterica serovar Typhimurium (Stm) strain SL1344 (18) and GFP-expressing methicillin-resistant Staphylococcus aureus (MRSA) strain USA300 JE2.

Adherent cells were harvested by trypsinization and gentle scraping in FBS and seeded in RPMI 1640 medium supplemented with 10% FBS and 2 mM L-Alanyl-L-Glutamine on Costar 96-well flat bottom culture plates (30,000 cells/well) or on Costar 24-well flat bottom culture plates (300,000 cells/well) (Corning, Amsterdam, The Netherlands) and incubated overnight at 37°C/5% CO₂. Mycobacterial cultures were diluted to an early log-phase corresponding with an OD600 of 0.25 one day prior to infection in 7H9 broth containing 10% ADC, 0.5% Tween-80 and 2% Glycerol. DsRed-expressing Stm was recovered from frozen stock and cultured in Difco™ Luria-Bertani (LB) broth (BD Bioscience) containing 100 µg/ml ampicillin (Sigma-Aldrich) overnight at 37°C. Bacterial suspension was diluted 1:33 in LB broth and grown for 3-4h to reach a log-phase with an OD600 between 0.4-0.6. GFP-expressing MRSA was recovered from frozen stock and cultured in Tryptic Soy (TS) broth (BD Bioscience) containing 5 µg/ml tetracycline hydrochlorine overnight at 37°C. Bacterial suspension was diluted 1:33 in TS broth and grown for 2-3h to reach a log-phase with an OD600 between 0.4-0.6. MRSA was harvested by centrifugation and resuspended in cold PBS supplemented with 5 mM EDTA. The indicated bacteria were diluted in RPMI 1640 medium (10% FBS and 2 mM L-Alanyl-L-Glutamine) to reach a multiplicity of infection (MOI) of 10. Accuracy of the MOI was validated by plating a serial dilution of the mycobacterial inoculum on Difco™ Middlebrook 7H10 agar (BD Bioscience) plates containing 10% oleic acid-albumin-dextrose-catalase (OADC, BD Bioscience) and 5% glycerol, by plating Stm on Difco™ LB agar plates (BD Bioscience) or by plating MRSA on TS agar (BD Bioscience). For mock infections, 7H9 broth was diluted in RPMI 1640 medium (10% FBS and 2 mM L-Alanyl-L-Glutamine) in equal concentrations (v/v) as the infection inoculum. Cells in flat-bottom 96-well plates containing 100 µl RPMI 1640 medium (10% FBS and 2 mM L-Alanyl-L-Glutamine) per well were inoculated with 100 µl bacterial suspension or mock solution. Cells in flat-bottom 24-well plates containing 500 µl RPMI 1640 medium (10% FBS and 2 mM L-Alanyl-L-Glutamine) per well were inoculated with 500 µl bacterial suspension or mock solution. Plates were centrifuged for 3 min at 800 rpm to increase bacterial uptake and incubated for 1h for Mtb, Mav or MRSA infection or for 20 minutes for Stm infection at 37°C/5% CO₂. Extracellular bacteria were removed by washing with fresh RPMI 1640 medium supplemented with 10% FBS, 2 mM L-Alanyl-L-Glutamine and 30 µg/ml gentamicin sulphate for 10 min. Cells were incubated at 37°C/5% CO₂ in RPMI 1640 medium supplemented with 10% FBS and 2 mM L-Alanyl-L-Glutamine and 5 µg/ml gentamicin sulphate in the presence of H-89 (10 µM), A-966492 (10 µM), niraparib (10 µM), pamiparib (10 µM), rucaparib (10 µM), 6-AN (100 µM), FX-11 (100 µM), Baf (10 nM), INH (0.4 μg/ml), RIF (0.05 μg/ml) or an equal amount of vehicle control DMSO (0.1% v/v) until readout. To calculate the coefficient of drug interaction (CID) the following formula was used: CID = AB/(AxB), where A indicates the ratio between antibiotic to the control group (DMSO, without antibiotics), B indicates the ratio between HDT compound to the control group and AB indicates the ratio between the treatments combined to the control group (63). A CID of 1 indicates an additive effect, <1 a synergistic effect and >1 an antagonistic effect.

Cells in 96-well flat bottom plates (30,000 cells/well) were washed with PBS and lysed in 0.05% sodium dodecyl sulfate (SDS) solution (Thermo Fisher Scientific). Serially diluted cell lysates were plated on 7H10 Agar containing 10% OADC and 5% Glycerol (Mtb and Mav), LB agar (Stm) or TS agar (MRSA) and incubated at 37°C. CFUs were determined from triplicate wells.

The number and percentage of dead cells based on plasma membrane integrity of the adherent cell population was quantified by analysis of microscopy images. Cells in 96-well flat bottom plates (30,000 cells/well) were stained with 2 µg/ml propidium iodide (Sigma-Aldrich) and 2 µg/ml Hoechst 33342 (H3570, Sigma-Aldrich) in 40 µl/well phenol red-free RPMI (Sigma-Aldrich) supplemented with 10% FBS and 2 mM L-Alanyl-L-Glutamine and incubated for 5 min at room temperature (RT). Cells were imaged using a Leica AF6000 LC fluorescence microscope (Leica Microsystems, Wetzlar, Germany) combined with a 10x dry objective. Total and dead cell numbers were quantified by respectively counting the nuclei and the number of propidium iodide-positive cells using ImageJ software (64). STSP (2.5 µM) was included as a positive control for cell death ( Supplemental Figure 4 ).

Compounds were diluted in Difco™ Middlebrook 7H9 Broth (Mtb, Mav), Difco™ LB Broth (Stm) or TS Broth (MRSA) and were added to log-phase bacterial cultures in flat bottom 96-well plates (OD₆₀₀ of 0.1). RIF (20 µg/ml) was added as positive control for reduction of Mtb and Mav growth and gentamicin (50 µg/ml) was added as positive control for reduction of Stm and MRSA growth. Bacterial plates were incubated at 37°C. Absorbance was measured directly after plating and at indicated times at a 550 nm wavelength on a Mithras LB 940 plate reader (Berthold Technologies, Bad Wildbad, Germany).

Cells (300,000 cells/well in 24-wells plates) were lysed with 100 µl/well EBSB buffer (10% v/v glycerol, 3% SDS, 100 mM Tris-HCl, pH 6.8) supplemented with one tablet of cOmplete™ EDTA-free protease inhibitor cocktail (Sigma-Aldrich) and one tablet of phosphatase inhibitor cocktail (PhosSTOP EASYpack, Sigma-Aldrich) per 10 ml. Cell lysates were boiled for 10 minutes at 95°C and stored at -20°C until use. Total protein concentrations were determined using a Pierce™ BCA protein assay kit (Thermo Fisher Scientific) according to manufacturer’ instructions and protein concentrations were equalized and diluted in Laemmli sample buffer (Biorad) containing β-mercaptoethanol (Sigma-Aldrich). Samples were loaded on a 15-well 4–20% Mini-PROTEAN^® TGX™ Precast Protein Gel (Bio-Rad Laboratories, Veenendaal, the Netherlands) and Amersham ECL Rainbow Molecular Weight Marker was added as reference (Sigma-Aldrich). Proteins were transferred to methanol-activated Immun-Blot PVDF membranes (Biorad) in Tris-glycine buffer (25 mM Tris, 192 mM glycine and 20% methanol). Membranes were blocked for 1h in polysorbate 20 tris-buffered saline (TTBS) supplemented with 5% w/v non-fat dry milk and incubated with the indicated antibodies in 5% non-fat dry milk/TTBS overnight at 4°C. Membranes were washed thrice for 15 min in TTBS and stained with secondary antibodies in 5% non-fat dry milk/TTBS for 2h at RT. Membranes were washed for 30 min with TTBS prior to revelation using enhanced chemiluminescence (ECL)™ Prime Western Blotting System reagent (GE Healthcare, Hoevelaken, The Netherlands). Imaging was performed on ChemiDoc XRS+ (Bio-Rad) or on an iBright Imaging System (Invitrogen, Breda, The Netherlands). Protein bands were quantified using ImageJ/Fiji software (64) and normalized to actin.

Human cytokine/chemokine levels were determined using the MilliPlex Human Cytokine/Chemokine magnetic bead premixed 41-plex kit (Millipore Billerica, MA, USA) as described before (60). Culture supernatants were collected and sterilized by using a 96-well filter plate containing a 0.2 µm PVDF membrane (Corning) following centrifugation. The following analytes were measured on a Bio-Plex 100 with Bio-Plex ManagerTM software v6.1 (Biorad): sCD40L, EGF, FGF-2, Flt3 ligand, Fractalkine (CX3CL1), G-CSF, GM-CSF, GRO (CXCL1), IFN-γ, IFN-α2, IL-1α, IL-1β, IL-1RA, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-8 (CXCL8), IL-9, IL-10, IL-12p40, IL-12p70, IL-13, IL-15, IL-17a, IP-10 (CXCL10), MCP-1 (CCL2), MCP-3 (CCL7), MDC (CCL22), MIP-1α (CCL3), MIP-1β (CCL4), PDGF-AB/BB, RANTES (CCL5), TGF-α, TNF-α, TNF-β, VEGF, Eotaxin (CCL11) and PDGF-AA.

Adherent and floating cells (300,000 cells/well in 24-wells plates) were washed in PBS/0.1% BSA (Merck, Darmstadt, Germany) and Fc receptors were blocked with 5% human serum (HS) in PBS for 10 min at RT. Cells were washed in PBS/0.1% BSA and stained with antibodies diluted in PBS/0.1% BSA for 30 min at 4°C. Cells were washed in PBS/0.1% BSA and fixed with 1% paraformaldehyde for 1h at 4°C. Fluorescence minus one (FMO) control samples were included to define background fluorescence for each stain. Fluorescence staining was measured using a FACSLyric™ flow cytometer with FACSDiva software (BD Bioscience). Geometric mean fluorescence intensity (gMFIs) was recorded for each sample. Data were analyzed using FlowJo v10 software.

Cell culture supernatants were collected and sterilized by using a 96-well filter plate containing a 0.2 µm PVDF membrane (Corning) following centrifugation. Undiluted sodium l-lactate (Sigma-aldrich) standard (0-8 mM, 5 µl) or sample (5 µl) was added to a flat-bottom 96-well plate (Greiner Bio-One, Alphen a/d Rijn, The Netherlands). 200 µl of reaction mix (0.74 mM NAD, Roche Applied Science, Woerden, The Netherlands; 0.4 mM glycine, Sigma-Aldrich; 0.4 M hydrazine hydrate, Sigma-Aldrich) was added to allow conversion of lactate to pyruvate by lactate dehydrogenase (LDH): Lactate + NAD⁺ <–> Pyruvate + NADH + H⁺. Hydrazine hydrate was added to avoid conversion of the newly formed pyruvate back to lactate. Baseline NADH levels were measured using the SpectraMax i3x plate reader at OD340 before the addition of 2 µl of three times diluted LDH from rabbit muscle (Roche Applied Science) to each well to initiate the lactate to pyruvate conversion. Following the addition of LDH, plates were incubated on a shaker at RT for 90 mins and the OD340 was measured again using a SpectraMax i3x plate reader.

For confocal microscopy, cells were cultured on pre-washed black glass bottom poly-d-lysine coated 96-well plates (no. 1.5, MatTek Corporation, Ashland, MA, USA) at a density of 30,000 cells/well in RPMI 1640 medium supplemented with 10% FBS and 2 mM L-Alanyl-L-Glutamine and incubated overnight at 37°C/5% CO₂. Cells were infected with DsRed-expressing Mtb H37Rv as described above and treated with compound (10 µM) or an equal volume of DMSO (0.1% v/v) for 4h. Culture medium was replaced with a solution of 75 nM LysoTracker^® Deep Red (Thermo Fisher Scientific) and 1x CYTO-ID^® Green 2.0 (1:500, Enzo Life Sciences, Bruxelles, Belgium) for 30 min at 37°C/5% CO₂ to stain lysosomal and autophagic vesicles, respectively. Cells were washed twice with PBS and fixed with Pierce™ 1% w/v formaldehyde (Thermo Fisher Scientific) for 1h at RT, washed twice with PBS and then stored at 4°C. Prior to imaging, plates were incubated with 2 µg/ml Hoechst 33342 for 5 min at RT and imaged using a SP8WLL confocal microscope (Leica) using a 63X oil immersion objective. Each treatment condition was performed in triplicate wells and three images were taken from each well. For image analysis, lysotracker and CYTO-ID channel background was subtracted using a rolling ball algorithm with a 10 pixel radius in Fiji/ImageJ (64). Lysotracker area and CYTO-ID area were specified for each image using Cellprofiler 3.1.9 (65) and were normalized to cell count based on Hoechst 33342 staining.

All statistical analyses were carried out using GraphPad Prism 8 software (Graphpad Software, San Diego, CA, USA). Normal distribution of data sets was evaluated using the Shapiro-Wilk normality test. Correlation was evaluated with a Spearman’s rank correlation test. Parametric paired data were tested with a paired t-test when comparing two groups and with RM one-way ANOVA or two-way ANOVA followed by Dunnett’s multiple comparison test versus control when comparing three or more groups. Nonparametric paired data were tested with Wilcoxon matched-pairs signed rank test when comparing two groups and with Friedman’s test followed by Dunn’s multiple comparison test versus control when comparing three or more groups. Statistical tests were considered significant when p<0.05 at 95% confidence interval.