Bazedoxifene Suppresses Intracellular Mycobacterium tuberculosis Growth by Enhancing Autophagy

Since current strategies for the treatment of multidrug-resistant tuberculosis (MDR-TB) and extensively drug-resistant tuberculosis (XDR-TB) have low efficacy and highly negative side effects, research on new treatments including novel drugs is essential for curing drug-resistant tuberculosis. Host-directed therapy (HDT) has become a promising idea to modulate host cell responses to enhance protective immunity against pathogens. Bazedoxifene (BZA), which belongs to a new generation of SERMs, shows the ability to inhibit the growth of M. tuberculosis in macrophages and is associated with autophagy. Our findings reveal a previously unrecognized antibacterial function of BZA. We propose that the mechanism of SERMs action in macrophages may provide a new potential measure for host-directed therapies against TB.

T uberculosis (TB), a chronic infectious disease caused by Mycobacterium tuberculosis, remains as a leading killer worldwide. In 2018, there were more than 10 million new cases and 1.57 million deaths from TB occurred (1). Although current antibiotics are effective at treating the majority of active TB cases, there is still a relapse rate of up to 9% even after completion of a 6-to 9-month treatment regimen (2,3). Moreover, the emergence of drug-resistant TB (4), as in the case of multiple-drug-resistant TB (MDR-TB) and extensively drug-resistant TB (XDR-TB) (5), has further impeded the effectiveness of the current regimen. This situation has led to the emergence of a new paradigm in TB drug discovery that involves therapeutic modulation of host cell functions in order to improve pathogen eradication, namely, host-directed therapy (HDT) (6). Since the mechanisms underlying HDT are completely different from those for antibiotics, which directly kill the mycobacteria, HDT has become an attractive strategy to shorten treatment regimens and to treat TB patients infected with drug-resistant M. tuberculosis strains (6,7). Tamoxifen (TAM), known as a selective estrogen receptor modulator (SERM), is widely used for the treatment of breast cancer (8). It has been initially suggested to have antimycobacterial activity using the computationally predicted "TB drugome" approach-a method for drug repositioning for tuberculosis treatment (9). In line with this, a recent study revealed that TAM exhibits antimycobacterial activity against drug-resistant M. tuberculosis and inhibits intracellular M. tuberculosis growth in macrophages (10). However, the underlying mechanism remains largely unknown. In particular, it is uncertain whether inhibition of intracellular M. tuberculosis growth by TAM is due to direct killing of bacteria or to modulation of bactericidal activity of macrophages. In addition, it remains to be elucidated whether other SERMs exhibit similar antimycobacterial effects in human macrophages.
In this study, we investigated whether bazedoxifene (BZA), a newer SERM that is effective in TAM-resistant breast cancer and has a high safety profile (11), has an antimycobacterial effect. We found that BZA efficiently inhibits intracellular M. tuberculosis growth in human macrophages.
As well, the autophagic machinery is a fundamental cellular response to defense against infectious pathogens, including M. tuberculosis infections (12). During autophagy, M. tuberculosis is sequestered within a double-membrane cytosolic vesicle called the autophagosome (13). After that, autophagosomes mature into autolysosomes through fusion with lysosomes, which leads to the digestion and death of the microbes (14) and subsequently activates the innate and adaptive immunity (15)(16)(17). In this study, we also identified that BZA inhibition is associated with enhanced autophagy.

RESULTS
BZA inhibits intracellular M. tuberculosis growth in macrophages. To investigate whether BZA has a direct antimycobacterial activity similar to TAM, we treated Mycobacterium smegmatis and M. tuberculosis H37Ra with BZA at different concentrations directly. Our data showed that BZA did not inhibit the growth of M. smegmatis and M. tuberculosis H37Ra at concentrations up to 20 M (10.613 g/ml) (Fig. 1A). Although the growth rate of M. tuberculosis H37Ra was statistically different from the control group when the concentration was up to 20 M, the bacterial growth inhibition efficiency was very low, less than 20% (Fig. 1B). While exhibiting a moderate inhibitory effect on M. smegmatis and M. tuberculosis H37Ra growth at a concentration higher than 200 M, BZA at this concentration caused significant cytotoxicity with 52% cell viability (see Fig. S1 in the supplemental material).
We next investigated whether BZA could inhibit M. tuberculosis growth indirectly through enhancing macrophage bactericidal activity. We observed that when used at a dose (5 M) that did not display apparent inhibitory effect on M. tuberculosis growth in 7H9 medium, BZA significantly inhibited both M. tuberculosis H37Ra and H37Rv growth in THP-1 macrophages as determined by counting CFU at 72 h postinfection ( Fig. 1C and D). These results indicated that BZA significantly suppresses intracellular M. tuberculosis growth through modulating macrophage functions.
BZA inhibits intracellular M. tuberculosis growth in macrophages through enhancing autophagy. We next investigated the mechanism underlying BZA inhibition of intracellular M. tuberculosis growth. Since autophagy is critical for macrophages to inhibit M. tuberculosis growth (18,19) and previous studies have demonstrated that BZA induces autophagy in cancer cells (20), we first investigated the effect of BZA on autophagy induction. As expected, we found that BZA treatment significantly increased autophagy in macrophages, regardless of M. tuberculosis infection, as evident from a significantly increased conversion of soluble LC3B-I to lipid bound LC3B-II ( Fig. 2A and Fig. S2A). To further examine the effect of BZA on autophagy flux, we monitored LC3B aggregation in autophagosomes and autolysosomes using the stably transformed monomeric red fluorescent protein (mRFP)-green fluorescent protein (GFP)-LC3 reporter THP-1 cells. Our result shows that BZA treatment significantly increased autophagosome and autolysosome forming punctate in H37Ra-infected cells compared with their mock-treated counterparts. (Fig. 2B to D), suggesting that enhanced autophagy contributes to BZA-mediated suppression of intracellular M. tuberculosis growth.
To determine the causal relationship between the effect of BZA on intracellular M. tuberculosis growth and its role in inducing autophagy, bafilomycin A1 (Baf), an autophagy inhibitor that prevents maturation of autophagic vacuoles by inhibiting fusion between autophagosomes and lysosomes (21), was used to block autophagy induced by BZA. As expected, the cumulative amount of LC3B-II increased after using Baf in H37Ra-or H37Rv-infected macrophages treated with BZA ( Fig. S2B and C), indicating that the increased accumulation of LC3B-II promoted by BZA is not due to the inhibition of downstream autophagic flow. Another autophagy inhibitor, 3-methyladenine (3-MA), a phosphatidylinositol 3-kinase (PI3K) inhibitor (22), was also used to block autophagy in the presence of BZA, and intracellular M. tuberculosis load was determined by CFU counting. As expected, 3-MA significantly reduced BZAinduced autophagy as determined by immunoblotting of LC3B-II conversion (Fig. 3A) as Values that are significantly different by one-way ANOVA and two-tailed Student's t test are indicated by asterisks as follows: *, P Ͻ 0.05; **, P Ͻ 0.01; ***, P Ͻ 0.001; ****, P Ͻ 0.0001 (ns, not significant). well as by counting and analyzing autophagosome and autolysosome formation in mRFP-GFP-LC3B reporter cells ( Fig. 3B to D). Importantly, the inclusion of 3-MA almost completely abrogated BZA-mediated suppression of M. tuberculosis growth within macrophages (Fig. 3E), indicating the effect of BZA on intracellular M. tuberculosis growth was dependent on its role in regulating autophagy.
To confirm the role of autophagy in BZA-mediated suppression of intracellular M. tuberculosis growth, we knocked down ATG5, a key protein in autophagy initiation and processing (23), and analyzed the effect of BZA on macrophage bactericidal activity. As demonstrated in Fig. 3F, ATG5 expression was significantly reduced following the treatment with ATG5-specific small interfering RNA (siRNA) for 36 h (Fig. 3F). In agreement with 3-MA treatment experiments, we found that ATG5 knockdown almost completely abolished BZA-mediated inhibition of intracellular growth of M. tuberculosis (Fig. 3G). However, we did not observe a significant increase of CFU in the siATG5 group compared to the control (Fig. 3G), which might be due to the knockdown efficiency, as a previous study showed that the complete knockout of ATG5 showed only a slight increase in CFU compared to wild-type (WT) macrophages (24). Taken together, these data confirmed that BZA suppresses M. tuberculosis growth in macrophages through enhancing autophagosome formation.  were calculated. (E) PMA-differentiated THP-1 macrophages were infected with H37Ra (MOI ϭ 10:1) for 6 h. After washing three times with prewarmed sterile phosphate-buffered saline (PBS) to remove extracellular bacteria, the infected THP-1 cells were treated with or without BZA (5 M) and 3-MA (10 mM) for another 72 h. Cells were lysed in 0.1% SDS and plated on 7H10 plates. The bactericidal activity of macrophages was assessed by determining CFU of intracellular H37Ra. (F) PMA-differentiated THP-1 macrophages were transfected with ATG5 small interfering RNA (siRNA) for 36 h. Scrambled siRNA was used as a negative control. These macrophages were then infected with H37Ra (MOI ϭ 10:1) and treated with BZA (5 M) as described above. The ATG5 protein level was analyzed by Western blotting. ATG5/␤-actin ratios are shown above the protein band. (G) The bactericidal activity of macrophages treated as described above was assessed by CFU. The data represent the means Ϯ standard deviations (SD) for three independent experiments. Values that are significantly different by one-way ANOVA are indicated by asterisks as follows: *, P Ͻ 0.05; **, P Ͻ 0.01; ***, P Ͻ 0.001; ****, P Ͻ 0.0001 (ns, not significant).

BZA-induced autophagy is associated with ROS production and phosphorylation of Akt/mTOR signal.
To determine the mechanisms by which the BZA modulates macrophage bactericidal capability against M. tuberculosis, we also investigated the effect of BZA on M. tuberculosis-induced reactive oxygen species (ROS) production, one of the pivotal mechanisms for the killing of intracellular M. tuberculosis by macrophages (25). We found that mitochondrial ROS (mROS) production (Fig. 4A) was significantly increased in M. tuberculosis-infected macrophages upon treatment with BZA at 24 h postinfection, whereas cytoplasmic ROS (cROS) was not significantly changed (Fig. 4B).
It has been reported that mitochondrial ROS is an important mediator of autophagy (26,27). We therefore used N-acetyl-L-cysteine (NAC), a ROS-scavenging agent, in the presence of BZA treatment to test whether increased mROS contributes to enhanced autophagy induced by BZA in M. tuberculosis-infected macrophages. In line with previous reports (28,29), we found that addition of NAC abolished the ability of BZA to induce autophagy (Fig. 4C).
Phosphorylation of mTOR and Akt is a key signaling step for the initiation of autophagy (30,31). To gain an insight into the signaling for BZA-induced autophagy, we determined the phosphorylation of mTOR, Akt, ULK, AMPK, and STAT3. There was no difference in the extent of ULK, AMPK, and STAT3 phosphorylation between the presence and absence of BZA treatment (Fig. 5A). In contrast, the levels of phospho-mTOR and phospho-Akt in BZA-treated M. tuberculosis-infected macrophages were significantly reduced (Fig. 5B) in comparison with those without BZA treatment. These results imply that BZA-induced autophagy is associated with phosphorylation of mTOR and Akt signaling.

DISCUSSION
With the emergence of drug-resistant TB, there is an urgent need for new TB treatment strategies. Repurposed FDA-approved drugs for anti-TB therapy are an attractive approach due to the safety profile of the approved products. In this study, we show that BZA, a newer SERM, significantly inhibits M. tuberculosis growth in human macrophages. This effect of BZA is mediated through enhancing autophagy in infected macrophages. We propose that BZA treatment increases mROS production and promotes autophagosome formation in macrophages, which ultimately enhances the passage of M. tuberculosis into the phagolysosome for pathogen destruction.
BZA is an indole-based estrogen receptor (ER) ligand with a unique structure, which differs from TAM and raloxifene (32) but exhibits similar functions (33). BZA is well tolerated and possesses improved inhibitory potency against the Y537S and D538G ERa mutants compared to TAM (34). In contrast to previous reports that TAM has direct antimycobacterial activity at a low concentration (16 M) (10), we found that BZA displays only moderate antimycobacterial activity in 7H9 culture medium at a high concentration (200 M). Therefore, these findings suggest that TAM and BZA limit mycobacterial growth through both shared and distinct mechanisms. The exact reason underlying the difference between TAM and BZA in extracellular bacterial killing is currently unknown, although this could be due simply to the intrinsic differences in their chemical structure. However, consistent with TAM and raloxifene (10,35,36), we found that BZA significantly enhanced the macrophage bactericidal activity against M. tuberculosis. More importantly, we confirmed that this effect of BZA is achieved through regulating autophagy activity in macrophages. By simultaneously analyzing the amount of LC3B-II and the formation of the autophagosome and autolysosome, we inferred that BZA facilitates autophagic activity and therefore enhances the elimination of intracellular M. tuberculosis via autophagosome synthesis and lysosome degradation.
How autophagy is regulated in M. tuberculosis-infected macrophages is complex. It has been reported that ROS can induce autophagy (37), one means of which is through the inhibition of Akt signaling, and mTOR (38). In line with this report, we found that mROS production was elevated after BZA treatment in M. tuberculosis-infected macrophages and that depletion of mROS by NAC abolished BZA-induced autophagy. However, our data have not addressed the exact signal pathway from ROS to autophagy during BZA treatment. Although we demonstrated that the disruption of autophagy abolished the effect of BZA and that BZA-induced autophagy was dependent on the upregulation of mROS production, we cannot exclude and discriminate the contribution of direct killing bacteria by mROS.
In summary, we have demonstrated that BZA significantly inhibits intracellular M. tuberculosis growth in macrophages through enhancing ROS production and autophagy. These findings suggest that BZA may be considered a host-directed drug candidate for TB treatment and that further in vivo investigations are warranted.

MATERIALS AND METHODS
Bacterial strains and culture conditions. M. smegmatis mc 2 155 and M. tuberculosis strains H37Ra and H37Rv were used. The bacteria were grown in Middlebrook 7H9 broth (BBL Microbiology Systems) supplemented with 10% oleic acid-albumin-dextrose-catalase (OADC; Becton, Dickinson), 0.05% Tween 80 (Sigma), and 0.2% glycerol (Sigma), for 5 to 7 days at 37°C with shaking to achieve mid-logarithmic phase (optical density at 600 nm [OD 600 ] ϭ 0.3 to 0.8).The bacteria were resuspended in serum-free RPMI medium and sonicated to obtain a single-cell suspension before being used in experiments. The concentration of bacteria was determined by the OD 600 as a function of CFU per milliliter.
Cell culture and chemicals. Cells of the human monocytic cell line THP-1 were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). THP-1 cells and THP-1 cells transformed with an mRFP-GFP-LC3B reporter (provided by Hongbo Shen, Institute Pasteur of Shanghai, China) were maintained in RPMI 1640 (Corning) culture medium supplemented with 10% fetal bovine serum (FBS; Invitrogen, Life Technologies) at 37°C and 5% CO 2 . The THP-1 cells were seeded at 5 ϫ 10 5 cells/ml in a 12-well plate in complete RPMI 1640 and differentiated using phorbol 12-myristate 13-acetate (PMA; Sigma, P8139) at 40 ng/ml for 24 h. Following removal of the PMA-containing medium, the cells were incubated in the fresh complete RPMI 1640 medium for 12 h at 37°C for further use.
In vitro antimycobacterial activity was tested with M. smegmatis and M. tuberculosis H37Ra. The bacteria were diluted with 7H9-OADC to an OD of 0.01 and transferred to a 96-well microtiter plate for the survival assay. Then, BZA with a serial dilution of 10-fold in 7H9 medium was added to the bacterial culture. DMSO was added as a negative control. The bacterial plates were incubated at 37°C with horizontal shaking (110 rpm). The absorbance (OD 600 ) was measured with an Epoch 2 microplate spectrophotometer (BioTek, USA) at indicated times, and the growth rate of bacteria was calculated.
Cell viability and proliferation assay. THP-1 cells (5 ϫ 10 3 cells/well) were incubated in a 96-well plate with 40 ng/ml PMA for 24 h. Cells were then maintained in fresh prewarmed complete RPMI 1640 for 12 h at 37°C for further use. Different concentration of BZA (200 M, 20 M, and 2 M) and DMSO, as a negative control, were added and cultured for an additional 72 h. WST-1 reagent (Beyotime, Shanghai, China) was added into the well (10 l/well) and incubated at 37°C for 2 h to measure cell viability and proliferation. The absorbance was measured at 450 nm and 690 nm with an Epoch 2 microplate spectrophotometer (BioTek, USA), and cell viability was calculated according to the manufacturer's instructions.
CFU assays. PMA-differentiated THP-1 macrophages (5 ϫ 10 5 cells) were infected with two laboratory M. tuberculosis strains (H37Ra and H37Rv) at a multiplicity of infection (MOI) of 10 for 6 h at 37°C, 5% CO 2 . After washing three times with prewarmed sterile phosphate-buffered saline (PBS) to remove extracellular bacteria, the infected THP-1 cells are treated with or without BZA (5 M) for another 72 h. Cells were washed three times with PBS and then lysed with 500 l of PBS containing 0.1% SDS. Triple experimental groups for each treatment were plated on Middlebrook 7H10 agar (Difco; Middlebrook) supplemented with 10% OADC and incubated vertically at 37°C for 3 weeks, and then the colonies were counted. CFU obtained from two or three dilutions was used to calculate the total number of CFU per milliliter. Data were presented as the amount of M. tuberculosis in 5 ϫ 10 5 cells and the percentage of mycobacterial survival in BZA-treated and untreated cells.
Measurement of H37Ra in lysosomes and colocalization with LC3B. mRFP-GFP-LC3B reporter THP-1 macrophages (1 ϫ 10 5 cells/ml) were differentiated using PMA and infected with mycobacterial strain H37Ra (MOI of 10:1) for 6 h. After washing with PBS, the cells were cultured with or without BZA (5 M) for 12 h. The cells were then washed twice in PBS and with 4% paraformaldehyde solution (Solarbio, China) for 15 min at room temperature. Cells were stained with Hoechst stain and mounted with antifade agent (Invitrogen, USA). The cells were visualized, and images were acquired with an Olympus FV1000 confocal microscope (Nikon A1R) and processed using Image J software (NIH, USA). A total of 30 to 40 infected cells were analyzed, and the level of autophagy was measured by enumerating the number of LC3 puncta per cell.
Measurement of ROS. PMA-differentiated THP-1 macrophages (5 ϫ 10 5 cells/ml) were cultured in 12-well plates and infected with H37Ra (MOI of 10:1) for 6 h. Extracellular bacteria were removed by washing 3 times with prewarmed sterile PBS. The cells were cultured in fresh medium with or without BZA (5 M) for 24 h. The generation of cytoplasmic ROS (cROS) was quantified using 5-(and -6)chloromethyl-2=,7=-dichlorodihydrofluorescein diacetate, acetyl ester (CM-H2DCF-DA) (Invitrogen, USA), according to the manufacturer's instructions. The generation of mitochondrial ROS (mROS) was quan-tified using MitoSOX red mitochondrial superoxide indicator (Invitrogen, USA). The cells were incubated with H2DCF-DA (10 M) and MitoSOX (10 M) for 15 min at 37°C in the dark. All fluorescence intensities of cells were measured by flow cytometry (BD FACSAria II flow cytometer) and analyzed by FlowJo 6.1 according to the manufacturer's protocol.
RNA interference and transfection. PMA-differentiated THP-1 macrophages (5 ϫ 10 5 cells/ml) were transfected with ATG5 siRNA (RiboBio; siB12531154855-1-5; China) using Lipofectamine RNAiMAX (Ther-moFisher, USA) according to the manufacturer's protocol. The siRNA and Lipofectamine complexes were prepared in Opti-MEM (Gibco) at a 1:1 ratio. Negative control (NC) vector without any insert was used as a negative control. The medium was replaced after 36 h, cells were washed twice with PBS before lysing, and the efficiency of knockdown was determined by Western blotting. For the CFU test, THP-1 cells were infected with H37Ra (MOI of 10:1) for 6 h after 36 h of treatment with siRNA silencing. Extracellular bacteria were washed 3 times with prewarmed serum-free RPMI and treated with BZA for 72 h.
Western blot analysis. PMA-differentiated THP-1 macrophages (5 ϫ 10 5 cells/ml) were cultured in 12-well plates and infected with H37Ra (MOI of 10:1) with or without BZA (5 M) for 6, 12, or 24 h. After washing with PBS, the cells were lysed in RIPA lysis buffer (Cell Signaling Technology). The protein concentration of the resultant lysates was measured with a bicinchoninic acid (BCA) protein kit (Beyotime, China). The protein samples were boiled for 10 min. Equal amounts of proteins were adjusted according to the molecular weight of the protein, electrophoresed on SDS-PAGE gels, and then transferred onto a polyvinylidene difluoride membrane (Merck/Millipore). The membrane was blocked with 5% skim milk powder (BD) solution in PBS with Tween 20 (PBST) for 1 to 2 h at room temperature and then incubated with antibodies against LC3B (Sigma, USA), p-mTOR (Abcam, United Kingdom), p-Akt (Abcam, United Kingdom), actin (Abcam, United Kingdom), p-ULK (Cell Signaling Technology, USA), p-AMPKa (Cell Signaling Technology, USA), p-STAT3 (Cell Signaling Technology, USA), and ATG5 (Cell Signaling Technology, USA) at 4°C for 12 h. After washing with 1ϫ PBST for 10 min, the membrane was incubated with appropriate secondary antibody, goat anti-rabbit IgG-horseradish peroxidase (HRP) (Abcam, USA). After washing with 1ϫ PBST, the protein band was visualized by enhanced chemiluminescence (ECL) detection solution (ThermoFisher). The digital images of the protein bands were acquired using an ImageQuant LAS 4000 (GE Healthcare Bio-Science) system. Statistical analysis. Statistical analysis was performed using GraphPad Prism 7 software (GraphPad Software, Inc.). Two-tailed Student's t test was used to evaluate the measurement of one parameter. One-way analysis of variance (ANOVA) with multiple comparisons was used to analyze more than two groups. Data were expressed as mean Ϯ standard deviation (SD). Differences between groups were considered statistically significant when P was Ͻ0.05.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.