MicroRNA-889 Inhibits Autophagy To Maintain Mycobacterial Survival in Patients with Latent Tuberculosis Infection by Targeting TWEAK.

TB remains a leading cause of morbidity and mortality worldwide. Approximately one-quarter of the world’s population has latent TB infection. TWEAK is a multiple-function cytokine and may be used as a target for the treatment of rheumatic diseases, cardiovascular diseases, and renal diseases. Here, we demonstrated a novel relationship between TWEAK and activation of the autophagic machinery which promotes antimycobacterial immunity. Additionally, TB infection is highly dynamic and determined by the interaction between the host and mycobacterium. We demonstrated a mechanism of fine-tuned balance between the mycobacterium and host for granuloma formation and/or maintenance in LTBI status. Once patients entered LTBI status, the upregulation of miR-889 was associated with TNF-α levels and granuloma formation to maintain mycobacterial survival. Adalimumab (a TNF-α inhibitor) reduced both TNF-α and miR-889 levels and caused LTBI reactivation and, thus, TWEAK enhancement. MiR-889 and TWEAK may become potential diagnostic biomarkers or therapeutic targets for LTBI and LTBI reactivation, respectively.

terium. We demonstrated a mechanism of fine-tuned balance between the mycobacterium and host for granuloma formation and/or maintenance in LTBI status. Once patients entered LTBI status, the upregulation of miR-889 was associated with TNF-␣ levels and granuloma formation to maintain mycobacterial survival. Adalimumab (a TNF-␣ inhibitor) reduced both TNF-␣ and miR-889 levels and caused LTBI reactivation and, thus, TWEAK enhancement. MiR-889 and TWEAK may become potential diagnostic biomarkers or therapeutic targets for LTBI and LTBI reactivation, respectively.
KEYWORDS autophagy, biomarker, latent tuberculosis infection (LTBI), microRNA-889, tumor necrosis factor-like weak inducer of apoptosis (TWEAK) A utophagy is a natural defense mechanism that is conducted to eliminate microbial infection, playing a crucial role in controlling the fate of invading pathogens (1). Tuberculosis (TB) is a major cause of morbidity and mortality worldwide. Mycobacterium tuberculosis has developed a mechanism that prevents the autophagy of immune cells so that it can survive in host cells and remain dormant for a longer period, which is responsible for latent TB infection (LTBI) (2). Most individuals infected with M. tuberculosis have an LTBI, and this population is an important reservoir for disease reactivation (3). Increased evidence indicates an elevated TB risk in patients with rheumatoid arthritis (RA) (4,5); the risk is even higher for those receiving anti-tumor necrosis factor alpha (TNF-␣) therapy (6). Gardam et al. (7) revealed that active TB in RA patients receiving anti-TNF-␣ therapy appears to be largely caused by LTBI reactivation. The tuberculin skin test (TST) and interferon gamma (IFN-␥) release assays (IGRAs) are currently the commonly used methods to screen for LTBI (8). However, the clinical utility of TST is not reliable in bacillus Calmette-Guérin (BCG)-vaccinated patients (9), and it has a low specificity. Although the specificity of IGRAs is enhanced, the cost of IGRAs is high. Additionally, neither the TST nor IGRAs can discriminate between LTBI and TB disease (10).
MicroRNAs (miRNAs) are key regulators that posttranscriptionally repress the expression of target mRNAs (11). miRNAs can be detected in body fluids and are emerging as novel disease biomarkers (12). Accumulating validations demonstrate that miRNAs are regulators during mycobacterial infections and can be potential biomarkers for the diagnosis of TB diseases (13,14). Most of these miRNAs were involved in TB pathogenesis by targeting autophagy-related genes (ATGs) or regulating autophagic activity, and thus mycobacterial survival was maintained. However, the role of miRNAs in modulating host defenses in the LTBI period and their clinical relevance are unclear.
In the present study, we investigated the candidate miRNAs that were significantly associated with LTBI. We also explored the biological roles of the candidate miRNAs using a cell-based assay and an in vitro human TB granuloma model. Finally, we investigated the pathogenic role of candidate miRNAs in LTBI and LTBI reactivation.

RESULTS
Clinical characteristics of RA patients. A total of 120 participants, including 97 RA patients and 23 healthy subjects, were enrolled. Among the RA patients, 35 with LTBI had positive QuantiFERON-TB Gold (QFT-G) results, 12 had nontuberculous mycobacterial (NTM) diseases, and 50 did not have mycobacterial infections. There were no significant differences in ages, proportions of females, rates of positivity for rheumatoid factor or anti-citrullinated peptide antibodies, DAS28 values, daily dosages of corticosteroids, proportions of disease-modifying antirheumatic drugs used, or frequencies of comorbidity among RA patients with LTBI, NTM, and those without infection (see Table S1 in the supplemental material).
Differentially expressed miRNAs using NGS and quantitative reverse transcription-PCR (QRT-PCR). In the first stage of miRNA next-generation sequencing (NGS) analysis, 4 out of 7 RA patients had LTBIs and 3 were noninfected subjects. After normalization, we observed 17 miRNAs that were distinctively expressed in peripheral blood mononuclear cells (PBMCs); 6 miRNAs were upregulated and 11 miRNAs were downregulated in LTBI patients compared with their levels of expression in noninfected patients ( Fig. 1A; Table S2). Among the six upregulated miRNAs, miR-889 exhibited the highest fold change, 4.02-fold. miR-889 expression increased in human PBMCs with M. bovis BCG infection (P Ͻ 0.01) (Fig. 1B, left) or M. tuberculosis H37Rv (P Ͻ 0.01) (Fig. 1B, right) at a multiplicity of infection (MOI) of 0.1, and the expression level was time dependent. Additionally, miR-889 expression increased after mycobacterial infection in a dose-dependent manner (Fig. S1A). Therefore, we focused on miR-889 for the further validation.
MiR-889 targets TWEAK. Bioinformatics analysis (http://www.microrna.org) (15) revealed TNF-like weak inducer of apoptosis (TWEAK; GenBank accession number NM_003809) as a potential seed match for miR-889 in its 3= untranslated region (3=UTR) ( Fig. 2A). To validate whether TWEAK is the target of miR-889, a luciferase reporter plasmid was constructed by cloning the predicted seed sequence in the human TWEAK 3=UTR into the pMIR-REPORT luciferase vector; the plasmid with the mutation at the putative binding site was used as a control. Our results showed that miR-889 mimics significantly decreased (P Ͻ 0.01), while those of miR-889 inhibitors significantly enhanced, the luciferase activity in cells transfected with the TWEAK 3=UTR plasmid, compared to that of the cells transfected with the inhibitor control (P Ͻ 0.05). No significant change in luciferase activity was observed in cells transfected with the mutant TWEAK 3=UTR construct or the pMIR-REPORT plasmid (Fig. 2B), indicating that TWEAK is the target of miR-889 and may be negatively regulated by miR-889. No significant difference in levels of TWEAK mRNA expression existed between miR-889 mimic-expressing cells and control cells or nontransfected cells (Fig. 2C). However, a decreased expression of TWEAK protein levels ( Fig. 2D) was detected in miR-889overexpressing cells compared with that in control cells (P Ͻ 0.01), indicating that TWEAK may be negatively posttranscriptional regulated by miR-889.
Mycobacterium-induced TWEAK was upregulated in early infection and then declined in patients with LTBI. Next, we validated the association between TWEAK expression and mycobacterial infection. TWEAK expression increased after mycobacterial infection in a dose-dependent manner (Fig. S1B). Increased intracellular (Fig. 2E)   To support the findings with a cell-based assay, we examined the TWEAK levels in the serum of one patient with M. tuberculosis infection before and after anti-TB therapy. The results showed higher levels of TWEAK (1,607.7 Ϯ 79.2 pg/ml) (Fig. 2I) in this patient at the time of active TB status (based on a positive TB culture result) than in healthy controls (1,017.0 Ϯ 52.5 pg/ml) (Fig. 2J). After anti-TB therapy, a significantly decreased TWEAK level was measured, compared to that before therapy (1,263.8 Ϯ 49.8 pg/ml, P Ͻ 0.01) (Fig. 2I). We further validated the association of TWEAK levels with LTBI in RA patients and revealed significantly lower levels of TWEAK in LTBI patients (630.1 Ϯ 53.2 pg/ml) than in noninfected patients (805.1 Ϯ 40.7 pg/ml, P Ͻ 0.05) and healthy controls (1,017.0 Ϯ 52.5 pg/ml, P Ͻ 0.005) (Fig. 2J). Additionally, we demonstrated a negative correlation between TWEAK levels and released IFN-␥ levels in the QFT-G assay of LTBI patients (␥ ϭ 0.43, P Ͻ 0.005) (Fig. S1C). Similarly, TWEAK levels were elevated in LTBI patients after anti-LTBI therapy (630.1 Ϯ 53.2 pg/ml versus 805.7 Ϯ 55.6 pg/ml, P Ͻ 0.05) (Fig. 2K).
TWEAK induces autophagy and autophagosome formation in macrophages. Autophagy is a major defense mechanism inhibiting mycobacterial survival in infected macrophages. Bhatnagar et al. (16) demonstrated that TWEAK augments the expression of ATGs in myotubes. To verify whether TWEAK induces autophagy, THP-1 cell-derived macrophages were treated with TWEAK (100 ng/ml) and the levels of autophagyrelated LC3 were measured by immunoblotting. We revealed that TWEAK could induce autophagy in macrophages in dose-and time-dependent manners (Fig. S2). The autophagy inhibitor 3-methyladenine (3-MA; 5 mM) rescued TWEAK-induced autophagy, while bafilomycin A1 (luminal acidification and autophagosome degradation inhibitor, 100 nM) increased LC3-II expression in cells treated with TWEAK ( Fig. 3A and Using an immunofluorescence assay (IFA), we demonstrated an elevated endogenous LC3 punctum formation in THP-1 cells treated with TWEAK (Fig. 3B). The autophagic activity was further examined using THP-1 cells stably expressing the red fluorescent protein (RFP)-green fluorescent protein (GFP)-LC3 fusion protein. The results showed that TWEAK induced a redistribution of the RFP-GFP-LC3 fusion protein from a diffuse to a punctate pattern (Fig. 3C). Additionally, numbers of both yellow and red puncta increased in cells with TWEAK treatment (P Ͻ 0.01), suggesting that TWEAK induces autophagy with an increased autophagic flux. Autophagosome maturation and acidification are crucial host processes against mycobacterial infection. To further examine whether TWEAK can induce autophagosome acidification, THP-1 cells were treated with TWEAK or IFN-␥ (10 ng/ml), an inducer of autophagosome acidification (17), for 24 h and then incubated with a specific fluorescent marker for autophagosome acidification (LysoTracker Green). A significantly increased autophagosome acidification ratio was detected in cells with TWEAK treatment compared to that of control cells     (Fig. S5). The TWEAK effect on enhanced autophagosome acidification was counteracted by treatment with the classical inhibitor of autophagosome formation, 3-MA (8.7 Ϯ 0.8%, P Ͻ 0.005).
miR-889 inhibits mycobacterium-induced autophagy to maintain mycobacterial survival by targeting TWEAK. To dissect the biological function of miR-889 in mycobacterial infection, THP-1 cells were infected with BCG in the presence of miR-889 to mimic transfection, and then TWEAK expression levels were examined and quantified using an IFA and enzyme-linked immunosorbent assay (ELISA), respectively. As shown in Fig. S7A, a markedly higher expression of miR-889 in THP-1 cells occurred after transfection with the miR-889 mimic, indicating its effective transfection. The results of both the IFA (Fig. S7B) and the ELISA (Fig. 4J) showed increased TWEAK levels in the cells with BCG infection (16.1 Ϯ 1.8 pg/ml) compared with those without infection (9.0 Ϯ 0.1 pg/ml, P Ͻ 0.01). Significantly lower levels of TWEAK were observed in miR-889-overexpressing cells after 24 h of infection (13.2 Ϯ 1.2 pg/ml) than with mimic control-expressing cells (17.8 Ϯ 0.7 pg/ml, P Ͻ 0.05) or nontransfected cells (16.1 Ϯ 1.8 pg/ml).
Adalimumab affected miR-889 and TWEAK expression in BCG-induced granulomas. We further analyzed the miR-889 and TWEAK expression in biologic-related LTBI reactivation. At 8 days postinfection, the granuloma-like structure was constructed, and cells were treated with biologics. At 24 h posttreatment, a destructured granuloma was observed in cells treated with adalimumab ( Fig. 5A) compared with that in dimethyl sulfoxide (DMSO)-or rituximab-treated cells. Additionally, decreased miR-889 expression (0.68-fold Ϯ 0.21-fold, P Ͻ 0.05) (Fig. 5B) and increased TWEAK levels (22.4 Ϯ 1.4 pg/ml versus 4.2 Ϯ 0.4 pg/ml, P Ͻ 0.01) (Fig. 5C) were detected simultaneously. Besides adalimumab levels, TWEAK levels were significantly increased in cells treated with another TNF-␣ inhibitor (etanercept at 17.0 Ϯ 2.1 pg/ml, P Ͻ 0.05), but no significant change was observed in cells treated with rituximab, a B-cell depletion biologic).

Significantly increased TWEAK expression in patients with LTBI reactivation.
To evaluate whether miR-889 or TWEAK expression reflects the different statuses of TB infection, we examined the dynamic change of miR-889 (Fig. 5D) and TWEAK (Fig. 5E) in the sera of one LTBI patient receiving adalimumab therapy. The examination revealed an elevation of miR-889 expression (1.52-fold Ϯ 0.07-fold) and a decrease of TWEAK levels (580.0 Ϯ 17.9 pg/ml) before TNF-␣ inhibitor therapy. However, TWEAK levels increased at the time of LTBI reactivation (736.9 Ϯ 28.4 pg/ml, P Ͻ 0.01 versus baseline) (Fig. 5E), and active TB disease developed. After anti-TB therapy was completed, both the miR-889 and TWEAK levels returned to baseline values.

DISCUSSION
The immunopathogenesis of M. tuberculosis is complex (19) and is attributed mainly to immune-evading strategies that allow the pathogen to remain dormant after primary infection, persisting in the host and reactivating its pathogenicity under favorable conditions (20). It is estimated that nearly one-quarter of the world's population has been latently infected with M. tuberculosis, and this population is an important reservoir for disease reactivation. LTBI is an equilibrium status between the host and mycobacterium during infection. Host responses in controlling the latent infection may include macrophage activation, maintenance of granuloma structure, and expression of IFN-␥, TNF-␣, CD4 T cells, and CD8 T cells (21). In the present study, miR-889 and TWEAK in Patients with LTBI ® we revealed a novel mechanism of fine-tuned balance between mycobacteria and the host for granuloma formation and/or maintenance in patients with the LTBI status. Our results showed that miR-889 expression was significantly higher in RA patients with LTBI than in those without infection or in healthy controls, and its expression was positively correlated with released levels of IFN-␥ in the QFT-G assay (r ϭ 0.63, P Ͻ 0.001). A significant decrease of miR-889 levels paralleled the prophylactic therapy for LTBI patients, indicating that miR-889 expression might be involved in M. tuberculosis infection, which is consistent with the conclusions of another study (22). Given that TWEAK acts as the target of miR-889, we found significantly lower TWEAK levels in miR-889-overexpressing cells and in LTBI patients who had high levels of expression of miR-889 than in other subjects. Moreover, TWEAK levels increased at the time of LTBI reactivation. These findings suggest that miR-889 and TWEAK may act as candidate biomarkers for LTBI and LTBI reactivation, respectively. Further large studies are required to confirm our data. Autophagy is a well-conserved lysosomal degradation pathway that plays a key role in the innate defense mechanism against mycobacteria (23). Dormant M. tuberculosis can suppress autophagy and then survive within macrophages for an extended period, which is responsible for LTBI (24). Our results showed that a mycobacterial component (e.g., Ag85A, Ag85B) induced TWEAK upregulation. TWEAK, a member of the TNF superfamily, regulates several cellular responses, including proinflammatory activity, angiogenesis, and apoptosis (25). TWEAK balances TNF-␣ activity by repressing the production of proinflammatory cytokines in modulating the transition from innate to adaptive immunity (26). In the present study, we demonstrated that TWEAK induced autophagy and promoted autophagosome maturation against mycobacterial infection through activation of AMPK. AMPK is an essential metabolic regulator that plays an important role in the maintenance of energy balance in response to stress (27). Additionally, AMPK plays a crucial role in the initiation of autophagy with subsequent autophagosome formation and maturation (28). Activated AMPK induces the phosphorylation of the serine residues (e.g., Ser555) of ULK1, which is a mammalian autophagy-initiating kinase that plays a key role in starvation-induced autophagy (29). A previous study demonstrated that ULK1 was strongly associated with LTBI and may play a crucial role in the regulation of autophagy and mycobacterial replication (30). In addition to demonstrating autophagy, recent studies demonstrated that AMPK is involved in metabolic responses, fatty acid ␤-oxidation, and the control of pathological inflammation in macrophages during M. tuberculosis infection (31). We identified a novel role of TWEAK in AMPK activation in antimycobacterial autophagy. More, indepth studies of the regulatory mechanism of TWEAK in AMPK-targeted effector networks in mycobacterial infection are needed to confirm our finding.
TB infection is highly dynamic and determined by the interaction between the host and the mycobacterium (32). Granuloma formation plays a crucial role in TB pathogenesis (33). The cellular factors that control granuloma formation and maintenance are multifaceted, involving a complex interplay between the host immune system and mycobacterium survival strategies (32). TNF-␣ is critical for granuloma formation (18). Previous studies demonstrated a fine-tuning of TNF-␣ production in the host during TB infection, which allowed mycobacterial persistence in granulomas without apparent disease (LTBI status) (18,(32)(33)(34). Agarwal et al. (35) validated that mycobacteria can use this host granulomatous response to continue infection. In this study, we found an increased expression of TWEAK during early mycobacterial infection, which induced autophagy and promoted mycobacterial autophagosome maturation through the activation of AMPK (Fig. 6A). Our results revealed that the role of TWEAK in mycobacterial infection is similar to that of TNF-␣, which has been known to induce autophagy and phagosome maturation in macrophages (17,36). Previous studies demonstrated that once the organism entered latent TB infection status, some latency-associated proteins of M. tuberculosis maintained mycobacterial survival in granulomas by stimulating the expression of TNF-␣ (37) or inhibiting autophagy (38). Our results showed that TNF-␣ inhibits TWEAK expression in macrophages through upregulating miR-889 miR-889 and TWEAK in Patients with LTBI ® expression to maintain granuloma formation (Fig. 6B), supporting the findings that TNF-␣ is responsible for granuloma formation and maintenance in subjects with LTBI (18). More studies are needed to dissect the regulatory mechanism between TNF-␣ and TWEAK, from TB infection to LTBI. Additionally, upregulated miR-889 inhibited mycobacterium-caused autophagy to maintain mycobacterial survival in granulomas, suggesting that miR-889 expression might be associated with latent M. tuberculosis infection. The association between miR-889 expression and latency-associated proteins of M. tuberculosis requires further studies to dissect.
Increasing evidence reveals that TNF-␣ inhibitors are associated with an elevated risk of LTBI reactivation (39,40). Since TNF-␣ is involved in granuloma maintenance, neutralizing TNF-␣ tends to disrupt the granuloma structure, allowing mycobacteria to emerge from dormancy and develop into active TB (41). A previous study inferred that TNF-␣ inhibitors may reduce autophagy; then, the repressive effect on autophagy may be responsible for the increased TB risk (36). We revealed an upregulation of miR-889 by soluble TNF-␣; adalimumab then reduced TNF-␣ levels and thus affected the expression of miR-889, resulting in granuloma destruction (Fig. 6C). Once granuloma destruction and LTBI reactivation occurred, TWEAK expression might have been enhanced by mycobacterium induction. Our findings were consistent with another mouse LTBI granuloma model showing increased TWEAK levels after treatment with TNFneutralizing antibody (42). Additionally, significantly elevated TWEAK levels were observed in our adalimumab-treated patient exhibiting LTBI reactivation, which may be associated with increased mycobacterial stimulation. These observations suggest that TWEAK may be a potential biomarker for predicting LTBI reactivation during the period of biologic therapy.
This study had some limitations. First, it lacked LTBI subjects without rheumatic disease, and the number of healthy control subjects was small. Previous studies demonstrated that deregulation of the autophagic pathway and TWEAK/Fn14 signaling are involved in the pathogenesis of RA (43,44), which may be risks of bias in this study. Additionally, the study was cross-sectional by design, and thus, the possibility that miRNA expression changed with therapeutic strategies cannot be excluded. However, we validated our observations by using an in vitro cell-based assay, suggesting that our results still provided valuable information.
The mycobacterium-containing granuloma represents a dynamic balance between host and pathogen (45). Modulation of host-specific immune responses can limit mycobacterial infection. TWEAK is a multifunctional cytokine and may be used as a therapeutic target for rheumatic diseases. The biologic agent that blocks TWEAK (BIIB023) has completed a phase II clinical trial in rheumatic-disease therapy (46,47). Here, we revealed a relationship between TWEAK and activation of the autophagic machinery, which promotes antimycobacterial immunity. In-depth studies are required to demonstrate the effect of a TWEAK inhibitor on TB disease and LTBI reactivation. Additionally, we explored an effect of miR-889 on antimycobacterial autophagy in patients with LTBI and LTBI reactivation. Given that the guidelines recommend LTBI screening for patients prior to biologic treatment (48,49), miR-889 and TWEAK may be diagnostic biomarkers of or therapeutic targets for LTBI or LTBI reactivation, respectively. Further studies are required to confirm and extend our findings for their clinical implications.

MATERIALS AND METHODS
Subjects. This prospective study was conducted at a medical center from 2014 to 2018. The Institutional Review Board of Taichung Veterans General Hospital approved this study (CE13330B), and the written consent of all participants was obtained according to the Declaration of Helsinki. Detailed definitions of subjects are available in Text S1 in the supplemental material.
MicroRNA isolation. Total RNAs were extracted using TRIzol reagent (Invitrogen, ThermoFisher Scientific, USA) and purified using an RNeasy MinElute cleanup kit (Qiagen, Germany), according to the manufacturer's instructions. Purified RNAs were quantified at an optical density at 260 nm (OD 260 ) and an OD 280 using an ND-1000 spectrophotometer (NanoDrop Technology, USA), and isolated miRNAs were qualified by capillary gel electrophoresis using a Bioanalyzer 2100 (Agilent Technology, USA).

QRT-PCR.
MicroRNA expression was measured and quantified using a TaqMan microRNA assay kit (Applied Biosystems, ThermoFisher Scientific, USA), according to the manufacturer's protocol. QRT-PCRs were performed on the StepOnePlus real-time PCR system (Applied Biosystems, ThermoFisher Scientific, USA), using a standard protocol. Detailed protocols are available in Text S1 in the supplemental material.
In vitro tuberculosis granuloma model and CFU assay. For granuloma formation, freshly isolated peripheral blood mononuclear cells (PBMCs) were infected with the M. bovis BCG or M. tuberculosis H37Rv strain at a multiplicity of infection (MOI) of 0.1; RPMI medium containing 20% human serum was added, and the mixture was incubated at 37°C in an incubator containing 5% CO 2 for up to 11 days (50). The medium and sera were replenished every 4 days. Mycobacterial growth was determined using CFU assays. Briefly, infected granuloma-like structures were lysed at different time periods ranging from 4 to 11 days postinfection. Lysates were serially diluted and plated on Middlebrook 7H11 agar plates and then incubated at 37°C to determine the number of CFU at 14 days.
Immunoblotting. The cells with different treatments were lysed in RIPA buffer (25 mM Tris-HCl, pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, and 0.1% sodium dodecyl sulfate [SDS]) containing a protease inhibitor cocktail (Complete, Roche, Germany). Twenty micrograms of total protein from exosome lysate was loaded and separated on a standard SDS-polyacrylamide gel electrophoresis (PAGE) gel and transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore, USA). The membranes were incubated with primary antibodies, followed by peroxidase-conjugated secondary antibodies. The results were detected using a charge-coupled device (CCD) camera-based imager (GE Healthcare Life Sciences, USA) after membrane incubation with enhanced-chemiluminescence (ECL) substrates (Millipore, USA).
Immunofluorescence assay. THP-1 cell-derived macrophages with individual treatment were fixed with 4% paraformaldehyde at room temperature for 10 min and then washed three times with phosphate-buffered saline (PBS). Cells were permeabilized in PBS containing 1% bovine serum albumin (BSA) and 0.2% saponin and then blocked for 1 h in PBS containing 2% BSA. Cells were then incubated with the primary antibodies, followed by a secondary antibody. Coverslips were mounted onto glass slides with DAPI (4=,6-diamidino-2-phenylindole)-containing SlowFade mounting medium (ThermoFisher Scientific, USA), and images were observed and recorded on an Olympus FV1000 laser-scanning confocal microscope. Images were analyzed by using FV10-ASW version 4.2 software. For quantification of the cells showing LC3-positive vesicles, approximately 50 cells were counted, and the cells with more than 20 LC3-labeled puncta were labeled as having formed an autophagosome.
Statistical analysis. An unpaired, two-tailed Student t test was used for between-group comparisons. A one-way analysis of variance (ANOVA) with the post hoc Bonferroni test was used for multiple comparisons. The correlation coefficient was calculated using Spearman's correlation test. P values of Ͻ0.05 were statistically significant, and tests were performed by using GraphPad Prism 7.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. TEXT S1, DOCX file, 0.03 MB.