Circular RNA TRAPPC6B inhibits intracellular Mycobacterium tuberculosis growth while inducing autophagy in macrophages by targeting microRNA‐874‐3p

Abstract Objectives Genetic and epigenetic mechanisms regulate antimicrobial immunity against Mycobacterium tuberculosis (Mtb) infection. Methods The present study assessed circular RNA TRAPPC6B (circTRAPPC6B) for antimicrobial immune functions and defined mechanisms wherein circTRAPPC6B regulates Mtb growth, autophagy and microRNA in macrophages. Results The Mtb infection of monocytes/macrophages resulted in a significantly decreased level of circTRAPPC6B that inhibited intracellular Mtb growth in macrophages. Conversely, circTRAPPC6B expression enhanced autophagy or autophagy‐associated protein LC3‐II production in Mtb‐infected macrophages. circTRAPPC6B‐enhanced autophagy aggregation or sequestration was also observed in fluorescence in situ hybridisation (FISH) analysis and confocal imaging. Mechanistically, circTRAPPC6B targets an inhibiting element miR‐874‐3p, as shown by bioinformatics, dual‐luciferase reporter gene analysis and pull‐down assay, respectively. Notably, miR‐874‐3p prohibited autophagy via suppressing autophagy protein ATG16L1 by binding to its 3′‐untranslated region (UTR) in Mtb‐infected macrophages and thus promoting intracellular Mtb growth. Concurrently, circTRAPPC6B enhanced autophagy in Mtb‐infected macrophages by blocking the ability of miR‐874‐3p to inhibit ATG16L1. Thus, circTRAPPC6B antagonises the ability of miR‐874‐3p to suppress ATG16L1 expression and activate and enhance autophagy sequestration to restrict Mtb growth in macrophages. Conclusion The current findings suggested that both circTRAPPC6B and miR‐874‐3p mechanisms can be explored as potential therapeutics against Mtb infection.

Objectives. Genetic and epigenetic mechanisms regulate antimicrobial immunity against Mycobacterium tuberculosis (Mtb) infection. Methods. The present study assessed circular RNA TRAPPC6B (circTRAPPC6B) for antimicrobial immune functions and defined mechanisms wherein circTRAPPC6B regulates Mtb growth, autophagy and microRNA in macrophages. Results. The Mtb infection of monocytes/macrophages resulted in a significantly decreased level of circTRAPPC6B that inhibited intracellular Mtb growth in macrophages. Conversely, circTRAPPC6B expression enhanced autophagy or autophagy-associated protein LC3-II production in Mtb-infected macrophages. circTRAPPC6B-enhanced autophagy aggregation or sequestration was also observed in fluorescence in situ hybridisation (FISH) analysis and confocal imaging. Mechanistically, circTRAPPC6B targets an inhibiting element miR-874-3p, as shown by bioinformatics, dual-luciferase reporter gene analysis and pull-down assay, respectively. Notably, miR-874-3p prohibited autophagy via suppressing autophagy protein ATG16L1 by binding to its 3 0 -untranslated region (UTR) in Mtb-infected macrophages and thus promoting intracellular Mtb growth. Concurrently, circTRAPPC6B enhanced autophagy in Mtbinfected macrophages by blocking the ability of miR-874-3p to inhibit ATG16L1. Thus, circTRAPPC6B antagonises the ability of miR-874-3p to suppress ATG16L1 expression and activate and enhance autophagy sequestration to restrict Mtb growth in macrophages. Conclusion. The current findings suggested that both circTRAPPC6B and miR-874-3p mechanisms can be explored as potential therapeutics against Mtb infection.

INTRODUCTION
Tuberculosis (TB) is a transmissible, airborne disease caused by Mycobacterium tuberculosis (Mtb), with approximately 9 million new cases and 1.5 million deaths annually. 1 Traditional anti-TB chemotherapy has shown low efficacy in TB control, mainly due to the prevalence of virulent and multidrug-resistant Mtb. 2 Recently, hostdirected therapies that modulate the host immune responses have received considerable attention because it avoids the development of multidrug resistance. 3,4 Thus, an improved understanding of immune responses to Mtb infection is critical for identifying novel therapeutic targets for TB control.
Resident alveolar macrophages are present on the lung epithelia, forming the first line of defence against Mtb infection. 5 Once inhaled, Mtb enters the alveolar macrophages via various phagocytic receptors and resides in the phagosomes that immediately mature to antimicrobial phagolysosomes. The delivery of Mtb into phagolysosomes relies on multiple trafficking pathways, including microtubuleassociated protein-1 light chain 3 (LC3)-associated phagocytosis and autophagy. During the lysosomal trafficking, Mtb activates various cellular processes, such as apoptosis, antigen presentation and autophagy, lethal to Mtb. 6 However, Mtb can evade these cellular processes, resulting in the development of TB disease. 7 Autophagy represents an intracellular process mediated by autophagosomes that transport cellular components to the lysosomes for degradation. Autophagosome formation requires the conversion of LC3-I (unconjugated cytosolic form) to LC3-II (autophagosomal membraneassociated phosphatidylethanolamine-conjugated form). Therefore, the amount of LC3-II reflects the number of autophagosomes, serving as the most widely used autophagy marker. 8,9 Autophagy acts as an innate immunity with a crucial role in eliminating intracellular Mtb. 10 The physiological activation of autophagy in macrophages facilitates mycobacterial phagosomes to mature into phagolysosomes, suppressing the intracellular survival of Mtb. 11 In contrast, impaired autophagy increases Mtb survival and correlates with poor outcomes in patients with Mtb infection. 10 During infection, the increased production of interferongamma (IFN-γ) activates macrophages to induce autophagy, facilitating the lysosomal degradation of Mtb by overcoming the Mtb-imposed block of phagosome maturation. 12 Many autophagyrelated proteins (ATGs), such as ATG5, ATG12 and ATG16L1, are involved in the innate immune clearance of Mtb. For example, ATG16L1 deficiency in macrophages protects Mtb from NADPH oxidase and phagocytosis. 13 MicroRNA (miR)-20a-mediated silencing of ATG7 and ATG16L1 inhibits autophagy and promotes the survival of TB vaccine Bacillus Calmette-Guérin (BCG) in macrophages. 14 Therefore, the activation of autophagy in macrophages represents a promising therapeutic strategy in anti-TB therapy.
Circular RNAs (circRNAs) are endogenous noncoding RNAs characterised by a covalently closed loop structure lacking 5 0 cap and 3 0 poly-A tail. [15][16][17] Previous studies reported that circRNAs are involved in multiple biological functions, including autophagy, by sponging miRNAs. 15,18 For example, circHIPK2 regulates astrocyte activation via the cooperation of autophagy and ER stress by targeting miR-124-2HG. 19 Intriguingly, miRNAs also play a critical role in autophagy by regulating host immunity during Mtb infection. 20 A recent study demonstrated that Mtb infection upregulates the expression of miR-23a-5p in macrophages, promoting Mtb survival by suppressing autophagy through the Toll-like receptor 2 (TLR2) signalling. 21 However, little is known about the interaction between circRNAs and miRNAs in macrophage autophagy during Mtb infection.
Our previous study revealed that the peripheral blood mononuclear cells (PBMCs) of patients with active TB have significantly decreased expression of circTRAPPC6B as compared to those of healthy controls, 22 suggesting that circTRAPPC6B might play an unfavorable role in Mtb survival. This study further investigated the role of circTRAPPC6B and its interaction with miRNAs in Mtb-infected macrophages. The findings suggested that circTRAPPC6B targets miR-874-3p to abolish its suppression on ATG16L1, thereby inducing autophagy and inhibiting Mtb growth in macrophages.

CircTRAPPC6B is significantly downregulated in PBMCs of patients with active TB and monocytes/macrophages with Mtb infection
In a previous study, we reported that PBMCs in patients with active TB had significantly downregulated circTRAPPC6B expression compared to healthy controls. 22 In the present study, we further confirmed the weak expression of circTRAPPC6B in PBMCs of 32 enrolled patients with active TB and 31 healthy volunteers (Figure 1a). One month of standardised anti-TB treatment rescues the expression of circTRAPPC6B in PBMCs (Figure 1b), suggesting an unfavorable role of circTRAPPC6B in Mtb growth. In addition, the area under the curve (AUC) of circTRAPPC6B in diagnosing active TB was 0.8609 (P < 0.0001; Figure 1c), indicating that circTRAPPC6B serves as a diagnostic marker for active TB. These results indicated that circTRAPPC6B was downregulated due to Mtb infection.
To confirm these conclusions in both ex vivo monocytes/macrophages and in vitro macrophage cell lines, we analysed the expression of circTRAPPC6B in Mtb-infected human PBMC and THP-1 macrophages. Also, due to the limited BSL-3 lab (P3) availability, we used BCG as a model of Mtb to explore the potential mechanism of circTRAPPC6B involved in TB and then further confirmed the results by Mtb (H37Rv)-infected macrophages. Because monocytes and alveolar macrophages constitute the first line of defence against TB, we detected the expression of circTRAPPC6B in BCG-infected human monocytes and macrophages. The data from real-time quantitative polymerase chain reaction (qRT-PCR) showed that BCG-infected human PBMCs exhibited significantly attenuated expression of circTRAPPC6B compared to the uninfected cells ( Figure 1d). Significantly attenuated expression of circTRAPPC6B was also observed in BCG-infected THP-1 macrophages (Figure 1e), which was also confirmed in H37Rv-infected THP-1 macrophages (Figure 1f). These results further confirmed our previous findings that weak expression of circTRAPPC6B was involved in Mtb infection.
CircTRAPPC6B is mainly located in the cytoplasm of macrophages Next, we explored the generation and cellular location of circTRAPPC6B in macrophages.
Bioinformatics revealed that circTRAPPC6B was derived from the back-splicing of exons 3 and 4 of the TRAPPC6B gene (circBase ID: hsa_circ_0005836) located at chromosome 14q21.1; the precise genomic location was 39,617,015-39,639,634. The length of mature circTRAPPC6B was 202 bp, according to the circBase database (http://www.c ircbase.org/) (Figure 2a). In order to exclude the possibility that the qRT-PCR products might originate from genomic rearrangements and transsplicing, we treated the RNAs with RNAse R before PCR. We observed that RNAse R digestion dramatically reduces the production of linear TRAPPC6B and GAPDH, but not that of circTRAPPC6B (Figure 2b), indicating that the splicing product was circular. Sanger sequencing further confirmed that the sequence of circTRAPPC6B produced by qRT-PCR matched that in the circBase database ( Figure 2a). Then, the cytoplasm and nucleus components of macrophages were extracted, and the expression of circTRAPPC6B was analysed, which indicated that circTRAPPC6B was mainly localised in the cytoplasm (Figure 2c). The FISH assay also confirmed the location of circTRAPPC6B in the cytoplasm of macrophages (Figure 2d, enlarged microscopy images are presented in Supplementary figure 1). These data collectively suggest that circTRAPPC6B, with a circular structure, is predominantly localised in the cytoplasm of macrophages.

CircTRAPPC6B inhibits intracellular Mtb growth in macrophages
To explore the roles of circTRAPPC6B involved in Mtb infection, we transfected the THP-1 macrophages with overexpression plasmid pHBAd-circTRAPPC6B (pHBAd-cir) with high transfection efficiency (Supplementary figure 2). Also, compared with pHBAd-control vector (vector), pHBAd-cir transfected cells demonstrated significant increase of circTRAPPC6B expression (Supplementary figure 3). To explore the role of circTRAPPC6B in intracellular Mtb growth, we performed a colony formation unit (CFU) assay in H37Rv-infected macrophages, of which the H37Rv infection was performed at a multiplicity of infection (MOI) of 1. At 4 h after infection, extracellular non-internalised bacilli were removed, and this time point was recorded as day 0. The results showed that after H37Rv infection, THP-1 macrophages overexpressing circTRAPPC6B had significantly decreased CFU compared to control cells at days 3 and 7, respectively ( Figure 3a). To further confirm these conclusions by tissue primary macrophages, we tested the effects of circTRAPPC6B on H37Rv growth in H37Rv infected macaque spleen primary macrophages. Similar results were observed in the primary macrophages in the macaque spleen ( Figure 3b). These findings suggested that the overexpression of circTRAPPC6B suppresses the intracellular Mtb growth in macrophages.

CircTRAPPC6B activates autophagy in Mtbinfected macrophages
Considering the critical roles of autophagy in eliminating intracellular Mtb, [10][11][12] we sought to investigate whether circTRAPPC6B affects autophagy in Mtb-infected macrophages. As a marker of autophagy, the expression of LC3-II was detected by Western blot. We found that the knockdown of circTRAPPC6B by siRNA transfection significantly attenuated LC3-II expression, whereas the overexpression of circTRAPPC6B by plasmid transfection remarkably enhanced the expression of LC3-II protein in BCG-infected THP-1 macrophages (Supplementary figure 4). Importantly, the knockdown of circTRAPPC6B partially but significantly reversed BCG infection-induced LC3-II upregulation in THP-1 macrophages (Supplementary figure 4a). Conversely, the overexpression of circTRAPPC6B further enhanced LC3-II upregulation in BCG-infected cells (Supplementary figure 4b). Similar results were observed in H37Rv-infected THP-1 macrophages (Figure 4c and d), which further suggested the (e, f) After THP-1 macrophages were infected with BCG (e) at MOI = 10 and H37Rv (f) at MOI = 1 at different time points (0, 1, 2, 3, 4 and 7 days), circTRAPPC6B expression was analysed by qRT-PCR. The data were obtained from three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 vs day 0. BCG, Bacillus Calmette-Guérin; HC, health control; MOI, multiplicity of infection; ns, no significance; PBMCs, peripheral blood mononuclear cells; qRT-PCR, quantitative real-time polymerase chain reaction; ROC, reactive operating characteristic curve; TB, tuberculosis. potential roles of circTRAPPC6B in regulating autophagy of Mtb-infected macrophages. Consistently, FISH assay was performed using biotin or digoxin-conjugated probes to detect the intracellular colocalisation.
Herein, we demonstrated that the overexpression of circTRAPPC6B resulted in the dramatically increased number of LC3B-positive puncta (red) in BCGinfected THP-1 cells compared to control vectors ( Figure 3e and f, enlarged microscopy images are presented in Supplementary figure 5). The overexpression of circTRAPPC6B also enhanced the colocalisation (yellow) of LC3B puncta and GFPexpressing BCG in THP-1 macrophages (Figure 3e and g, enlarged microscopy images are presented in Supplementary figure 5). Taken together, these data suggest that circTRAPPC6B activates autophagy in Mtb-infected macrophages.

CircTRAPPC6B specifically targets miR-874-3p
circRNAs can specifically bind to miRNAs to counter miRNA-mediated gene silencing. To investigate the mechanism underlying the role of circTRAPPC6B in Mtb growth and autophagy in macrophages, we predicted its potential miRNA targets using public databases CircInteractome, miRDB and RegRNA 2.0. The intersection of these databases revealed that circTRAPPC6B harboured a binding site for miR-874-3p (Figure 4a and b). Then, biotin-conjugated circTRAPPC6B and digoxin-conjugated miR-874 probes were synthesised, hybridised with the cells, and incubated with anti-biotin CY3-conjugated or anti-digoxin FITC-conjugated secondary antibody to visualise the colocalisation by RNA FISH assay. Subsequently, we observed that circTRAPPC6B was colocalised with miR-874-3p in the cytoplasm of THP-1 cells (Figure 4c, enlarged microscopy images are presented in Supplementary figure 6), which suggested that circTRAPPC6B could bind with miR-874 in the cytoplasm of macrophages. Dualluciferase reporter gene assay demonstrated that the overexpression of wild-type circTRAPPC6B reduced the luciferase activity of miR-874-3p mimics, whereas circTRAPPC6B with a mutated miR-874-3p binding site could not (Figure 4d). In addition, the biotin-coupled RNA complex was used in the circRNA and miRNA pull-down assay. We demonstrated that either circTRAPPC6B probe or miR-874-3p probe could pull down significantly enriched miR-874-3p and circTRAPPC6B as compared to the control probe (Figure 4e and f). Taken together, these findings indicated that circTRAPPC6B physically binds to miR-874-3p, acting as a miR-874-3p sponge.

MiR-874-3p promotes intracellular Mtb growth while inhibiting autophagy in Mtbinfected macrophages
A recent study showed that miR-874-3p could inhibit autophagy by reducing LC3-II expression in HeLa cells 23 ; however, whether miR-874-3p affects Mtb growth and autophagy in Mtb-infected macrophages remains to be explored. In contrast to the circTRAPPC6B expression, we found that PBMCs from patients with active TB had significantly increased miR-874-3p level compared to the respective controls ( Figure 5a). Moreover, Spearman's analysis showed a significant negative correlation between circTRAPPC6B and miR-874-3p expression in PBMCs (Figure 5b). BCG infection also induced significantly increase of miR-874-3p level in THP-1 macrophages (Supplementary figure 7). In addition, significant increase of miR-874-3p level was also found in H37Rv-infected THP-1 macrophages (Figure 5c). These findings suggested that miR-874-3p might play opposite roles to circTRAPPC6B in TB.
Strikingly, the results of the CFU assay showed that miR-874-3p inhibitor significantly suppressed the intracellular H37Rv growth in both THP-1 and macaque spleen macrophages (Figure 5d and e), indicating an essential role of miR-874-3p in promoting Mtb growth in macrophages. In addition, miR-874-3p inhibition resulted in significantly elevated number of LC3B-positive puncta and enhanced colocalisation of LC3B and GFP-BCG in GFP-BCG-infected THP-1 macrophages (Figure 5f-h, enlarged microscopy images are presented in Supplementary figure 8). miR-874-3p mimics dramatically suppressed, whereas miR-874-3p inhibitor facilitated the conversion of LC3-I to LC3-II in THP-1 macrophages irrespective of BCG (Supplementary figure 9a and b) and H37Rv infection (Figure 5i and j), suggesting that miR-874-3p is also essential for suppressing autophagy in macrophages. These data collectively suggested that miR-874-3p plays opposite roles to circTRAPPC6B in TB by promoting Mtb growth while inhibiting autophagy in macrophages.
CircTRAPPC6B regulates autophagy in Mtbinfected macrophages, possibly via miR-874-3p/ATG16L1 Next, we sought to investigate whether miR-874-3p mediates the regulatory role of circTRAPPC6B in macrophage autophagy during Mtb infection. We found that miR-874-3p mimics reverse the LC3-II conversion induced by circTRAPPC6B overexpression in BCG-infected (Supplementary figure 10) and H37Rv-infected THP-1 macrophages (Figure 6a). Consistent results were observed in immunofluorescence assay obtained in BCGinfected THP-1 macrophages (Figure 6b and c, enlarged microscopy images are presented in (e) CircRNA pull-down assay was performed in THP-1 macrophages using biotin-conjugated circTRAPPC6B to assess the interaction between circTRAPPC6B and miR-874-3p, followed by qRT-PCR to detect circTRAPPC6B and miR-874-3p enrichment. The data were obtained from three independent experiments. Data are expressed as mean AE SEM. ***P < 0.001 vs control probe. (f) miRNA pull-down assay was performed in THP-1 macrophages using digoxin-conjugated miR-874 probes, followed by qRT-PCR to detect miR-874-3p and circTRAPPC6B enrichment. The data were obtained from three independent experiments. Data are expressed as mean AE SEM. **P < 0.01, ***P < 0.001 vs control probe. FISH, fluorescence in situ hybridisation. figure 11).

Supplementary
These findings suggested that miR-874-3p could block the inductive role of circTRAPPC6B in macrophage autophagy during Mtb infection.
To further investigate the mechanism underlying the role of miR-874-3p in macrophage autophagy, we conducted a bioinformatics analysis to identify its potential targets using TargetScan, miRanda and miRDB databases. It was found that the 3 0 -UTR of ATG16L1 mRNA harboured a potential miR-874-3p binding site (Figure 7a). Dual-luciferase reporter assay showed that compared to the negative control, miR-874-3p mimics significantly reduced the luciferase activity of wild-type ATG16L1 3 0 -UTR reporter vectors but not that of mutant 3 0 -UTR reporter vectors (Figure 7b). miR-874-3p mimics also significantly attenuated, whereas its inhibitor enhanced the expression of ATG16L1 protein in THP-1 macrophages regardless of the BCG infection (Figure 7c and d), suggesting that miR-874-3p can suppress the expression of ATG16L1 by binding to its 3 0 -UTR. Moreover, the PBMCs of patients with active TB significantly decreased ATG16L1 mRNA levels as compared to those of healthy controls (Figure 7e). Similar results were also observed in H37Rv-infected THP-1 macrophages (Figure 1f). A correlation analysis revealed that ATG16L1 expression was significantly and positively correlated with circTRAPPC6B expression (Figure 7f) but negatively with miR-874-3p expression (Figure 7g). Western blot assay demonstrated that the silencing of circTRAPPC6B decreased the level of ATG16L1 protein in BCG-infected macrophages (Figure 7h). Interestingly, the overexpression of circTRAPPC6B  increased the expression of ATG16L1 protein (Figure 7i). Thus, the data demonstrated that circTRAPPC6B antagonises miR-874-3p to counter its suppression on ATG16L1 expression, thereby activating autophagy to eliminate Mtb in macrophages.

DISCUSSION
In this study, consistent with the previous findings, 22 we demonstrated that circTRAPPC6B expression is significantly reduced in PBMCs of TB patients, BCG-infected human peripheral monocytes, and BCG-infected THP-1 macrophages as compared to that in corresponding controls. Furthermore, our previous study demonstrated that the overexpression of circTRAPPC6B inhibits intracellular Mtb growth while activating autophagy in Mtb-infected macrophages. In addition, circTRAPPC6B targets miR-874-3p and might abolish its suppression on ATG16L1, thereby activating autophagy to eliminate Mtb in macrophages.
CircRNAs are widely expressed in mammalian cells. The unique loop structure of circRNAs makes them resistant to RNase R degradation 24 and are more stable biomarkers than linear RNAs. 25 In the present study, bioinformatics analysis showed that circTRAPPC6B is derived from the back-splicing of exons 3 and 4 of the TRAPPC6B gene. RNAse R digestion fails to reduce the production of circTRAPPC6B by PCR, hinting at a circTRAPPC6B that has a circular structure. In addition to the weak expression of circTRAPPC6B in PBMCs of TB patients, we further found that anti-TB treatment can rescue its expression. Additionally, the AUC of circTRAPPC6B in active TB was 0.8609. These findings suggested that circTRAPPC6B may serve as a biomarker for diagnosing and monitoring treatment responses in TB.
The low expression of circTRAPPC6B in PBMCs, monocytes and macrophages with Mtb infection suggested an unfavorable role of circTRAPPC6B in Mtb growth. As expected, our data showed that the overexpression of circTRAPPC6B suppresses Mtb growth in BCG-infected THP-1 macrophages and H37Rv-infected primary spleen macrophages in macaques. Because autophagy plays a vital role in eliminating intracellular Mtb, 11 we hypothesised that the overexpression of circTRAPPC6B might activate autophagy to eliminate Mtb in macrophages. Intriguingly, the gain-and loss-of-function assays showed that the overexpression of circTRAPPC6B significantly enhances, whereas knockdown of circTRAPPC6B attenuates LC3-II expression in THP-1 macrophages regardless of BCG and H37Rv infection, suggesting that circTRAPPC6B activates autophagy in macrophages. Although we could not establish a direct link between autophagy activation and Mtb elimination in circTRAPPC6B-overexpressing macrophages, accumulating evidence suggested that the stimulation of autophagy suppresses, whereas the impairment increases Mtb survival in macrophages. 10,11 Therefore, we speculated that circTRAPPC6B suppresses Mtb growth in macrophages by activating autophagy, serving as a potential therapeutic agent in TB treatment.
Considering the well-established interaction between circRNAs and miRNAs, we further identified the potential miRNA targets of circTRAPPC6B to investigate the mechanisms underlying its role in Mtb growth and autophagy in macrophages. Bioinformatics analysis demonstrated that circTRAPPC6B harbours a miR-874-3p binding site. Dual-luciferase reporter assay, RNA FISH and RNA pull-down further confirmed that circTRAPPC6B acts as a miR-874-3p sponge in macrophages. These findings suggested that circTRAPPC6B and miR-874-3p exhibit opposite expression patterns in macrophages. As expected, both PBMCs of patients with active TB and THP-1 macrophages with Mtb infection show significantly increased miR-874-3p level compared to the corresponding controls, contrary to that observed in circTRAPPC6B expression. Moreover, circTRAPPC6B and miR-874-3p expression is significantly and negatively correlated in PBMC samples. These findings prompt us to explore further whether miR-874-3p plays opposite roles to circTRAPPC6B in Mtb growth and autophagy.
MiR-874-3p functions as a tumor suppressor in human cancers 26 ; however, its role in Mtb growth and macrophage autophagy remains unknown. A recent study reported that miR-874-3p inhibits autophagy in HeLa cells. 23 Consistently, our results showed that miR-874-3p mimics dramatically suppresses, whereas miR-874-3p inhibitor facilitates the conversion of LC3-I to LC3-II in THP-1 macrophages regardless of BCG and H37Rv infection, suggesting that miR-874-3p plays an essential role in suppressing autophagy in macrophages. In addition, miR-874-3p is also essential for Mtb growth in macrophages, as evident from suppressed intracellular H37Rv growth in both THP-1 and macaque spleen macrophages exposed to miR-874-3p inhibitor. Furthermore, miR-874-3p can block the inductive role of circTRAPPC6B in macrophage autophagy during Mtb infection. To the best of our knowledge, this is the first study reporting the essential roles of miR-874-3p and its interaction with circTRAPPC6B in suppressing autophagy while promoting Mtb growth in macrophages.
Reportedly, miR-874 targets ATG16L1 in gastric cancer cells and HeLa cells. 23,27 Consistently, our results showed that miR-874-3p suppresses ATG16L1 expression by binding to its 3 0 -UTR. In addition, PBMCs of active TB patients have significantly decreased ATG16L1 mRNA levels compared to those of healthy controls. The mRNA expression of ATG16L1 is significantly and positively correlated with that of circTRAPPC6B but negatively correlated with that of miR-874-3p in PBMCs. Therefore, we speculated that circTRAPPC6B antagonises miR-874-3p to counter its suppression of ATG16L1 expression. ATG16L1, ATG5 and ATG12 form a large protein complex (the ATG16L1 complex) that is essential for autophagosome formation. 28 ATG16L1 is involved in host immune responses against intracellular bacteria and viruses via autophagy. 29 siRNA targeted to ATG16L1 impairs autophagy in macrophages infected with adherent-invasive Escherichia coli, in turn, increasing the proinflammatory cytokine secretion and intracellular bacterial population. 30 In addition, knock-in mice harbouring a missense ATG16L1 variant showed a defective clearance of the ileal pathogen Yersinia enterocolitica and elevated inflammatory responses due to diminished autophagy. 31 Based on these findings, we speculated that circTRAPPC6B overexpression might abrogate the suppression of miR-874-3p on ATG16L1 expression, thereby activating autophagy to eliminate Mtb in macrophages. However, we did not establish a direct causative correlation between the expression of circTRAPPC6B and ATG16L1, which should be addressed in future studies.
In summary, we demonstrated that circTRAPPC6B inhibits intracellular Mtb growth while inducing autophagy in Mtb-infected macrophages. CircTRAPPC6B acts as a novel competitive endogenous RNA to sponge miR-874-3p in macrophages, suggesting that circTRAPPC6B abolishes the suppression of miR-874-3p on ATG16L1 expression to regulate Mtb growth and autophagy in macrophages (Figure 8). This study provides circTRAPPC6B as a potential therapeutic agent for TB control.

Ethics statement
The protocol of using human blood samples in vitro was approved by the Institutional Review Boards for Human Donors' Research and Institutional Biosafety Committee at Institute Pasteur of Shanghai, Guangdong Medical University, and Dongguan 6th Hospital (Dongguan, China). The protocols of using macaque spleen tissue samples and Mycobacterium H37Rv strain were approved by the Institutional Review Boards for Biosafety Committee at the University of Illinois, Chicago College of Medicine (IL, USA). All studies complied with the guidelines of the Office for Human Research Protection. All patients and healthy controls provided written informed consent before initiation of the study.

Patients and cell samples collection
A total of 32 patients (17-70-year-old) with active pulmonary TB admitted to Dongguan 6 th Hospital were recruited in this study. The diagnosis was made, as described previously. 32 According to the standard methods, the sputum samples of all were tested by Ziehl-Neelsen acid-fast staining and Lowenstein-Jensen slant culture. A total of 31 healthy volunteers (20-65 years of age) without bacteriological and clinical evidence of TB were enrolled as controls (Supplementary table 1). Patients with active TB received individualised anti-TB treatments, including isoniazid, rifampicin, pyrazinamide and ethambutol.

Cell lines and Mycobacterium
The human THP-1 monocyte cell line was grown in RPMI 1640 supplemented with L-glutamine (2 mM), sodium pyruvate (1 mM) and 10% heat-inactivated foetal bovine serum (FBS; Life Technologies, Carlsbad, CA, USA). Before mycobacterial infection, THP-1 cells were treated with 50 ng mL −1 phorbol 12-myristate 13-acetate (PMA; Sigma-Aldrich, St. Louis, MO, USA) for 48 h to allow differentiation into macrophages. The cells were then washed three times with prewarmed phosphate-buffered saline (PBS) and maintained in antibiotic-free media at 37°C in a humidified atmosphere of 5% CO 2 for subsequent use.
All the cell lines were tested for Mycoplasma infection using a 16S-based PCR. New cultures were established monthly from frozen stocks.
Macaque spleen macrophages were enriched from macaque spleen tissues. The spleen macrophages were grown in RPMI 1640 supplemented with L-glutamine (2 mM), sodium pyruvate (1 mM) and 10% heat-inactivated FBS for 7 days.
Human PBMCs were prepared, as described previously. 32,33 Peripheral blood samples (10 mL) were collected in EDTA-tubes from TB patients and healthy volunteers and separated by density gradient centrifugation using Ficoll-Paque PLUS medium (GE Healthcare, Chicago, IL, USA). Human monocytes were sorted from PBMCs by immunomagnetic positive selection (STEMCELL Technologies, Vancouver, BC, Canada).

RNA preparation, RNAse R treatment and PCR
Total RNAs were isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. An equivalent of 2 μg total RNA was incubated with 2 U RNAse R (Epicenter Technologies, San Diego, CA, USA) in 1× RNase R reaction buffer for 10 min at 37°C, followed by reverse transcription into cDNA using PrimeScript RT Reagent (TaKaRa, Japan). Subsequently, qRT-PCR was performed using a TransStart Tip Green qPCR SuperMix (Transgen, Beijing, China) on a QuantStudio ™ 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). GAPDH and RNU6-1 were used as internal controls for circRNA, mRNA and miRNA, respectively. All qRT-PCR reactions were performed in triplicate. The data were analysed using the 2 ÀΔΔCt method to calculate the relative expression of the target gene. PCR primers are listed in Supplementary table 2.

Western blot analysis
Proteins were isolated from THP-1 cells using a protein extraction kit (Beyotime, Shanghai, China) according to the manufacturer's protocol. The protein concentration was measured using a BCA kit (Beyotime). An equivalent of protein samples was separated using SDS-PAGE and then transferred onto a PVDF membrane (Millipore, Burlington, MA, USA). After blocking with 5% skim milk in TBST buffer for 2 h, the membrane was probed at 4°C overnight with monoclonal antibody against LC-3 (1: 1000, Proteintech, Rosemont, IL, USA) or ATG16L1 (1: 1000, Proteintech). Following 4 × 5 min washes with TBST, the membrane was incubated with a corresponding HRP-labelled secondary antibody for 2 h at room temperature. Subsequently, the immunoreactive signals were visualised using an enhanced chemiluminescence detection system with a chemiluminescence substrate (Thermo Scientific, Waltham, MA, USA). β-Actin (Beyotime, Shanghai, China) was used as an internal control.

Mycobacteria CFU assay
Transfected THP-1 and macaque spleen macrophages were infected with H37Rv at an MOI of 1. At 4 h after infection, extracellular non-internalised bacilli were removed by washing three times with prewarmed PBS, and this time point was recorded as 0. Then, the supernatant of the cells was removed, and the infected cells were lysed in 1 mL sterile water containing 0.03% SDS for 20 min. A 10-fold serial dilution was performed for quantitative culturing. Aliquots (100 mL) were plated in triplicate on Middlebrook 7H11 agar plates supplemented with 10% OADC enrichment for 2-3 weeks until colonies were sufficiently large for counting. The viability of Mycobacteria was quantified by counting CFU.

FISH
To detect the intracellular location of circTRAPPC6B and miR-874 as well as the colocalisation of circTRAPPC6B and miR-874, RNA FISH was performed according to the protocol of Guangzhou Geneseed Technology (Guangzhou, China). The biotin-conjugated circTRAPPC6B and digoxinconjugated miR-874-3p probes were synthesised by Guangzhou Geneseed Biotechnology. Briefly, THP-1 cells were rinsed in PBS and fixed in 4% formaldehyde with RNase-free PBS for 5 min at room temperature. Then, the cells were incubated with absolute ethyl alcohol for 1 min. circTRAPPC6B and miR-874-3p probes were hybridised with the samples in the dark at 37°C overnight. The cells were blocked with 3% bovine serum albumin (BSA) for 30 min and incubated with anti-biotin CY3-conjugated or antidigoxin FITC-conjugated secondary antibody for 1 h at 37°C. The cells were visualised using laser-scanning confocal microscopy (Leica TCS SP2 AOBS, Germany).

CircRNA pull-down assay
All the following experiments were performed at room temperature except for those mentioned otherwise. The probes are listed in Supplementary table 2. The cells were lysed using lysis buffer containing Superase-In and protease inhibitors for Western blot and IP (P0013, Beyotime, Shanghai, China).
An equivalent of 1 × 10 7 THP-1 cells was fixed with 1% formaldehyde for 10 min, lysed for 10 min at 4°C, and sonicated, followed by centrifugation at 10 000 g for 10 min. A volume of 50 μL supernatant was used for input analysis. The remaining volume was incubated with 3 μg circTRAPPC6B specific probes or control probes at 30°C for 1 h, rotating at 10 rpm. The streptavidin magnetic beads (Life Technologies) were washed twice with 500 μL lysis buffer and resuspended in 1000 μL lysis buffer. The biotincoupled RNA complex was pulled down by incubating the cell lysates with 50 μL beads for 3 h on the rotator at 10 rpm, followed by RNA extraction using TRIzol and qRT-PCR detection.

MiRNA pull-down assay
Approximately 2 × 10 6 cells were transfected with 50 μM of biotinylated-miRNA mimics or nonsense control (GenePharma, Shanghai, China) at 50% confluency. At 24 h after transfection, the cells were harvested, washed in PBS, and lysed with lysis buffer. A volume of 50 μL of washed streptavidin magnetic beads was blocked for 2 h and then added to each reaction tube to pull down the biotincoupled RNA complex. Subsequently, the tubes were incubated for 4 h on the rotator at a low speed of 10 rpm, and then, washed with lysis buffer five times. TRIzol LS (Life Technology, USA) was used to recover RNAs, specifically interacting with miRNA. The abundance of circTRAPPC6B in the bound fractions was evaluated by qRT-PCR and agarose gel electrophoresis.

Immunofluorescence analysis
THP-1 cells were seeded on a Nunc glass-bottom dish (Thermo Fisher Scientific) and incubated in RPMI 1640 medium at 37°C in a humidified atmosphere with 5% CO 2 overnight. The cells were washed with PBS three times after corresponding treatment, fixed with 4% paraformaldehyde for 15 min, and permeabilised with 0.25% Triton X-100 for 10 min. After blocking with 5% skim milk in TBST for 30 min, the cells were incubated with anti-LC3B (Cell Signal Technology, Danvers, MA, USA) for 2 h at room temperature, followed by incubation with Alexa Fluor ® 594conjugated anti-Rabbit IgG (Thermo Fisher Scientific) for 1 h at room temperature. The nuclei were stained with DAPI for 5 min. After mounting, fluorescence images were acquired using a confocal laser-scanning microscope (Zeiss, Jena, Germany). The experiments were repeated three times. A total of 100 cells were selected to calculate the average value at each step, and the average obtained from three independent experiments was plotted.

Statistical analysis
All data were analysed using SPSS17.0 (IBM, Armonk, NY, USA) and plotted using GraphPad Prism 5 software (GraphPad, USA). Data were expressed as mean AE standard error of the mean (SEM). The Student's t-test and the paired t-test were used to assess the statistical significance between the two groups. ANOVA with Dunnett's multiple comparisons test was used to compare more than two groups. Pearson's analysis was used to analyse the correlation between the two groups. A P-value < 0.05 was considered to be statistically significant.