Mycobacterium massiliense Induces Macrophage Extracellular Traps with Facilitating Bacterial Growth

Human neutrophils have been known to release neutrophil extracellular traps (NETs), antimicrobial DNA structures capable of capturing and killing microbes. Recently, a similar phenomenon has been reported in macrophages infected with various pathogens. However, a role for macrophages extracellular traps (METs) in host defense responses against Mycobacterium massiliense (M. mass) has yet to be described. In this study, we show that M. mass, a rapid growing mycobacterium (RGM), also induces the release of METs from PMA-differentiated THP-1 cells. Intriguingly, this process is not dependent on NADPH oxidase activity, which regulates NET formation. Instead, M. mass-induced MET formation partially depends on calcium influx and requires phagocytosis of high bacterial load. The METs consist of a DNA backbone embedded with microbicidal proteins such as histone, MPO and elastase. Released METs entrap M. mass and prevent their dissemination, but do not have bactericidal activity. Instead, they result in enhanced bacterial growth. In this regard, METs were considered to provide interaction of M. mass with cells and an environment for bacterial aggregation, which may facilitate mycobacterial survival and growth. In conclusion, our results demonstrate METs as an innate defense response against M. mass infection, and suggest that extracellular traps play a multifaceted role in the interplay between host and bacteria.


Introduction
The Mycobacterium abscessus (M. abscessus) complex is a group of rapidly growing nontuberculous mycobacteria (NTM) that cause inflammatory disease ranging from localized abscess to disseminated infection in patients with immune deficiency such as cystic fibrosis or immunocompromised patients [1,2,3]. The M. abscessus complex comprises of three genetically related species: M. abscessus (sensu stricto), M. massiliense, and M. bolletii [4]. M. abscessus and M. massiliense (M. mass) are often isolated from patients with respiratory disease, and are known to be hard to treat and take longer to cure because they often acquire resistance to multiple antibiotics [5]. by performing only soft spin centrifugation of the same bacterial cultures. The number of viable bacteria in stored bacterial vials was counted on Luria-Bertani (LB) agar (BD Biosciences, Franklin Lakes, NJ, USA).

Cell culture and reagents
The human acute monocytic leukemia THP-1 cell line was maintained in RPMI media supplemented with 10% FBS (Gibco, Carlsbad, California, USA). Differentiation of THP-1 cells into macrophages was performed by incubation with 25 ng/ml phorbol 12-myristate 13-acetate (PMA) (Sigma-Aldrich, St. Louis, MO, USA) for 2 days at 37°C in a humidified atmosphere with 5% CO 2 . In these experiments, THP-1 cells were grown in 24-well tissue culture plates (Corning, Corning, NY, USA), and PMA-differentiated THP-1 macrophages were selected by keeping only the adherent cells.

Immunofluorescence microscopy
To perform immunofluorescence staining, M. mass R-infected cells were fixed with 4% paraformaldehyde (PFA) for 15 min, and permeabilized with 0.1% Triton X-100 for 5 min and blocked in 1% BSA/PBS for 1 hr at room temperature. The samples were subsequently incubated with rabbit anti-histone H4, rabbit anti-myeloperoxidase (MPO) or rabbit anti-elastase antibody (Santa cruz, Dallas, TX, USA) at 1:50-1:100 dilution for 90 min at room temperature. They were then washed and incubated with Alexa 488-conjugated anti-rabbit IgG (Thermo Fisher, Waltham, MA, USA) at 1:200 dilution for 90 min. Lastly, the slides were stained with 1μM TO-PRO-3 for 30 min and examined using fluorescence.

Cytokine treatment of macrophage
Differentiated THP-1 macrophages were pretreated with 10ng/ml IFN-γ (ProSpec, East Brunswick, NJ, USA) and TNF-α (R&D Systems, NE Minneapolis, MN, USA) or untreated for 24 hr. The cells were washed and infected with M. mass R or CIP (MOI 5) for 24 hr. Then, MET formation in each sample was examined.

Cytokine analysis
For cytokine analysis, the supernatants of THP-1 macrophages infected with M. mass R or CIP (5 MOI) were harvested at 24hr post infection, and stored at -80°C until use. The levels of inflammatory cytokines (TNF-α, IL-1β and IL-6) were quantified using an enzyme-linked immunosorbent assay (ELISA) kit (Roche, Basel, Switzerland). The assays are performed according to the manufacturer's instructions.

Reactive Oxygen Species (ROS) Detection
Differentiated THP-1 macrophages were pretreated with 20 μM DPI or left untreated for 30 min before stimulation with 1 mM hydrogen peroxide or M. mass R (MOI 5) for 24 hr. The cells were washed and stained with intracellular indicator, 5 μM 2', 7 dichlorofluorescein diacetate (H2DCF-DA) (Thermo Fisher, Waltham, MA, USA) for 30 min. After washing, labeled cells were treated with 5 mM EDTA for 15 min, and collected. The samples were subsequently washed and analyzed using BD FACSCalibur and data analysis was done using FlowJo software (BD Biosciences, Franklin Lakes, NJ, USA).

Lactate dehydrogenase (LDH) release assay
To assess cell lysis in infected cells, PMA-differentiated THP-1 cells were cultured in 24 well plates (2×10 5 cells/well), and infected by M. mass R (MOI 1, 3, or 10) for 24 hr. The supernatants were collected and Lactate dehydrogenase (LDH) release detection was performed using cytotoxicity detection kit (Roche, Mannheim, Germany) according to the manufacturer's instructions.

Annexin V and Propidium iodide staining for flow cytometry
To determine necrotic cell death, differentiated THP-1 macrophages were infected with M. mass R (MOI 5) or left uninfected for 1 day. Collected cells were washed and then stained with Annexin V FITC (BD Biosciences, Franklin Lakes, NJ, USA) and Propidium iodide (BD Biosciences, Franklin Lakes, NJ, USA) for 15 min. Stained cells were analyzed using BD LSR Fortessa (BD Biosciences, Franklin Lakes, NJ, USA), and data analysis was done using FlowJo software (BD Biosciences, Franklin Lakes, NJ, USA).

Statistical analyses
ANOVA and Student's t-test were performed to determine statistically significant differences between groups using GraphPad Prism (GraphPad Software, La Jolla, CA, USA). A P value of < 0.05 was deemed to be statistically significant and is indicated in the figures by an asterisk. P values of <0.01 and <0.001 are indicated by two and three asterisks, respectively.

Microscopic examination of MET formation induced by M. mass infection
In a previous study, M. abscessus infection induced the formation of DNA extracellular trap structures by human peripheral blood mononuclear cells (PBMC) [30]. Similarly, when PMAdifferentiated THP-1 macrophages were infected by M. mass cells released extracellular fiber structures that merged together ( Fig 1A). These structures were positively stained by a DNAstaining dye, TO-PRO, and were morphologically similar to DNA extracellular traps ( Fig 1B  and 1C). We observed similar structures in PBMC-derived human macrophages infected with M. mass (data not shown). To confirm whether these structures can function like METs and entrap bacteria, we infected macrophages with CFSE-stained M. mass. Fluorescence microscopic examination revealed CFSE-stained M. mass co-localization with extracellular DNA fibers 1 day after infection, suggesting that extracellular M. mass was entrapped and bound to the MET structures (Fig 1D and 1E). In addition, we used scanning electron microscopy (SEM) to examine the ultrastructure of the METs induced by M. mass (Fig 1F). In SEM examination, most of extracellular M. mass were shown to be attached to macrophages. Some mycobacteria were linked to METs released from adjacent macrophages. MET structures were more common near clustered M. mass rather than single or a few bacilli. Thus, we confirmed that M. mass infection induces the formation of DNA extracellular traps from differentiated THP-1 macrophages and that M. mass-induced METs are morphologically similar to the extracellular structures released by activated neutrophils and other immune cells.

M. mass with rough colony morphology strongly induces MET formation
Various stimuli have been shown to induce neutrophil extracellular traps including PMA, LPS, hydrogen peroxide, etc [6,31]. We examined the effects of such stimuli on inducing MET formation by PMA-differentiated THP-1 macrophages (Fig 2A). LPS (10 μg/ml) and hydrogen peroxide (1 mM) treatments for 1 day did not induce extracellular traps from macrophages, and high concentrations of stimuli (LPS ! 100 μg/ml; H 2 O 2 ! 10mM) caused extensive cell death. In addition, we infected THP-1 macrophages with two different strains of M. mass, those that form rough and smooth colonies, to determine whether colony type affects MET formation or not. We infected the cells with each M. mass at 20 MOI, to induce extreme responses by each colony type for 1 day and compare them with the effects of other stimuli. After infecting macrophages with each strain of M. mass, the sample infected with the rough strain, M. mass R, showed MET formation in 14% of the cells (Fig 2B). On the other hand, less than 6% of the macrophages infected by the smooth strain, M. mass CIP, formed METs. This result is similar to that of previous study using human monocytes infected with M. abscessus [30]. We examined whether only intracellular infection without extracellular bacteria induces MET formation. The sample with only intracellular infection showed much lower levels of MET formation than the sample infected with both intracellular and extracellular bacteria (S1 These results suggest that DNA extracellular trap production by macrophages is not due to simple inflammatory or oxidative stimuli, but is dependent on the colony type of M. mass, indicating that interactions with distinct antigens or surface structures of M. mass may cause the release of extracellular traps from macrophages. M. mass R-induced METs contain mitochondrial DNA, nuclear DNA, and microbicidal proteins Extracellular traps formed by different immune cells are known to consist of nuclear or mitochondrial DNA embedded with histones and various enzymatic peptides [6,16,32]. To investigate the origin of the DNA released by macrophages infected with M. mass, we infected THP-1 cell with M. mass R and performed PCR analysis to detect nuclear signals (β-Actin and Gapdh) and mitochondrial signals (Atp6 and Nds1). PCR analysis of the supernatants from M. mass Rinfected cells, containing visible MET structures, showed that the extracellular DNA contained sequences of both nuclear and mitochondrial genes (Fig 3A). It is possible that the extracellular DNA in supernatants is due to the release of cell contents from cell death associated with M.  mass R-induced MET production. In addition, we performed immunofluorescence staining to identify the DNA-embedded components in M. mass R-induced METs (Fig 3B). We used specific antibodies to test for potential candidates such as histones and antimicrobial enzymes (MPO and elastase), and found that histones colocalized with the extracellular DNA of METs while antimicrobial enzymes colocalized with METs. The latter suggests that M. mass Rinduced METs also have enzyme-mediated bactericidal functions. However, we also observed that portions of METs did not contain histones or enzymes. Collectively, we show that M. mass R-induced METs have a backbone of nuclear and mitochondrial DNA that is embedded with histones and enzymatic proteins, suggesting they have similar functions in host defense against foreign pathogens to extracellular traps of other immune cells.
M. mass R-induced MET formation does not depend on NADPH oxidase, but rather calcium influx Extracellular trap formation is known to be closely associated with the activation of NADPH oxidase [6,33,34]. To investigate whether NADPH oxidase is required for M. mass R-induced MET formation, we treated infected macrophages with diphenylene iodonium (DPI), an inhibitor of NADPH oxidase enzymes (Fig 4A). Although treatment of infected macrophages with DPI reduced ROS production, M. mass R infection did not augment ROS production compared to uninfected cells (S4 Fig). Microscopic examination showed that DPI treatment of M. mass R-infected cells did not significantly reduce MET formation, suggesting that inhibition of ROS generation by NADPH oxidase does not affect extracellular trap release from macrophages by M. mass R infection. Although hydrogen peroxidase treatment enhanced ROS generation in macrophages, it did not induce MET formation. In a previous study using M. tuberculosis, MET formation was reported to be associated with elastase activity [26]. We tested using various concentrations of N-Methoxysuccinyl-Ala-Ala-Pro-Val p-nitroanilide (AAPV), elastase inhibitor, but it did not affect M. mass R-induced MET formation (S5 Fig).
Recent studies have reported that calcium influx also triggers the formation of extracellular traps via an NADPH oxidase-independent mechanism [35,36]. To investigate the role of calcium influx in M. mass R-induced MET formation, we measured calcium influx in infected and uninfected THP-1 macrophages by staining cells with Fluo-4 AM (Fig 4B). Fluorescence intensity measurement showed that M. mass R infection triggers a slow increase in the level of intracellular calcium that apparently exceeds uninfected cells at 6 hr after infection. The chelation of intracellular or extracellular calcium using BAPTA or EGTA, respectively, decreased calcium influx in M. mass R-infected macrophages. To evaluate the effects of calcium influx on METs formation, we monitored the formation of extracellular traps by M. mass R-infected macrophages after pretreatment with BAPTA or EGTA (Fig 4C). Although both treatment with BAPTA and EGTA reduced intracellular calcium levels of infected cells at early stages, however, BAPTA and EGTA did not completely abolish MET formation. Conversely, although treatment with ionomycin increased calcium influx, it could not trigger significant release of METs. Collectively, these results suggest that production of extracellular traps is partially affected by calcium-mediated inflammatory signals induced by M. mass R infection, and not calcium influx, in an NADPH oxidase-independent manner.

MET formation is dependent on phagocytosis of M. mass R
Considering the previous results, we supposed that an intracellular direct interaction between macrophages and M. mass R may contribute to MET formation. To analyze the dependency of MET formation on phagocytosis of M. mass R, we pretreated THP-1 macrophages with a phagocytosis blocker, cytochalasin D (CytD), prior to M. mass R infection and analyzed changes in MET formation. Treated cells showed decreased phagocytosis, represented by decreased intracellular CFUs, compared to untreated cells (Fig 5A). This was correlated with reduced MET formation in CytD-treated cells, suggesting that induction of MET is dependent on phagocytosis of M. mass R (Fig 5B). To corroborate this, we investigated MET formation after phagocytosis of different numbers of bacteria (Fig 5C). We infected the cells with M. mass R at 10, 3 and 1 MOI, to examine gradually induced MET formation and cell lysis for 1 day. Indeed, the extent of MET formation depended on the initial amount of bacteria that was phagocytosed (Fig 5D). In addition, LDH release data showed that cell lysis occurred after infection with M. mass R in a manner that is also correlated with bacterial burden (Fig 5E).  mass R, rather than an extracellular interaction with bacteria. Moreover, M. mass R induces more active cell necrosis and MET formation than M. mass CIP, and the subsequent responses associated with cell necrosis seem to contribute to the release of DNA.
METs have no bactericidal effects on M. mass R Extracellular traps released from neutrophils or other immune cells are known to be microbicidal by entrapping microbes [6,8,21]. To determine whether METs induced by M. mass R infection have anti-bacterial activity, we examined bacterial survival after treatment of infected macrophages with DNase to degrade METs. Interestingly, disintegration of METs by DNase caused a decrease in cell-associated M. mass R compared to untreated macrophages (Fig 6A). On the other hand, there was an increase in extracellular bacteria recovered from culture supernatant in DNase-treated samples (Fig 6B). These results suggest that METs have a role in entrapping microbes and thereby reduce extracellular CFUs in the supernatant, but may not have direct bactericidal activity on M. mass R. The collective amounts of M. mass R recovered from DNase-treated groups were significantly higher than those of untreated groups (Fig 6C). A previous study investigating METs induced by M. tuberculosis reported similar results that bacterial aggregation and growth increased as MET formation was enhanced [23]. To test whether M. mass R aggregation was induced by METs, we performed AFB stain on infected cells (Fig 6D). Indeed, we observed intracellular growth and aggregation of M. mass R were decreased in DNase-treated samples. Instead, extracellular bacteria attached to cell surface were detected more often in the same samples. These data indicate that METs have no bactericidal activity on M. mass R, but rather may increase MET-mediated interaction of bacteria with cell and facilitate stable aggregation of entrapped M. mass R.

Discussion
Macrophages are distributed through-out many body tissues and contribute to tissue homeostasis and immune defense against foreign agents. Upon infection by pathogens, macrophages generally perform the first innate immune responses, engulfing and killing pathogens, and evoking subsequent responses by other immune cells. Recent studies report that macrophages can generate extracellular trap structures consisting of DNA as a backbone with microbicidal proteins, and entrap extracellular bacteria, fungi, and parasites [22,24,25]. Similar to extracellular traps of other immune cells, METs are known to be induced by various stimuli, such as microbes, microbial products or cytokines [21,37]. However, the characteristics and host defense mechanisms of METs remain to be elucidated.
Investigations on the interactions of macrophages with mycobacteria previously focused on intracellular responses after phagocytosis [38,39]. However, M. tuberculosis was recently shown to induce extracellular trap formation by human PBMC-derived macrophages, providing evidence of a novel role of macrophages in extracellular immune defense against mycobacterial infection [23]. In this study, we showed that M. mass, a rapid growing virulent mycobacterium, also induces the formation of extracellular traps from PMA-differentiated THP-1 macrophages. Notably, M. mass R, a rough isolate of M. mass, significantly induced more MET release than a smooth isolate of M. mass (Fig 2). The main difference between the rough and smooth strains is the loss of glycopeptidolipid (GPL) in cell surface, caused by loss or mutation of the genes related to GPL biosynthesis [29]. Rough strains are known to be more virulent and associated with acute respiratory disease [40,41,42]. Because loss of GPL unmasks bioactive cell wall lipids, rough strains induce higher inflammatory reponses than smooth strains [43,44]. This mechanism is considered to be associated with MET formation. Similarly, in a previous study, the cord-forming rough isolate of M. abscessus was found to mainly induce extracellular trap formation by human monocytes [30]. In addition, nonsonicated aggregates of M. mass R was shown to induce more MET formation comparing sonicated single bacteria. This response was also examined in human macrophages infected with M. tuberculosis [23]. Collectively, these results indicate colony type and bacterial structures modulate interaction with macrophages and extracellular trap formation.
Previous study using M. tuberculosis showed IFN-γ pretreatment of human macrophages readily induced rapid necrotic death and MET formation [23]. These responses were strictly dependent on ESX-1. However, our results showed that pretreatment of THP-1 macrophages with IFN-γ or TNF-α do not enhance MET formation. Because M. mass do not have ESX-1 secretion system, M. mass-induced MET formation may be affected by not pretreatment with proinflammatory cytokines but other mechanisms. In addition, chemical agents, such as PMA, LPS or hydrogen peroxide, did not stimulate MET release by human THP-1 macrophages (Fig  2), indicating that extracellular trap release from macrophages is closely associated with signal transductions mediated by specific microbes or antigens [21,23,24].
The formation of extracellular traps by neutrophils or other immune cells is known to mainly depend on NADPH oxidase [6,33,34]. However, our study revealed that the activity of NADPH oxidase did not affect M. mass-induced MET formation (Fig 4A). A previous study reported similar results that E. coli and Candida albicans induce MET production by murine macrophages in an NADPH oxidase-independent manner [24]. On the other hand, another study using bovine macrophages showed that Mannheimia haemolytica-induced MET formation is dependent on the activation of NADPH oxidase [21]. Collectively, these data indicate that the requirements for NADPH oxidase and therefore mechanisms for NET and MET extracellular trap formation differ depending on the stimulus [45].
Recent studies suggest that calcium influx is a major regulator of NOX-independent NETosis and that calcium activated NETosis is induced faster than NOX-dependent NETosis [35,36]. M. mass infection indeed increased the levels of intracellular calcium and MET formation was partially dependent on calcium influx. These results seem to be consistent with M. tuberculosis, which caused calcium influx and extracellular trap formation in neutrophil by secretory antigen, ESAT-6-mediated cytolysis [46]. In addition, calcium-activated NET formation is known to be associated with mitochondrial ROS and activation of the MAPK pathway [35]. However, such mechanisms have yet to be confirmed in MET formation induced by bacterial infection. Further studies are needed to elucidate the molecular mechanism of calcium-mediated MET formation.
Neutrophils generally release extracellular traps toward extracellular microbes, such as Shigella flexneri and Staphylococus aureus, in a phagocytosis-independent manner [6]. However, a recent study reported M. tuberculosis-induced NET formation depends on phagocytosis [47]. In the study, M. tuberculosis-infected neutrophils after pretreatment with CytD closely resembled the uninfected control. Similarly, M. mass-induced MET formation was reduced by inhibition of phagocytosis with CytD, suggesting that phagocytosis of M. mass by macrophages is required for the production of extracellular traps. A previous study reported similar results that CytD-treated bovine macrophages showed a decrease in METs formation in response to M. haemolytica infection [21]. It was also shown that infection of human macrophages with high burden of M. tuberculosis is linked to more extracellular trap formation [23]. These results suggest that extracellular trap production by macrophages mainly depends on inflammatory responses following phagocytosis of M. mass.
NET-associated cell death, designated as NETosis, is a process that is morphologically and functionally distinct from apoptosis or necrosis [11,13]. A previous study reported that neutrophils release nuclear DNA while maintaining membrane integrity and phagocytic responses without cell lysis [13]. However, unlike NETosis, necrotic cell death was observed during MET formation after phagocytosis of M. mass R. Similarly, M. tuberculosis also induced MET formation via their secretory system ESX-1-dependent macrophage necrosis [23]. However, although M. mass CIP induced macrophage cell death, infected cells showed low levels of MET formation. Collectively, these results reveal that M. mass-induced MET formation is accompanied by necrotic cell death, which is different from NETosis and may be affected by specific antigens related to bacterial strains [21,23,24].
Embedded histones and antimicrobial enzymes in NETs are known to have microbicidal effects on entrapped microbes [6,48,49]. However, a recent study reported that some microbes, such as Streptococcus pneumoniae or S. pyogenes, have mechanisms to evade entrapment and destruction by extracellular traps [50,51]. In our study, METs released from macrophages did not show bactericidal activity on M. mass, but rather enhanced bacterial growth.
Previous study investigating METs induced by C. albicans also report similar result [24]. In another study, survival and growth of M. tuberculosis were not inhibited by METs, and bacterial aggregation and burden increased as extracellular trap formation from human macrophages was enhanced by stimulation with IFN-γ [23]. We also found that M. mass growth was enhanced in macrophages with METs. A recent study reported macrophages are able to process extracellular trap via the phagocytosis route [52]. In this regard, we consider entrapping by METs may enhance phagocytosis of M. mass by macrophages and bacterial aggregation, which may facilitate mycobacterial survival and growth. However, the mechanism for evading the bactericidal action of METs and for enhancing mycobacteria survival is not yet fully understood.
In conclusion, our study revealed that rapid growing virulent mycobacteria M. mass infection causes differentiated THP-1 macrophages to release extracellular DNA traps (Fig 7). M. mass-induced METs have the distinct characteristics compared to NETs. Phagocytosis of M. mass by macrophages induces calcium influx and intracellular bacterial growth. And increase in intracellular bacterial number is followed by macrophage cell necrosis and subsequent METs release. In addition, instead of killing entrapped bacteria, METs may enhance intracellular growth of M. mass. This response may be closely related to sequelae of extracellular traps, such as chronic inflammation or autoimmune diseases [53,54,55]. However, METs retain the primary functions of entrapping M. mass and preventing bacterial spread. Furthermore, considering the importance of adaptive immune responses against M. mass, METs may enhance immune responses, priming microbes to be targeted by other immune cells and evoking subsequent immune responses including T cell recruitment and activation [56,57,58]. Thus, our data suggest that METs have complex functions on host interactions with mycobacteria, and more studies are required for elucidating their functions in the context of various immune cells in vivo.