Mycobacterium tuberculosis exploits host ATM kinase for survival advantage through SecA2 secretome

(Mtb) produces inflections in the host signaling networks to create a favorable milieu for survival. The virulent Mtb strain, Rv caused double strand breaks (DSBs), whereas the non-virulent Ra strain triggered single-stranded DNA generation. The effectors secreted by SecA2 pathway were essential and adequate for the genesis of DSBs. Accumulation of DSBs mediated through Rv activates ATM-Chk2 pathway of DNA damage response (DDR) signaling, resulting in altered cell cycle. Instead of the classical ATM-Chk2 DDR, Mtb gains survival advantage through ATM-Akt signaling cascade. Notably, in vivo infection with Mtb led to sustained DSBs and ATM activation during chronic phase of tuberculosis. Addition of ATM inhibitor enhances isoniazid mediated Mtb clearance in macrophages as well as in murine infection model, suggesting its utility for host directed adjunct therapy. Collectively, data suggests that DSBs inflicted by SecA2 secretome of Mtb provides survival niche through activation of ATM kinase.


Introduction
In response to the damage, the host activates an intricate and indispensable signaling cascade entitled 'DNA damage response' (DDR), which not only detects and repairs the damaged lesions in DNA but also regulates the activation of effectors that determine the fate of the cell. Ataxia telangiectasia mutated (ATM), ATM-and Rad3-related protein (ATR) and DNA-dependent protein kinase catalytic subunit (DNA-PKcs) are three drivers of DDR which belong to the family of phosphoinositide 3-kinase like kinases (PIKKs). PIKKs respond to the DNA damage by activating their downstream substrates leading to cell cycle delay/arrest and DNA repair and depending on the extent of damage can also lead to apoptosis. ATM is activated through autophosphorylation of S1981 residue and is subsequently recruited to the double strand breaks (DSBs) through a sensor complex, MRN (MRE11, NBS1 and Rad50) (Lee and Paull, 2005). DSBs are also the site of recruitment and activation of DNA-PKcs, facilitated by a DSB-bound heterodimer Ku70-80 (Gottlieb and Jackson, 1993;Spagnolo et al., 2006). ATR-ATRIP complex is recruited in response to RPA-coated ssDNA (Ball et al., 2005). Even though all the three kinases coordinate their functions independently, crosstalk among these kinases is known to exist, such as ATM-dependent activation of ATR (Shiotani and Zou, 2009); ATM and ATR-dependent phosphorylation of DNA-PKcs Hirokawa et al., 1992) and DNA-PKcs-mediated modulation of ATM (Peng et al., 2005;Zhou et al., 2017). Moreover, these kinases also share substrates such as p53 (Canman et al., 1998;Lakin et al., 1999;Lees-Miller et al., 1990). Phosphorylation of H2AX at the serine 139 (termed gH2AX), at the chromatin regions flanking the damage site is considered as a marker for DNA damage (Burma et al., 2001;Rogakou et al., 1998). While phosphorylation of H2AX is predominantly modulated by ATM, ATR and DNA-PKcs also contribute either partially or entirely (Burma et al., 2001;Royo et al., 2013;Wang et al., 2005). Phosphorylation of H2AX acts as the foundation for recruitment of other DDR mediator proteins eventually leading to DNA repair (Ciccia and Elledge, 2010). Thus, DDR is an indispensable mechanism that restores the genomic integrity and regulates the cellular response by modulating DNA repair, cell cycle progression, apoptosis, or senescence in response to DSBs in the cell. Thus, DDR is also a preferred target of pathogenic bacteria.
Pathogenic Escherichia coli and Helicobacter pylori are capable of imparting notable DNA damage to the host and subsequently impair the DDR to avoid premature cell death (Cuevas-Ramos et al., 2010;Toller et al., 2011). Chlamydia trachomatis triggers DSBs in the host, and gH2AX induction but simultaneously it impairs the DDR by inhibiting the recruitment of 53BP1, ensuing inadequate signal amplification (Chumduri et al., 2013). Listeriolysin O (LLO) secreted by Listeria monocytogenes induces degradation of crucial DNA damage sensor, MRE11. This results in impaired DDR, curtailing the host ability to halt cell cycle, thus successfully promoting multiplication and survival of the pathogen (Samba-Louaka et al., 2014). H. pylori impedes nucleotide repair by downregulating the proteins and their respective transcripts that are involved in mismatch and base excision repair (Kim et al., 2002;Machado et al., 2009). Neisseria gonorrhoeae, H. pylori and C. trachomatis also downregulate p53 levels to promote host cell survival and inhibit apoptosis (Buti et al., 2011;Wei et al., 2010;Vielfort et al., 2013).
Since ancient times Mycobacterium tuberculosis (Mtb) has been steadily evolving sophisticated tactics to dodge the defense responses of the host. One of the remarkable survival strategies of Mtb is to intervene with the fundamental signaling events of the host cell (Koul et al., 2004) and to facilitate these manipulations Mtb secretes an enormous number of characterized and uncharacterized effectors inside the host. These effectors modulate host cellular processes such as phagosome maturation, apoptosis, autophagy, calcium homeostasis, activation of pro-inflammatory responses and TLR, TNFa, MAPK signaling pathways (Dey and Bishai, 2014). However, till date, the role of ATM kinase in the survival of Mtb inside the host has not been investigated.
In this study, we demonstrate that Mtb causes DSBs and determine its impact on the activation of host DDR. SecA2 secretome is necessary and sufficient for inflicting DSBs in the host. We show that instead of classical ATM-Chk2 pathway, Mtb gains survival advantage through activation of ATM-Akt signaling cascade that results in the inhibition of apoptosis. In a chronic mice infection model, Mtb infected lungs showed significant DSBs and activation of ATM. Combining ATM inhibitor, KU55933 with INH resulted in better clearance of Mtb compared with INH treatment alone in the lungs and spleen of infected mice. This study reveals novel exploitation mechanism utilized by Mtb, wherein the pathogen inflicts persistent DSBs in the host to activate ATM-Akt signaling pathway and thereby inhibiting apoptosis and accentuating cell growth.

Mtb inflicts DNA damage in the host cell
To address the question of whether Mtb infection leads to the damage of the host DNA we used PMA differentiated THP-1, RAW264.7 (RAW) macrophages and primary murine peritoneal macrophages (PF). The cells were infected with the virulent Mtb strain, H37Rv (Rv) and the gH2AX levels, the hallmark of DNA damage, was evaluated. Results showed considerable DNA damage in the infected cells compared to the corresponding uninfected control and these observations were consistent across all the three cell types (Figure 1a-c). The damage could be observed as early as 1 hr post infection (p.i) ( Figure 1b) and persisted even at 72 hr ( Figure 1a). We assessed if the observed DNA damage is dependent on the presence of live bacteria by infecting cells with live or heat killed Rv. Results showed that (Figure 1d) only the live bacilli could cause damage to the host genome. To evaluate the role of Mtb virulence in inflicting genotoxicity, we performed infection experiments with Rv or its avirulent counterpart H37Ra (Ra). Based on its characteristics such as reduced survival under anaerobic conditions, inability to produce persistent infection in mice and guinea pigs, Ra is considered as attenuated, avirulent strain of Mtb (Alsaadi and Smith, 1973;North and Izzo, 1993). While both Ra and Rv infected THP1 cells showed considerably higher gH2AX levels compared with the uninfected control, there was consistent and noticeably higher levels of gH2AX levels in Rv compared with the Ra infected cells (Figure 1e; compare 4, 24 and 48 hr time points). However, in case of RAW cells the higher gH2AX levels in Rv was apparent only in the early time points (Figure 1f and g: up to 4 hr). Subsequent to DNA damage, H2AX gets phosphorylated on the chromatin flanking the site of damage, which appears as foci in the nucleus. We performed immunofluorescence experiments using antibodies specific for gH2AX and observed significantly higher numbers of gH2AX foci at both 4 and 24 hr p.i ( Figure 1h) compared with the uninfected controls. Consistent results were observed in infected THP1 and PF cells (Figure 1-figure supplement 1a and b). Collectively, the data suggests that both Rv and Ra possess genotoxic characteristics and virulent Rv strain triggers rapid damage post infection. (a) PMA differentiated THP-1 cells were infected with Rv for 4 hr. Cells were washed thrice with 1XPBS to remove extracellular bacilli and supplemented with fresh media for further time points which were calculated with respect to 4 hr point. (b and c) RAW264.7 macrophages (RAW) (b) and murine peritoneal macrophages (PF) (c) were infected with Rv as described in (a) except that the extracellular bacilli were washed off after 1 and 2 hr post-infection (p.i) For the remaining time points, 4 hr serves as the reference point. (d) RAW264.7 macrophages were infected with Rv or heat-killed Rv (HK-Rv) and cell lysates were prepared at 4 and 24 hr p.i (e-g) THP-1 or RAW264.7 cells were infected with Rv and Ra. Cell lysates were prepared as indicated above after 1, 2, 4, 24, 48 or 72 hr p.i. (a-g) Whole cell lysates (WCL) were resolved on SDS-PAGE, transferred to nitrocellulose membrane and probed with a-gH2AX(S139) and a-b-actin/a-b-tubulin antibodies. (h) Representative immunofluorescence image showing gH2AX foci (AlexaFluor594-Red) and nucleus (DAPI-blue) in RAW264.7 infected with GFP expressing Rv and Ra (Green) independently at 4 and 24 hr p.i. Images were captured at 63X magnification with scale bar of 5 mm. The online version of this article includes the following figure supplement(s) for figure 1: Host cell activates DNA damage response Subsequent to the challenge to genome integrity, eukaryotic cells activate the DDR, which initiates either cell survival or cell death signals contingent upon the magnitude of damage. Depending on the type of damage, -DSBs or single-stranded DNA (ssDNA; due to replication fork stalling) either ATM-Chk2 or ATR-Chk1 pathways, respectively, are activated ( Figure 2a; Lee and Paull, 2007;Zou and Elledge, 2003). To determine the type of DDR involved in Ra and Rv inflicted damage, lysates prepared at 4 and 24 hr p.i were examined for the phosphorylated forms of ATM, Chk2, ATR and Chk1, indicative of their activation. Infection by Ra led to the formation of distinct nuclear pRPA2 ( Figure 2c Shiloh, 2003). To corroborate these findings, we evaluated gH2AX foci for the presence of mediator proteins 53BP1, which accumulates at the DNA damage site. Distinct focal pattern and colocalization of gH2AX and 53BP1 foci in Rv infected PF supported the argument that infection with Rv induces DSBs ( Figure 2f).

Intracellular Mtb causes continuous DSBs
Mtb infection not only elicits early DNA damage ( Figure 1f) but that also persists at higher levels for prolonged duration (Figure 1a, e and Figure 3a left panel). DDR is necessary for sensing, amplifying and eventually repairing the damage. However, persistent DSBs indicate either inefficient DDR or continued presence of the factors released by Rv that are responsible for the damage. To determine whether there is a direct correlation between the levels of gH2AX and Rv load, we evaluated gH2AX levels in the absence or presence of isoniazid (INH), the front line anti-tuberculosis antibiotic (Nuermberger and Grosset, 2004).  Figure 3b). Thus, the data suggests that even though DDR is active, inefficient repair is due to continuous generation of DSBs during prolonged Rv infection.
Apart from activating signals for repair, the DDR stalls the DNA replication and cell division by activating the cell cycle checkpoints (Smith et al., 2010). To examine the consequence of Rvinduced DSBs on the host cell cycle progression, RAW cells were infected with GFP-expressing Rv and the populations of infected and uninfected cells were analyzed through flow cytometry ( Figure 3d). Flow cytometry profile of uninfected control cells would show single population, whereas Rv-infected RAW cells would have two populations: GFP+ve cells harboring intracellular GFP-Rv and GFP-ve uninfected bystander cells that lack intracellular GFP-Rv ( Figure 3d). Interestingly, compared with the uninfected cell population, Rv infected GFP+ve population showed decreased accumulation of cells in G1 phase, soon after the infection (4 h p.i). Lowered accumulation of cells in the G1 phase suggest alterations in the progression of cell cycle in the cells harboring intracellular bacilli ( Figure 3e). Importantly, at late stages of infection we observed concomitant increase in the sub-G1 cell population, indicative of cell death (Figure 3e). DDR activation maintains the genomic stability by delaying the occurrence of cell division till DNA is repaired, else in case of irreversible damages; key regulators such as p53 mediates apoptosis (Figure 3f). To corroborate flow cytometry results, we evaluated key molecular markers that modulate the transition of cells through different cell cycle phases (Abraham, 2001;Matsuoka et al., 1998). Elevated levels of p53 and p21 upon Rv infection are reflective of molecular level alterations in the cell cycle leading to Mtb SecA2 secretome is necessary and sufficient for DNA damage Pathogenic bacteria often release genotoxins and cyclomodulins to modulate host cellular processes (Nougayrède et al., 2005). Whi, WhiB3, a redox regulator in Mtb has been shown to regulate the production of specific polyketides and lipid cyclomodulins in the host (Cumming et al., 2017). We speculated that the Mtb secretome might play a role in promoting DNA damage. Culture filtrate (CF) of Mtb contain proteins that are translocated into the extracellular milieu ( Figure 4a). Addition of Mtb CF resulted in robust gH2AX induction, even higher than those observed upon Rv infection ( Figure 4b). Importantly as little as 1 mg/ml CF treatment induced significant gH2AX ( Figure 4c). In Mtb, proteins are secreted using both classical secretory pathways such as TAT and SecA1, which are essential for in vitro growth; and accessory secretion pathways such as SecA2 and Type VII secretion systems, which are necessary for survival and virulence in the host (Feltcher et al., 2010;Ligon et al., 2012;Miller et al., 2017). The TypeVII secretion system and its effectors are encoded by Region of difference 1 (RD1) in Mtb. Host cells infected with Rv or RvDRD1(Rv carrying RD1 deletion) or RvDCE (Rv strain with CFP10-ESAT6 deletion) had comparable gH2AX levels ( (Braunstein et al., 2003;Kurtz et al., 2006;Miller et al., 2017), we were unable to rule out the possibility that the decreased gH2AX levels could be due to better clearance of the pathogen. To address this issue, cells were treated with CFs prepared from either Rv or RvDRD1 or RvDsecA2 and the gH2AX levels were evaluated. While the gH2AX levels in cells treated with CF from RvDRD1 were comparable to those observed upon Rv-CF treatment; cells treated with RvDsecA2-CF showed negligible gH2AX levels ( Figure 4g). Treatment of cells either with Rv-CF or RvDRD1-CF led to the activation of DDR players, ATM and Chk2. However, treatment with RvDsecA2-CF cells did not show ATM-Chk2 activation. Analogous results were obtained when the experiment was performed with PF ( Figure 4h). Together, results suggest that effectors secreted by SecA2 pathway are necessary and sufficient for inflicting genotoxic stress in the host.

ATM activation confers survival advantages to Mtb
DNA damage mediates the activation of ATM kinase, which in addition to activating DDR is also known to modulate other cellular processes such as apoptosis, DNA repair, cell division, autophagy and inflammatory responses (Kastan and Lim, 2000). We assessed the impact of host ATM kinase activation in regulating the survival of Mtb with the help of KU55933 (ATM-I), a specific inhibitor of ATM activation (Hickson et al., 2004). Treatment of infected cells with ATM-I inhibitor resulted in substantial decrease in the gH2AX levels and abrogation of ATM and Chk2 activation ( Figure 5a). This was also reflected in disappearance of pATM foci upon ATM-I treatment in Rv-infected cells ( Figure 5b). Notably, ATM-I treatment of Rv-infected cells resulted in reduced survival of the pathogen in a dose dependent manner ( Figure 5c and  signalling cascade coordinating the DNA damage, RAW264.7 macrophages infected with Rv, Ra or HK-Rv. WCL prepared at 4 and 24 h p.i. were subjected to immunoblotting with a-gH2AX, a-pATM(S1981), a-ATM, a-pCHK2(T68), a-CHK2 (left panel) and a-b-actin, a-ATR(S428), a-ATR, a-pCHK1 (S345), a-CHK1 (right panel) antibodies. (c) Representative immunofluorescence showing presence of pRPA2-S4/8 foci (AlexaFluor594-Red) in uninfected or GFP-Rv/GFP Ra (Green) infected RAW264.7 macrophages at 24 hr (d) WCL prepared from PF infected with Rv for 4 and 24 hr were immunoblotted with a-gH2AX, a-pATM(S1981), a-ATM, a-pCHK2(T68), a-CHK2 and a-b-actin antibodies. (e-f). Representative images of immunofluorescence showing (e) pATM(S1981) (AlexaFluor594-Red) and (f) gH2AX, 53BP1 foci formation (AlexaFluor594-Red and AlexaFluor647-yellow) in PF infected with GFP expressing Rv (Green) for 24 hr. Nuclei were stained with DAPI. Images were captured at 100X magnification with 2.9 X optical zoom in LSM510 Meta System (Zeiss, Germany) confocal microscope. Scale bar: 2 mm. The online version of this article includes the following figure supplement(s) for figure 2:  . Cells at indicated time points were lysed with 0.05% SDS, serially diluted and plated on 7H11 plates to enumerate bacillary load. Bar graph on y-axis represents Mean CFU (log 10 ) ± SD of one of the three independent experiments performed in triplicates. **p 0.005, ***p 0.0005, (c). WCL were prepared at indicated time points were probed with a-gH2AX, a-pATM(S1981), a-ATM, a-pChk2(T68), a-Chk2 and a-b-actin antibodies. (d) Schematic diagram illustrating the experimental approach for the cell cycle analysis using flow cytometry. RAW264.7 cells were infected with GFP expressing Rv. FACS profile showing GFP+ve population containing Rv and GFP-ve population that lacks intracellular Rv (bystander uninfected cells). Uninfected control cells are GFP-ve. (e) Pie chart showing percentage of uninfected and GFP+ve infected cells in different phases of cell cycle. Infection was performed at the indicated timepoints. (f) Schematic representation of signaling activated due to DSB occurrence in the host cell. ATM phosphorylates its self and H2AX. Subsequently it activates the downstream effector, Chk2 at T68. Activated Chk2 stabilizes p53 which in turn elevates p21 levels. Upregulation of p53 can divert cells toward apoptosis while p21 leads to cell cycle arrest (g). WCLs prepared from RAW264.7 infected with GFP expressing Rv for indicated time points were subjected to immunoblotting with a-gH2AX, a-p21, a-p53, a-pChk2(T68), a-Chk2 and a-b-actin antibodies. The online version of this article includes the following source data for figure 3:

Inhibition of ATM-Akt axis hypersensitizes host cells to apoptosis
In addition to ATM and ATR, H2AX can also be phosphorylated by DNA-PK catalytic subunit (DNA-PKcs), the third PIKK that also coordinates DDR. DNA-PKcs is recruited to the DSBs at the heterodimer complex of Ku70/80 to form the holoenzyme DNA-PK, wherein it initiates the repair. Infection of PF with Rv resulted in the activation of DNA-PKcs, albeit to much lower levels and at later time points compared with the ATM (Figure 6a). Results in Figure 5a showed presence of residual gH2AX despite ATM-I treatment, thus we sought to assess the effect of combinatorial inhibition of ATM and DNA-PK or ATR on the levels gH2AX. PF treated with ATM-I or ATM-I + DNA-PK-I (NU7441; inhibitor of DNA-PK) or ATM-I + ATR-I(VE-281; inhibitor of ATR) were infected with Rv and the extent of H2AX phosphorylation was evaluated (Figure 6b). PF irradiated at 2Gy was used as the positive control ( Figure 6-figure supplement 1). While the addition of ATR-I to ATM-I did not alter the residual levels of gH2AX, addition of DNA-PK-I to ATM-I resulted in further decrease in the levels of gH2AX, corroborating data in Figure 6a. Next, we assessed the role of DNA-PK and ATR in the survival of pathogen inside the host. Toward this, host cells treated with inhibitors individually were infected with Rv and were enumerated for CFUs (Figure 6c and Figure 6-source data 1). It is evident that addition of DNA-PK-I or ATR-I did not alter the CFUs, suggesting that even though DNA-PK seems to be activated upon Mtb infection, it does not influence the survival of the pathogen (Figure 6c and Figure 6-source data 1). Taken together, the data suggests that ATM might be feeding in to some parallel pathway which is necessary for the intracellular survival of Mtb.  showing pATM(S1981) foci formation (AlexaFluor594-Red) in PF infected with GFP-Rv (Green) in the presence or absence of ATM-I for 24 hr p.i. Nuclei were stained with DAPI. Images were captured at 100X magnification. Scale bar: 2 mm. (c) PF were infected with Rv and treated with 1, 5 or 10 mM of ATM-I as described above. Cells were lysed with 0.05% SDS to release the intracellular Mtb bacilli and CFUs were enumerated. (d-e) PF were pretreated with 10 mM ATM-I or Chk1/2 inhibitor (Chk2-I) or DMSO (vehicle). Cells were lysed either with RIPA or 0.05% SDS. (d) RIPA WCL were subjected to immunoblotting using a-pATM, a-ATM, a-pChk2(T68), a-Chk2 and a-GAPDH antibodies. (e) SDS lysates were serially diluted and plated on 7H11 plates to determine CFUs. The online version of this article includes the following source data and figure supplement(s) for figure 5: Source data 1. ATM activation provides survival advantages to Mtb. Figure 5 continued on next page An established survival strategy employed by Mtb is to circumvent host apoptosis and upregulate pro-survival signals (Srinivasan et al., 2014). PI3K/Akt pathway, which upon activation inhibits apoptosis and promotes cell survival, has been shown to be up regulated during Mtb infection (Figure 6d; Lachmandas et al., 2016). Western blot analysis showed that treatment with either ATM-I or Akt-I inhibited the activation of ATM and Akt, respectively (Figure 6e). Data suggests that administration of Akt inhibitor (Akt-I) alone or in combination with ATM-I significantly reduced the intracellular survival of Mtb (Figure 6f). Treatment with a combination of ATM-I + Akt-I showed survival defects similar to Akt-I alone indicating that ATM mediated signals are channeling through Akt (Figure 6f and Figure 6-source data 1). While treatment of cells with ATM-I resulted in reduced activation of Akt, addition of Akt-I did not impact activation of ATM, suggesting that ATM activation is upstream of Akt activation (Figure 6e). Interestingly, the impact of Akt inhibitor on the survival of Rv was higher compared with ATM-I (Figure 6f; Figure 6-source data 1). Moreover, while the addition of ATM-I decreased activation of Akt, the reduction with Akt-I was higher, suggesting that other signaling pathways could also be involved in activating Akt (Figure 6e).
There exists a possibility that the survival differences observed could be due to reduced viability of host cells in the presence of inhibitor. To negate the cytotoxic effects of inhibitors, we evaluated the survival of uninfected and Rv infected PF in the presence of different concentrations of ATM-I or Akt-I or Chk2-I ( Figure 6-figure supplement 2). Treatment of uninfected PF with ATM-I or Akt-I showed reduced viability, however, infection of PF with Rv marginally decreased the survival, compared with the UI cells, suggesting that the modulation of pathogen survival is not due to cytotoxicity (Figure 6f and Figure 6-figure supplement 2). To analyze the impact of ATM-I or Akt-I or ATM-I + Akt-I on the apoptosis, we performed flow cytometry analysis of UI and Rv infected cells in the absence and presence of inhibitors using Annexin V/7-AAD staining. While the addition of inhibitors did not impact the percent live cells (Figure 6g; top panel and Figure 6-source data 1) in UI, their presence decreased the percent live population in Rv-infected cells (Figure 6g; top panel and Figure 6-source data 1). We observed concomitant increase in the early and late apoptotic cells in the presence of ATM-I or Akt-I or ATM-I + Akt-I (Figure 6g; bottom panel and Figure 6-source data 1). Thus, the data suggests that Rv inflicts DSBs to activate ATM which in turn activates Akt resulting in anti-apoptotic and pro-survival signals which favors Mtb survival (Figure 7g).

ATM kinase is a potential therapeutic target for host directed TB therapy
Persistent DNA damage drives genomic instability (Tubbs and Nussenzweig, 2017). Continuous DNA damage triggered during prolonged Mtb infection in established and primary cell lines encouraged us to examine the occurrence of possible DSBs in Mtb infected mice lungs. BALB/c mice challenged with Rv strain through aerosolic route were evaluated for bacillary load at 1-and 56 days p.i (Figure 7a and Figure 7-source data 1). We observed appearance of granulomatous lesions and significant bacillary load at 56 days p.i suggesting that Mtb has established chronic infection in the lungs (Figure 7-figure supplement 1a). Subsequently, tissue lysates prepared from uninfected and Rv infected mice lungs were subjected to western blotting to evaluate the expression of molecular markers corresponding to DNA damage and DDR. While the levels of gH2AX, pATM, pChk2, and pAkt was negligible in the uninfected mice lungs, we observed significantly elevated levels of gH2AX, pATM, pChk2 and pAkt in the lungs of Rv infected mice, suggesting that Mtb infection produces sustained DSBs and Akt mediated proliferation (Figure 7b and Figure 7-figure supplement  1b). In a similar vein, we also observed increased levels of gH2AX in the lysates obtained from spleen of infected mice (Figure 7b; bottom panels). CFUs in spleen after 56 days of infection suggests dissemination of Rv from lungs (Figure 7a). Collectively the results suggest that pathogenic Mtb induces genotoxicity both ex vivo and in vivo resulting in deleterious DSBs in the host genome.
DSBs activate ATM, which in turn phosphorylates H2AX to mark the damaged site and activates Chk2 and p53 to alter the cell cycle progression. In parallel, it also activates Akt to delay cell death and promote survival of the pathogen (Figure 7b). We hypothesized that Mtb mediated  To determine the activation of DNA-PKcs upon Rv infection, PF were infected with Rv as indicated and WCL were immunoblotted with a-pDNA-PKcs-S2056, a-DNA-PKcs, a-pATM(S1981), a-ATM, and a-b-actin. WCLs prepared from PF irradiated at 2Gy was used as positive control. (b) PF were pre-treated with 10 mM ATM-I or 10 mM ATM-I + 1 mM DNA-PK-I or 10 mM ATM-I + 10 mM ATR-I for 1 hr prior to Rv infection. 4 h p.i, extracellular bacilli were washed off and cells were replenished with fresh media containing above inhibitors with indicated combinations for 24 and 48 hr. Cells were lysed with RIPA to prepare WCL which were then subjected to immunoblotting against a-gH2AX and a-b-actin antibodies. (c) PF were infected with Rv and treated with DMSO or 10 mM ATM-I or 1 mM DNA-PK-I or 10 mM ATR-I individually as described above. Cells were harvested at indicated time points and CFUs were enumerated as described previously. (d) Figure 6 continued on next page genotoxicity provides a survival niche through activation of ATM kinase (Figure 7b). We sought to explore the possibility of utilizing ATM inhibitor ATM-I as a possible candidate that can be used in combination with INH towards adjunct host directed therapy (HDT) (Rayasam and Balganesh, 2015). To investigate this possibility, PF were infected with Rv and the infection was allowed to establish for 24 hr.

Discussion
The present study was designed to address the following questions: i) Does Mtb target the genomic integrity of the host? If so, do virulent and avirulent Mtb target the genomic integrity of the host differentially, and in that case, how? ii). How the host cell responds to the DNA damage mediated through Mtb? iii). What is the role of the Mtb secretome, if any, in imparting DNA damage? iv). What is the role of ATM kinase activation and subsequent downstream signaling in the survival of pathogen within the host? vi). Is this Mtb mediated DNA damage manifested in the mice model of in vivo infection?
Results show that Mtb infection promotes gH2AX upregulation in ex vivo models of murine/ human origin and murine primary macrophages. The induction of gH2AX and subsequent formation of foci could also be detected by immunofluorescence (Figure 1a-h). These results are in accordance with earlier studies wherein induction of gH2AX foci and genomic instability was detected through immunofluorescence and in situ hybridization, respectively, in the ex vivo model of Mtb infection (Castro-Garza et al., 2018;Mohanty et al., 2016). While the induction of gH2AX is dependent on the presence of the live bacteria, Ra a non-pathogenic counterpart of Rv could also induce gH2AX levels, albeit at comparatively slower rates (Figure 1d-g). In addition to temporal differences, Ra and Rv appear to inflict distinct types of damages resulting in the activation of ATR and ATM pathways, respectively (Figure 2b). The infection with Rv results in lingering gH2AX levels even at later time points, suggesting the sustained presence of pathogenic Mtb continues to inflict DNA damage (Figure 3a). In consonance with this hypothesis, elimination of intracellular Mtb drastically Figure 6 continued WCL from PF infected with Rv for indicated time points were immunoblotted with a-pAkt (S473), a-Akt antibodies (e-f) PF were pre-treated with either 10 mM of ATM-I or Akt inhibitor (Akt-I) or ATM-I + Akt-I for 1 hr prior to Rv infection. 4 hr p.i., extracellular bacilli were washed off and cells were replenished with fresh media containing inhibitors. At defined time points, cells were lysed either with RIPA or 0.05% SDS. (e) WCL were subjected to immunoblotting with a-pATM(S1981), a-ATM, a-pAkt (S473), a-Akt and a-b-actin antibodies. (f) Serially diluted SDS lysates were plated on 7H11 to determine CFU. (g) At 48 hr time point cells were scraped in PBS and subsequently stained with Annexin V-FITC and 7AAD. Cells were analyzed by flow cytometry to determine the percentage of live cells (upper panel) or cells undergoing early/late apoptosis (lower panel). Inhibitor treatments were as described above. Readings are average of three biological replicates. Error bar, SD. *, p 0.05; **, p 0.005; ***, p 0.0005. Significance was calculated with respect to Rv. The online version of this article includes the following source data and figure supplement(s) for figure 6: Source data 1. ATM inhibition hypersensitizes host cells to apoptosis.  reduced gH2AX levels/DNA damage, suggesting Mtb may be inducing DNA damage at a faster frequency relative to the host's ability to repair them (Figure 3c).
In response to the DSBs generated upon pathogenic E. coli infection, host activates phosphorylation of Cdc25 through ATM-Chk2 pathway to hamper cell cycle progression. However, at higher MOI, host cell undergoes p53 SUMOylation through SENP1 downregulation to promote senescence (Cougnoux et al., 2014). On the other hand, N. gonorrhoeae decelerate the host cell cycle by reducing p53 and concurrent elevation of p21 and p27 (Vielfort et al., 2013). L. monocytogenes in addition to impairing host MRE11 also activates DDR in ATM/ATR independent fashion eventually resulting in cell cycle delay (Leitão et al., 2014). Taken together, it is evident that pathogens utilize Figure 7. ATM kinase is a potential therapeutic target for host directed TB therapy. (a) Graphical representation depicting Rv load at day 1 (n = 4) and day 56 (n = 5) of infection in lungs and spleen of mice. Lung homogenates prepared at indicated time points were plated on 7H11 plates to enumerate CFU per lung or per spleen (b) Lung and spleen lysates were prepared from uninfected and Rv infected mice 56 days p.i. 100 mg lung lysates were subjected to immunoblotting with a-gH2AX, a-pATM(S1981) a-ATM, a-pCHK2(T68), a-CHK2, a-pAkt (S473), a-Akt, a-p53 and a-b-actin antibodies. (c) PF were infected with Rv at high MOI (1:10) as described above. 24 h p.i, cells were treated with either 3.6 mM isoniazid (INH) or 10 mM ATM-I alone or INH + ATM-I together. CFUs were enumerated at 24, 48 and 72 hr post treatment. (d) Schematic representation of the mice infection and drug treatment protocol used. (e-f) CFUs were enumerated in lungs of mice at Day 1, and in the lungs (e) and spleen (f) on day 15 and 30 of infection. Number of mice in each batch was 6 (n = 6) except in Rv at Day 1(n = 5). Error bar, SD. *, p 0.05; **, p 0.005; ***, p 0.0005. (g) Model depicting the findings. Rv induces genotoxicity and causes deleterious DSBs in the host genome through SecA2 secretome. Host cell in response to the occurrence of DSBs activate ATM kinase and is recruited at the site of damage by the sensor, MRN complex. Activated ATM autophosphorylates itself and phosphorylates H2AX in the chromatin flanking the sites of DNA damage, which becomes the foundation for the recruitment of mediator protein MDC1, thus amplifying DDR. pATM promote recruitment of 53BP1 at the damage site. pATM as a part of DDR also activates downstream effectors, Chk2 and p53, which are responsible for alterations in the host cell cycle. pATM in a parallel pathway also activates Akt, which is known inhibit apoptosis and promote cell survival. Activation of ATM and Akt and subsequent inhibition of apoptosis provides survival advantage to Rv. Inhibition of ATM or Akt activation through inhibitors, ATM-I or Akt-I, respectively, promote host cell apoptosis, which impedes the bacilli growth. Phosphorylation status of proteins is depicted with a P in a circle. The online version of this article includes the following source data and figure supplement(s) for figure 7: Source data 1. ATM kinase is a potential therapeutic target for host directed TB therapy.  diverse and distinct mechanisms to perturb the host cell cycle to avoid untimely death. We observed activation of ATM-Chk2 pathway (Figure 2b) and subsequent decrease in the cells in G1 phase (Figure 3e), an indicator of differed host cell cycle at the early stages of Mtb infection. However, at the later time points, infected cells showed increased cell death, most likely an effort by the host to curb the transmission of genotoxic effects (Figure 3e). Interestingly, cell cycle perturbations were limited to the cells possessing intracellular bacilli but not extended to bystander-uninfected cells (Figure 3e). Thus, our results demonstrate that Mtb mediated DNA damage disturbs the homeostatic coordination among the cellular components participating in cell cycle.
Toxins like Cytolethal distending toxin (CDT), typhoid toxin and colibactin from pathogenic bacteria are examples of characterized toxins that are reported to block the host cell cycle progression after inducing DNA damage in the host (El-Aouar Filho et al., 2017). Persistent presence of DSBs and modulation of host cell cycle in cells harboring bacilli, propelled us to examine the role of Mtb secretory factors in imparting genotoxicity. The proteins exported from Mtb are considered a source of virulence, however not many secreted proteins have been individually characterized. Mtb secretes >300 secretory factors with the help of secretory systems such as classical TAT and SecA1; accessory SecA2 and TypeVII. Results presented in Figure 4e-h suggest that SecA2 mediated secretory factors are principally responsible for the DSBs in the host. Mass spectrometry-based study has identified the secretion of 37 proteins, including Mce1, Mce4 family and multiple solute binding lipoproteins to be dependent on the SecA2 pathway. In addition to these, secretion of protein kinase G (PknG), superoxide dismutase (SodA) and phosphatase SapM proteins are also through SecA2 pathway (Feltcher et al., 2015;Miller et al., 2017;Sullivan et al., 2012;Zulauf et al., 2018). As polyketides and lipids from mycobacteria were suggested to be cyclomodulins; therefore, we speculate that one or more lipoproteins secreted through the SecA2 pathway could be critical for genotoxicity (Cumming et al., 2017;George et al., 1998). Macrophages generate reactive oxidation species as an anti-microbial activity to combat the survival of Mtb intracellularly. Mice lacking p47 and gp91 subunit of NADPH oxidase (gp91 phox-/-) are susceptible to Mtb as the macrophages in these mice fail to generate ROS (Cooper et al., 2000). While the primary purpose of generated ROS is to clear the intracellular bacilli, it can however damage its own DNA as exemplified in case of C. trachomatis infection (Chumduri et al., 2013;Yu and Anderson, 1997). On the contrary, H. pylori, L. monocytogenes mediated DSBs are ROS-independent. SodA Mtb , which is secreted through SecA2 pathway plays an important role in detoxification of ROS generated inside the host thus partially neutralizing the anti-microbial property of ROS (Chan et al., 1992;Lau et al., 1998;Piddington et al., 2001). In RvDSecA2 mutant SodA Mtb secretion is expected to be compromised and hence one would expect higher levels of host ROS upon infection. The fact that RvDSecA2 mutant fails to inflict DSBs suggest that Mtb mediated damage is unlikely to be ROS-dependent. In agreement with this, addition of N-acetyl cysteine (NAC), a scavenger of ROS, to the infected cells did not alter the gH2AX, suggesting that DNA damage is not mediated through ROS (Figure 7-figure supplement 1c). Our future study would be aimed towards identifying the specific effector(s) and the mechanism of how the effector brings about DSBs in the host.
Mtb is known to modulate multiple host signaling pathways to establish favorable environment in the host for its survival. Moreover, a previous study (Hinchey et al., 2007) showed that RvDsecA2 strongly induced markers of apoptosis in both human and murine macrophages suggesting a critical role played by SecA2 secretome in preventing apoptosis. Thus, we investigated the role of Mtb induced host DNA damage and subsequent activation of ATM-Chk2 pathway in facilitating and enhancing its survival. While presence of ATM inhibitor significantly compromised the Mtb survival, addition of inhibitor for one of the downstream effectors, Chk2 which is associated to cell cycle checkpoints activation, failed to do so (Figure 5e). These results suggest that while the survival signals for Mtb are dependent on activation of ATM, it is not channeled through the classical DDR. High-throughput study revealed that following DSBs, ATM-dependent phosphorylation of~700 substrates regulate multitude of cellular processes such as cell cycle, cellular differentiation etc., emphasizing its global regulatory role. Importantly, DDR was suggested to crosstalk with the IGF-1/PI3K/ Akt pathway to elicit cell survival signals, which is also supported by data that demonstrated ATMdependent phosphorylation of Akt at S473 (Khalil et al., 2011;Matsuoka et al., 2007). Previous reports suggested that inhibiting Akt diminished Mtb survival inside host cells and proposed it to be a novel target for host directed therapy (Kuijl et al., 2007). Thus, we hypothesized a probable divergence of ATM signaling towards Akt activation may play a role in influencing Mtb survival. We observed that addition of Akt-I compromises the survival of Mtb (Figure 6f). Furthermore, combinatorial addition of both ATM and Akt inhibitors did not show either additive or synergistic impact on the Mtb survival, suggesting that survival signals are routed through ATM-Akt axis (Figure 6f).
Mtb infection of the host cells can induce host cell death through both apoptotic and non-apoptotic pathways. Apoptotic death is considered to be advantageous for the host as it results in the elimination of pathogen (Fratazzi et al., 1999;Gan et al., 2008;Molloy et al., 1994). Induction of non-apoptotic death of the host cells is beneficial for the bacilli as it fails to eliminate the pathogen (Dobos et al., 2000;Keane et al., 2000;Lee and Paull, 2005;Martin et al., 2012;Park et al., 2006). Recently, it was reported that internalization of Mtb aggregates by the host cell results in killing of the host and there was a direct correlation between the size of aggregates and cell death. However, the death in such cases was through non-apoptotic pathway (Mahamed et al., 2017). In our experiments we kept MOI at one to circumvent aggregation and non-apoptotic cell death of the host. We observed significantly increased apoptosis upon the addition of inhibitors for either ATM or Akt or both (Figure 6g). These results are in line with the previous findings, wherein bacterial pathogens such as Salmonella typhimurium and Coxiella burnetii have been shown to get survival advantage through Akt activation (Kuijl et al., 2007;Voth and Heinzen, 2009).
We hypothesize that persistent DSBs caused by Mtb could be a means to constitutively activate ATM, which in turn would activate Akt and inhibit apoptosis thus facilitating the survival of the pathogen. We corroborated our finding in the murine model of chronic tuberculosis infection, wherein the infected lungs showed the activation of gH2AX and ATM-Chk2 signaling (Figure 7a-b). Addition of ATM inhibitor to the INH regime substantially reduced the bacilli load in the host establishing ATM as a potent target for HDT (Figure 7c-f). In conclusion, our study demonstrates that Mtb through SecA2 secretome challenges the host fitness by damaging the genome (Figure 7g). In an effort to repair the damage, the host cell activates the ATM kinase mediated signaling cascade, which is exploited by Mtb to survive. We believe we have identified a novel survival mechanism utilized by Mtb, wherein the pathogen constantly challenges the host genome leading to the activation of pro-survival ATM-Akt signals (Figure 7g). Hence, we propose the use of ATM inhibitors as adjunct for HDT in the treatment of tuberculosis.

Mtb strains, cell culture and Mtb infection
Mtb strains, H37Rv (Rv) and H37Ra (Ra), RvDRD1 and RvDSecA2 were propagated in Middlebrook 7H9 broth (Difco) supplemented with 10% ADC (Albumin-Dextrose-Catalase complex), 0.2% glycerol (v/v), 0.05% (v/v) Tween-80 with shaking at 100 rpm at 37˚C. Rv and Ra strains were electroporated with pSC-301-GFP plasmid (Cowley and Av-Gay, 2001) to generate GFP expressing strains. Human acute monocytic leukemia cell line, THP-1 (ATCC) and RAW264.7 (ATCC), peritoneal macrophages (PF) derived from Balb/c mice were grown in the cell culture media (RPMI-1640 constituted with 1X Anti-Anti and 10% heat inactivated (HI)-fetal bovine serum) at 37˚C. THP-1 cells were treated with 10 ng/mL PMA for 24 hr followed by 12 hr resting without PMA before Mtb infection. RAW264.7 cells were seeded 12-15 hr before Mtb infection. PF were extracted from thioglycolate injected Balb/c mice as described earlier (Zhang et al., 2008). PF isolated from the peritoneal cavity were resuspended in cell culture media and infected 12-15 hr post plating with Mtb. THP1 cells used in the study have been authenticated by STR profiling. All the cell lines used were free of mycoplasma contamination. Single-cell suspensions of actively growing Mtb strains were prepared and all infections were performed at MOI of 1, else indicated. Four hours p.i extracellular bacilli were removed by washing thrice with PBS and the cells were supplemented with fresh cell culture media for defined periods of time. For the inhibitor studies, RAW or PF were pre-treated with different doses (1, 5 or 10 mM) of ATM kinase inhibitor (KU55933, SML1109, Sigma; S1092, Selleckchem) or Chk2 inhibitor (C3742, Sigma) or Akt1/2 kinase inhibitor (A6730, Sigma), ATR inhibitor (VE-281, SML1415, Sigma), DNA-PK inhibitor (NU7441, 18003649897, Cayman chemical company). ATM-I, DNA-PK-I and ATR-I were used at reported dose of 10 mM, 1 mM and 10 mM, respectively. 4 hr p.i cells were supplemented with fresh media containing inhibitors. For the clearance of intracellular Mtb from infected RAW cells, cells were treated with 1 mg/ml (7.3 mM) of isoniazid (INH). Colony forming units (CFUs) were enumerated at 24, 48 and 72 hr p.i to evaluate the bacterial load. RAW/PF were lysed in 0.05% SDS and different dilutions of the lysates were plated on 7H11 agar plates supplemented with 10% OADC (oleic acid-Albumin-Dextrose-Catalase complex). After 21 days, CFUs of Mtb were enumerated.

Preparation of lysates, western blotting, and immunofluorescence
Cell lines were lysed and snap-frozen mice lungs and spleen were macerated in cold RIPA buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 1.0% NP-40, 0.5% Sodium deoxycholate, 0.1% SDS, freshly supplemented with PhosSTOP and Complete protease inhibitor (Roche). The lysates were resolved in 10-15% SDS-PAGE followed by transfer to nitrocellulose membrane (Biorad). Antibody dilutions, incubation time and temperature conditions were according to the manufacturers' protocol. Blots were developed on autoradiograms using chemiluminescent HRP substrate (ECL, Millipore). For immunofluorescence assay, 5 Â 10 5 cells were seeded on sterile coverslips in 6-well tissue culture plates and infected with GFP expressing Rv/Ra. At defined time points, cells were fixed with 2% PFA and immunostaining was performed according to the manufacturers' protocol provided by respective antibody. Coverslips were mounted on slides with DAPI containing Vectashield mountant (H-1200, Vector Labs). Images were captured in Carl Zeiss LSM510 Meta confocal microscope.

Animal infection
Actively growing Rv cultures were used for aerosolically infecting mice to implant 200 CFU/lung of Balb/c mice (6-8 weeks old of either sex) inside Madison Aerosol Chamber (University of Wisconsin, Madison, WI). Infection load in lungs and spleen was determined at day 1 and 56 p.i to evaluate deposition and commencement of chronic infection of Mtb, respectively. One lung and part of spleen of infected mice were homogenized and plated on 7H11 plates supplemented with PANTA (BD Bioscience) to enumerate CFUs. Second lung and part of the spleen from infected and uninfected mice were used preparation of tissue lysates. In another experimental set up, Balb/c mice were infected with Rv as stated above. Post 15 days of infection, mice were either kept untreated or treated with Vehicle (5% DMSO in sterile PBS), INH (prepared in sterile water, 25 mg/kg/dose through oral gavage), KU55933 (prepared in DMSO and further diluted in PBS such that DMSO is 5% (v/v) in PBS before injecting in the peritoneal cavity; administered every third day at 10 mg/Kg/ dose [Batey et al., 2013] Cell cycle analysis and apoptosis assay RAW cells infected with GFP expressing Rv at MOI of 1 for indicated time points. Cells were fixed with 70% ethanol for overnight at 4˚C followed by PBS washes and staining with 10 mg/ml propidium iodide (PI) solution (PBS+0.05% Triton-X + 100 mg/ml RNase A). Cells were analyzed through flow cytometry (FACS Calibur, Becton Dickinson, USA) to assess the distribution of cells in different phases of the cell cycle. For the apoptosis assay, PF infected with Rv and treated with the indicated inhibitors were stained with FITC Annexin V-7AAD according to manufacturer's protocol (BioLegend). Following staining, cells were fixed with 2% PFA before assessing the cell percentage undergoing apoptosis through flow cytometry.

Culture filtrate protein (CFP) preparation
Mtb strains were cultured in modified sauton's medium (2.9 mM KH 2 PO 4 , 4.2 mM MgSO 4 , 10.4 mM citric acid, 0.2 mM Ferric ammonium citrate, 6% Glycerol, 30.5 mM asparagine, pH 7.2) till mid-log phase. Culture supernatants were passed through 0.22 mM filter and concentrated to 1/100 th of the original volume using Amicon concentrators with a 3 kDa cut off (Rosenkrands and Andersen, 2001). The proteins concentration in the obtained culture filtrates was quantitated.

Statistical analysis
GraphPad Prism eight software and MS office Excel 2010 was used to execute statistical calculations. Unpaired student's t test was used to determine the p-values. Significant p values of the data sets were considered significant *p<0.05, **p<0.005, ***p<0.0005 and not significant (ns) if p>0.05. Densitometric analysis of important western blots (Figure 7-figure supplement 2) was performed using Image J software (Schneider et al.).