Countervailing, time-dependent effects on host autophagy promote intracellular survival of Leishmania

Autophagy is essential for cell survival under stress and has also been implicated in host defense. Here, we investigated the interactions between Leishmania donovani, the main etiological agent of visceral leishmaniasis, and the autophagic machinery of human macrophages. Our results revealed that during early infection—and via activation of the Akt pathway—Leishmania actively inhibits the induction of autophagy. However, by 24 h, Leishmania switched from being an inhibitor to an overall inducer of autophagy. These findings of a dynamic, biphasic response were based on the accumulation of lipidated light chain 3 (LC3), an autophagosome marker, by Western blotting and confocal fluorescence microscopy. We also present evidence that Leishmania induces delayed host cell autophagy via a mechanism independent of reduced activity of the mechanistic target of rapamycin (mTOR). Notably, Leishmania actively inhibited mTOR-regulated autophagy even at later stages of infection, whereas there was a clear induction of autophagy via some other mechanism. In this context, we examined host inositol monophosphatase (IMPase), reduced levels of which have been implicated in mTOR-independent autophagy, and we found that IMPase activity is significantly decreased in infected cells. These findings indicate that Leishmania uses an alternative pathway to mTOR to induce autophagy in host macrophages. Finally, RNAi-mediated down-regulation of host autophagy protein 5 (ATG5) or autophagy protein 9A (ATG9A) decreased parasite loads, demonstrating that autophagy is essential for Leishmania survival. We conclude that Leishmania uses an alternative pathway to induce host autophagy while simultaneously inhibiting mTOR-regulated autophagy to fine-tune the timing and magnitude of this process and to optimize parasite survival.

The leishmaniases are a group of vector-borne infectious diseases that are primarily endemic to tropical and sub-tropical regions of the world (1). The populations affected tend to be from lower socioeconomic strata, and there are projected to be approximately 12 million people affected worldwide (2).
Beyond this, there are an estimated 350 million people across 88 countries living at risk for contracting leishmaniasis (3). A multitude of factors, including poor sanitary conditions, lack of vector control, rapid environmental changes, increased travel, and resistance to commonly used drugs, have contributed to rising incidence rates.
Depending on the Leishmania species, the severity and form of leishmaniasis range from the relatively limited cutaneous leishmaniasis to a progressive, lethal form of visceral leishmaniasis that involves the liver, spleen, and bone marrow. Cutaneous leishmaniasis can be characterized by superficial lesions and ulcers that cause moderate to severe disfigurement. Visceral leishmaniasis, in contrast, results in internal organ damage that can be fatal when left untreated. It has been estimated that the incidence rate of visceral leishmaniasis is in the range of 200,000 -400,000 cases per year (1). The main etiological agent for human visceral leishmaniasis is Leishmania donovani.
Among various phagocytic cells, Leishmania primarily target macrophages. Like all Leishmania species, L. donovani has a digenetic life cycle, transitioning from the motile promastigote form within the sandfly gut to the non-motile amastigote form inside macrophages. Both life cycle stages have evolved to use multiple strategies to resist host microbicidal functions and to evade the immune system (4). For example, we showed that Leishmania infection of both murine and human macrophages hijacks the PI3K 3 /Akt pathway (5), leading to the inactivation of glycogen synthase kinase-3␤ (GSK-3␤) and the induction of IL-10 production, via enhanced activity of the transcription factor cAMP-response element-binding protein (5).
In this study, we sought to characterize other macrophage functional programs that might be affected downstream of the PI3K/Akt pathway in infected cells. One candidate of particular interest because of its pleiotropic regulatory properties is the mammalian target of rapamycin (mTOR), which is positively regulated by Akt. mTOR is a conspicuous kinase that functions as a master regulator of numerous cellular processes, including autophagy (6,7). It was this context that prompted us to ask whether Leishmania infection modulates host cell autophagy via an mTOR-dependent pathway and, importantly, how this impacts intracellular survival.
Autophagy encompasses a spectrum of conserved, catabolic processes in which cellular debris is removed and degraded. The most commonly addressed form is macroautophagy, herein referred to as autophagy. It is characterized by the active degradation of cytoplasmic constituents that are engulfed by double-membrane structures, known as autophagosomes. These distinctive structures ultimately fuse with lysosomes to form autophagolysosomes. It is at this stage that the intravesicular contents are degraded (8). More than 30 autophagyrelated proteins (ATGs) have been identified. Among these, the lipid-conjugated protein marker, microtubule-associated protein 1 light chain 3b (LC3-II)/ATG8, associates with autophagosomes and can be detected using various techniques. In fact, LC3-II has been used extensively as an indicator of autophagy in a wide variety of cells and tissues (9).
Autophagy can be regulated via multiple signaling pathways. Broadly, the two commonly defined pathways are either mTOR-dependent or mTOR-independent. As mentioned previously, PI3K/Akt activates mTOR leading to inhibition of cellular autophagy, and this is considered to be the classical pathway for regulation. In addition to this pathway, mTORindependent regulation of autophagy has also been recently studied (8). For example, inositol-lowering agents, such as lithium, induce autophagy independent of any change in mTOR activity (10).
Autophagy has long been considered to be a major recycling mechanism used by the cell. However, recent research has found that it has other functions, including roles in innate immunity and antimicrobial defense. Notably, autophagy in macrophages attenuates survival of numerous pathogens such as Mycobacterium tuberculosis, Shigella flexneri, Listeria monocytogenes, and Toxoplasma gondii (11).
Current knowledge around host autophagy and Leishmania pathogenesis is a focus of interest. One early study suggested that the transfer of dextran, from macrophage cytosol to Leishmania mexicana phagosomes, occurred via autophagy (12). Another study reported the accumulation of LC3-II in human bone marrow cells during L. donovani infection (13). Furthermore, induction of autophagy in infected macrophages has been linked to increased growth and parasite load of Leishmania amazonensis (14,15). Most recently, it has been found that Leishmania major uses macrophage autophagy to inhibit T-cell responses and prevent parasite clearance (16). The molecular mechanism(s) involved in Leishmania-mediated regulation of host autophagy have begun to emerge. For example, recent evidence suggests that select host microRNAs (miRNAs) may participate in regulation of host autophagy in response to Leishmania infection (17,18). It has also been shown that autophagy induction through endosomal Toll-like receptors plays a role in macrophages conferring resistance against L. major infections (19).
In this study, we report that Leishmania actively inhibits the induction of host classical autophagy via the early and sustained activation of mTOR. In striking contrast, as infection progressed, significant induction of autophagy was seen at later stages of infection. In fact, we present evidence that Leishmania engages an alternative mTOR-independent signaling pathway to induce host autophagy while at the same time maintaining tonic control of this process via activated mTOR. Our findings suggest a model in which Leishmania infection brings about countervailing, time-dependent effects on host autophagy through distinct pathways thereby promoting intracellular survival.

Infection by L. donovani initially inhibits and then induces host macrophage autophagy in a biphasic manner
Autophagosomes are formed from a cup-shaped doublemembrane sac called the isolation membrane in the cytoplasm. During maturation of autophagosomes, specific insertion of lipidated LC3-I (also known as LC3-II) occurs, which remains associated with autophagosomes. Thus, accumulation of LC3-II has been extensively used as a marker of autophagy, and increased levels of LC3-II may be detected by Western blot assay or by immunofluorescence microscopy as punctate vesicular LC3-II (20). We used both readouts to investigate the induction of autophagy in response to L. donovani infection in PMA-differentiated THP-1 cells that have been extensively used as a model system to study human visceral leishmaniasis (21).
To examine whether L. donovani induces autophagy, THP-1 cells were infected with stationary phase Leishmania promastigotes, and autophagy was monitored by Western blottings using LC3-II-specific antibodies over 48 h. As shown in Fig. 1, expression of LC3-II early after infection (2-8 h) was not affected relative to control cells. However, over the course of longer infection (24 -36 h), there was a clear increase in LC3-II expression in infected cells.
These increased levels of LC3-II could have been due either to enhanced de novo synthesis (autophagy induction) or to inhibition of turnover of LC3-II, due to a potential infection-induced blockade of the autophagic flux. To address this question, infected cells were treated or not with the lysosomal degradation inhibitors pepstatin A and E64d (9) and were assessed by Western blotting for LC3-II levels. If autophagic flux was normal in infected cells, then their levels of LC3-II should have been increased in the presence of lysosomal degradation inhibitors because the transit of LC3-II through the autophagic pathway would have been blocked (22). Indeed, the results presented in Fig. S1 show that each alone-infection or inhibitor treatment-were equipotent, and together they were additive in boosting LC3-II levels. These results, assuming near-complete reduction in flux by inhibitor treatment, support the conclusion that infection promotes the induction of autophagy per se with the attendant increases in de novo synthesis of LC3-II. To further strengthen these results, we also monitored levels of Sequestosome-1 (SQSTM-1/p62), an alternative autophagy marker, during Leishmania infection. p62 is a ubiquitin receptor protein that can bind to specific cargo and sequester it in autophagosomes. The degradation of p62, through the autophagy pathway, has been extensively used as an indicator of an increased autophagic flux. Conversely,

Leishmania exploits host cell autophagy
increased levels of p62 are indicative of inhibition of autophagy (23)(24)(25)(26)(27). As shown in Fig. S2, late Leishmania infection resulted in the depletion of p62 levels, compared with control, confirming an induction of autophagic flux. Lysosome-mediated autophagic degradation was confirmed by treating control and infected cells with pepstatin A and E64d, as shown in Fig. S2. Treatment with pepstatin A and E64d, in Leishmania-infected cells, resulted in the restoration of p62 levels, as expected. Taken together, the LC3-II and p62 results show that Leishmania induces host autophagy at a late stage of infection. We also investigated the level of p62 at the early stage of infection ( Fig.  S3, lanes 1 and 2). The results showed the accumulation of p62, indicative of a blockade in autophagic flux.

L. donovani actively inhibits rapamycin-induced host autophagy during early infection
As shown in Fig. 1, Leishmania did not induce host autophagy at early time points post-infection (2-8 h). This could have been passive, due to "silent entry" of Leishmania into host cells or to Leishmania actively attenuating host autophagy during early infection. We favored the second possibility because it is known that Leishmania activates the host PI3K pathway, and activation of PI3K is inhibitory for autophagy through its positive effects on mTOR (28). Moreover, the accumulation of p62 during the early stage of Leishmania infection was indicative of active autophagic inhibition (Fig. S3, lanes 1 and 2). To test the possibility that that Leishmania actively inhibited host autophagy in the early stages of infection, dTHP-1 cells were infected with Leishmania for 6 h, followed by treatment with the well characterized autophagy inducer rapamycin for 2 h (29,30). As shown in Fig. 2, A and B, Leishmania infection nearly completely abrogated rapamycininduced autophagy at 6 h post-infection. dTHP-1 cells treated with chloroquine (12.5 M for 2 h) were used as positive controls to confirm LC3-II induction. These findings were strengthened by monitoring levels of p62 under the conditions above, as shown in Fig. S3, lanes 3 and 4. Notably, Leishmania did not affect accumulation of LC3-II in response to chloroquine (Fig. 2, A and B), which acts through a distinct pathway from that of rapamycin (30,31). These findings further indicate that the inhibitory effect of Leishmania infection on autophagy is selective.
To confirm inhibition of rapamycin-induced autophagy by Leishmania using an orthogonal readout, we examined autophagosome formation using confocal microscopy and LC3-II-specific immunostaining of LC3-II puncta within the dTHP-1 cells. For this assay, we also used carboxyfluorescein succinimidyl ester (CFSE)-positive Leishmania for direct detection of organisms inside macrophages. Cells were infected with L. donovani for 6 h and then treated with rapamycin for 2 h followed by intracellular LC3-II staining. Cells treated with rapamycin alone for 2 h were used as positive controls. Noninfected cells at the 6-h time point were used to determine the baseline number of autophagosomes. The images taken were analyzed for LC3-II-positive puncta, and the mean immunofluorescence index was determined for 100 cells over three independent experiments. Representative images and quantification of puncta are shown, respectively, in

Leishmania exploits host cell autophagy
B, during the first 6 h infection with Leishmania was associated with inhibition of rapamycin-induced autophagosome formation ( Fig. 2, C and D). Taken together, these results clearly show that host autophagy is inhibited during early stage infection with Leishmania. Inhibition of host autophagy during early infection (Fig. 2) and induction of host autophagy as infection progresses ( Fig. 1) show dynamic biphasic regulation of host autophagy in response to Leishmania.
It was of interest to investigate the role of the regulatory pathway(s) that Leishmania targets to inhibit host autophagy. As it is well established that Leishmania activates host Akt, it was reasonable to link the involvement of this pathway in the active inhibition of host autophagy (5,28). To investigate this possibility, cells were first treated with Akt1/2 inhibitor for 4 h; subsequently, the cells were infected with Leishmania for 6 h. Subsequently, the cells were treated with rapamycin for 2 h. At the end of the experiment, the level of LC3-II was determined by Western blot assay. The results in Fig. 3 show that prior treatment with the Akt inhibitor prevented Leishmania's ability to inhibit rapamycin-induced autophagy. These findings directly link the utilization of the host Akt pathway to inhibit autophagy by Leishmania. It should be pointed out that the h. Subsequently, the cells were incubated either chloroquine (12.5 M) or rapamycin (12.5 g/ml) for 2 h. Whole-cell lysates were then collected and analyzed by immunoblotting for LC3-II. The same membrane was stripped and reprobed for actin as a loading control. A, Western blot for LC3-II and actin levels. B, densitometry analysis of three independent experiments. Data are presented as mean Ϯ S.D.; ns, not significant; *, p Ͻ 0.05. C, immunofluorescence and mean fluorescence index of LC3ϩ puncta, during an early infection. dTHP-1 cells were incubated with CFSE-positive L. donovani promastigotes (green) for 6 h. The cells were then washed and given new media containing rapamycin (12.5 g/ml) for 2 h. Uninfected control and infected cells with and without rapamycin were then fixed and stained for LC3-II (red) and the nuclei (blue). Confocal microscopy was performed to acquire images using LSM 790 Zeiss microscope, and immunofluorescence analysis was done using ImageJ software. The zoomed confocal photomicrographs are shown to improve clarity of LC3-II decorated autophagosomes. D, quantification of puncta from 100 dTHP-1 cells from each group over three independent experiments is shown as a histogram. Data are presented as mean Ϯ S.D.; ns, not significant; *, p Ͻ 0.05.

Leishmania exploits host cell autophagy
concentration of Akt inhibitor used to pre-treat cells did not affect the internalization of Leishmania (data not shown).

L. donovani induces host autophagy at later stages of infection despite actively attenuating mTOR-dependent autophagy
As shown in Fig. 1, Leishmania induced autophagy at later stages of infection, and it was of interest to ask whether this involved the use of the classical mTOR pathway or possibly an alternative mTOR-independent mechanism. We investigated whether delayed induction of autophagy by Leishmania correlated with down-regulation of mTOR, an important negative regulator of classical autophagy. Two direct downstream sequelae of mTOR activation include the phosphorylation of S6 kinase at Thr-389 and the phosphorylation of the eIF4E inhibitor, 4E-BP1, at multiple sites (32). As shown in Fig. 4, infection of dTHP-1 cells did not affect either the abundance of mTOR or lead to any changes in the phosphorylation states of either host 4E-BP1 or S6 kinase. In contrast, rapamycin when used as a positive control did down-regulate mTOR activity as expected (Fig. 4). These results indicate that neither degradation of mTOR nor inhibition of host mTOR activity was likely to be the mechanism by which Leishmania induced delayed host cell autophagy.
To examine this question further, cells infected for 24 h were then treated with rapamycin and analyzed for LC3-II levels.
Interestingly, the levels of LC3-II in infected cells, plus/minus rapamycin, were equivalent (Fig. 5, A and B), and these results were confirmed by confocal microscopy (Fig. 5, C and D). The finding that infection and treatment with rapamycin were not additive suggests, as one possibility, that the induction of autophagy in response to Leishmania may have reached its maximal potential. Alternatively, the lack of summation raised the possibility that Leishmania is still able to inhibit classical mTOR-regulated autophagy through 24 h of infection, as it had through 6 h of infection, by activating host Akt. To examine this possibility, we measured levels of phospho-Akt at 24 h postinfection and found these to be strikingly high (Fig. 6A). These results, along with the findings reported in Fig. 2 and Fig. S3, strongly suggest that Leishmania actively inhibited classical mTOR-dependent autophagy at both early and late stages of infection. To examine this directly, cells were infected for 24 h and then treated with the Akt1/2 inhibitor (Sigma). If Leishmania actively inhibited classical autophagy by activating the Akt-mTOR pathway, then inhibition of Akt should prevent mTOR activation, thereby removing this brake on induction of host autophagy. This prediction is shown to be correct in Fig.  6B, where it is also shown that Leishmania-induced autophagy was significantly enhanced in the presence of the Akt inhibitor. These findings provide support for a model in which Leishmania uses the PI3K-Akt-mTOR pathway to down-regulate classical autophagy, while at the same time using an alternative mTOR-independent pathway to induce autophagy in response to infection.

Host IMPase activity is reduced in response to Leishmania infection
Apart from the regulation of autophagy by mTOR, various mTOR-independent autophagy pathways have been described that are sensitive to chemical perturbations (33). One of the

Leishmania exploits host cell autophagy
first of these to be reported is linked to an inositol-signaling pathway where elevated levels of IP 3 inhibit the generation of autophagosomes and negatively regulate autophagy. Conversely, inositol-lowering agents, such as mood-stabilizing drugs like lithium induce autophagy without inhibiting mTOR activity (10). Our interest in this mTOR-independent pathway arose from our previous finding of reduced IP 3 levels in cells infected with L. donovani (34). Thus, we hypothesized that Leishmania might engage an mTOR-independent pathway by reducing the concentrations of IP 3 leading to induction of delayed autophagy. IMPase is the key enzyme required to gen-erate free inositol that is essential for the inositol-signaling pathway to function (35). Therefore, we measured the enzymatic activity of IMPase in Leishmania-infected cells and found that it was significantly reduced (Fig. 7). These findings identify one potential mTOR-independent pathway that may be used for the induction of delayed autophagy by Leishmania.

Dynamic regulation of host cell autophagy by Leishmania and impact on survival
To address the biological relevance of bi-directional regulation of host cell autophagy by Leishmania, we investigated

Leishmania exploits host cell autophagy
whether attenuation of autophagy during early infection (Ͻ12 h) is beneficial to Leishmania. Here, we used various concentrations of rapamycin for 2 h to induce autophagy in dTHP-1 cells prior to infection. After 2 h of rapamycin treatment, cells were infected with L. donovani for 24 h. At the end of the experiment, infected cells were extensively washed, and internalized parasites were released from the infected cells by mild treatment with SDS as described under "Experimental procedures." The effect of SDS was neutralized by adding Leishmania growth media, and freed amastigotes were allowed to transform in motile promastigotes and counted. The parasite rescue results presented in Fig. 8 clearly show that pre-treatment of host cells with rapamycin to induce autophagy prior to infection was inhibitory to survival of parasites in a concentrationdependent manner. These findings suggest that early inhibition of host autophagy by Leishmania is beneficial to promastigotes as they are not yet fully equipped to survive inside the hostile environment of phagolysosomes. It should be pointed out that rapamycin concentrations up to 25 g/ml did not affect internalization of promastigotes and were not toxic to promastigotes in culture (data not shown).
Next, we investigated the possibility that once infection is established, the induction of autophagy by Leishmania may confer a parasite survival advantage. To address this, we elected to study two autophagy-related proteins. One protein is part of the two-conjugation system, autophagy protein 5 (ATG5), although the other autophagy protein 9A (ATG9A) is not. ATG5 is an important protein for autophagic activity and is essential for autophagosome formation. It is required for LC3-I conjugation to phosphatidylethanolamine to form LC3-II and for the elongation of autophagic membranes (36). ATG9A is important for adding membrane to the autophagosome during its formation (37). We used siRNAs to down-regulate ATG5 or ATG9A. Treatment of cells with specific siRNAs prior to infec-

Leishmania exploits host cell autophagy
tion resulted in significantly decreased ATG5 or ATG9A levels (Fig. 9, A and C). Cells transfected with scrambled siRNAs were used as negative controls. Cells that were down-regulated for either ATG5 or ATG9A and control cells were then infected with L. donovani for 24 h. At the end of the experiment, the infected cells were lysed for parasite rescue. The results presented in Fig. 9, B and D, show that reduced ATG5 or ATG9A levels, respectively, correlated with decreased survival of Leish-mania. Taken together, these results indicate that delayed induction of host cell autophagy by Leishmania is critical for optimal intracellular survival.

Discussion
Diverse intracellular microbes, including Leishmania (38 -41), Yersinia (42), Coxiella (43), Mycobacterium (44) and others, have evolved mechanisms to modulate numerous macrophage functions to facilitate their survival within host cells. One of these mechanisms, used by Leishmania, is to create an immunosuppressive environment by promoting the production of the cytokine IL-10 (45). In a recent study, we examined the molecular mechanism used by Leishmania to induce host IL-10 production and defined a key role played by the PI3K/Akt pathway (5). Given that this pathway also regulates mTOR-dependent autophagy (46,47) and the importance of the latter to host defense (48), in this study we investigated what impact activation of the host PI3K/Akt pathway by L. donovani might have on autophagy.
Autophagy is a conserved, dynamic process in which intracellular components are packaged within autophagosomes that ultimately fuse with lysosomes leading to degradation of vesicular cargo. Classical regulation of autophagy involves mTOR, which functions as a negative regulator (8). Various growth factors can modulate mTOR activity, and a major signal cascade for its regulation is via the PI3K/Akt pathway (49).
To study Leishmania-induced autophagy further, in this study, we established a robust model of autophagy induction in response to L. donovani infection (Fig. 1). This involved delayed kinetics with induction of macrophage autophagy requiring at least 24 h to appear. Strikingly, we found that the apparent unresponsiveness of host cells at earlier time points (6 h) was not simply a passive encounter, but rather it was due to active inhibition of autophagy by Leishmania. This conclusion is based on the finding that during early stages of infection Leishmania was able to completely abrogate rapamycin-induced autophagy ( Fig. 2 and Fig. S3). These results suggested that as a survival strategy, Leishmania may actively inhibit early stage autophagy possibly by using the PI3K/Akt pathway to activate mTOR. In fact, further investigation showed the clear involvement of this pathway in the inhibition of autophagy at early stages of Leishmania infection. This is based on the finding that pre-treatment of host cells with the Akt1/2 inhibitor impeded the ability of Leishmania to attenuate rapamycin-induced autophagy, at early stages of infection (Fig. 3). However, this does not preclude the engagement of other autophagy regulatory pathways. For example, recent findings have shown the role of miRNAs in autophagy regulation, particularly MIR-30A-3p, in Leishmania-infected macrophages (18). As a corollary, it was of interest to investigate whether early induction of host autophagy would be deleterious to Leishmania survival.

Leishmania exploits host cell autophagy
To examine this possibility, first autophagy was induced using rapamycin for 2 h, followed by L. donovani infection for 24 h. At the end of the experiment, the viability of internalized parasites was tested by transforming them into motile promastigotes (Fig. 8). This showed that early induction of autophagy was in fact deleterious to Leishmania and suggested that attenuation of autophagy at early stages of infection is a pathogen survival strategy. Although rapamycin has been used extensively in the literature as an inducer of autophagy by inhibiting mTOR, there is the potential for off-target effects that may have affected the survival of intracellular Leishmania. However, rapamycin did not directly affect the growth and replication of L. donovani promastigotes in culture (data not shown). This is not surprising as the L. donovani TOR homolog shares only 23% homology with human mTOR. The infectiveness of rapamycin on growth and proliferation of L. major promastigotes in culture has also been previously reported (50). To the best of our knowledge, the ability of L. donovani to actively inhibit classical mTORregulated autophagy at early stages of infection is a novel strategy that promotes pathogen survival (Figs. 2 and 3 and Fig. S3).
Nevertheless, recent reports and the present findings make it clear that Leishmania does activate autophagic machinery in host cells, albeit with delayed kinetics (Fig. 1), in the case of L. donovani (13,18). In respect to the latter, it was of interest to establish whether delayed autophagy involves mTOR-dependent cell signaling or an alternative signaling pathway that is mTOR-independent. To address the mechanism of late stage (24 h) autophagy induction by L. donovani, we examined whether this correlated with a decrease in mTOR activity based upon phosphorylation of S6 kinase and 4EBP-1 as surrogate markers (32,(51)(52)(53). Surprisingly, Leishmania infection did not lead to a decrease in mTOR activity (Fig. 4). Recently, proteolytic inactivation of mTOR in L. major-infected bone marrowderived macrophages by Leishmania metalloprotease glycoprotein 63 was reported (54). Therefore, we examined mTOR levels from L. donovani-infected cells by Western blottings. In contrast to the findings by Jaramillo et al. (54), L. donovaniinfected dTHP-1 cells showed no evidence of proteolytic inactivation of mTOR (Fig. 4, bottom panel). This result indicates that unlike the case for L. major and murine bone marrowderived macrophages, proteolytic degradation of mTOR was not likely to be involved in the induction of autophagy in dTHP-1 cells infected with L. donovani. Taken together, these results show that Leishmania does not use an mTOR-dependent pathway to induce late stage autophagy. In fact, Leishmania appears to inhibit mTOR-dependent autophagy at late stages of infection. This interesting result is supported by three findings. First, Leishmania markedly enhanced phosphorylation of host Akt at a late stage of infection, which would be expected to inhibit mTOR-dependent autophagy (Fig. 6A). Second, and consistent with the latter, Leishmania-induced autophagy was significantly increased in the presence of the Akt1/2 inhibitor (Fig. 6B). Third, cells infected with Leishmania for 24 h remained resistant to rapamycin-induced autophagy (Fig. 2). These findings so far suggest the possibility that Leishmania uses an alternative mTOR-independent pathway to induce delayed host autophagy.
In addition to regulation of autophagy by mTOR, various mTOR-independent pathways have been reported (8,55). These pathways are sensitive to chemical perturbations, including one that is sensitive to cellular concentrations of inositol (10). It has recently been shown that an inositol-signaling path- Figure 9. ATG5 or ATG9A knockdown in host macrophages results in reduced survival of Leishmania. THP-1 cells were transfected with indicated siRNAs for ATG5 or ATG9A for 48 h. Control and ATG knockdown cells were differentiated with PMA, and subsequently incubated with L. donovani promastigotes for 24 h. At the end of the experiment, infected cells were washed, and internalized parasites were released by lysing cells using a mild concentration of SDS, followed by transfer to the transformation medium as described under "Experimental procedures." On day 5, transformed motile promastigotes were counted (B for ATG5 and D for ATG9A). In parallel, cells treated with siRNAs were confirmed for low ATG5 protein levels and low LC3-II protein levels (A for ATG5 and C for ATG9A). Data are presented as mean Ϯ S.D.; *, p Ͻ 0.05; **, p Ͻ 0.01.

Leishmania exploits host cell autophagy
way negatively regulates mTOR-independent autophagy (56). Interestingly, inositol-lowering agents, such as lithium chloride, induce autophagy without affecting mTOR activity. We hypothesized that Leishmania may induce delayed autophagy by impeding the phosphoinositol-signaling pathway. Support for this hypothesis came from our previous finding of reduced IP 3 levels in cells infected with L. donovani (34). In this report, we selected IMPase for study as this key enzyme is involved in inositol recycling and de novo synthesis of inositides (8,56). We reasoned that if the activity of this enzyme is reduced in Leishmania-infected cells, then this would result in reduced cellular inositides and lead to induction of autophagy. In fact, this turned out to be the case. Leishmania significantly attenuated IMPase activity in infected cells (Fig. 7) supporting a model in which Leishmania induces autophagy at late stages of infection by modulating an mTOR-independent inositol-signaling pathway. However, this finding does not rule out the possibility that Leishmania also uses other mTOR-independent pathway(s) in addition to the inositol-signaling pathway to induce delayed host autophagy.
An important question that arises from these findings is whether host autophagy is involved in defense against Leishmania or whether it functions to promote parasite survival. Several reports focused on pathogens other than Leishmania have drawn opposing conclusions in this regard. For example, in Helicobacter pylori (57), L. monocytogenes (58), and M. tuberculosis infections (59, 60), autophagy appears to limit growth of these pathogens. In contrast, T. gondii (48), hepatitis C virus (61), and Coxiella burnetii (62) appear to take advantage of autophagy to support infection. In the context of Leishmania infection, it should be pointed out that the impact of host autophagy on parasite survival appears to depend upon the host and strain. For example, induction of autophagy correlated with increased parasite loads of L. amazonensis in BALB/c but not in C57BL/6 mice (14). A recent study using L. major-infected bone marrow-derived macrophages showed host autophagy was detrimental to pathogen survival (17). In contrast, it was recently reported that in a stationary phase, the L. major inoculum containing a significant number of apoptosis-like parasites activated human macrophage autophagy machinery, thereby dampening T-cell responses and promoting parasite survival (16). In this study, inhibition of autophagy by down-regulating either ATG5 or ATG9A (both are essential constituents of autophagosome formation) reduced the survival of Leishmania inside the infected macrophages. This provided unambiguous evidence to show that autophagy is beneficial for the survival of L. donovani in human macrophages at a late stage of infection. In fact, Leishmania-induced autophagy in host macrophages could be an important determinant of nutrient supply where survival is largely dependent on host resources. In support of this suggestion, acquisition of host cell macromolecules by L. mexicana involved an autophagy-like mechanism (12).
A highly novel finding of this study is that Leishmania regulates host autophagy in a biphasic, time-dependent manner. Active attenuation of host autophagy by Leishmania at early stages of infection appears to be beneficial to the survival of as yet not fully differentiated organisms (Fig. 8). In contrast, the delayed kinetics of a robust autophagic signal, which occurred around 24 h post-infection, suggest the possibility that this response may have been elicited by increased nutrient demand. This was not supported by a decrease in mTOR activity, which has been shown in many studies to be the mechanism of autophagic stimulation by nutrient deprivation. An alternative possibility to consider is that delayed onset, Leishmania-induced autophagy may result from an active event brought about by infection. This could be linked to active inhibition of host IMPase resulting in up-regulation of mTOR-independent autophagy.
Autophagy was initially characterized as a non-selective bulk degradation pathway induced by nutrient deprivation (8,63). However, it is now becoming clear that the recycling of damaged organelles, removal of dysfunctional protein aggregates, and elimination of intracellular pathogens are highly selective processes that require cargo recognition by the specialized autophagy machinery (64,65). Therefore, it is reasonable to propose that Leishmania prefers cargo-selective autophagy, induced by inhibition of host IMPase activity for its optimal intercellular growth and survival. Support for this hypothesis will require characterization of autophagosomes and their cargo induced in response to Leishmania infection.
In summary, our results indicate that Leishmania uses dual strategies to exert countervailing effects on host autophagy. This presumably enables fine-tuning of autophagy to optimally promote parasite survival. On the one hand, Leishmania inhibits mTOR-dependent autophagy at both early and late stages of infection, most likely by sustained activation of host Akt and mTOR (Fig. 10). On the other hand, Leishmania infection induces mTOR-independent autophagy, at later stages of infection, likely by down-regulating host IMPase activity (Fig. 10). It is tempting to speculate that Leishmania might derive a number of advantages through this bidirectional, time-dependent regulation of host autophagy. First, inhibition of autophagy at early stage infection by activating host Akt would provide an opportunity for Leishmania promastigotes to begin to differentiate into amastigotes to adapt the harsh conditions of phagolysosomes. Second, sustained activation of host Akt would contribute to inhibition of apoptosis, which is a potent pathogen clearance mechanism. In this context, it is known that Leishmania confers host cell resistance to apoptosis by activating Akt (28,66). Third, activation of mTOR-independent autophagy, in the background of sustained activation of the PI3K-Akt pathway, would allow for fine-tuning of autophagy for optimal acquisition of essential nutrients from the host.
The molecular triggers used by Leishmania to either resist (early infection) or promote (later infection) autophagy are not completely known. Recently, we and others have shown that Leishmania contains specialized secretory vesicles, termed exosomes, that release multiple parasite-encoded proteins into the host cell during infection (67,68). It is possible that one or more of these exported proteins regulates host cell autophagy. Currently, work is in progress to investigate whether Leishmania exosomes have the potential to regulate host autophagy. In fact, our initial findings show that incubation with Leishmania Leishmania exploits host cell autophagy exosomes activates macrophage Akt. 4 This should be a fruitful area for ongoing studies. In addition, our host-parasite system presents a novel opportunity for the coordinated investigation of apoptosis, autophagic signals, and Leishmania pathogenesis.

THP-1 cell culture
THP-1 cells, obtained from ATCC (TIB-202TM), were incubated and cultured at 37°C, 5% CO 2 in RPMI 1640 media (HyClone) containing 10% heat-inactivated fetal calf serum (Gibco), 10 mM HEPES (Sigma), 100 units/ml penicillin/streptomycin (Sigma) and 2 mM L-glutamine (Gibco). Cells in suspension were passaged every 2-3 days to maintain a density between 3 ϫ 10 5 and 8 ϫ 10 5 cells/ml. For differentiation, THP-1 cells were treated with 10 ng/ml PMA for 16 -18 h. After adherence, cells were washed three times with Hanks' balanced salt solution (HBSS) (Sigma) and given fresh media not containing PMA. The cells were rested for 6 h before being used experimentally. For infections, day 5 stationary promastigotes were used at an m.o.i. of 20:1. Figure 10. Dynamic regulation of macrophage autophagy in response to Leishmania. Shown is a model illustrating Leishmania involving PI3K/Akt/mTORsignaling pathway and IMPase of phosphoinositol-signaling pathway to regulate host autophagy in infected macrophages. Continuous Leishmania-mediated activation of macrophage Akt sustains mTOR activity. This in turn suppresses autophagy at early and late stages of infection. At later stages of infection, Leishmania engages IMPase to induce host autophagy, which is independent of mTOR activity, to promote their survival. Potential mechanism for Leishmaniamediated host autophagy program involving Leishmania secretory factors is shown.
Whole-cell lysates were separated by SDS-PAGE and transferred to appropriate transfer membranes (Bio-Rad). For LC3-II and phospho-4EBP1, whole-cell lysates were subjected to Tris/Tricine, 15% SDS-PAGE and transferred onto polyvinylidene difluoride membrane. For SQSTM-1/p62, phospho-p70-S6 kinase, ATG5, and ATG9A, whole-cell lysates were subjected to Tris-glycine, 10% SDS-PAGE and transferred to nitrocellulose membrane. For mTOR, whole-cell lysates were subjected 4 -20% gradient SDS-PAGE and transferred to nitrocellulose membrane. Transferred proteins were probed with appropriated antibodies, according to the manufacturer's instructions. Protein bands were either observed on Blue X-ray film (Carestream) using ECL Select TM Western blotting detection reagent from GE Healthcare (RPN2235) for enhanced chemiluminescence or using Odyssey CLx Imaging System (LI-COR Biosciences) for infrared fluorescence.

Determination of intracellular parasite burden
For parasite infection rate and burden, infected dTHP-1 cells were briefly fixed using ice-cold 2% paraformaldehyde in phosphate-buffered saline (PBS),for 15 min and protected from light. The fixed cells were washed twice with PBS and placed onto Prolong TM Diamond Antifade Mountant with DAPI (Life Technologies, Inc.). DAPI was used to stain macrophage and parasite nuclei to determine the number of infected macrophages. Cell images were acquired at ϫ40 using Zeiss Axioplan 2 imaging microscope. At least 100 cells were counted for each condition to determine the average number of parasites per macrophage and percent macrophages infected.

Confocal microscopy
For the infection, stationary phase (day 5) L. donovani promastigotes were spun down, washed twice with PBS, and resuspended in 1 ml of PBS with 6 l of 1 M CellTrace TM CFSE. After a 30-min incubation at 37°C, parasites were spun down and resuspended in 10% fetal calf serum in PBS. Promastigotes were then spun down and resuspended into PBS to infect dTHP-1 at an m.o.i. of 20:1.
For the fixation, cells were washed once with HBSS and twice with PBS. Then the cells were fixed with ice-cold 2% paraformaldehyde in PBS for 15 min, protected from light. The cells were then washed twice with PBS. The cells, with the respective treatments, were stained for LC3-II and DAPI. The cells were blocked with 5% NGS, 0.3% Triton X-100 in PBS for 1 h. The primary antibody for LC3-II was used at a 1:400 dilution in 1% BSA, 0.3% Triton X-100 in PBS, overnight at 4°C. The secondary antibody, Alexa Fluor 594 goat anti-rabbit IgG (HϩL), was used at a 1:250 dilution in 1% BSA, 0.3% Triton X-100 in PBS for 2 h. The cells were placed onto Prolong TM diamond antifade mountant with DAPI. Cells were imaged using Zeiss LSM 780 confocal microscope under ϫ63 magnification and Zen soft-ware. For the analysis of LC3-II puncta, ImageJ macro plugin software was used as described previously (69). The analysis was done on 100 cells over three independent experiments.

Parasite rescue and transformation assay
For this assay, dTHP-1 cells were infected with L. donovani promastigotes at an m.o.i. of 20:1. After the desired period of infection, cells were extensively washed with HBSS to remove non-internalized parasites. Controlled lysing of infected cells was performed using 0.01% SDS as described previously (70). Quantification of the infection was performed through transformation of live, rescued Leishmania amastigotes to log phase promastigotes in M199 media by incubating the plates in 26°C for 48 h. The evaluation of their growth was performed by manual counting of transformed promastigotes using trypan blue solution (0.4% w/v in PBS) and a hemocytometer. Counts were taken from each group in triplicates.

IMPase assay
IMPase activity was assayed by measuring the conversion of inositol 1-phosphate into inositol and inorganic phosphate. The release of inorganic phosphate was then measured using malachite green (71). Control and Leishmania-infected cells were extensively washed with warm HBSS to remove serum and non-internalized parasites. Then, cells were washed three times with hypotonic buffer (20 mM Tris-HCl, pH 7.8), and the plate was placed on ice. Cells were then dislodged and disrupted in ice-cold extraction buffer (50 mM Tris-HCl, pH 7.8, 250 mM KCl, 3 mM MgCl 2 supplemented with aprotinin, leupeptin, and PMSF) by passing several times through a 22-gauge needle. The resulting cell extracts were left on ice for 10 min and then clarified at 10,000 ϫ g for 10 min at 4°C. Equal amounts of proteins from control and infected cells were assayed for phosphatase activity using 0.4 mM inositol 1-phosphate (Sigma) as the substrate, at 37°C for 30 min, and the reaction was stopped with malachite green reagent. Inorganic phosphate present in each well was calculated by reading the OD 620 against a standard curve. Enzyme activity was then calculated by subtracting the inorganic phosphate formed in wells with cell extract and inositol 1-phosphate from inorganic phosphate formed in corresponding wells with cell extract not containing inositol 1-phosphate.

siRNA knockdown
The oligoribonucleotides targeting the cDNA sequence of human ATG5 and human ATG9A, as well as non-specific control siRNA, were obtained from OriGene (catalog no. SR306286 for ATG5 and SR312320 for ATG9A). THP-1 cells in 24-well dishes were transfected with non-specific or ATG5 siRNA or ATG9A siRNA (50 pmol/well) using HiPerFect transfection reagent (Qiagen) according to the manufacturer's instruction. After 48 h of transfection, the cells were differentiated with 10 ng/ml PMA and infected with L. donovani (m.o.i. 20:1) for 24 h before preparation of whole-cell lysates or of controlled lysis for parasite rescue.

Statistical analysis
The data of three independent experiments were determined using a paired t test on GraphPad Prism 6.0 software. The val-Leishmania exploits host cell autophagy ues were considered statistically signification at *, p Ͻ 0.05; **, p Ͻ 0.01; ****, p Ͻ 0.0001.
Author contributions-S. A. T., D. N., N. E. R. designed the research. S. A. T., D. N., and J. K. performed the experiments and analyzed data. S. A. T., D. N., and N. E. R. prepared and wrote the manuscript.