Leishmania donovani Induces Autophagy in Human Blood–Derived Neutrophils Free

The study of the biology of neutrophils has attracted significant attention in recent times because of the realization of the complex nature of these cells and their capability to carry out a large range of specialized functions. Forming the highest percentage of granulocytes, also called polymorphonuclear neutrophils (PMN) with the capability of phagocytosis, they are the first cell population to be massively recruited to the infection site (1). Evidently, PMNs are the first line of defense against pathogens, and the success or failure of an infection depends largely on the proper functionality of these cells employing several antimicrobial mechanisms like NETosis, inflammation, oxidative burst, phagocytosis, and antimicrobial peptide secretion (2). Although involvement of neutrophils in bacterial infections is relatively well understood, their participation in infections with trypanosomatid parasites like the Leishmania has not been explored well. Very few reports are available about neutrophil–Leishmania interactions (3, 4). Apart from the defense mechanisms of phagocytic cells, other cellular processes like autophagy and apoptosis (5) play important roles in the transmission of pathogens and sustenance of infection. Autophagy eliminates intracellular pathogens by direct killing or silencing their replication cycle by modulating transcriptional and translational machinery of the host cell (6). Coevolving pathogens also succeed in adapting strategies to impair or counteract deleterious effects directly by avoiding or inhibiting autophagy (5). Whether pathogens use the process of host cell autophagy for successful transmission is an area that needs to be explored in much detail. Despite increasing evidence on the role of autophagy in eliminating Leishmania parasites in macrophages (7), there are still insufficient shreds of evidence that explain the behavior of PMNs during transmission of the parasite to the regular host cells.

Visceral leishmaniasis, caused by the parasite L. donovani, inflicts significant morbidity and mortality in humans and, in the absence of a vaccine and suitable drugs, remains a problem in the tropics. About 50,000 to 90,000 new cases of visceral leishmaniasis occur annually (8). A parasite with a digenetic life cycle living between an insect vector and the mammalian host, the Leishmania parasite divides and grows within the mature phagolysosomal compartment of the macrophage (9). This capability to endure and survive the defense arsenal of the host cell within the phagolysosome provides the parasites with opportunities to use the phagocytic cells for transmission and propagation. Prior knowledge describes how the Leishmania parasites use PMNs as vehicles for transfer to macrophages through the Trojan Horse mechanism (10). Although the Trojan Horse mode of transmission is recognized, the associative processes involved in regulating the actual transfer are not clear. There are several reports about the unusual behavior of PMNs in mice during Leishmania infections (11); however, intrinsic differences in the biology of human and mouse neutrophils (12) restrict the extrapolation of the observations into the human system. Therefore, the absence of a proper animal model and the lack of human neutrophil cell lines makes research in this area more challenging and interesting from the point of view of revealing novel events during host–pathogen interactions.

The observation of autophagy increase in human PMNs (hPMNs) after Leishmania infection prompted us to explore the trigger for autophagy and its influence in the resultant uptake of the PMNs by the macrophages. Accordingly, we used modulation of autophagy to explore how macrophage uptake of neutrophils was affected. Both canonical and noncanonical autophagy was triggered upon infection in which noncanonical autophagy was followed by the occurrence of canonical autophagy. The trigger for autophagy was parasite interaction with the hPMNs through Leishmania surface lipophosphoglycan (LPG) inducing ERK and protein kinase B (Akt) phosphorylation, leading to NADPH oxidase (NOX) activation accompanied by reactive oxygen species (ROS) generation. Lipidation of microtubule-associated proteins 1A/1B L chain 3 (LC3) and formation of autophagosomes was blocked when ERK, NOX, or ROS were inhibited. Inhibition of both types of autophagy or noncanonical autophagy alone resulted in reduced uptake of PMNs by the macrophages. This uptake of the PMNs was orchestrated by reduction and diffusion of CD47 or “do not eat me” signals with no change in phosphatidylserine (PS) or “eat me signal” expression. Interestingly, autophagy induction by rapamycin treatment of hPMNs elevated the uptake of infected PMNs by the macrophages. Therefore, we present a new finding of a role of neutrophil autophagy in the uptake of hPMNs by the macrophages that is also important for the uptake of Leishmania parasite–infected neutrophils as well.

All experiments with human material (blood) were performed in compliance with guidelines duly approved by the Institutional Human Ethics Committee of the National Institute of Immunology (New Delhi, India) (project no. IHEC 83/14).

Neutrophils and PBMCs were isolated from the blood of healthy human volunteers (n = 15) using Polymorphprep (Alere Technologies, Axis-Shield, Norway) per manufacturer’s protocol. Venous blood was separated into PBMCs, neutrophils, and erythrocytes on Polymorphprep based upon their differential densities. Upper layered PBMCs were used for macrophage culture, whereas middle layered PMNs were used for neutrophil-related experiments. Neutrophil purity assessment was performed using Giemsa staining (Riedel-de Haën, Seelze, Germany) and by flow cytometric analysis using Anti-Human Neutrophil Peptide-1–specific Ab (Santa Cruz Biotechnology, Santa Cruz, CA). Briefly, isolated neutrophils were stained with 10% Giemsa stain solution, mounted on a slide, and visualized under the microscope. The >95% pure hPMNs were checked for viability by trypan blue (Loba Chemie, Mumbai, India) membrane exclusion method. The >95% viable hPMNs and human monocyte–derived macrophage (hMDMs) were cultured in RPMI 1640 media (Sigma-Aldrich, St. Louis, MO) supplemented with 10% heat-inactivated FBS (Biological Industries, Kibbutz Beit Haemek, Israel) and penicillin (100 U/ml), streptomycin (100 U/ml) (Sigma-Aldrich,) and gentamycin (50 μg/ml) (Abbott Laboratories Argentina, Buenos Aires, Argentina) at 37°C in humidified air (95% relative humidity) containing 5% CO₂. Additionally, 300 ng/ml of recombinant human M-CSF (Shenandoah Biotechnology, Warwick, PA) and 5% human serum were used as supplements in hMDM culture periodically to complete the differentiation of monocytes into macrophages. Monocytes were allowed to differentiate for at least 7 d before initiating experiments with the macrophages (13).

For inhibition of the autophagic machinery, neutrophils were pretreated with an early-stage autophagy inhibitor 3-methyladenine (3-MA) (Sigma-Aldrich) (5 mM for 1 h) that interferes with autophagy by blocking autophagosome formation through the inhibition of class III PI3K (14). For late-stage autophagy inhibition, Bafilomycin A1 (Sigma-Aldrich) (500 nM for 2 h), capable of blocking maturation of autophagic vacuoles by inhibiting fusion between autophagosomes and lysosomes, was used (15). Rapamycin, inhibitor of mammalian target of rapamycin (mTOR), was used (500 nM) for 1 h to stimulate autophagy (16). The inhibition of signaling pathways was achieved treating hPMN cultures with U0126, a MAP kinase/ERK inhibitor (10 μM for 10 min) (Cell Signaling Technology [CST], Beverly, MA); LY294002, a selective PI3K inhibitor (50 μM for 10 min) (CST); SP600125, a JNK inhibitor (50 μM for 1 h); and SB203580, selective inhibitor p38/MAPK (Sigma-Aldrich) (1 μM for 30 min). For reducing the generation of ROS, diphenyleneiodonium chloride (DPI), a NOX inhibitor (Sigma-Aldrich) (10 μM for 1 h), was used. For inhibition of TLR2 and 4, TIRAP inhibitor peptide (TIRAPinh), a TLR2- and 4-specific inhibitor that blocks the interaction between TLR2/TLR4 TIR domain with their adaptor protein TIRAP/Mal (17) (Novus Biologicals, Littleton, CO) (40 μM for 1 h) and TAK242, TLR4-specific small molecule inhibitor that inhibits interaction between TLR4 and its adaptor molecules TIRAP and TRIF-related adaptor molecule (TRAM) (18) (Calbiochem, Darmstadt, Germany) (1 μM for 10 min) were added to cultures prior to infection for 5 h. MRT68921, Unc-51–like kinase (ULK1/2) inhibitor (19) (Cayman Chemical Company, MI); Vps34-IN1, vacuolar protein sorting 34 kinase activity inhibitor (Cayman Chemical Company) (10 μM for 1 h) (17); and LPS (25 μg/ml for 30 min) (Sigma-Aldrich) were also used for different experiments. All inhibitor treatments were carried out for specific time periods prior to infection. Reagents required for SDS-PAGE were purchased from Sigma-Aldrich.

L. donovani wild-type (strain AG83, BHU1260; India), L. major (MHOM/Su73/5ASKH), and L. donovani LD1S Sudan strain wild-type, lpg1-knockout (KO), complemented mutant of lpg1-KO (lpg1-KO+LPG1) strains [courtesy of Prof. A. Descoteaux (20)] were cultured in M199 media supplemented with 10% heat-inactivated FBS, 40 mM HEPES (pH 7.4), 100 μM hypoxanthine, 5 μM hemin, 3 μM biopterin, 1 μM biotin (Sigma-Aldrich), and penicillin streptomycin antibiotic solution maintained at 26°C. Wild-type parasites were also maintained on brain heart infusion agar (Laboratorios CONDA, Madrid, Spain) supplemented with normal rabbit blood provided by the animal facility of the National Institute of Immunology in New Delhi. Complemented mutants were grown in M199 media supplemented with 10% FBS and 80 μM of Zeocin (InvivoGen, San Diego, CA). Metacyclic promastigote-enriched stationary phase Leishmania culture was used in all infection studies (21). Parasites were used until the third passage after they exhibited appropriate infectivity in vitro measured through staining of the macrophages for the presence of the parasites. Macrophages and neutrophils were infected with Leishmania parasites with multiplicity of infection of 10 and 5, respectively. Engulfment was counted at 4 h after staining the cells with DAPI (10 μg/ml) (Thermo Fisher Scientific, Tokyo, Japan). Infection rate for neutrophils was calculated as: Total number of parasites infected hPMNs under given condition per 500 hPMNs.

To prepare the sample for indirect immunofluorescence assay, treated and vehicle-treated hPMNs were fixed with 3.7% paraformaldehyde for 30 min at 4°C. Cells were permeabilized with 0.25% Triton X-100 for 10 min (Sigma-Aldrich) for vesicular LC3-II, Beclin-1, and RUN domain protein as Beclin-1 interacting and cysteine-rich containing (Rubicon) staining. Following permeabilization, cells were incubated with primary Abs against LC3-II (1:100) (Molecular Probes, Eugene, OR), Rubicon (1:100) (Abcam, Cambridge, U.K.), Beclin-1 (1:100) (BD Biosciences, San Jose, CA), and CD47 (B6H12.2) (1:100) (Thermo Fisher Scientific) overnight at 4°C. For LC3-II punctae detection, cells were directly incubated with secondary Ab conjugated with Rhodamine Red-X fluorophore (1:300) (Molecular Probes). For Beclin-1 and Rubicon colocalization study and CD47 avidity analysis, cells were incubated with respective biotinylated secondary anti-rabbit and mouse Abs (1:200) (Jackson ImmunoResearch Laboratories, West Grove, PA), followed by incubation with tertiary Ab conjugated to streptavidin and fluorophore Alexa Fluor 488 and Alexa Fluor 594 (1:250) (Invitrogen, Carlsbad, CA). For colocalization analysis, Pearson coefficient was calculated using Leica Application Suite Advanced Fluorescence version 2.7.3.9723 software (Leica Microsystems, Mannheim, Germany) by minimizing basal noise. Pearson coefficient above 0.6 was considered a significant colocalization event.

All immunofluorescence images were taken with a confocal microscope (Leica TCS SP5 II; Leica Microsystems, Wetzlar, Germany) using HCX APO L U-V-I 63×/0.90 W (2.2 mm) in oil for all studies as described previously (13). In some analyses, z stacking was carried out for the slice with step size of 0.17 μm (for LC3-II, Beclin-1, and Rubicon staining) and 0.13 μm (for CD47 staining) with a pin hole (m) at 95.5 μm and pin hole (airy) at 1.00. All image analysis and brightness adjustments were carried out with Leica Application Suite Advanced Fluorescence version 2.7.3.9723 software.

For measuring protein expression, cells from the various experimental groups were lysed in Laemmli sample buffer, and total protein was estimated using CB-X Protein Assay Kit (G-Biosciences, Saint Louis, MO) as described previously (22). SDS-PAGE was done under denaturing conditions as described earlier (22). Subsequently, the resolved proteins on gels were transferred onto the 0.45 μm nitrocellulose membrane (MilliporeSigma, Darmstadt, Germany), and the blots were treated for reduction of nonspecific staining with 5% skimmed dry milk (Santa Cruz Biotechnology), followed by probing with specific Abs at suitable dilutions. Blots were then incubated with relevant secondary Abs at suitable dilutions and probed with Femto LUCENT PLUS-HRP–sensitive chemiluminescence immuno-detection system (G-Biosciences). Protein bands were visualized on x-ray films (Carestream Health, Rochester, NY). To verify equal loading, the same lysates were processed simultaneously on independent membranes and checked using anti–β-actin Ab. Relative density of blots was quantified by using Lab Works Image acquisition and analysis software version 4.0.0.8 (Lab Works, Analytik Jena, Upland, CA) as described earlier (22). Dilutions of Abs used for Western blot analysis were as follows: LC3-II polyclonal rabbit Ab (1:2000; Novus Biologicals and CST); phospho-autophagy–related protein 14 (pATG14) (S29) rabbit mAb (1:1000); ATG 14 rabbit mAb (D1A1N) (1:1000); phospho-p42/44 MAPK (T202/Y204) (pERK1/2) rabbit mAb (D13.14.4E) (1:1000); p42/44 MAPK (ERK1/2) rabbit mAb (137F5) (1:1000); phospho-Akt (S473) (pAkt) rabbit mAb (D9E) (1:1000); pan Akt polyclonal Ab (C67E7) (1:1000) rabbit mAb; phospho-5′ AMP-activated protein kinase (pAMPK-α) (T172) rabbit mAb (40H9) (1:1000); AMPK-α rabbit polyclonal Ab (23A3) (1:1000); phospho–Beclin-1 (pBeclin-1) (S93) rabbit mAb (D9A5G) (1:1000) (CST); p62 mouse Ab (1:1000) (BD Biosciences); Beclin-1 mouse Ab (1:1000) (BD Biosciences); β-actin mouse mAb (AC-15) (1:10,000) (Sigma-Aldrich); CD47 (1:500); and phospho-SAPK/JNK (pJNK) (T183/Y185) (Thermo Fisher Scientific). The secondary Abs were anti-mouse HRP-conjugated Ab (1:10,000) and anti-rabbit HRP-conjugated Ab (1:10,000) from Jackson ImmunoResearch Laboratories.

The hPMNs were pretreated with a cell-permeable ROS indicator H₂DCFDA (25 μM for 45 min, Invitrogen) seeded onto BMG black 96-well plate with cell density of 2 × 10⁶ cells/well. Cells were exposed to different treatments after the baseline ROS reading was recorded with a CLARIOstar fluorometer (BMG Labtech, Ortenberg, Germany) with excitation/emission wavelength at 530-10/590-10 nm as described earlier (23). Postinfection, ROS kinetics was measured for 7 h duration at 15 min intervals. PMA (25 ng/ml) and DPI plus PMA–treated cells were taken as positive and negative controls, respectively. Data were analyzed in MARS data analysis version 3.10 R6 software for change of ROS in PMNs during the course of infection.

After 7 h of parasite exposure, hPMNs under various experimental conditions were washed two to three times with chilled 1× PBS and centrifuged at 250 relative centrifugal force for 5 min to remove extracellular unbound parasites. Cells were then labeled with Alexa Fluor 488 Annexin V using Alexa Fluor 488 Annexin V/Dead Cell Apoptosis kit (Invitrogen) as described earlier (24). H₂O₂ (Merck Biosciences, Darmstadt, Germany)-treated cells were used as positive controls. Detection of a significant amount of PS by Annexin V labeling on the outer leaflet of cells identifies the cell as early apoptotic. After labeling, the reaction was stopped by adding 1× Annexin V–binding buffer, and cells were analyzed using the FL-1 channel in BD FACSCalibur (BD Biosciences). Logarithmic fluorescence intensity was plotted against forward scatter. Number of Annexin V–positive cells was measured using FlowJo version 8.7.10 software (FlowJo, Ashland, OR). Data are representative of fold change in Annexin V–positive PMNs with respect to unexposed and untreated Leishmania cells.

For engulfment assay, 2 × 10⁶ human blood–derived monocytes were functionally differentiated into matured hMDMs in cell culture–compatible 35-mm petri dishes as reported previously (13). The hPMNs were labeled with CellTracker Orange (25 μM for 30 min) (Thermo Fisher Scientific) and exposed to parasites for 7 h. All the respective inhibitor treatment was carried out prior to Leishmania infection for a respective period as mentioned previously. After this infection step, the hPMNs were washed with complete media two times to remove extra unbound parasites. Parasite-exposed PMNs were further cocultured with macrophages for 1 h at 37°C under 5% CO₂ and air. At the end of 1 h, the cells were fixed with 4% paraformaldehyde for 30 min at 4°C. The engulfment was visualized under 400× magnification under a fluorescence microscope (Nikon Eclipse TE2000-E inverted microscope; Nikon, Melville, NY). The number of hMDM-engulfed PMNs were counted (by unrelated individual to avoid visual bias), and fold change in engulfment rate was calculated as follows: (Total number of neutrophils engulfed under given condition per 500 macrophages)/(Total number of neutrophils engulfed in Leishmania-infected condition per 500 macrophages) × 100 = % change in phagocytosis with respect to infection.

All the statistical analyses were carried out using SigmaPlot 12.5 (Systat Software, San Jose, CA). Data were analyzed by Student unpaired t test for two samples with equal variance. One-way ANOVA has been used to compare one group versus different treatment groups. Kinetics experiments were analyzed by using two-way ANOVA. Events with p < 0.05 were considered statistically nonsignificant (p > 0.05 or NS), and p < 0.0001 was considered as highly significant events (*p ≤ 0.05, **p < 0.01, and ***p < 0.001). All experiments were repeated using human blood from different individual donors. The data represent the mean ± SEM from three to four independent experiments.

The hPMNs are temporary hosts for trypanosomatid parasites like the Leishmania, and initial neutrophil response could have a substantial bearing on the eventual outcome of the infection.

The current study describes a new finding on the role of parasite-induced autophagy of hPMNs in the regulation of macrophage uptake of the neutrophils. In this context, the occurrence of autophagy in hPMNs postinfection was analyzed by checking a hallmark of autophagy, the conversion of LC3-I protein to LC3-II and its subsequent localization to the autophagosomes (25). The presence of higher number of fluorescent LC3-II puncta and increased expression of LC3-II in the hPMNs indicated an increase in autophagy (Fig. 1A, 1B). Treatment of PMNs with 3-MA, an autophagy inhibitor (14), prior to exposure to the parasites significantly lowered the conversion of LC3-I to LC3-II, evidenced by a noticeable reduction of fluorescent puncta in cells and lessening of LC3-II protein concentration on Western blots (Fig. 1A, 1B). Because starvation induces autophagy (26), starved cells were used as positive controls in which increased puncta formation and higher LC3-II protein concentration was observed (Fig. 1A, 1B). Interestingly, autophagy was observed in uninfected bystander cells as well (Supplemental Fig. 1A), which raised the issue of the kind of stimuli these cells received that could enhance their autophagic activities. Likely, the stimulus would be parasite-secretory products or parasite-neutrophil–combined secretory products; however, exposure of hPMNs to conditioned media from parasite culture or parasite plus neutrophil–combined cultures did not increase autophagy (Supplemental Fig. 1B–D). This suggested that secretory products were not involved in autophagy initiation. The induction of autophagy in hPMNs during parasite exposure appeared to be independent of phagocytosis as treatment of the cells with cytochalasin D, a potent phagocytosis inhibitor, during parasite challenge did not interfere with the induction of autophagy (Supplemental Fig. 1E, 1F). Therefore, the above observations indicated that parasite engulfment or secretions were not necessary for autophagy increase. Time-kinetic analysis of autophagic activities revealed a gradual increase in autophagy postinfection over a period of 12 h, autophagy being maximal at 12 h among the investigated time period (Fig. 1C, 1D). One of the questions was whether a continuous presence of parasites was required for autophagy escalation. Parasites were removed at every hour until 7 h, and autophagy was checked. Even the exposure of 1 h was sufficient to trigger an increase in autophagy (Supplemental Fig. 1G), suggesting that a small period of parasite exposure was sufficient for initiation of signaling for autophagy.

It is essential to measure the autophagic flux in any system for investigating autophagy. The p62 protein or sequestosome 1 (SQSTM1) targets misfolded cellular proteins for autophagy, and therefore, the p62 levels are lower because of turnover in conditions of high autophagy (27). In consonance with such observations, the p62 levels decreased at 12 h when autophagy was highest (Fig. 1D), but when fusion of autophagosome with lysosome was blocked in the presence of Bafilomycin A1, the p62 protein accumulated (Fig. 1E) because of low turnover, suggesting the existence of normal autophagic flux. A few other autophagy markers, like phospho–AMPK-α (early-stage autophagy) (28), phospho–Beclin-1, and phospho-ATG14L (late-stage autophagy) (25), in hPMNs exposed to L. donovani were checked. Increased phosphorylation of AMPK-α, Beclin-1, and ATG14L at amino acid Thr172, Ser93, and Ser29, respectively, was detected (Fig. 1F, 1G). In summary, the above data established a time-dependent increase of autophagy in hPMNs when exposed to the Leishmania parasite. Typical markers of autophagy were displayed with LC3-II formation like p62 decrease with an increase in autophagy and phosphorylation of AMPK-α, Beclin-1, and ATG14L. The autophagy observed was independent of parasite engulfment but dependent on the presence of the parasites around the hPMNs.

Although canonical autophagy or xenophagy is known to occur as a response to infections (5), over the last decade, data has emerged on the occurrence of noncanonical forms of autophagy during bacterial, fungal, and parasitic infections (29). Noncanonical autophagy shares some of the machinery used for canonical autophagy but is functionally distinct and does not require the activity of the preinitiation complex (30). The kind of autophagy initiated in hPMNs during early response to infection by Leishmania is not known and to understand the complexity of the early events, it was deemed suitable to explore the relative existence of the canonical and the noncanonical autophagy. In the preinitiation step of canonical autophagy, ULK1 is involved (25, 30), which is not required for noncanonical autophagy (30). Based on this premise, we used a small molecule inhibitor to ULK1/2 (MRT68921) to stop the formation of the preinitiation complex prior to infection and checked for the autophagic response. Contrary to the expectations, no inhibition of conversion of LC3-I to LC3-II was observed in the presence of MRT68921 (Fig. 2A, 2B), suggesting the possible occurrence of a noncanonical form of autophagy at the initiation of infection. At later time points from 3 h onwards, some inhibition of autophagy was observed with MRT68921 (Fig. 2A, 2B), possibly indicating the presence of a canonical form of autophagy from that point onwards. Western blots of ULK inhibitor–treated cell extracts showed lesser LC3-II formation at 7 h, corroborating the observations with LC3-II staining on autophagosomes. Notably, LC3-I level was also less at 7 h. To ensure that MRT68921 was competent to inhibit canonical autophagy, treatment with MRT68921 was given during inhibition of starvation-induced autophagy by Bafilomycin A1. As expected, addition of MRT68921 resulted in much less accumulation of autophagic punctae (Fig. 2C, 2D), demonstrating the competency of the MRT68921 in inhibiting canonical autophagy. In addition, lesser LC3-II levels was observed on Western blots of MRT68921-treated cells (Fig. 2D). Another confirmation of reduction of autophagy due to ULK1/2 inhibition was the status of phospho–Beclin-1 (S93) and phospho-ATG14 (S29). Beclin-1 phosphorylation was downregulated during ULK inhibition, confirming effective inhibition of ULK kinase activity by the inhibitor (Fig. 2D). Surprisingly, no inhibition of phosphorylation of ATG14 was observed, suggesting ULK-independent activation of ATG14L.

ULK1 activates the class III PI3K complex (PI3KC3), consisting of the PI3K vacuolar protein sorting 34 (VPS34), p150, Beclin-1, ATG14L (25), and the activating molecule in Beclin-1–related autophagy 1 (AMBRA1) protein (25). Both Beclin-1 and VPS34 are required for both types of autophagy (31), and the interaction of Beclin-1–VPS34/PI3KC3 with its molecular partners decides the fate of autophagy. ATG14L in association with Beclin-1 positively regulates autophagy, but the reverse effect is observed when PI3KC3/Beclin-1 interaction occurs with Rubicon (32). Rubicon showed a transient colocalization with Beclin-1 after 1 h infection with the Leishmania parasites (Fig. 2E), confirming the presence of the noncanonical form of autophagy during the early hours of Leishmania infection. The Rubicon–Beclin-1 complex is known to interact with VPS34, reducing its lipid kinase activity and impairing autophagosome maturation (32). The inhibitor to VPS34, Vps34-IN1, prevented autophagy (Fig. 2F, 2G), which is visible through the reduction of autophagy puncta and lessening of LC3-II levels (Fig. 2F, 2G), but it did not interfere with Beclin-1 and Rubicon interaction (Fig. 2H), suggesting that Beclin-1–Rubicon interaction was independent of VPS34 activation. Rubicon, considered as the molecular switch between suppression of canonical autophagy and activation of the noncanonical form of autophagy (33), appears to have played an important role in the early events ensuing after initiation of infection. Therefore, the early part of infection appears to be predominated by noncanonical autophagy with canonical autophagy taking over at later hours.

To initiate infection-induced intracellular signaling, surface molecules on both host and pathogens are involved (34). A variety of surface molecules on the Leishmania parasite have been identified, of which LPG, a constituent of the surface glycocalyx of the Leishmania parasite, plays a major role in host–parasite interactions (35). The predominant components of the glycocalyx of the infective promastigotes are the GPI-anchored phosphoglycosylated glycans and the LPG (35). LPG is composed of a GPI anchor, a glycan core, and a linear phosphoglycan chain (i.e., [Galβ1,4Manα1-PO₄] repeating units and a terminal oligosaccharide cap) (36). A mutant parasite (lpg1-KO) expressing the repeating units but lacking the ability to bring together functional LPG glycan core on the surface (35) was unable to induce any autophagy (Fig. 3A, 3B), suggesting the involvement of the LPG in inducing downstream signaling for autophagy. This observation was reconfirmed by using LPG add-back line of the parasite (lpg1-KO+LPG1) in which induction of autophagy was observed after 5 h of LPG1-complemented parasite exposure (Fig. 3F). The interaction and subsequent uptake of LPG-deficient promastigotes with the neutrophils remains unaffected even in the absence of LPG, suggesting a dispensable role during parasite engulfment by neutrophils. Interestingly, we observed more uptake of parasites in a LPG-deficient condition (Fig. 3E). Prior knowledge shows binding of LPG to TLR2 on host macrophages (37, 38) and NK cells (39) and also engagement of the TLR2 and four receptors by LPG on mouse neutrophils to trigger the innate immune responses (40). Blocking of TLR2 and 4 signaling with TIRAPinh, a cell-permeable MyD88-specific inhibitory peptide (41) (Fig. 3C, 3D), and TAK242 (18) (Supplemental Fig. 2A), capable of blocking only TLR4 signaling, did not result in the blockade of infection-induced autophagy in the inhibitor-treated cells in comparison with the cells without inhibitor treatment (Fig. 3C). The lack of inhibition was confirmed by no change in the LC3-II formation on Western blots of TIRAPinh-treated cell extracts versus the vehicle-treated infected cells (Fig. 3D) and no change in the LC3 puncta. The efficacy of TIRAP inhibitor was confirmed by analyzing the phosphorylation status of JNK1/2 in the presence of LPS stimulation (Supplemental Fig. 2B, 2C). Effective inhibition of phosphorylation of JNK2 (i.e., p46) was observed in the presence of TIRAPinh under LPS stimulation. In summary, based on the above data, it can be inferred that the surface LPG of parasites was essential to trigger autophagy, but this was independent of TLR2/4 engagement by the LPG.

At the time of infection, a large amount of ROS catalyzed by NOX is generated for the elimination of pathogens, bringing about changes in the intracellular milieu (42). The past few years have witnessed the increase of literature detailing the interconnectedness between autophagy and oxidative stress (16, 43). A notable ROS increase as measured by CM-H₂DCFDA fluorescence was observed from 2 h onwards postinfection (Fig. 4A). This ROS was largely generated by NOX activity because NOX inhibitor DPI and apocyanin (data not shown) were able to inhibit this increase (Fig. 4A). One other ROS quenching agent, Trolox, was also able to repress the increase (data not shown). This is in contrast to macrophages in which a reduction of ROS generation is observed over a period of time (44), presumably because of suppression caused by the pathogens. The issue was whether ROS was the inducer of autophagy, and if it was, arguably, scavenging of the ROS would result in interference with autophagy induction. Because inhibition of NOX through DPI addition interfered with the conversion of LC3-I to LC3-II, it was a clear indication of involvement of ROS with autophagy (Fig. 4B, 4C). Previous studies from other laboratories with mouse neutrophils suggest opposite roles of PMNs during visceral and cutaneous leishmaniasis caused by L. donovani and L. major, respectively (11). However, in our experiments, a similar autophagic response as that with L. donovani was induced in hPMNs with L. major infection (Fig. 4B, 4C). Although we did not find any involvement of TLR2/4 in inducing autophagy, we sought to measure if ROS production was influenced through TLR2/4. The presence of TIRAPinh during infection did not interfere with ROS generation (Fig. 4D), showing no involvement of the TLR system in ROS generation as well as autophagy. Similarly, the involvement of LPG in ROS generation was checked, and lpg1-KO parasites that did not stimulate autophagy were unable to induce ROS as well (Fig. 4E), suggesting requirement of an intact LPG on the parasite surface to activate events related to ROS and autophagy generation. Unlike the NO generation in macrophages postinfection with the Leishmania parasite (42), the hPMNs did not show any increase in NO, and there was no Ca²⁺ influx either (data not shown) during the first 7 h of infection. Therefore, hPMNs did not use NO to counter the parasites, instead using ROS as a defense arsenal. From the above studies, a distinct role of ROS was established in the induction of autophagy following infection, the stimulation of ROS being TLR2/4 independent.

In the induction of a process like autophagy, multiple signaling pathways may be involved. An increase in the level of the lipidated form of LC3 (LC3-II) was observed in Leishmania-infected hPMNs under the inhibition of JNK and p38 signaling pathways using SB203580 (p38 MAPK inhibitor) and SP600125 (JNK inhibitor), respectively. The lack of inhibition of autophagy suggested that neither JNK nor p38 signaling were involved in transducing signals (Fig. 5A, 5C). In contrast, the involvement of ERK and PI3K pathways was indicated by a reduction of LC3-II formation when ERK/MAPK- and PI3K-specific inhibitors U0126 and LY294002 were used (Fig. 5A, 5B). An additive effect on reduction of LC3-II expression was not observed even when ERK and PI3K pathways were blocked simultaneously (Fig. 5B), indicating the convergence of the two pathways. Phosphorylation of ERK was observed in hPMNs as early as 5 min after exposure to the parasite, which reconfirms activation of the ERK/MAPK pathway (Fig. 5D). Literature suggests phosphorylation of AMPK-α leads to activation of mTORC2, which in turn induces autophagy through phosphorylation of Akt at Ser-473 position (45). The highest level of phosphorylation of Akt (S473) was observed at 30 min of incubation with parasites, supporting the reports of involvement of PI3K/Akt pathway in autophagy (Fig. 5E). Our experiments with wild-type and lpg1-KO parasites showed that there was ERK or Akt phosphorylation in the case of wild-type, but with lpg1-KO parasites, there was no phosphorylation of ERK and Akt, suggesting the requirement of an intact LPG on the parasite for phosphorylation events to ensue. In summary, the above data provides evidence for involvement of ERK/MAPK and PI3K/Akt pathways in the transduction of signals required for the triggering of autophagy. The fact that ERK and Akt phosphorylation was reduced when NOX was inhibited (Fig. 5F) clearly showed the involvement of ROS in phosphorylation events (Fig. 5F–J). ERK1/2 phosphorylation did not occur when lpg1-KO parasites were used for infection; however, TLR2/4 inhibition did not have any effect on ERK1/2 phosphorylation. Similarly, Akt phosphorylation could be blocked by NOX inhibition but not by TLR2/4 inhibition with TIRAPinh. Collectively, the data provides evidence of LPG involvement in the transduction of signals through the signaling pathways postinfection mediated by ROS through the activation of ERK/MAPK and PI3K/Akt.

The Leishmania parasite infects neutrophils at the infection site, and if they survive within these cells, they are taken up by the macrophages while residing within the neutrophils. Subsequently, the parasites are released inside the host cells without evoking an immune response, and this method of transfer is known as the Trojan Horse mechanism (46). As there was a very distinct event of autophagy associated with the engulfment of the parasite, we sought to understand if infection-induced autophagy in some way influenced the process of uptake of the infected hPMNs. Arguably, if autophagy was responsible for increased uptake, the inhibition of autophagy should reduce the hPMNs engulfment by the macrophages. The rate of hPMN uptake decreased when autophagy was inhibited by 3-MA prior to the addition of the neutrophils to macrophage cultures (Fig. 6A), suggesting a possible requirement of autophagy for neutrophil engulfment. If inhibition of autophagy postinfection could reduce uptake of the cells whereas cells with higher autophagy are engulfed, we sought to understand whether this was an infection-induced phenomenon or whether an increase in autophagy induced by any other means would act similarly. When uninfected hPMNs were treated with an autophagy inducer like rapamycin (mTOR inhibitor), resulting in an increase in autophagy, there was a distinct increase in the rate of hPMN intake (Fig. 6A). This showed that autophagy was essential for neutrophil uptake by macrophages; the means of induction could be different. When engulfment of parasite-exposed cells was checked after the inhibition of canonical autophagy with ULK1/2 inhibitor, there was a 40% decrease in the engulfment, suggesting a significant role of canonical autophagy. The extent of inhibition achieved with 3-MA was significantly more, indicating the contribution of noncanonical autophagy as well. The combined data on the lowering of hPMN engulfment when autophagy was restrained and escalation of intake when autophagy was induced suggested clearly that the uptake of neutrophils was related to the process of autophagy. We did not observe any reduction in neutrophil phagocytosis when neutrophils were treated with Bafilomycin A1 (Supplemental Fig. 2D). As the experiments revealed autophagy to be induced by ROS, we tested if ROS inhibition would affect the uptake as a consequence of lower autophagy. Expectedly, inhibition of ROS generation in hPMNs resulted in reduced engulfment of the neutrophils (Fig. 6A).

It is known that the uptake of cells by phagocytes depends on the relative concentration of eat me or do not eat me signals on the cell surface (46). The uptake of the hPMNs should therefore be a direct outcome of the balance of these signals. Studies describe PS externalization on cell surface as the eat me signal for engulfment (46); however, the hPMNs showed comparable PS externalization in uninfected cells as well as infected cells (Fig. 6B), suggesting no involvement of PS. Both PI and TUNEL staining corroborates this data, suggesting that cells were probably at a very early stage of apoptosis, or they were still not apoptotic (data not shown). Because eat me signals did not change, it was a strong possibility that do not eat me signals were active. CD47 or the do not eat me signal is a self-recognition marker that interacts with signal regulatory protein-α (SIRP-α) expressed on macrophage surface initiating inhibitory signals that blocks macrophage engulfment of the interacting cells. During apoptosis, the concentration of the CD47 is reported to remain unaltered, but the distribution pattern changes (47). Our results show an alteration in membrane distribution of CD47 in infected cells bordering changes of a punctate appearance to a diffused one without significant change in its expression (Fig. 6C, 6D). Interestingly, rapamycin treatment induced a diffuse distribution of CD47 on neutrophil surface, suggesting lowering of do not eat me response, supporting the fact that autophagy induction decreased surface distribution of CD47 on cell surface (Fig. 6C), thereby emanating signals for uptake. In summary, the data support the observation of an increased uptake of neutrophils on enhanced autophagy and decreased uptake when pharmacological inhibitors were used to inhibit autophagy (Fig. 7).

During the initial phase of Leishmania infection, neutrophils home to the site of tissue injury caused by the bite of infected sandfly (48). After the engulfment of parasites, neutrophils can either commence killing of the parasites or act as reservoirs (i.e., acting as Trojan Horses for parasite transfer to the host). Therefore, the intricacies of neutrophil function during early infection are of great interest as that determines the success of eventual pathogenesis of the disease, but much of it remains to be elucidated. In this report, we detail a new finding of involvement of infection-induced neutrophil autophagy in determining the efficiency of parasite-exposed neutrophil uptake by the macrophages.

A mechanism of circumventing host defense by using the Trojan Horse strategy has been reported for many pathogens, including the Leishmania parasite. A study by Nathan and coworkers suggests facilitation of infection by neutrophils by quarantining parasites from other phagocytic cells or releasing transitional-stage parasites better adapted for intramacrophage survival (49). Conversely, a recent study by Ribeiro-Gomes and coworkers demonstrated the uptake of infection-induced apoptotic neutrophils by dendritic cells after intradermal infection with L. major (49). They showed that most of the dendritic cells obtained the parasites at the infection site through the engulfment of the infected neutrophils. Although no uptake of infected neutrophils by macrophages could be detected, it was speculated that rapid phagosomal degradation of enhanced GFP signal may have masked the actual uptake (50). Our study corroborates previous reports of increased uptake of neutrophils by the macrophages.

Although Leishmania parasite–induced autophagy in neutrophils has not been demonstrated, multiple reports confirm macrophage autophagy upon Leishmania infection (51–53) The observed increase in hPMN autophagy postinfection with the Leishmania parasites is in line with reports of autophagy induction by other classes of pathogens (5, 54). To explore the functional role of autophagy, it was important to establish the kinetics of the process in hPMNs. The gradual increase in autophagy with advancing time of infection substantiates the adjustment of the cells to the invading parasites. Cellular macroautophagy can be initiated through two different pathways, a canonical or a noncanonical pathway described in relatively recent times (25, 30). In the absence of an inhibitor blocking the noncanonical pathway, an inhibitor blocking only the canonical pathway, namely ULK1/2 inhibitor (19), was used, whereas 3-MA was used for blocking both pathways. Interaction of PI3KC3 with ULK1/2 initiates the canonical autophagic process by forming the preinitiation complex (25). The repression of total autophagy by 3-MA within 3 h but lack of inhibition by ULK1/2 inhibitor clearly suggested the prevalence of noncanonical autophagy in the early hours. Rubicon–Beclin-1 interaction observed at 1 h postinfection confirmed this. Noncanonical autophagy can proceed without the preinitiation complex but requires the presence of Rubicon (33). Rubicon, also described as a VPS34-interacting protein, inhibits VPS34 lipid kinase and consequent autophagosome formation via the canonical pathway (32). However, the Beclin-1–Rubicon interaction was independent of the kinase activity because inhibition of the VPS34 kinase activity did not interfere with the interaction of the two proteins. Rubicon can also commit a cell to another pathway of noncanonical autophagy termed as LC3-associated phagocytosis or LC3-associated phagocytosis (32). LC3-associated phagocytosis is reported for pathogens in which a single-layered autophagosomal membrane is involved in contrast to the autophagosomes with double membrane, a product of the canonical pathway (33). Because our data suggest the occurrence of two forms of autophagy, the lesser accumulation of LC3-II and expression of LC3-I might be due to the codominance of noncanonical and canonical autophagy. However, once the prevalence of noncanonical autophagy lowered from 3 h onwards, canonical autophagy dominated, and therefore, a continuous accumulation of LC3-I/II occurred from 3 h onwards. This observation is supported by experiments in which after ULK inhibition, LC3-II started appearing from 1 h postinfection at a time when noncanonical autophagy was dominant. Importantly, although the noncanonical form of autophagy could be induced by parasite invasion, it is also possible that suppression of the canonical pathway could have triggered the noncanonical pathway (32), leaving the dual possibility of both canonical and noncanonical autophagy being triggered at early infection. The observed nonsignificant difference in LC3-I and II in ULK-inhibited and noninhibited conditions at 7 h was possibly due to the overlapping presence of both canonical and noncanonical autophagy at this time point. Therefore, from the data, it is evident that the Leishmania parasites were able to induce both the canonical and noncanonical pathways of autophagy in the hPMNs.

Induction of ROS is frequently associated with host response to pathogens in which ROS is used as a weapon against the invaders. NOX2 activity is required for ROS production, and in these experiments, NOX2 was shown to be involved in generating ROS repressible by the NOX2 inhibitors (33). Rubicon promotes this pathway, but as it is not possible to transfect hPMNs, the part Rubicon played had to be interpreted from the indirect evidences of the appearance of Rubicon at 1 h associated with Beclin-1. In the absence of ROS or in NOX2 knockout cells, recruitment of downstream constituents like LC3-II, ATG5, and ATG16L1 are reported to be diminished (33). Suppression of ROS leading to inhibition of autophagy was a clear indication of ROS association with infection-induced autophagy. These findings corroborate observations of neutrophil autophagy using PMA, a known stimulator of oxidative burst (16).

It is known that Leishmania-infected neutrophils and the engulfment of the infected neutrophils promote parasite dissemination by the Trojan Horse mechanism (10). The lowering of hPMN engulfment when both canonical and noncanonical autophagy was blocked by 3-MA clearly indicated involvement of hPMN autophagy in their engulfment by the macrophages. Arguably, if engulfment of cells were prevented because of inhibition of autophagy, cells with higher autophagy would be picked up better by the macrophages. Accordingly, hPMNs with high autophagy induced by rapamycin showed higher engulfment by macrophages, indicating that irrespective of the stimulus, autophagy increase induced an escalation in hPMN engulfment. The difference in engulfment rates achieved by inhibition of total autophagy and selectively the canonical form by ULK1/2 inhibitor suggests that both canonical and noncanonical autophagy have a role in neutrophil engulfment; the contribution of the canonical pathway is more significant.

The uptake of cells by macrophages is dependent on alterations on cell surface molecules like the PS and CD47 (46). Whereas PS acts as a eat me signal, the interaction of CD47 with SIRP-α conveys a do not eat me signal, limiting the clearance of the concerned cells. A change in CD47 distribution from aggregates to a diffuse pattern on the cell surface affecting SIRP-α interaction primes cells for removal (47). In our experiments, there was no change in the pattern of CD47 expression. Without any change in PS or eat me signal distribution, the primary impetus for engulfment of the cells appears to be the lowering or change of distribution of CD47. Events in the absence of CD47 overexpression to substantiate this was not possible with hPMNs; therefore, observations are primarily based on the changes in localization.

Pathogens are recognized by receptors on host cell surfaces, and they form an important component of the initiation of signaling within the host cells. Prior data report recognition of the Leishmania parasites by the macrophages through TLR2/4, mannose, fibronectin, Fc γ, and complement receptors (55), but the receptors on neutrophils remain unknown. The lack of prevention of downstream events postinfection when TLR2/4 inhibitors were used suggested the existence of a different receptor for Leishmania in the neutrophils. Inhibition of Dectin-1 (56) did not interfere with activation of autophagy. Therefore, the identity of the receptor on the host cell involved in Leishmania–hPMNs interaction remains unknown. The host cell receptors are engaged by the pathogens through ligands on their surface. The macrophage receptors engaged by the Leishmania parasite is through the abundant promastigote glycolipid, the LPG (35); however, the interacting ligand to neutrophil receptors is not known. The failure of lpg1-KO parasites to trigger signaling events and consequent autophagy and the ability of lpg1 add-back line to trigger autophagy is strongly indicative of a role of LPG in interaction of the parasite with the neutrophil. The possibility remains that LPG is not the direct trigger but may be facilitating the promastigote–neutrophil interaction. Autophagy induction by lpg2-KO, which is deficient in golgi GDP-sugar transporter (Galβ1,4Manα1‐PO4‐defective mutants), indicates that the presence of the glycan core is sufficient to initiate autophagy (D.M. Pitale and C. Shaha, unpublished observations; A. Descoteaux provided lpg2-KO parasites). Therefore, LPG has a distinct role in autophagy trigger.

Once surface receptors are engaged, the signaling events are initiated within the host cells (57). Because ERK phosphorylation occurred at 5 min postinfection followed by Akt phosphorylation at 15 to 30 min, it appears that ERK phosphorylation drove Akt phosphorylation, which is the known primary driver for autophagy (57). Inhibition studies confirm ERK phosphorylation as the primary upstream event. It was interesting to observe the global effect the parasite incubation had in inducing autophagy both in the infected as well as the bystander cells. It is possible that bystander cells do have effects on the progression of infection, but the changes induced in the bystander cells could be due to changes in the niche caused by secretions from parasites or neutrophils themselves. Because conditioned media from parasite and neutrophil–parasite incubation did not induce autophagy, it appeared that the changes might have been inflicted by some transient contact with the parasite (34).

In summary, this study clearly shows a significant role of neutrophil autophagy induced by infection or other means in determining neutrophil engulfment by the macrophages. In the context of Leishmania infection, the data suggest the contribution of both the canonical and noncanonical pathways in initiating infected neutrophil engulfment. The stimulus for autophagy induction is initiated by parasite LPG through ERK and Akt phosphorylation and consequent NOX2 activation responsible for elevated ROS acting as a stimulus for autophagy occurring through the canonical pathway. This study, therefore, establishes a role of neutrophil autophagy in the engulfment of parasite-laden cells by the macrophages (Fig. 7).

The authors have no financial conflicts of interest.