Leishmania donovani Exploits Macrophage Heme Oxygenase-1 To Neutralize Oxidative Burst and TLR Signaling–Dependent Host Defense

Reactive oxygen species (ROS) is one of the major host defense arsenals against invading pathogens used by macrophages (1). The enzyme primarily responsible for ROS production is multisubunit NAD(P)H-dependent phagocytic oxidase (Phox or NOX2), expressed in professional phagocytes such as neutrophils and monocytes/macrophages (2). Phox complex is usually dormant in quiescent cells, in which its components are segregated into the cytosolic and membrane compartments. NAD(P)H oxidase is comprised of two membrane-localized subunits, gp91phox and p22phox, which form a tightly associated heterodimer referred to as flavocytochrome b558. Four additional essential regulatory components of the NAD(P)H oxidase enzyme, p40phox, p47phox, p67phox, and Rac2, are mostly found in the cytosol and are associated with the membrane-bound components upon activation (3). The catalytic subunit of the NAD(P)H oxidase, gp91phox is synthesized as a 58-kDa polypeptide, which eventually gets glycosylated in endoplasmic reticulum and emerges as a 65-kDa protein. Getting glycosylated further in Golgi and acquiring heme, the 65 kDa protein finally emerges as a matured molecule of 91 kDa (4). This processing of gp91phox increases its affinity for the other membrane resident subunit, p22phox. The incorporation of heme is important for the assembly of the cytochrome b558, and the stability of phagocyte p22phox and gp91phox is dependent on both heme incorporation and heterodimer formation (4).

Intramacrophage pathogens neutralize early ROS production for their successful survival (5). There are several antioxidant macrophage enzymes like superoxide dismutase (SOD), catalase, glutathione peroxidase (GPX), and heme oxygenase-1 (HO-1) used by microbes for this neutralization (6). In various studies, the role of HO-1 in ROS neutralization have been reported. HO-1 is a stress responsive enzyme that metabolizes heme and releases free iron, CO, and biliverdin (BV), which rapidly undergoes conversion to bilirubin (BL) (7). Overexpression of HO-1 protected against TNF-α–mediated airway inflammation via the downregulation of oxidative stress in both cultured human tracheal smooth muscle cells and the airways of mice (8). HO-1 attenuates the cisplatin-induced apoptosis of auditory cells via downregulation of ROS generation (9). During myocardial hypoxia-reoxygenation injury, HO-1 overexpression induces autophagy to protect the stability of the mitochondrial membrane exerting a protective effect (10). Apart from ROS neutralization, HO-1 has an important role in the modulation of innate and adaptive immunity in mammalian systems (11). The immunomodulatory effects of HO-1 can drive both beneficial and detrimental consequences in host immunity against infectious agents (12). Several studies have emphasized that HO-1 can overcome pathogenesis of a variety of immune system–mediated inflammatory conditions, such as malaria (13), ischemia/reperfusion injury (14), intrauterine fetal growth restriction (15), sepsis (16), graft rejection (17), and sickle hemoglobin (Hb) (18). In contrast, HO-1 protects Plasmodium-infected hepatocytes, thereby promoting the persistence of the parasites (19). Overexpression of HO-1 in dendritic cells inhibits their LPS-induced maturation and proinflammatory functions (20), suggesting its role as an anti-inflammatory mediator. HO-1 gene knockdown results in improved control of infection with intracellular bacteria, which can be attributed to impaired bacterial survival, mainly as a consequence of iron limitation for intramacrophage microbes (21).

Leishmania spp. are obligate intramacrophage parasites and are reported to regulate oxidative stress-mediated host response in a differential manner. Whereas L. major stationary phase promastigotes were found to trigger superoxide production in macrophages (22), purified metacyclic promastigotes elicited minimal superoxide production (22). Moreover, L. major promastigotes inhibit phagosomal oxidative activity by inactivation of the SNARE VAMP8 (23). In the case of L. donovani, the causative agent of visceral leishmaniasis (VL), there is limited ROS generation when macrophages were infected with either promastigotes or amastigotes (24), and both the forms inhibit NAD(P)H oxidase assembly at the phagosome membrane (24). Interestingly, although L. donovani promastigotes induced p47phox phosphorylation (25), it was almost undetectable during infection with amastigotes (26). Impaired p47phox phosphorylation in turn led to defective p67phox and p47phox recruitment to phagosome membrane and hampered ROS production (26). All these evidences collectively indicate the ability of Leishmania in resisting NAD(P)H oxidase–dependent early oxidative burst of the host cell. Involvement of HO-1 in facilitating parasite survival has also been addressed in two independent studies with L. chagasi and L. pifanoi (22, 27). However, detailed molecular characterization on the interrelationship between infection-induced HO-1 induction and suppression of oxidative burst is yet to be elucidated.

In the current study, we aimed to gain mechanistic understanding of the molecular events used by the parasite to counteract ROS-mediated host defense. Experiments performed in cell lines, primary cell and mouse models of VL, demonstrated that Leishmania could effectively use host antioxidant enzyme HO-1 not only for ROS neutralization but also for the maintenance of an anti-inflammatory milieu necessary for disease propagation.

L. donovani (MHOM/IN/1983/AG83) promastigotes were maintained in Medium 199 (M199; Invitrogen) supplemented with Hanks salt containing HEPES (12 mM), l-glutamine (20 mM), 10% heat-inactivated FBS, 50 U/ml of penicillin, and 50 μg/ml of streptomycin (Invitrogen). The murine macrophage cell line RAW 264.7 was maintained at 37°C, 5% CO₂ in DMEM (Invitrogen), supplemented with 10% FBS, penicillin (100 U/ml), and streptomycin (100 μg/ml). Bone marrow–derived macrophages (BMDM) were isolated from the femurs and tibiae of euthanized BALB/c mice (6–8 wk old) as described earlier (28). In in vitro infection experiments, RAW 264.7 cells and BMDM cells were infected with stationary phase L. donovani promastigotes at a 10:1 parasite/macrophage ratio (28). Infection was allowed to proceed for 4 h, the noninternalized parasites were removed by washing the plates with 1× PBS, and cells were cultured. Determination of intracellular parasite numbers were done by fixing the cells in methanol and then stained with DAPI (29).

HO-1 Ab was purchased from Abcam (Cambridge, U.K.) and gp91phox Ab from Santa Cruz Biotechnology (Santa Cruz, CA). The rest of the Abs were purchased from either Santa Cruz Biotechnology or Cell Signaling Technology. Alkaline phosphatase–conjugated anti-mouse, anti-goat, and anti-rabbit secondary Abs were purchased from Sigma-Aldrich (St. Louis, MO). The chemicals tin protoporphyrin IX dichloride (SnPP) and heme were purchased from Santa Cruz Biotechnology. Remaining chemicals were purchased from Sigma-Aldrich, unless otherwise mentioned.

Accumulation of intracellular ROS in cells were measured using the oxidant-sensitive green fluorescent dye 2′,7′-dihydrodichlorofluorescein diacetate (H₂DCFDA) (Molecular Probes) (24). The cells (1 × 10⁶) postinfection or during treatment were harvested and suspended in H₂DCFDA (10 μM final concentration) and incubated for 20 min at 37°C dark. ROS level was measured by counting at least 10,000 events per test using a FACSCalibur flow cytometer (BD Biosciences) with an FITC filter, and the cells were gated out based on their fluorescent property. Samples were examined by FACSCalibur, and the results were analyzed using CellQuest software (BD Biosciences).

ROS production was detected via an NBT assay. NBT was reduced by ROS to a dark-blue, insoluble form of NBT called formazan (30). The cells were incubated for 90 min in PBS containing 0.2% NBT. Formazan was dissolved in 50% acetic acid, and the absorbance was determined at 570 nm.

HO-1 activity was measured by spectrophotometric determination of the amount of BL produced from hemin added as the substrate (31). Cells were lysed with a hypotonic buffer (0.1 M potassium phosphate, 2 mM MgCl₂, and complete protease inhibitor mixture; Roche Diagnostics, Milano, Italy [pH 7.4]) for 15 min on ice. After brief sonication, 0.6 M sucrose was added to cell lysates to obtain a hypotonic solution (final 0.25 M sucrose). Lysates were centrifuged at 1000 × g for 10 min at 4°C to pellet nuclei, and supernatant was centrifuged at 12,000 × g for 15 min at 4°C to pellet mitochondria. Finally, the supernatant was ultra-centrifuged at 1,05,000 × g for 1 h at 4°C. Microsomal fractions were resuspended in 100 mM potassium phosphate buffer (pH 7.4) containing 2 mM MgCl₂ and complete protease inhibitor. Protein concentration was determined using a small aliquot of these suspensions (Bio-Rad Laboratories). The microsomal supernatant fraction (cytosol) from the liver of a normal rat served as the source of BV reductase. Liver supernatant was prepared fresh from rat liver by homogenization in 0.1 M sodium citrate buffer (pH 5) containing 10% glycerol. HO-1 activity assay was carried out by incubating 600 μg microsome proteins with a reaction mixture containing 1 mM NAD(P)H, 2 mM glucose-6-phosphate, 1 U glucose-6-phosphate dehydrogenase (Sigma-Aldrich), 25 μM hemin, 2 mg of rat-liver cytosol, and 100 mM potassium phosphate buffer (pH 7.4) (400-μl final volume). The reaction was conducted in the dark for 1 h at 37°C and terminated by placing tubes on ice for 2 min. The amount of BL was determined by the difference in absorption between 464 and 530 nm (extinction coefficient, 40 mM⁻¹ cm⁻¹ for BL). HO-1 activity was expressed in picomoles of BL formed per milligram microsomal protein per hour.

Cells were harvested and lysed using ice-cold lysis buffer (Cell Signaling Technology) supplemented with 3 mM PMSF and 3 mM protease inhibitor mixture. Protein concentration in the clear supernatant was determined using Bradford assay. From each sample, an equal amount of protein (50 μg) was resolved by 10% SDS-PAGE and transferred to nitrocellulose membrane, followed by blocking the membrane with 5% BSA and incubation overnight with primary Ab. After washing, membranes were incubated with alkaline phosphatase–conjugated secondary Ab and detected by hydrolysis of 5-bromo-4-chloro-3-indolylphosphate chromogenic substrate, according to the manufacturer’s instructions. Immunoprecipitation was performed as described previously (32). Briefly, precleared cell lysates (500 μg) were incubated overnight with specific primary Ab at 4°C. Twenty-five microliters of protein A/G plus agarose beads (Santa Cruz Biotechnology) were added to the mixture and incubated for 4 h at 4°C. Immune complexes were collected and washed three times with ice-cold lysis buffer and once with lysis buffer without Triton X-100. The immunoprecipitated samples and cell lysates were resolved by 10% SDS-PAGE and then transferred to nitrocellulose membrane (MilliporeSigma). Thirty micrograms of protein from the whole cell lysate of each sample were loaded as input. The proteins were then immunoblotted as already mentioned.

Stable HO-1 knocked-down cell line was kindly provided by Dr. A.-M. Mitterstiller (33) and maintained as described. Inducible short hairpin RNA (shRNA) transduction was done by treatment of the HO-1 knocked-down RAW macrophages (hmox.sh) with 1 μg/ml doxycycline 24 h prior to experiments. The percentage of efficiency knockdown was compared with the luciferase control (ctrl.sh).

For intracellular heme content assessment, cells were cultured, and after treatment, the cells were washed twice with PBS, and assay was performed 2 h later. Heme content was determined according to the method of Motterlini et al. (34). Briefly, cells were washed with PBS and solubilized by adding 1 ml of concentrated formic acid. The heme concentration of the formic acid solution was determined spectrophotometrically at 398 nm (extinction coefficient 1.56 × 10⁵ M⁻¹ cm⁻¹). Heme content was expressed as picomoles per milliliter.

NAD(P)H oxidase–dependent ROS formation was determined by measuring lucigenin-derived chemiluminescence (35) using GloMax Luminometer (Promega, Madison, WI). Briefly, RAW 264.7 cells were seeded onto cell culture plates. After overnight culture, cells were infected with Leishmania and then lysed with HBSS containing 0.1% Triton X-100 and 0.1 N NaOH. Cellular lysates were incubated with NAD(P)H (200 mM) and lucigenin (100 mM) in HBSS for 30 min at 37°C in the dark. Chemiluminescence was measured in relative light units every 5 min over a period of 60 min.

MTT assay was performed to see the effect of heme on cell viability. A total of 1 × 10⁴ cells were grown in a 96-well plate and incubated overnight. The cells were infected with increasing concentrations of heme (1–100 μM) followed by L. donovani (1:10) infection. Thereafter, MTT (5 mg/ml) was added to the plates and incubated at 37°C for 4 h. Finally, formazan crystals were solubilized in solubilization buffer, and absorbance was measured at 570 nm. All assays were performed in triplicate. The extent of cell viability was measured as the percentage of viability with respect to untreated cells.

ELISA was performed using a sandwich ELISA kit (Quantikine M; R&D Systems, Minneapolis, MN). The detection limit of these assays was >5.1 and >2.5 pg/ml for TNF-α and IL-12, respectively.

Animal maintenance and experiments were performed in accordance with the guidelines provided by the Committee for the Purpose of Control and Supervision of Experiments on Animals (New Delhi, India). The protocol was approved by the Institutional Animal Ethics Committee (IAEC) of Bose Institute (Kolkata, India) (IAEC approval no. IAEC/BI/82/2017). For in vivo infection, female BALB/c mice (∼20 g) were injected via the tail vein with 1 × 10⁷ stationary phase L. donovani promastigotes. Infection was assessed by removing the liver and spleen from infected mice up to 6 wk. Parasite burden was determined from Giemsa-stained impression smears (32). Liver and spleen parasite burdens, expressed as Leishman–Donovan units (LDU), were calculated as the number of amastigotes/1000 nucleated cells × organ weight (in grams) (36). Splenic macrophages from BALB/c mice were isolated and cultured as described earlier (37). Briefly, the spleens were taken out and mildly glass ground, residual erythrocytes were lysed with hypotonic buffers, splenocytes were washed and counted, and 1 × 10⁷ cells/ml were then seeded on tissue culture plates and incubated at 37°C. Fresh DMEM was then used to wash the adherent splenic macrophage cells. Adherent cells were then stimulated with soluble leishmanial Ag (20 μg/ml) for 48 h, followed by measurement of proinflammatory cytokines by ELISA (38).

Isolated livers were fixed in 10% formalin (Merck) and embedded in paraffin wax. Tissue sections (5 mm) were made with microtome (Leica Biosystems) and stained with H&E to study their microarchitecture by light microscopy.

Densitometric analysis for all experiments were carried out using ImageJ software. Band intensities were quantitated densitometrically, and the values obtained were normalized to endogenous control and expressed in arbitrary densitometric units. The ratios of OD of particular bands to endogenous control are represented as bar graphs adjacent to figures.

Most of the experiments were performed at least three times. Data are shown as mean ± SD of three independent experiments unless otherwise stated as n values given in the legend. Student t test was used to assess the statistical differences among pairs of datasets with a p value <0.05 considered to be significant.

Because one of the early responses of macrophages toward microbial invaders is the production of ROS, RAW 264.7 cells were infected with live or paraformaldehyde-fixed L. donovani metacyclic promastigotes (PFA Fx) for various time periods, and ROS level was measured by DCFDA-based FACS analysis (Fig. 1A). Infection with live L. donovani was found to cause maximum ROS production (68.8%) at 0.25 h postinfection, which gradually reduced to basal level at 1 h postinfection (15.9%) (Fig. 1A). In contrast, ROS level in macrophages infected with PFA Fx was found to be persistently high even after 1 h of infection (54.5%) (Fig. 1A). To find out whether ROS level affects intramacrophage survival of L. donovani parasites, infected macrophages were pretreated with increasing concentration of ROS-inducer H₂O₂ for 1 h, and the number of parasites was counted at 24 h postinfection. Treatment of RAW cells with an increasing concentration of H₂O₂ up to 400 μM did not affect the cell health (data not shown). H₂O₂ treatment showed a dose-dependent decrease in parasite survival with maximum suppression obtained at the concentration of 400 μM (53.5% decrease of parasite count over infected control; p < 0.001) (Fig. 1B). However, when ROS production was inhibited by using ROS quencher N-acetyl cysteine (NAC, 5 μM) (39) along with 400 μM H₂O₂, reduction in parasite count was significantly reversed (Fig. 1B), thereby suggesting that neutralization of ROS is necessary for parasite survival. To determine how Leishmania parasites were able to suppress macrophage ROS production, we screened macrophage antioxidant enzymes, catalase, Cu/Zn SOD, GPX, and HO-1, which might neutralize the early burst of ROS level. Of all the antioxidant enzymes studied, HO-1 only showed an early and significant induction at the protein level within 0.5 h of infection (4.5-fold induction compared with uninfected control; p < 0.001), and the level persisted as studied up to 4 h of infection (Fig. 1C). Similar results were also observed in infected BMDM cells (2.1-fold induction compared with uninfected control; p < 0.001) (Supplemental Fig. 1A). At the mRNA level also, HO-1 showed a similar pattern of induction in both RAW and BMDM cells with maximum expression observed at 0.5 h postinfection (4.7- and 4.3-fold induction in RAW and BMDM cells, respectively, compared with control; p < 0.001) (Fig. 1D, Supplemental Fig. 1B). HO-1 activity was also significantly induced with a maximum at 0.5 h postinfection (3.3- and 3.6-fold increase over control in RAW and BMDM cells, respectively; p < 0.001) and maintained a steady level as observed up to 4 h (Fig. 1E, Supplemental Fig. 1C). Taken together, these results suggest that L. donovani infection causes suppression of macrophage ROS during early hours of infection possibly via upregulation of host antioxidant enzyme HO-1.

To ascertain the role of HO-1 in neutralization of early ROS level in Leishmania-infected cells, we transduced mouse macrophage cell line RAW264.7 with lentiviruses encoding doxycycline-inducible HO-1 shRNA to establish a stable HO-1 knocked-down cell line (hmox.sh) (33). RAW cells transduced with lentiviruses encoding control shRNA served as control (ctrl.sh). Doxycycline induction resulted in 78.9% reduction of HO-1 expression as assessed by immunoblotting (Fig. 2A). Interestingly, along with a marked increase in ROS production (53.6% increase over ctrl.sh macrophages in RAW cells; p < 0.001) (Fig. 2B), parasite survival was found to be significantly decreased in infected hmox.sh cells (64.6% decrease compared with ctrl.sh macrophages; p < 0.01) (Fig. 2C, Supplemental Fig. 2A), indicating a prominent role of HO-1 in parasite survival. The increase in ROS production in L. donovani–infected hmox.sh cells was not because of low infectivity as the extent of parasite internalization in hmox.sh and ctrl.sh macrophages was comparable with infected control as observed at 4 h postinfection (Supplemental Fig. 2B). As HO-1 is a heme-degrading enzyme (7), increased HO-1 activity during infection (Fig. 1E, Supplemental Fig. 1C) prompted us to measure total cellular heme content of Leishmania-infected macrophages. Heme content of infected cells showed a marked reduction at 0.5 h postinfection (26.8 and 30.1% in infected RAW and BMDM cells, respectively), which persisted as observed up to 4 h postinfection (Fig. 2D). However, administration of 20 μM SnPP, a known inhibitor of HO-1 (40), reversed infection-induced decrease in heme content in both RAW and BMDM macrophages. An 80.2% inhibition of HO-1 activity was found in the presence of inhibitor compared with control (data not shown). Moreover, correlation analysis highlighted a reciprocal relationship between HO-1 activity and total cellular heme content in infected RAW cells (Supplemental Fig. 2C) and further suggested that the induction of HO-1 during infection leads to a reduction of ROS production along with a decrease of total cellular heme content. Because heme is an essential component of the major ROS-producing macrophage enzyme NAD(P)H oxidase (4), activity of the enzyme was assayed in L. donovani–infected macrophages. In both RAW and BMDM cells, enzyme activity was found to be maximum at 0.25 h postinfection, significantly decreased at 0.5 h postinfection (68.7 and 69.5% reduction compared with 0.25 h infected cells in RAW and BMDM cells, respectively) and thereafter maintained a steady level as observed up to 4 h postinfection (Fig. 2E, Supplemental Fig. 2D). NAD(P)H oxidase activity in L. donovani–infected cells was found to be increased by treatment with increasing concentrations of SnPP, with a maximum increase observed at 20 μM SnPP (Fig. 2F). Increased ROS production observed in hmox.sh-infected cells was found to be significantly decreased (64.5% decrease compared with ctrl.sh-infected cells; p < 0.01) when cells were treated with NAD(P)H oxidase–specific inhibitor apocynin (50 μM) (41) (Fig. 2G). Interestingly, the effect of apocynin was much less pronounced in infected and ctrl.sh-infected cells having functional HO-1 (Fig. 2G), thus indicating that HO-1 probably decreased cellular ROS by inhibiting NAD(P)H oxidase activity. Furthermore, a significant increase in parasite count was also observed in apocynin pretreated hmox.sh cells compared with untreated hmox.sh macrophages (Fig. 2H), whereas like the ROS level, the effect of apocynin administration was negligible on parasite survival in infected and ctrl.sh-infected macrophages. All these results collectively suggest that HO-1 regulates NAD(P)H oxidase activity for reduction of early oxidative burst, which helps in intramacrophage survival of L. donovani.

The NAD(P)H oxidase is a multimeric enzyme complex containing both cytosolic (p40phox, p47phox, p67phox, and Rac2) and membrane-bound (p22phox and gp65 [gp91phox]) proteins. Assembly of all the components to form the NAD(P)H oxidase complex on the target membrane is essential for the local release of optimal amounts of superoxide (3). Incorporation of heme is a prerequisite for the formation of the gp91phox–p22phox heterodimer, and both gp91phox and p22phox are degraded in the absence of heme, leading to instability of NAD(P)H oxidase (4). To validate the role of HO-1–induced decrease in cellular heme in the regulation of NAD(P)H oxidase complex, we checked mRNA and protein levels of different subunits of NAD(P)H oxidase in RAW264.7 cells. The transcript levels of all the subunits did not change significantly as studied up to 4 h postinfection (Supplemental Fig. 3A). In contrast, the protein levels of membrane-bound subunits gp91phox and p22phox showed a decreasing pattern of expression as studied up to 4 h of infection (65.1 and 78.2% reduction compared with 0.25 h infected cells for gp91phox and p22phox, respectively). However, the protein level of cytosolic subunits p67phox, p47phox, p40phox, and Rac2 remained almost unaltered (Fig. 3A). In contrast, infected hmox.sh cells displayed increased gp91phox and p22phox (3.8- and 4.3-fold increase over control shRNA-treated infected cells; p < 0.001) (Fig. 3B, 3C), further validating the role of infection-induced HO-1 in regulating NAD(P)H oxidase subunit expression. To determine whether heme replenishment may revert the effects of HO-1, infected macrophages were subjected to increasing concentrations of heme (1–100 μM), and NAD(P)H oxidase activity was determined (Supplemental Fig. 3C). No apparent toxicity was found up to 100 μM (Supplemental Fig. 3B), and 75 μM heme was chosen as the optimum dose. Pretreatment with 75 μM heme led to significantly increased NAD(P)H oxidase activity in 1 h infected cells (2.8-fold; p < 0.001) (Fig. 3D), along with a marked increase in ROS production (2.9-fold over infected cells) (Fig. 3E). Expression of gp91phox and p22phox was also found to be increased in heme-pretreated infected cells (Fig. 3F). Because ROS production was found to be significantly increased when macrophages were infected with PFA Fx at 1 h postinfection, we checked all the above parameters in PFA Fx also. Likewise, in heme-pretreated infected cells, PFA Fx showed increased NAD(P)H oxidase activity (3.5-fold increase over infected cells) (Fig. 3D), enhanced ROS generation (3.3-fold over infected cells) (Fig. 3E), and increased gp91phox and p22phox expression (Fig. 3F). Infected cells replenished with heme further documented decreased intracellular parasite numbers (67.2% in RAW cells over infected control; p < 0.001) (Fig. 3G), but heme replenishment had no significant effect on parasite survival in hmox.sh macrophages over control shRNA-treated cells (Fig. 3G). These results suggested that HO-1–mediated heme-dependent instability of NAD(P)H oxidase subunits resulted in reduced ROS production, leading to increased intramacrophage parasite survival.

Although oxidative burst is an early host response and we found that HO-1 plays an important role in neutralizing it, expression of HO-1 at the protein level continued to remain upregulated even at a late phase of L. donovani infection (Fig. 4A). A similar finding was observed in infected BMDM cells (Fig. 4B). These observations led us to investigate whether HO-1 plays any other role in parasite survival apart from neutralizing ROS. VL is known to be associated with upregulation of disease-progressing Th2 cytokines and concomitant downregulation of disease-resolving Th1 cytokines. To determine whether HO-1 has any effect on cytokine balance during infection, we checked for the effect of HO-1 silencing on the levels of IL-12 and TNF-α production in L. donovani–infected cells as well LPS plus L. donovani–infected cells. LPS induced increased production of IL-12 and TNF-α (Fig. 4C), which were significantly decreased in LPS plus L. donovani–infected cells. In contrast, levels of IL-12 and TNF-α were increased in hmox.sh-infected cells both in the presence or absence of LPS, thereby implicating a role of HO-1 in regulating cytokine balance. To find out whether heme degradation by HO-1 is responsible for downregulation of cytokines, the above experiment was conducted in the presence of heme, which did not significantly alter the cytokine level (data not shown). Next, to ascertain whether the HO-1–mediated effect on proinflammatory cytokines is attributed to its reaction products CO, BL, BV, and Fe, hmox.sh-infected cells were treated with 75μM CO-releasing compound (CORM2), 10 μM BL, 10 μM BV, and 400 μM iron chelator deferoxamine mesylate (DFO) (42). A significant inhibition of IL-12 and TNF-α was observed only in CORM2-treated macrophages (63.6 and 57.4% decrease, respectively, compared with infected hmox.sh cells; p < 0.001), whereas BL, BV, and DFO did not have any effect on the increased proinflammatory cytokine production found in hmox.sh-infected cells (Fig. 4D). To further ascertain that CO is responsible for the downregulation of IL-12 and TNF-α, LPS-treated macrophages were also treated with CORM2, which showed a significant decline in the levels of IL-12 and TNF-α (51.7 and 64%, respectively, over LPS-treated cells; p < 0.001) (Fig. 4E). Because CORM2 contains ruthenium, RuCl₂ was used as a control to show that the change in proinflammatory cytokine level is not due to the presence of ruthenium. When cells were pretreated with CO quencher Hb (5 μM) for 0.5 h before CORM2 administration and subjected to LPS treatment, CORM2-mediated decreased cytokine levels were restored and became almost comparable with LPS-treated cells (Fig. 4E). Hb (5 μM) pretreatment of Leishmania-infected cells also led to an increase in TNF-α and IL-12 production (3.5- and 2.8-fold, respectively, over infected cells; p < 0.001) (Fig. 4E). These results suggest that suppression of proinflammatory response by HO-1 is mediated by CO, generated as a by-product of heme catabolism.

Infection-induced HO-1 reaction produced CO-mediated downregulation of proinflammatory cytokine production prompted us to investigate the effect of CO on the transcription factors, which primarily regulate TLR4-induced proinflammatory cytokine production (43). The major ones involved were found to be NF-κB and IFN regulatory factor (IRF) 3 (44), and therefore, we analyzed the expression level of phosphorylated NF-κB p65, p50, and IRF3 in whole cell extract of LPS- or L. donovani–treated cells (Fig. 5A, 5B). As expected, LPS-treated samples depicted increased expression of p-p65, p-p50, and p-IRF3 (Fig. 5A), but an almost undetected expression was observed in L. donovani–infected cells at early time points (Fig. 5B). The observation that the CO-mediated downregulation of proinflammatory cytokine production was reversed by Hb (CO quencher) treatment (Fig. 4E) led us to check the expression levels of p-p65, p-p50, and p-IRF3 in L. donovani–infected cells pretreated with 5 μM Hb. The decrease in the expression levels of p-p65, p-p50, and p-IRF3 in L. donovani–infected cells was reversed by treatment with Hb (Fig. 5C), indicating a direct role of CO in inhibiting the phosphorylation-mediated activation of both NF-κB and IRF3. The expression levels of p-p65, p-p50, and p-IRF3 were also increased in hmox.sh-infected cells and were downregulated when these cells were treated with 75 μM CORM2 (Fig. 5D). TLR4 is known to interact with its adapter molecules MyD88 and TIR domain–containing adapter-inducing IFN-β (TRIF) to activate NF-κB and IRF3, respectively (41). Because the activation of both the transcription factors were hampered during L. donovani infection, the association of TLR4 with MyD88 and TRIF were assessed by coimmunoprecipitation using anti-TLR4 Ab (Fig. 5E). Unlike LPS-treated cells, which showed a strong interaction, a weak association was observed between both MyD88 and TRIF with TLR4 in L. donovani–infected ctrl.sh macrophages (Fig. 5E). On the contrary, infected hmox.sh cells showed greater association of both MyD88 and TRIF with TLR4, and this interaction was inhibited in presence of CORM2 (Fig. 5E), although no detectable change in the association was observed in CORM2-administered ctrl.sh-infected cells (Fig. 5E). The inhibitory effect of CORM2 was found to be reversed in hmox.sh-infected cells pretreated with Hb (Fig. 5E). A little increase in the association of TLR4 with MyD88 and TRIF in the presence of Hb was also found in ctrl.sh-infected cells. Protein levels of TLR4, MyD88, and TRIF remained unaltered in whole cell lysate during all the treatments (Fig. 5E). These results suggest that in L. donovani–infected macrophages, CO, a by-product of heme degradation by HO-1, hampers the association between TLR4 with the adapter proteins MyD88 and TRIF, resulting in decreased activation of NF-κB and IRF3, leading to the downregulation of TNF-α and IL-12.

To evaluate the role of HO-1 on parasite survival in experimental BALB/c mouse model of VL, L. donovani–infected mice were administered i.p. with HO-1 inhibitor SnPP. The dose-titration experiments assessed the efficacy of different doses of SnPP against established (7 d) infections with a dose range of 1–10 mg/kg per day given daily for 7 d. The infection was allowed to proceed for 6 wk, after which the animals were sacrificed, and the antileishmanial potency was assessed in terms of parasite burden in liver and spleen. During the experimental period, all the animals remained healthy. Maximum suppression of organ parasite burden was obtained at a dose of 5 mg/kg per day (67.1 and 69.5% in liver and spleen, respectively) (Fig. 6A, 6B). Next, to find out the effectiveness of SnPP on the progression of VL, infected BALB/c mice were injected with SnPP at a dose of 5 mg/kg per day daily for 7 d starting at 1 wk postinfection, and the liver and spleen parasite burden (LDU) was determined after 2, 4, and 6 wk of infection (Fig. 6C, 6D). In the case of SnPP-treated mice, the parasite burden was found to be lowered (51.9 and 66.8% in liver and 61.3 and 69.6% in spleen at 4 and 6 wk postinfection, respectively, compared with the infected control) (Fig. 6C, 6D). H&E staining of the liver sections of infected BALB/c mice at 4 wk postinfection showed a partial loss in tissue architecture and detectable granuloma formation (Fig. 6E). In contrast, photomicrographs from SnPP-treated infected mice revealed well-defined zones of granuloma formation (Fig. 6E). Expression levels of HO-1 and gp91phox were determined in splenic macrophages isolated from control, infected, and SnPP-administered infected mice after 2, 4, and 6 wk postinfection (Fig. 6F). Unlike control mice, which showed a steady expression of gp91phox at all time points, splenic macrophages isolated from infected mice showed a decreasing pattern of gp91phox with the maximum decrease obtained at 4 wk postinfection (Fig. 6F). HO-1 expression, however, showed an increased expression at 2, 4, and 6 wk postinfection (Fig. 6F). Treatment with the HO-1 inhibitor SnPP in mice significantly decreased HO-1 expression along with increased expression of gp91phox, in which maximum expression was obtained at 4 wk postinfection (Fig. 6F). The levels of proinflammatory cytokines TNF-α and IL-12 were analyzed in the supernatant of splenic macrophages isolated from control, infected, and SnPP-treated infected mice at 2, 4, and 6 wk postinfection. In the case of SnPP treatment, a marked increase in the level of TNF-α and IL-12 (758.1 ± 43.3 compared with 103.3 ± 33.3 pg/ml for TNF-α and 1016.2 ± 95.3 compared with 100.3 ± 36.6 pg/ml for IL-12, respectively; p < 0.001) were observed at 4 wk postinfection compared with the infected control (Fig. 6G, 6H). These results suggest the potential of SnPP as an effective therapeutic agent in VL as the inhibition of HO-1 by SnPP markedly reduced parasite burden in BALB/c mouse model, along with a substantial increase in the proinflammatory cytokine level (Fig. 7).

To successfully survive within host macrophages, intracellular parasite Leishmania employs efficient strategies to subvert or impair the antimicrobial arsenals of macrophages (45, 46). Immediately after its entry inside the cell, the first line of defense put forward by the host is oxidative burst (24, 39, 47). Several studies have shown the role of ROS in microbial killing (48). However, a significant reduction in ROS level followed by live Leishmania infection indicated active suppression of ROS. In the current study, we have shown that upon L. donovani infection, there is rapid and significant upregulation of the antioxidant enzyme HO-1. Suppression of HO-1 during an infected condition led to the inhibition of ROS production along with reduced parasite survival. Our study clearly documented that infection-induced HO-1 exerted its antioxidant effect through heme degradation. Heme was indispensable for the maturation of the catalytic component of NAD(P)H oxidase gp91phox (49), and restoring heme in an infected condition reversed inhibitory effect of HO-1. We provided further evidence that HO-1 also helps in creating a parasite-favorable anti-inflammatory milieu by releasing CO during heme degradation. CO inhibited the association of TLR4 with MyD88 and TRIF, thereby preventing activation of NF-κB– and IRF3-dependent proinflammatory cytokine production. Using cell culture– and animal model–based methods, the current study underscores the crucial role of HO-1 in parasite survival, leading to the establishment of successful infection (Fig. 7).

HO-1 is a heme-metabolizing enzyme and plays an important role during oxidant-induced injury (50). The role of HO-1 has been found to be responsible for the successful survival of Mycobacterium abscessus (51) and Burkholderia pseudomallei (52). There are reports that HO-1 also facilitates parasite survival of Leishmania sp. (22), but none of these studies investigated its role during an early time point of infection. HO-1 is known to upregulate the level of another antioxidant enzyme SOD, leading to downregulation of the ROS level (18). However, no significant upregulation in the level of other antioxidant enzymes, SOD, catalase, and GPX, was observed, precluding their involvement in neutralizing early oxidative burst. Along with the increased HO-1 expression, infected macrophages displayed decreased heme content, which prompted us to check the activity of major ROS-producing enzyme NAD(P)H oxidase, because the maturation of its catalytic component gp91phox is heme dependent. The significant decrease in NAD(P)H oxidase activity along with decreased protein expression of gp91phox during early Leishmania infection strengthened our hypothesis. This is in accordance with the study in which it has been shown that HO-1 activation limits heme availability for the maturation of the gp91phox subunit and assembly of the functional NAD(P)H oxidase (50). Another study showed the role of HO-1 in Salmonella infection (33), in which HO-1 knockout–infected cells documented uniform gp91phox expression. Impaired NAD(P)H oxidase activity has earlier been reported also in Leishmania infection. Defective phosphorylation of the NAD(P)H oxidase cytosolic component p47phox and its exclusion from the phagosome has been documented in the impairment of NAD(P)H oxidase during L. donovani infection (26). In the case of L. pifanoi infection, impairment of NAD(P)H oxidase has been attributed to HO-1–induced heme degradation (27), and furthermore, absence of heme was shown to result in defective maturation of gp91phox from its precursor gp65 (27). The present study attested this observation by Pham et al. (27) and further extended it by showing that in absence of heme, expression of both the critical subunits of NAD(P)H oxidase, gp91 phox and p22 phox, decreases. Moreover, restoration of heme in the presence of HO-1 leads to the functional assembly of NADPH oxidase and ROS production. This is in line with earlier work in which HO-1 induction depicted a reciprocal relation with gp91phox expression, and the effect was reversed by exogenous heme loading (50).

ROS generation by the host is an early antimicrobial defense. However, sustained HO-1 expression at later hours of infection led us to study its effect on proinflammatory cytokines. TLRs recognize the pathogen-associated molecular pattern and lead to the production of proinflammatory cytokines TNF-α and IL-12, ultimately causing parasite clearance by creating a host favorable environment (44). Our study showing that the level of TNF-α and IL-12 was higher in hmox.sh-infected macrophages suggested a role of HO-1 in curtailing inflammation. HO-1 also promotes the persistence of L. chagasi infection and is strongly associated with inflammatory imbalance during VL (22). Because replenishing heme during the infection did not restore inflammatory cytokines, our attention shifted toward heme-degraded metabolites, CO, Fe, and BV (42). BV is later converted to BL with the help of BV reductase (47). BL, BV, free Fe (53), and CO all have been shown to downregulate proinflammatory cytokines TNF-α and IL-12, but depletion of none except CO marked any effect in cytokine level during infection. CO has also been shown to inhibit the production of TNF-α, IL-1β, and MIP-1β by LPS-activated mouse macrophages and increased IL-10 expression (54). CO negatively controlled TLR signaling pathways by inhibiting translocation of TLR to lipid rafts (43). Detailed analysis of the TLR pathway in presence or absence of CO during infection revealed that the transcription factors NF-κB and IRF3 were the main targets, as their impaired nuclear translocation were restored upon treatment with Hb, the CO quencher. IRF3 and NF-κB activation is mediated by TLR4 signaling, and CO derived from HO-1 has already been shown to inhibit TLR2, 4, 5, and 9 signaling but not TLR3-dependent signaling in macrophages. Moreover, the impairment of TLR4 signaling is well documented in Leishmania infection. Our observation further attested this finding in which CO hampered the association between TLR4 with the adapter proteins MyD88 and TRIF, ultimately leading to the downregulation of TNF-α and IL-12. CO has previously been found to inhibit trafficking of not only TLRs but also of adaptor molecules, resulting in the suppression of downstream signaling pathways (43). There are reports in which CO was shown not only to suppress LPS-induced proinflammatory cytokine production but also decrease mortality in a mouse model of sepsis (55).

The role of HO-1 in parasite survival was also verified in the BALB/c mouse model of VL, in which infected mice were subjected to the administration of SnPP, a potent HO-1 inhibitor. SnPP, apart from inhibiting HO-1, significantly increased liver and spleen parasite burden. Moreover, in the presence of SnPP, the disease propagating anti-inflammatory environment was shifted toward host favorable proinflammatory condition. In summary, the current study demonstrates that HO-1 plays a crucial role in the survival and persistence of L. donovani within the hostile environment of host cell and could be a potential therapeutic target in VL.

We thank Dr. Pijush K. Das for providing the opportunity to carry out a few experiments in the Indian Institute of Chemical Biology (Kolkata, India). We acknowledge Dr. Amrita Saha for helping with microscopy. We sincerely express our gratitude towards Subrata Modak, Avishek Nath, Dr. Ranajoy Ghosh, and Dr. Aniket Haldar, Medical Technologists Laboratory, School of Digestive and Liver Diseases, the Institute of Post-Graduate Medical Education and Research (Kolkata, India) for histopathological facilities.

The authors have no financial conflicts of interest.