Female Sprague Dawley (SD) rats, BALB/c mice, and C57BL/6 mice were purchased from Zhongshan School of Medicine, Sun Yat-Sen University (SYSU), Guangzhou. Female Nos2^−/− SD rats were generated by TALENs and bred in the animal facility of SYSU (Shen et al., 2017). All the animals were maintained in specific pathogen-free conditions. Protocols for the use of animals were approved by the Institutional Review Board for Animal Care at Sun Yat-Sen University (NO:31472058).

Monoclonal antibodies used for phenotypic and intracellular cytokine analyses: isothiocyanate (FITC)-anti-CD3, PE-Cy7-labeled anti-CD4, and peridinin-chlorophyll protein (PerCP)-labeled anti-CD8 were purchased from BD Biosciences (San Jose, CA). Rat ELISA kits for cytokine IFN-γ and IL-4 were purchased from BD Biosciences.

Leishmania amazonensis LV78 (RAT/BA/74/LV78) causes only cutaneous infection, a clone expressing enhanced green fluorescent protein (eGFP) was kindly provided by Dr. K. P. Chang. This parasite showed no difference in infection virulence to macrophages or mice compared to the wild-type strain (Chan et al., 2003). The amastigotes of the LV78 strain were freshly isolated from footpad lesions of previously infected BALB/c mice and transformed into promastigotes in Schneider’s insect medium (SDM, pH 6.9) supplemented with 10% FBS (Hyclone) at 26°C. The parasites at stationary phase (day 5 culture) were used for in vitro or in vivo infection. Only 2 to 9 passages were used in subculture of the parasites in order to make sure that virulence of the parasites was maintained for infection.

Naïve peritoneal macrophages were isolated by peritoneal lavage with ice-cold D-hanks buffer as described preciously (Li et al., 2012). Briefly, WT and Nos2^−/− rats (females, 9weeks old) were sacrificed by CO₂ asphyxia. A total of 20ml ice-cold D-hanks buffer were injected into the peritoneal cavity and then withdrawn. Cells were collected by centrifugation and resuspended in RPMI-1640 (without phenol red) supplemented with 10% FBS and 100U/ml penicillin and streptomycin. Then, 5×10⁵ macrophages were seeded into each well of a 24-well plate with 500μl medium. Four hours later, unattached cells were removed by washing twice with pre-warmed D-hanks buffer to obtain macrophages. Macrophages were allowed to culture overnight before infection. LV78 promastigotes from stationary culture were washed 3 times in D-hanks buffer, resuspended in complete RPMI-1640 medium, and adjusted to a final concentration of 2.5×10⁷. Aliquots of 100μl were put into each well of a 24-well plate with a ratio of parasites to macrophages of 5:1. They were then incubated at 35°C with 5% CO₂ for 6h, and free promastigotes were removed by washing 3 times with pre-warmed D-hanks. Cells were treated with or without various combinations of IFN-γ (50ng/ml) and aminoguanidine (AG, 5mM). Culture supernatants were collected at indicated time points to determine the production of NO. At 0, 24, 48, and 72h postinfection, and parasite proliferation was determined by counting at least 300 macrophages to enumerate the percentage of cells infected and amastigotes per 100 macrophages.

Culture supernatants were collected from treated cells. Proteins were removed by adding ZnSO₄ to a final concentration of 30mg/ml, and the liquid was thoroughly vortexed and centrifuged at 8,000g for 10min. Then, 100μl of total supernatant was mixed with 100μl Griess reagent and incubated in the dark for 10min; the absorbance was read at 540nm, and nitrite concentrations were calculated against a standard curve.

Arginase activity of macrophages was measured by a colorimetric method as described (Corraliza et al., 1995). Tenmm MnCl₂ and 0.5ML-arginine were successively added to macrophage lysates for 1h at 37°C. The reaction was stopped by addition of an acid solution (H₂SO₄: H₃PO₄: H₂O=1:3:7), and the urea generated by arginase was analyzed by addition of a-isonitrosopropiophenone at 100°C for 45min. The colored product was quantified by absorption at 550nm in an ELISA reader. Arginase activity was determined as the amount of urea produced from total protein of peritoneal macrophages.

Groups of WT, heterozygous (Het), and Nos2^−/− SD rat (n=12, in each group, 4weeks old) were subcutaneously infected in the left ear with 10⁷ stationary phase L. amazonensis LV78 promastigotes, using a 30G needle, in a volume of 10μl PBS, while the control rat ear received 10μl PBS without parasites. Uninfected rats of the same age were used as controls. The lesion thickness and diameter were measured every 4 or 8days with a dial caliper. Lesion thickness was calculated by subtracting the thickness of the control ear from the infected ear, while lesion diameter was measured as the longest axis of the ear lesion.

At 32 and 64days postinfection, 3 rats per group were sacrificed and the parasite load in the infected ears was estimated by using a limiting dilution assay as previously described with modification (Belkaid et al., 2000). Briefly, the infected ears were cutoff, and their surfaces were disinfected in 70% ethanol, and the dorsal and ventral part of the ear were separated with forceps. The two ear leaflets were placed dermal side down on SDM containing 1mg/ml collagenase 1 (Sigma) and digested for 2h in 37°C. Then, the dermal sheets were cut into small pieces and homogenized in a glass tissue homogenizer with 2ml SDM followed by the liquid being passed through a 70μm cell strainer (Falcon, BD) to obtain a single cell suspension. Serial 10-fold dilutions were made with SDM supplemented with 10% FBS and 5% defibrinated rabbit whole blood lysate, and aliquots of 100μl were distributed in 96-well plates with 12 wells per dilution. The plates were kept at 26°C for up to two weeks, and the number of parasites was determined by the highest dilution for which the parasites could be grown.

Infected ears were collected and fixed in 4% paraformaldehyde for 24h. They were dehydrated with gradient solutions of ethanol and embedded in paraffin. The tissue specimens were sliced to a thickness of 5μm and stained with hematoxylin and eosin. All histological sections were observed by light microscopy (Leica DM 2500B, Germany).

L. amazonensis LV78 promastigotes were grown to stationary phase in SDM and were collected and washed three times in PBS and then resuspended in PBS at 10⁹/ml. The cells were subjected to 10cycles of rapid freezing (−70°C) and thawing at (37°C) for lysis and centrifuged at 12,000rpm at 4°C for 10min, and the supernatants were then collected and passed through a 0.20-μl filter assembly. The protein concentration was determined by a BCA kit. SLA was stored at −70°C until use.

At 32days postinfection, immune cells from Leishmania-infected WT and Nos2^−/− SD rats were isolated for analysis. PBMCs were isolated by Ficoll-Hypaque (Tianjin Hao Yang Biological Manufacture, Tianjin, China) density gradient centrifugation. Briefly, cells were suspended in Hank’s buffer and gently added on top of the Ficoll solution, and the ratio of cell suspension to Ficoll was 4:3 and centrifuged at a speed of 2,200rpm for 20min. The cells in the second upper layer were harvested for mononuclear leukocytes and washed twice with D-Hanks. Spleen and lymph nodes cells were separated by traditional methods. Briefly, the spleens and mesenteric lymph nodes from rats were mashed to release the cells and they were then passed through a 70μm cell strainer to obtain single cells. Red blood cells were removed using ammonium chloride as a lysing reagent, and immune cells were then washed with D-Hanks buffer and resuspended in RPMI-1640 supplemented with 10% FBS. Cells were counted and seeded into each well of a 96-well plate with 2× 10⁵ cell/ml. Cells were stimulated with SLA (25μg/ml) for 48h. IFN-γ and IL-4 production in the supernatants were determined by ELISA.

Supernatant from SLA-stimulated cells were assayed for IFN-γ and IL-4 with rat ELISA kits (BD Biosciences). Absorbances were read at 405nm using a microtiter plate reader. Concentrations were determined using standard curves for each cytokine. The detection limits of IFN-γ and IL-4 assay kits were 31.25 and 1.6pg./ml, respectively.

Rat serum was collected at 16, 48, and 64days postinfection, and cytokines were analyzed by Multiplex Luminex using ProcartaPlex Rat Th Complete Panel (ProcartaPlex, eBioscience), according to the manufacturer’s instructions.

Bone marrow-derived macrophages (BMDMs) were generated as previously described with modification (Boltz-Nitulescu et al., 1987). Briefly, bone marrow cells were obtained by flushing femurs of adult SD rats with D-Hanks buffer, cells were then dispersed by passing through a needle, and erythrocytes were removed by lysis with 0.015M Tris-NH₄Cl (pH 7.4). The cells were then washed and resuspended in DMEM supplemented with 10% FBS and 30% L929 cell supernatant. The cells were cultured in a 75cm² TC flask for 3days with a change in fresh medium for another 4days. Then, the BMDMs were harvested by digestion with 0.5% trypsin containing 0.03% EDTA. The harvested BMDMs were washed 3 times in PBS and adjusted to 10⁷ cells/ml in PBS. The cells were i.v. injected into the recipient Nos2^−/− rat at 0, 8, 16, and 24 DPI.

Statistical differences between groups were analyzed with Student’s t test using a built-in module of GraphPad Prism v5.0 (GraphPad software), using p<0.05 as the level of significance.

To investigate the mechanism of rat resistance against Leishmania infection, macrophages were targeted since they play a crucial role in the host innate immune response and are the major cell type inhabited by Leishmania during infection (Rittig and Bogdan, 2000). Naïve peritoneal macrophages were isolated from mice and Sprague Dawley (SD) rats and infected with L. amazonensis in vitro. In accordance with previous studies (Giannini, 1985; Soong et al., 1997), macrophages from both BALB/c and C57BL/6 mice are susceptible to L. amazonensis infection as indicated by unrestrained intracellular parasite replication during infection and by NO being absent in the culture supernatant (Figures 1A,D). In contrast, macrophages from SD rats showed restrained L. amazonensis growth, with intracellular amastigotes remaining low and only showing a slight increase at 72h postinfection (HPI). A considerable amount of NO was detected in the culture supernatant (Figures 1A,D). Nitric oxide (NO), produced by NOS2, has been shown to be the key antimicrobial agent found in macrophages, and its induction is critical in the control of many intracellular pathogens (Li et al., 2012). Therefore, we tested whether NO is important for conferring resistance to Leishmania infection in rat macrophages by inhibition of NOS2 activity with a specific inhibitor aminoguanidine (AG; Griffiths et al., 1993). When NOS2 was inhibited by AG, NO was not detected and parasite replication increased significantly, suggesting that NOS2 activity is indispensable in resistance (Figures 1A,D). To further dissect the role of NOS2, we generated an Nos2 gene knockout SD (Nos2^−/−) rat using transcription activator-like effector nuclease technology (TALENs; Tesson et al., 2011). TALENs resulted in a two-nucleotide deletion in the open reading frame of the Nos2 gene (Supplementary Figure 1A). Disruption of the Nos2 gene was demonstrated by the absence of NOS2 proteins in naïve or activated peritoneal macrophages and the absence of NO production in the culture supernatant of the activated macrophages (Supplementary Figures 1B,C), while arginase activity showed no significant difference between the wild-type and Nos2^−/− macrophages (Figure 1F). Macrophages from Nos2^−/− rats showed a significant increase in parasite burden compared to the macrophages from the wild-type rats at 72h postinfection (HPI; Figure 1C). More importantly, they failed to eliminate amastigotes even in the presence of IFN-γ, unlike the macrophages from the wild-type rats (Figures 1B,C). These observations clearly demonstrate that the parasite growth in NOS2 inhibited wild-type or Nos2^−/− macrophages was caused by the lack of NO production (Figure 1). The slight reduction in the parasite burden in the Nos2^−/− rat macrophages after IFN-γ activation indicates that NO-independent killing of L. amazonensis might exist. However, it must play a minor role, since most of the parasites remain intact, indicating that NO is the predominant leishmanicidal effector in rat peritoneal macrophages.

The above results demonstrate the critical role of NOS2 for rat macrophage resistance against L. amazonensis infection in vitro. We employed a rat ear cutaneous infection model to further investigate the role of NOS2 in vivo. Wild-type (WT), heterozygous (Het), and Nos2^−/− SD rats (Nos2^−/−) were subcutaneously infected with 10⁷ stationary phase promastigotes of L. amazonensis. Interestingly, the results showed that all the Nos2^−/− SD rats were highly susceptible to L. amazonensis infection compared to their wild-type SD and Het counterparts. In all Nos2^−/− rats, lesions began to appear at 8days postinfection (DPI). Afterward, the lesions became progressively worse as shown by an increase in thickness and diameter. The increase in lesion pathology (thickness and diameter) occurred rapidly from 8 to 40 DPI, then levelled off, and did not show any sign of recovery (Figures 2A,B). In contrast to mouse strains and Nos2^−/− rats, the resistance to infection in wild-type SD rats was demonstrated by observing only a temporary mild swelling or even absence of swelling, in some animals, during the course of the experiment. At the end of the experiment (56–64 DPI), no lesions were found in any WT rats. Het rats developed moderate lesions at the early stages of infection but were also able to heal afterward. Hematoxylin-Eosin (HE)-stained sections of ear lesions at 32 DPI revealed that, in the Nos2^−/− rat, the cutaneous tissue in the thickened ear was dominated by an increased infiltration of inflammatory cells. However, unlike observations in mice (Vasconcellos and Sotto, 1997), granulomas were not observed at the site of the infection in rats even though there was extensive cell infiltration. Furthermore, the general structure of the ear was not affected. Some recruited cells in the Nos2^−/− rats harbored numerous amastigotes, indicating their inability to kill parasites (Figures 2D,E,G). In sharp contrast, the structure of the dermal tissue from the ear of wild-type rats remained either normal at 32 DPI or with a few cells infiltrating if mild swelling was present. However, we were unable to observe any amastigotes in the sections of WT ears, indicating that most of the injected parasites were eliminated (Figures 2D,F). The parasite burden of infected ears was estimated at 32 and 64 DPI, by a limiting dilution assay. The parasite load in lesions was much higher in Nos2^−/− rats than in the wild-type and Het rats. At 32 DPI, parasite load in the Nos2^−/− rats was almost 330-fold greater than that observed in WT rats. Furthermore, the parasite burden declined rapidly to an undetectable point in WT rats at 64 DPI while it persisted in Nos2^−/− rats. The parasite load in the Het rats was higher than that observed for WT rats at 32 DPI (p=0.027), but they were almost eliminated at 64 DPI (Figure 2C). The above results demonstrate that NOS2 was also indispensable for rat innate resistance to L. amazonensis in vivo.

The above results have shown that NOS2 plays a pivotal role in rat resistance against Leishmania infection, both in vitro and in vivo. Since NO has been shown to participate in both antimicrobial activity and immune regulation in mice (Taylor-Robinson et al., 1994; Diefenbach et al., 1999), it is important to know whether rats deficient in NOS2 are capable of mounting an effective immune response toward Leishmania. Naïve Nos2^−/− rats did not show any observed abnormalities in the differentiation of CD4⁺ or CD8⁺ T-cell subsets (Supplementary Figure 2), as has previously been reported for NOS2-deficient mice (Wei et al., 1995). Both wild-type and Nos2^−/− groups of rats were each subcutaneously infected with 10⁷l. amazonensis via the dermal tissue in the ear. Then, peripheral blood mononuclear cells (PBMCs), splenocytes, and mesenteric lymph node (LN) cells were isolated 32 DPI and stimulated with soluble Leishmania antigens (SLA). Culture supernatants were assayed for IFN-γ and IL-4 production by ELISA. Splenocytes and LN cells from both WT and Nos2^−/−-infected rats produced comparable amounts of IFN-γ, while a significantly higher concentration of IFN-γ in PBMCs was detected in the Nos2^−/− group than that in the WT group (p=0.0014; Figure 3), indicating that Nos2^−/− rats are able to mount a strong Leishmania-specific immune response. The higher level of production of IFN-γ in Nos2^−/− rats is presumably due to the persistence of greater numbers of parasites in the ear dermis, thus eliciting more parasite antigen to be recognized by the immune cells. IL-4 was also detected in both Nos2^−/− and WT rat cells.

Serum cytokines profiles were measured to better understand the immune status in these rats. Inflammatory cytokines were significantly higher in Nos2^−/− rats but low to undetectable in WT rats (Figure 4). IFN-γ, TNF-α, IL-12, IL-1β, and IL-6 are important pro-inflammatory mediators during tissue inflammation. An increase in these cytokines was observed in the Nos2^−/− rats at all the time points tested, indicating that tissue inflammation persisted and was intense. Furthermore, this further supports the evidence that the Nos2^−/− rats were capable of mounting a specific immune response to Leishmania.

The above results provided strong evidence for the important role of NOS2 against Leishmania infection in rats. Since the immune regulatory network in Nos2^−/− rats is unimpaired, we therefore sought to investigate the role of macrophages in the rat model by adoptive transfer of BMDM from WT rats to the Nos2^−/− rats. In the Nos2^−/− rats that received WT BMDMs, the onset of the lesion development was delayed and shifted from 8 DPI to 16 DPI (Figure 5). Meanwhile, both the lesion thickness and diameter were significantly lower than in the untreated group prior to 20 DPI. However, lesions in the treated group developed progressively at a later time point (16–32 DPI). The lesion thickness increased to the same level as the untreated group but the diameter was still lower at 32 DPI. Nevertheless, the Nos2^−/− rats receiving wild-type BMDMs harbored significantly less parasites in lesions compared with the untreated group. These results provided further evidence that macrophages with NOS2 function play an important role in the control of parasites in vivo.