The NLRP3 inflammasome activation has been reported to have a protective or detrimental role during infection with different Leishmania spp., depending on the spp. or strain [16,35]. To determine whether NLRP3 activation has an impact on L. mexicana pathology, we infected Nlrp3^-/- and C57BL/6 wild-type (WT) control mice intradermally (i.d.) with 10⁶ L. mexicana metacyclic promastigotes and observed lesion development by assessing the score, which includes the inflammation as well as the length and width of the lesion [36]. Interestingly, no significant difference in lesion score was observed 5 weeks post infection (p.i.), when differences between susceptible and resistant mice usually appear [37]. No difference between both genotypes was observed up to 18 weeks p.i., with both groups exhibiting similar chronic, non-progressive lesion development, with no significant differences in lesion score, lesion size (Fig 1A–1C), and in parasite load at the infection site as determined by limited dilution analysis (LDA) (Fig 1D). At the end of the experiment, immune cell infiltration at the site of infection, including neutrophils, monocytes, and dendritic cells was analyzed by flow cytometry as shown in the gating strategy (S1A Fig). The number and frequency of cells was similar between the two genotypes (Figs 1E and S1B) except for a small difference in dendritic cell frequency that was not statistically significant in other independent experiments. These data indicate that the NLRP3 inflammasome does not play a significant role in L. mexicana-induced pathology.

Non-canonical inflammasome activation through the direct recognition of Leishmania by caspase-11 was previously reported for L. amazonensis [19]. We thus assessed whether L. mexicana infection could also activate caspase-11. We infected Casp-11^-/- mice with L. mexicana and compared lesion evolution and pathology with that of C57BL/6 mice. Similar lesion development and parasite burden were observed in both groups (S2A–S2C Fig). At the end of the experiment, there was no difference in the number and frequency of CD45⁺ cells (S2D Fig) and in neutrophil, monocyte, and dendritic cell number between Casp-11^-/- and C57BL/6 mice (S2E Fig). These data show that L. mexicana-induced pathology is not influenced by non-canonical inflammasome activation.

NLRP3, as well as other inflammasome sensors lacking a CARD domain, rely on the adaptor protein ASC for the recruitment of caspase-1. We thus assessed if other ASC-dependent inflammasomes could be triggered by L. mexicana. To this end, we infected Asc^-/- mice i.d., and measured lesion development for several weeks. Asc^-/- mice displayed a comparable lesion score and size to C57BL/6 controls (S2F–S2G Fig). Both genotypes showed similar parasite burden at the site of infection (S2H Fig). Additionally, the number and frequency of CD45⁺ cells, including neutrophils, monocytes, and dendritic cells present was similar between both genotypes as analyzed by flow cytometry (S2I and S2J Fig). These findings suggest a negligible role of ASC-dependent inflammasomes during L. mexicana infection.

The NLRC4 inflammasome can function in absence of ASC, through direct interaction of its CARD domain with the caspase-1 CARD domain [38]. To assess a putative role of this inflammasome, Nlrc4^-/- mice were infected with L. mexicana, and disease development was evaluated. Mice genetically deficient for the NLRC4 inflammasome developed similar lesion score, size, and parasite burden as control C57BL/6 mice (S2K–S2M Fig). As previously observed in infected Nlrp3^-/- and Asc^-/- mice, no difference in the number and frequency of immune cell infiltration at the site of infection was detected compared to control C57BL/6 mice (S2N and S2O Fig). Together, these results rule out a role for the activation of classical ASC-dependent inflammasomes, the non-canonical inflammasome, and NLRC4 during L. mexicana infection.

The NLRP1 inflammasome contains a CARD domain and can bypass the absolute requirement of ASC to activate caspase-1 [39,40]. Moreover, NLRP1 activation by Toxoplasma gondii, another protozoan parasite, has been reported [22,41]. In C57BL/6 mice, three tandem paralogs (Nlrp1a, Nlrp1b and Nlrp1c) encode for NLRP1. We thus infected Nlrp1^-/- mice, which are deficient in all three forms, i.d. with L. mexicana and measured lesion development. Five weeks post-infection, Nlrp1^-/- mice developed a significantly larger lesion score than control C57BL/6 mice (Fig 1F). This difference increased thereafter, as shown in representative pictures and lesion size at 8 weeks p.i. (Fig 1G and 1H). Nlrp1^-/- mice displayed significantly higher parasite burden at the site of infection compared to C57BL/6 mice (Fig 1I). Of interest, there was a significant increase in neutrophil number at the lesion site of Nlrp1^-/- mice compared to control mice, with similar number and frequency of CD45⁺ cells, monocytes and dendritic cells (Figs 1J and S3A).

To explore the potential impact of NLRP1 on adaptive immunity, we analyzed by flow cytometry the T helper cytokine profile in the draining lymph nodes (dLNs) and infected ears 8 weeks p.i. No difference was observed in the frequency of CD4⁺ IFN-γ⁺ T cells, CD8⁺ IFN-γ⁺ T cells or CD4⁺ IL-4⁺ T cells in dLNs and at the infection site between Nlrp1^-/- mice and control mice (S3B and S3C Fig). Collectively, these data show that NLRP1 protects against lesion exacerbation and contributes to parasite control but does not influence the development of an adaptive immune response.

NLRP1 activation causes downstream caspase-1 clustering and autoproteolytic activation. To determine if the observed phenotype was due to a classical NLRP1-Caspase-1 axis, we infected mice genetically deficient in caspase-1 (Casp-1^-/-) with L. mexicana. After 5 weeks of infection, Casp1^-/- mice exhibited a significantly higher lesion score, lesion size, and parasite burden than control C57BL/6 mice, and this difference continued to increase until 7 weeks p.i. (Fig 2A–2D). The number and frequency of CD45⁺ cells were slightly increased with a neutrophil number detected at the site of infection that was significantly more elevated in Casp1^-/- mice compared to control mice, while that of monocytes, and dendritic cells was not significantly increased (Figs 2E and S4A). At the same timepoint, caspase-1 did not impact the development of the adaptive immune response, as no difference was observed in the frequency of CD4⁺IFN-γ⁺ T cells, CD8⁺IFN-γ⁺ T cells and CD4⁺IL-4⁺ T cells in the dLNs and at the site of infection of Casp1^-/- and C57BL/6 control mice (S4B and S4C Fig).

Gasdermin D (GSDMD) is a key substrate for caspase-1 and a central player in pyroptotic cell death. Neutrophil cell death mechanisms, like NETosis, have been shown to involve GSDMD [10]. Additionally, GSDMD has recently been shown to protect against Leishmania lesion aggravation [20]. Following long-term infection with L. mexicana, Gsdmd^-/- mice also displayed exacerbated lesions (Fig 2F–2H) and an increased parasite burden (Fig 2I). Furthermore, a threefold increase in neutrophil number was detected at the site of infection, together with increased CD45⁺cells and neutrophil frequency (Figs 2J and S4D). The frequency of CD4⁺IFN-γ⁺ T cells, CD8⁺IFN-γ⁺ T cells, and CD4⁺IL-4⁺ T cells in the dLNs and at the site of infection of Gsdmd^-/- and C57BL/6 control mice (S4E and S4F Fig) remained unchanged, as determined by flow cytometry. The exacerbated pathology observed following infection of Gsdmd^-/- mice with a high dose (10⁶ metacyclic promastigote parasites) of L. mexicana was confirmed using a lower infection dose (10⁴ metacyclic promastigote parasites), which better mimics the inoculum deposited by a sandfly during a natural infection [42]. Similar to our observations following a high-dose infection, the lesion score, size, and the parasite burden were elevated in Gsdmd^-/- compared to control mice (S5A–S5D Fig).

GSDMD pores can facilitate the release of IL-1β through the cell membrane. The role of GSDMD in macrophage pyroptosis is well-defined, but its involvement during infection with Leishmania has only recently been described [20]. To assess if L. mexicana infection triggers inflammasome activation in macrophages, we quantified IL-1β release in vitro by ELISA. IL-1β release was detected 16 hours p.i. in bone marrow-derived macrophages (BMDMs) in a dose-dependent manner, and this secretion was decreased in the absence of caspase-1 and GSDMD, but not in the absence of NLRP1 (S6 Fig). These findings suggest that in vitro, IL-1β release in macrophages is independent of NLRP1. Consequently, the NLRP1-dependent phenotype observed following infection likely results from NLRP1 activation in other cell types.

Neutrophils play a pivotal role during L. mexicana infection and early neutrophil presence can negatively impact disease progression, preventing the parasite from exploiting the neutrophil niche for lesion establishment [27]. To assess whether L. mexicana was able to induce inflammasome activation in neutrophils, we infected LPS-primed C57BL/6 bone marrow-derived neutrophils (BMNs) with L. mexicana promastigotes for 16 hours and analyzed GSDMD activation by Western Blot. We observed a dose-dependent cleavage of full-length GSDMD (p53) into the active p30 N-terminal part in cell lysates, suggesting inflammasome activity in infected neutrophils (Fig 3A). To investigate if GSDMD cleavage is a common process in other Leishmania spp., we infected LPS-primed C57BL/6 BMNs for 16 hours with New-World Leishmania spp., including L. mexicana, L. amazonensis, L. panamensis and L. guyanensis and Old-World Leishmania spp. and strains, including L. major LV39, L. major WR288, L. major Seidman and L. donovani (Fig 3B). Infection with L. mexicana induced GSDMD cleavage, contrary to the other New World spp. tested, except for L. panamensis, where GSDMD cleavage was observed at a lower level. Among the Old-World spp. tested, GSDMD activation in neutrophils was selectively induced by L. major Seidman, in line with the reported importance of the NLRP3 inflammasome and neutrophils in the pathology induced by this L. major strain [17]. Thus, the highest induction of GSDMD cleavage was observed following L. mexicana infection. To rule out possible variations due to long-time in vitro culture, we used the same L. mexicana strain after demonstrating its vector competency through the development of mature transmissible infections in sand flies (S7A Fig.). Infection with both L. mexicana strains resulted in similar GSDMD cleavage showing that the phenotype observed did not arise during in vitro culture of the parasites (S7B Fig.). We also analyzed GSDMD cleavage in neutrophils infected with another L. mexicana strain, TAB3, isolated from a patient with diffuse cutaneous lesions [43]. Similar GSDMD cleavage was observed in neutrophils infected with both strains (Fig 3C). In the mammalian host, infectious Leishmania promastigotes rapidly transform into amastigotes, its intracellular replicating form. To investigate if L. mexicana amastigotes would also induce neutrophil pyroptosis, axenic L. mexicana amastigotes were prepared, and LPS-primed neutrophils were infected at a MOI of 10. In parallel, neutrophils were similarly infected with L. mexicana promastigotes. We observed GSDMD cleavage following infection with both life stages, although much lower levels of GSDMD cleavage were observed after infection with amastigotes than promastigotes (Fig 3D). As a correlate of cell membrane permeability, uptake of propidium iodide (PI), a nucleic acid dye, was measured in response to promastigotes and axenic amastigotes at multiplicity of infection 5 (MOI 5). Both L. mexicana life stages induced PI uptake, but at a significantly lower level in response to amastigotes, in line with the lower GSDMD cleavage observed (Fig 3E). Furthermore, infection with amastigotes also induced IL-1β release by neutrophils, again at lower levels than those infected by promastigote forms of the parasites (Fig 3F). Collectively our data reveal that both forms of L. mexicana induce neutrophil pyroptosis, with a higher induction by the early promastigote stage.

Our in vivo data demonstrated a role for the NLRP1 inflammasome in the control of infection. To investigate whether NLRP1 activation in neutrophils was responsible for L. mexicana-induced GSDMD cleavage, we infected LPS-primed Nlrp1^-/- and C57BL/6 BMNs with L. mexicana for 16 hours and analyzed GSDMD cleavage by Western Blot. A strong reduction of GSDMD cleavage was detected in Nlrp1^-/- neutrophils (Fig 4A), indicating that NLRP1 is upstream of GSDMD activation during neutrophil infection, in line with our in vivo results. In addition, we examined if GSDMD cleavage was mediated by caspase-1. We infected LPS-primed Casp1^-/- and Casp11^-/- BMNs and analyzed GSDMD cleavage by Western Blot. We could not detect any cleaved p30 N-terminal GSDMD in Casp1^-/- neutrophils, while GSDMD cleavage was detected in Casp11^-/- neutrophils (Fig 4B). Altogether, these data strongly support the activation of a NLRP1-caspase-1-GSDMD axis in neutrophils upon L. mexicana infection. LPS-primed BMNs released IL-1β upon L. mexicana infection 16 hours p.i., a process abrogated in Gsdmd^-/- neutrophils (Fig 4C). In addition, to assess whether GSDMD activation in neutrophils led to pore formation in infected neutrophils, LPS-primed BMNs were infected with L. mexicana promastigotes at MOI of 5, and uptake of PI was measured over 24 hours as a correlate of cell permeability. L. mexicana-induced PI uptake was significantly reduced in Gsdmd^-/- neutrophils compared to C57BL/6 controls (Fig 4D). We then quantified lactate dehydrogenase (LDH) release at 16 hours p.i, to measure if pore formation led to cell lysis and pyroptosis. LDH release was lower in Gsdmd^-/- compared to C57BL/6 infected neutrophils, showing GSDMD-dependent LDH release (Fig 4E). These results indicate that L. mexicana induces GSDMD-dependent, lytic pyroptosis. Of note, no difference in apoptosis was observed between L. mexicana-infected Gsdmd^-/- and control neutrophils, as analyzed by Annexin-V / DAPI staining. (S8A and S8B Fig). These findings suggest that infection of neutrophils with L. mexicana leads to GSDMD cleavage and pyroptosis, with no impact on non-lytic, apoptotic cell death. Collectively our data show that L. mexicana induces GSDMD activation in neutrophils, leading to pyroptotic cell death.

Human NLRP1 differs from murine NLRP1 as far as the number of paralogs coding for NLRP1 (3 in C57BL/6 mice and 1 in human), requirement of ASC (dispensable in mNLRP1b) and cell expression, with human NLRP1 reported to be mostly expressed in epithelial cells and keratinocytes, while murine NLRP1 is mainly expressed in myeloid cells [44]. To assess if L. mexicana-also induces pyroptosis in human neutrophils, we isolated peripheral blood neutrophils from healthy human donors and infected them with L. mexicana. We first assessed inflammasome-dependent cell permeability by measuring PI uptake. Neutrophils were primed with LPS and exposed to L. mexicana, or to L. mexicana and the caspase-1 inhibitor ac-YVAD-cmk. PI uptake was observed following infection, a process that was markedly reduced in presence of caspase-1 inhibitor (Fig 4F), revealing inflammasome activation in L. mexicana-infected human neutrophils. Furthermore, the NLRP1 activator Val-boroPro similarly induced PI uptake in neutrophils, a process that was also reduced upon inhibition of caspase-1 (Fig 4G), suggesting that NLRP1 can also be activated in human neutrophils. In line with these results, induction of IL-1β release by human neutrophils was observed in response to both L. mexicana or Val-boroPro, and secretion was decreased in presence of the caspase-1 inhibitor (Fig 4H). Collectively, these data reveal that the parasites can trigger pyroptosis in both murine and human neutrophils and suggest that NLRP1 can be triggered in human neutrophils.

GSDMD has been implicated in neutrophil extracellular trap (NET) formation [10,45]. We thus assessed dsDNA release in vitro through Picogreen dsDNA detection and detected a small, but statistically significant GSDMD-dependent increase in extracellular dsDNA release in response to L. mexicana (S9A Fig). To further evaluate the impact of GSDMD on L. mexicana-induced NET formation, Gsdmd^-/-, Casp1^-/- and C57BL/6 bone marrow-derived neutrophils and human neutrophils isolated from peripheral blood of healthy donors were infected with L. mexicana. The formation of NETs was visualized and quantified by confocal microscopy following DAPI and MPO staining. L. mexicana induced NET formation in neutrophils from all mouse genotypes and in human neutrophils (S9B and S9C Fig). No statistically significant difference was observed in the frequency of NETs between L. mexicana-infected Gsdmd^-/-, Casp1^-/- and C57BL/6 neutrophils (S9B Fig). These data suggest that the small decrease observed in dsDNA release in Gsdmd^-/- neutrophils likely comes from reduced pyroptosis.

GSDMD had no major impact on other neutrophil functions, as we did not observe changes in radical oxygen species (ROS) production levels in Gsdmd^-/- compared to C57BL/6 BMNs 2 hours p.i (S10A and S10B Fig). Furthermore, no impact of GSDMD on L. mexicana internalization at 2 hours p.i. (S10C Fig) was observed, with a small impact on infection rate observed 24 hours p.i. Additionally, we did not observe an impact of GSDMD on neutrophil activation status (S10D Fig) as compared to C57BL/6 control cells.

L. mexicana has evolved ways to exploit neutrophils as a niche to establish initial infection and interfere with a protective immune response [27,46]. To assess if neutrophil recruitment was impacted by GSDMD in vivo, we monitored neutrophil frequency in the blood, BM, and at the site of infection (Fig 5B). At 24 hours post-Leishmania infection, neutrophil frequency in the BM and the blood was similar in C57BL/6 and Gsdmd^-/- mice, indicating that GSDMD has no impact on the initial neutrophil recruitment at this timepoint (Fig 5B). A small but not significant increase in neutrophil frequency was detected at the site of infection (Fig 5B). To determine neutrophil dynamics at the site of infection, we assessed the number of neutrophils at 24, 48 and 72 hours p.i. (Fig 5C). As previously noted, the number of neutrophils was comparable in Gsdmd^-/- and C57BL/6 mice at 24 hours p.i., and decreased in both genotypes at 48 hours p.i., but at a significantly slower rate in Gsdmd^-/- compared to C57BL/6-infected ears. By 72 hours p.i., neutrophil numbers returned to basal levels in both groups (Fig 5C), correlating with the frequency of cells (S11A Fig). Over the same time course, we analyzed the corresponding number of infected neutrophils using dsRed-expressing L. mexicana parasites. A significantly higher number of infected neutrophils was detected at 48 hours and 72 hours post-infection in Gsdmd^-/- mice compared to the control group (Figs 5D and S11B). This increase in infected neutrophils correlated with an increased parasite burden observed in the infected ear 48 hours post infection (Fig 5E), as determined by limiting dilution analysis (LDA). Altogether, these data show that GSDMD acts as a negative regulator of neutrophil survival during L. mexicana infection, thereby reducing the neutrophil niche exploited by the parasite.

Following infection with L. mexicana, neutrophils are rapidly recruited to the site of infection and return to basal levels by 3 days p.i. A second wave of neutrophils arises around 3 weeks post infection [37], a time at which the size of the inflammatory lesion between Gsdmd^-/- and C57BL/6 mice does not differ significantly yet (Figs 2F and 7A), and only the amastigote form is present. The parasite burden measured by LDA in the infected Gsdmd^-/- ears was higher than control C57BL/6 mice (Fig 5F). These data suggest that L. mexicana-induced pyroptosis also contributes to the regulation of parasite load during the second wave of neutrophil recruitment.

To strengthen these findings, we conducted an adoptive transfer experiment in neutropenic Genista mice [47], designed to assess neutrophil survival at the site of infection, in absence of neutrophil recruitment. Gsdmd^-/- and C57BL/6 BMNs were isolated and stained with distinct cell tracker dyes and co-injected with L. mexicana into the ear dermis of neutropenic Genista mice. The frequency of Gsdmd^-/- and C57BL/6 neutrophils was analyzed 24 hours later, by flow cytometry (Fig 6A). Genista mice showed no mature neutrophils as expected, in the blood (Fig 6B) and in the ear (Fig 6C). Following adoptive transfer at the time of infection, a higher frequency of Gsdmd^-/- compared to C57BL/6 neutrophils was observed 24h later in the infected ear (Fig 6D and 6E), indicating prolonged survival of GSDMD-deficient neutrophils. As we showed that caspase-1 and NLRP1 are upstream of GSDMD cleavage in L. mexicana-infected neutrophils in vitro, we performed similar adoptive transfer experiments with a mix of C57BL/6 neutrophils and either Casp1^-/- or Nlrp1^-/- neutrophils. Again, a higher frequency of either Casp1^-/- and Nlrp1^-/- neutrophils was observed in the infected ear dermis 24 hours p.i. (Fig 6F and 6G). Together, these results confirm that during L. mexicana infection in vivo, NLRP1-dependent caspase-1 activation of GSDMD, and subsequent pyroptosis control the neutrophil pool at the site of infection.

To confirm that the increased pathology observed in Gsdmd^-/- mice was driven by neutrophils, we generated Mrp8^creGsdmd^lox/lox mice, hereafter referred to as Gsdmd^ΔPMN mice, that have a neutrophil-specific deletion of GSDMD. We then infected Gsdmd^ΔPMN, Gsdmd^-/- and control Mrp8^WTGsdmd^lox/lox mice with L. mexicana and compared lesion and pathology development. Both Gsdmd^ΔPMN and Gsdmd^-/- mice exhibited similar, significantly increased lesion score, size, and exacerbated pathology as compared to control mice (Fig 7A–7C). Similarly, enhanced parasite burden was observed at the site of infection in both Gsdmd^ΔPMN and Gsdmd^-/- mice, as compared to control mice (Fig 7D). We visualized the higher parasite load observed in Gsdmd^ΔPMN mice by immunofluorescence of infected ears 8 weeks p.i. and detected a higher density of neutrophils in infected ears (Figs 7E, 7F and S12A). Locally, higher levels of CD45⁺ cells were observed by flow cytometry in both Gsdmd^ΔPMN and Gsdmd^-/- infected ears (Fig 7G). Additionally, we observed an increase in neutrophil number and frequency, as well as an increase in dendritic cell and monocytes numbers and frequency in Gsdmd^ΔPMN compared to C57BL/6 infected ears (Figs 7G and S12B). The immune response was similar in the three genotypes with comparable frequency of CD4⁺IFNγ⁺, CD4⁺ IL-4⁺ and CD4⁺IL-17⁺ T cells observed in dLN of cells 11 weeks and 21 days p.i. (S12C and S12D Fig). Furthermore, at 21 days p.i. the levels of arginase and iNOS in dermal macrophages, as detected by flow cytometry of infected ear cells, were similar in the three genotypes (S12E Fig). These data suggest that the increased number of infected neutrophils observed in absence of GSDMD does not result from impaired microbicidal activity of macrophages but from reduced neutrophil pyroptosis. These results highlight the protective role of the NLRP1-caspase-1-GSDMD axis in neutrophils during L. mexicana infection and subsequent skin pathology.

To further elucidate the role of pyroptosis in the early wave of neutrophils in the long-term pathology, we injected either Gsdmd^-/- or C57BL/6 neutrophils in Genista neutropenic mice at the time of infection and monitored lesion development over 10 weeks (Fig 7H and 7I). Genista mice infected with L. mexicana are able to heal their lesion and clear their parasites, revealing a deleterious role for neutrophils in this infection. Furthermore, healing was linked to the absence of neutrophils during the first days of infection, as a similar healing phenotype was observed following depletion of neutrophils at the onset of infection in C57BL/6 mice [27]. Here, conversely, the transient adoptive transfer of C57BL/6 neutrophils in Genista mice early in infection was sufficient to restore the nonhealing phenotype observed in WT C57BL/6 infected mice. Remarkably, mice that received Gsdmd^-/- neutrophils at the onset of infection developed a significantly larger inflammatory lesion with elevated parasite burden compared to mice that received C57BL/6 neutrophils (Fig 7J and 7K). No difference was observed in CD4⁺IFNγ⁺, CD4⁺IL-4⁺and CD8⁺IFNγ⁺ T cells between Genista mice transferred with C57BL/6 or Gsdmd^-/- neutrophils (S12F Fig). These results show that GSDMD-induced neutrophil pyroptosis plays a crucial role at the onset of infection and underline the significance of neutrophils in shaping the outcome of the disease, as summarized in Fig 8.

Healthy, informed and consenting human blood donors participated in this study. Verbal consent was obtained from the donors. All procedures were approved by the Ethical Committee of the Canton of Vaud (number: CER-VD2017-00182 to FTC) and conducted in accordance with the legislation of the Canton of Vaud, the Swiss Confederation and the Declaration of Helsinki. Animal experimental protocols were approved by the veterinary office regulations of the Canton of Vaud, Switzerland, under authorization number 3476.d to FTC and performed in compliance with Swiss laws for animal protection and the principles of the declaration of Basel.

In this study, we investigated the effect of NLRP1-induced GSDMD activation in neutrophils during L. mexicana. Sample size of all assays was based on pilot experiments conducted in our laboratory, to ensure reliable data for statistical analyses and reproducibility. Each experiment was independently replicated at least two to three times, with the number of individual replicates specified in the figure legends. Animals were randomly assigned to different experimental groups.

C57BL/6 mice were obtained from Envigo (Cambridgeshire, United Kingdom) or bred in our Epalinges facility. Nlrp1^-/- mice were kindly provided by S. Masters, the Walter and Eliza Hall Institute of Medical Research (Parkville 3052, Australia). These mice are homozygous for the deletion of the entire Nlrp1 locus (Nlrp1a, Nlrp1b and Nlrp1c) [80]. Casp1 ^-/- mice [81] were kindly provided by O. Gross (Institute of Neuropathology, University of Freiburg). Genista neutropenic mice were generated by Bernard Malissen [47]. Nlrp3^-/- [82], Nlrc4^-/-, Asc^-/- [83], Gsdmd^-/- [5] and Casp11^-/- mice have been described previously. All mouse strains used in this study were either backcrossed to or generated on a C57BL/6 background. All mice were bred and housed in specific pathogen-free facilities at the Epalinges Center. All experiments were performed using 6 to 10-week-old age- and sex-matched mice.

To generate conditional Gsdmd knockout (MRP8^cre Gsdmd^lox/lox) mice, loxP sites were inserted up and downstream of exon 2 of Gsdmd using the 2 following gRNAs (TCCACGGGTTCTATAGACGG’TGG and TCTACTACTCCACTCCTCTG’GGG). Injection of the gRNAs and Cas9 protein into C57BL/6 embryos was done as described before [84]. Biopsies for genotyping were taken at an age of 10–12 days. DNA extraction was performed using the KAPA HotStart Mouse Genotyping Kit according to the manufacturer’s protocol. Genotyping PCR was done using Q5 Polymerase (NEB), carrying out 3 PCR reactions with 2 primer sets: PCR-A (CGCTTCCCTTACCTTGAGCA, GTGTCTAGGTGGTTGTGGGG) covering the binding site of gRNA1, and PCR-B (CAGCCCTACTTGCTCTAGCC, AGCCAAAACACTCCGGTTCT) covering the binding site of gRNA2. The expected fragment sizes were 339 bp for PCR-A and 388 bp for PCR-B in animals harboring a WT allele, and 373 bp and 423 bp in mice harboring the loxP insertions. To obtain neutrophil-specific genetic deletion of GSDMD, GSDMD^lox/lox mice were crossed to MRP8^cre/WT cre-deleter mice (MRP8-Cre-ires/GFP, Jackson AX, stock #021614), deletion of exon 2 was verified by PCR using primers CAGCCCTACTTGCTCTAGCC and GTGTCTAGGTGGTTGTGGGG). Mrp8^{cre +}; GSDMD ^flox/WT offspring were selected and bred with GSDMD^lox/lox mice to generate GSDMD^lox/lox MRP8cre ^+/- mice. Breeding pairs were established with male MRP8^cre+/- and female GSDMD^lox/lox mice.

L. mexicana (WHO strain MYNC/BZ/62/M379) was used in this study. We thank Prof. Scott M. Landfear and Shaden Kamhawi for sharing with us the same strain cultured in their laboratory. L. mexicana DsRed [85] was kindly donated by Prof. Tony Aebischer (Robert Koch-Institute, Berlin). Both strains were used in all experiments unless indicated. Other used species and strains were L. mexicana MHOM/ MX/00/Tab3, isolated from a human patient with diffuse leishmaniasis, L. (v) panamensis MHOM/COL/86/1166LUC, L. amazonensis (LTB0016), L. guyanensis (MHOM/BR/75/M4147), L. major LV39, L. major Seidman [86] and L. infantum (MHOM/CH/2016/BELA). L. donovani (MHOM/SD/62/1S) and L. major (WR 2885) were donated by Dr. Shaden Kamhawi. Parasites were cultured at 26°C in M199 medium (Gibco), supplemented with 10% fetal calf serum (FCS), 4% HEPES, and 2% PSN (Penicillin, Streptomycin and Neomycin, Bioconcept, Allschwill). Transgenic DsRed parasites were grown under the same conditions, with added 50 μg/ mL Hygromycin B (Sigma-Aldrich). Metacyclic promastigote parasites were selected by density centrifugation, as previously described [87]. Axenic amastigotes were isolated as previously described [46].

10⁶ metacyclic promastigotes were injected intradermally (i.d.) using a 20G needle. In selected cases, a lower dose (10⁴ metacyclic parasites) was injected. Lesion development was monitored biweekly using a caliper, and lesion score was calculated as previously described [36]. In brief, this scoring system considers early inflammation as well as late necrosis, with naïve ears scoring 0, and early inflammation scoring 0.5. Once the lesion outline is clear, lesion size is determined using a caliper. Necrosis in the ear lesion will lead to a maximum score of 8 and immediate sacrifice of the animal.

To determine the population of living parasites, single-cell suspensions obtained from infected ears were subjected to serial dilution in supplemented M199 media as previously described [37]. Detection of parasites in each well was performed using microscopy and the quantification of parasite number was carried out using the ESTIMFRE software, following established procedures.

Ears were collected at the specified time points. The two dermal layers were separated, homogenized, and digested with 0 .2 mg/mL liberase TL (Roche) at 37°C for 2 hours. Digested ears were then filtered through 40 μm filters (Falcon). Draining lymph nodes were recovered and homogenized in DMEM (Gibco). Blood was obtained by cardiac puncture and mixed with 1% EDTA in PBS. Blood, bone marrow, and spleen underwent erythrocyte lysis using ACK buffer.

Four to six days old Lutzomyia longipalpis (Jacobina colony) female sand flies reared at the Laboratory of Malaria and Vector Research, NIAID, NIH, were artificially fed as previously described [88]. Briefly, rabbit defibrinated blood (Noble Life Science) was centrifugated at 600xg for 10 min. The serum was separated, and complement was inactivated for 30 min at 56°C. Reconstituted blood was spiked with 5x10⁶/ mL Leishmania mexicana (MYNC/BZ/62/M379 strain) procyclic promastigotes passaged the day before sand fly infection in Schneider’s media (Gibco) supplemented with 20% fetal bovine serum (Thermo-fisher Scientific) and 1% penicillin/streptomycin (Gibco) and kept at 26°C. Sand flies were starved overnight prior to infection. Post-infection, sand flies were kept on 30% sucrose and maintained at 26°C with 75% humidity with 12 hours light/dark cycle. At day 10–14 post-infection, sand fly midguts (n = 5–10) were dissected in 30 μl of PBS (Lonza), macerated and counted using a hemocytometer. The total number of parasites and percent infectious metacyclic promastigotes per midgut were calculated in two independent experiments.

Whole blood was collected from healthy donors in heparinized tubes at the Interregional Blood Transfusion SRC (Epalinges, Switzerland). Neutrophils were isolated by centrifugation over a polymorphprep (Axis-Shield) gradient as previously described [89]. Purity of neutrophils (>95%) was assessed by cytospin and Diff-quick staining (RAL Diagnostics).

Cell surface antigens in single-cell suspensions were detected with the following rat anti-mouse antibodies: anti-CD45-PerCPCy5.5, anti-CD11c-PeCy7, anti-Ly6C-FITC, anti-Ly6G-PE, and anti-Ly6C-APC (all from BD Bioscience); anti-Ly6G-APC/Cy7 and anti-CD4-AF700 (Biolegend); anti-CD8-APC, anti-CD11b-PB, anti-IFNγ-PeCy7; anti-IFNγ-PE, anti-IL-4-FITC, anti-IL17-PercpCy5.5, anti-Annexin V-PeCy7 (eBioscience), anti-iNOS-AF700 and anti-Arginase1-BV421 (all from eBioscience) diluted in supernatant from 2.4G2 hybridoma cells. Viable cells were gated using Aqua Live/dead (Invitrogen). Parasite infection was determined based on the intrinsic DsRed fluorescence of transgenic parasites. Stained cells were acquired on an LSR-Fortessa (BD Bioscience) and analyzed using FlowJo software (Tree StarA).

For IL-1β detection, bone marrow-derived neutrophils and human neutrophils from peripheral blood were seeded at 0.5 x 10⁶ cells per well in a total volume of 200 μL and primed for 4h at 100 ng/mL LPS (mouse) or 1 hour at 500 ng/mL LPS (human) (LPS Ultrapure; Invivogen). After infection with L. mexicana as described previously, supernatants were collected and released mouse IL-1β (IL-1β ELISA kit, Invitrogen) or human IL-1β (human IL-1β/IL-1F2 ELISA DuoSet Kit, R&D) were analyzed according to the manufacturer’s instructions.

Femora and tibia of naïve mice were collected, and bone marrow was obtained by centrifugation. Neutrophils were purified by negative magnetic-activated cell sorting (MACS) for dsDNA release assays, to avoid unwanted activation and positive MACS for other experiments where higher purity and lower background apoptosis were preferred. Isolations were performed by using the appropriate anti-Ly6G MicroBeads UltraPure isolation kit, according to the manufacturer’s instructions (Miltenyi Biotec). Neutrophil purity (>95%) was verified flow cytometry. Neutrophils were seeded at 0.5 x 10⁶ cells / well in Optimem + 5% FCS (Gibco). For macrophage differentiation, bone marrow cells were cultured for 7 days in RPMI (10% heat-inactivated fetal bovine serum, 1% HEPES (Amimed) and 100U/ mL penicillin and 100μg/ mL streptomycin, Invitrogen) with 25ng/ mL M-CSF (Immunotools- 12343117) to generate BM-derived macrophages. LPS priming (LPS Ultrapure; Invivogen) was performed at 100 ng/ mL in OptiMEM (Gibco) + 5% FCS for 4h at 37°C when indicated.

Cell-membrane permeabilization was determined by measuring the uptake of Propidium Iodide (PI) (1 μg/mL; Thermo Fisher Scientific) over time using a fluorescent plate reader (Cytation5; Biotek) or an IncuCyte live-cell analysis system (Sartorius). Cell lysis was quantified by measuring LDH release into the supernatant (Cytotx kit, Promega). Both values of PI and LDH were calculated as a percentage of a 100% lysis control and normalized to the respective untreated sample.

Cells were lysed using lysis buffer containing protease inhibitor cocktail (Complete Ultra Tablets, Roche). Protein separation was performed using 12% gels and transferred onto nitrocellulose conform standard procedures. Antibodies for immunoblots were against GSDMD (EPR19828; Abcam; 1:1000) and beta-actin (Sigma; 1:5000).

Double-stranded DNA release was quantified in the supernatants by Picogreen detection as previously described [56]. Briefly, neutrophils were obtained through negative MACS selection and primed with GM-CSF for 20 min at 37°C, prior to infection with stationary-phase L. mexicana promastigotes. After 4 hours of infection, stimulation was stopped with the addition of DNAse I (20 μg/mL; Roche) and EDTA (2.5 mM). Supernatants were collected, transferred to 96-well plates (Perkin Elmer), and light-sensitive Picogreen dye was added. DsDNA content was then measured at 480nm excitation and 520 nm emission using a spectrophotometer (Molecular Devices, SpectraMax MiniMax 300).

Total reactive oxygen species (ROS) production by bone marrow-derived neutrophils was measured through a luminol-based chemiluminescence assay as previously described [56]. Briefly, BMNs were incubated in X-vivo (Lonza) medium in a white opaque 96-well plate (White Opaque 96-well Microplate, PerkinElmer) and infected with L. mexicana promastigotes. ROS production was measured by adding 20 μg/mL of luminol (Carbosynth) and measuring chemiluminescence at all wavelengths over time, using a Spectramax plate reader (Molecular Devices, SpectraMax MiniMax 300). To measure ROS by flow cytometry, BMNs were incubated with 1.2 μM DHR123 molecular probe (Thermo Fisher Scientific) for 45 min at 37°C.

BMNs from the indicated genotypes were stained with either CMRA (Cell tracker Orange, Life Technologies) or CMFDA (Cell tracker Green, Life Technologies), for 15 min at 37°C, in DMEM (Gibco) containing 1% FBS. For each replicate, dyes were switched to ensure no bias due to dye color. After staining, neutrophils were counted and mixed at an equal 1:1 ratio. At the time of infection, 10⁶ neutrophils were injected with 10⁶ metacyclic promastigotes into the ear dermis of Genista mice. Following infection in WT mice, neutrophils peak at the site of infection between 4 and 24 hours [27,89]. 24 hours after neutrophil adoptive transfer and co-infection, at a time that may correspond to 28 hours or more p.i. at the site of infection in WT C57BL/6 mice, mice were sacrificed, and ears and blood were processed as previously described.

OCT cryosections of infected ears were prepared and thawed for 30 minutes at room temperature before fixation for 10 minutes in 4% PFA. After 30 minutes in blocking buffer (0.5% BSA, 5% donkey serum, 0.3% Triton X-100, 0,1% NaN3), slides were incubated with primary Ly6G⁺ antibody (Biolegend) overnight at 4°C. Washing was performed for 30 minutes with 0.3% Triton X-100 in PBS. Secondary antibodies were diluted 1:500 in blocking buffer and staining was performed for 1 hour at RT. After washing, slides were mounted using Fluoromount-G, containing DAPI (Invitrogen). Images were acquired with an inverted confocal microscope Zeiss LSM 880 with Airyscan and processed in Imaris 9.0 (Bitplane). Total area of DAPI, Ly6G⁺ and parasite dsRed was quantified, and total area of Ly6G⁺ and parasite dsRed was normalized to DAPI in each field of view.

Statistical analyses were performed using Prism 8 (GraphPad) software. When data sets were normally distributed, parametric unpaired t-tests were used, if not, a non-parametric Mann-Whitney U test was used to analyze the data. Adoptive transfer of neutrophils of several genotypes was analyzed with Wilcoxon matched pairs signed rank test. When more than 2 groups were considered, a parametric 1-way ANOVA or non-parametric Kruskal-Wallis test was performed, comparing each group back to the control variable. Long-term mice infections, as well as kinetic PI absorbance, were analyzed using a 2-way ANOVA model with a time variable. P-values were considered significant at p ≤ 0.05.