All animal experimentation protocols described in this study were approved by the Swiss Federal Veterinary Office (SFVO), under authorization numbers VD2113.2 and VD3551. Animal handling and experimental procedures were undertaken with strict adherence to the ethical guidelines given by the SFVO and under inspection by the Department of Security and Environment of the State of Vaud, Switzerland.

Wild-type (WT) (C57BL/6JOlaHsd) mice were purchased from Envigo (Netherlands) and Nlrx1-deficient mice (Nlrx1^-/-) (B6.129-Nlrx1tm1 Tsc) were generated previously by replacing the first four coding exons with a neomycin cassette, that was later removed (38). The mice were previously backcrossed onto the C57BL/6JOlaHsd background for at least 5 generations. Mice were genotyped by PCR using tissue-isolated genomic DNA using the KAPA Mouse Genotyping Kit (KAPA Biosystems). Mice were maintained at the animal facility of the Center for Immunity and Immunology Lausanne (CIIL) (Switzerland) in a pathogen-free environment. Males and females (6-9 weeks old) were used for experiments. In vivo experiments were performed at a biosafety level 2 (BSL-2) animal facility at the CIIL. Cages were enriched with one igloo, two carboard tubes, one wood stick, and tissues. Experiments were performed after one week of acclimation in the BSL-2 animal facility. Food (SAFE or KLIBA NAGAF) and water were provided ad libitum. Light cycle was maintained at 13 hours light and 11 hours darkness, temperature was set at 21°C ± 2 and humidity was kept at 55% ± 10. The oligonucleotides used for genotyping of Nlrx1-deficient mice were:

Nlrx1 “WT For”: 5’-TTA GAC TGG TGT TAC GGG AGA CTG-3’

Nlrx1 “Common Rev”: 5’-CCC AGG CAC TGT TGT CCT ACA-3’

Nlrx1 “KO For”: 5’-TAA GGG TTC GCG TAC GGT G-3’

Two isogenic clones of Leishmania guyanensis (Lgy) were used. A LRV1-bearing (LRV1⁺ LgyM4147/SSU : IR2SAT-LUC(b)c3) and LRV1-cured (LRV1^- LgyM4147/SSU : IR2SAT-LUC(b)c3) Lgy (named LgyLRV1+ and LgyLRV1-, respectively) were obtained by drug treatment of LRV1+ strain of Lgy M4147 (MHOM/BR/75/M4147) containing a firefly luciferase (ffLUC) gene as described previously (39). Lgy parasites were cultured at 26°C in Schneider’s Drosophila medium (Gibco) supplemented with 20% of Fetal Bovine Serum (FBS, Gibco), 1% penicillin/streptomycin (BioConcept), 2% HEPES buffer (BioConcept) and 0.6 μg/ml 6-Biopterin (Sigma-Aldrich) and 0.2% Hemin folate (Sigma-Aldrich, Fluka). For infection, parasites were cultured for 6 days to obtain stationary phase infectious metacyclic promastigotes.

BMDMs were isolated from tibias and femurs of non-infected female and male WT and Nlrx1 ^-/- mice. Macrophages were cultured at 37°C and 5% CO₂ in complete Dulbecco’s modified Eagle Medium (DMEM) supplemented with 10% FBS, 1% penicillin/streptomycin, 1% HEPES buffer and 50 ng/ml of murine recombinant mouse macrophage colony stimulating factor (rmM-CSF, Immunotools) for 6 days. At day 3, fresh complete DMEM supplemented with rm-MCSF was added. At day 6, adherent BMDMs were isolated and plated at a concentration of 1.25 x 10⁶ cells/ml one day prior of infection. BMDMs were infected at 35°C and 5% CO₂ with stationary phase parasites with a multiplicity of infection (MOI) of 5 parasites per macrophage or stimulated with 2 μg/ml of polyinosinic-polycytidylic acid (poly I:C) (Immunotools).

Age-matched (6-9 weeks old) female or male mice were injected in the hind footpads with 3 x 10⁶ stationary phase Lgy promastigotes in 50 μl of Dulbecco’s Phosphate-Buffered Saline (dPBS, Gibco). To follow disease progression, footpad swelling was measured weekly using a Vernier caliper. When required, to quantify inflammation and parasite burden, mice were injected intraperitoneally (i.p.) with 200 mg/kg of Luminol sodium salt (Carbosynth) or 150 mg/kg of VivoGlo Luciferin (Promega) diluted in dPBS, respectively. Bioluminescence from mouse footpads was measured by In-vivo Xtreme II (BRUKER) and quantified using Molecular Imaging (MI) software (BRUKER) as described previously (40).

Footpads were collected and fixed overnight at 4°C with 4% paraformaldehyde (PFA, Fluka). Following fixation, tissue samples were dehydrated and included in paraffin. 3.5 μm paraffin sections were generated using a Microm HM355 microtome (Thermo Scientific) and stained with hematoxylin (J.T Baker) and eosin (Merck) stain. Sections were visualized using a NanoZoomer S60 (Hamamatsu Photonics K.K.) scanner with Nikon Plan Apochromat 40x objective using brightfield contrast and analyzed using NPD.scan3.3 (Hamamatsu Photonics K.K.). Representative images of the sections are shown.

Footpads from infected WT or Nlrx1^-/- mice were collected and snap-frozen in liquid nitrogen and kept at -80°C for storage. For RNA extraction, tissues were lysed in TRI Reagent (Molecular Research Center, inc) using a TissueLyser system (Qiagen). RNA was isolated by chloroform/isopropanol/ethanol phase separation protocol as described previously (41). RNA was quantified using NanoDrop 2000 (ThermoFisher Scientific) and 2000 ng/μl of RNA was used for cDNA synthesis. Alternatively, BMDMs were lysed with PRImeZOL Reagent (Canvax) and RNA was isolated using Direct-zol-96 RNA (Zymo Research) according to the manufacturer’s instructions. cDNA was synthesized using SuperScript II Reverse Transcriptase (Invitrogen). Real-time quantitative PCR (qRT-PCR) was performed using the LightCycler 480 (Roche). The results were analyzed using the threshold cycle (C_T) method (2^-ΔΔCt) for relative quantification of gene expression and normalized to L32 housekeeping gene encoding for 60S ribosomal protein. The oligonucleotides used were:

L32: 5′-AAG CGA AAC TGG CGG AAA C-3’ and 5′-TAA CCG ATG TTG GGC ATC AG-3’

Il6: 5’-TCC AGT TGC CTT CTT GGG AC-3’ and 5’-GTC TAA TTA AGC CTC CGA CT-3’

Tnfa: 5’-CAT CTT CTC AAA ATT CGA GTG ACA A-3’ and 5’- TGG GAG TAG ACA AGG TAC AAC CC-3’

Ifnb: 5’-AAC CTC ACC TAC AGG GC-3’ and 5′-CAT TCT GGA GCA TCT CTT GG-3’

Nlrx1: 5’-CAT GGA AAC TCG GCA GAC AG-3’ and 5’-GGC TAA ACC ACT CGG TGA GG-3’

BMDMs were lysed with 1.5x Laemmli’s Sample Buffer in H₂O and incubated at 95°C for 3 min. Cell lysates were size-fractioned by 8% SDS-PAGE and wet-transferred to a nitrocellulose membrane. Membranes were blocked with 5% non-fat dry milk in Tris buffered saline with 0.1% Tween-20 (TBST) at room temperature. Western blotting was performed using the following antibodies: anti-NLRX1 (1/1000, Proteintech, 17215-1-AP), anti-y-TUBULIN (1/10 000, Sigma-Aldrich, T5326), goat anti-rabbit IgG (H+L) HRP (1/2500, Promega, W4011) and goat anti-mouse IgG (H+L) HRP (1/2500, Promega, W4021). ECL Western Blotting detection reagent (GE Healthcare Life Sciences) was used for revelation.

The concentrations of IFNβ (Thermo Fisher, 424001), IL-6 (Invitrogen, 88-7064-88) and TNF-α (Invitrogen, 88-7324-88) in collected cell-free supernatants at 24 hours from infected or stimulated BMDMs were determined by ELISA following the manufacturer’s instructions. Optical density was read on a Synergy HT Multi-Mode Plate Reader (BioTek Instruments) at 450/570 nm.

BMDMs were seeded in μ-Plate 96 Well Black (Ibidi) at a concentration of 1.25 x 10⁶ cells/ml. Cells were infected or stimulated for 8 and 24 hours and fixed with 4% PFA (Fluka). Cells were subsequently stained with 4′,6-diamidino-2-phenylindole (DAPI) (Molecular Probes) and Alexa Fluor 488 phalloidin (Molecular probes) to stain the nuclei and cytoplasm, respectively. Images were acquired using ImageXpress Micro Confocal (Molecular Devices) with a 40x objective. Parasite and cell number per well were quantified using MetaExpress custom Module Editor (Molecular Devices) as described previously (42).

BMDMs from age-matched WT and Nlrx1^-/- mice were infected with Lgy parasites or stimulated with poly I:C for 8 and 24 hours. RNA was extracted using a RNeasy Kit (Qiagen) following manufacturer’s instructions. RNA quality and concentrations were determined by Fragment Analyzer and Ribogreen QubIT quantification, respectively, and libraries for sequencing were then prepared at the Lausanne Genomic Technologies Facility (GTF). Statistical analysis was performed for genes independently in R (R version 4.0.3). Genes with low counts were filtered out according to the rule of 1 count(s) per million (cpm) in at least 1 sample. Differential expression was computed with limma by fitting data to a linear model (43). Weighted gene co-expression network analysis (WGCNA) was performed on normalized data in R (package WGCNA 1.69). Modules were identified by dynamic tree cut with a minimum module size=20. Module eigengenes (MEs) that are the first principal component of the module were calculated and relationship of module eigengenes with infection status was assessed with a regression analysis. Module eigengenes average predictions were plotted as a heatmap. Gene Ontology (GO) enrichment analysis was performed for gene co-expression modules against GO categories using the topGO R package (topGO 2.26.0) and gene ontology database (07.2019). Only biological processes (BP) were considered for the analysis. For all modules, genes were ranked according to their connectivity within a single module as measured by the kwithin-index.

Metabolism measurement was performed with a Seahorse XFe96 Extracellular Flux Analyzer (Agilent). BMDMs were plated in a Seahorse XF96 cell culture microplate (Agilent) overnight. Cells were then infected with Lgy parasites for 8 hours and then pre-incubated in assay medium (Seahorse XF DMEM pH 7.4, Agilent) supplemented with 2 mM L-glutamine (Gibco), 1 mM pyruvate (Gibco) and 25 mM glucose (Gibco) for 1h at 37°C without CO₂. To measure mitochondrial metabolism and glycolysis, a Mito Stress Test and a Glycolytic Rate Assay were performed, respectively, according to manufacturer’s instructions. Cells were treated with 1 μM Oligomycin (Sigma), 2 μM FCCP (Sigma), 0.5 μM/0.5 μM Rotenone/Antimycin A (Sigma) and 50 mM 2-deoxy-D-glucose (Sigma). Results were analyzed using Wave Desktop software (Agilent) and data was normalized to total protein concentration per well. Briefly, post-assay, cells were lysed with a mixture of RIPA Buffer IV (Biotech) and a complete protease inhibitor cocktail tablet (Roche) in H₂O. Protein concentration was quantified using Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) following manufacturer’s instructions. To assess the effect of estradiol, cells were pre-treated with 200 pg/ml of 17-β estradiol (Sigma) for 2 hours and estradiol was kept in the assay medium for the duration of the assay.

Intracellular and mitochondrial ROS production were measured by the superoxide indicator dihydroethidium (DHE) (Thermo Fisher Scientific) and the mitochondrial superoxide indicator MitoSOX Red (Thermo Fisher Scientific). Briefly, cells were infected with Lgy parasites or treated with poly I:C for 8 hours, after which cells were labelled with 5μM DHE or 5 μM MitoSOX Red in PBS with 5 mM glucose (Gibco) for 20 min at 37°C. After incubation, cells were washed, and fluorescence was measured at 518/606 nm (MitoSOX Red) and 510/590 nm (DHE) using the Spectramax i3 plate reader (Molecular Devices). Measures were normalized to total protein concentration per well using Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) following manufacturer’s instructions.

After 8 hours of infection, cells were fixed in 2.5% glutaraldehyde solution (Fluka) in PBS for 1 hour at room temperature (RT), then postfixed with a mixture of 1% osmium tetroxide (EMS) and 1.5% of potassium ferrocyanide (Sigma) in PBS for 1 hour at RT. Samples were washed in distilled water, spin down in low melting 2% agarose (Sigma) in H₂O (Sigma), let to solidify on ice, cut in 1 mm³ cube and dehydrated in acetone solution (Sigma) at graded concentrations (30%, 40 min; 50%, 40 min; 70%, 40 min; 100%, 2x1 hour). This was followed by infiltration in Epon (Sigma) at graded concentrations (Epon 1/3 acetone, 2 hours; Epon 3/1 acetone, 2 hours; Epon 1/1, 4 hours; Epon 1/1, 12 hours) and finally polymerized for 48 hours at 60°C. Ultrathin sections of 50 nm were cut on a Leica Ultracut (Leica Mikrosysteme GmbH) and picked up on a copper slot grid 2x1 mm (EMS) coated with a polystyrene film (Sigma). Sections were poststained with 2% uranyl acetate (Sigma) in H₂O for 10 minutes, rinsed several times with H₂O followed by Reynold’s lead citrate in H₂O (Sigma) for 10 minutes and rinsed several times with H₂O. Singles micrographs were taken with a transmission electron microscope Philips CM100 (Thermo Fisher Scientific) at an acceleration voltage of 80kV with a TVIPS TemCam-F416 digital camera (TVIPS GmbH). To determine the percentage of mitochondria volume per cell volume, a grid (500 nm spacing) was applied on each micrograph and each intersection was defined as being part of the mitochondria, nucleus, or cytoplasm. The stereology analysis was performed using 3dmod and its stereology plugin (44).

All graphs and statistical tests were generated in GraphPad Prism [version 9.3.1 (350)]. Either unpaired Student’s t-test or two-way ANOVA with multiple comparisons was used for bar graphs, while repeated-measure two-way ANOVA with Bonferroni’s post-test correction was used for x/y curves. Significance was reached with p values ≤ 0.05. p values are shown as * for p < 0.05, ** for p < 0.01, *** for p < 0.001 and **** for p < 0.0001.

To investigate whether NLRX1 modulated the pathogenicity of Lgy and affected disease progression, we first infected female C57BL/6 wild-type (WT) or NLRX1-deficient (Nlrx1^-/- ) mice with Lgy parasites containing the dsRNA LRV1 virus (LgyLRV1+) and monitored lesion development weekly. In comparison to WT female mice, Nlrx1^-/- infected female mice showed significantly increased footpad swelling ( Figure 1A ) and signs of inflammation as measured by in vivo bioluminescence imaging following luminol injection ( Figure 1B ). However, no significant differences were observed in parasite load as measured by bioluminescence of luciferase expressing parasites ( Figure 1C ). Thus, in female mice, NLRX1 seemed to attenuate LRV1 mediated inflammation independently of the parasite load.

To further confirm the role of NLRX1 in regulation of inflammation, we collected non-infected and infected footpads at the peak of infection. We did not observe any differences or abnormalities in skin structure of non-infected Nlrx1-deficient mice compared to WT ( Figure 1D , upper panel). However, correlating to the lesion severity, we observed an important increase in thickness and in immune cell infiltration in the dermis (“D”) of LgyLRV1+ infected Nlrx1 ^-/- mice ( Figure 1D , bottom panel) compared to WT mice. In addition, we observed that at the peak of infection pro-inflammatory markers Ifnb and Il6 ( Figures 1E, F ) were significantly upregulated in lesions of Nlrx1^-/- mice infected with LgyLRV1+ parasites, while no differences were observed in Tnfa expression as measured by qRT-PCR ( Figure 1G ). Thus, taken together these results supported a role for NLRX1 in controlling inflammation by limiting immune cell infiltration and tissue damage as well as type I IFN and IL-6 expression in lesions of female mice, while not affecting TNFα expression known to be responsible for parasite killing (36).

Several studies have reported a downregulation of NLRX1 in different experimental models such as viral and bacterial infections and brain injury (27, 45, 46). Therefore, we investigated whether infection with Lgy regulated Nlrx1 expression. Thus, we infected bone marrow derived macrophages (BMDMs) isolated from WT female mice with LgyLRV1+ parasites or stimulated them with polyinosinic-polycytidylic acid (poly I:C), a known synthetic dsRNA agonist of TLR3. We observed a downregulation of Nlrx1 mRNA in both infected and poly I:C treated BMDMs at 8 hours post-infection (p.i.) when infection is established, however with a fold-change inferior to 2 ( Figure 2A ). This downregulation was no longer observed at 24 hours p.i. ( Supplementary Figure S1A ). Similarly, we did not observe any significant changes in NLRX1 protein levels at 8 or 24 hours p.i. ( Supplementary Figure S1B ), suggesting a transcriptional regulation of NLRX1 only in the early phase of infection with LgyLRV1+ parasites or after poly I:C treatment.

To further characterize the increased tissue damage and inflammation observed in Nlrx1^-/- female mice, we next sought to determine any possible effect of NLRX1-deficiency on inflammation and Lgy infection in in vitro infected BMDMs. Thus, we infected WT and Nlrx1 ^-/- female BMDMs with LgyLRV1+ or stimulated them with poly I:C and measured pro-inflammatory cytokines in the cell-free supernatant at 24 hours post-infection. As expected, no cytokines were detected in non-infected cells. In contrast, NLRX1-deficiency resulted in a significantly increased production of IFNβ in LgyLRV1+ infected or poly I:C treated cells ( Figure 2B ). We did not observe any differences in IL-6 or TNFα production ( Figures 2C, D ) suggesting that IFNβ was produced by infected macrophages whereas, upon tissue damage, Il6 up-regulation measured in lesions was likely produced by other sources such as keratinocytes, dendritic cells and fibroblasts (47).

In infectious models, loss of NLRX1 has been reported to promote either pathogen survival or clearance depending on the model (46, 48–52). To investigate whether NLRX1 affected the number of parasites per BMDM, we infected WT and Nlrx1-/- female BMDMs with LgyLRV1+ parasites for 8 and 24 hours. We did not observe differences in parasite burden in the establishment of infection (8 hours p.i.) ( Supplementary Figure S1C ) suggesting that NLRX1 did not affect the phagocytic capacity of BMDMs. In contrast, at 24 hours post infection, the absence of NLRX1 resulted in a decreased number of parasites per cell ( Figure 2E ). This decrease was associated to an increased macrophage survival in absence of NLRX1 that was observed only at 24 hours p.i. ( Supplementary Figures S1D, E ) (27, 53). However, as shown in Figures 1A, C , despite an increased macrophage survival with a lower number of parasites per cell in vitro, in vivo Nlrx1-deficient female mice developed larger lesions independently of the parasite load.

To better define the role of this mitochondrial sensor in macrophages isolated from female mice, we performed a transcriptomics analysis of WT and Nlrx1^-/- female BMDMs infected with LgyLRV1+ parasites or stimulated with poly I:C for 8 or 24 hours. We then performed a global weighted correlation network analysis (WGCNA) to group genes with similar expression patterns into modules. WGCNA has been used previously to identify key biological processes and gene modules associated with the studied disease (54–57). The underlying hypothesis is that genes involved in the same function or pathway or that are co-regulated are expected to be in the same module named by a color. Genes that do not group to any module form the “grey” module and are discarded from the analysis. The association of the WGCNA modules with the different conditions is represented as heatmaps at 8 and 24 hours ( Figure 2F , Supplementary Figure S1F ). The genes in each module are listed in Supplementary Tables 1, 2 . We performed a gene ontology (GO) enrichment analysis for each module to identify the biological processes associated to each module. For each module GO terms are listed in Supplementary Tables 3 and 4 . Based on the GO analysis, we could group most of the modules into 2 main categories of GO terms (1): inflammation and infection and (2) mitochondria and metabolism ( Supplementary Figure S1G ). These categories were based on a selection of several keywords such as “interleukin” or “biosynthesis”, respectively. Identification of such biological processes was not surprising because of the choice of an infectious model and a knock-out gene for a mitochondrial receptor described to play a role in both categories. In contrast, both at 8 and 24 hours p.i., more than a third of the modules (34.8% and 45.5%, respectively) were enriched in GO terms associated to sex hormone signaling ( Figure 2F , Supplementary Figures S1F, G ) such as “regulation of androgen receptor signaling pathway” (GO:0060765) or “cellular response to estrogen stimulus” (GO:0071391). These GO terms were rather unexpected and suggested a potential sex-bias in the NLRX1 phenotype.

Although no association between NLRX1 and sex hormone signaling has been described to our knowledge, many studies have shown that sex hormones may play a role in the regulation of innate immune cell activation, inflammation, and mitochondrial function (1, 58). Based on the transcriptomics analysis of female BMDMs, we decided to investigate whether the Nlrx1^-/- phenotype was dependent on sex. Thus, we infected male C57BL/6 WT or Nlrx1^-/- mice with LgyLRV1+ parasites. In contrast to female mice, we found no significant differences in the development of lesions ( Figure 3A ). Moreover, at the peak of infection, male mice did not display any differences in inflammation as measured by in vivo bioluminescence imaging ( Figure 3B ). However, we observed a reduced parasite burden in Nlrx1^-/- mice in comparison to WT ( Figure 3C ) that we did not observe in female mice ( Figure 1C ). Consistent with the absence of difference in inflammation and contrary to female mice, we observed no significant differences in skin structure, dermal thickness or immune cell infiltration between non-infected ( Figure 3D , upper panel) or LgyLRV1+ infected male mice ( Figure 3D , bottom panel). Finally, we observed a significant reduction in Ifnb mRNA levels in the lesions of Nlrx1^-/- male mice infected with LgyLRV1+ ( Figure 3E ) in comparison to WT, but no differences in Il6 or Tnfa mRNA levels as measured by qRT-PCR ( Figures 3F, G ). Taken together, NLRX1 did not modulate lesion development or inflammation in male mice contrarily to results obtained with female mice suggesting a strong sex bias in the function of NLRX1.

To better understand the sex-bias in NLRX1 function, we decided to investigate whether isolated macrophages also displayed a sex-dependent phenotype. First, we infected both female and male WT BMDMs with Lgy parasites or stimulated them with poly I:C for 8 and 24 hours to analyze whether NLRX1 expression was affected by sex. At 8 or 24 hours p.i., we did not observe any significant differences in Nlrx1 mRNA ( Figures 4A, B ) and protein levels ( Supplementary Figure S2A ) between females and males, suggesting that biological sex did not directly affect NLRX1 expression in BMDMs. Based on the in vivo data, the role of NLRX1 in the regulation of inflammation and infection was highly dependent on sex. Thus, we measured IFNβ, IL-6 and TNFα cytokine production at 24 hours post-infection in both female and male WT and Nlrx1^-/- BMDMs infected with LgyLRV1+ parasites or stimulated with poly I:C. No cytokines were detected in non-infected cells. Consistent with the lesions, increased IFNβ production ( Figure 4C ; Supplementary Figure S3A ) was specific to female Nlrx1^-/- BMDMs infected with LgyLRV1+ or treated with poly I:C, whereas we did not observe any differences between male BMDMs. Interestingly, the Nlrx1^-/- female BMDMs showed a similar IFNβ production compared to males. Surprisingly and in contrast to the lesions, NLRX1-deficiency resulted in a significantly increased production of both IL-6 and TNFα levels ( Figures 4D, E ; Supplementary Figures S2B, S3B, C ) in Nlrx1^-/- male BMDMs in comparison to WT male BMDMs.

Since we observed differences in cytokine profiles of female and male BMDMs, we investigated whether also parasite burden in BMDMs was affected by sex. As previously, we infected WT and Nlrx1 ^-/- female and male BMDMs with LgyLRV1+ parasites for 8 and 24 hours. At 8 hours p.i. female and male BMDMs showed similar parasite burdens independently of the genotype ( Supplementary Figure S2C ). However, at 24 hours p.i. WT female BMDMs maintained a higher parasite burden, whereas female Nlrx1^-/- BMDMs and both male BMDMs showed a significantly reduced and similar parasite burden ( Figure 4F ; Supplementary Figure S3D ). Although we did not observe statistically significant differences in macrophage survival, at 24 hours but not at 8 hours p.i., both male BMDMs and Nlrx1-deficient female BMDMs seemed to survive better in comparison to WT female cells correlating to the lower parasite burden observed in these cells ( Figure S2D, E ). Taken together, these results supported a strong association between sex and NLRX1 function.

NLRX1 was shown to play a role in the regulation of both OXPHOS and glycolysis, the two major metabolic pathways for energy production (31, 59). In addition, based on the transcriptomics analysis of female BMDMs, NLRX1 seemed to have a strong effect on cellular metabolism and mitochondrial function. Thus, we sought to investigate whether the loss of NLRX1 resulted in a change in OXPHOS or glycolysis, the two major energetic pathways of the cell. To examine if NLRX1 affected OXPHOS, we determined the oxygen consumption rate (OCR) in both female and male WT and Nlrx1^-/- BMDMs infected with LgyLRV1+ parasites for 8 hours. In non-infected BMDMs, we observed very low OCR levels and no differences between groups, suggesting a low metabolic activity in unstimulated cells after 8 hours ( Figure 5A ). Moreover, NLRX1-deficient female BMDMs infected with LgyLRV1+ showed a significantly increased basal mitochondrial respiration compared to WT female, but that was comparable to basal respiration of male BMDMs. In contrast we did not observe differences in basal respiration between WT and Nlrx1-deficient male BMDMs infected with LgyLRV1+ ( Figure 5A ; Supplementary Figures S4A , S5A ). To investigate whether NLRX1 also affected glycolysis in BMDMs, we measured the basal glycolytic rate in the same conditions. In non-infected cells, we observed a very low glycolytic rate, and no differences between groups ( Figure 5B ). Similarly to the pattern observed for mitochondrial respiration, female Nlrx1^-/- BMDMs infected with LgyLRV1+ showed a significantly increased glycolytic rate compared to WT female, but that was similar to the glycolytic rate of male BMDMs. As before for mitochondrial respiration, we did not observe differences in glycolytic rate between WT and Nlrx1-deficient male BMDMs ( Figure 5B ; Supplementary Figure S5B ). Taken together, these results suggested a global effect of NLRX1 on female, but not on male macrophage metabolism.

NLRX1 is localized at the mitochondria, and its function in the control of inflammation has been linked to mitochondria and in the modulation of mtROS production (21, 23, 24, 50, 60, 61). This modulation occurs potentially through the interaction with Ubiquinol-Cytochrome C Reductase Core Protein 2 (UQCRC2), a subunit of the complex III of the respiratory chain (38, 62). We measured mtROS and cellular ROS accumulation, by MitoSOX Red and DHE respectively, in female and male BMDMs infected with LgyLRV1+ parasites or treated with poly I:C for 8 hours. Globally mtROS was reduced in Nlrx1^-/- BMDMs independently of sex, whereas no differences were observed in cellular ROS ( Figure 5C ; Supplementary Figure S4B, S5C ). To verify whether mitochondrial morphology was affected by NLRX1-defiency, we performed transmission electron microscopy (TEM) on non-infected and LgyLRV1+ infected female and male BMDMs. We did not observe any differences in mitochondrial morphology between WT and Nlrx1^-/- BMDMs and no mitochondrial defects were observed ( Figure 5D ). On the other hand, quantification of TEM images revealed a reduced mitochondrial density and number in both female and male Nlrx1^-/- BMDMs compared to WT ( Figure 5E ; Supplementary Figure S4C, S5D ) potentially explaining the reduced mtROS production observed in these cells.

Since neither mtROS production nor mitochondrial structure could explain differences observed in metabolism, we examined whether sex differences in cellular bioenergetics could be modified by the female hormone 17-β estradiol. Similarly to the study done by Gupta et al. (2020) (3), estradiol treatment of male BMDMs significantly reduced mitochondrial respiration ( Figure 5F ). Surprisingly, estradiol treatment increased basal respiration of WT female BMDMs but reduced basal respiration of Nlrx1-deficient female BMDMs similarly to males ( Figure 5F ). This male-like pattern of Nlrx1-deficient female cells after estradiol treatment was specific to mitochondrial respiration since estradiol increased the glycolytic rate only in females but not in male BMDMs ( Supplementary Figure S4D ). Taken together, these results suggested that macrophage metabolism can be modulated by the female sex hormone and differences in estradiol response might contribute to the sex bias observed in the bioenergetics between WT and Nlrx1-deficient BMDMs.

To better understand the complex interplay between NLRX1 genotype and sex, we decided to perform a second RNA sequencing and WGCNA analysis including both female and male BMDMs. To further confirm the role of NLRX1 in the control of TLR3- and type I IFN-mediated inflammation, we added a non-inflammatory strain of Lgy that does not carry the LRV1 virus (LgyLRV1-) to the analysis. Infection with LgyLRV1- strain does not induce NF-κB mediated pro-inflammatory cytokines nor a potent type I IFN response thus leading to a less severe form of the disease (37, 63). We infected BMDMs with both Lgy parasites or stimulated them with poly I:C for 8 or 24 hours. The relationship between the different modules and the experimental conditions was assessed with a regression analysis and module eigengenes average predictions are shown as heatmaps at 24 and 8 hours ( Figure 6A ; Supplementary Figure S6A , respectively). The genes in each module are listed in Supplementary Tables 5, 6 . In each condition, the heatmap clusters the groups according to their similarity. Supporting the male-like behavior of Nlrx1-deficient female cells, we could observe that at 24 hours post-infection, in all condition except with poly I:C treatment, the Nlrx1^-/- female BMDMs clustered more closely to the WT and Nlrx1-deficient male BMDMs than to the WT female ( Figure 6A ). At 8 hours p.i., this clusterization pattern was only observed with infection with LgyLRV1+ ( Supplementary Figure S6A ). As previously, we performed a gene ontology (GO) enrichment analysis for each module to identify the biological processes associated to each module. For each module GO terms are listed in Supplementary Tables 7, 8 . As expected, GO enrichment analysis of the different modules revealed that most modules were enriched in GO terms associated to (1) Inflammation and infectivity and (2) Mitochondria and metabolism. In addition, at 8 hours p.i. more than half of the modules (60%) were enriched in GO terms associated to (3) sex hormone signaling, while at 24 hours p.i. we observed this enrichment in all modules, except the green module ( Figure 6A ; Supplementary Figure S6A, S6B ).

We next performed pairwise comparisons between the different groups and plotted the number of significantly differentially expressed genes both at 8 and 24 hours using a cut-off of fold-change [-2:2] and an adjusted p-value < 0.05 ( Figure 6B–D ; Supplementary Figures S6C–G ). We identified only a few sex- or genotype-specific genes except in non-infected BMDMs in which at 8 hours p.i. we identified a significant number of genotype-specific genes ( Supplementary Figure S6C ). In the last two comparisons, we compared opposite genotypes with opposite sex. Interestingly, we observed a common pattern that when we compared Nlrx1^-/- female BMDMs to a WT male, independently of the condition or time-point, we observed less differences than when Nlrx1^-/- male BMDMs were compared to a WT female, suggesting that the transcriptomic profile of female Nlrx1^-/- cells was closer to a male, as suggested by the WGCNA analysis.

To investigate whether we would be able to explain why Nlrx1^-/- female BMDMs had a male-like transcriptomics profile, we wanted to identify genes that in Nlrx1^-/- female BMDMs were expressed at a similar level than in male BMDMs. To do so, we looked for genes that were differentially expressed only in the WT female BMDMs, but not between the three other groups, Nlrx1^-/- female, Nlrx1^-/- male and WT male BMDMs. We used a threshold of adjusted p-value inferior to 0.1. The analysis identified only 3 genes in non-infected BMDMs at 8 hours timepoint ( Figure 6E ): the lysophosphatidic acid receptor 1 (Lpar1), the vascular endothelial growth receptor factor 2 (Kdr) and the cyclic AMP-responsive element-binding protein 3-like protein 2 (Creb3l2). These genes play a role in inflammation (64–66), angiogenesis (67) and collagen synthesis (68, 69), respectively, aspects that are hallmarks of Leishmania infection (70, 71) and where in other experimental models sex differences have been described (72–74).