Inducible nitric oxide synthases (iNOS) is a member of a family of three enzymes responsible for conversion of L-arginine to L-citrulline, an important metabolic process accompanied by the release of nitric oxide (NO). Unlike endothelial NOS (eNOS) and neuronal NOS (nNOS), two constitutively expressed Ca2+ dependent isoforms, expression of iNOS does not require Ca2+ intake and does not occur in resting cells. Several inflammatory stimuli can trigger iNOS expression and the production of large amounts of NO. This molecule has many different functions and was described, amongst others, to have regulatory effects on angiogenesis, tumor metastasis and cell adhesion (Bogdan, 2015). Most importantly, NO plays a crucial role in the immune protection against intracellular pathogens as it is directly responsible for their elimination. Thus, iNOS is the key player mediating the immune response to many microbial infections.

Toll-like receptors (TLRs) are the most known and best-characterized family of pattern-recognition receptors (PRRs) (Akira et al., 2006). They play a crucial role in detection of pathogenic intruders. TLRs initiate microbial clearance and production of pro-inflammatory effector cytokines and chemokines that further direct the adaptive immune response (Majewska and Szczepanik, 2006). Most of the TLRs signal via the myeloid differentiation primary response protein 88 (MyD88), whilst TLR3 is the only TLR that acts solely via the TIR-inducing IFN-β (TRIF) molecule. TRIF signaling promotes an antiviral innate response via activation of the interferon regulatory factor 3 (IRF-3), which leads to the transcription of type I interferons (type I IFNs). Both MyD88 and TRIF signaling cascades lead to the nuclear translocation of the nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) and activator protein-1 (AP-1) transcription factors.

NF-kB is the key player in inflammation and immunity. This complex encompasses five different members containing the Rel homology domain: NF-kB1 (p105/p50), NF-kB2 (p100/p52), RelA (p65), RelB and c-Rel (Hoffmann et al., 2006). The most abundant member is the p65p50 heterodimer (Hayden and Ghosh, 2012) which is usually associated with activation of transcription. This in turn leads to the consequent transcription of pro-inflammatory cytokines, such as tumor necrosis factor alpha (TNF-α) and Interleukin 6 (IL-6) (Kawai and Akira, 2006). As a major negative feedback regulator of NF-kB-mediated inflammation, tumor necrosis factor alpha-induced protein 3 (TNFAIP3) (also named A20) is induced very early in the NF-kB response (Coornaert et al., 2009; Catrysse et al., 2014).

Leishmania is a genus of protozoan parasites that exists in two morphological forms. The flagellated promastigote forms are transmitted by a sand fly vector to the mammalian hosts. Upon inoculation, parasites are engulfed by tissue resident macrophages, where they convert into the round amastigote form. Leishmania survive and proliferate inside the macrophages causing different types of pathologies depending on the infecting species and immune competence of the host. Termed leishmaniasis, this range of diseases affects over 12 million people in 98 countries worldwide (Kaye and Scott, 2011; Alvar et al., 2014). Mucocutaneous leishmaniasis (MCL) and disseminated cutaneous leishmaniasis (DCL) develop mainly due to infection with species of Leishmania Viannia subgenus, such as Leishmania guyanensis (Lgy) and Leishmania brasiliensis (Lbr). These pathologies arise as complications of cutaneous leishmaniasis (CL) and are characterized by metastatic lesions coming from primary sites of infection to the mucosal nasopharyngeal tissues in the case of MCL, and other skin areas in the case of DCL.

In some cases, parasites bear a cytoplasmic double stranded RNA (dsRNA) virus termed Leishmania RNA virus (LRV). There are two known subgenera of LRVs. LRV1 has been found in New World species such as Lgy (Tarr et al., 1988) and Lbr (Zangger et al., 2013), while another species of LRV, termed LRV2 has been found in Old World L. major (Scheffter et al., 1994), L. aethiopica (Zangger et al., 2014) and L. infantum (Hajjaran et al., 2016). The difference between LRV1 and LRV2 lies within the genetic organization and replication mechanisms (Lee et al., 1996; Se et al., 2005; Fujimura and Esteban, 2011; Kostygov et al., 2021).

Although the existence of such endoprotozoan viruses of the Totiviridae family have been discovered and described almost 30 years ago (Tarr et al., 1988; Stuart et al., 1992; Weeks et al., 1992), their role in disease progression remained elusive for years. It is now known that upon its release from dying parasites or possibly due to exosomal transfer (Atayde et al., 2019; Olivier and Zamboni, 2020), this viral endosymbiont is recognized by endosomal TLR3, leading to increased production of pro-inflammatory cytokines and chemokines. Paradoxically, this TLR3-mediated innate response exacerbates the disease as mice infected with LRV1-bearing Lgy (LgyLRV1+) develop higher lesions and display more parasite burden at the peak of infection when compared to their LRV1-negative counterparts (LgyLRV1-) (Ives et al., 2011). In part this phenomenon is explained by the activation of Akt-dependent pro-survival mechanisms within the host macrophage. Prolonged survival of infected cells promotes parasite persistence, therefore contributing to the exacerbated phenotype (Eren et al., 2016). We recently showed that LRV1-dependent dissemination and formation of metastatic lesions is linked to low levels of IFN-γ and is mediated by IL-17A (Hartley et al., 2016). Clinical significance of LRV1 presence within Leishmania parasites relies on the selection of correct treatment schemes. We and others recently described that patients, diagnosed with MCL and DCL, caused by LgyLRV1+ or LbrLRV1+, require prolonged therapy and are more prone to relapse upon treatment (Adaui et al., 2015; Bourreau et al., 2015). We further identified that type I IFNs, induced downstream of TLR3 signaling, either by LRV1, or co-infecting viruses, are the key molecules that contribute directly to the exacerbated disease phenotype (Rossi et al., 2017).

Being an extremely successful pathogen, Leishmania has developed various strategies to evade the host immune response. These include, facilitating its own capture by the host cells and resisting the internal hostile environment together with subverting various signaling pathways and exploiting immune mechanisms in order to facilitate its own survival and progression, thus making Leishmania one of the most insidious pathogens (Gupta et al., 2013; Rossi and Fasel, 2018). Some examples describe various strains of Leishmania promoting host-cell A20 induction and therefore evading inflammatory reactions as well as suppressing inflammasome activation (Srivastav et al., 2012; Gupta et al., 2017; Hartley et al., 2018).

In the case of LgyLRV1+ infection, viral dsRNA serves as an additional potent immunogen, whose recognition triggers a plethora of signaling cascades. Herein we described how LgyLRV1+ infection initiates a parasiticidal signaling pathway. Consequently, we showed how parasites mitigated these effects and reported an observation implying that LgyLRV1+ may exploit these anti-Leishmania mechanisms to facilitate its dissemination.

Although transmission of most CL-causing Leishmania spp is zoonotic by nature (Gramiccia and Gradoni, 2005), generations of man-made impact on environmental setting and urbanization have led to the disease being acquired in various types of human settlements (Reithinger et al., 2019). This led to wider geographical distribution and increased incidence of human cases that were first described as early as 7^th century BC (Cox, 2002; Zink et al., 2006). Threatening humanity for thousands of years, Leishmania parasites have excelled at survival and propagation. In a long lasting evolutional race they have developed various mechanisms allowing them to evade and subvert defensive mechanisms of the host (Aguirre-Garcia et al., 2018; Rossi and Fasel, 2018). One strategy lies within the cooperative relationship with dsRNA viruses such as LRV1 that are harbored inside the parasite. Innate immune recognition of LRV1 within Lgy leads to exacerbation of the disease (Ives et al., 2011), prolonged survival of the host cell and persistence of intracellular parasites (Eren et al., 2016) that in turn facilitates parasite dissemination and metastasis (Hartley et al., 2016). Similarly, infection with exogenous viruses can seriously worsen the course of disease, promote metastasis and relapse of leishmaniasis caused by LRV1-free Lgy (Rossi et al., 2017).

An indispensable role of iNOS in immune protection against leishmaniasis in mice has been long known (Wei et al., 1995), and widely reviewed (Bogdan, 2015; Olekhnovitch and Bousso, 2015). Host defense against leishmaniasis primarily relies on the actions of a mounted adaptive immune response. Several weeks upon infection, C57BL/6 mice develop a strong type 1 helper (Th1) CD4+ T cell response. These mice are able to efficiently control infection whereas BALB/c mice develop a type 2 helper (Th2) CD4+ T cell response and fail to recover. These observations provided evidence that immune protection against leishmaniasis is governed by actions of CD4+ T cells (Sacks and Noben-Trauth, 2002). Upon polarization, CD4+ T cells migrate to the lesion where they accumulate (Filipe-Santos et al., 2009). Consequently, they begin secreting large amounts of IFN-γ of the type II IFN family. Sensed by macrophages, this cytokine triggers expression of iNOS which in turn generates high amounts of NO leading to parasite elimination in an auto- and paracrine manner (Liew et al., 1991; Olekhnovitch and Bousso, 2015).

In this study, we revealed that one of the major impacts of LRV1 recognition by TLR3 was up-regulation of iNOS expression. However, LRV1-triggered iNOS failed to control infection in vitro and in vivo. We further showed that production of iNOS in response to LRV1 required the action of type I IFNs, that have been previously shown to govern the exacerbated disease phenotype (Rossi et al., 2017). We then investigated the strategy employed by the parasites to evade death by type I IFN-mediated iNOS. We showed that Leishmania suppressed NF-kB activation, another pathway required for a robust iNOS induction. Finally, we report an observation suggesting that low levels of LRV1-triggered iNOS and, the production of NO that might correlate with parasite propagation and metastasis. Our findings provide the link between disease exacerbating LRV1-elicited type I IFNs and IL-17A mediated metastasis, occurring in the absence, or in the presence of low levels of protective IFN-γ.

Type I IFNs that are mainly associated with innate immune response are also known to be upstream of iNOS induction. Their role in pathogenesis of leishmaniasis is controversial and, most likely, depends on the species of Leishmania. In particular, IFN-β has been shown to be protective against Leishmania major (Lm) of the Leishmania subgenus. This effect has been associated with iNOS induction (Diefenbach et al., 1998; Mattner et al., 2004). On the other hand, species of the Leishmania Viannia subgenus such as Lbr and Leishmania amazonensis (Lam) benefit from the actions of type I IFNs in an NO-independent manner. In the latter case, type I IFNs led to up-regulation of superoxide dismutase I (SOD1) and consequent reduction of parasiticidal superoxide (Khouri et al., 2009). We recently reported that LRV1-induced type I IFNs contributed to disease progression in the case of Lgy infection (Rossi et al., 2017).

The exact mechanism leading to type I IFN-mediated iNOS induction is based on cooperation of two transcriptional pathways and has been well characterized (Farlik et al., 2010; Wienerroither et al., 2015). On the one hand it involves the type I IFN pathway, whereby ligation of IFNAR leads to phosphorylation of STAT1 and STAT2 by the receptor-associated Janus tyrosine kinases (JAK). The tyrosine-phosphorylated STATs form heterodimers and also bind to the interferon regulatory factor 9 (IRF9). This results in the formation of a trimeric complex, known as interferon-stimulated gene factor 3 (ISGF3). ISGF3 alone may be both necessary and sufficient for the transcription of some interferon-stimulated genes (ISGs). However, often and depending on a promoter, it requires input from additional signaling pathways such as NF-kB (Levy and Darnell, 2002). This becomes the second arm of the aforementioned transcriptional cooperation, which is necessary for iNOS induction. NF-kB binds to the gene promoter following induction by PRR signaling, this in turn leads to the recruitment of various proteins involved in the formation of a bigger transcriptional complex. Consequent assembly and binding of ISGF3 leads to the ligature of previously recruited proteins to their binding partners and later recruitment of complexes containing the RNA polymerase II (Pol II) (Farlik et al., 2010; Wienerroither et al., 2015). Our data lied within this mechanism whereby LRV1-induced type I IFNs could be sensed by BMDMs in an autocrine and paracrine manner. However, the NF-kB pathway appeared to be controlled, or insufficient to lead to strong iNOS induction in vitro. Our data demonstrated that the above described system was not impaired functionally as additional NF-kB stimulation via TLR-2/6 signaling boosted LRV1-mediated iNOS induction ( Figure 2 ).

It is evident that the overall impact of the LRV1-TLR3-IFNAR signaling axis benefited disease progression and contributed to the pathological phenotype (Rossi et al., 2017). iNOS can also be induced by activation of TLR4 (Abdul-Cader et al., 2016), or as shown by our data, by combined ligation of TLR2 and TLR3 ( Figures 3A, B ). We previously demonstrated that TLR2 and TLR4 do not play major roles in the innate response to Lgy (Ives et al., 2014). This was also supported by another study which showed no protective role of TLR2 in response to Lbr, another species of the Leishmania Viannia subgenus (Vargas-Inchaustegui et al., 2009). These findings were opposite to what is known in regard to TLR2 function against the Leishmania subgenus and Lm in particular (Halliday et al., 2016). Observed species-specific differences can be explained by the fact that Leishmania parasites express different types, and up to 10-fold less lipophosphoglycan (LPG), a surface molecule, which is known to activate TLR-2 (Muskus et al., 1997; De Assis et al., 2012). Taken together these results lead to the assumption that the symbiotic partnership between Leishmania parasites and LRV1 occurred preferentially in those parasite species where the chances to induce parasitotoxic iNOS were minimized and thus allowed beneficial effects of co-habitation that overweighed the detrimental ones.

As described above, BMDM infection with LgyLRV1+ alone led to very weak iNOS production at the protein level and no detectable nitrites in the culture medium that did not have an impact on the level of intracellular infection in vitro ( Figures 1E–G ). This observation combined with the in vivo data ( Figures 4C, D ) led us to hypothesize that LgyLRV1+ parasites developed a strategy to minimize type I IFN induced iNOS, allowing them to benefit from other effects of type I IFN signaling (Rossi et al., 2017). As we were able to greatly enhance LgyLRV1+ induced iNOS induction by providing additional inflammatory stimuli ( Figures 2A–F ), we assumed that the iNOS-coping strategy of LgyLRV1+ lied within suppression of NF-kB activation. While artificial in our in vitro setting, such additional inflammatory stimuli can naturally arise from harmless commensals present in the skin and known to impact inflammatory reactions (Naik et al., 2012; Belkaid and Tamoutounour, 2016; Gimblet et al., 2017) or from sandfly gut microbiota egested together with the parasites (Dey et al., 2018). Such immunologically cross-reactive environment would serve as an additional stimulus for a parasite to evolutionary develop iNOS-mitigation mechanisms.

TNFAIP3, also known as A20, is a ubiquitin-modulating enzyme known to regulate NF-kB activity by removing the K63-linked polyubiquitin chains from RIPK1 and thus interfering with its interaction with IKK-γ (Karin and Greten, 2005), a crucial element in the classical NF-kB pathway. Several studies reported that Leishmania infection is able to promote A20 induction and thus control NF-kB activity and subsequently evade iNOS induction (Srivastav et al., 2012; Gupta et al., 2017). We recently showed that both, LgyLRV1+ and LgyLRV1- strains, induce A20 expression (Hartley et al., 2018). Notably, A20 production was more potent in the first hours of infection and was intensified in case of LgyLRV1+. This, together with the fact that A20 expression is itself under the control of NF-kB (Catrysse et al., 2014) suggested two waves of signaling. Firstly, infection with LgyLRV1+ led to production of type I IFN and activation of NF-kB that consequently led to A20 induction. Secondly, A20 suppressed NF-kB activity before the LRV1-triggered type I IFNs were sensed by IFNAR and led to formation of ISGF3 and induction of iNOS. To test this hypothesis we used BMDMs obtained from mice conditionally lacking A20 in myeloid lineage as full A20 knockout mice rapidly become cachectic and died within a few weeks of age, as a result of uncontrolled inflammation in several organs (Lee et al., 2011). We revealed that, indeed, cells lacking A20 expressed higher amounts of pro-inflammatory cytokines, iNOS and NO and were able to resist to LgyLRV1+ but not to LgyLRV1- infection ( Figures 5A–E ). While we could prove this hypothesis in vitro, in vivo experiments were impossible for ethical reasons as A20 deficient mice spontaneously developed polyarthritis over time (Vande et al., 2014) and any parasitic infection would extremely aggravate this effect. Taken together, we concluded that LgyLRV1+ parasites efficiently evaded LRV1-TLR3-IFNAR-triggered iNOS induction by activating A20 expression and thus suppressing NF-kB which is also required for potent iNOS production. Therefore, we provided, for the first-time evidence of how a symbiotic relationship between the protozoan parasite Leishmania and its dsRNA virus, LRV1, can lead to potentially mutually-detrimental effects of this cohabitation and how this mechanism was circumvented.

The exact mechanisms that govern parasite dissemination processes in MCL and DCL complications of the cutaneous form of the disease are largely unknown and the role of LRV1 is still controversial. Host-related immunological and environmental as well as parasite species-specific factors all have been linked to mediation of infectious metastasis, suggesting that this process is multifactorial and depends on a large number of players and their interaction (Hartley et al., 2014). LRV1, nested within Lgy and Lbr has been shown to be one of such factors and, perhaps, the main one among those associated with the parasite. In vivo experiments showed that presence of LRV1 strongly enhances metastatic activity of the parasites in hamsters and mice (Martinez et al., 2000; Ives et al., 2011; Hartley et al., 2016) although the latter were immunocompromised. LRV1 was found in various clinical isolates of Lgy and Lbr from Neotropical regions and was linked to the metastatic phenotype in patients (Cantanhêde et al., 2015), although this association was not exclusive (de Oliveira Ramos Pereira et al., 2013; Saberi et al., 2019; Valencia et al., 2022). These discrepancies further suggest that host-related factors such as the strength of the immune system seem to play a major role in mediating parasitic metastasis. We recently reported that patients, suffering from chronic forms of leishmaniasis caused by Lgy had decreased levels of IFN-γ and high levels of the inflammatory cytokine IL-17A. Moreover, these ratios were strictly dependent on whether parasites contained LRV1 or not (Hartley et al., 2016). Based on this we earlier established a murine model to study metastatic leishmaniasis and observed that absence, or low levels of IFN-γ, was a prerequisite for the metastatic effect whilst presence of LRV1 significantly intensified it. In the absence of IFN-γ, levels of IL-17A in the draining lymph nodes were highly elevated and seemed to govern the development of metastasis as genetic deletion or pharmacological blockade of this cytokine abolished the effect (Hartley et al., 2016). Similarly, involvement of IL-17A in metastatic processes is well documented in cancer research (Coffelt et al., 2015; Song et al., 2020; Zhao et al., 2020).

In our study we wished to determine whether LRV1-TLR3-IFNAR induced iNOS is able to control Leishmania infection in vivo. As STAT-1 activation, required for Nos2 transcription, could be induced by either type I or type II IFN signaling, we needed to exclude the latter in order to answer the aforementioned question. Thus, we used IFN-γ-deficient mice, in which the LRV1-TLR3-IFNAR axis should remain the predominant signaling path to induce iNOS. Additionally, we generated Ifn-γ ^-/- Inos ^-/- DKO mice that would lack iNOS and have the same genetic background. Both genotypes responded to infection in a similar pattern with no difference in footpad swelling or parasite burden ( Figures 4C, D ) and thus confirming the previously stated conclusion that LRV1-triggered iNOS is insufficient to control the infection due to suppression of NF-kB by parasite-induced A20. However, the difference that we observed was in the formation of metastatic lesions on the tail that should arise in IFN-γ-deficient mice as described above (Hartley et al., 2016). DKO mice developed less severe nodules on the tail and their formation was delayed by 1-2 weeks ( Figures 6A-C ). In line with this delay we observed that DKO mice had decreased levels of metastasis-mediating IL-17A in the draining lymph nodes compared to Ifn-γ ^-/- deficient mice ( Figure 6D ). This effect was transient as at the later time point no significant difference was observed ( Figure 6E ). This is consistent with the appearance of the metastatic nodules on the tails ( Figure 6C ). Although, we could not observe significant reduction of the parasite load on the secondary sites (tail) ( Supplementary Figure 4 ), our findings suggest that LRV1-triggred iNOS and the NO produced by it may have pro-metastatic effects in IFN-γ-deficient mice. This observation correlates well with the role of NO in the field of cancer metastasis. Pro-metastatic effects of NO are specifically attributed to the low concentrations (Tozer et al., 1997; Felley-Bosco, 1998; Cheng et al., 2014), while high concentrations have an opposite, anti-metastatic and tumoricidal effect (Juang et al., 1995; Dong et al., 1994). Our data lies within these concentration-dependent effects as LRV1-TLR3-IFNAR induced iNOS and subsequently produced NO are barely detectable in in vitro experiments and their amounts are insufficient to control the infection.

In our experiments we observed that depletion of the Nos2 gene in addition to Ifn-γ^-/- leads to decreased IL-17A production and delayed metastasis. There is little and contradicting evidence of a direct cause-effect association between iNOS and IL-17A. While some studies show the positive correlation between iNOS and stability of Th17 cells as well as production of IL-17A (Obermajer et al., 2013; Zhong et al., 2018), others show that NO negatively regulates these processes (Niedbala et al., 2011). While these discrepancies might also relate to the magnitude of NO, it is important to mention that we were unable to detect the exact T cell population that was responsible for IL-17A production in LgyLRV1+ infection (Hartley et al., 2016). Thus, the exact relation between LRV1-triggred iNOS and IL-17A production in LgyLRV1+ infection of Ifn-γ-/- mice is still unclear and requires further investigation.

In conclusion we demonstrated that LRV1-bearing Leishmania guyanensis induced A20 to minimize NF-kB activation and evaded death via iNOS induced by its viral endosymbiont. Additionally, we report that low amounts of NO induced as a result of the LRV1-TLR3-IFNAR signaling axis might correlate with the amounts of metastasis mediating IL-17A and thus impact parasite metastasis in an immunocompromised setting.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm.nih.gov/, GSE201120.

The animal study was reviewed and approved by All animal protocols were approved by the Swiss Federal Veterinary Office (SFVO), under the authorization licenses 2113 and 3551. Animal experimentation procedures were conducted with strict compliance with the ethical guidelines set out by the SFVO and under inspection by the Department of Security and Environment of the State of Vaud, Switzerland.

DK and NF designed research. DK, CD, FP, MR, RM, TS, and SC performed research. RO, L-FL, MP and SB contributed animals, reagents/analytic tools. DK analyzed data. DK and NF wrote the paper. All authors contributed to the article and approved the submitted version.

This work is funded by the grants from the Swiss National fund for research (http://www.snf.ch/en/Pages/default.aspx) (No. 310030_173180 and IZRJZ3_164176/1 to NF) and the National Institute of Health (NIH, https://www.nih.gov/) (R01AI-31078 and R01AI-30222-02, SB).

We thank Dr. Dimitri Moreau from the NCCR Geneva Access platform for his help with high-content imaging analysis. We are very grateful to Dr. Stephanie Claudinot and all the veterinary and animal care-taking staff of the University of Lausanne and especially Margaux Baudet for maintaining the highest ethical standards for animal experimentation. We are thankful to Rebecca Balz for technical help with some experiments and her cheerful spirit. We thank Dr. Nathalie Isorce, Dr. Amel Bekkar and Dr. Filipa Teixeira for providing RNAseq data. We also thank Dr. Slavica Masina and for careful reading of the manuscript.

The authors declare that the research was carried out in the absence of any commercial or financial relationships that could result in a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.