ROS regulate differentiation of visceralizing Leishmania species into the virulent amastigote form

Leishmania virulence and disease development critically depends on the ability of Leishmania promastigotes to infect, differentiate into amastigote forms and replicate inside mammalian host macrophages. Understanding changes associated with amastigote differentiation in axenic culture conditions is a key to identifying virulence factors. Here we compared efficiency of the conventional pH – temperature-dependent shift method to induce amastigote differentiation with the recently identified trigger for differentiation mediated by mitochondrial reactive oxygen species (ROS). Using two different visceral leishmaniasis species, L. infantum and. L. donovani , we show that ROS-generating methods such as iron deprivation or exposure to sub-lethal concentrations of hydrogen peroxide (H 2 O 2 ) or menadione are significantly more effective in promoting promastigote – amastigote differentiation than the low pH – high temperature shift, leading to higher survival rates, morphological changes and gene expression patterns characteristic of the amastigote stage. Notably, both H 2 O 2 and menadione-mediated differentiation did not require upregulation of the mitochondrial electron transport chain-associated protein p27, suggesting that treatment with oxidants bypasses the necessity to upregulate mitochondrial activity, a precondition for mROS generation. Our findings confirm that ROS-induced differentiation occurs in multiple Leishmania species, including the medically important visceralizing species, and provide mechanistic rationale for earlier reports demonstrating markedly increased virulence of L. infantum promastigotes pre-treated with oxidative reagents.


Introduction
Leishmania is an eukaryotic intracellular protozoan parasite that causes a broad spectrum of human diseases referred to as leishmaniasis, which currently affect about 12 million people and threaten more than 350 million globally (WHO, 2010). Of the three main pathological forms of the diseasecutaneous, mucocutaneous and visceralthe visceral form caused by L. infantum (in South America and in the Mediterranean) and L. donovani (India and parts of Europe and Africa) is the most severe. If left untreated, visceral leishmaniasis has a high fatality rate, ranking next to malaria as the deadliest protozoan human disease (Torres-Guerrero et al., 2017). The life cycle of Leishmania parasites alternates between insect vectors and mammalian hosts, with intra-species transfer occurring through sand fly bites. Parasites in the promastigote phase are elongated in shape and have long, easily distinguishable flagella, whereas those in the amastigote phase are more spherical and have almost indistinguishable flagella. The transition from the flagellated, non-virulent promastigote insect stages to the rounded, virulent amastigote forms that replicate inside phagolysosomes of mammalian macrophages is critical for Leishmania virulence (Killick-Kendrick, 1990), and remains the 'holy-grail' for the identification of virulence factors and therapeutic targets. The ability to successfully induce promastigote to amastigote differentiation in axenic culture by mimicking some of the intra-phagolysosomal conditions inside host macrophages (namely, low pH and elevated temperature) has greatly aided the understanding of cellular changes that accompany parasite differentiation (Bates et al., 1992;Zilberstein and Shapira, 1994;Barak et al., 2005). However, these temperature-/pH-dependent protocols for axenic differentiation and culture of amastigote forms have been difficult to adapt for many Leishmania species, including isolates of the clinically important L. infantum and L. donovani visceralizing species. Although the successful conversion of L. chagasi (same as L. infantum) promastigotes to amastigote forms in axenic culture was previously reported (Hsiao et al., 2008), several published methods for differentiating L. infantum promastigotes into amastigotes in axenic culture involved a pH shift that varied between 5.5 and 6.5 and a temperature elevation that ranged from 26 to 37°C (Sereno and Lemesre, 1997;Somanna et al., 2002), and most studies lacked a demonstration that the parasites retained viability under these conditions. Sudden changes in pH and temperature have been reported to result in loss of Leishmania viability as a result of a high reactive oxygen species (ROS) surge (Alzate et al., 2006(Alzate et al., , 2007. In spite of the general perception that ROS radicals are largely harmful and deleterious for cells, it has become increasingly clear that low concentrations of ROS such as superoxide and hydrogen peroxide (H 2 O 2 ) are essential for maintaining homoeostasis and for cell fate determination (Tsukagoshi et al., 2010;Tatapudy et al., 2017). The role of ROS molecules, particularly H 2 O 2 , as a differentiation signal has been demonstrated in many eukaryotic cells (Marinho et al., 2014;Sies, 2014;Lennicke et al., 2015), including the nonvisceralizing Leishmania species L. amazonensis (Mittra et al., 2013(Mittra et al., , 2017. H 2 O 2 -mediated signalling in L. amazonensis has been shown to play a critical role in the development of infectivity for bone-marrow-derived macrophages or mice, by promoting promastigote to amastigote differentiation independently of a shift in pH and temperature (Mittra et al., 2013). Successful differentiation into amastigotes form was observed under conditions that promote conversion of mitochondria-generated superoxide radicals into H 2 O 2 by the iron-dependent superoxide dismutase (SODA), through a process that can occur metabolically or via iron depletion from the culture medium (which results in an increase in iron uptake by upregulating transporters). Axenic differentiation into amastigotes was also achieved by directly treating L. amazonensis promastigotes with H 2 O 2 or sub-lethal concentrations of menadione, a drug that generates mitochondrial superoxide (Mittra et al., 2013).
In this study, we examined if H 2 O 2 -mediated signalling is also involved in triggering amastigote differentiation of other Leishmania species, in particularly the medically important visceralizing species. Our study was inspired by a previous study that reported significant enhancement of virulence in mice following treatment of L. chagasi (now designated as L. infantum) promastigotes with sub-lethal levels of menadione or H 2 O 2 (Wilson et al., 1994). In light of the more recent findings that ROS triggers the differentiation of infective amastigotes in L. amazonensis, we hypothesized that the increase in virulence reported by Wilson et al. (1994) in L. infantum might be explained by a similar mechanism. To test this hypothesis, we compared the efficacy of the conventional pH/temperature-mediated differentiation protocol to treatments that induce ROS signalling. We found that both H 2 O 2 and menadione at specific concentrations induce the promastigote to amastigote morphological transition, which is marked by loss of the long flagella and reorganization into rounded forms at 24 h following treatment. Expression of amastigote-specific genes, including the gene encoding A2, is also upregulated by exposure to the ROS-generating agents. Thus, our findings provide important mechanistic insight into the previously reported oxidative stress-induced enhancement of L. infantum virulence (Wilson et al., 1994), and demonstrate that the previously described role of ROS in regulating the development of virulent amastigote forms (Mittra et al., 2013) can be applied more broadly to include the visceralizing species L. infantum and L. donovani, which cause very severe forms of human disease.

Primers
All primers used in this study were designed based on Leishmania species-specific genomic sequences from TriTrypDB (http://tritrypdb.org/tritrypdb/). Primer efficiency was calculated by performing quantitative PCR after serial dilution of the template genomic DNA and calculating the efficiency using the Bio-Rad CFX manager programme. All primers were within an acceptable range for qPCR (90-110% as recommended by the manufacturer). See online Supplementary Table S1 for sequences and primer efficiency.

ROS treatment
Promastigotes in the log phase of growth were seeded in media containing either H 2 O 2 (diluted with sterile water) or menadione (Sigma-Aldrich, St. Louis, MO, USA; diluted with DMSO), at an initial concentration of 5 × 10 6 cells mL −1 . Menadione is a drug that causes accumulation of superoxide in the mitochondria through redox cycling. Cells that are denoted as a 'no drug' sample were treated with an equal volume of sterile water or DMSO, respectively. Samples were incubated in a total volume of 2 mL per sample at 26°C.
Cell viability assays and determination of elongated/ flagellated vs. rounded/non-flagellated cells Cell viability was determined by staining with fluorescein diacetate (FDA) (Sigma-Aldrich, St. Louis, MO, USA), a membranepermeable molecule that becomes fluorescent after being cleaved intracellularly in metabolically active cells (Sacks and Melby, 2001). FDA was prepared at a stock concentration of 5 mg mL −1 in acetone. This stock was diluted in PBS to get a working concentration of 0.02 mg mL −1 . Equal volumes of sample and a working stock of FDA were mixed and incubated for 5 min at room temperature. Ten microlitres of this FDA/cell mixture was loaded into a haemocytometer. Cells that fluoresced green were counted microscopically using 40× (0.75) objective of a Nikon Eclipse 200 microscope, and noted as being elongated/flagellated or rounded/non-flagellated based on the cell morphology and visibility of flagellum.
Fluorescence and scanning electron microscopy DIC and FDA fluorescence images were acquired with a Nikon Eclipse Ti using a 40× (1.40) objective. After the various treatments, the parasites were washed twice with 1× PBS and resuspended in 100 µL of 1× PBS. An equal volume of FDA (0.02 mg mL −1 ) was added, followed by imaging.
Scanning electron microscopy (SEM) was performed as previously described (Mittra et al., 2016). Briefly, 3 × 10 6 cells were centrifuged at 2000 × g for 5 min. Cells were resuspended in 3 mL of PBS+/+ (supplemented with calcium and magnesium), added to poly-L-lysine-coated coverslips and incubated for 15 min. After two washes with PBS+/+, samples were fixed in 2.5% (v/v) glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.4 for 60 min and post-fixed with osmium tetroxide in the same buffer for 1 h at room temperature. After sputter coating with Au/Pd, the preparations were imaged using a variable pressure Hitachi SU3500 SEM.

Iron depletion and supplementation
Iron-depleted promastigote growth media was prepared as described before (Mittra et al., 2013). Iron-depleted FBS was generated by adding 10 mM ascorbic acid to 100 mL of FBS for 6 h at 37°C, until the OD reading at 405 nm was reduced to 0.6-0.8. Following this, 5 g 100 mL −1 Chelex® (Bio-Rad) was added and stirred at 50 rpm for 3-4 h and then filtered through Whatman paper and dialysed (2000D cut-off) against 4 L of cooled, sterile PBS with three buffer changes every 6 h. Iron-depleted FBS was then filter sterilized (0.22 µm filter pore size) and stored at −20°C. Iron-depleted media was generated in the same manner as culture media, except that iron-depleted FBS was used and hemin (another source of iron) was excluded. The media was stirred with 5 g 100 mL −1 Chelex® for 1 h at room temperature and filter sterilized. Iron supplementation was done by adding Fe-NTA at a concentration of 8 µM.

Determination of gene expression by qPCR
Equal numbers of cells per biological replicate (1-5 × 10 7 cells) were spun down at 2000 × g for 5 min. Cells were washed with PBS twice and RNA was isolated using the RNeasy kit (QIAGEN, Hilden, Germany) as per the manufacturer's protocol. Both on-column and post-isolation RNase-free DNase (QIAGEN, Hilden, Germany) digestion were performed to ensure complete removal of any contaminating genomic DNA. Equal amounts of RNA for each biological set were used for cDNA synthesis (Bio-rad iScript). cDNA was diluted to 1 ng µL −1 concentrations for each sample and amplified with gene-specific primers using the Sybr Green master mix, according to the manufacturer's protocol. A CFX96 qPCR instrument (Bio-Rad, Hercules, CA, USA) was used and three technical replicates as well as a no template control were run with each plate. Primer efficiency was calculated (see online Supplementary Table S1) and gene expression was normalized to ubiquitin hydrolase, which is known to be expressed constitutively (Rochette et al., 2008;Depledge et al., 2009).

Iron deprivation generates amastigote-like parasites more effectively than traditional methods
We tried several conditions for differentiating L. infantum amastigotes in axenic culture, all involving shifting of promastigotes in culture (pH 7.4/26°C) to low pH/high temperature to mimic intra-macrophage conditions that are conducive for amastigote growth and replication (Sereno and Lemesre, 1997;Somanna et al., 2002). All the protocols we tried, which mainly differed in amastigote media composition with small variations in pH or temperature, were not efficient in generating viable amastigotes (data not shown) when viability was estimated using FDA staining assay (Sacks and Melby, 2001). This variability between our results and those reported in the literature could be related to differences in the parasite strains used in different laboratories. Our best results were achieved using an amastigote media composition similar to that used for L. amazonensis (Huynh et al., 2006), but with 5.5 as final pH and incubation at 36°C for culture growth. Transfer of log-phase promastigotes to amastigote culture conditions (36°C/pH 5.5) for 24 h resulted in about 43% decrease in the number of FDA-stained viable cells, when compared with the 0 h time point (Fig. 1A). Microscopic examination showed that about 60% of the FDA-positive cells had lost their long flagella and developed the rounded, amastigote-like morphology (Fig. 1B). On the other hand, cells kept in promastigote growth conditions (26°C/pH 7.4) replicated normally, with about 4-fold increase in viable cell count between 0 and 24 h and showing no change in the flagellation pattern ( Fig. 1A and C). Microscopic examination confirmed that L. infantum promastigotes maintained in promastigote culture conditions maintained their elongated, flagellated morphology and divided actively, while a shift to amastigote growth conditions caused cells to round up and lose visible flagella. However, a part of the cell population also showed increased darkening and loss of FDA staining, indicating loss of viability (Fig. 1D). This demonstrates the traditional low pH, high temperature shift method is not effective for generating viable axenic amastigotes of L. infantum, at least for the isolate used in our study.
We also transferred L. infantum promastigotes from regular culture media to iron-depleted media, and monitored changes in growth rate and cell morphology (loss of elongated flagella in particular) for several days. As previously reported, the transfer of L. amazonensis promastigotes growing in regular promastigote media to iron-depleted media is effective in initiating the amastigote stage differentiation process, through a ROS signalling pathway (Mittra et al., 2013(Mittra et al., , 2016(Mittra et al., , 2017. Similarly, transfer to iron-depleted media lowered the growth rate of L. infantum promastigotes ( Fig. 2A) and induced the appearance of distinct amastigote-like morphology on day 4, with approximately 70% of the population becoming non-flagellated by day 5 (Fig. 2B and C). Positive staining

Parasitology Open 3
with the viability dye FDA confirmed that there was no significant loss of viability under these conditions (Fig. 2B). As previously reported for L. amazonensis (Mittra et al., 2013), add-back of iron (8 µM of Fe-NTA) to the iron-depleted media promoted promastigote growth to levels similar to regular media, and reduced the number of rounded/non-flagellated forms on days 4 and 5 of culture. This result strongly suggests that the iron-dependent, ROS-mediated amastigote differentiation signalling pathway previously described for L. amazonensis might also be functional in visceral leishmaniasis species.
Exposure of L. infantum to ROS causes them to differentiate into an amastigote-like form To directly test the potential role of ROS signalling in initiating L. infantum amastigote differentiation, we treated log-phase promastigotes cultured in regular iron-containing media (20% FBS) with an incremental concentration gradient of H 2 O 2 . Viable cells were counted at the 24 h time point following the treatment, which corresponded with the highest incidence of morphological alterations. The appearance of rounded/non-flagellated forms was less prominent at later time points. As H 2 O 2 is rapidly degraded by catalase present in the FBS component of the parasite growth media, the loss of effectivity after 24 h is not surprising. Based on our previous findings with L. amazonensis, if H 2 O 2 is a signal for promastigote to amastigote differentiation, we expected promastigotes to lose their elongated flagella, assume an amastigotelike rounded morphology and exhibit a decreased doubling time at the optimal H 2 O 2 concentration. We also expected that overtreatment would shift the amastigote differentiation signal to signals for apoptosis or necrosis. Thus, this treatment with a gradient would determine the ideal conditions for morphological change and viability. Consistent with the stated hypothesis, the treatment caused an arrestment of cell division and eventually viability in L. infantum, commensurate with the concentration of H 2 O 2 added (Fig. 3A). As the number of viable cells decreased, the number of non-flagellated cells increased as a function of the concentration of H 2 O 2 (Fig. 3B). Parasites treated with no drug or 100 µM H 2 O 2 did not exhibit a significant change in morphologyretained their flagella, elongated shape and divided at the normal promastigote rate (12 h doubling time). Doublets of head-to-head joined cells, representing actively dividing Leishmania promastigotes, were visible under both conditions (no drug or 100 µM H 2 O 2 ). Based on both viability and morphological change, the optimal concentration of H 2 O 2 to induce L. infantum amastigote differentiation appeared to be 200 µM. At this concentration, approximately 42% of the parasite population had lost their elongated flagella and rounded up into an amastigote-like form, indicating successful initiation of differentiation, without any loss in viability. The majority of the parasites treated with 250 µM H 2 O 2 showed a distinct amastigote-like phenotype, but also showed decreased viability (compared with the 0 h time point), with cell-clumping becoming more apparent. Higher concentrations of H 2 O 2 (300 and 500 µM) resulted in almost 100% cell death at 24 h following treatment (Fig. 3C). While H 2 O 2 treatments represented an exogenous source of ROS signal, we also treated L. infantum promastigotes with varying concentrations of menadione, a mitochondrial superoxidegenerating drug, to mimic metabolic mitochondrial ROS signals generated through the electron transport chain (ETC) activity (Loor et al., 2010). Menadione, a redox cycler, induces superoxide production by transferring electrons from the ETC to available oxygen molecules, which are then converted to H 2 O 2 by the mitochondrial SODA (for Leishmania spp.). Log-phase L. infantum promastigotes cultured in regular media (20% FBS) were treated with a gradient of menadione concentrations, and cell viability was determined using FDA staining at 24 h after treatment. Treatment of promastigotes with DMSO alone (no drug), or 1 µm of menadione did not cause any marked change in viability or morphology (Fig. 4A and B), as the cells remained elongated and divided actively. The optimal condition was achieved at a menadione concentration between 3 and 5 µM. Starting with 3 µM menadione, parasite growth rate was slowed down and around 21% of the cells lost the elongated flagella and exhibited amastigote-like morphology. At 5 µM menadione concentration, there was a complete cessation of cell division, with >80% of the parasites losing their elongated flagella and rounding up. No significant loss in viability (as compared with the 0 h time point) was detected under these conditions, even though low amounts of cell clumping was present in some of the replicate samples. Parasite populations treated with 7.5 and 10 µM of menadione showed complete loss of elongated flagella, but cell clumping became more apparent (Fig. 4C). These data clearly demonstrate the effectivity of ROS-generating treatments in initiating L. infantum amastigote differentiation, and is in accordance with our previous reports of menadione-induced promastigote to amastigote differentiation in L. amazonensis (Mittra et al., 2013(Mittra et al., , 2016(Mittra et al., , 2017. SEM was used to confirm that the morphological changes observed under DIC microscopy were consistent with that of an amastigote form. Leishmania infantum was treated with either agent, 200 µM H 2 O 2 or 5 µM menadione. As expected, untreated cells remained elongated and flagellated, while parasites exposed to either H 2 O 2 or menadione experienced a loss of long flagella, cell rounding and a decrease in size, all hallmarks of the amastigote form (Fig. 5).
Amastigote stage-specific gene transcripts are upregulated following ROS treatment or shift to low-pH/high-temperature conditions To confirm if ROS-induced amastigote-like morphological transition of L. infantum promastigotes is orchestrated through genetic controls, we compared changes in expression levels of several stage-specific transcripts at 24 h following exposure to H 2 O 2 or menadione by quantitative RT-PCR (qRT-PCR). Steady-state transcript levels for A2, a protein exclusively expressed during the amastigote stage of visceral Leishmania strains like L. infantum, but barely detectable in promastigote form (Zhang et al., 1996;Zhang and Matlashewski, 1997) were compared between untreated cells and cells treated with either 200 µM H 2 O 2 (Fig. 6A) or 3 µM menadione (Fig. 6B), as parasite health and viability was uncompromised at these concentrations. Upregulation of A2 transcript levels was detected in H 2 O 2 and menadionetreated cells (>3-and >2-folds, respectively). Orthologues of AML1 and AML2, the 'amastin-like' genes known to be

Parasitology Open 5
upregulated during amastigote differentiation (Lahav et al., 2011;Mittra et al., 2013), also showed a relative increase in transcript levels but was not statistically significant. No noticeable change in transcript levels for ATG8 was noted. Notably, an increased transcript level for this autophagy gene was observed in L. amazonensis under low iron conditions and following menadione treatment, but not with H 2 O 2 (Mittra et al., 2013). Although we noticed a trend in gene expression changes similar to what occurs during differentiation of the amastigote stage, the changes were not significant in the samples treated with 3 µM menadione. This is likely due to the fact that this lower drug concentration was not sufficient to initiate differentiation. However, when L. infantum promastigotes were treated with a higher concentration of menadione (10 µm), all amastigote stage-specific genes tested showed considerable upregulation ( Fig. 7) with the exception of AML1. As expected, transcript levels for the para-flagellar rod protein (Pr2C) were downregulated, consistent with the loss of the elongated flagella typical of promastigotes.
Comparative analysis of transcript levels using L. infantum promastigotes induced to differentiate under low-pH/elevated temperature conditions showed a similar trend for A2, AML1, AML2 and ATG8 upregulation, while Pr2C was downregulated (Fig. 6C), even though RNA samples were isolated from cell samples collected at a much earlier time point (8 h, compared with 24 h). Analysis of 24 h samples was avoided because of the reduced viability at that time. A very interesting difference in the gene expression pattern induced by ROS exposure, compared with the conventional low pH/high temperature shift method for amastigote differentiation, was observed for transcript levels of the virulence-determining protein p27, which is an essential component of the active cytochrome c oxidase (COX) at the amastigote stage. Expression of p27 increases the overall ETC activity and promotes oxidative phosphorylation for ATP production, as the parasites shift towards a mitochondria-dependent mode of energy generation (Dey et al., 2010). A 6-fold upregulation of p27 expression was observed with the low-pH/high-temperature shift, whereas treatment with either 200 µm H 2 O 2 or 10 µm menadione resulted in downregulation (Figs 6D and 7). This result suggests that while p27-mediated enhancement of ETC activity is necessary for generating superoxide during low-pH/high-temperature exposure, but its expression is not necessary and can be bypassed during treatment with the superoxide-inducing drug menadione or direct exposure to H 2 O 2 .

Leishmania donovani also differentiates into amastigotes in response to ROS treatment
To further validate our findings that ROS exposure can trigger amastigote differentiation and virulence in visceralizing Leishmania species, we tested L. donovani, the causative agent of the highly lethal kala-azar. Leishmania donovani promastigotes in culture (pH7.4; 26°C temperature) were treated with either 200 µM H 2 O 2 or 5 µM menadione, exactly as described earlier.
Both H 2 O 2 and menadione treatments resulted in attenuation of cell division (compare 24 h carrier treatments as control to 24 h H 2 O 2 or menadione-treated samples) but did not affect overall viability of the parasites in culture (compare 0 h carrier treatments as control with 24 h H 2 O 2 or menadione-treated samples), was confirmed by FDA staining (Fig. 8A, B and E). A significant increase in the number of non-flagellated cells, consistent with promastigote to amastigote transition, was also observed in promastigote cultures 24 h following H 2 O 2 or menadione treatment (Fig. 8C-E), similar to that seen with L. infantum. SEM examination of parasites clearly showed loss of the elongated flagellar structure, reduction in body length and appearance of amastigotelike morphology following both ROS treatments (Fig. 8F). Collectively, our results identify ROS-mediated signalling as a critical step for differentiation and virulence development in visceralizing Leishmania species.

Discussion
There has long been evidence for an interplay between ROS and Leishmania infections. On their initial encounter with host macrophages, invading pathogens can be exposed to a macrophage-mediated ROS burst that acts as a first line of innate immune defence. Leishmania, however, has developed a unique 'Trojan-horse' mechanism for muting this ROS burst, gaining entry inside and colonize macrophages (Murray, 1981). Successful   6. Both ROS and low-pH/high-temperatureinduced differentiation result in upregulated amastigote-specific genes but differ in p27 dependence. Mid-log phase L. infantum promastigotes were resuspended at a concentration of 5 × 10 6 cells mL −1 in regular medium and treated with either 200 µM H 2 O 2 (A) or 3 µM menadione (B) for 24 h, or resuspended in amastigote media at 36°C for 8 h (C). Changes in expression levels of genes associated with amastigote differentiation were determined by qRT-PCR using RNA isolated from cells collected at the indicated time points. Fold changes were expressed relative to water as control for H 2 O 2 -treated cells, (A) DMSO as control for menadione-treated cells and (B) cells in promastigote media (26°C/pH7.4) as control for transfer to amastigote media (pH 5.5) at 36°C (C). (D) Determination of changes in expression levels of mitochondrial COX-associated gene p27. Results are representative of four biological replicates for H 2 O 2 treatment, three biological replicates for menadione treatment and shift to amastigote conditions, respectively. Error bars represent standard error of measurement. Student's two-tailed t-test was used to assess significance. P values are indicated in figure. invasion depends on the ability of invading parasites to upregulate their anti-oxidant defence systemas evidenced by increased expression of SODA and peroxidases during infective metacyclic promastigote and subsequent amastigote stages. Priming of the anti-oxidative system primarily occurs via increased metabolic oxidative stress and is considered as key for virulence development (Alzate et al., 2007). Notably, a role for subtle changes in ROS levels in controlling cell fate has been demonstrated, particularly for the H 2 O 2 -induced triggering of promastigote to amastigote differentiation (Mittra et al., 2017). The molecular targets of H 2 O 2 and other details of this signalling cascade remain to be elucidated, but the effect of ROS on integral differentiation processes such as autophagy is well established (Williams et al., 2009;Mittra et al., 2013).
A role of mitochondria-generated superoxide and H 2 O 2 in determining virulence have also been demonstrated in L. amazonensis (Mittra et al., 2013(Mittra et al., , 2016(Mittra et al., , 2017. Superoxide production occurs through mitochondrial ETC activity and is enhanced under stress. Thus, the available evidence indicates that the ability of L. amazonensis to regulate the surge in superoxide levels, through enzymatic conversion into H 2 O 2 enables differentiation of promastigotes into amastigotes in axenic culture. Previous studies also showed that conditions that induce mitochondrial stress and ROS production in promastigotes, namely iron depletion or transfer to low pH/elevated temperature, also efficiently generate axenic amastigotes in L. amazonensis. Direct implication of ROS in such stress-related differentiation mechanism comes from direct exposure to H 2 O 2 or menadione, which also promoted amastigote differentiation and virulence development in those studies. Here we showed that when subjected to similar conditions of stress including iron depletion, L. infantum and L. donovani show upregulation of amastigote stage-specific transcripts, downregulation of promastigote-specific transcripts and a clear transition into the amastigote morphology, thereby confirming that ROS-mediated regulation of virulence is also functional in visceralizing Leishmania strains. Importantly, our findings also provide a mechanistic explanation for earlier reports of enhanced infectivity of L. infantum (at the time referred to as L. chagasi) promastigotes following treatment with ROS inducers. Parasites treated with H 2 O 2 or menadione showed a dose-dependent increase in infectivity, compared with the DMSO-treated control, in assays quantifying the parasite load in the liver and spleen of mice (Wilson et al., 1994). Pre-exposure of the parasites to a sub-lethal oxidative stress resulted in increased viability and resistance to exposure to increasing concentrations of H 2 O 2 . However, no promastigote to amastigote morphological change associated with differentiation was reported under their experimental conditions. Interestingly, upregulation of the heat-shock protein hsp70 expression observed in that study suggested a possible link between traditional differentiation methods (heat shock) and ROS accumulation, a suggestion later confirmed by reports of mitochondrial hyper-polarization and associated increase in superoxide levels during heat shock of L. infantum promastigotes (Alzate et al., 2006(Alzate et al., , 2007. Our results provide further evidence for roles of both ROS and heat shock in L. infantum differentiation. By carefully monitoring cell morphology and stage-specific gene expression during exposure to incremental concentrations of H 2 O 2 and menadione, we have established that at optimal concentrations, both treatments are very effective in inducing promastigote to amastigote differentiation without a loss in viability. Our failure in repeated attempts to generate viable axenic amastigotes using the low pH/elevated temperature method described in earlier reports could possibly be ascribed to strain differences, which might respond differently to the variable pH and temperature conditions used in those studies (Sereno and Lemesre, 1997;Somanna et al., 2002). On the other hand, we did observe a morphological change and upregulation of amastigoteassociated gene expression in L. infantum after 8 h of pH/temperature shift, consistent with initiation of the differentiation process. However, the marked loss in viability we observed under these conditions, which may have been overlooked in previous studies, might have resulted from oxidative damage, if the levels of ROS triggered by the pH/temperature stress were high enough to overwhelm the parasite's anti-oxidative machinery.
An additional finding that emerged from this study is the inverse trend in p27 expression observed between pH/temperaturemediated and ROS-induced differentiation. This observation reveals a difference between the two processes in their dependence on mitochondrial ETC for the generation of superoxide radicals. While p27-mediated upregulation of COX activity is necessary Fig. 7. Leishmania infantum parasites treated with 10 µM menadione fully differentiate into amastigotes. DIC microscopy and FDA staining confirms that parasites treated with 10 µM menadione are fully differentiated, with some cell clumping. All amastigote markers, with the noted exception of p27, are upregulated and the promastigote marker, Pr2C, is downregulated. Experiments consist of three biological replicates assayed in triplicate. Error bars represent standard error measurement. Bar = 14 µm. Student's two-tailed t-test was used to assess significance. P values are indicated in figure. 8 Yousuf A. Khan et al.
for enhancing ETC function, to meet the increasing demand for energy production in metacyclic promastigote and amastigote stages (Dey et al., 2010;Saunders et al., 2014), it also serves as a potent source of mitochondrial superoxide that is critical for initiating the ROS signalling cascade. Direct treatment of promastigotes with H 2 O 2 or the superoxide generating drug menadione, on the other hand, bypasses that requirement. In fact, the downregulation of p27 expression observed with higher concentrations of menadione may represent a regulatory mechanism to control ROS levels, by slowing down ETC activity. Taken together, our findings significantly expand the importance of the H 2 O 2 signalling pathway in promoting Leishmania virulence, by demonstrating the effectiveness of this pathway for triggering amastigote differentiation in two clinically important visceralizing species. Our study also outlines a simple and effective method for generating axenic amastigotes in these Leishmania species without the need for stressors such as pH and temperature changes, and opens up the possibility for exploring anti-oxidants as anti-leishmanial therapeutic agents.