Serine proteases profiles of Leishmania (Viannia) braziliensis clinical isolates with distinct susceptibilities to antimony

Glucantime (SbV) is the first-line treatment against American Tegumentary Leishmaniasis. Resistance cases to this drug have been reported and related to host characteristics and parasite phenotypes. In this study, 12 Leishmania (Viannia) braziliensis isolates from patients that presented clinical cure (Responders—R) and relapse or therapeutic failure (Non-responders—NR) after treatment with antimony, were analyzed. These parasites were assessed by in vitro susceptibility to SbIII and SbV, serine proteases activity measured with substrate (z-FR-AMC) and specific inhibitors (TLCK, AEBSF and PMSF). In vitro susceptibility of axenic amastigotes to SbIII showed a significant difference between R and NR groups. The protease assays showed that TLCK inhibited almost 100% of activity in both axenic amastigotes and promastigotes while AEBSF inhibited around 70%, and PMSF showed lower inhibition of some isolates. Principal component and clustering analysis performed with these data yielded one homogeneous cluster with only NR isolates and three heterogeneous clusters with R and NR isolates. Additionally, differential expression of subtilisins (LbrM.13.0860 and LbrM.28.2570) and TXNPx (LbrM.15.1080) was evaluated in promastigotes and axenic amastigotes from both groups. The results showed a higher expression of LbrM.13.0860 and LbrM.15.1080 genes in axenic amastigotes, while LbrM.28.2570 gene had the lowest expression in all isolates, regardless of the parasite form. The data presented here show a phenotypic heterogeneity among the parasites, suggesting that exploration of in vitro phenotypes based on SbIII and serine proteases profiles can aid in the characterization of L. (V.) braziliensis clinical isolates.


Introduction
Leishmaniasis is a neglected tropical disease that affects around 1 million people worldwide, causing 26 000 to 65 000 deaths per year (1). Leishmania spp. are eukaryotic protozoan parasites responsible for a broad range of clinical manifestations, such as cutaneous, mucocutaneous and visceral forms, which depend on the parasite species and immune state of the mammalian host (2). In the American continent, Leishmania mexicana complex and subgenus Viannia species cause cutaneous and mucocutaneous forms that can develop into localized, disseminated or diffuse lesions, which is known as American Tegumentary leishmaniasis (ATL) (1).
A pentavalent antimony formulation -Glucantime® (Sb V ) -is the rst-line treatment against ATL and has not changed in the last 70 years (1). In South America, this treatment doses (low and high dose) vary according to therapeutic response patterns seen in speci c geographical areas (3)(4)(5). Its mechanism of action is still not completely understood, but, it has been well accepted the hypothesis that Sb V needs to be reduced to its trivalent form (Sb III ), either directly by the parasite or within the host macrophages, to excel leishmanicidal activity (6)(7)(8).
Moreover, the most abundant Leishmania spp. thiol, trypanothione reductase [T(SH) 2 ], is part of the parasite trypanothione reductase system which is a unique detoxi cation defence mechanism that relies on four key enzymes: trypanothione synthetase (TryS), trypanothione reductase (TR) and tryparedoxin (TXN) and tryparedoxin peroxidase (TXNPx) (9). This system is involved in the detoxi cation of metal ions, including reduction of Sb V to Sb III , within the macrophage phagolysosome (6,7). It is also hypothesised that Sb V reduction might happen through the oxidation of T(SH) 2 forming a stable complex -Sb III (TS) 2 -that could be pumped out of the cells through parasite membrane transporters (6,9).
Interestingly, parasite's resistance towards Sb V has been associated with increased trypanothione levels and a decreased capacity to reduce Sb V (10)(11)(12). Indeed, increased abundance of TXNPx has been correlated with an enhanced thiol redox potential in Sb V resistant parasites, not only in lab-generated strains (13,14) but also in clinical isolates (15). Furthermore, some studies have shown that parasite enzymes are involved in the reduction of Sb V to Sb III and Leishmania spp. resistance towards Sb V (10,16,17). Particularly, serine proteases have been extensively studied in American Leishmania spp. due to their functional interrelationship in parasite physiology and their potential as therapeutic targets (18)(19)(20)(21)(22)(23)(24). Experimental evidence has demonstrated that these enzymes can act as signal peptidases, maturases of other proteins, and can have a metacaspase-like activity (20,25). Among them, subtilisins are proposed to have a functional interrelationship with other Leishmania spp. proteins, suggesting key functions of these enzymes in parasite physiology such as modulation of the trypanothione reductase system through direct action over TXNPx (18,25).
In this context, subtilisins are correlated to the balance of cytosolic and mitochondrial TXNPx levels in L. (L.) donovani clinical isolates. Subtilisin knockout promastigotes of these parasites failed to differentiate into viable amastigotes and their TXNPx peptide abundance was decreased, which suggested that subtilisins can act as maturases of speci c proteins or pathways (25). Although, in L. (V.) braziliensis, subtilisins have distinct subcellular distributions and expression (26) it has not been assessed yet if these enzymes excel a similar role during parasite differentiation; and if impairment of trypanothione reductase system has any relation with Sb V resistant phenotypes.
This study aims to contribute in the understanding of Glucantime® susceptibility based on L. (V.) braziliensis clinical isolates. Gathered data here incorporate new information on the heterogenic pro le of clinical isolates, assessing their in vitro susceptibility towards Sb V and Sb III as a direct relation to serine proteases of this parasite. Additionally, transcripts of some enzymes of the parasite detoxi cation system as subtilisins (LbrM.13.0860 and LbrM.28.2570) and TXNPx (LbrM.15.1080) were accessed.

Results
In vitro susceptibility to antimony Both, Sb III and Sb V , were unable to discriminate susceptibility pro les between promastigote forms of responders (R) and non-responders (NR) isolates. The mean IC 50 value of R versus NR group was very close which did not allow discrimination among them ( Fig. 1A and 1B). Meanwhile, the results showed a signi cant difference between the axenic amastigotes grouped as NR and R exposed to Sb III (p < 0.05, Fig. 1C). Conversely, these forms did not respond well to Sb V exposure and we were not able to discriminate among the groups (Fig. 1D). The IC 50 values are summarised in Supplementary le 1.
The inhibition percentage of each serine inhibitor was calculated using each isolate W/i as a comparison (Supplementary le 3). TLCK showed the strongest inhibition with almost 100% of inhibition for R1, R2, R3, R4, NR1 and NR2 both in promastigotes and axenic amastigotes extracts. Similarly, AEBSF strongly inhibited the same group of isolates (70 %) but only for the promastigotes' extracts, while for the axenic amastigotes extracts the inhibition was dispersed, only R1, R3 and NR6 were 70 % inhibited. On the contrary, PMSF showed the lowest inhibition rates. In promastigotes extracts of R3, R4, NR6 and NR7, and axenic amastigotes extracts of R3, R5, NR3, NR6 and NR7 it inhibited less than 35 %. Altogether these results indicated that these isolates have a distinct quantitative pro le for serine protease activities in both assayed parasite forms.

Cluster analysis
The rst three PCs explained approximately 80 % of the data variance. Based on our analysis of the total sum of squares as a function of the number of clusters (Supplementary le 4) we opted for using ve clusters: Cluster 1, Cluster 2, Cluster 3, Cluster 4, and Cluster 5 (Fig. 3). There was one cluster that contained, exceptionally, only one clinical isolate (Cluster 3: NR3). One homogenous cluster containing two isolates from the same clinical group (Cluster 2: NR1, NR4) while the remaining three were heterogeneous (Cluster 1: NR6, R5; Cluster 4: R1, R3, NR5, NR7; and Cluster 5: NR2, R2, R4).
Interestingly, in the heterogeneous Cluster 1, the distances between both members for all variables ranged from 0.2 to 0.7, except for TLCK inhibition over promastigotes (0.04) and axenic amastigotes (0.9) (Supplementary le 5). Moreover, within Cluster 4, another heterogeneous cluster with two R and two NR isolates, TLCK inhibition between R and NR was the highest for both promastigotes and axenic amastigotes, while, PMSF and AEBSF values ranged from 0.06 to 0.85 without an established pattern (Supplementary le 6). Also, in Cluster 5, a cluster containing two R and one NR isolate, there was a pattern among variables with all distance values lower than 0.56, exceptionally, TLCK inhibition of NR3 versus R6 axenic amastigotes being the highest distance (0.72) (Supplementary le 7). Regarding Sb III distance within Cluster 4 and Cluster 5, there is no value higher than 0.43.

Discussion
The hypothesis that L. (V.) braziliensis clinical isolates have, indeed, different response pro les towards antimony was corroborated in this study by in vitro susceptibility experiments using Sb V and Sb III . It is known that antimonial therapy failure and resistance do not only depend on host characteristics (nutrition, immune status, comorbidities, inadequate drug doses and treatment follow-up) but also on parasite factors (strains innate susceptibility, virulence factors, biologic pro le) (12)(13)(14)(15)(27)(28)(29). In this context, Leishmania spp. capacity to alternate between clonal and sexual reproduction increases their diversity, plasticity and biological capacity to expand under different stress conditions (30). Exposure and resistance towards several drugs could be a developed capacity to adapt and survive within hosts.
The IC 50 values obtained in this study showed that L. (V.) braziliensis clinical isolates were better distinguished due to the Sb III IC 50 values over the axenic amastigote forms. Promastigotes showed a 1.9fold difference while axenic amastigotes presented a signi cant 3.1-fold difference between the R and NR group. However, R5 and NR2 promastigotes did not follow this observation, since they had almost the same IC 50 value. Sb V was toxic for promastigotes of each clinical isolate and showed a 2-fold difference between the R and NR group. These results are in concordance with another study using L. (V.) braziliensis promastigotes, isolated from clinical samples of the same endemic area, that showed Sb V IC 50 minimum and maximum values of 0.37 ± 0.09 and 5.75 ± 0.26 mg/mL (13). Similarly, another study showed great variance among Sb V IC 50 values, with a 3-fold difference between promastigotes from poor/bad and cured/good clinical response to antimonial therapy (31). It has been observed that L. (V.) braziliensis isolates, circulating in the state of Rio de Janeiro, share common genetic traits but have different responses to Glucantime® treatment (32). The varied in vitro susceptibility showed in this study reinforces the phenotypical heterogeneity reported for this species.
Even though intracellular amastigote is the "gold standard" model to test susceptibility, axenic amastigotes have been previously used to evaluate response towards Sb V , which was found to be stage and strain-speci c in L. (L.) donovani (33) and L. (L.) infantum (34). Furthermore, it is important to remember that assays with intracellular amastigotes are di cult to standardize and strongly depend on the type of host cell and the medium used (15,35), which might bias the nal results. Therefore, axenic amastigotes are a feasible model to characterize in vitro phenotypes since they maintain similar morphology, metabolic and virulence genes expression pro les as intracellular amastigotes (36)(37)(38)(39)(40) The fact that promastigotes and axenic amastigotes have different in vitro susceptibilities, led us to evaluate the possible correlation of these phenotypes with virulence factors -serine proteases -under a biochemical approach. The serine protease activity measured by Z-FR-AMC, a substrate for serine proteases such as cathepsins, kallikrein and plasmin (41), was relatively the same among promastigotes and axenic amastigotes. Serine enzymatic activity is generally con rmed by using PMSF, AEBSF and TLCK inhibitors (26). Besides, PMSF and AEBSF are known to inhibit a broad range of serine proteases including subtilisins while TLCK has a greater preference for Leishmania spp. oligopeptidases (20,(42)(43)(44). The inhibition pro les seen in this study were different depending on each inhibitor assayed, suggesting that there are different groups or isoforms of serine proteases as part of the protease-network of each clinical isolate, and its respective biological forms.
Furthermore, cluster analysis was performed to investigate if the in vitro phenotypes correlate with the clinical response of each clinical isolate. Cluster 3 contained only one isolate, NR3, which was closely similar to the isolates from homogeneous Cluster 2 (NR1, NR4). Cluster analysis shows the associations between in vitro susceptibility and clinical response. However, additional studies with more isolates from other geographical areas are necessary to explore and reinforce the associations found here.
On the other hand, the presence of heterogeneous clusters was better understood once we compared the pairwise distance between the members of each cluster. This analysis showed a similar pattern of PMSF and AEBSF inhibition over both parasite biological forms while TLCK inhibition was signi cantly more varied among the heterogeneous clusters. This may suggest that independently of each isolate clinical response, they have common enzymatic and in vitro Sb III susceptibility traits. This observation supports the hypothesis that parasites of the subgenus Viannia are a polyclonal population with high genetic variability and, consequently, phenotypic diversity (37,45). Other studies have shown that genetic and phenotypic characteristics among different L. (V.) braziliensis strains are associated with different clinical manifestations and drug resistance (32,46). Additionally, the variation among serine proteases inhibition indicates speci c traits among the conformation of their serine protease network since the speci city of PMSF, AEBSF and TLCK is different (42,47). Therefore, L. (V.) braziliensis isolates can be composed of distinct enzymatic pro les that in uence host-parasite interactions and, consequently, the success or failure of speci c drugs.
Additionally, based on the previous role seen for subtilisins as maturases of TXNPx (18,26), our study examined the expression of subtilisin transcripts from SB clan and S8 family (LbrM.13.0860 and LbrM.28.2570) and TXNPx (LbrM.15.1080). In general, almost all isolates (promastigotes and axenic amastigotes) expressed the three mentioned genes. This nding suggests that subtilisins and TXNPx are expressed by L. (V.) braziliensis clinical isolates, but its level of expression varies depending on the isolate phenotype. Interestingly, a proteomic study using lab-generated L. (V.) braziliensis NR/resistant strains showed that TXNPx abundance was signi cantly higher than R/susceptible strains (48). Additionally, another study using L. (L.) donovani clinical isolates correlated TXNPx higher ampli cation levels with antimony resistance phenotypes (49). However, is important to bear in mind that this study did conventional PCR which is a qualitative method, thus, qPCR and proteomics experiments, which are more sensitive quantitative methods, need to be performed to quantify these transcripts expression and explore their relationship with different in vitro phenotypes.
Leishmania spp. resistance towards antimony treatment is an increasing multifactorial phenomenon, and several virulence factors have been studied and related to resistant phenotypes in clinical isolates (15,(50)(51)(52). The present study adds to this discussion by bringing proteases to focus, one of the most studied parasite virulence factors in the last 30 years (18,21,53). The data presented here show the possibility to use serine proteases to in vitro characterize L. (V.) braziliensis clinical isolates with different responses towards antimony. There is an important heterogeneity among the assayed isolates suggesting that some of them have different innate abilities to adapt to different environments and biological lters. Moreover, serine proteases activity and differential transcripts expression suggested that each isolate may have independent means to adapt to their niches and that in vitro response to antimony needs further characterization, especially in clinical isolates causing ATL. Laboratory (LaPClinVigiLeish) at INI -Fiocruz were responsible for this classi cation following the criteria reported for ATL patient's treatment in the state of Rio de Janeiro -Brazil (4,5,54). All the isolates included in this study were previously characterized as L. (V.) braziliensis by multilocus enzyme electrophoresis (MLEE), according to procedures described elsewhere (55). For this study, each isolate was named after its patient' response: Responder (R1, R2, R3, R4, R5) and Non-responder (NR1, NR2, NR3, NR4, NR5, NR6, NR7).

Parasites culture and in vitro differentiation
Parasites were cultured in biphasic Novy-MacNeal-Nicolle (NNN) medium with 10 % of inactivated FBS. Then, they were expanded in Schneider´s insect medium at pH 7.2 supplemented with 20 % of inactivated FBS, 200 IU penicillin and 200 mg/mL streptomycin and maintained at 26 °C. To obtain each isolate growth curve, 3x10 5 /mL promastigotes were initially cultured and maintained in 25 cm 2 asks containing 5 mL of the medium described above. Daily, for eight days, a 10 µL aliquot was taken to determine the number of viable parasites using a Neubauer chamber (data not shown). Each parasite isolate did not have more than 7 passages since isolation. Differentiation from promastigotes to axenic amastigotes was performed as previously described elsewhere (38), with a few modi cations. Brie y, 5x10 5 per mL of log-phase promastigotes were cultured in Schneider medium (pH 5.5) supplemented with 20 % of FBS, 60 IU penicillin, 60 mg/mL streptomycin and maintained at 26 °C for 2 days. Then, to complete differentiation, each culture was subjected to temperature shock at 32 °C for 2 days. After this period, full differentiation was veri ed under an optical microscope (Labomed, Labo America, Inc.) and these parasites -named as one-day axenic amastigotes -were immediately used for all experimental assays (Supplementary le 8).
In vitro susceptibility assays Log-phase promastigotes and one-day axenic amastigotes forms, of each isolate, were tested against Sb V and Sb III to measure the half-maximal inhibitory concentration (IC 50 ) induced by each drug. The IC 50 was determined by AlamarBlue™ reduction assay as previously described (56), with some modi cations. Brie y, each parasite form was seeded in 96-well plates in triplicate at adequate conditions: log-phase promastigotes (4x10 6 parasites/mL) and one-day axenic amastigotes (5x10 5 parasites/mL) in 0.1 mL of Schneider medium (pH 7.2 for promastigotes or pH 5.5 for axenic amastigotes) supplemented with 20 % of FBS and each drug in decreasing concentrations, leaving one column without any drug to serve as the control. Sb V concentrations ranged from 20 mg/mL to 6x10 -4 mg/mL; Sb III concentrations ranged from 0.196 mg/mL to 5x10 -6 mg/mL, with a 2:1 dilution factor between each one. After parasites incubation (Promastigotes: 26 °C, 48 h; Axenic amastigotes: 32 °C, 24 h), AlamarBlue™ reagent was added to each well (10 µL) followed by a new incubation at their respective temperatures for 4 hours. Then, each plate was read on a Spectramax 190 microplate spectro uorometer (Molecular Devices Corporation) at 570 excitation and 590 nm emission wavelengths, and the percentage of reduction of AlamarBlue™ was determined.

Protein extraction
Log-phase promastigotes (10 8 to 10 9 parasites/mL) and one-day axenic amastigotes (10 8 parasites/mL) of each isolate were separately used to obtain whole protein extracts as it follows. Parasites were washed by centrifugation (3 000 × g, 4 °C, 10 min) in sterile cold PBS pH 7.2. Then, the pellets were re-suspended in 1 mL of lysis buffer (100 mM Tris-HCl pH 8.0, 150 mM NaCl, 10% glycerol and 0.6% Triton X-100) and subjected to a minimum of 5 freeze-thaw cycles. After parasites lysis, con rmed by optical microscopy, the soluble fraction was obtained by centrifugation (25 000 x g, 4 °C, 30 min) and the supernatant stored at -80 °C until further use. The parasites total protein concentrations were determined by the Lowry method using BSA as a standard protein (57).

Enzymatic assays
The serine protease activity of the whole protein extract, 5 µg of total protein, was assessed in activation buffer (Tris-HCl [10 mM], pH 7.5) using a speci c uorogenic peptide substrate, Z-FR-AMC [1 mM], at a nal volume of 60 mL. Samples were incubated (37 °C, 60 min), and the variance in the relative uorescence was monitored on a Molecular Devices SpectraMax spectrophotometer (Gemini XPS).
The substrate cleavage rate was de ned as follows: v = Ds/Dt, where v = velocity, Ds = substrate concentration variation and Dt = total reaction time, as determined elsewhere (36). The self-degradation of the uorescent peptide substrate was controlled throughout the assay to avoid incorrect readings; the enzymatic activity is expressed as mmol min -1 .mg of protein -1 .

Primers design
The primers used in this study were previously designed and used to detect serine proteases and tryparedoxin peroxidase transcripts (26).

RNA extraction and cDNA synthesis
Log-phase promastigotes (10 8 to 10 9 parasites/mL) and one-day axenic amastigotes (10 7 parasites/mL) were separately lysed in 1mL TRIzol containing 200 mL of chloroform. For RNA extraction the samples were centrifuged (10 000 × g, 4 °C, 10 min) and the supernatants containing RNA were dissolved in Isopropanol (12 000 × g, 4 °C, 20 min) and washed with 70 % ethanol (9 000 × g, 4 °C, 5 min promastigotes Sb III IC 50 and (iii) axenic amastigotes, (iv) Z-FR-AMC protease activity substrate over promastigotes and (v) axenic amastigotes, (vi) PMSF inhibition over promastigotes and (vii) axenic amastigotes, (viii) AEBSF inhibition over promastigotes and (ix) axenic amastigotes, (x) TLCK inhibition over promastigotes and (xi) axenic amastigotes were normalized and subjected to principal component analysis (PCA). Here we have 11 variables per isolate and biological form which is why PCA analysis helped us reduce dimensionality and perform further cluster analysis (59). The rst three PCs (PC1, PC2, PC3) were used to cluster the clinical isolates using the K-means algorithm clustering method. To determine the optimal number of clusters we used the total-within cluster sum of squares (twcss), as a function of the number of clusters, where the squared distances between each cluster centroid () and each of its cluster members ( are summed over each cluster , Nc is the total number of clusters, equation (1). Additionally, with the previously normalized data, we explored the characteristics of the clusters by calculating the pairwise distance between clusters using the normalized values of Sb III IC 50 , PMSF, AEBSF and TLCK inhibition over promastigotes and axenic amastigotes. The statistical analysis was carried out using R (version 1.1.463). Figure 1 Promastigotes and axenic amastigotes in vitro susceptibility pro le to antimony. Both parasite forms of each isolate (n=12), 4 × 106 promastigotes/well and 5 × 105 axenic amastigotes/well were exposed to serial dilutions of trivalent (SbIII) and pentavalent (SbV) antimonial for 48 hours and 24 hours in 96-well plates, respectively. The half-maximal inhibitory concentration (IC50 [mg/mL]) of parasites viability was measured using a uorescence method with AlamarBlue reagent. Parasites isolated before treatment of patients with ATL cured after antimonial therapy (R: •) or with poor clinical response to therapy, either therapeutic failure or relapse (NR: ○). The data is presented by boxplot diagrams as the mean of three biological replicates for each isolate. Asterisks indicate statistically signi cant differences: * p < 0.05