In Vitro Susceptibility to Miltefosine of Leishmania infantum (syn. L. chagasi) Isolates from Different Geographical Areas in Brazil

Treatment of visceral leishmaniasis in Brazil still relies on meglumine antimoniate, with less than ideal efficacy and safety, making new therapeutic tools an urgent need. The oral drug miltefosine was assayed in a phase II clinical trial in Brazil with cure rates lower than previously demonstrated in India. The present study investigated the susceptibility to miltefosine in 73 Brazilian strains of Leishmania infantum from different geographic regions, using intracellular amastigote and promastigote assays. The EC50 for miltefosine of 13 of these strains evaluated in intracellular amastigotes varied between 1.41 and 4.57 μM. The EC50 of the 73 strains determined in promastigotes varied between 5.89 and 23.7 μM. No correlation between in vitro miltefosine susceptibility and the presence of the miltefosine sensitive locus was detected among the tested strains. The relatively low heterogeneity in miltefosine susceptibility observed for the 73 strains tested in this study suggests the absence of decreased susceptibility to miltefosine in Brazilian L. infantum and does not exclude future clinical evaluation of miltefosine for VL treatment in Brazil.


Introduction
Visceral leishmaniasis (VL) remains a worldwide public health concern, mainly affecting poor populations across Asia, East Africa, South America, and the Mediterranean region. Over 600 million people live in areas at risk of developing the disease [1]. Overall annual incidence is estimated to be from 50,000 to 90,000 cases, with ten countries concentrating more than 95% of the reported cases: Bangladesh, Brazil, China, Ethiopia, India, Kenya, Nepal, Somalia, South Sudan, and Sudan [1]. In Latin America, VL is a zoonosis caused by Leishmania infantum (syn. L. chagasi) with dog as the main reservoir of infection and Lutzomyia longipalpis the principal vector. It is endemic in 13 countries with an average of 3470 new reported cases every year. In 2019, Brazil reported 2529 new cases, accounting for 97% of cases in Latin America, with around 35% of those in children under 10 years of age [2], and a lethality of 9% [2,3], one of the highest in the world. Another concern is the increasing incidence of HIV/VL coinfection, rising from 0.7% of reported VL cases in Brazil in 2001 to 11.1% in 2019 [2,4].
First line treatment for VL in Brazil is meglumine antimoniate (MA) at a dose of 20 mg intravenous Sb5 + /kg/day for 20 days. In 2013, the Brazilian Ministry of Health

Leishmania infantum (syn L. chagasi) Strains
Seventy-three strains isolated between 2002 and 2015 from humans (n = 60) and dogs (n = 13) were included in this study. In addition, one L. infantum (NLC/IOCL3241) and one L. donovani strain were used as internal controls. All L. infantum strains employed in this study are deposited in the Leishmania Collection at the Oswaldo Cruz Foundation (CLIOC), and were shipped to the ICB-USP Leishmaniasis Laboratory. Isolates were chosen to represent different endemic areas of Brazil (Figure 1), and included six isolates from patients who were enrolled in the clinical trial to assess the efficacy and safety of miltefosine in Teresina [15]. Information provided by the institutions that conducted parasite sampling included the absence of previous treatment before sampling. The activities performed with the Leishmania strains were duly registered in SisGen (Sistema Nacional de Gestão do Patrimonio Genético) under the identifier AE5FEC8, as determined by article 20 of Decree No. 8.772, in accordance with the Brazilian Biodiversity Law (Law No. 13.123/2015).  [22]. Numbers inside each Brazilian state represent the number of strains that originated there. The reference strain NLC is represented in red and underlined. (*) Strains were allocated according to the geographic region in which the parasite was isolated, when the state in which the infection occurred was not confirmed. Strains used in the initial amastigote assays are highlighted in red.

Parasite Culture
L. infantum promastigotes were grown in sterile M199 medium (Sigma-Aldrich, St. Louis, MO, USA), prepared by diluting the powder in distilled water. Medium was supplemented with 10% heat-inactivated fetal calf serum, 0.25% hemin, 100 μg/mL penicil-  [22]. Numbers inside each Brazilian state represent the number of strains that originated there. The reference strain NLC is represented in red and underlined. (*) Strains were allocated according to the geographic region in which the parasite was isolated, when the state in which the infection occurred was not confirmed. Strains used in the initial amastigote assays are highlighted in red.

Parasite Culture
L. infantum promastigotes were grown in sterile M199 medium (Sigma-Aldrich, St. Louis, MO, USA), prepared by diluting the powder in distilled water. Medium was supplemented with 10% heat-inactivated fetal calf serum, 0.25% hemin, 100 µg/mL penicillin/streptomycin, and 2% sterile human urine. Cultures were maintained in an incubator at 25 • C and weekly passages were performed.

Intracellular Amastigote Assay
To determine drug activity in intracellular L. infantum amastigotes, bone marrowderived macrophages (BMDM) from BALB/c mice were obtained using the method previously described [23]. Briefly, 3 × 10 5 BMDM were left to adhere to round glass coverslips in 24-well culture plates. After incubation in a 5% CO 2 atmosphere for 24 h at 37 • C, wells were washed with PBS to remove non-adhered cells. Macrophages were then exposed to stationary-phase promastigotes of L. infantum at a multiplicity of infection (MOI) of 15 parasites per macrophage for 4 h at 37 • C in a 5% CO 2 atmosphere.
Non-internalized parasites were removed by successive washings with PBS, and infected macrophages were incubated for 72 h at 37 • C at 5% CO 2 in RPMI medium with increasing miltefosine concentrations (0.5 to 40 µM). Coverslips were washed with PBS, stained with the Instant Prov kit (Newprov, Pinhais, PR, Brazil) and examined by optical microscopy (Nikon Eclipse E200; Nikon Corporation, Tokyo, Japan). The percentage of infected macrophages, number of amastigotes per infected macrophage, and infectivity index (percentage of infected macrophages multiplied by the average number of amastigotes per macrophage) were determined by counting 100 cells in each triplicate slide. The EC 50 was determined using the values of the infectivity index through the nonlinear, dose response, sigmoidal model using the GraphPad Prism 6 program. At least two independent experiments were performed in triplicate for each miltefosine concentration and L. infantum NCL/IOCL3241 and L. donovani Ld15 strains were used as controls in some experiments.

Promastigote Susceptibility Assay
Drug activity was determined by incubating promastigotes of the L. infantum strains in the presence of increasing miltefosine (2.5 to 200 µM) and amphotericin B (6.25 to 400 nM) concentrations. After 24 h incubation, viability was assessed by MTT (3-[4,5-dimethyl-2thiazolyl]-2,5-diphenyl-2H-tetrazolium bromide) assay as previously described [23]. EC 50 values were obtained by plotting optical density values measured by absorbance at 595 and 690 nm in nonlinear, dose response, sigmoidal models using the GraphPad Prism 6 program. At least three independent experiments were performed in triplicate and L. infantum NCL/IOCL3241 and L. donovani Ld15 strains were used as controls.

PCR Detection of the Presence/Absence of Miltefosine Sensitive Locus (MSL)
Leishmania strains were characterized as MSL-or MSL+ either by whole genome deep sequence analyses or by qPCR. Both methodologies and results are available at Schwabl et al. [24]. Briefly, to obtain DNA, Leishmania strains were grown in biphasic NNN + Schneider's medium until the end of the log phase, and genomic DNA was extracted using a commercial kit following the manufacturer's instructions (Qiagen, Valencia, CA, USA). qPCR amplification of the MSL on chromosome 31 or whole genome sequencing was accomplished according to the strategies published [24].

Statistical Analysis
All statistical analyses were performed using Graph Pad Prism 6.0. Correlation between infectivity and EC 50 of intracellular amastigotes was tested using Spearman's correlation test. For the analysis of correlation between the presence/absence of MSL locus and EC 50 to miltefosine, Mann Whitney test was used. Analysis of possible correlations between susceptibility and year of isolation were done using two-tailed unpaired t-test with Welch's. Statistical significance was considered in all these cases for p-value under 0.05 (p < 0.05).

Results
The susceptibility to miltefosine was determined in 13 strains isolated before treatment from 10 humans and 3 dogs, from 5 different Brazilian states and in a control strain (NLC), that was included in these experiments as a reference. The EC 50 determined in intracellular amastigotes varied from 1.41 to 4.57 µM between the strains (Figure 2A and Table 1). The EC 50 obtained for the L. infantum reference strain NLC to miltefosine was 1.79 ± 0.42 µM ( Table 1). Infectivity was also analyzed (Table 1), and correlation tests between EC 50 of intracellular amastigotes and the percentage of infection revealed that these two variables were not correlated (Spearman's correlation test: R = −0.2992 and p = 0.2965).
In previous studies that evaluated the susceptibility of Leishmania strains to miltefosine, we and others observed that, for this drug, there was a direct correlation between susceptibility in amastigotes and promastigotes [21,25]. As promastigote assays are more amenable for larger samples, sensitivity to miltefosine was determined for promastigotes in a total of 73 strains of L. infantum, including the 13 strains tested in intracellular amastigotes.     Sensitivity to miltefosine was determined for promastigotes in a total of 73 strains of L. infantum, sampled before treatment from 60 humans and 13 dogs. A reference strain was again used as an internal control for all experiments (strain NLC). The L. donovani strain LD-15 was also employed as an internal control ( Figure 2B, Table 1). The EC 50 for miltefosine in promastigotes varied from 5.89 to 23.7 µM ( Figure 2B, Table 1). The EC 50 for miltefosine in the L. infantum reference strain NLC and the L. donovani strain LD-15 were 10.98 ± 1.95 and 9.00 ± 0.80 µM respectively. Calculation of the activity indexes (ratio between the EC 50 of the clinical strains and EC 50 of the reference strain) revealed that the highest EC 50 observed among the clinical strains was 2.16-fold higher than the EC 50 of the NLC reference strain.
Sensitivity to amphotericin B was also determined in these strains, and a 2.8-fold variation between the lowest (25 nM) and the highest (70 nM) EC 50 values was observed ( Figure 2C, Table 1). The EC 50 of the L. infantum reference strain NLC and of the L. donovani LD-15 strain to amphotericin were 51.00 ± 5.00 and 35.00 ± 0.50 nM, respectively.
Forty-nine strains (out of 73; 67%) were genotyped for the presence or absence of MSL, recently described and associated with the treatment outcome in Brazilian patients and L. infantum in vitro susceptibility to miltefosine [15,16]. Twenty strains (2 from dogs and 18 from humans) were positive (presented) for MSL and 29 (7 from dogs and 22 from humans) were deleted for the locus.  Statistical analysis was also conducted using only the human L. infantum strains from Piauí, Maranhão and Minas Gerais (amastigotes: 2MSL-; 5MSL+; promastigotes: 9MSL-, 18MSL+), aiming to mimic the parasite diversity analyzed previously [16], but even in this restricted population, no statistical differences were observed in the median EC50 of MSLand MSL+ parasites (p = 0.3810 and p = 0.4948, for amastigotes and promastigotes, respectively) ( Figure S1).
Analysis of possible correlations between susceptibility and year of isolation detected no significant trends. The same was true for samples isolated from humans and dogs (Figure 4). The analysis of correlation between EC50 and geographical origin was not reliable due to low representation for some states (Table 1). Statistical analysis was also conducted using only the human L. infantum strains from Piauí, Maranhão and Minas Gerais (amastigotes: 2MSL-; 5MSL+; promastigotes: 9MSL-, 18MSL+), aiming to mimic the parasite diversity analyzed previously [16], but even in this restricted population, no statistical differences were observed in the median EC 50 of MSL-and MSL+ parasites (p = 0.3810 and p = 0.4948, for amastigotes and promastigotes, respectively) ( Figure S1).
Analysis of possible correlations between susceptibility and year of isolation detected no significant trends. The same was true for samples isolated from humans and dogs ( Figure 4). The analysis of correlation between EC 50 and geographical origin was not reliable due to low representation for some states (Table 1). restricted population, no statistical differences were observed in the median EC50 of MSLand MSL+ parasites (p = 0.3810 and p = 0.4948, for amastigotes and promastigotes, respectively) ( Figure S1).
Analysis of possible correlations between susceptibility and year of isolation detected no significant trends. The same was true for samples isolated from humans and dogs (Figure 4). The analysis of correlation between EC50 and geographical origin was not reliable due to low representation for some states (Table 1).

Discussion
Inadequate efficacy and serious side effects are major and widespread problems of VL treatment. The range of drugs for VL is narrow and fraught with limitations [26]. Miltefosine, though displaying some less than ideal characteristics such as potential teratogenicity, gastrointestinal side effects, and a long plasma half-life of 150-200 h [27], is and will remain for the next few years the only oral drug available for leishmaniasis treatment. Miltefosine also presents a series of advantages compared to the other available treatment options, and care must be taken to avoid the loss of this oral drug. Employed as monotherapy in the Indian subcontinent for several years and in combination therapy with paromomycin as the second line treatment in the framework of the Kala-azar Elimination program [28], miltefosine is not currently in use in Brazil for human visceral leishmaniasis [29,30].
Only one clinical trial to test miltefosine monotherapy for treatment of VL has been conducted in Brazil, with patients enrolled at Montes Claros and Teresina. An overall cure rate of 59% was documented in a sample of 42 treated patients. This study was carried out in two different locations and using two different treatment schedules. In all cases, adults were treated with a 100 mg miltefosine/day total dose, which did not necessarily correspond to the desired dose of 2.5 mg/kg/day [15]. The efficacy in this trial was lower than the 94% cure rate observed in India when miltefosine was implemented [31] and the 90% cure rate observed after a decade of use [29].
Differences in efficacy of the same drug or treatment schedule found in different geographical areas have been demonstrated, but there is no clear understanding of the reasons for this, even given the differences between L. donovani and L. infantum parasite species [28]. This information, and the results obtained in the Brazilian miltefosine trial, raised concerns about the possibility of Brazilian L. infantum being intrinsically tolerant to miltefosine. In this context, we gathered efforts to investigate the susceptibility to miltefosine of a panel of clinical strains of L. infantum from different geographic regions of Brazil.
The susceptibility to miltefosine of 13 clinical strains was determined for intracellular amastigotes. The EC 50 varied from 1.41 to 4.57 µM, or 3.2-fold, between the least and the most susceptible isolates. The EC 50 values in these isolates were strikingly homogenous compared to, for example, Brazilian L. (V.) braziliensis clinical isolates, which presented 6and 15-fold differences in the EC 50 between the most and the least susceptible parasites, in promastigotes and amastigotes, respectively [21]. This low variability was also reported by Faral-Tello, et al. [32] that evaluated the infectivity and susceptibility of 5 L. infantum clinical strains to leishmanicidal drugs including miltefosine and found EC 50 values that varied from 3.4 to 5.3 µM.
The characterization of susceptibility to miltefosine in a total of 73 strains was done in cultured promastigotes. The promastigote model is less laborious and eliminates the need of animals for macrophage isolation. The susceptibility to amphotericin B was also determined to widen the scope of the study and also to serve as an independent variable in case a heterogeneous behavior toward miltefosine was discovered. The analysis in promastigotes confirmed a low degree of variation (5.89 to 23.7 µM), when compared to the differences found among L. braziliensis clinical strains (22.9 to 144.2 µM) [21].
The evaluation of a small percentage of strains in the amastigote stage is a limitation of this study. A correlation between miltefosine susceptibility in promastigotes and amastigotes of L. braziliensis [21] and L. donovani [25] had already been demonstrated. This correlation was not present among the 13 strains in this study and we hypothesize this could be due to the low variability in the EC50 values determined. However, it would be interesting to widen the spectrum of strains tested as amastigotes in the future. Another aspect that remains to be addressed in future studies is to improve the distribution of strains between the different geographical areas and continuing the evaluation with samples collected more recently. In our sample, some states and samples isolated after 2013 were underrepresented.
The analysis in promastigotes, confirmed a low degree of variation (5.89 to 23.7 µM), when compared to the differences found among L. braziliensis clinical strains (22.9 to 144.2 µM) [21].
Recently Carnielli, et al. [16] reported that a deletion of four genes on chromosome 31 found by whole genome sequencing of 26 L. infantum strains isolated from cured and relapsed patients from two Brazilian endemic regions was correlated to the treatment outcome. An association between the presence of this locus and a positive response to miltefosine treatment was found, and the genomic region was named miltefosine sensitive locus (MSL) [16]. The susceptibility profile of intracellular amastigotes from strains isolated from patients that relapsed or were cured in the Montes Claros/Teresina clinical trial [15] identified a mean EC 50 in strains from cured patients of approximately 5 µM whereas for relapsed patients this value was approximately 13 µM. The threshold for predicting treatment outcome was calculated as 8 µM.
In our sample, all the strains presented miltefosine EC 50 against amastigotes below 5 µM. Due to the high prevalence of MSL deletion in L. infantum strains in the Americas [24] more than 60% of the strains studied were genotyped for the presence or absence of the MSL following the previously published protocol [24]. The prevalence of MSL deletion detected in the present study (59%) was in agreement with that found previously (54%) [16]. However, no associations were found between L. infantum in vitro susceptibility to miltefosine and the presence/absence of the MSL after testing 49 strains from 11 localities. This lack of association was still observed when the L. infantum population analyzed was reduced to geographically mimic the sample previously studied (only from Minas Gerais, Piauí and Maranhão).
Even with the relative homogeneity in the determined miltefosine EC 50 found in our sample, an important limitation of our study is that the evaluation of in vitro susceptibility does not immediately correlate with the clinical efficacy. Nevertheless, infections of BALB/c mice with the reference strain NLC were successfully treated with 20 mg/kg/day miltefosine for 5 days. When followed up for 4 weeks, these mice showed complete resolution of clinical disease and absence of parasites (M. Galuppo and S. Uliana, personal communication) which supports interpreting the results of the present study as indicating the overall susceptibility to miltefosine of the 73 isolates.
The lack of concordance between data from this and the previous study might reflect differences in methodology, the number of in vitro passages of parasites used, or differences in the populations studied, meaning that there is a need to extend the analysis to a wider sample of strains. However, it is worth considering whether the high failure rates observed in the previous Brazilian clinical trial might have been due to factors other than parasite susceptibility. In the case of L. donovani VL, it has been shown that the exposure to miltefosine is lower in children than in adults; indeed, by using an allometric dose in children, efficacy increases from 59% to 90% [19]. Further studies performed to understand the pharmacokinetics and pharmacodynamics of miltefosine in different populations found that the probability of VL relapse was strongly correlated with the time during which the plasmatic concentrations of miltefosine were higher than the EC 90 values or above 10 × EC 50 [20]. In African patients, the 2.5 mg/kg/day for 28 days schedule resulted in concentrations above the desired threshold for 51.4 days (CI 30.5-77.1). To analyze the likelihood of failure, the authors employed miltefosine EC 50 and EC 90 for L. donovani intracellular amastigotes of 4.95 and 25.9 µM, respectively [17]. These values are comparable to the median EC 50 values for L. infantum clinical isolates found in the present study. If pharmacokinetic and pharmacodynamics characteristics in Brazilian patients are comparable to those previously characterized in Nepal and Africa, it is likely that the necessary plasmatic concentrations will be achieved upon use of allometric dosing.
In conclusion, this study indicates that isolates of L. infantum obtained over a wide geographical area in Brazil do not appear to be significantly heterogeneous in their susceptibility to miltefosine. Further clinical evaluation of miltefosine for treatment of VL in Brazil is advisable, using an allometric dose in children and associated with pharmacokinetic studies, possibly in combination with other leishmanicidal drugs, in order to evaluate the possibility of incorporating this drug into the therapeutic arsenal for VL treatment in Brazil.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/microorganisms9061228/s1. Figure S1: Susceptibility to miltefosine in L. infantum strains presenting (MSL+) or not (MSL-) the MSL locus. Informed Consent Statement: Patient consent was waived by the Institutional Review Boards because parasites used in this work were available at the IOC Culture Collection. Part of the sample had been obtained from patients participating in research projects approved by an Ethical Committee and part was from parasites isolated as part of routine diagnostic procedures in humans or dogs. These latter samples were not attached to patient information beyond clinical form of the disease and geographical origin.

Data Availability Statement:
The data presented in this study are available in the article and supplementary material.