BALB/c mice (Mus musculus) and golden hamsters (Mesocricetus auratus) were maintained and bred under pathogen-free conditions. The use of both mice and hamsters was approved by the Institutional Animal Ethics Committees of Indian Institute of Chemical Biology, India. All animal experiments were performed according to the National Regulatory Guidelines issued by CPSEA, IAEC (Committee for the Purpose of Supervision of Experiments on Animals, IICB/AEC-15-2008, 10.06.2008; biosafety clearance number, IITKGP RCGM: BT/BS/17/27/97-PID, 11.02.21).

Mouse peritoneal macrophages (PECs) were harvested from BALB/c mice by lavage 48 h after an intraperitoneal injection of 2% soluble starch (Sigma). For convenience, the term PECs and macrophages are used interchangeably. The macrophages were plated on sterile 22-mm² coverslips (Bluestar) in 35-mm disposable petri plates (Tarsons) at a density of 1 × 10⁵ cells/coverslip in RPMI 1640 medium (Sigma) supplemented with 10% fetal bovine serum (FBS; Sigma) and 100 U PenStrep (Gibco) (RPMI complete medium). The cells were left to adhere for 48 h at 37°C and 5% CO₂ before infection.

Promastigotes of 13 and six previously characterized clinical L. donovani isolates showing resistance or susceptibility to antimonial drugs, respectively, were maintained in M-199 complete medium. The respective years of isolation of the parasites and the treatments received by the patients are mentioned previously (Mukhopadhyay et al., 2011). Furthermore, all the clinical LD isolates used in this study were clonal isolates whose EC50 against SSG was determined and matched with a previously reported EC50 value. The host cells (PECs) were infected with metacyclic promastigotes at a ratio of 1:10 as described previously (Mukhopadhyay et al., 2011). Briefly, all experimental infections were performed with stationary-phase early-passage promastigotes. Similarly, for impedance and thiol measurement, early-passage (passage 2–5) stationary-phase promastigotes were used as described previously (Mukhopadhyay et al., 2011). All promastigotes were checked for their EC₅₀ against SSG, and isolates showing comparable EC₅₀ values with no significant difference from their previously reported susceptibility levels towards SSG (Mukhopadhyay et al., 2011) were further characterized against Mil, AmpB, and Paro.

A 2-mM stock of Amp-B (SIGMA) was prepared in dimethyl sulfoxide and subjected to sterile filtration. Mil (received through Kaladrug-R consortium) and Paro sulfate (Sigma) stock solutions were prepared at 20 mM in water and filter-sterilized. SSG (Albert David, Kolkata) was prepared at 1 mg/ml of 30.30% Sb^V in 1X phosphate-buffered saline (PBS); and dissolution of the powder was being aided by incubation at 37°C overnight, and the solution was then subjected to sterile filtration. All aqueous drug solutions appeared clear and fully dissolved.

The drugs were threefold serially diluted over six concentrations in RPMI 1640 medium supplemented with 10% FBS and 100 U PenStrep in triplicate at each concentration. Stock solutions and dilutions were prepared fresh for each use. The infected macrophages were incubated with drug dilutions for a total of 48 h at 37°C and 5% CO₂. The untreated macrophages received RPMI complete medium, and infection was determined at 48 h post-treatment. Before the experiment, the dose of each drug that was tolerated by the macrophage was determined as described previously (Mukhopadhyay et al., 2011). At the endpoints, the coverslips were washed with PBS, dried and fixed with 100% methanol (Merck), stained with 10% Giemsa (SIGMA), and examined microscopically. The EC₅₀ values of each isolate were estimated (Kremb et al., 2010; Mukhopadhyay et al., 2011) against each drug by the method recommended by the Kaladrug-R consortium. The experiment was repeated thrice.

The dose of each drug that was tolerated by the macrophage was already determined. Briefly, uninfected PEC (0.5 × 10⁶) cells were treated with each of the dilutions of the four drugs for a total of 48 h at 37°C and 5% CO₂. The percentage of cell viability was checked 48 h later by trypan- Blue staining and counting under a light microscope. The cells were washed with PBS and incubated with 0.4% Trypan blue (Thermo Fisher Scientific) solution in PBS for 15 min (Strober, 2015). The concentration ranges for the experimental sets were adjusted to cover EC₅₀ values where possible. In parallel experiments, parasites were incubated with different doses of Mil (μM), and then their susceptibility to the drug was observed by MTT (Sigma) and Alamar Blue (ThermoFisher) assay according to the protocol of the manufacturers.

Two million stationary-phase parasites were used for the study. Each of the samples was washed with sterile PBS, incubated with 5 μM of 5-chloromethylfluorescein-diacetate (Molecular Probes, Carlsbad, CA, USA), incubated for 15 min at 35°C in PBS in the dark as described elsewhere, and subjected to flow cytometric analysis (Sarkar et al., 2009).

The impedance measurement of the parasites suspended in PBS was carried out in an ECIS-8W1E DD (Applied BioPhysics, USA) cell culture well. An Agilent precision impedance analyzer 4294-A, in the frequency range of 40 Hz to 100 MHz, was used to record the impedance data from the ECIS device by applying 10-mV AC excitation voltage. The ECIS-8W1E DD culture well consists of eight separate mini-culture wells having an individual circular working electrode and common counter-electrode made of a thin gold film as described previously (Das et al., 2012). The impedimetric responses were acquired with 400 µl of PBS for measuring baseline and subsequently 400 µl of PBS containing 10⁶ promastigotes to study the activity of LD isolates.

Promastigotes of LD strains BHU782, BHU814, BHU777, and BHU816 were used to infect 4- to 6-week-old golden hamsters. At 45 days post-infection, 17.5 mg/kg Mil was administered orally for 30 days. A human-equivalent dose of Mil for the hamster was used as described elsewhere (Reagan-Shaw et al., 2008; Fortin et al., 2016). At 7 days post-completion of treatment, splenic and liver parasite burden was determined by the stamp–smear method as described previously (Basu et al., 2005).

Each experiment was performed three times independently in duplicates, and the results are expressed as mean ± SD. Statistical significance was calculated by implementing t-test, and a P-value <0.05 was considered statistically significant. Values were considered most significant (represented as ****) if p <0.0001, extremely significant (represented as ***) if p <0.001, very significant (represented as **) if p =0.001–0.01, or significant (represented as *) if p =0.01–0.05; non-significance was represented as “ns”. To establish a correlation between the EC₅₀ with other parameters, the Spearman rank correlation coefficient was used and expressed as r. Data were analyzed using Graphpad Prism software.

The previously characterized clonal population of field isolates, obtained from patients in the SSG-unresponsive zone, was subjected to test for susceptibility to three different antileishmanial drugs at intracellular amastigote stages (Seifert et al., 2010). The EC₅₀ values of different isolates at their intracellular amastigote stages are shown in Table 1 . The EC₅₀ values for AmpB at the amastigote stage ranged from 0.01 ± 0.004 to 0.368 ± 0.199 µM, with a mean EC₅₀ of 0.189 ± 0.10 µM. The field isolates showed various levels of susceptibility to Mil, with EC₅₀ values at the amastigote stage ranging from 0.04 ± 0.026 to 6.13 ± 0.315 µM, with a mean EC₅₀ of 3.085 ± 0.17 µM. For Paro, the EC₅₀ values ranged from 1.01 ± 0.168 to 96.95 ± 16.96 µM, with a mean EC₅₀ of 48.98 ± 8.56. This indicates a very high variation of field resistance among different isolates with respect to all three drugs. The antileishmanial activity of SSG indicates a medium positive correlation (r = 0.407, p < 0.001 and r = 0.431, p < 0.001) with Paro and Mil, respectively, and low positive correlation (r = 0.119, p < 0.001) with AmpB. Among 13 previously characterized SSG-R LD, two (BHU575 and BHU814) showed high EC₅₀ against all three anti-leishmanial drugs, while BHU770, BHU872, and BHU782 showed high EC₅₀ against Mil and Paro. BHU744 showed high EC₅₀ against Mil, while BHU796 and BHU569 showed high EC₅₀ against Paro and AmpB, respectively. Interestingly, some SSG-R LD (BHU568, BHU764, BHU572, BHU573, and BHU574) showed low EC₅₀ to all the three drugs. However, all previously characterized SSG-S LD (BHU816, BHU777, BHU581, BHU741, Ag83, and BPK206) showed low EC₅₀ to all three drugs as presented in Table 1 .

For a comparative evaluation of IC₅₀ and EC₅₀ against Mil, two different dyes (MTT and Alamar Blue) were used. The MTT and Alamar Blue results show that the isolates having high EC50 against Mil in amastigote stage also showed the highest IC₅₀ against MIL in MTT and Alamar Blue assay. Those resistant to SSG alone show moderate IC₅₀ against MIL, while SSG-sensitive isolates show low IC₅₀ to MIL ( Table 2 ).

It has been well established that SSG-R isolates have a higher content of intracellular thiols than the sensitive strains (Rai et al., 2013). However, whether these higher thiol contents can also influence resistance to other antileishmanial drugs is yet to be defined. To evaluate this based on EC₅₀ data, the LD isolates were grouped in seven distinct categories, and intracellular thiol content was determined for each representative group: group 1 representing high EC₅₀ against all drugs; group 2 with moderate EC₅₀ values against SSG, along with high resistance towards Mil/Paro; group 3 representing high EC₅₀ values against SSG and Mil; group 4 with high EC₅₀ values against SSG and Paro; group 5 with high EC₅₀ values against SSG and AmpB; group 6 having high EC₅₀ against only SSG; and group 7 was representative SSG-LD with low EC₅₀ against all other drugs. Our analysis suggests that isolates with high EC₅₀ against SSG tend to have a significantly higher intracellular non-protein thiol as compared to those isolates having low EC₅₀ against SSG and are not affected by EC₅₀ against any of the other three drugs, although it is enhanced in case of high EC₅₀ against all the drugs mentioned ( Figure 1A )—for instance, the level of non-protein thiol is significantly higher in BHU568 which shows high EC₅₀ against only SSG and low EC₅₀ against all other drugs. Thiol content is highest in BHU814 with high EC₅₀ against SSG, Mil, and Paro in contrast to BHU770 with lower thiol content and having moderate EC₅₀ against SSG but high EC₅₀ against both Mil and Paro. Similarly, BHU569, which exhibits a high EC₅₀ against AmpB, exhibits much lower non-protein thiol content owing to its moderate and low EC₅₀ against SSG, respectively. Taken together, our data suggest that isolates with high EC₅₀ against SSG independent of EC₅₀ values of all other drugs tend to have a high non-protein thiol content in contrast to isolates with lower (as in the case of SSG-S isolates BHU816 and BHU777) EC₅₀ against SSG.

Previously, fluorescence anisotropy has been used to characterize between SSG-S and SSG-R isolates, with SSG-R parasites showing higher membrane fluidity (Mukhopadhyay et al., 2011). As intracellular non-protein thiol content can distinguish LD isolates with only high EC₅₀ against SSG but fails to distinguish between other drugs used in this study, we decided to check whether we can use membrane fluidity as a possible parameter to distinguish between LD isolates showing varying degrees of susceptibility towards Mil, Amp B, and Paro. By using impedance-based spectral analysis, we determine resistance and capacitance for members of group 1 (BHU575 and BHU 814), group 2 (BHU770 and BHU 872), group 3 (BHU744), group 4 (BHU796), group 5 (BHU569), group 6 (BHU 568 and BHU 573), and group 7 (BHU816, BHU777, Ag83, and DD8). These parameters are then fitted into a Bode plot using an equivalent circuit model ( Figure 1B ) in the EC-lab software to extract the resistance and capacitance of cell membrane (Das et al., 2014). The equation is given as follows: overall impedance = R _sol || C _sol || (R_i +C _mem), where R _sol and C _sol are solution resistance and solution capacitance respectively. R_i represents cytoplasm resistance, and C _mem represents membrane capacitance. Here changes in R_i and C _mem are mainly considered for analysis, as solution resistance (R _sol) and capacitance (C _sol) do not affect the overall impedance. R _sol and C _sol depend on the conductivity of the media solution where the parasites were suspended. In case of media, as no cells are present in the solution, the overall equivalent circuit becomes R _sol || C _sol. A lower value of R _sol confirms the higher conductivity of media, as it contains more number of ions. R _sol in LD strains increases significantly, as the incubated parasites produce insulating byproducts, and thus drastically reduces the conductivity of the solution. C _sol, in case of media, will not vary significantly, as it is present due to the electrode–electrolyte interface. Our result suggests that SSG-R isolates having a high EC₅₀ value only against Mil, irrespective of their susceptibility towards other drugs, exhibited a significantly higher membrane fluidity owing to its higher membrane capacitance ( Figure 1C ). This is particularly evident for BHU814, BHU 872, BHU744, BHU575, and BHU 770 which have higher EC₅₀ towards Mil and showed a higher value for C _mem (nF), decreasing the overall impedance denoted by |Z| (kOhm), as shown in Figure 1C ; however, cytoplasmic resistance (Ri), despite being the variable, does not change much in different parasites ( Table 3 ).

To determine whether impedance-based spectral analysis can be used to determine low Mil susceptibility among clinical LD isolates, a correlation analysis was performed. It is evident from our analysis that, while EC₅₀ values show a positive correlation with IC₅₀ values against Mil-R LD isolates, a significant correlation of EC₅₀ was also observed with C _mem (r = 0.96716, p < 0.001) and C _sol (r = 0.85427, p < 0.001). Hence, our result suggests that impedimetric parameters can be used to detect low miltefosine susceptibility in LD isolates.

The results of our previous drug sensitivity and bioimpedance analysis implied that there are six LD isolates (BHU575, BHU814, BHU770, BHU872, and BHU782) with low sensitivity to Mil and BHU816, BHU777, BHU581, BHU741, Ag83, and BPK206 show high sensitivity towards Mil. Out of these, two isolates (BHU814 and BHU782) that are less sensitive to Mil and two isolates (BHU777 and BHU816) showing high sensitivity towards Mil were used to infect golden hamsters. After the indicated time of Mil treatment, the infected hamsters and both the Mil-treated and untreated groups were sacrificed, and splenic and liver parasite number was determined. There were no detectable parasites in either the spleen or liver of the hamsters infected with BHU777 or BHU816 after Mil treatment; however, a significant number of residual parasites were observed in the spleen and liver of hamsters infected with BHU782 or BHU814 even after 1 month of treatment with Mil (seven times higher than the human-equivalent dose in hamsters), as shown in Figure 1D (i and ii), suggesting that these isolates do exhibit low susceptibility towards Mil in vivo.