Biogenic nanoporous silicon carrier improves the efficacy of buparvaquone against resistant visceral leishmaniasis

Visceral leishmaniasis is a vector-borne protozoan infection that is fatal if untreated. There is no vaccination against the disease, and the current chemotherapeutic agents are ineffective due to increased resistance and severe side effects. Buparvaquone is a potential drug against the leishmaniases, but it is highly hydrophobic resulting in poor bioavailability and low therapeutic efficacy. Herein, we loaded the drug into silicon nanoparticles produced from barley husk, which is an agricultural residue and widely available. The buparvaquone-loaded nanoparticles were several times more selective to kill the intracellular parasites being non-toxic to macrophages compared to the pure buparvaquone and other conventionally used anti-leishmanial agents. Furthermore, the in vivo results revealed that the intraperitoneally injected buparvaquone-loaded nanoparticles suppressed the parasite burden close to 100%. By contrast, pure buparvaquone suppressed the burden only by 50% with corresponding doses. As the conclusion, the biogenic silicon nanoparticles are promising carriers to significantly improve the therapeutic efficacy and selectivity of buparvaquone against resistant visceral leishmaniasis opening a new avenue for low-cost treatment against this neglected tropical disease threatening especially the poor people in developing nations.

Introduction AD VP sample injector and an FCV-10 AL UV/vis detector. The HPLC analysis was carried at room temperature using Inertsil ODS-3 C 18 , 4×150 mm (GL Sciences Inc.) analytical column. The mobile solution was 50 mM NaOAc and 85% MeOH at pH 3.4. The samples were diluted with a mobile solution before running. The flow rate was 1 mL/min, injection volume was 0.1 mL, and the analysis time was 20 min per sample. The samples were monitored at 262 nm. The BPQ concentrations were determined by measuring the peak areas compared to the calibration curves of BPQ standards.

Synthesis of the nanoparticles (NP)
Barley husk ash (32 g) was washed in 375 mL aqueous solution of 10 wt.% HCl in 100˚C water bath for 2 h to remove inorganic impurities. The washed powder was then filtered out and rinsed with copious amounts of water and dried at 120˚C for 2 h. The sample was then calcined at 600˚C in air for 3 h to remove organic impurities. The purification procedure was repeated once and altogether 24 g of purified porous SiO 2 was obtained.
Nanostructured PSi was produced from the purified SiO 2 powder using magnesiothermic reduction [15]. Briefly, 4 g of silica powder was mixed with 3.24 g of Mg and 4.0 g of NaCl, and milled in a planetary ball mill (Fritsch Pulverisette 7 Premium line) using 1 mm zirconia balls at 400 rpm for 4 min. The milled powder was loaded into a quartz tube and placed into a tube oven at 300˚C in a nitrogen (N 2 ) atmosphere. The sample was then slowly heated up to 600˚C with a heating rate of 5˚C/min and kept at 600˚C for 4 h. The magnesiothermic reduction procedure was repeated four times and the products were pooled together. After reduction, the side products (MgO, Mg 2 Si, Mg) were dissolved first in 400 mL of 37% HCl at 70˚C for 2 h followed by washing with water. Remaining SiO 2 was removed by washing in 1:1 HF:EtOH solution in an ice bath for 5 min. The PSi powder was then filtered out and dried at 65˚C for 1 h yielding 2.8 g of PSi particles.
Nanoparticles were prepared by further milling the PSi powder in 18 ml of EtOH at 800 rpm for 140 min at 5 min intervals. After milling, the nanoparticles were washed with the 1:1 HF:EtOH solution to remove SiO 2 produced during the milling. The particles were separated from the solution by centrifugation and dried at 65˚C for 1 h.
Thermal hydrocarbonization of PSi (THCPSi) was performed through reaction with acetylene at elevated temperature. First, the PSi powder (0.95 g) was flushed with 1 L/min N 2 gas for 30 min using a quartz tube. Acetylene gas (1 L/min) was introduced into the tube while maintaining the N 2 flow. After 15 min, the tube was placed in a tube oven preheated to 500˚C for 14 min 30 s. The acetylene flow was stopped and after 30 s in N 2 flow at 500˚C, the tube was removed from the tube oven and let to cool down to room temperature in the N 2 flow.
Carboxylic acid (-COOH) functional groups were grafted on the THCPSi surfaces by following the method described previously [14,16]. Briefly, THCPSi was mixed with 30 mL of undecylenic acid into the quartz tube. The sample was placed in an oven and incubated at 120˚C for 16 h while maintaining the nitrogen flow. The undecylenic acid functionalized THCPSi was washed with chloroform on a filter and dried at 65˚C for 1 h. The particles were then redispersed by sonicating in 1% Tween solution for 5.5 h, and separated from the suspension by centrifugation at 7000 g for 10 min. The nanoparticles were then washed with EtOH to remove Tween, and finally, the medium was changed to deionized water. The synthesized UnTHCPSi nanoparticles are denoted as NP.
BPQ loading/release study and freeze-drying of the nanoparticles NP (20 mg) was washed with 20 mL EtOH using centrifugation (20 min at 13200 rpm, Eppendorf Centrifuge 5424, Eppendorf AG). After the washing, NP was mixed with 20 mg BPQ suspended in 20 mL EtOH and incubated for 2 h at room temperature. NP was centrifuged, and the BPQ concentrations in the supernatant were measured with HPLC before and after mixing. The loading degree (% w/w) of BPQ in the nanoparticles (BPQ+NP) was calculated using following equation where m d is the mass of loaded BPQ and m NP represents the mass of UnTHCPSi nanoparticles. The m d was calculated from the BPQ concentrations in the supernatant, before and after the loading. For the long-term storage, the nanoparticles were freeze-dried using mannitol, which is a typical excipient among pharmaceuticals [17]. The aliquots of nanoparticles (1 mg/mL) in aqueous solution of 4 wt.% mannitol, was sonicated for 5 minutes, kept in a freezer (-40˚C) for 45 min and then freeze-dried for 48 h through lyophilization (Edwards Lyophilizator). The freeze-dried cakes of the nanoparticles were stored at 4˚C.
The freeze-dried nanoparticles were used to test the release kinetics of the BPQ. The experiment was done in EtOH, Milli-Q water, and cell culture medium (DMEM). For each release medium, 0.5 mg of the BPQ+NP was redispersed in 1 mL of the respective medium and incubated at room temperature. At pre-defined time points (0 h, 0.5 h, 1 h, 2 h, 4 h, and 24 h), NP was centrifuged, and 500 μL of release medium was taken. Fresh 500 μL of the corresponding medium was added to the NP and incubated further until the next time point. The EtOH and water samples were evaporated and dissolved with MeOH. The DMEM samples were filtrated with Amicon Ultrafree-Cl filters cut off 10 000 (Millipore), evaporated, and then dissolved with MeOH. The released BPQ concentrations were measured with HPLC as described above.

Characterization of the nanoparticles
The crystallite size of NP was determined with an X-ray diffractometer (XRD, Bruker D8 DISCOVER) equipped with a Cu-anode X-ray tube (λ = 1.54 Å). Freshly synthesized NP was dried at 65˚C overnight, and the measurements were done on a 2θ range of 20˚-80å nd a step size of 0.031˚. The measurement time was 9 min 32 s. The X-rays were limited on the primary side with a motorized slit set to illuminate 2 cm length on the sample, and a fixed 3˚anti-scatter slit on the secondary side. Axial Soller slits at 2.5˚were used on both the primary and the secondary sides. The K β radiation was removed with a 0.02 mm Ni-filter. The signal was collected with an LYNXEYE 1D detector with the detector stripes fully open. A corundum standard was measured with identical setup to determine the instrumental effects on the peak profile. The lattice parameters for crystalline silicon were obtained from the PDF-2015 database and verification was performed with DIFFRAC.EVA 3.1 software. TOPAS 4.2 software was used to perform a full profile Pawley fit to the measured data, where the instrumental effects determined with the corundum sample were utilized. From the fitted curve, the crystallite size was calculated from the Scherrer equation using integral breadth as the peak width, which eliminates the requirement for Scherrer constant approximation [18].
The porous structure of the NP was investigated by N 2 physisorption using Micrometrics TriStar 3000 instrument. The sample was degassed (at 65˚C in vacuum) and filled with controlled increments of N 2 at a temperature of 77 K. The N 2 adsorption/desorption isotherms were used to calculate the surface area using BET-theory [19], whereas single point pore volume, and pore size distribution were calculated based on the BJH-theory [20].
The surface chemistry of NP was studied with a Fourier transform infrared spectroscope (FTIR, Thermo Scientific Nicolet 8700), and the sample was prepared as tablets by mixing 250 mg KBr with 0.2 mg of the NP.
The particle size, polydispersity, and zeta potential of NP and BPQ+NP were measured at 25˚C by dynamic light scattering (DLS, Zeta-sizer Nano ZS, Malvern Instruments, UK). Samples were washed thrice with Milli-Q water and resuspended in 2 mL water, and the measurements were done both before and after the freeze-drying.

Nanoparticle internalization into macrophages
Mouse RAW 264.7 macrophages were used to study the cellular uptake of NP. The macrophages were grown in DMEM supplemented with 10% FBS, 1% penicillin/streptomycin, and 1% L-glutamine. The cells were maintained at 37˚C, 95% air, and 5% CO 2 atmosphere by changing the culture medium every second day and the cells were divided after reaching 80-90% confluency.
The intracellular distribution of NP in the macrophages was visualized by using transmission electron microscope (TEM, Jeol JEM 2100F). The cells were plated in a 12-well plate (Corning) (50 000/cm 2 ), and the NP (10 μg/mL) were incubated for 1, 4, and 24 h. Cells were washed twice with Dulbecco's Phosphate Buffered Saline (DPBS, Bio Whittaker) and fixed with 2.5% glutaraldehyde buffer solution (0.1 M phosphate buffer at pH 7.4) for 2 h and in 1% osmium buffer solution (0.1 M phosphate buffer at pH 7.4) for 1 h. The cells were washed between and after the fixations two times with the same buffer. Subsequently, the cells were dehydrated using EtOH gradient and embedded with EPON epoxy resin (LX-112). The EPON was polymerized at 60˚C for 2-3 days, and approximately 70 nm thick slices were cut from the casts afterward. The sample slices were stained with 1% uranyl acetate for 30 min and 1% lead citrate for 2 min. The sample slices were placed on standard 200 mesh copper grids and imaged using TEM with an acceleration voltage of 200 kV.

Generation of drug-resistant L. donovani amastigote strains
SSG and PMM resistant L. donovani amastigotes were generated by using the methods described previously [24]. Briefly, the wild-type AG83 promastigotes were cultured in medium 199 (with supplements), in the presence of drug concentration corresponding to the 50% inhibitory concentration of the drugs (IC 50 values) against the parasite, which was 3.6 μg/mL and 10 μM for SSG and PMM, respectively. The cultures were stabilized for three subcultures. Then, the drug concentration was increased until the cell population decreased approximately to 20% for each batch. Finally, when 90% cell population of the initial count was reduced, the phenotype generated was plated on medium 199 agar plates containing the same drug concentration, and a single colony was picked for culture in medium 199 liquid media at the same drug concentration [25,26]. The cultured promastigotes are converted into amastigotes as described above. Drug pressure was withdrawn, and stability was checked at one, two, and four months. SSG and PMM resistant phenotype clones were labeled as AG83-R,SSG and AG83-R, PMM, respectively.

Resident peritoneal macrophage isolation
Resident peritoneal macrophages were isolated from BALB/c mice. For the isolation, 1.5 mL of sterile thioglycolate medium was injected into the peritoneal cavity to induce non-infectious inflammation. Animals were sacrificed after 5 days, and peritoneal exudates were harvested by intraperitoneal lavage. Cold RPMI (6 mL) was injected into the peritoneal cavity. After 10 min, the injected fluid was withdrawn by the same syringe. From each animal, approximately 3-4 mL of medium was recovered. After centrifugation at 200 g for 15 min, macrophages were collected. Hemocytometer was used to count the total number of isolated macrophages, and it was found that approximately 84% of total isolated cells were macrophages, as calculated by using the method previously reported [27].

Monitoring the viability of amastigotes and macrophages
Before starting the drug sensitivity assays, the viabilities of macrophages, amastigotes and promastigotes were checked with the Trypan blue exclusion method in order to verify that the cells and the parasites were good for the experiments [28]. The number of the viable cells was verified by their intact cell membranes that excluded the dye (Trypan blue), whereas dead cells did not. The experimental protocol was such that the cell suspension of 100 μL was mixed with 100 μL of 0.4% trypan blue in phosphate buffer saline at pH 7.2 for macrophages and promastigotes, and pH 5.5 for amastigotes. The mixture was incubated for 2 min at room temperature. The sample was loaded into a hemocytometer and examined immediately under a light microscope. The number of blue staining cells were counted. The viability was more than 95% for all types of cells.

Evaluation of antileishmanial activities against axenic amastigotes
The antileishmanial activity of the drug specimens (BPQ+NP, BPQ, SSG, AmB, PMM, and MF) against the axenic amastigotes were evaluated based on the IC 50 value, which is the concentration of the specimen required to kill 50% of the parasites. The parasites were first seeded into 96-well plates at a density of 1x10 5 amastigotes/well in 200 μL medium. Then, 10 μL of the drug candidates at various concentrations were added and the final volume was adjusted to 300 μL medium. The number of amastigotes in the wells treated with each specimen were compared to those in the untreated control wells.

Cytotoxicity studies
The cytotoxicity of the drug specimen was evaluated by the CC 50 value, which is the concentration required to kill 50% of the host cell population. Macrophages were first isolated from peritoneal lavage as described above. The cells were seeded into 96-well plates at 1x10 6 cells/mL in 200 μL of RPMI-1640 supplemented with 10% FCS, 20 mM L-glutamine, 16 mM NaHCO 3 , penicillin (50 U/mL) and streptomycin (50 μg/ mL) at 37˚C in 5% CO 2 for 4 h for adherence. To remove the non-adherent cells, each plate was washed two times with phosphate buffer saline (pH 7.2). Plates were treated with 2-fold serial dilutions of each drug specimen. After 48 h of exposure at 37˚C in 5% CO 2, the viable macrophages were counted microscopically using a hemocytometer.

Evaluation of antileishmanial activity against intracellular amastigotes
The intracellular antileishmanial activity was evaluated based on the reduction in the amastigote numbers in the infected macrophages after the treatment. The method was adapted from our previous work [24]. Briefly, 4x10 6 amastigotes in 500 μL of RPMI-1640 suspension were added to each well containing macrophages with the macrophage: parasite ratio of 1:10. The plates were incubated for 4 h at 37˚C in 5% CO 2 . After the parasite internalization period, cells were washed at least twice with fresh media to remove non-phagocytosed parasites. Then, the drug specimen suspended in 1 mL RPMI-1640 was added, and the samples were incubated for 72 h at 37˚C in 5% CO 2 . The medium was aspirated, the coverslips were removed, and then fixed with methanol and air-dried. After staining with Giemsa, 100 cells on the glass discs were counted. Three independent experiments in quadruplicate for each concentration were performed.

Determination of IC 50 and CC 50 values
Microscopic examinations were performed in duplicates with various concentrations of pure drugs or buparvaquone-loaded nanoparticles and also for control wells. The amastigotes cells were counted for all cases using a hemocytometer. The mean of the control wells was considered as 100% survival. The counts in the treated wells at various concentrations were converted into percentages. Giemsa-stained slides of the macrophages infected with amastigotes were observed under a light microscope. Then the numbers of amastigotes and infected macrophages per 100 cells were counted for each slide. The percentage (%) of inhibition was calculated by using the equation where C is the number of amastigotes/infected macrophages per 100 cells in the control well and T is the number of amastigotes/infected macrophages per 100 cells in the drug-treated well. Origin version 6.1 software was used to calculate IC 50 and CC 50 values by plotting a graph of the percentage of inhibition at various concentrations of all the observations.

Resistance index and selectivity index
Resistance index (RI) is calculated using the ratio of the IC 50 against drug resistant strain to the IC 50 against wild-type strain (AG83). The lower RI value indicates higher efficiency against resistant strains. The selectivity index (SI) is the ratio between the cytotoxicity and the antiparasitic activity of a drug (SI = CC 50 /IC 50 ). The higher the SI value, the safer and more effective the drug is during the in vivo treatment. It is considered that a drug specimen presents a promising activity when the SI value is �10 [29].

Antileishmanial activity in vivo; mice infection and treatment
Leishman-Donovan Unit (LDU) assay. BALB/c mice were infected through the lateral tail vein with 2×10 7 amastigotes of wild-type AG83 L. donovani isolated from infected mice spleen. To examine the potency of infection, spleen biopsies were performed by randomly selecting 3-4 mice from each group on day 30 post-infection. Spleen smears were stained with Giemsa to confirm infection. The infected animals were divided into three groups consisting 4-5 animals per group. The groups were classified as the infected control group, BPQ treated group and BPQ+NP treated group. On the next day, the treatment was initiated by intraperitoneal administration of 0.2 mL of BPQ or BPQ+NP at a dose of 1 mg/kg body weight for 5 consecutive days. The infected control group was injected an equal volume of PBS during the entire dosing period. On day 7 post-treatment, the weight of the spleen was measured immediately after autopsy. After that spleen impression smears were prepared on glass slides and stained with Giemsa to analyze infected visceral organ (spleen) microscopically. The parasite burdens were determined by evaluating the LDU values, which were calculated by multiplying the number of amastigotes counted per 500 nuclei with the spleen weight in mg [30]. The % of amastigote inhibition was calculated using the equation and the % of amastigote replication suppression was calculated by using the equation where NB is the number of amastigotes per 500 nuclei in the spleen before treatment, NA is the number of amastigotes per 500 nuclei in the spleen after treatment, and NC is the number of amastigotes per 500 nuclei in the spleen after treatment in the control group. Culture microtitration assay. Three animal groups consisting 4-5 mice per group were divided as the infected control group, BPQ treated group, and BPQ+NP treated group. Mice were infected with wild-type L. donovani AG83 parasites by injecting 0.2 mL sterile PBS containing 2×10 7 amastigotes cells in the lateral tail vein. After 30 days of infection, the infected mice were administered with 0.2 mL of sterile BPQ or BPQ+NP samples through the intraperitoneal route with a dose of 2 mg/kg at a 5-days interval. The infected control group was injected with 0.2 mL of sterile water during the entire dosing period. Seven days post-treatment, the animals were sacrificed, and their spleens were aseptically removed and weighed for determination of amastigote burden by microtitration assay as described elsewhere [31]. Wells were examined for viable and motile promastigotes. The final titer was the last dilution for which the well contained at least one parasite. The number of amastigotes per gram of spleen tissue (amastigote burden) was calculated by the following equation

Amastigote burden ¼
Geometric mean of reciprocal titers from each duplicate Weight of homogenized cross section � 400: ð5Þ

Evaluation of in vivo toxicity by biochemical markers
BALB/c mice were grouped into three categories containing 5 animals in each group; normal group, infected group and BPQ+NP treated group. Infection was established as described above. The BPQ+NP treated group received 4 doses of 0.2 mL of the drug loaded nanoparticles, intraperitoneally with each dose of 2 mg/kg at 5-days interval. Normal group and infected group received 0.2 mL of water for injection by similar way. After the completion of dosing schedule, blood samples were collected by cardiac puncture. Then the levels of different enzymes like alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN) and creatinine (CRE), were analyzed using commercially available biomarker enzymatic assay kit (Thermo Fischer Scientific).

Statistical analysis
GraphPad Prism 5 version was used for statistical analysis. An unpaired t-test (two-tailed) was applied for the determination of significance of IC 50 and CC 50 values of BPQ+NP over BPQ in axenic amastigotes, intracellular amastigotes, inhibition of infected and non-infected macrophages. It was also used to analyze amastigote inhibition in spleen and amastigote replication suppression. The Mann-Whitney test (two-tailed) was used for the determination of significance of mean amastigotes per 500 nuclei in spleen and LDU values.

Characterization of the nanoparticles and cellular uptake
Phytoliths, or biogenic porous SiO 2 particles, are accumulated in significant amounts in many plants, including rice, horsetail, and barley. Here, phytoliths from barley husks, produced in high volumes as agricultural residues, were used to produce the silicon nanoparticles. The porous SiO 2 powder was purified from barley husk ash and reduced to PSi by magnesiothermic reduction [15]. The material was thermally hydrocarbonized (THCPSi) to produce a hydrophobic surface enabling efficient loading of BPQ and slow release. The hydrocarbon surface was further functionalized with undecylenic acid (UnTHCPSi) to make the particles negatively charged. The composition and crystallite size of UnTHCPSi (NP) were measured using XRD. The diffractogram consisted of broad Si peaks indicating line broadening caused by a small crystallite size of (11.4 ± 0.2) nm (Fig 2A). No other diffraction peaks were observed. In the FTIR

PLOS NEGLECTED TROPICAL DISEASES
Porous silicon nanovectors of buparvaquone against visceral leishmaniasis therapy spectra (Fig 2B), typical Si-OH groups at 3500 cm -1 was observed. The peaks around 2850-2970 cm -1 was ascertained to be the C-H x groups. Characteristic stretching of C = O at 1714 cm -1 revealed the conjugation of carboxylic acid groups. The peak at 1630 cm -1 was attributable to the O−H bending of H 2 O molecules adsorbed on the sample surfaces. Peak around 1100 cm -1 was assigned to Si−O−Si stretching and around 600 cm -1 to Si-H 2 terminations.
The pore structure was characterized by N 2 sorption (Fig 2C). The NP had a high specific surface area of 183 ± 3 m 2 /g characteristic to nanoporous materials. The pore volume was 0.793 ± 0.006 cm 3 /g, and the pore size distribution was wide, consisting mainly of pores < 100 nm in diameter, the majority of them being < 40 nm (Fig 2D). It should be noted that the sample consisted of both inter and intraparticle pores and therefore the pore volume and pore size distribution are not directly indicative of the properties of the dispersed particles.
The loading degree of BPQ in the nanoparticles was 7% w/w. After the loading, the hydrodynamic size of the nanoparticles slightly increased (ca. 10%), but it did not affect the surface charge, as measured by DLS ( Table 1). The increase in the particle size can be attributed to the adsorption of BPQ molecules on the surface of the nanoparticles and slight aggregation of small particles. The measured polydispersity index (PDI) values of all samples were between 0.1-0.2, indicating monodispersed nanoparticle suspension.
Even after the freeze-drying there were no essential differences in the particle sizes, PDI values, nor the zeta potentials of the nanoparticles. Also, the visual inspection did not show agglomeration. The results affirmed that the characteristics of the nanoparticles were well-preserved, and freeze-drying facilitated the long-term storage of the nanoparticles. Thus, the freeze-dried nanoparticles were utilized for further investigations.
In vitro release kinetics of BPQ (Fig 3) showed that nearly 80% of BPQ was released in EtOH within 1 h from the freeze-dried nanoparticles. By contrast, the release in water was negligible (< 2%) due to the hydrophobic nature of the drug even in 24 h. The release in cell culture medium (DMEM, pH 7.4) was reasonably low (< 10% in 1 h and < 30% in 24 h), implying that the majority of BPQ is still loaded inside the nanoporous structure of the carrier during phagocytosis into the macrophages. The cellular internalization process via phagocytosis started already at 1 h through the accumulation of the nanoparticles on the cell membranes (Fig 4). The internalization is seen within 4 h and the number of the nanoparticles increased inside the cells up to 24 h.

Anti-leishmanial activity against axenic and intracellular amastigotes
The efficacies of the drug specimens (BPQ+NP, BPQ, SSG, AmB, PMM, and MF) were examined against four different strains of L. donovani amastigotes both in the axenic and the intracellular model. Two main strains of the amastigotes were used: AG83 and GE1. The AG83 is a wild-type strain sensitive to SSG, whereas GE1 is SSG resistant isolate. The wild-type AG83 strain was further generated into SSG resistant (AG83-R, SSG) and PMM resistant (AG83-R, PMM) strains to broaden the spectrum of anti-leishmanial activity of the drug specimen. In the axenic model, the anti-leishmanial activity of BPQ was essentially improved when loaded into the nanoparticles (BPQ+NP). The activity was measured by the IC 50 values, which were higher with pure BPQ in comparison to BPQ+NP against all types of L. donovani strains ( Table 2). This means that high dose of pure BPQ was required to achieve the same inhibitory effect as with BPQ+NP. The concentrations of pure BPQ were higher than the BPQ+NP by 3, 8, 7 and 4-fold against the wild-type AG83, (AG83-R, SSG), (AG83-R, PMM), and GE1 respectively. Nevertheless, the IC 50 values with BPQ+NP were not as low as with AmB and MF, regardless of the type or the strain of L.donovani. But the efficiency of BPQ+NP in terms of the RI values were comparable with AmB and MF, in which low RI value indicated high efficiency of the drug against the amastigotes.
In the cellular model, pure BPQ reached the CC 50 values over 300 μM and IC 50 values at the range of only 2-5 μM (Table 3). These values are consistent with the values of >229.75 in peritoneal macrophages and 1.5 μM against L. infantum infection for CC 50 and IC 50 , respectively, reported by Reimao et al. [32]. These high CC 50 values suggest that the BPQ has only little in vitro toxicity to the healthy macrophages, whereas the small IC 50 values indicate the high activity in killing the intracellular amastigotes. In case with BPQ+NP, the CC 50 values were even larger than with pure BPQ, implying that the BPQ+NP was less toxic to the host cells (Table 3). This means that the nanoparticles are functional carriers to deliver BPQ inside the cells, and most of the drug is consumed in killing of the intracellular amastigotes. Like in axenic model, the IC 50 values with BPQ+NP was lower than pure BPQ indicating that the BPQ+NP was more active than the pure BPQ to kill the amastigotes. The IC 50 values of pure BPQ were larger than with BPQ+NP by 3, 5, 5, and 4-fold against AG83, (AG83-R, SSG), (AG83-R, PMM), and GE1 strain, respectively. The efficacy to selectively kill the L. donovani amastigotes were assessed by  Table 2. Drug sensitivity profiles against axenic L.donovani amastigotes (mean ± σ, n = 4).

(μM) a Resistance index (RI) b
Wild-type AG83 c AG83-R, SSG d AG83-R, PMM e GE1 f AG83-R, SSG AG83-R, PMM GE1 the selectivity index (SI); the higher the SI value, the more selective the drug was to kill the amastigotes while the cellular toxicity was diminished. According to the SI values, BPQ+NP were clearly more selective in comparison with pure BPQ as well as other conventional drugs against both the wild-type and the drug-resistant strains. It should be noticed in Tables 2 and 3 that there was cross-resistances of (AG83-R, PMM) to SSG and (AG83-R, SSG) to PMM.

Treatment efficacy on in vivo intracellular amastigotes
LDU assay. In LDU experiment spleens were taken for determination of organ weight and parasite burden to the immediate effects of drugs and nanoformulation on intracellular amastigotes ( Table 4). The mice were treated for 5 consecutive days (1 mg/kg). Parasite burden in vivo was assessed by microscopic observation of amastigotes against host cell nuclei on tissue imprints and expressed as LDU values. The amastigote burden in spleen in the groups treated with BPQ and BPQ+BP was low. The amastigote burden in spleen was reduced more effectively by 20-fold and 170-fold in the BPQ and BPQ+NP treated groups, respectively when compared with the control group. Crucially, the inhibition of the amastigotes with BPQ+NP reached nearly 100% while with pure BPQ, it was only around 50%. Nevertheless, the amastigote replication suppression by BPQ was over 90% either with or without the use of the nanoparticles.
Culture microtitration assay (CMA). Using the quantification of the amastigotes by simple microscopic observation may lack some amastigotes in stained organ. We carried out more Table 3. Drug sensitivity profiles against intracellular L.donovani amastigotes (mean ± σ, n = 4).

Drug CC 50 (μM) a Cellular IC 50 (μM) Selectivity index (SI) b
Wild-type AG83 AG83-R, SSG AG83-R, PMM GE1 AG83 AG83-R, SSG AG83-R, PMM GE1 sensitive culture microtitration assay (CMA) to evaluate the parasite burden by determining the number of live amastigotes present in the infected spleen. The mice were dosed either 2 or 4 times at 5 days interval (2 mg/kg). After seven days post-treatment the number of amastigotes per gram of spleen was calculated using CMA method. With free BPQ, 50% of the parasite burden was reduced, whereas the BPQ+NP removed the burden almost by 100% ( Table 5). The number of dosing strengthen the effect both with BPQ (52% vs. 64%) and with BPQ+NP (89% vs. 98%). These numbers are in concordance with the results in LDU method indicating that BPQ+NP was more effective than BPQ in the VL therapy in mice.

In vivo toxicity
The levels of biological markers, alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), and creatinine (CRE) were analyzed to study the in vivo toxicity to liver and kidney (Fig 5). The biomarker levels of the normal mice were considered 100%, to which the toxicity levels of the infected mice and the BPQ+NP group were compared. With the normal mice infected with L. donovani, the biomarker levels increased even though the differences are not statistically different (p > 0.05). However, all four biochemical markers remained at the normal level with the BPQ+NP group. The results indicate that BPQ+NP were not toxic in repeated dosing, and there was no kidney or liver injury.

Significances of the nanocarriers and possible mechanism of action
Ever since the anti-leishmanial activity of BPQ was discovered, new strategies to improve its bioavailability has been introduced. For example, derivation of more water soluble BPQ prodrugs (buparvaquone-3-phosphate and 3-phosphonooxymethyl-buparvaquone) were developed [12]. The buparvaquone-3-phosphate was shown to be the most effective prodrug but only small reduction (34%) in the liver parasite burden in BALB/c mice was achieved [12,33]. Apart from BPQ, new therapeutics such as bisnaphthalimidopropyl derivatives [34] and plant-based extracts e.g., artemisinin [35] have been shown as potential anti-leishmanial drugs but their utility in native form is limited because of their low bioavailability and high toxicity. However, when the drugs were incorporated with nanoparticles, their efficacies were significantly improved [34,35]. For example, artemisinin loaded with poly( D,L -lactic-co-glycolic acid) PLGA nanoparticles reduced the parasite burden infected with L.donovani to 85% (liver) and 82% (spleen) while the free drug reduced the burdens only moderately (~50%) [35]. Similarly, Lima et al., reported the anti-leishmanial activity of bisnaphthalimidopropyl loaded PLGA nanoparticles to be 80% (spleen) and up to 50% (liver) more effective than the free form of the drug against L. infantum infections [34]. As the nanoparticles themselves were non-toxic, the activities of the payloads were significantly elevated to combat the VL infections.
Despite the therapeutics, such as BPQ, being nonresistant to VL, the toxicity hinders their use in native form and requires a delivery platform to make them safer and improve their bioavailability and efficacy. Drug formulations based on functionalized nanocarrier is an option to facilitate their therapeutic use. For active drug targeting, the nanoparticle surfaces have been anchored with a variety of specific moieties like mannose [36][37][38] and phosphatidylserine (PS) [32,[39][40][41][42]. The crafted functional moiety operated as active ligands that specifically bind with the receptor of infected cells, and therefore, enhanced the selectivity and uptake of the nanocarriers towards the target site. For example, mannose anchored nanoparticles were more effective to transport AmB to the diseased site in comparison with the bare nanoparticles or pure AmB [36][37][38]. In vivo studies examined by Shahnaz et al., showed that the efficacy of AmB loaded in mannose grafted thiolated chitosan nanoparticles was as high as 89%, while the non-targeted thiolated chitosan nanoparticles showed only 36% reduction in the L.donovani burden in spleen [36]. Similarly, doxorubicin loaded in PS grafted nanocapsules inhibited the parasite burden by 85% in spleen whereas the efficacy of doxorubicin was lower (73%) when loaded with unfunctionalized nanocapsules [42]. Another study conducted by Khatik et al., revealed that AmB loaded in PS modified gelatin nanoparticles reduced the parasite burden infected with L. donovani in spleen (85%) whereas with bare gelatin nanoparticles, the efficacy reduced to 71% at corresponding doses of AmB [41].
In case with BPQ, when the drug was encapsulated with PS grafted liposomes, the efficacy of BPQ was substantially improved. For instance, the parasite burden in the spleen and in the liver were reduced up to 98 and 96%, respectively [39]. The improved efficacy was attributed to greater interaction of negatively charged PS-liposomes with "scavenger receptors" of the macrophages leading to effective internalization of the encapsulated BPQ [32,39,40].
Even though surface functionalization has shown to improve the targeting efficacy of the nanocarriers, their wider use is meager due to the complex processes involved in the grafting of the functional component [43]. Consequently, the production cost of the nanomedicine is susceptible to escalate. Herein, we produced affordable biogenic PSi nanocarriers from lowcost agricultural residue (barley husk), alongside valorizing the biomass. Conventionally, PSi nanoparticles are produced by electrochemical etching of Si wafers, but their production cost is relatively high and may not be viable in VL therapy since the disease is mostly endemic in developing countries. Furthermore, biocompatibility of PSi in drug delivery systems has been explored in several biomedical applications. The PSi is biodegradable material that does not induce toxicity at moderate concentrations since the degradation product, orthosilicic acid, is non-toxic to human body, which is eventually excreted through the urine [44][45][46][47].
In the present study, the biogenic PSi have been shown to be biocompatible since the cytotoxicity level of BPQ were significantly lowered when encapsulated in nanoparticles (BPQ +NP) in comparison with pure BPQ and other conventionally used drugs. The in vitro results revealed that the BPQ-loaded nanoparticles were effectively internalized into the cells, and thereafter, enhanced the permeability of BPQ inside the cells. Subsequently, the BPQ-loaded in the nanoparticles had reasonable anti-leishmanial activity both against the wild-type and the drug resistant L. donovani strains. In vivo inhibition studies examined in spleen also revealed that the efficacy of BPQ was significantly improved when the drug was encapsulated in the biogenic PSi nanoparticles. The parasite burden was almost completely inhibited with BPQ+NP (~100%), while the inhibition was only 50% with the native BPQ.
For leishmaniasis chemotherapy trypanothione system, surface lipids, polyamine and sterol biosynthesis, and efflux pumps are putative drug targets [48,49]. In addition, P-glycoprotein in the infected cells plays major role for multidrug resistance which limits the leishmaniasis chemotherapy. Nanotechnology based drug delivery systems have the ability to overcome drug resistance and improve cellular uptake due to the nano size of the particles which makes them capable to penetrate through biological barriers, enhance cellular uptake and deliver the drug in infected tissues [50][51][52]. Furthermore, it has been well established that the nanoparticles overcome multidrug resistance by down-regulating the overexpression of P-glycoprotein [53,54].
It can be proclaimed that the pH-responsive coordination interactions between the hydrocarbonized surfaces of the nanoparticles and BPQ not only improved its loading efficiency but also helped in its controlled release in the acidic phagolysosomal environment. Moreover, the internalized nanocarriers typically inhibit the efflux pump proteins to remove the drug molecules from the cells making the treatment more effective alleviating the drug resistance [55]. So, after the endocytosis of BPQ+NP into the infected cells, P-glycoprotein could not pump out the nanoparticles and get rid of being effluxed. This could be the main reason for the improved antileishmanial activity of BPQ when loaded in biogenic nanocarriers against drug resistant strains of VL.
Nevertheless, specific surface functionalization of the PSi nanoparticles may further improve their selectively and uptake towards the targeted site. In future, thorough investigations of the drug loaded nanoparticles in terms of different administration routes, toxicity, and their side effects shall be studied in more detail.

Conclusion
Visceral leishmaniasis is a neglected tropical disease for which no vaccine is available. Novel treatments against leishmaniasis are globally focused on repositioning of approved drugs. Buparvaquone has been used long in veterinary treatment against bovine theileriosis. It has been shown to have excellent in vitro activity against L. donovani, but its lipophilic nature hinders its use in vivo. Herein, we loaded the buparvaquone in biogenic porous silicon nanoparticles to improve its bioavailability. The nanoparticles were synthesized from barley husk, which is an agricultural residue and widely available. This approach to develop low-cost biogenic nanocarrier system is also an efficient way to valorize the biomass. The nanoparticles were able to encapsulate buparvaquone adequately and release it in a controlled manner intracellularly. The in vitro antileishmanial studies revealed that the buparvaquone-loaded nanoparticles were susceptible to the wild-type and different resistant strains of the L. donovani amastigotes with high degree of selectivity. The present work reports for the first time the successful utilization of low-cost biogenic nanomaterial in the antileishmanial nanomedicine having minimal toxicity, improved cellular uptake and high degree of selectivity against drug-resistant visceral leishmaniasis. New affordable and low-toxic multifunctional nanomedicine was developed as an alternative to treat the macrophage associated neglected tropical disease that affects mainly the poor people in the developing countries. The present study opens the route to further develop the porous silicon nanoparticles, for instance, by anchoring specific ligands on the particle surface to improve targeting. Such nanomedicine has high potential to be developed from the laboratory to the clinic, and to establish a new treatment regimen for antileishmanial chemotherapy especially in case of the drug resistant visceral leishmaniasis.
Supporting information S1 Scheme. Semantic diagram representing the buparvaquone loaded nanoparticles (left) and their internalization in macrophage infected with visceral leishmaniasis (right). (TIF)