Leishmanicidal Effect of Synthetic trans-Resveratrol Analogs

Background Stilbene-based compounds show antitumoral, antioxidant, antihistaminic, anti-inflammatory and antimicrobial activities. Here, we evaluated the effect of the trans-resveratrol analogs, pterostilbene, piceatannol, polydatin and oxyresveratrol, against Leishmania amazonensis. Methodology/Principal Findings Our results demonstrated a low murine macrophage cytotoxicity of all four analogs. Moreover, pterostilbene, piceatannol, polydatin and oxyresveratrol showed an anti-L. amazonensis activity with IC50 values of 18 μM, 65 μM, 95 μM and 65 μM for promastigotes, respectively. For intracellular amastigotes, the IC50 values of the analogs were 33.2 μM, 45 μM, 29 μM and 30.5 μM, respectively. Among the analogs assayed only piceatannol altered the cell cycle of the parasite, increasing 5-fold the cells in the Sub-G0 phase and decreasing 1.7-fold the cells in the G0-G1 phase. Piceatannol also changed the parasite mitochondrial membrane potential (ΔΨm) and increased the number of annexin-V positive promastigotes, which suggests incidental death. Conclusion/Significance Among the analogs tested, piceatannol, which is a metabolite of resveratrol, was the more promising candidate for future studies regarding treatment of leishmaniasis.


Introduction
Leishmaniasis is a public health problem that affects 98 countries on 5 continents, and approximately 1.1 to 1.7 million cases of leishmaniasis occurs each year [1]. Pentavalent antimonials are the first-line drugs for treating leishmaniasis; amphotericin B, pentamidine, and paramomycin, are secondary options for the treatment of resistant cases [2,3]. All of these drugs have problems that limit their use, such as side effects, the induction of parasite resistance, in-patient

Ethics Statement
BALB/c mice (8-10 weeks) obtained from Federal Fluminense University animal facility were housed at Federal University of Rio de Janeiro in temperature-controlled room (23 ± 1°C), with a 12/12 h light/dark cycle, and water and food ad libitum. Euthanasia was carried out in a saturated atmosphere of isoflurane, and all efforts were made to minimize animal suffering. All experiments with animals were performed in strict accordance with the Brazilian animal protection law (Lei Arouca number 11.794/08) of the National Council for the Control of Animal Experimentation (CONCEA, Brazil). This study protocol was approved by the Committee for Animal Use of the Universidade Federal do Rio de Janeiro (Permit Number: IMPPG 024).

Anti-amastigote Activity
Mice peritoneal macrophages obtained after stimulation with thioglycolate for 3 days were harvested in RPMI 1640 medium (LGC Biotecnologia, São Paulo, Brazil), 20 μg/mL of gentamycin and were plated onto 13 mm 2 coverslips inside 24-well plates for adherence for 2 h at 37°C, 5% CO 2 . Non-adherent cells were removed, and the macrophages were incubated overnight in RPMI supplemented with 10% FCS, as described above. Adhered macrophages were infected with L. amazonensis promastigotes (stationary growth phase) at a 10:1 parasite/macrophage ratio and incubated for 1 h at 34°C, 5% CO 2 . Free parasites were washed out with 0.01 M phosphate buffered saline (PBS), and the cultures maintained for 24 h at 37°C, 5% CO 2 . Infected macrophage cultures were treated with different concentrations of the analogs for an additional 24 h at 37°C, 5% CO 2 . Monolayers were then washed with PBS at 37°C, fixed in methanol, and stained with Giemsa. The number of amastigotes and the percentage of infected macrophages were determined by counting at least 200 cells in duplicate cultures. Endocytic indices were obtained by multiplying the percentage of infected macrophages by the mean number of amastigotes per infected macrophage. The results were expressed as the percentage of surviving macrophages based on endocytic indices of treated and untreated macrophages as reported previously [19].

Selectivity Index Calculation
The selectivity index was calculated as the ratio of murine peritoneal macrophages CC 50 by L. amazonensis intracellular amastigotes IC 50 .

Nitric Oxide Production
Thioglycolate-stimulated peritoneal macrophages were obtained as described above (10 6 cells/ well in 24-well plate) and were activated with 1 μg/ml LPS (Sigma-Aldrich) or left untreated. After incubating the cells for 24 h at 37°C, 5% CO 2 , they were treated with 33.2 μM pterostilbene, 45 μM piceatannol, 29 μM polydatin and 30.5 μM oxyresveratrol. The nitrite concentrations in the culture supernatants were determined using the Griess method. The reaction was read at 540 nm, and the concentration of NO 2 determined using a standard curve of sodium nitrite. The results are expressed as micromolar concentrations of nitrite [22].

Nitric oxide-trapping capacity
A cell-free system with an NO donor was used to test the capacity of resveratrol analogs to trap NO. SNAP (s-nitroso n-acetyl DL-penicillamine, Sigma) liberates nitric oxide in solution, which is then transformed to nitrite in the medium. The addition of a NO scavenger to the SNAP solution results in nitrite decay in the supernatant. Using this protocol, 33.2 μM pterostilbene, 45 μM piceatannol, 29 μM polydatin and 30.5 μM oxyresveratrol were incubated with 1 mM SNAP. One milimolar rutin (Sigma), a known NO scavenger, was used as a positive control for this assay. After 6 h of incubation, the nitrite concentration was determined using the Griess method. The results are expressed as the μM concentration of nitrite calculated in comparison with the sodium nitrite standard curve [18].

Detection of Reactive Oxygen Species (ROS)
Mice peritoneal macrophages were adhered to 96-well opaque culture plates and then infected or not with L. amazonensis promastigotes. Twenty-four hours post-infection, they were treated with pterostilbene, piceatannol, polydatin and oxyresveratrol at their respective IC 50 concentrations and were stimulated or not with 1 μg/mL phorbol 12-myristate 13-acetate (PMA, Sigma). The cells were then stained with 50 μM dihydrorhodamine 123 (DHR 123, Life Technologies), and ROS was measured immediately using 500/526 nm excitation/emission wavelengths.

Cell Cycle Analysis
Promastigotes were incubated in Schneider's complete medium, with or without 17.7 μM pterostilbene, 65 μM piceatannol, 95.5 μM polydatin, 65 μM oxyresveratrol and 1% DMSO for 48 h. The cells were washed with PBS and fixed in 70% (v/v) ice-cold methanol/PBS for at least 1 h at 4°C. The fixed cells were washed once with PBS and then incubated in PBS supplemented with 10 μg/mL propidium iodide (PI) and 20 μg/mL RNAse at 37°C, 45 min, as reported [23]. For each sample, 10000 events were collected on a BD FACScalibur (Becton and Dickson) and analyzed using CellQuest software.

Annexin V binding assay
Double staining for annexin V-fluorescein isothiocyanate (FITC) and PI was performed with the Annexin-V apoptosis detection kit (Molecular Probes). Promastigotes were treated or not with different concentrations of piceatannol for 48 h. They were then washed twice in cold annexin V-buffer and centrifuged at 2760 x g for 10 min. Pellets were resuspended in 20 μL of annexin V FITC, and after 15 min of incubation in the dark, 480 μL of annexin V-buffer was added according to the manufacturer instructions. Annexin V-FITC labeling was recorded on a BD FACScalibur (Becton and Dickson) and analyzed using CellQuest software.

Statistical Analysis
Data were analyzed using Student's t-test when comparing two groups or one-way ANOVA for more than two groups using the software GraphPad Prism. P values of less than 0.05 were considered significant.

Anti-promastigote Activity
Initially, the activity of trans-resveratrol analogs on Leishmania amazonensis promastigotes survival was assayed (Table 1). Our results demonstrated an anti-leishmanial activity of these analogs with IC 50 values of 18, 65, 95.5 and 65 μM for pterostilbene, piceatannol, polydatin and oxyresveratrol, respectively, after 48 h of treatment. The IC 50 of Amphotericin B used as a control was 0.1 μM (Table 1).

Host cell viability assays
To test the safety of trans-resveratrol analogs on host cells, we evaluated the dehydrogenases activity of macrophages using the XTT method. Our results indicated that the cytotoxic concentration for 50% of the population (CC 50 ) was 181, >400, 150, and 128 μM for pterostilbene, piceatannol, polydatin and oxyresveratrol, respectively ( Table 1). The CC 50 of Amphotericin B was > 5 μM.

Anti-amastigote activity
Next, we tested the effect of trans-resveratrol analogs on amastigotes, which are the parasite forms present in the mammalian host and are responsible for the infection. The anti-amastigote activity was tested after 24 h of treatment, and our results showed that the IC 50 value of pterostilbene, piceatannol, polydatin and oxyresveratrol was calculated as 33.2 μM, 45 μM, 29 μM and 30.5 μM, respectively ( Table 1). The IC 50 of Amphotericin B was 8.8 μM.

Nitric Oxide (NO) Production and Nitric Oxide-trapping Capacity
NO is recognized as an important mediator of Leishmania death, therefore we investigated whether trans-Resveratrol analogs were able to modulate NO production by macrophages. Our results showed that pterostilbene decreased the NO production of uninfected macrophages stimulated or not with LPS by 5.3-and 3.8-fold, respectively (Fig 2A). Piceatannol decreased the NO production only in the uninfected macrophages stimulated with LPS by 5.0-fold ( Fig  2C). The treatment of uninfected and unstimulated macrophages with polydatin reduced NO production by 2.8-fold, and polydatin or oxyresveratrol treatments decreased the NO production of infected macrophages by 6.1-and 2.0-fold, respectively (Fig 2E and 2G). We used a cell-free system to rule out a possible NO scavenging effect of the trans-resveratrol analogs. Thus, s-nitroso n-acetyl-DL-penicillamine (SNAP), a NO donor, was incubated in the presence or absence of pterostilbene, piceatannol, polydatin and oxyresveratrol. As a positive control, rutin, a NO scavenger, was added to the SNAP solution, and it decreased NO levels by 72%. Addition of piceatannol, polydatin or oxyresveratrol at 45 μM, 29 μM and 30.5 μM, respectively, did not reduce the levels of NO, indicating that the decrease in NO production that resulted from the treatment with these compounds was not due to a scavenging effect of the drugs (Fig 2D, 2F and 2H). However, 33.2 μM pterostilbene decreased NO levels by 22%, indicating that this analog has a NO scavenging effect (Fig 2B).

Annexin-V Binding Assay
The annexin V assay evaluated the potential induction of incidental death in promastigotes by piceatannol (Fig 5). Our results showed that piceatannol induced a dose-dependent increase in annexin-V labeling (Fig 5C-5E and 5G), and 100 μM piceatannol increased the percentage of annexin-V positive promastigotes by 2.5-fold in relation to untreated control (Fig 5G). Moreover, piceatannol induced a dose-dependent increase in PI labeling (Fig 5C-5E). Miltefosine, used as a positive control, increased the annexin-V binding to promastigotes by 13.2-fold ( Fig  5F and 5G).

Parasite Mitochondrial Alterations
Leishmania is a single mitochondrion parasite and this organelle is involved in apoptotic cell death. In order to address the parasite mitochondrial toxicity, reduction of the mitochondrial membrane potential (ΔCm) was assessed in piceatannol-treated promastigotes using the JC-1 assay (Fig 5). Our results showed a reduction of 2.8-fold in the mitochondrial membrane potential in parasites treated with 65 μM piceatannol (Fig 5H). Treatment with 1% DMSO did not change the mitochondrial membrane potential.

In Silico Evaluations
The rate of drug-likeness indicates whether a compound has certain essential features observed in most of the drugs on the market. Thus, this index shows the potential of a certain compound as a drug candidate. Potential drug-likeness of the trans-resveratrol analogs was calculated, and AMB was used as a control because it is a drug widely used for leishmaniasis treatment. Among the analogs, only piceatannol showed a potential drug-likeness to AMB (Fig 6A). The Drug-Score index is a parameter that combines characteristics of the compound (ClogP, solubility, theoretical toxicity, Drug-Likeness and molecular weight) and is calculated using Osiris software, which evaluates whether a certain compound might be a future drug candidate. According to established criteria, values closer to 1 indicate a higher probability of a compound being a future drug candidate. The trans-resveratrol analog that had the best drugscore was piceatannol (0.31). Pterostilbene, polydatin and oxyresveratrol had drug-score values of 0.24, 0.25 and 0.28, respectively. AMB had a drug-score value of 0.27 (Fig 6B).
The prediction of toxicological risks (mutagenic, tumorigenic, irritant and toxic risk in reproduction) of these compounds was also analyzed. All analogs showed no tumorigenic, irritant or mutagenic effects. AMB did not show any of the toxic effects evaluated (Fig 6C).
The results presented in Table 3 showed that there was no violation of the Lipinski's rules [25], which evaluate physico-chemical parameters, suggesting the importance of the resveratrol analogs as drug candidates with good oral bioavailability.

Discussion
Stilbenes are natural products with low molecular weights isolated from several plant species, and resveratrol is by far the most studied among them, with several of its biological effects having already been described [26,27]. Among its effects, we recently described the leishmanicidal activity of resveratrol against L. amazonensis [7]. Here, we describe the leishmanicidal effect of four trans-resveratrol analogs: pterostilbene, piceatannol, polydatin and oxyresveratrol. Our study showed, for the first time, that the analogs polydatin and oxyresveratrol are active against L. amazonensis promastigotes and amastigotes. Considering the CC 50 determined using the XTT assays, the therapeutic indices for amastigotes were 5.2 and 4.3 for polydatin and oxyresveratrol, respectively. The therapeutic index for pterostilbene was 5.4, and that for piceatannol was >8.9. Interestingly, although it has been reported that piceatannol causes a loss of mitochondrial potential and the release of cytochrome c [28], it was found to be the less toxic of the analogs tested with XTT, which measures mitochondrial dehydrogenases. Duarte and colleagues [29] demonstrated that piceatannol has IC 50 values of 4.2 μg/mL (17.2 μM), 3.9 μg/mL (16 μM) and 5.7 μg/mL (23.3 μM) for promastigotes of L. donovani, L. infantum and L. major, respectively. They also found an IC 50 of 7.4 μg/mL (30.3 μM) for piceatannol against amastigotes of L. donovani in RAW macrophages. We determined that piceatannol is effective against L. amazonensis with an IC 50 of 65 μM for promastigotes and an IC 50 of  45 μM for intracellular amastigotes. Tolomeo and colleagues [30] showed that pterostilbene has LD 50 values of 7.6 μg/mL (29.6 μM) and 10.3 μg/mL (40.2 μM) for the promastigote and amastigote forms of L. infantum. Corroborating these data, we found IC 50 values of 17.7 μM and 33.2 μM for pterostilbene against L. amazonensis promastigotes and amastigotes, respectively. To evaluate the leishmanicidal mechanisms activated by the trans-resveratrol analogs, we tested their ability to modulate NO production in murine macrophages. Our results demonstrated that pterostilbene and piceatannol decrease NO production by LPS-stimulated macrophages. Hsu et al. [31] showed that pterostilbene decreased the mRNA expression of iNOS in TNF-α-induced 3T3-L1 adipocytes. Moreover, resveratrol and pterostilbene inhibited the production of nitric oxide and the secretion of TNF-α by LPS-stimulated RAW 264.7 cells. Resveratrol and pterostilbene also inhibited the expression of TNF-α, IL-1β, IL-6, and iNOS genes in LPS-stimulated RAW 264.7 cells and LPS-stimulated peritoneal macrophages from both C57BL/6 and BALB/c mice [32]. Interestingly, we demonstrate that at least a portion of the NO inhibition by pterostilbene could be due to its capacity to scavenge NO, a property that it does not share with the other three analogs.
We showed that piceatannol reduces ROS production in agreement with Kim and colleagues [33], who demonstrated that 10 μM piceatannol attenuated hydrogen peroxide and peroxynitrite-induced apoptosis of a rat pheochromocytoma cell (PC12). Pterostilbene also decreased ROS in activated polymorphonuclear leukocytes [34]. It has been suggested that resveratrol and oxyresveratrol have protective effects against reactive oxygen and nitrogen species in a murine microglial cell line, by considerably decreasing NO and ROS levels [35]. Our results showed that oxyresveratrol decreased the production of ROS, but polydatin was a weak ROS inhibitor in unopsonized zymosan-stimulated murine macrophage Raw264.7 cells, human monocytes, and neutrophils in comparison to resveratrol [36]. In our assay conditions, piceatannol and oxyresveratrol were more potent ROS inhibitors than pterostilbene and polydatin. These data suggest that the leishmanicidal effect of the resveratrol analogs could not be attributed to ROS and NO generation in macrophages.
Previously, we showed that resveratrol increases promastigotes in the sub-G0/G1 phase of their cell cycle and that it induces mitochondrial membrane depolarization, suggesting that the parasites undergo incidental death [7]. Therefore, in the present study, the effect of trans-resveratrol analogs in the modulation of the parasite cell cycle was examined. Among the analogs tested, we found that only piceatannol affected the promastigote cell cycle, increasing the number of cells in the sub-G0 phase and decreasing the number of cells in the G0-G1 phase. Similarly to resveratrol, piceatannol changed the parasite mitochondrial membrane potential, suggesting incidental death as its killing mechanism. In agreement with our findings, Kita and colleagues [37] demonstrated that 25 and 50 μM piceatannol induced cell cycle arrest at the G2/M phase in AH109A hepatoma cells. Furthermore, piceatannol failed to induce apoptosis at 50 μM in AH109A hepatoma cells; however, at 100 μM, it induced strong apoptosis in those cells [37]. Another feature of the promastigote incidental death we observed was an increase in annexin-V binding to the surface of piceatannol-treated promastigotes.
The theoretical in silico pharmacokinetics approach is widely used in the initial study of ADMET (Absorption, Distribution, Metabolism, Excretion and Toxicity) properties (S1 Table) with the aim of reducing the costs of biological assays of compounds and a high probability of identifying future problems and toxicity; this saves both time and money [38]. It has been demonstrated in vitro that hepatic biotransformation of trans-resveratrol is dependent of the activity of trans-resveratrol isoform(s) and liver cytochrome P450 (CYP) [39]. The hepatic metabolism of trans-resveratrol into piceatannol and another tetrahydroxystilbenes has been shown to be catalyzed by recombinant human CYP1A1, CYP1A2 and CYP1B1. Our in silico ADMET properties validated that piceatannol may act allosterically to inhibit CYP1A2. We showed that resveratrol acts synergistically with amphotericin B against L. amazonensis [7]. Interestingly, CYP1A2 was not inhibited by amphotericin B; thus, it did not harm the metabolism of resveratrol (S1 Table).
The replacement of the hydroxyl group at carbon-3 with one on carbon-2 was not able to change the leishmanicidal effect of resveratrol as oxyresveratrol and piceatannol had similar IC 50 values on promastigotes of L. amazonensis. However, their mechanisms of action are different; piceatannol alters the parasite cycle and oxyresveratrol the effect of is not clear. This highlights the importance of studying the correlations between structural changes and biological effects.
The results of our study show that trans-resveratrol analogs have anti-Leishmania amazonensis activity of in vitro and that piceatannol is capable of inducing promastigotes' incidental death, suggesting that these compounds are promising candidates for future studies regarding leishmaniasis treatment.