Antileishmanial Activity and Structure-Activity Relationship of Triazolic Compounds Derived from the Neolignans Grandisin, Veraguensin, and Machilin G

Sixteen 1,4-diaryl-1,2,3-triazole compounds 4–19 derived from the tetrahydrofuran neolignans veraguensin 1, grandisin 2, and machilin G 3 were tested against Leishmania (Leishmania) amazonensis intracellular amastigotes. Triazole compounds 4–19 were synthetized via Click Chemistry strategy by 1,3-dipolar cycloaddition between terminal acetylenes and aryl azides containing methoxy and methylenedioxy groups as substituents. Our results suggest that most derivatives were active against intracellular amastigotes, with IC50 values ranging from 4.4 to 32.7 µM. The index of molecular hydrophobicity (ClogP) ranged from 2.8 to 3.4, reflecting a lipophilicity/hydrosolubility rate suitable for transport across membranes, which may have resulted in the potent antileishmanial activity observed. Regarding structure-activity relationship (SAR), compounds 14 and 19, containing a trimethoxy group, were the most active (IC50 values of 5.6 and 4.4 µM, respectively), with low cytotoxicity on mammalian cells (SI = 14.1 and 10.6). These compounds induced nitric oxide production by the host macrophage cells, which could be suggested as the mechanism involved in the intracellular killing of parasites. These results would be useful for the planning of new derivatives with higher antileishmanial activities.


Introduction
Cutaneous leishmaniasis (CL) is a parasitic infectious disease that affects the skin, cartilage and mucosa of the upper respiratory tract. The resulting ulcers can develop into destructive and disabling injuries, making the illness a serious public health problem [1]. There have been between 0.7 and 1.3 million new annual cases throughout the world, and 95% occur in the Americas, Mediterranean basin, Middle East and Central Asia [2]. Leishmania (Leishmania) amazonensis Lainson and Shaw 1972 is one of the etiological agents of CL. It produces the typical localized lesions or the diffuse form of the disease, in which the parasite spreads due to impaired cell-mediated immune response. Thus, the lesions tend not to heal spontaneously and to be more resistant to treatment [3,4]. The first choice drugs for CL are the pentavalent antimonials, which are associated to hepato, cardio and nephrotoxicity [5]. When these drugs are ineffective or cannot be prescribed, drugs such as amphotericin B, pentamidine or paramomycin are indicated, despite their degree of toxicity [6]. Resistance and high cost are other aspects that lead to the urgent need for new therapeutic options arising from natural products [7][8][9][10]. However, these products may have undesirable properties such as high toxicity, low solubility and bioavailability, which can be minimized by the development of synthetic derivatives [11].
Cassamale et al. [12] have synthetized a series of triazole derivatives from the tetrahydrofuran neolignans veraguensin 1, grandisin 2, and machilin G 3. These compounds have been considered important scaffolds in molecular modification studies due to their antileishmanial and antichagasic activities [13][14][15][16]. Cassamale et al. [12] demonstrated the antileishmanial activity of these triazole derivatives on promastigote forms; now the present work shows their activity on L. (L.) amazonensis intracellular amastigotes, searching for structure-activity relationship information to support the development of new drug candidates for CL.
The ability to inhibit the growth of parasites apparently depends on the presence and ratio of lipophilic/hydrophilic substituents on aromatic rings [18]. The transport of a compound across membranes may be influenced by the molecular hydrophobicity described by the octanol/water partition coefficient (ClogP) [19]. According to Lipinski's Rule of Five [20][21][22], compounds with logP < 5 have better absorption and permeation in vivo. Daunes and D'Silva [23] verified that among a series of molecules, the ones with higher logP (>2.7) were the most active against trypanosomatids once entering the host cell more easily. In other words, the molecular hydrophobic character improves the antileishmanial activity, which may indicate that the active compound must interact with a target system such as an enzyme or receptor, where the binding site is generally hydrophobic.
In our study, synthetic derivatives showed lipophilicity 2.8 < ClogP < 3.4 (Table 1), reflecting an adequate lipophilicity for the transport across membranes and resulting in potent antileishmanial activity.
We observed that ring B containing the trimethoxy substituent has influenced the antileishmanial activity independently of the substituent on the ring A (compounds 6, 10, 14 and 19) (Table 1).
Grandisin derivative 6 was active on intracellular amastigotes (IC 50 value of 9.4 µM), with the highest selectivity index. This compound was 66 times more toxic to amastigotes than to mammalian cells (Table 1). The substitution of tetrahydrofuran ring by triazolic ring may have resulted in increased activity, once it was described an IC 50 value of 98.05 µM (42.4 µg¨mL´1) for grandisin on L. (L.) amazonensis promastigotes [24]. Furthermore, its coefficient of hydrophobicity is lower than of its precursor 2 (2.8 and 3.7, respectively), which may have resulted in a better solubility and consequent in vitro activity. Indeed, more soluble grandisin derivatives have been synthetized in order to reduce its lipophilicity, which may limit in vivo studies [25]. It is important to note that the intracellular amastigote form is the target for drug candidates in the mammalian host.  Treatment with compound 6 had significantly decreased the infection index at the concentration of 6.25 µg¨mL´1, reaching 98.6% reduction (p < 0.0001) at the concentration of 50 µg¨mL´1 ( Figure 1A). 6 has induced an increased production of nitric oxide (NO) compared to untreated infected cells (control) (Figure 2A). Thus, NO production may be suggested as the mechanism of leishmanicidal action, especially because 6 does not seem to have a direct action on L. (L.) amazonensis once it was proven inactive on promastigote forms [12]. Treatment with compound 6 had significantly decreased the infection index at the concentration of 6.25 µg·mL −1 , reaching 98.6% reduction (p < 0.0001) at the concentration of 50 µg·mL −1 ( Figure 1A). 6 has induced an increased production of nitric oxide (NO) compared to untreated infected cells (control) (Figure 2A). Thus, NO production may be suggested as the mechanism of leishmanicidal action, especially because 6 does not seem to have a direct action on L. (L.) amazonensis once it was proven inactive on promastigote forms [12].  The other compounds from trimethoxy series (with substituents on ring A) were active (11) and moderately active (15 and 18), but less selective, with important cytotoxicity on mammalian cells Treatment with compound 6 had significantly decreased the infection index at the concentration of 6.25 µg·mL −1 , reaching 98.6% reduction (p < 0.0001) at the concentration of 50 µg·mL −1 ( Figure 1A). 6 has induced an increased production of nitric oxide (NO) compared to untreated infected cells (control) (Figure 2A). Thus, NO production may be suggested as the mechanism of leishmanicidal action, especially because 6 does not seem to have a direct action on L. (L.) amazonensis once it was proven inactive on promastigote forms [12].  The other compounds from trimethoxy series (with substituents on ring A) were active (11) and moderately active (15 and 18), but less selective, with important cytotoxicity on mammalian cells The other compounds from trimethoxy series (with substituents on ring A) were active (11) and moderately active (15 and 18), but less selective, with important cytotoxicity on mammalian cells ( Table 1). All of them were able to significantly decrease infection index, mainly 18, reaching 92.1% reduction (p < 0.0001) ( Figure 1A). This compound, however, did not induce a significant production of NO. 11 acted variably on NO release, with a production lower than untreated cells at the lowest concentrations, but higher than control at the highest concentrations tested (Figure 2A).
Compound 5 with a dimethoxy group as substitution pattern on rings A and B (veraguensin 1 derivative) was moderately active on intracellular amastigotes (IC 50 value of 21.3 µM) with relative selectivity (SI = 3.6). Silva Filho et al. [16] have demonstrated the activity of veraguensin on L. donovani promastigote forms (48.3 µM; 18 µg¨mL´1). Once again we may educe that the insertion of triazole ring may have improved the antileishmanial activity due to reduction of octanol/water partition coefficient (3.1 (5) versus 4.2 (1)). Compound 5 has significantly reduced the infection index from the concentration of 12.5 µg¨mL´1 ( Figure 1B). Furthermore, the mechanism of leishmanicidal action seems also to be independent of NO, once NO production was lower than control at the concentration of 25 µg¨mL´1 ( Figure 2B). These data corroborate those obtained by Konishi et al. [26], who verified the inhibition of NO production from LPS-activated murine macrophages by three veraguensin's position isomers.
Compound 14 was active and selective (IC 50 = 5.6 µM; SI = 14.1) (Table 1), and reduced the infection index at all concentrations tested (86.4% at the highest concentration, p = 0.0003) ( Figure 1B). It's important to note that 14 was a hybrid from veraguensin 1 and grandisin 2, and it had the best activity of the whole dimethoxy series. Furthermore, it significantly induced NO production at the lowest concentrations tested, and this may be suggested as a possible mechanism of leishmanicidal action ( Figure 2B).
Compounds from methoxy series showed IC 50 values from 13.1 to 16.8 µM (4, 8, 10, and 12, Table 1). Compound 4 was the most active of this series, and highly selective (IC 50 = 13.1 µM; SI = 66.9). As well as 12, this compound was able to induce an increase of NO production at the highest concentration tested ( Figure 2C).
The most active compound on intracellular amastigotes (19) came from methylenedioxy series (IC 50 = 4.4 µM). Compound 19 is a hybrid from machilin G 3 and grandisin 2 and was also quite selective (SI = 10.6) ( Table 1). 19 has significantly reduced the infection index and in addition it induced NO production twice as high than control at the highest concentrations tested ( Figure 2D). Cassamale et al. [12] have demonstrated this compound as highly active against L. (L.) amazonensis promastigote forms (IC 50 = 7.2 µM), and this suggests its direct action on the parasite. On the other hand, position isomer 18 (IC 50 = 29.8 µM, Table 1) was less active than 19, indicating the role of minor structural differences on the antileishmanial activity of these compounds.
Compound 13 did not show potential activity and selectivity (Table 1). Despite this, 13 was able to induce an increase in NO production ( Figure 2D), as well as its position isomer 12 ( Figure 2C). We should point out that all compounds from methylenedioxy series have induced NO production in greater or lesser degree ( Figure 2D). This implies that this group may be associated to cellular activation and/or cytotoxicity. For example, compound 17 (machilin G 3 analog) was toxic for the macrophages (SI = 0.3) with moderate antileishmanial activity (Table 1).

General Procedure for the Synthesis of Triazoles 4-19
To a solution of terminal acetylenes 25a-d (2 mmol, 1.0 equiv) and azides 27a-d (2 mmol, 1.0 equiv) in dichloromethane (2 mL) and water (2 mL), were added CuSO4·5H2O (0.128 mmol, 0.064 equiv) and sodium ascorbate (0.352 mmol, 0.176 equiv). The reaction mixture was stirred for 24 h. Then it was added a saturated solution of NH4Cl (30 mL) and the product was extracted with dichloromethane (3 × 20 mL). The organic phase was dried over anhydrous MgSO4, and the solvent was evaporated under reduced pressure. The products were purified by recrystallization from ethyl acetate.

ClogP
In order to estimate the molecular hydrophobicity of triazole derivatives 4-19, theoretical values of logP (ClogP) were calculated with ChemAxon software. The program estimated octanol/water partition coefficient based on a modified version of the method of Viswanadhan et al. [34].

General Procedure for the Synthesis of Triazoles 4-19
To a solution of terminal acetylenes 25a-d (2 mmol, 1.0 equiv) and azides 27a-d (2 mmol, 1.0 equiv) in dichloromethane (2 mL) and water (2 mL), were added CuSO 4¨5 H 2 O (0.128 mmol, 0.064 equiv) and sodium ascorbate (0.352 mmol, 0.176 equiv). The reaction mixture was stirred for 24 h. Then it was added a saturated solution of NH 4 Cl (30 mL) and the product was extracted with dichloromethane (3ˆ20 mL). The organic phase was dried over anhydrous MgSO 4 , and the solvent was evaporated under reduced pressure. The products were purified by recrystallization from ethyl acetate.

ClogP
In order to estimate the molecular hydrophobicity of triazole derivatives 4-19, theoretical values of logP (ClogP) were calculated with ChemAxon software. The program estimated octanol/water partition coefficient based on a modified version of the method of Viswanadhan et al. [34].

Parasites and Peritoneal Macrophages
In vitro antileishmanial activity was evaluated on peritoneal macrophages infected with. L. (L.) amazonensis intracellular amastigotes. Parasites (IFLA/BR/1967/PH8 strain) were routinely isolated from BALB/c mice and maintained as promastigotes at 25˝C in Schneider's Insect Medium (Sigma-Aldrich ® , St. Louis, MO, USA) supplemented with 20% fetal calf serum (FCS-Cultilab ® Campinas, Brazil) and 140 µg¨mL´1 gentamicin (Sigma-Aldrich ® ). Macrophages were obtained from peritoneal wash of BALB/c mice after euthanasia. 10 mL of RPMI 1640 (Sigma-Aldrich ® ) supplemented with 2% L-glutamine, 2.8% bicarbonate buffer, 100 U¨mL´1 penicillin and 100 mg¨mL´1 streptomycin were injected into the peritoneal cavity. After massage area, liquid was aspirated and transferred to tubes on ice. Peritoneal cells were quantified in Neubauer chamber after cellular exclusion with Trypan Blue staining (Sigma-Aldrich ® ).

In Vitro Antileishmanial Activity on Intracellular Amastigotes
Peritoneal cells (1ˆ10 5 cells/well) were added to 24-well plates containing circular coverslips. Plates were incubated for one hour at 37˝C/5% CO 2 to allow cell adhesion and then 1ˆ10 6 L. (L.) amazonensis promastigotes were added to each well. Plates were incubated at 35˝C/5% CO 2 for four hours and then cells were treated for 24 h with synthetic compounds 4-19 (6.25-50 µg¨mL´1). Amphotericin B (Sigma-Aldrich ® ) was used as the reference drug (0.25 to 2 µg¨mL´1) and untreated cells were used as negative control. Coverslips were processed as described by Rizk et al. [35]. The overall number of amastigotes was determined by counting 100 cells in six replicates. The half maximal inhibitory concentration (IC 50 ) was calculated using a nonlinear regression curve. Infection index was obtained as described by Paladi et al. [36].

Nitric Oxide Production
To evaluate the nitric oxide production (NO) by infected and treated peritoneal cells, 50 µL of culture supernatant from the antileishmanial assay were collected and incubated with equal volume of Griess reagent (1% sulfanilamide /0.1% naphthalene diamine in 5% phosphoric acid) for 10 min at room temperature. According to Ding et al. [37], the absorbance was determined at 540 nm and converted to NO 2´( µM) by comparing to a standard curve of known concentrations of sodium nitrite (1-10 µM) in RPMI medium (Sigma-Aldrich ® ).

Cytotoxicity Assay
Murine macrophages (J774.A1, Rio de Janeiro Cell Bank, Brazil) were seeded in 96-well plates (1ˆ10 5 mL´1) and incubated with compounds at 37˝C/5% CO 2 for 48 h at concentrations of 0.25-250 µg¨mL´1 to estimate IC 50 . Amphotericin B (Sigma-Aldrich ® ) was used as the reference drug (0.025-25 µg¨mL´1). Cell growth was evaluated according to Skehan et al. [38] using the sulforhodamine B assay. Dimethyl sulfoxide (DMSO, Vetec ® , Rio de Janeiro, Brazil) was used as negative control at the concentration used to solubilize the highest concentration of compounds. IC 50 was calculated by nonlinear regression curve. Selectivity index was calculated according to Medeiros et al. [39].

Ethical Aspects
This study received approval from the local Animal Experimentation Ethical Committee (CEUA/UFMS) under protocol 503/2013.

Conclusions
Among 16 synthetic derivatives of tetrahydrofuran neolignans veraguensin 1, grandisin 2 and machilin G 3, 15 showed high or moderate antileishmanial activity. Compounds 14 and 19, containing a trimethoxy substituent on ring B, were the most active against intracellular amastigotes, with low cytotoxicity on mammalian cells. These compounds induced nitric oxide production by the host macrophage cells, which could be suggested as the mechanism involved in the intracellular killing of parasites. These results would be useful for the planning of new derivatives with higher antileishmanial activities.