Exploration of ethylene glycol linked nitrofurantoin derivatives against Leishmania: Synthesis and in vitro activity

Leishmaniasis is a neglected tropical disease that is caused by the Leishmania parasite. It is estimated that there are more than 350 million people at risk of infection annually. Current treatments that are in clinical use are expensive, have toxic side effects, and are facing parasitic resistance. Therefore, new drugs are urgently required. In the quest for new, safe, and cost‐effective drugs, a series of novel ethylene glycol derivatives of nitrofurantoin was synthesised and the in vitro antileishmanial efficacy of the compounds tested against Leishmania donovani and Leishmania major strains. Arylated ethylene glycol derivatives were found to be the most potent, with submicromolar activity up to 294‐fold greater than the parent compound nitrofurantoin. Analogues 2j and 2k had the best antipromastigote activities with submicromolar IC50 values against L. major IR‐173 and antimonial‐resistant L. donovani 9515 strains.

of a vaccine against leishmaniasis. [7] Therefore, the control of leishmaniasis currently relies on chemotherapy. The first-line treatment drugs are pentavalent antimonials; meglumine antimoniate, and sodium stibogluconate. [8] However, due to the development of antimonial resistance, second-line treatment drugs including amphotericin B, miltefosine, pentamidine, and paromomycin have been used as alternatives. These drugs are all toxic, present with numerous adverse effects, and are expensive. [9] Therefore, this accentuates the need for new safe and cost-effective antileishmanial drugs.
A fast-track avenue that is often explored in finding new drug treatments is drug repurposing. [10] It involves the testing of existing drugs for use against other ailments and takes advantage of preexisting pharmacokinetic data to reduce the time taken to utilise these drugs as new treatment approaches. Drugs with multiple or generalised biological activities may serve as good targets for repurposing. [11] For example, it is well known that redox-active drugs can cause enzymatic stress thus killing parasites; [12] therefore, the use of an existing drug that contains a redox-active scaffold may also be beneficial in overcoming drug resistance development of infectious pathogens. In this study, a repurposing strategy was accordingly adopted for antileishmanial drug discovery.
Previously disregarded due to toxicity concerns, the interest in recently summarised in a review by Zuma et al. [13] Thus, some nitroaromatic scaffolds are currently in clinical use while others have been incorporated into molecules in an attempt to develop good, active, and safe novel therapeutics for the treatment of mammal infections, including leishmaniasis.
The nitrofuran scaffold is among these privileged pharmacophores and has been proven to be highly effective in treating infectious diseases of bacterial origin. Its excellent biological activity is primarily attributed to the nitro group, which can induce toxic metabolites under both aerobic and anaerobic conditions, leading to microorganism death. [14] Nitrofurantoin (NFT), in particular, is a redox-active antimicrobial drug that falls under a class of therapeutics referred to as clinical nitrofurans (cNFs; Figure 1). It is effective against both aerobic and anaerobic pathogens, [14] and its activity is credited to the reduction of the nitro group, the main pharmacophore by nitroreductases present in microorganisms. Nitroreduction can occur in the presence of oxygen (type II) or in the absence of oxygen (type I). [15] These reactions result in the formation of toxic species such as hydroxylamine, nitroso intermediates, and reactive oxygen/nitrogen (RON) free radicals that can bind to microbial DNA and other biological molecules, and ultimately lead to pathogen death. [12] Additionally, cNFs possess a second pharmacophore, the hydrazone moiety (Figure 1 in blue), containing a zwitterionic carbon that stabilises the nitrofuran ring (Figure 1 in red). The azo bond of the diazenyl functional group can be reduced by azoreductases into active primary metabolites, semicarbazides that can covalently bind to cellular proteins Azoreductases are NAD(P)H-dependent flavoenzymes present in a variety of bacteria found in the gut that can activate azo prodrugs into active metabolites. [16] In humans, azoreductase homologues are NAD(P)H quinone oxidoreductases.
They confer cytotoxicity through the reduction of cNFs similar to the azoreduction by azoreductases. [17] Semicarbazides are weakly toxic and potentially carcinogenic, [18] accummulate, distribute in high concentrations and persist for long periods of time in animal tissues, [19] and are found in edible animal products such as milk and meat. The potential risk of carcinogenicity F I G U R E 1 Clinical nitrofuran drugs for the treatment of infectious diseases. motivated the ongoing collective ban of cNFs as veterinary antibiotics for livestock antibacterial treatments by EMA in 1995 [20] and the US Food and Drug Administration (FDA) in 2002. [21] However, despite the potential carcinogenicity shortcoming, the development of parasitic resistance against cNFs is rare as they target multiple metabolic pathways, including carbohydrate and protein synthesis, DNA replication, and transcription. [22] Hence, cNF moiety may stand as a reasonable candidate for drug repurposing.
NFT has poor water solubility and low oral bioavailability, which limits its pharmacological use. [23] These drawbacks may be overcome by anchoring NFT with hydrolytically and enzymatically stable hydrosoluble groups.
Munsimbwe et al. [24] previously investigated the trypanocidal activity of NFT and its N-alkyl derivatives. Analogues 1 and 2 ( Figure 2) with IC 50 ˂ 0.34 µM were uncovered as trypanocidal hits against a variety of Trypanosoma strains. However, these hits appeared insoluble in the in vivo testing media which resulted in poor to no activity in the mice study. [24] Zuma et al., [25] conducted another study wherein the antileishmanial activity of novel NFTtriazole hybrids was assessed against the second major species of the kinetoplastidae family, that is, Leishmania, which is taxonomic to Trypanosoma. Propargyl intermediary analogue 3 and hydrid 4 ( Figure 2) were identified as in vitro antileishmanial leads combining with good safety and activity profiles.
These results motivated us to further explore the NFT scaffold and in this study, the antileishmanial activities of novel ethylene oxide groups appended to NFT were thus investigated. Ethylene oxide oligomers (EO) are extensively used in the pharmaceutical industry due to their good physical properties. [26] These groups have a high solubility in water and are biocompatible. They are nontoxic, [27] nonimmunogenic, and resistant to biodegradation. [28] Therefore, we hypothesise that the drug repurposing of NFT utilising these groups might improve its pharmacokinetic and physicochemical properties.
Apart from the ethylene glycols, other groups including pyrrolidone and substituted alcohols and propargyl were used. Herein, the synthesis and biological activity of these novel NFT EO are discussed.

| Chemistry
A series of novel ethylene glycol derivatives of NFT were obtained via a two-step synthetic route as depicted in Scheme 1. In the first step, monosubstituted ethylene glycols were brominated using the Hell-Volhard-Zelinsky (HVZ) halogenation as applied by Li et al. [29] at 60°C to form intermediates (1a-o) which were obtained in good to excellent yields (70%-90%). In the second step, triethylamine was used to deprotonate NFT, which was then followed in situ by a nucleophilic substitution (SN) reaction with 1a-o intermediates to afford the targeted ethylene glycol NFT derivatives (2a-o) that were obtained in poor (9%) to moderate yields (57%) after purification by F I G U R E 2 Structure of antileishmanially active synthesised nitrofurantoin analogues.

| Predicted physicochemical and pharmacokinetic properties
Lipinski's rule is an important guide in drug discovery and development. It describes the physicochemical (lipophilicity, molecular weight, hydrogen bond acceptors and donors, charge, and polar surface area) requirements for drugs to be administered orally. [30] Lipophilicity is the most important physical parameter as it affects the distribution, potency, and elimination of drugs. The calculated log P (octanol-water partition) values that determine the lipophilicity of compounds, [31] ranged from 1 to 2.3 with the values increasing with n, except for compound 2g (n = 3) for the synthesised analogues. For the phenoxy-containing analogues (2i-l), the log P values increased with decreasing electronegativity. Log P values that range from 1 to 3 have potentially good bioavailability and absorption and this makes them ideal candidates. [32] All the compounds complied with Lipinski's rule, except for 2c, 2d, 2f, 2g, 2h, and 2l which violate one rule (N OR O > 10); hence they were predicted to be druglike since their physicochemical properties were within the required range. Furthermore, all the compounds except for 2d, 2f, 2g, and 2l were confirmed to be potential oral compounds with PSA values of less than 140 Å 2 and less than 10 rotatable bonds. [33]

| Pharmacology/biology
The Leishmania parasite can exhibit two morphological forms; that is, the amastigote in macrophages of the mammalian host and the promastigote in the gut of sand fly vectors. [34] Studies have indicated that while some antileishmanial drugs are active against amastigotes they may not be active against promastigotes [35] and only 4% of distinguished antipromastigote hits are affirmed as antiamastigote hits in screening. [36] Furthermore, the utilisation of axenic or intracellular amastigotes also produces different screening outcomes, with axenic anti-amastigote hits correlating with antipromastigote hits rather than with intracellular amastigote assays. [35] Due to this correlation between axenic amastigote and antipromastigote hits and the high growth rate and cost effectiveness of promastigote cultures, the synthesised analogues were screened against the promastigote and intracellular amastigote forms for a potential indication of axenic and intracellular antiparasitic activity, respectively. The antipromastigote results may provide parasite form-specific activity, but may also have the potential to serve as a warning sign for potential host cell interference, for example, cell membrane permeability or metabolism preventing accessibility to the parasite, indicating a potential drug design challenge.
The synthesised NFT analogues were initially screened at 10 μM for potential antipromastigote and anti-amastigote activities against the L. major strain IR-173 and antimony-resistant L. donovani strain 9515.
[37] L. major and L. donovani are known to cause CL and VL, respectively, and in this study, they were used to determine the specificity of the synthesised compounds towards parasite stages and species. Furthermore, for the intracellular amastigote assay, a modified method of Jain et al. [38] on parasite rescue and transformation was utilised to focus on the detection of cidal compounds. The  Table 1.
The parent compound NFT was included for comparison, and it showed very poor inhibition (0%-34%) against both strains. The analogues were distinctively more active against the promastigotes than they were against the amastigotes with the growth inhibition ranging from 38% to 81% against the former while a poor 0%-39% range was observed against the latter, regardless of the strain. It could be noticed that the inhibition decreases with the increasing chain length for the same alkyl terminal R group as seen from 2b-d (R = methyl) and 2e-g (R = ethyl). Furthermore, the growth inhibition of the methylated derivatives against the L. donovani strain was higher than those of their ethylated counterparts (2e-g) whereas the inverse was observed against L. major strain. The pyrrolidone-ring-containing analogue (2h) also displayed poor growth inhibition against both strains. On contrary, the arylated analogues (2i-l) demonstrated good to excellent inhibition against the promastigote form of both strains. Among these, the halogenated analogues (2j and 2k) showed the highest inhibitory potential of ca. 80% while the 4-nitrophenoxy containing analogue (2l) possessed the least growth inhibition regardless of the strain. Allyl (2m) and propargyl (2n) derivatives showed good inhibition against both strains and last, the 3-bromopropyl analogue (2o) also had excellent growth inhibition against both strains.
All the synthesised compounds had very poor growth inhibition against the amastigote form of both strains, indicating that these analogues were not cidal in nature; however, they may potentially have static antileishmanial activity. Previously, Zuma et al. [25] screened compound 4 (analogue 2n in this study) for antileishmanial activity. A modified method of Njanpa et al. [39] was used, which is similar to the NDLOVU ET AL. | 5 of 13 assay of Jain et al. [38] except the plates were incubated for only 24 h after macrophage lysis to measure amastigote viability instead of amastigote recovery and transformation. Compound 4 was identified as a potential antileishmanial lead, with an IC 50 of 0.39 µM against intracellular L. donovani amastigotes (strain 9515). However, in this study, this activity was lost, with the 72 h amastigote recovery period postlysis resulting in the reduced growth inhibition of 2n to only 37%.
This loss of activity between the two assays with 24 and 72 h postlysis incubation periods, respectively, may be indicative of 2n having static antileishmanial activity, allowing the amastigotes to recover from treatment and transform into fast proliferating promastigotes.
According to Siqueira-Neto et al., [36] compounds with antipromastigote growth inhibition >70% qualify for IC 50 determination, whereas those with anti-amastigote growth inhibition >60% qualify for IC 50 determination. [35] Based on these criteria, only seven compounds (2b, 2i, 2j, 2k, 2m, 2n, and 2o) qualified for antipromastigote IC 50 screening, whereas none of the analogues qualified for further screening against the intracellular amastigotes. The antipromastigote data, cytotoxicity, and selectivity indexes of the screened compounds are provided in Table 2.
Apart from 2o that presented with weak toxicity at IC 50  Since the analogues were more active against the promastigotes, their IC 50 values of activity against this developmental form were used to assess the potential structure-activity relationship (SAR) within the series. The data in Tables 1 and 3 present evidence that the arylated ethylene glycol derivatives (2i-k) have good to excellent activity against the promastigotes of both strains. Furthermore, the presence of an electron-withdrawing group (EWG) such as Br and Cl on the phenyl ring in analogues 2j and 2k, respectively, was shown to increase the potency. However, the phenoxy derivative with the strongest EWG, NO 2 (2l) had a lower growth inhibition, which T A B L E 1 Growth inhibition (%) of analogues against Leishmania donovani and Leishmania major promastigotes and intramacrophage amastigotes. countries. The application of these criteria eliminated the analogues as potential antileishmanial hits/leads. However, when these criteria were extended to promastigotes, all seven selected analogues, that is, 2b, 2i, 2j, 2k, 2m, 2n, and 2o, were identified as antipromastigote hits (IC 50 ˂ 10 µM and SI > 10) with analogues 2j and 2k qualifying further as potential antipromastigote leads (IC 50 ˂ 1 µM and SI > 100). Hence, the propensity of these analogues for the promastigote form and their inactivity against the amastigotes corroborate the early findings of Siqueira-Neto et al. [36] The difference in susceptibility of the parasitic forms to the analogues may be attributed to the choice of host cell line. [45] Indeed, the origin of the macrophages used has been shown to affect drug activity against both Trypanosoma cruzi and L. donovani. Furthermore, the THP-1 cell line has been shown to be relatively tolerant to drug exposure due to the inherent metabolic status and/or hostile cytoplasmic environment of differentiated THP-1 cells, which may affect drug pharmacodynamics and thus affect antiparasitic efficacy. [45] Reduced accessibility to the parasite due to host cell membrane permeability may also be deemed as a possible reason for this low activity. [46] Indeed, in promastigote cultures, the parasites are directly exposed to treatment, whereas the amastigote is encased within the host macrophage THP-1 cell. Treatments must, accordingly, be able to enter and accumulate within the host cell to impact the parasite. [46,47] Moreover, intramacrophage amastigotes can utilise natural defense mechanisms that alter host cell permeability. For example, parasite-mediated alterations to the host cell efflux transporters, multidrug resistance-associated protein 1 (MRP1), and permeability glycoprotein (P-gp), result in the non-accumulation of compounds within the host cells, a known mechanism of antimony resistance in Leishmania. [46,48] Furthermore, the in vitro cytotoxicity of the analogues was evaluated on Vero cells, and all six compounds showed no toxicity against mammalian cells with IC 50 > 100 µM, except 2o that it found to be moderately toxic with IC 50 < 100 µM. T A B L E 2 Antipromastigote data and cytotoxicity (μM ± Standard deviation) of compounds against Leishmania donovani and Leishmania major promastigotes.
m According to Teague [43] : 250 < MW < 350, LogP o/w < 3.5 and RB < 7. All values in this table were calculated using SwissADME web tool, http://www. swissadme.ch. [44] water (10 ml) was added to quench the reaction, which was then extracted with DCM (3 × 20 ml), washed with sodium hydrogen carbonate (1 × 20 ml), and then water (2 × 20 ml). The organic phase was dried over anhydrous magnesium sulphate, filtered, and concentrated which resulted in a clear oily crude. This crude was used in the next step without further purification.

| Cytotoxicity assay
The resazurin assay was used to assess the cytotoxicity of the synthesised compounds on Vero (African green monkey kidney) cells. In this assay, viable cells irreversibly reduce oxidised blue resazurin dye into highly fluorescent pink resorufin via an enzymatic reaction. [49] Hyclone Dulbecco's modified Eagle's medium (DMEM) with 4.5 g/L glucose (HyClone) supplemented with 1% non-essential amino acids, L-glutamine, and streptomycin-penicillin, as well as 10% foetal serum from bovine (Themofisher Scientific) was used to culture the Vero cells. The cells were kept in a humid environment with 5% CO 2 at 37°C. The compound is dissolved in 100% DMSO at 50 mM stock concentration, then diluted with growth medium to 1% DMSO during the serial dilutions.
To perform the resazurin assay, 96 well plates containing 100 μl of cell suspension (60,000 cells/ml) were prepared and incubated for 24 h. The cells were then exposed to the following:

| Antipromastigote assay
The synthesised compounds were evaluated for antipromastigote activity using two Leishmania strains and the resazurin assay.
A modified approach of Siqueira-Neto et al. [36] and [50] was used.  Compounds with more than 70% growth inhibition are eligible for additional IC 50 testing. [36] Cell viability was assessed to determine the IC 50 as indicated in the cell viability assay.

| Intramacrophage antileishmanial assay
The two Leishmania strains utilised for the antipromastigote assay were also used to assess the efficacy of the synthesised compounds against intramacrophage parasites using an altered method of Njanpa et al. [39] and Jain et al. [38] Human monocytic leukaemia (THP1) cells (Cellonex, South Africa) were grown in RPMI 1640 medium that contained L-glutamine (Sigma Aldrich) with 1% streptomycin/penicillin solution (Lonza), 2 g/L sodium bicarbonate, and 10% FBS (Thermofisher Scientific). The cells were maintained in 5% CO 2 at 37°C in a humidified environment. To perform the intramacrophage assay, 25 ng/ml phorbol 12myristate 13-acetate (PMA) was added to a cell suspension of 500,000 cells/ml. The suspension was seeded at 200 μl/well in 96 well plates (Nunc, Thermofisher Scientific). The medium blank did not contain cells. The plates were incubated in 5% CO 2 at 37°C for 48 h in a humid atmosphere to allow the suspended cells to differentiate into adherent macrophages.
After the incubation, the non-adherent cells were carefully removed from the plates using warmed PBS and a preprepared respectively, were determined as indicated in the antipromastigote assay. Compounds with more than 60% growth inhibition are deemed eligible for additional IC 50 testing. [36] However, no compounds qualified for IC 50 determination in this study, and no further screening was accordingly performed.

| Statistical analysis
Nonlinear regression analysis was used to calculate the IC 50 values of the cytotoxicity and in vitro antileishmanial assays. The mean ± the standard deviation (SD) of the triplicate experiments performed for each assay was used to represent the results. Prism V5 software (GraphPad)) and Skanlt 4.0 Research Edition software (Thermofisher Scientific) were used to perform statistical analysis. The reported data was significant at p < 0.05.