PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of aacPermissionsJournals.ASM.orgJournalAAC ArticleJournal InfoAuthorsReviewers
 
Antimicrob Agents Chemother. 2010 May; 54(5): 1712–1719.
Published online 2010 February 16. doi:  10.1128/AAC.01634-09
PMCID: PMC2863604

Inhibitors of Leishmania GDP-Mannose Pyrophosphorylase Identified by High-Throughput Screening of Small-Molecule Chemical Library [down-pointing small open triangle]

Abstract

The current treatment for leishmaniasis is based on chemotherapy, which relies on a handful of drugs with serious limitations, such as high cost, toxicity, and a lack of efficacy in regions of endemicity. Therefore, the development of new, effective, and affordable antileishmanial drugs is a global health priority. Leishmania synthesizes a range of mannose-rich glycoconjugates that are essential for parasite virulence and survival. A prerequisite for glycoconjugate biosynthesis is the conversion of monosaccharides to the activated mannose donor, GDP-mannose, the product of a reaction catalyzed by GDP-mannose pyrophosphorylase (GDP-MP). The deletion of the gene encoding GDP-MP in Leishmania led to a total loss of virulence, indicating that the enzyme is an ideal drug target. We developed a phosphate sensor-based high-throughput screening assay to quantify the activity of GDP-MP and screened a library containing ~80,000 lead-like compounds for GDP-MP inhibitors. On the basis of their GDP-MP inhibitory properties and chemical structures, the activities of 20 compounds which were not toxic to mammalian cells were tested against ex vivo amastigotes and in macrophage amastigote assays. The most potent compound identified in the primary screen (compound 3), a quinoline derivative, demonstrated dose-dependent activity in both assays (50% inhibitory concentration = 21.9 μM in the macrophage assay) and was shown to be nontoxic to human fibroblasts. In order to elucidate signs of an early structure-activity relationship (SAR) for this class of compounds, we obtained and tested analogues of compound 3 and undertook limited medicinal chemistry optimization, which included the use of a number of SAR probes of the piperazinyl aryl substituent of compound 3. We have identified novel candidate compounds for the design and synthesis of antileishmanial therapeutics.

Leishmania is a protozoan parasite that shuttles between sand fly vectors and mammalian hosts, causing a spectrum of diseases known as leishmaniases (13). Leishmaniasis is prevalent in Africa, Latin America, Asia, the Mediterranean basin, and the Middle East, with an estimated 12 million people currently being infected and a further 350 million people in 88 countries being threatened by the disease. For many years, the public health impact of leishmaniasis has been underestimated, as a substantial number of cases were never recorded. The expansion of leishmaniasis and the sharp rise in its prevalence are related to both environmental changes and the migration of nonimmune people to regions of endemicity (33). Moreover, there is an increase in the overlap between HIV infection and visceral leishmaniasis, especially in intravenous drug users in both southwestern Europe and Brazil (6, 27). The situation may be much worse in Africa and Asia, where the prevalence and rates of detection of HIV and Leishmania coinfections are still largely underestimated.

Current treatment is based on chemotherapy, which relies on a handful of drugs with serious limitations, such as high cost and toxicity, the difficult route of administration, and a lack of efficacy in areas of endemicity (7). Extensive evidence from studies with animal models indicates that solid protection can be achieved by immunization; however, to date no vaccine is available in clinical practice. Therefore, there is an urgent need to develop new, effective, and affordable antileishmanial therapeutics in order to control different forms of the disease.

Leishmania synthesizes a range of mannose-rich glycoconjugates, which are considered essential for both parasite virulence and survival. A prerequisite for glycoconjugate biosynthesis in Leishmania, as in all eukaryotes, is the conversion of monosaccharides to activated sugar nucleotides. This process is catalyzed by several enzymes, including phosphomannose isomerase (PMI), phosphomannomutase (PMM), GDP-mannose pyrophosphorylase (GDP-MP), and dolicholphosphate-mannose synthase (DPMS) (9). The consecutive action of PMM and GDP-MP transforms mannose 6-phosphate (M-6-P) to GDP-mannose (GDP-Man), which acts as the mannose donor in the synthesis of molecules such as lipophosphoglycan (LPG), proteophosphoglycan (PPG), glycoinositolphospholipids (GIPLs), and N-glycans. Gene deletion studies with Leishmania have indicated that PMM and GDP-MP constitute attractive targets for the development of novel therapeutics. Both PMM and GDP-MP mutants are unable to establish infection in mice or survive in macrophages in vitro, suggesting that these proteins are essential for the clinically relevant, intracellular form of the parasite, the amastigote. However, these enzymes are not essential for the survival of the promastigotes, the insect vector stage of the parasite (10, 11). We have shown that Leishmania GDP-MP forms a hexamer in solution, whose stability might be compromised at both low ionic strength and high pH (8). We have also reported that leishmanial PMM shows a high degree of similarity to its human isoforms, PMM1 and PMM2, suggesting that the development of parasite-selective inhibitors would not be an easy task (21). Here, we report on the development of a highly sensitive, 384-well plate-based enzyme activity assay that uses phosphate detection technology and the high-throughput screening (HTS) of ~80,000 small, lead-like molecules to identify inhibitors of leishmanial GDP-MP. Although some of these compounds demonstrated in vitro antileishmanial activity on both parasite life stages—promastigotes and amastigotes—and may therefore be considered generally cytotoxic, one class of quinoline derivatives showed promising and selective activity for the amastigote stage. This class is exemplified by compound 3, which inhibited mouse macrophage infection in a dose-dependent manner (50% inhibitory concentration [IC50] = 21.9 μM) and was nontoxic to human fibroblasts at concentrations up to 100 μM.

MATERIALS AND METHODS

Chemical library.

Our lead discovery chemical library is a collection of ~80,000 compounds purchased from commercial vendors and stored in neat dimethyl sulfoxide (DMSO) at a final compound concentration of 10 mM. The compounds represent a diverse set of molecules, as judged by Tanimoto dissimilarity analysis (Tanimoto dissimilarity T value ≤ 0.85), and although simple filters based on the Lipinski criteria were not used in the selection process, 89% of the compounds in the library are Lipinski compliant (24) and 81% conform with Oprea's criteria for “lead likeness” (26).

GDP-MP high-throughput screening assay.

A high-throughput screening assay was conducted to identify lead-like compounds that inhibited Leishmania major GDP-MP from catalyzing the conversion of M-1-P to GDP-mannose. The assay ultimately detected inorganic phosphate produced from the enzymatic cascade shown in Fig. Fig.1.1. Inorganic pyrophosphatase was added to the reaction mixture in excess, to minimize the chance of identifying inhibitors of this enzyme.

FIG. 1.
High-throughput screening enzymatic cascade. The assay ultimately detected the production of inorganic phosphate. Inorganic pyrophosphatase was added to the reaction mixture in excess, to minimize the chance of the identification of inhibitors of this ...

Recombinant L. major GDP-MP was expressed as a six-His-tagged protein in Escherichia coli BL21(DE3)pLysS. Purification of the fusion protein was performed by affinity chromatography with Talon metal affinity resin, according to the manufacturer's instructions (BD Biosciences). The purity and integrity of the protein were assessed on SDS-polyacrylamide gels stained with Coomassie blue, and enzymatic activity was assessed as described previously (8). GDP-MP (4 mg/ml) was stored in aliquots at −80°C. The aliquots were thawed on ice immediately prior to use. The assay buffer components Tween 20, albumin from chicken egg white (OVA), inorganic pyrophosphatase from baker's yeast, α-d-(+)-mannose-1-phosphate, and GTP were purchased from Sigma Aldrich. Additional buffer components Anal water, NaCl, and MgCl2·6H2O were purchased from Merck. Tris-HCl and phosphate sensor were purchased from Invitrogen. Phosphate sensor is a purified form of recombinant E. coli phosphate binding protein labeled with an N-[2-(1-maleimidyl)ethyl]-7-(diethylamino)coumarin-3-carboxamide (MDCC) fluorophore that is sensitive to changes in its environment (15). Upon binding to phosphate, the fluorescence intensity of the bright blue MDCC fluorophore increases dramatically. Phosphate binding is tight (Kd ~ 100 nM), enabling the detection of phosphate at submicromolar concentrations. In addition, as the binding is rapid, this technology enables the detection of phosphate production in real time.

The assay buffer comprised 25 mM Tris-HCl (pH 8.0), 4 mM MgCl2, 150 mM NaCl, and 0.02% (vol/vol) Tween 20; and the pH was adjusted to 7.5. On each experimental day, OVA was added to a final concentration of 0.4 mg/ml and the buffer was then filtered through a 0.45-μm-pore-size filter (Millipore).

Individual enzymatic assays were performed in a total reaction volume of 5 μl in low-volume, black, 384-well plates (Greiner). Optimization experiments determined the Km for M-1-P to be 4.6 μM and that for GTP to be 1.5 μM (Fig. (Fig.22).

FIG. 2.
Determination of the Km values. The Km for M-1-P (A) was determined to be 4.6 μM and the Km for GTP (B) was determined to be 1.5 μM, as indicated by the lines.

High-throughput screening reactions were performed with concentrations close to the Km for both substrates, with the final M-1-P concentration being 5 μM and the final GTP concentration being 1.3 μM. Initially, 2.5 μl of 10 μM M-1-P with 5 μM phosphate sensor in assay buffer was dispensed into columns 1 to 23 of a 384-well plate, and a solution lacking M-1-P (negative control) was dispensed into column 24. Subsequently, 22 nl (14% interplate coefficient of variance, determined by fluorimetric quality assurance protocols) of individual 10 mM library compounds was added to columns 1 to 22 of the 384-well plate, and 22 nl of neat DMSO vehicle was added to the remaining columns (columns 23 and 24, which comprised the positive and negative controls, respectively). Finally, 2.5 μl of 0.0652 U/ml inorganic pyrophosphatase, 4 μg/ml GDP-MP, and 2.6 μM GTP in assay buffer was added to all wells of the plate. This mixture started the enzymatic reaction in all wells of the assay plates except the negative control wells. The final library compound and DMSO concentrations were 44 μM and 0.44% (vol/vol), respectively. We determined that up to 1% (vol/vol) DMSO had no effect on the rate of the reaction. Fluorescence was determined each minute for 10 min in an Envision 2103 multilabel plate reader (Perkin Elmer) with Ex430/8 nm (excitation at 430/8 nm) and Em480/30 nm (emission at 430/30 nm) filters. The rate of phosphate production was quantified with ActivityBase (version 5.4) software and the XLfit program (ID Business Solutions, Ltd., Survey, United Kingdom). “Hit” compounds were determined on a plate-by-plate basis and were defined as those compounds that reduced the rate of phosphate production by >50% relative to the rates in both the positive and the negative control wells. The Z′ value, a measure of high-throughput assay quality that takes into consideration both the signal window and the variation within the positive and the negative controls, was determined for each plate. Tests for plates with Z′ values of <0.4 were repeated. Figure Figure33 shows the average positive control values, the average negative control values, and the calculated Z′ values for individual library compound plates.

FIG. 3.
Average control values for compound plates in chronological order of plates screened. (A) Average positive (squares) and negative (triangles) controls; (B) calculated Z′ values for individual library compound plates screened. A Z′ value ...

Initial confirmation of active compounds.

Initial confirmation experiments were performed in the same manner used for the primary screen. Primary screen hit compounds, determined as described above, were picked from a separate 10 mM DMSO solution of the library compound for use in confirmation assays.

Dose-response determination and counterscreen.

Confirmed inhibitors of GDP-MP were repurchased from the original commercial vendor. These compounds were prepared in pure DMSO to a final compound concentration of 10 mM. The compounds were titrated 1:2 in DMSO and were subsequently added to the enzymatic assay in the same manner used for the primary screen, enabling quantification of the IC50 for each compound. To determine whether the hit compounds were interfering with the phosphate sensor detection technology, a counterscreen was employed in which titrated compounds were assayed in the presence of 2.5 μM phosphate sensor spiked with 1 μM Na2HPO4 in assay buffer.

In vitro testing of Leishmania parasites.

L. major V121 parasites (MHOM/IL/67/JERICHO II; WHO Reference Centre for Leishmaniases, Jerusalem, Israel) were maintained as promastigotes at 26°C in M199 medium (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum (HI-FBS; Trace Biosciences). Amastigotes were harvested from lesions of L. major-infected hypothymic CBA nu/nu mice, as described previously (12). The amastigotes were resuspended in M199 medium and immediately used for in vitro experiments.

The CellTiter blue cell viability assay (Promega, Madison, WI) was used to screen the compounds for their antileishmanial activities, as described previously (20). The compounds were dissolved in DMSO at a 10 mM working concentration and titrated across a range of concentrations (2-fold dilutions from 100 μM to 49 nM) in culture medium. The assay was performed in duplicate in 96-well plates containing 106 parasites per well. Cell viability was assessed spectrophotometrically at 550 nm with a reference wavelength of 630 nm. Amphotericin B (Sigma) was used as a reference antileishmanial agent.

The macrophage invasion assay was performed as described previously (20). Briefly, mouse bone marrow-derived macrophages were plated on coverslips placed in 24-well plates and were allowed to adhere to the plates for 48 h. Adherent cells were exposed to promastigotes (20:1, parasite/cell ratio) for 5 h, and nonattached cells and free promastigotes were washed off, followed by the addition of medium containing the compound. The intracellular proliferation drug sensitivity assay was performed with amastigotes for macrophage invasion, as described by Vermeersch et al. (30). Infected cells (duplicate coverslips) were exposed to the compounds for 48 h and then washed and fixed with methanol. Following Giemsa staining of the cells, the coverslips were examined microscopically and the percentage of infected cells was counted.

Toxicities of candidate compounds.

Primary human foreskin fibroblasts (a gift from Christopher Tonkin, Walter+Eliza Hall Institute of Medical Research [WEHI]) were maintained in Dulbecco modified Eagle medium containing 10% HI-FBS at 37°C in 5% CO2. The CellTiter blue cell viability assay (Promega) was used to examine the toxicities of the compounds for fibroblasts.

RESULTS

High-throughput screen to identify inhibitors of GDP-MP.

The high-throughput screen of ~80,000 compounds from the WEHI Bio21 lead discovery library identified 442 compounds that inhibited the activity of GDP-MP by more than 50% at a compound concentration of 44 μM, yielding a hit rate of ~0.6%. These compounds were subsequently picked from a fresh 10 mM copy of the compound library and retested in triplicate at 44 μM. Of the 442 compounds retested, 223 demonstrated confirmed activity, yielding a hit confirmation rate of 50%.

Compound dose-response curves for the 223 primary hits were conducted by the GDP-MP assay and a counterscreen assay in which the assay mixture contained only buffer, phosphate sensor, and 1 μM Na2HPO4. Compounds that exhibited similar IC50s in both the GDP-MP and the counterscreen assays were designated to have false-positive results and were eliminated from further consideration. The most potent compounds were found to have IC50s less than 1 μM in the GDP-MP assay and no activity in the counterscreen assay.

In vitro antileishmanial activity.

Following counterscreening and validation of the original hits, we chose 20 compounds for further testing of their activities against the parasites, primarily on the basis of their inhibitory potencies against GDP-MP. We obtained 39 analogues from commercial sources to develop structure-activity relationships (SARs). The structures of the 25 compounds discussed in more detail below are given in Fig. Fig.44.

FIG. 4.
Structures of compound focus sets discussed herein. Three distinct classes can be discerned: 4-pyrizinylquinolines (A), thiadiazole-like (B), and pyrazolin-3,5-diones (C). Me, methyl; Et, ethyl.

We initially investigated whether these compounds affected the survival and proliferation of wild-type L. major promastigotes in vitro. GDP-MP-knockout parasites are viable in culture; therefore, we were not expecting to see inhibition of the growth of wild-type parasites if the compounds were specifically targeting the parasite GDP-MP. The parasites were incubated for 48 h in the presence of the test compounds, and their growth rate was compared to that of the nontreated controls. Compound 21 showed relatively potent antipromastigote activity, while three others (compounds 7, 17, and 23) displayed leishmaniostatic activity (confirmed by microscopy), whereas none of the other compounds had antipromastigote activity (data not shown).

We next examined the 20 most potent GDP-MP inhibitors for their activities against L. major lesion-derived ex vivo amastigotes, the clinically relevant form of the parasite. The amastigotes were purified from lesions of infected hypothymic CBA nu/nu mice (12) and were incubated for 24 h in the presence of different concentrations of the compounds, and their growth rates were compared to the growth rate of the nontreated controls. The majority of compounds did not display significant antileishmanial activity at the concentrations tested; however, nine compounds inhibited parasitic growth (Table (Table1).1). The most active compounds were compound 20 and compound 17, for which the in vitro IC50s were 23.7 μM and 25.1 μM, respectively. Compounds 24, 3, and 25 were characterized by in vitro IC50s of 52.6 μM, 62.3 μM, and 73.4 μM, respectively, whereas four other compounds demonstrated inhibitory activity only at 100 μM. Amphotericin B was used as a reference antileishmanial agent (IC50 = 1.5 μg/ml), and DMSO used at a range of concentrations (0.0005 to 1%) showed no adverse effect on amastigote viability (Table (Table11).

TABLE 1.
Summary of antileishmanial activities of GDP-MP inhibitors

Human primary fibroblasts were used for the determination of compound toxicity. The 10 compounds tested showed no significant effects on the fibroblasts across the range of concentrations investigated (Table (Table11).

Nine compounds with activity against Leishmania lesion-derived amastigotes were also tested for their activities in intracellular assays. Initially, promastigote-infected macrophages (macrophage invasion assay) were incubated for 48 h in the presence of test compound at 30 μM, a concentration chosen on the basis of the IC50s of the two most potent compounds in the amastigote assay. A significant reduction in the number of promastigote-infected macrophages was observed for compounds 3 (P < 0.001), 25 (P = 0.003), and 17 (P = 0.01), whereas the other compounds showed only moderate activity in this assay (Fig. (Fig.5A).5A). An amastigote invasion assay (intracellular drug sensitivity assay) yielded identical results; however, statistically significant inhibition (P < 0.001) was also observed for compounds 6 and 24 (Fig. (Fig.5D).5D). Even though they had no adverse effect on human fibroblasts, compounds 20 and 13 displayed adverse effects on mouse macrophages at 30 μM (data not shown).

FIG. 5.
Treatment of L. major-infected, mouse bone marrow-derived macrophages with GDP-MP inhibitors. (A and D) Infected macrophages were treated for 48 h with 30 μM test compounds. Data are plotted for seven compounds (two additional compounds tested ...

Compound 3 showed moderate activity against amastigotes in vitro but showed excellent activity in intracellular drug sensitivity assays. Therefore, we were particularly interested in comprehensive testing of its efficacy. This compound was further tested in the macrophage assay at 10, 20, 30, and 40 μM, together with purchased structural analogues (compounds 2 and 4), which were tested at 10, 20, and 30 μM. With the exception of compound 2 at 10 μM, treatment with all three compounds at all concentrations tested produced significant reductions in the numbers of infected macrophages (P < 0.001) and comparable activities against the ex vivo amastigotes. This reduction occurred in a dose-dependent manner for all three compounds, regardless of the parasite stage used for macrophage invasion, either promastigotes (Fig. (Fig.5B)5B) or amastigotes (Fig. (Fig.5E).5E). A third analogue, compound 5, was subsequently synthesized and examined more thoroughly for its activity against amastigotes. As shown in Table Table1,1, this compound displayed antileishmanial activity against ex vivo amastigotes (IC50 = 32.6 μM) and displayed excellent selectivity, producing no signs of toxicity for fibroblasts even at concentrations of 100 μM. Compound 5 also showed excellent activity in the macrophage invasion assay (IC50 = 17.6 μM) (Fig. (Fig.5C)5C) and intracellular proliferation assay (IC50 = 11.7 μM) (Fig. (Fig.5F).5F). The results of all tests are summarized in Table Table11.

DISCUSSION

The pentavalent antimonials, such as sodium stibogluconate and meglumine antimoniate have been recommended for use for the treatment of leishmaniasis for over 70 years. It is thus not surprising that the rate of resistance to this class of drugs is increasing, and in some areas of endemicity, their use is limited due to a lack of efficacy. The second-line drugs used in the treatment of leishmaniasis include aromatic diamidines (pentamidine) and amphotericin B, but similar to the pentavalent antimonials, these drugs are toxic and produce severe (sometimes life-threatening) side effects (22). Since the introduction of miltefosine at the beginning of this century, which is the only drug orally active against leishmania, no new antileishmanial compounds have been approved for use for the treatment of humans. Therefore, the development of safe, effective, and affordable antileishmanial chemotherapies is a critical global public health priority.

Leishmania synthesizes unique mannose-rich glycoconjugates that are essential for the survival of the Leishmania parasites in the mammalian host. In search of novel antileishmanial compounds, we have employed a novel high-throughput screen involving a small-molecule chemical library of ~80,000 drug-like molecules to identify inhibitors of leishmanial GDP-MP, an important enzyme in mannose-containing glycoconjugate biosynthesis (8). Several compounds that inhibited GDP-MP activity at submicromolar or low-micromolar IC50s were identified and were therefore worthy of further investigation. We have successfully screened the activities of compounds against other parasitic targets, with the guiding philosophy behind the selection criteria for the design of the chemical library being that lead-like simplicity rather than drug-like complexity is better for furnishing more optimizable screening hits, which are better starting points for drug development and better cover diversity space. We have commented in more detail on this elsewhere (17, 18). Lead-like chemical structures have been defined as having simple molecular structures that are chemically unreactive, that are synthetically accessible, that have drug-like properties, and that might bind only relatively weakly in the low-micromolar range to target proteins but that are also highly optimizable.

Figure Figure44 depicts the structures of the 20 most potent GDP-MP inhibitors that were tested for activity against ex vivo amastigotes, along with those of some purchased and synthesized analogues that were also investigated further. Eight of the compounds are singletons and belong to their own structural class. However, the remaining compounds populate three distinct classes: pyrazinylquinolines (class A), thiadiazole-like derivatives (class B), and pyrazoline-3,5-diones (class C). Thiadiazole-containing compounds, in particular, 1,3,4-thiadiazoles, comprise by far the largest group of compounds. Scrutiny of the structures did not yield clear SARs, nor did potent GDP-MP activity translate to good antiamastigote activity in vitro. In particular, we note that compounds that are almost identical to those discussed here and that belong to the same subclass of 1,3,4-thiadiazoles bearing a ring-fused 1,3,4-triazole have recently been reported as screening (wet and in silico) hits for p38 MAP kinase (5), factor-inhibiting HIF-1 (23), CYP17 (2), and shikimate kinase (28). We therefore view these as likely panassay interference compounds (3) and, in light of this, did not progress these compounds any further. However, we note that the 1,3,4-thiadiazole heterocycle per se is present in compounds with good oral bioavailability (14), as well as in marketed drugs (4); and it may be that specific subclasses of 1,3,4-aminothiadiazoles may be optimizable screening hits, while others, such as those fused with a 1,3,4-triazole, may be panassay interference compounds.

Of the 20 compounds (initial hits) tested for their activities against lesion-derived amastigotes, the 9 most active compounds were tested in a murine macrophage model of L. major infection. The most potent, compound 3, belongs to the quinoline class of GDP-MP inhibitors that we have identified here. As shown in Fig. 5A and D, this compound reduced macrophage infection in a dose-dependent manner by about 70 to 80% at a concentration of 30 μM. This is more efficacious than its anti-ex vivo amastigote activity might suggest (IC50 = 62.3 μM). While basic compounds such as piperazines can accumulate intracellularly in a parasitological setting (see reference 17 and references therein), it is also possible that this apparent discrepancy is simply due to the use of different assay formats. However, it is also possible that ex vivo amastigotes maintained in axenic medium underwent conversion to the promastigote form, which is not affected by the compounds tested in this study (data not shown). Previous studies suggested that about 30% of L. major lesion-derived amastigotes become flagellated while they are maintained for 24 h at 37°C and pH 7.3, conditions very similar to those used in the current study (16). Nevertheless, our data suggest (unpublished observations) that ex vivo amastigotes maintained in axenic medium have the ability to infect dendritic cells but that this ability is lost following 48 h of incubation at 37°C. This is consistent with the findings described in previous reports indicating that amastigotes, but not promastigotes, have the ability to invade dendritic cells (31). Nonetheless, the results obtained from the intracellular proliferation assay are of greater importance since they reflect the activities of compounds against the true clinically relevant form of the parasite, an amastigote proliferating in the parasitophorous vacuole.

In order to elucidate signs of early SARs for this class of compounds, we obtained two analogues of compound 3, namely, compounds 2 and 4 (the structures are given in Fig. Fig.4).4). These compounds showed activity similar to that of compound 3 in the intracellular assays and were also inhibitory in a dose-dependent manner (Fig. 5B and E). This suggests that an aryl substituent on the piperazine ring may be important for activity. In order to test this hypothesis, we made the truncated analogue (compound 1) shown in Fig. Fig.4.4. In support of the hypothesis that an aryl substituent is important, compound 1 was completely inactive, providing a useful negative SAR. Such a negative SAR is important because it supports the notion that most of the features of the original hit are essential for activity and that this activity arises from the specific engagement of the test compound with GDP-MP. On this basis, we undertook some limited medicinal chemistry optimization, which included a number of SAR probes of the piperazinyl aryl substituent. The best of these was compound 5, which bore a para-phenyl ring from the benzyl group. As shown in Table Table11 and Fig. 5C and F, this compound showed good activity against ex vivo amastigotes and in an intracellular drug sensitivity assay (IC50s = 32.6 μM and 11.7 μM, respectively) and yet was completely nontoxic to fibroblasts at concentrations as high as 100 μM. This compound has been selected for further testing and optimization. The key to optimizing this compound will be to increase its ligand-binding efficiency (19) by improving its activity without significantly increasing its molecular weight.

Two other classes of compounds showed statistically significant activity in the murine macrophage model, namely, compounds 25 and 17. These belong to the phenolic Mannich base and pyrazoline-3,5-dione classes, respectively. Recent analysis of our historical screening data has shown the compounds in these classes to be problematic in high-throughput screening, in that they may appear to be genuine hits but in fact may be selectively protein reactive: in the former class via the formation of quinone methides and in the latter class by acting as a Michael acceptor for protein nucleophiles. We have coined the acronym PAINS (panassay inteference compounds) for such compounds and caution readers that these compound classes are likely to represent poor choices for optimization. In addition, compounds 6 and 24 showed significant activity in reducing amastigote infection in macrophages. Compound 6 has been assigned to the problematic thiadiazole category (group B), but in fact, it is distinct in structure from the other compounds in this group that are fused with 1,2,4-triazole. This compound might still be of potential interest for further optimization and, interestingly, resembles the antihelmintic drug levamisole (32), which also shows putative anticancer efficacy through its immunomodulatory action (25). Compound 24 belongs to the class of cyanopyridones that we have found to be PAINS, possibly by being protein reactive and/or redox reactive. Therefore, as a candidate for further optimization, we would treat it cautiously.

In general, the compounds which showed excellent inhibitory activities in the enzyme-based assay in the primary screen were less active in antiparasite assays. Possible explanations might be inefficient uptake of the compounds by the parasites, rapid intracellular degradation, or compound exclusion due to permeability constraints. The lack of correlation between enzyme inhibition and antiparasitic activity was not unexpected and is not uncommon when assay formats for drug discovery are altered and, as such, has been reported previously (29). In addition, without biomarker information for GDP-MP inhibition, it cannot be ruled out that antiparasitic activity may arise through off-target activity. In any case, it is unlikely that the nonoptimized screening hits would exhibit potent antileishmanial activities without further iterative rounds of medicinal chemistry optimization of their enzyme inhibition activity, permeation, and solubility. Interestingly, our macrophage invasion assay, which used promastigotes for invasion over 5 h, provided outcomes equivalent to those of the intracellular proliferation assay recently labeled as a “gold standard” for in vitro testing of antileishmanial compounds (30).

Studies with parasites from which the gene for GDP-MP had been deleted validated GDP-MP as a drug target that is involved in a biosynthetic pathway essential for parasite survival in the mammalian host. Given the increasing incidence of leishmaniasis and the need for potent yet minimally toxic treatment alternatives, the identification of several GDP-MP inhibitors amenable to medicinal chemistry optimization is encouraging and warrants further investigation. All these inhibitors belong to a class of 4-pyrazin-4-yl quinolines, are drug like, and are selectively toxic toward Leishmania amastigotes.

Acknowledgments

This work was supported by an NHMRC project grant 406631 and by program grant 406601.

Footnotes

[down-pointing small open triangle]Published ahead of print on 16 February 2010.

REFERENCES

1. Reference deleted.
2. Armutlu, P., M. E. Ozdemir, S. Ozdas, I. H. Kavakli, and M. Turkay. 2009. Discovery of novel CYP17 inhibitors for the treatment of prostate cancer with structure-based design. Lett. Drug Des. Discov. 6:337-344.
3. Baell, J. B., and G. A. Holloway. 2010. New substructure filters for removal of pan assay interference compounds (PAINS) from screening libraries and for their exclusion in bioassays. J. Med. Chem. [Epub ahead of print.] doi:.10.1021/jm901137j [PubMed] [Cross Ref]
4. Bertucci, A., A. Innocenti, D. Zoccola, A. Scozzafava, S. Tambutte, and C. T. Supuran. 2009. Carbonic anhydrase inhibitors. Inhibition studies of a coral secretory isoform by sulfonamides. Bioorg. Med. Chem. 17:5054-5058. [PubMed]
5. Cheeseright, T. J., M. Holm, F. Lehmann, S. Luik, M. Gottert, J. L. Melville, and S. Laufer. 2009. Novel lead structures for p38 MAP kinase via FieldScreen virtual screening. J. Med. Chem. 52:4200-4209. [PubMed]
6. Cruz, I., J. Nieto, J. Moreno, C. Canavate, P. Desjeux, and J. Alvar. 2006. Leishmania/HIV co-infections in the second decade. Indian J. Med. Res. 123:357-388. [PubMed]
7. Davis, A. J., and L. Kedzierski. 2005. Recent advances in antileishmanial drug development. Curr. Opin. Invest. Drugs 6:163-169. [PubMed]
8. Davis, A. J., M. A. Perugini, B. J. Smith, J. D. Stewart, T. Ilg, A. N. Hodder, and E. Handman. 2004. Properties of GDP-mannose pyrophosphorylase, a critical enzyme and drug target in Leishmania mexicana. J. Biol. Chem. 279:12462-12468. [PubMed]
9. Freeze, H. H., and M. Aebi. 1999. Molecular basis of carbohydrate-deficient glycoprotein syndromes type I with normal phosphomannomutase activity. Biochim. Biophys. Acta 1455:167-178. [PubMed]
10. Garami, A., and T. Ilg. 2001. Disruption of mannose activation in Leishmania mexicana: GDP-mannose pyrophosphorylase is required for virulence, but not for viability. EMBO J. 20:3657-3666. [PubMed]
11. Garami, A., A. Mehlert, and T. Ilg. 2001. Glycosylation defects and virulence phenotypes of Leishmania mexicana phosphomannomutase and dolicholphosphate-mannose synthase gene deletion mutants. Mol. Cell. Biol. 21:8168-8183. [PMC free article] [PubMed]
12. Glaser, T. A., S. J. Wells, T. W. Spithill, J. M. Pettitt, D. C. Humphris, and A. J. Mukkada. 1990. Leishmania major and L. donovani: a method for rapid purification of amastigotes. Exp. Parasitol. 71:343-345. [PubMed]
13. Handman, E. 1999. Cell biology of Leishmania. Adv. Parasitol. 44:1-39. [PubMed]
14. Hilfiker, M. A., N. Wang, X. Hou, Z. Du, M. A. Pullen, M. Nord, R. Nagilla, H. E. Fries, C. W. Wu, A. C. Sulpizio, J. P. Jaworski, D. Morrow, R. M. Edwards, and J. Jin. 2009. Discovery of novel aminothiadiazole amides as selective EP(3) receptor antagonists. Bioorg. Med. Chem. Lett. 19:4292-4295. [PubMed]
15. Hirshberg, M., K. Henrick, L. L. Haire, N. Vasisht, M. Brune, J. E. Corrie, and M. R. Webb. 1998. Crystal structure of phosphate binding protein labeled with a coumarin fluorophore, a probe for inorganic phosphate. Biochemistry (Mosc.) 37:10381-10385. [PubMed]
16. Hodgkinson, V. H., and L. Soong. 1997. In vitro maintenance of Leishmania amastigotes directly from lesions: advantages and limitations. J. Parasitol. 83:953-956. [PubMed]
17. Holloway, G. A., J. B. Baell, A. H. Fairlamb, P. M. Novello, J. P. Parisot, J. Richardson, K. G. Watson, and I. P. Street. 2007. Discovery of 2-iminobenzimidazoles as a new class of trypanothione reductase inhibitor by high-throughput screening. Bioorg. Med. Chem. Lett. 17:1422-1427. [PMC free article] [PubMed]
18. Holloway, G. A., W. N. Charman, A. H. Fairlamb, R. Brun, M. Kaiser, E. Kostewicz, P. M. Novello, J. P. Parisot, J. Richardson, I. P. Street, K. G. Watson, and J. B. Baell. 2009. Trypanothione reductase high-throughput screening campaign identifies novel classes of inhibitors with antiparasitic activity. Antimicrob. Agents Chemother. 53:2824-2833. [PMC free article] [PubMed]
19. Hopkins, A. L., C. R. Groom, and A. Alex. 2004. Ligand efficiency: a useful metric for lead selection. Drug Discov. Today 9:430-431. [PubMed]
20. Kedzierski, L., J. M. Curtis, M. Kaminska, J. Jodynis-Liebert, and M. Murias. 2007. In vitro antileishmanial activity of resveratrol and its hydroxylated analogues against Leishmania major promastigotes and amastigotes. Parasitol. Res. 102:91-97. [PubMed]
21. Kedzierski, L., R. L. Malby, B. J. Smith, M. A. Perugini, A. N. Hodder, T. Ilg, P. M. Colman, and E. Handman. 2006. Structure of Leishmania mexicana phosphomannomutase highlights similarities with human isoforms. J. Mol. Biol. 363:215-227. [PubMed]
22. Kedzierski, L., A. Sakthianandeswaren, J. M. Curtis, P. C. Andrews, P. C. Junk, and K. Kedzierska. 2009. Leishmaniasis: current treatment and prospects for new drugs and vaccines. Curr. Med. Chem. 16:599-614. [PubMed]
23. Ko, S., M. K. Lee, D. Shin, and H. Park. 2009. Structure-based virtual screening approach to the discovery of novel inhibitors of factor-inhibiting HIF-1: identification of new chelating groups for the active-site ferrous ion. Bioorg. Med. Chem. 17:7769-7774. [PubMed]
24. Lipinski, C. A., F. Lombardo, B. W. Dominy, and P. J. Feeney. 2001. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv. Drug Deliv. Rev. 46:3-26. [PubMed]
25. Mitchell, M. S. 2003. Combinations of anticancer drugs and immunotherapy. Cancer Immunol. Immunother. 52:686-692. [PubMed]
26. Oprea, T. I., A. M. Davis, S. J. Teague, and P. D. Leeson. 2001. Is there a difference between leads and drugs? A historical perspective. J. Chem. Infect. Comput. Sci. 41:1308-1315. [PubMed]
27. Rabello, A., M. Orsini, and J. Disch. 2003. Leishmania/HIV co-infection in Brazil: an appraisal. Ann. Trop. Med. Parasitol. 97(Suppl. 1):17-28. [PubMed]
28. Segura-Cabrera, A., and M. A. Rodriguez-Perez. 2008. Structure-based prediction of Mycobacterium tuberculosis shikimate kinase inhibitors by high-throughput virtual screening. Bioorg. Med. Chem. Lett. 18:3152-3157. [PubMed]
29. St. George, S., J. V. Bishop, R. G. Titus, and C. P. Selitrennikoff. 2006. Novel compounds active against Leishmania major. Antimicrob. Agents Chemother. 50:474-479. [PMC free article] [PubMed]
30. Vermeersch, M., R. Inocencio da Luz, K. Tote, J. P. Timmermans, P. Cos, and L. Maes. 2009. In vitro susceptibilities of Leishmania donovani promastigote and amastigote stages to antileishmanial reference drugs: practical relevance of stage-specific differences. Antimicrob. Agents Chemother. 53:3855-3859. [PMC free article] [PubMed]
31. von Stebut, E., Y. Belkaid, T. Jakob, D. L. Sacks, and M. C. Udey. 1998. Uptake of Leishmania major amastigotes results in activation and interleukin 12 release from murine skin-derived dendritic cells: implications for the initiation of anti-Leishmania immunity. J. Exp. Med. 188:1547-1552. [PMC free article] [PubMed]
32. Weikert, R. J., S. Bingham, M. A. Emanuel, E. B. Fraser-Smith, D. G. Loughhead, P. H. Nelson, and A. L. Poulton. 1991. Synthesis and anthelmintic activity of 3′-benzoylurea derivatives of 6-phenyl-2,3,5,6-tetrahydroimidazo[2,1-b]thiazole. J. Med. Chem. 34:1630-1633. [PubMed]
33. World Health Organization. 2008. Initiative for vaccine research—leishmaniasis. World Health Organization, Geneva, Switzerland. http://www.who.int/vaccine_research/diseases/soa_parasitic/en/index3.html.

Articles from Antimicrobial Agents and Chemotherapy are provided here courtesy of American Society for Microbiology (ASM)