Development of a New Antileishmanial Aziridine-2,3-Dicarboxylate-Based Inhibitor with High Selectivity for Parasite Cysteine Proteases

Abstract

Leishmaniasis is one of the major neglected tropical diseases of the world. Druggable targets are the parasite cysteine proteases (CPs) of clan CA, family C1 (CAC1). In previous studies, we identified two peptidomimetic compounds, the aziridine-2,3-dicarboxylate compounds 13b and 13e, in a series of inhibitors of the cathepsin L (CL) subfamily of the papain clan CAC1. Both displayed antileishmanial activity in vitro while not showing cytotoxicity against host cells. In further investigations, the mode of action was characterized in Leishmania major. It was demonstrated that aziridines 13b and 13e mainly inhibited the parasitic cathepsin B (CB)-like CPC enzyme and, additionally, mammalian CL. Although these compounds induced cell death of Leishmania promastigotes and amastigotes in vitro, the induction of a proleishmanial T helper type 2 (Th2) response caused by host CL inhibition was observed in vivo. Therefore, we describe here the synthesis of a new library of more selective peptidomimetic aziridine-2,3-dicarboxylates discriminating between host and parasite CPs. The new compounds are based on 13b and 13e as lead structures. One of the most promising compounds of this series is compound s9, showing selective inhibition of the parasite CPs LmaCatB (a CB-like enzyme of L. major; also named L. major CPC) and LmCPB2.8 (a CL-like enzyme of Leishmania mexicana) while not affecting mammalian CL and CB. It displayed excellent leishmanicidal activities against L. major promastigotes (50% inhibitory concentration [IC₅₀] = 37.4 μM) and amastigotes (IC₅₀ = 2.3 μM). In summary, we demonstrate a new selective aziridine-2,3-dicarboxylate, compound s9, which might be a good candidate for future in vivo studies.

INTRODUCTION

Leishmaniasis is one of the 17 neglected tropical diseases (NTDs) assigned by the World Health Organization (WHO). NTDs affect 1 billion people worldwide (1). The primary occurrences are in low-income countries in sub-Saharan Africa, Asia, and Latin America, but the Mediterranean countries of Europe are also concerned (2). Among the NTDs is the group of “most neglected diseases,” affecting the poorest, mainly rural areas, including leishmaniases, sleeping sickness (African trypanosomiasis), and Chagas' disease (3). These three NTDs have the highest rates of death. However, the NTD drug discovery pipeline is almost empty, thus leading to a lack of efficient and safe drugs (2, 4). Because of climate warming and tourism, the occurrence of leishmaniasis is also reported in states around the Mediterranean Sea (1).

Leishmaniasis is caused by more than 20 species of protozoan parasites belonging to the genus Leishmania. The parasite life cycle is characterized by two morphological stages: extracellular flagellated promastigotes, occurring in the insect vector, and intracellular aflagellated amastigotes, occurring in the mammalian host. The promastigotes are transmitted by an insect bite into the skin of the host, where they are internalized by macrophages, dendritic cells, neutrophils, and fibroblasts and differentiate into amastigotes residing and replicating in parasitophorous vacuoles of these phagocytes. The parasites disseminate through the lymphatic and vascular systems. During the blood meal of an (uninfected) sand fly, amastigotes are transmitted back from the infected mammalian host to the insect vector and differentiate again into promastigotes (5, 6).

The clinical outcome of leishmaniasis depends on the complex interactions between the virulence characteristics of the infecting species and the type of immune response of the host. There are three clinical forms: cutaneous, mucocutaneous, and visceral leishmaniases (6).

Concerning the treatment of leishmaniasis, it is obvious that new drugs must circumvent the limitations of currently established chemotherapies, i.e., toxicity, long courses of treatment, the frequent need for parenteral administration, high costs in countries where the disease is endemic, and the emergence of resistance. Therefore, it is important not only to test and apply combinations of existing drugs to avoid resistance but also to develop new potential leishmanicidal compounds with alternative mechanisms, as well as vaccination strategies (7, 8).

Cysteine proteases (CPs) of parasites such as Plasmodium, Trypanosoma, and worms are druggable targets for developing a new promising strategy for chemotherapy based on protease inhibition (9–12). Therefore, the identification and synthesis of highly selective protease inhibitors might be a promising means for the treatment of such infections in future. In recent years, we have been working on the development of inhibitors of papain-like CPs belonging to the CAC1 family (13–18). These proteases may represent attractive targets because of their key roles in parasite infections (9–12). The Leishmania major genome encodes a total of 65 CPs, grouped into 4 clans [CA, CD, CF, and PC(C)] and 13 families. Leishmanial CPs belonging to the CAC1 family are the lysosomal cathepsin L (CL)-like enzymes CPA and CPB, as well as the cathepsin B (CB)-like enzyme CPC (19). They share some homology with the related mammalian enzymes; however, their substrate specificities are different. While human CB accepts an Arg at the P2 position (benzyloxycarbonyl-Arg-Arg-7-amino-4-methylcoumarine [Cbz-Arg-Arg-AMC] is a quite good substrate), the leishmanial homolog CPC does not, preferring Phe in that position, comparable to human and leishmanial CL enzymes (20). In the case of Leishmania CPs, it was shown that these enzymes are essential for parasite growth, differentiation, pathogenicity, and virulence (19, 21, 22). However, the extent to which the additional inhibition of related host cathepsins may have an anti-infective effect or, in contrast, may even support the infection is not yet fully understood (23–25). Therefore, it is necessary to develop inhibitors selective for Leishmania cysteine proteases.

In previous studies, we identified two peptidomimetic aziridine-2,3-dicarboxylate-based inhibitors, Boc-(S)-Leu-(R)-Pro-(S,S)-Azi(OBn)₂ (compound 13b) and Boc-(R)-Leu-(S)-Pro-(S,S)-Azi(OBn)₂ (compound 13e), exerting excellent antileishmanial activities, in a series of inhibitors of CL and CL-like CPs (15, 16, 26, 27). Both aziridines targeted the leishmanial CB-like enzyme LmaCatB (L. major CPC), as documented with a biotin-tagged derivative of 13b (27). The inhibitor compound 13b induced an accumulation of undigested debris in autophagy-related lysosome-like vacuoles in L. major, followed by parasite cell death (27). An in vivo experiment was carried out using the BALB/c mouse model of L. major infection. After application of compound 13b, a weak exacerbation of the infection was observed; this was characterized by a significantly increased secretion of the Th2 cell cytokine interleukin 4 by murine splenic cells. This effect was probably caused by inhibition of murine CL (data not shown). This is in accordance with studies by the Katunuma group indicating that inhibition of human CL results in the potentiation of Th2-type immune responses and thus leads to an exacerbation of inflammation (23–25). These studies also showed that CB-specific inhibitors can switch T-cell development from Th2- to Th1-type immune responses in mice, resulting in an amelioration of infection. In summary, there is an urgent need for inhibitors which selectively inhibit the CL-like parasite CPs and do not affect the mammalian equivalents.

There is no X-ray structure available for leishmanial papain-like CPs, making the development of selective inhibitors a matter of “trial and error” by synthesis and testing of a broad variety of related inhibitors. Therefore, we extended our study by synthesizing a series of aziridine-2,3-dicarboxylates based on compounds 13b and 13e as lead structures. This series comprises structural isomers (s11 to s14), derivatives with ethyl ester moieties (s1 to s8), a derivative with an extended peptide chain (s15), and derivatives with nonproteinogenic amino acids within the peptide sequence in order to improve hydrolytic stability (β-Ala in s21, α-aminoisobutyric acid [Aib] in s22, and norvaline [Nva], norleucine [Nle], cyclohexylglycine [Chg], cyclohexylalanine [Cha], and phenylglycine [Phg] in s26 to s30 and s32). The influence of the configuration of the three-membered aziridine ring (R,R or S,S) on affinity and selectivity was investigated for most of the structural isomers (s16 to s19) and for the lead compounds 13b and 13e (s9 and s10). Additionally, the Leu residue in 13b was replaced by other neutral amino acids (Gly in s20, Ala in s23, Val in s24, Ile in s25, Phe in s31, and Trp in s33). On the other hand, the Pro residue in 13b was replaced by the amino acids Orn in s34, (NO₂)Arg in s35, and nipecotic acid (Nip) in s38, with the latter containing brominated benzyl esters at the aziridine ring in order to improve the properties for X-ray diffraction studies of enzyme-inhibitor complexes.

The compounds were tested against mammalian CL and CB, the recombinantly expressed CB-like protease LmaCatB (L. major CPC), and a recombinantly expressed CL-like protease from Leishmania mexicana (LmCPB2.8).

Furthermore, selected compounds were tested for the ability to inhibit proteolytic activity in L. major promastigote lysates. This was done with the compounds alone and in combination with the standard cysteine protease inhibitors E64 and CA074 in order to evaluate the extent to which the proteolytic activity is further decreased by the addition of aziridine-based cysteine protease inhibitors.

The most promising compounds were analyzed for the ability to inhibit the growth and viability of L. major promastigotes and amastigotes in vitro and for cytotoxicity against the macrophage cell line J774.1.

MATERIALS AND METHODS

Syntheses.

Synthesis of the potential inhibitors was performed as depicted in Fig. 1. The preparation was carried out through fragment coupling of Boc-protected dipeptides or amino acids to the trans-configured aziridine-2,3-dicarboxylates.

FIG 1.

General synthesis of N-acylated trans-aziridine-2,3-dicarboxylates. Reagents and conditions are as follows: i, PPA, ethyl acetate, or dimethylformamide, 1 h at 0°C and 24 h at room temperature (s1 to s37); and ii, PPA, triethylamine, or ethyl acetate, 1 h at 0°C, 4 h at 40°C, and 7 days at room temperature (s38).

Diethyl and dibenzyl aziridine-2,3-dicarboxylates in either the (S,S) or (R,R) configuration were prepared stereoselectively as described before (28). In the same manner, (2S,3S)-bis(4-bromobenzyl) aziridine-2,3-dicarboxylate was synthesized as a building block for s38, using (4-bromophenyl)methanol and l-(+)-tartaric acid as the starting materials. Dipeptides for fragment coupling were synthesized by using standard peptide coupling procedures (29). Racemic Nip was synthesized as a building block for compounds s36, s37, and s38 by microwave-assisted hydrogenation of nicotinic acid (30). N-Acylation of the aziridines with Boc-protected fragments or dipeptides was accomplished via propylphosphonic anhydride (PPA) as a coupling reagent (31). General structures of compounds s1 to s37 and the structure of compound s38 are presented in Fig. 2.

FIG 2.

Structures of the synthesized N-acylated trans-aziridine-2,3-dicarboxylates s1 to s37 and structure of the dibromo derivative s38.

Table 1 summarizes all synthesized N-acylated aziridine-2,3-dicarboxylates. Detailed analytical data can be found in the supplemental material.

TABLE 1.

Inhibition of CL, CB, LmCPB2.8, and LmaCatB, antileishmanial activity against L. major promastigotes, and cytotoxicity by trans-aziridine-2,3-dicarboxylates 13b, 13e, and s1 to s38^c

^aDiastereomeric ratio (dr) = 1:0.59.

^bDiastereomeric ratio (dr) = 1:0.55.

^cBn, benzyl ester; Et, ethyl ester; pBr-Bn, para-bromo-benzylester; NI, no inhibition; ND, not determined; ND^p, not determined because of precipitation; J774.1, macrophage cell line; Nip, nipecotic acid; CB, cathepsin B; CL, cathepsin L.

Cloning and site-directed mutagenesis of LmaCatB (L. major CPC).

The coding sequence of LmaCatB was cloned from extracted genomic DNA of L. major (clinical isolate MHOM/IL/81/FE/BNI) into the vector pGAPZαA (Invitrogen, Darmstadt, Germany) for expression in the yeast Pichia pastoris. LmaCatB was amplified by PCR as a proform according to the method of Chan et al. (32), with the sense primer 5′-AGAGAGGCTGAAGCTAAGCCGAGTGACTTTCCGCTTC-3′ and the antisense primer 5′-ATGATGGTCGACGGCCTCCTGCGCGGGTATGCCAG-3′ for expression with a hexahistidine tag and the primer 5′-ATGATGGTCGACGGCCTACTCCTGCGCGGGTATGCCAG-3′ with a stop codon for expression without the tag. The purified PCR product was cloned into pGAPZαA by sequence- and ligation-independent cloning (33), and the resulting construct was used for transformation of Escherichia coli XL1-Blue cells. Transformants were selected on LB agar plates containing 25 μg/ml Zeocin (InvivoGen, Toulouse, France) and verified using colony PCR. Plasmids isolated from individual clones were sequenced in both directions (Seqlab, Goettingen, Germany) with pGAP forward and 3′AOX1 primers. A sufficient amount of plasmid was linearized for transformation of P. pastoris by digestion for 2 h at 37°C with the restriction enzyme AvrII. P. pastoris X-33 cells were transformed with the linearized plasmids by electroporation (Gene Pulser MXcell; Bio-Rad, Munich, Germany) at 1,500 V, 400 Ω, and 25 μF. P. pastoris colonies were selected on yeast extract-peptone-dextrose (YPD) agar plates containing 100 to 200 μg/ml Zeocin. Colony PCR did not produce reliable results with P. pastoris cells; therefore, genomic DNA was extracted for verification of proper integration of the construct into the P. pastoris genome.

Expression, purification, and activation of LmaCatB.

Recombinant P. pastoris clones were screened for expression in small-scale cultures (5 ml YPD) after 24 h, 48 h, and 72 h at 30°C. Genes under the control of the GAP promoter of pGAPZαA are transcribed constitutively, and the expressed proteins are secreted into the medium. The expressed proteins were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting using murine anti-His antibodies. After 72 h of expression, the supernatant from each culture was harvested by centrifugation at 5,000 × g for 15 min, followed by vacuum filtration through a glass microfiber filter (grade GF/A; Whatman [commercially available from Sigma-Aldrich]) to remove residual P. pastoris cells. The pH was adjusted to 8.0 by addition of Tris-HCl to a final concentration of 10 mM. Subsequently, the supernatant was loaded onto an XK16 column packed with Q Sepharose Fast Flow resin (GE Healthcare, Freiburg, Germany). Bound protein was eluted in a concentration gradient between buffer A (10 mM Tris-HCl, pH 8.0) and buffer B (10 mM Tris-HCl [pH 8.0],1 M NaCl). Fractions containing the recombinant protein were determined by SDS-PAGE, pooled, and concentrated by ultrafiltration in a 10-kDa-cutoff concentrator (Vivaspin 20; Sartorius AG, Goettingen, Germany). The two major bands on the gel, at 35 and 43 kDa, were confirmed as LmaCatB by electrospray ionization-liquid chromatography (ESI-LC)-mass spectrometry (LTQ Orbitrap; Thermo Scientific, Darmstadt, Germany), using the peptides from the bands after digestion with trypsin. As the final purification step, the protein was loaded onto a size-exclusion chromatography column (Superdex XK26/60; GE Healthcare) equilibrated with 20 mM sodium citrate (pH 5.0) and 250 mM NaCl. The protein-containing fractions were concentrated and then buffer exchanged into activation buffer (100 mM sodium citrate [pH 5.0], 100 mM NaCl, 10 mM dithiothreitol [DTT], 1 mM EDTA). The protein was then incubated for 24 h at 4°C to convert any remaining proform enzyme into the mature form by releasing its N-terminal propeptide. Finally, the buffer was exchanged into storage buffer (10 mM sodium citrate [pH 5.0], 1 mM DTT, and 1 mM EDTA), and aliquots of LmaCatB were flash frozen in liquid nitrogen and stored at −80°C.

Parasites.

The virulent L. major isolate (strain MHOM/IL/81/FE/BNI) was maintained by continuous passages in female BALB/c mice (permission number 55.2-2531.01-26/12 from the Government of Lower Franconia, Germany). Promastigotes were isolated from BALB/c mouse lesions and finally grown in blood agar cultures at 27°C, 5% CO₂, and 95% humidity.

Enzyme assays with recombinantly expressed Leishmania proteases and mammalian proteases.

Activity assays were carried out as described previously (27, 34, 35). LmCPB2.8 was recombinantly expressed as described previously (36). CL and CB were purchased from Calbiochem (Schwalbach, Germany). The fluorimetric substrate Cbz-Phe-Arg-AMC was purchased from Bachem (Bubendorf, Switzerland). The assay buffer for CL and CB contained 50 mM Tris (pH 6.5), 5 mM EDTA, 200 mM NaCl, and 0.005% polyoxyethyleneglycol dodecyl ether (Brij 35). The enzyme buffer for CL and CB contained 50 mM Tris (pH 6.5), 5 mM EDTA, 200 mM NaCl, and 2 mM DTT. For LmCPB2.8 and LmaCatB, the assay buffer consisted of 50 mM phosphate buffer (pH 6.5) and 5 mM EDTA. The enzyme buffer for LmCPB2.8 and LmaCatB consisted of 50 mM phosphate buffer (pH 6.5), 5 mM EDTA, and 5 mM DTT. Substrates and inhibitor stock solutions were prepared in dimethyl sulfoxide (DMSO) and diluted with assay buffer (to a final DMSO concentration of 7.5%). A Varian Cary Eclipse fluorescence spectrophotometer (Varian, Darmstadt, Germany) with 96-well plates was used, with an excitation wavelength of 380 nm and an emission wavelength of 460 nm.

Fluorimetric assays for inhibition of proteolytic activity of promastigote lysates.

For preparation of promastigote lysates, stationary-phase L. major promastigotes were harvested from blood-agar plates, washed twice with phosphate-buffered saline (PBS), and pelleted by centrifugation at 3,000 × g for 10 min. Afterwards, the pelleted cells were resuspended in acidic sodium acetate buffer (pH 5.5). Finally, the promastigotes were disrupted by freezing in liquid nitrogen and thawing at 37°C three times, followed by centrifugation at 700 × g for 15 min at 4°C. The supernatant was aliquoted into fresh tubes and stored at −20°C until use. Final protein concentrations of the lysates were determined with a bicinchoninic acid protein assay kit (Pierce Biotechnology, Inc., Pittsburgh, PA). The assay buffer for L. major promastigote lysates contained 200 mM sodium acetate, 1 mM EDTA, 0.05% Brij 35, and 0.5 mM DTT. Cbz-Phe-Arg-AMC (see above) was also used as a fluorimetric substrate.

Determination of K_i values.

The hydrolysis of the substrate was monitored over 10 min in the presence of inhibitor. K_i^app values were calculated using the following equation: y = v₀/{1 + ([I]/K_i^app)^s} (2-parameter logistics), with y indicating enzyme activity (df/min increase of fluorescence over time as a result of substrate hydrolysis), v₀ indicating enzyme activity in the absence of inhibitor, [I] indicating the inhibitor concentration, and s indicating the Hill coefficient. The correction to a zero substrate concentration necessary for competitive inhibitors was done by considering substrate concentrations and the affinity of the substrate for the target enzyme (K_m values) by using the equation K_i = K_i^app/(1 + [S]/K_m) (37) with the following values: [S] of 6.25 μM and K_m of 6.5 μM for CL, [S] of 100 μM and K_m of 150 μM for CB, [S] of 10.0 μM and K_m of 5.0 μM for LmCPB2.8, and [S] of 25.0 μM and K_m of 7.0 μM for LmaCatB. GraFit software (38) was used to calculate the K_i^app values.

IC₅₀ value determination for L. major promastigotes and amastigotes.

The half-maximal inhibitory concentrations (IC₅₀s) of compounds against L. major promastigotes were determined by the alamarBlue assay as described previously (26, 39). Stationary-phase promastigotes were seeded into 96-well plates at a density of 1 × 10⁷ ml⁻¹ in RPMI medium without phenol red and with 10% fetal calf serum (FCS), in the absence or presence of increasing concentrations of compounds. Parasites were then incubated for 24 h at 27°C, 5% CO₂, and 95% humidity. Following the addition of 20 μl of ready-to-use alamarBlue solution (Trinova Biochem, Giessen, Germany) per well, the plates were incubated again and the optical densities measured after 48 h.

A recently described amastigote drug screening assay against intracellular amastigotes (39) was applied to determine the IC₅₀s of compounds against L. major amastigotes. Bone marrow-derived macrophages (BMDM) were generated and infected with luciferase-transgenic L. major promastigotes at a ratio of 1:15 as recently described (39). Compounds were added to BMDM 24 h after infection, when the differentiation of promastigotes into amastigotes was complete. Control BMDM were incubated for the same amount of time in phenol red-free RPMI medium with 10% FCS and 1% DMSO. BMDM were then incubated at 37°C, 5% CO₂, and 95% humidity for another 24 h. After cell lysis with a luciferin-containing buffer, the IC₅₀s of the compounds used against L. major amastigotes were determined by the resulting luminescence.

Promastigote staining after treatment with inhibitor s9.

Promastigotes with a cell density of 1 × 10⁸ ml⁻¹ were treated with 100 μM s9 for 180 min at 27°C. Control cultures were incubated in 0.5% DMSO-containing RPMI medium. Cells were harvested and transferred to microscopic slides by centrifugation for 5 min at 1,500 rpm, using a Cytospin 3 Shadon centrifuge (Thermo Electron Corporation, Waltham, MA). Parasites were fixed and stained using a Diff-Quik kit (Medion Diagnostics, Duedingen, Switzerland) according to the instructions in the kit's manual.

TEM of s9-treated L. major amastigotes.

BMDM were generated and infected with promastigotes at a ratio of 1:15 as recently described (39). After 24 h of coculture, complete differentiation from the extracellular promastigote stage to the intracellular amastigote stage was observed (39). Finally, amastigote-infected macrophage cultures were incubated in RPMI medium containing 0.5% DMSO for control cultures or in RPMI medium containing compound s9 at a concentration of 10 μM. After incubation for 30 min or 60 min, s9-treated and control amastigote-infected macrophages were subjected to transmission electron microscopy (TEM) as recently described (27).

RESULTS

Aziridine-2,3-dicarboxylate-based inhibitors selectively inhibited parasite CPs.

In fluorescence enzyme assays, the inhibitory effects of the potential inhibitors were evaluated against human CL and CB, the CB-like protease LmaCatB (L. major CPC), and the CL-like enzyme LmCPB2.8 from L. mexicana (Table 1).

In contrast to the lead compounds 13b and 13e, which were active against human CL in the single-digit micromolar range, most of the new compounds showed no or only weak inhibition of CL and CB (i.e., K_i > 10 μM). The exception was s35 being active against CB (K_i = 5.4 μM). In agreement with earlier studies on CPs (15), most of the compounds containing ethyl ester moieties, namely, compounds s1 to s8, did not inhibit or only weakly inhibited the enzymes. Only compounds s5 and s8 showed weak inhibition of LmaCatB (L. major CPC). The structural isomers and the stereoisomers of 13b (s9 to s14 and s16 to 19) inhibited the CL-like enzyme LmCPB2.8, and most of them also inhibited the CB-like enzyme LmaCatB. Interestingly, a higher degree of selectivity between mammalian and parasite enzymes was achieved on the one hand with the stereoisomers of 13b and 13e, namely, s9 and s10, which were R,R-configured at the aziridine ring, and on the other hand with the S,S-configured structural isomers s11 to s14. Since s9 turned out to be the most selective inhibitor concomitantly displaying antileishmanial activity against promastigotes (IC₅₀ = 37.4 μM against L. major) (Table 1), the compound was further modified by exchanging the amino acid (S)-Leu, yielding the compounds s15 and s20 to s35. Elongation of the amino acid sequence of s9 yielded the tripeptide derivative s15, which was a quite good inhibitor of the parasite proteases that maintained antileishmanial activity against L. major promastigotes (IC₅₀ = 34.2 μM) (Table 1). Of these compounds, only those with lipophilic or bulky groups showed considerably improved inhibition (for s31 [with Phe], IC₅₀ = 1.7 μM; and for s32 [with hPhe], IC₅₀ = 1.5 μM) (Table 1). Interestingly, these compounds did not inhibit the cathepsin B-like L. major enzyme CPC (LmaCatB) but only the cathepsin L-like protease LmCPB2.8. The exchange of the (R)-Pro residue in s9 against (R)-Orn(Boc) and (R)-Arg(NO₂) (s34 and s35) resulted in two strong inhibitors of the parasite protease LmaCatB, which may be explained by the preference of the enzyme for peptides with Arg at the P1 position. Compounds with a Nip residue, i.e., s36 to s38, were quite good inhibitors of LmCPB2.8, with selectivity over LmaCatB, making the brominated compound a good candidate for cocrystallization with the target enzyme.

Selective inhibitors of parasite CPs displayed highly significant antileishmanial activity in vitro.

The antiparasite activities of selected inhibitors were evaluated against L. major promastigotes (Table 1) and, for the most promising inhibitors, also against L. major amastigotes (Table 2) (27). Since previous studies showed that diethyl esters were not active in cell assays (15), probably due to poor membrane permeability, only the dibenzyl esters were tested. Cytotoxicity against host cells was determined using the macrophage cell line J774.1 (Table 1). We recently demonstrated (27) that the broad-spectrum inhibitor E-64 (40, 41), the CB-selective inhibitors CA074 (42) and CA074ME (42), and paromomycin have no or only weak effects against promastigotes. The IC₅₀s of 13b and 13e against promastigotes were comparable to those of pentamidine and miltefosine. Only amphotericin B was more effective against L. major promastigotes (27). Within the series of new dibenzyl esters, compounds s9, s15 to s19, s23 to s25, s28, and s31 showed inhibitory potency against L. major promastigotes (Table 1). The IC₅₀s are in the same range as those of 13b, 13e, pentamidine, and miltefosine (27) (Table 1).

TABLE 2.

Antileishmanial activities of trans-aziridine-2,3-dicarboxylates 13b, 13e, s9, s17, s24, and s25 and of standard inhibitors against L. major amastigotes

Compound	L. major IC50 (μM)
13b	2.2 ± 1.5
13e	2.7 ± 0.7
s9	2.3 ± 0.6
s17	1.6 ± 0.3
s24	2.2 ± 0.6
s25	2.0 ± 0.6
E-64d	39.8 ± 11.3
CLIK-148	>100
CA074ME	>100

Inhibitor s25 displayed the best inhibition of growth and viability of L. major promastigotes (IC₅₀ = 9.8 μM) (Table 1). At the concentrations used, none of the tested compounds was cytotoxic against the macrophage cell line J774.1 (Table 1). With compound s9, changes in the morphology of promastigotes were studied. Rounding of L. major promastigotes after treatment with s9 for 180 min was observed before cell death induction (see Fig. S2 in the supplemental material).

Selected compounds, namely, 13b, 13e, s9, s17, s24, and s25, together with the epoxides E-64d (the cell-permeative prodrug form of E64c, which is similarly active to E64), CLIK-148 (a CL-selective inhibitor), and CA074ME, were additionally tested for antileishmanial activity against L. major amastigotes (Table 2). All aziridine-based inhibitors displayed high antileishmanial activity, with IC₅₀s in the low micromolar range, in contrast to the epoxide-based inhibitors E-64d, CLIK-148, and CA074ME (Table 2). This is in agreement with previous results obtained with the aziridines, which showed better effects on amastigotes than on promastigotes (27). With IC₅₀s of >250 μM for s17, s24, and s25 on macrophages, the selectivity indices are excellent (SI_s17 = 156, SI_s24 = 114, and SI_s25 = 125), matching the identification criteria for hits of protozoan diseases of the WHO (43, 44).

The aziridine-2,3-dicarboxylate-based inhibitor s9 showed enzyme inhibition of L. major promastigote protein lysates similar to that by E-64.

For further evaluation, the highly selective compound s9 (Table 1) was chosen to characterize its potential to inhibit leishmanial CPs in promastigote protein lysates. With this inhibitor, fluorescence proteinase activity assays with protein lysates obtained from stationary-phase promastigotes were performed. For comparison, the standard CP inhibitors E-64, CLIK-148, and CA074, as well as the lead aziridine-based inhibitors 13b and 13e, were included. Proteinase activities were determined by proteolytic cleavage of the substrate Cbz-Phe-Arg-AMC. Protein lysates were incubated with either DMSO or the inhibitors in the first incubation step, and in the second step an incubation with DMSO followed. The residual proteolytic activity after treatment with E-64 was 3.2%, that after treatment with the CB-selective inhibitor CA074 was 20.1%, and that after treatment with the CL-selective inhibitor CLIK-148 was 8.9% (Fig. 3A). Compounds 13b and 13e provoked only moderate inhibition (residual activity after treatment, 47.0% with 13b and 61.6% with 13e) (Fig. 3A). For both inhibitors, it was demonstrated previously that they specifically reduced the activity of the CB-like enzyme CPC in protein lysates of L. major promastigotes (27). This result was confirmed in the present study with recombinantly expressed LmCPB2.8 (Table 1).

FIG 3.

Assay for proteolytic activity of promastigote protein lysates. (A and B) Protein lysates obtained from stationary-phase promastigotes were preincubated in the first incubation step (1st Inc.) with DMSO, 200 μM E-64, 200 μM CA074, 200 μM CLIK-148, 200 μM compound 13b, 200 μM compound 13e, or 200 μM compound s9. In the second incubation step (2nd Inc.), protein lysates were incubated with either DMSO, 200 μM compound 13b, 200 μM compound 13e, or 200 μM compound s9. Proteinase activities were determined by proteolytic degradation of the fluoropeptide Cbz-Phe-Arg-AMC.

Treatment with s9 resulted in a residual enzyme activity of 5.6%, which was comparable to that with E-64 (Fig. 3A). The result clearly showed that s9 caused additional inhibitory effects compared to its isomers 13b and 13e. For detailed analyses of the selectivity of the inhibitors, protein lysates were preincubated in the first incubation step with E-64 (broad-spectrum CP inhibitor; inhibition of leishmanial CPA, CPB, and CPC) or CA074 (CB-selective CP inhibitor; inhibition of leishmanial CPC) (Fig. 3B). In the second incubation step, protein lysates were incubated with DMSO, 13b, 13e, or s9. In the case of 13b and 13e, no further effect on activity was observed after preincubation with E-64 or CA074 (Fig. 3B), which clearly confirmed that only CPC was affected. However, there was a significant decrease of activity after additional incubation with s9 for preincubation with CA074. These data suggested that s9 might inhibit not only the CB-like CPC of L. major but also the CL-like CPA and/or CPB of L. major.

The CP inhibitor s9 induced an accumulation of lysosome-like vacuoles followed by cell death in amastigotes.

TEM studies were performed to analyze how cell death of L. major amastigotes was induced by s9 (see Fig. S3 in the supplemental material). We recently described that treatment with the aziridine-based inhibitor 13b resulted in cell death, characterized by an inhibition of digestion in lysosome-like vacuoles and hallmarked by an accumulation of debris in these organelles (27). Based on this fact, we expected a similar phenotype for s9-treated amastigotes. An accumulation of lysosome-like vacuoles in s9-treated amastigotes compared to the case in control macrophages was observed after 30 min and 60 min of incubation (see Fig. S3). Such lysosome-like vacuoles have been described to contain CPA, CPB, and CPC. Surprisingly, the phenotype was slightly different from that induced by 13b (27) in terms of the vacuoles, which were more numerous in s9-treated than 13b-treated amastigotes. Finally, cell death of amastigotes was observed after 60 min of treatment with s9 (see Fig. S3).

DISCUSSION

CPs of parasites are attractive targets for developing new leishmanicidal drugs. Leishmania species express the CL-like proteases CPA and CPB and the CB-like enzyme CPC. We previously identified two aziridine-2,3-dicarboxylate-based inhibitors, 13b and 13e, with antileishmanial activity (26, 27). Since inhibition of host cell CL may lead to compensation of the positive effects caused by inhibition of Leishmania cathepsins, the aim of the present study was the development of inhibitors with selectivity for Leishmania enzymes. Using compound 13b as the lead structure, a second series was synthesized and is presented in this study. The series contains structural isomers, stereoisomers, derivatives with ethyl ester moieties, and derivatives with nonproteinogenic amino acids within the peptide sequence. In most cases, the compounds of this second series showed selective inhibition of parasite CPs, while the mammalian proteases CL and CB were not affected. Since no X-ray structure of Leishmania papain-like CPs has been published so far, docking studies to identify possible binding modes and to explain the selectivity would be possible only with homology models, which is a rather uncertain method. In previous studies, we suggested possible binding modes for CL- and CB-selective aziridine-based inhibitors (15). We also performed docking studies with the related parasite enzyme cruzain from Trypanosoma cruzi (unpublished data), which in principle are in agreement with the previous findings. According to those results, aziridines consisting of at least two large, hydrophobic moieties interact with the hydrophobic S₂ and/or S₁′ binding pocket of a CL-like enzyme, while the other residues (N-terminal protecting group and a second benzyl ester) are widely solvent exposed during the binding process and do not have defined contacts with amino acids of the protein. Based on these findings, two of the three hydrophobic residues of the aziridine-2,3-dicarboxylate-based inhibitors (two benzyl esters and one hydrophobic amino acid side chain) are involved in binding, whereby the proline residue and the configuration of the aziridine ring determine the positions of these groups relative to each other. Hence, different binding modes are theoretically possible depending on the ligand and target enzymes. Interestingly, the structural isomers of 13b and 13e with an (R,R)-configured aziridine ring (s16 to s19) are not selective between mammalian and parasitic enzymes, while those with an (S,S)-configured ring (s11 to s14) are highly potent and selective inhibitors of Leishmania enzymes, especially of LmaCatB (L. major CPC). However, these compounds do not affect the growth of L. major promastigotes. In contrast, the less selective inhibitors s16 to s19 display antileishmanial activities in the same range as that of 13b, 13e, and s9. The compounds active against promastigotes are also active against the amastigotes, with even lower IC₅₀s. Fortunately, the compounds do not show cytotoxicity against host cells.

To elucidate the antileishmanial activity of s9, which combines selective inhibition of both leishmanial enzymes with good antileishmanial activity against both promastigotes and amastigotes, fluorescence protease activity assays with lysates of L. major promastigotes were carried out. Compound s9 produced stronger inhibition of the leishmanial enzyme activity than did 13b and 13e after a first incubation step with the inhibitors and a second incubation step with DMSO. After preincubation of the lysates with the broad-spectrum CP inhibitor E-64 or the CB-selective inhibitor CA074, a significant further reduction of the proteolytic activity was observed after incubation with s9, in contrast to that after incubation with 13b or 13e. This clearly demonstrates that s9 targets additional proteases compared to 13b or 13e, and also compared to E64. This may also explain why the phenotype of amastigotes after treatment with s9 is slightly different from that observed after treatment of amastigotes with 13b (27).

To sum up, the present study extends our previous knowledge about aziridine-2,3-dicarboxylate-based inhibitors with leishmanicidal activity as potential targets. We achieved exclusive selectivity of inhibition between parasite CPs and the related mammalian proteases. Furthermore, we identified a new lead structure with the highly selective inhibitor s9.

Funding Statement

The German Research Foundation (DFG) provided funding to Heidrun Moll, Tanja Schirmeister, and Uta Schurigt under grant number SFB 630. Alicia Ponte-Sucre was supported by the Alexander von Humboldt Foundation.