Cysteine Protease Inhibitors Cure an Experimental Trypanosoma cruzi Infection

J774 macrophages were cultured in RPMI-1640 medium with 5% heat- inactivated FCS (RPMI medium). For growth inhibition assays, J774 macrophages were irradiated (3,000 rad) to arrest cell growth and cultured on coverglasses within six-well plates for 24 h at 37°C. After infection with T. cruzi trypomastigotes of the Y strain for 3 h, monolayers were washed with RPMI medium and treated with inhibitors at 20 μM in RPMI medium. Inhibitor stocks were made at 20 mM in DMSO and all assays included DMSO (0.01–0.02%, vol/vol) controls. FMK inhibitors (Mu-F-hF-VSφ, Mu-F-K-VSφ, Mu-F-V-VSφ, Mu-F-S(OBzl)-VSφ, Mu-L-hF-VSφ, Mu-Yii-hF-VSφ, Boc-tic-hF-VS, Mu-tic-hF-VS, Mu-Y-hF-VS, Mu-F-hF-FMK, Mu-F-hF-VAmBzl, N-Pip-F-hF-VSφ, Z-F-A-FMK, Mu-bsu-hF-FMK, and Mu-F-hF-FMK) were provided by Prototek (Dublin, CA) and vinyl sulfones were provided by Axys Pharmaceuticals (South San Francisco, CA). CPI were evaluated in T. cruzi–infected macrophage cultures for 21– 30 d. Trypomastigote output, indicative of the completion of the intracellular cycle, was then assayed in treated and untreated cultures to determine growth inhibition of intracellular T. cruzi amastigotes (20).

After this initial inhibitor screen, T. cruzi–infected macrophages were treated with 20 μM Mu-F-hF-VSφ and Mu-F-V-VSφ for up to 76 h. Monolayers were washed, fixed with 4% paraformaldehyde, and then Giemsa stained at determined intervals. To evaluate treatment, the percentage of infected macrophages and the total number of intracellular amastigotes in 100 infected macrophages were quantified. A decrease in the number of intracellular generations indicated inhibition of intracellular growth of T. cruzi amastigotes and was calculated from the total number of intracellular amastigotes per 100 infected macrophages (n = 3).

3-wk-old female C3H mice weighing initially between 17–19 g were used in all experiments. In the first experiment (see Fig. 1), mice (five animals per lot) were infected with 10⁵ trypomastigotes of the Y strain and treated with a 1-mg i.p. injection of FMK inhibitors (Mu-F-hF-FMK, Mu-bsu-hF-FMK, and Z-F-A-FMK) twice per day. Controls included intraperitoneal injection with equal volume of DMSO. Treatment was initiated 24 h after infection and continued until death of the animals or the end of the experiment, as appropriate. Parasitemias were determined every 48 h for each animal on alternating days from 5 μl of blood extracted from the tail and diluted 1/4 (vol/vol) in RPMI medium. The numbers of parasites per milliliter, calculated in a Neubauer chamber, were expressed as a mean of two or three animals per day. The experiment was terminated 18 d after infection.

Approximate dosing regimes and different inhibitor chemistries were then analyzed (see Table 3, Experiment 1). Mice (five animals per lot) were infected (intraperitoneally) with 4 × 10⁶ tissue culture–derived trypomastigotes. 24 h after infection, mice were treated with two daily doses of 1 mg i.p. Mu-F-hF-FMK and Mu-F-hF-VSφ for up to 16 d. In a third experiment (see Table 3, Experiment 2), animals were infected with 10⁶ trypomastigotes and treated with the inhibitors with the same regimen for 12 d.

Finally (see Table 3, Experiment 3), mice were infected with 10⁵ T. cruzi trypomastigotes and treated with three daily intraperitoneal doses of N-Pip-F-hF-VSφ (2.1 mg/d) for 24 d. Blood (hemo) cultures of untreated controls (n = 6) and N-Pip-F-hF-VSφ–treated animals (n = 10) were performed 16, 22, and 46 d after infection using arrested macrophages as host cells. In brief, 6-μl aliquots of blood were resuspended in RPMI medium with 5% FCS and antibiotics. Irradiated J774 macrophages were infected with blood dilutions, and incubated for up to 30 d at 37°C in a 5% CO₂ atmosphere. Blood cultures were considered positive if T. cruzi infected macrophages and/or free trypomastigotes were observed. Hemocultures were considered negative if no infected host cells and no free trypomastigotes were observed for up to 30 d, after which macrophages died.

Six animals from Table 3, Experiment 3, survived the acute phase of the infection. Three out of six mice had consistently negative blood cultures while the remaining three out of six were positive. After being allowed to establish chronic infection for 3 mo, the latter three mice were retreated with three daily doses (2.1 mg/d i.p.) of N-Pip-F-hF-VSφ for 21 d (see Table 4, Experiment 4). Hemocultures from the six animals were performed repeatedly as described above but with a larger volume of blood sampled (6–50 μl).

T. cruzi–infected, CPI-treated mice, noninfected CPI-treated mice, and untreated T. cruzi–infected animals were necropsied to evaluate toxicity of the inhibitor regimen. Mice were weighed, examined grossly, and all major organs also examined by routine hematoxylin-eosin histopathology. The toxicity of Mu-F-hF-VSφ and N-Pip-F-hF-VSφ was further evaluated in uninfected C3H mice treated for 45 d with 4 mg/d i.p., and 12 mg/d per oral (p.o.), respectively.

Irradiated macrophages were infected with T. cruzi trypomastigotes for 2 h at 37°C. Monolayers were washed 24 h after infection and reincubated with or without the addition of the CPI Mu-F-hF-VSφ (10 μM). After 48 h, culture medium containing Bodipy FL ceramide (Molecular Probes, OR) was substituted for 15 min at 37°C. Monolayers were subsequently processed according to manufacturer's instructions and observed by fluorescence microscopy (21).

T. cruzi–infected, irradiated J774 macrophages were treated or not with Mu-F-hF-VSφ (10 μM) for 48 h. Monolayers were trypsinized, washed, fixed, and processed for electronmicroscopy. Cells were fixed with 1.5% glutaraldehyde in 0.66 M sodium cacodylate buffer, pH 7.4, at room temperature for 2 h, embedded in EPONATE 12 (Ted Pella, Inc., Redding, CA), sectioned, stained, and then observed with a Zeiss 10C electronmicroscope (Carl Zeiss Inc., Thornwood, NY; reference 22). The techniques (23) and reagents used for immunocytochemistry have been described previously for T. cruzi epimastigotes (21). Heart muscle from a mouse treated 6 d after infection with Mu-F-hF-VSφ (2 mg/d i.p. for 3 d) was fixed and processed as described above.

Irradiated J774 macrophages were infected with Y strain trypomastigotes and incubated at 37°C for 48 h to allow intracellular development of amastigotes. T. cruzi–infected macrophages, uninfected macrophage controls, and T. cruzi epimastigotes were radiolabeled with 20 μM [¹⁴C]Mu-F-hF-VSφ for 2 h at 37°C or 26°C, as appropriate. For competition experiments, duplicate cultures were treated with 20 μM Mu-F-hF-VSφ for 3 h before addition of the radiolabeled inhibitor. Infected macrophages were scraped and centrifuged at 1,700 g for 12 min at 4°C. Pellets were resuspended and lysed in a tissue grinder to release intracellular amastigotes, centrifuged at 260 g for 5 min, and the amastigote-containing supernatant centrifuged at 2,000 g for 12 min. Amastigote pellets were resuspended, transferred to Eppendorf tubes, and washed three times with PBS before sonication. Radiolabeled, noninfected macrophages and T. cruzi epimastigotes were sonicated as above. Purified recombinant cruzain was also radiolabeled as a control. Samples were boiled in sample buffer (24), electrophoresed in 10% acrylamide gels, and autoradiographed (21). Western blots were developed with anti-cruzain antibody (21).

The efficacy of a number of CPI on the mammalian stages of the life cycle of T. cruzi was evaluated as the percentage of growth inhibition of intracellular amastigotes in the presence of the inhibitors for up to 21 d (Table 1). The first intracellular cycle of untreated controls was completed within 6 d after infection. Mu-F-hF-VSφ was the most effective compound and inhibited growth 100%. Mu-F-V-VSφ and BOC-tic-hF-VS, which are less effective inhibitors of the protease itself, produced 30% and 60% growth inhibition, respectively.

Based on this initial screening of inhibitors, Mu-F-hF-VSφ with Mu-F-V-VSφ for comparison were further tested (Table 2). 20 μM concentrations of Mu-F-hF-VSφ cured T. cruzi–infected macrophages. Although intracellular amastigotes were still visible in Giemsa-stained cultures 24– 78 h after infection, their morphology was extremely abnormal 50 h after treatment. Cultures were amastigote free after 12 d. No intracellular amastigotes or release of trypomastigotes were observed in cultures treated for 12 d and maintained without CPI for up to 30 d in numerous independent experiments, indicating that Mu-F-hF-VSφ not only blocked the intracellular development of T. cruzi, but eventually eliminated all parasites. In contrast, amastigotes divided at normal rate in the presence of the chemically related, but less effective inhibitor of cruzain, Mu-F-V-VSφ.

Levels of parasitemia in T. cruzi–infected C3H mice were first analyzed after treatment with Z-F-A-FMK, Mu-bsu-hF-FMK, and Mu-F-hF-FMK (Fig. 1). These FMK inhibitors were chosen based on results of tissue culture screens (20, Table 1) and included a natural amino acid dipeptide (Z-F-A-FMK) and pseudopeptides designed to increase in vivo half-life (Mu-bsu-hF-FMK and Mu-F-hF-FMK). Control animals died at day 12 after infection while all animals treated with Mu-bsu-hF-FMK and Mu-F-hF-FMK survived throughout the experiment (18 d). The levels of parasitemia were lowest in mice treated with Mu-F-hF-FMK and ranged from 10–500 trypomastigotes per milliliter of blood. This corresponded to a reduction of 3 log units from untreated controls. Mice treated with the related but less effective (in protease substrate assays) dipeptide inhibitor Mu-bsu-hF-FMK had persistent high parasitemia in the range of 4 × 10^4–5 trypomastigotes per milliliter of blood. Not only was Z-F-A-FMK ineffective, but mice had higher parasitemia and died earlier than controls.

From tissue culture screens (Tables 1 and 2), vinyl sulfone derivatized pseudopeptides with high efficacy but less toxicity than FMK inhibitors were selected for further evaluation in a mouse model of acute Chagas' disease. The effect of CPI-treatment on the survival of T. cruzi infected C3H mice is shown in Table 3. Untreated controls infected with a very high dose (4 × 10⁶) of tissue culture– derived trypomastigotes all died by 4–5 d after infection. All five mice treated with Mu-F-hF-VSφ survived for 14– 16 d at which time the experiment was terminated and animals killed (Table 3, Experiment 1). When the infectious dose was reduced to 1 × 10⁶ T. cruzi trypomastigotes and mice treated with one 24-d regimen of Mu-F-hF-VSφ, two out of six mice survived up to 180 d after infection at which time the experiment was terminated (Table 3, Experiment 2).

From these pilot studies, the most promising lead compound was then evaluated in a new dosing regimen based on a pharmacokinetics analysis of Mu-F-hF-VSφ (25). Mice were treated three times daily with 2.1 mg/d i.p. of N-Pip-F-hF-VSφ (Table 3, Experiment 3). As indicated in Table 3, all 10 mice survived at least 7 d longer than untreated mice and 6 out of 10 mice survived over 270 d after infection. Three of the six long-term surviving mice had consistently negative blood-parasite cultures over the entire treatment period while the remaining three out of six were still positive, although well below controls. After being allowed to establish chronic infection for 3 mo, this latter group of 3 mice were retreated with 2.1 mg/d i.p. of N-Pip-F-hF-VSφ for 21 d (Table 4, Experiment 4). Blood cultures (10–50 μl) were negative for the retreated animals. In summary, 5 of 10 treated mice from this study have now survived for 9 mo without symptomatology and with negative parasitemia and hemocultures (one mouse died of accidental trauma after 100 d). Similar results were obtained with 3-wk-old outbred Swiss mice weighing initially 19– 21 g (data not shown).

In the studies described above, CPI treatment did not induce gross or microscopic abnormalities in either infected or uninfected mice. To further assess toxicity, doses of 4 mg/d i.p. of Mu-F-hF-VSφ and 12 mg p.o. of N-Pip-F-hF-VSφ for 4 d were also evaluated. No gross or microscopic abnormalities were observed in necropsied animals.

CPI have been shown to produce a characteristic Golgi abnormality in the extracellular epimastigote stage of T. cruzi (21).

To confirm that the same effect was produced in the treated amastigotes, amastigotes within host macrophages that had been treated with Mu-F-hF-VSφ (10 μM) for 48 h were examined after labeling with Bodipy FL ceramide that accumulates in the Golgi compartment. Cells were observed by contrast phase (Fig. 2, A and B) and fluorescence microscopy (Fig. 2, C and D). No Bodipy FL labeling was apparent in untreated intracellular amastigotes (Fig. 2,C), whereas the larger Golgi complex (G) of untreated host cells was visible. In contrast, the Golgi apparatus (g) of Mu-F-hF-VSφ–treated amastigotes had abnormally large, fluorescent vesicles consistent with ultrastructural alterations in the Golgi complex previously observed in treated epimastigotes (Fig. 2,D). To confirm these results, the normal morphology of control intracellular amastigotes (Fig. 3,A) was compared with the ultrastructure of CPI-treated tissue culture amastigotes (Fig. 3,B). Golgi complex and cytoplasmic vesicle alterations found in amastigotes (Fig. 3,B) resembled those described for the epimastigote stage, and consisted of significant dilation of cisternae (21). Similar ultrastructural alterations were evident in amastigotes isolated from heart muscle of T. cruzi–infected mice treated with Mu-F-hF-VSφ (Fig. 3,C). The amounts of cruzain expressed on the surface of Mu-F-hF-VSφ–treated and untreated amastigotes were quantified by immunoelectronmicroscopy (Fig. 4). A marked decrease in cruzain expressed on the cell surface of CPI-treated cells (Fig. 4,B) was evident compared with untreated amastigotes (Fig. 4 A).

T. cruzi intracellular amastigotes were incubated with or without cold inhibitor before incubation labeling with [¹⁴C]Mu-F-hF-VSφ. T. cruzi epimastigotes and recombinant cruzain were also radiolabeled and autoradiographed as standards (Fig. 5 A). ¹⁴C-inhibitor labeling of amastigote cruzain (lane 1) was abolished by preincubation with unlabeled Mu-F-hF-VSφ (lane 2). In amastigotes isolated from infested host cells two protease species were labeled. One comigrated with the epimastigote cruzain species (∼50 kD) known to contain both the catalytic and COOH-terminal domains (13) while the second comigrated with recombinant cruzain containing only the catalytic domain and with a macrophage protease (∼30 kD).

High toxicity and low efficacy make current chemotherapy for Chagas' disease highly unsatisfactory. Moreover, commercial nifurtimox production has been discontinued and benznidazole is at present the only treatment available. Until recently, it was still controversial as to whether drug treatment would have any effect on the more common chronic stage of Chagas' disease because it was unclear whether parasites were still present. However, there is now a clear association between parasitic burden and degree of myocardial damage (26). A recent follow up study in Argentina of benznidazole-treated versus untreated chronic chagasic patients showed less cardiomyopathy in the first group (27). Electrocardiogram patterns of benznidazole-treated patients confirmed an important reduction in disease progression (27). Andrade et al. (28) reported a correlation between heart disease and persistent parasitemia in mice. These studies emphasize the importance of the development of more effective and less toxic chemotherapy for both the acute and the chronic phase of Chagas' disease.

We have targeted cruzain (a.k.a. cruzipain, gp57/51), the major cysteine protease of T. cruzi (reviewed in 13, 29, 30), for the development of new chemotherapy. Previously tested dipeptide-based cruzain inhibitors that mimic substrate (20) were further modified to increase affinity for the catalytic site of cruzain, increase half-life in vivo, provide oral bioavailability, and reduce toxicity. In a previous report, we showed that these CPI induced death in the extracellular epimastigote stage of T. cruzi as a consequence of blocking the autocatalytic processing of cruzain precursor protein. This resulted in Golgi complex and endoplasmic reticulum abnormalities secondary to accumulation of unprocessed cruzain precursor molecules in the vesicle compartments (21). We now present evidence that these CPI also inhibit T. cruzi amastigote growth within macrophages by the same mechanism. Among the inhibitors tested, Mu-F-hF-VSφ most effectively blocked the intracellular cycle of the parasite (Table 2) and resulted in amastigote death in vitro. None of the inhibitors produced abnormalities in the host cells at the concentrations necessary to interrupt the parasite life cycle.

Subsequent testing of a number of CPI in an experimental mouse model of acute Chagas' disease showed that the inhibitors also disrupted the life cycle of T. cruzi in vivo. Initially, T. cruzi–infected animals were treated with the fluoromethyl ketone–derivatized peptidomimetics Mu-F-hF-FMK, Mu-bsu-hF-FMK, and Z-F-A-FMK (Fig. 1). Treatment with Mu-F-hF-FMK reduced parasitemia by 3 log units which resulted in the survival of infected mice throughout the experiment, terminated 18 d after infection. In contrast, animals treated with Mu-bsu-hF-FMK, a less effective inhibitor of the protease itself, and Z-F-A-FMK, a peptide with natural amino acids, had ∼100- and 10,000-fold higher parasitemias, respectively. Survival in the Z-F-A-FMK group (8 d) was significantly lower than in controls (12 d). The increased mortality and parasitemia probably result from a toxic metabolite of the natural amino acid inhibitor. Z-F-A-FMK is cleaved between the phenylalanine and alanine in vivo by an as yet unknown mammalian protease releasing alanine-FMK which, in turn, enters and inhibits the Krebs cycle (31). This fluoride-dependent toxicity results in hypothermia in mice, and T. cruzi has a higher growth rate at 35°C than at 37°C (32, 33). The peptidomimetics containing at least one nonnatural amino acid analogue (e.g., hF) do not undergo this cleavage and metabolism (25). To completely avoid the potential toxicity of FMK derivatives, vinylsulfone analogues were tested and found to extend survival to 120–180 d after infection with 10⁶ trypomastigotes. We further evaluated the more aqueous soluble derivative N-Pip-F-hF-VSφ that is absorbed after oral dosing. 5 out of 10 treated mice have now survived for over 9 mo, 4 of them with repeatedly negative hemocultures indicative of parasitological cure. No toxicity was observed at doses two- to sixfold higher than the therapeutic doses administered p.o. or intraperitoneally for 20 d or when mice were treated for 45 d with the therapeutic dose of 2 mg/d inhibitor.

To elucidate the mechanism of action of CPI versus intracellular amastigotes, the pathogenic stage of T. cruzi, we infected host cells whose cell cycles were arrested. Intracellular amastigotes treated with Mu-F-hF-VSφ showed major morphological alterations as early as 24–48 h after treatment, and macrophage cultures were amastigote free within 12 d. Mu-F-hF-VSφ produced abnormalities in the protein trafficking pathway (nuclear membrane, ER, Golgi complex) and induced the appearance of double-membrane vacuoles with similar morphology to autophagosome vacuoles (Fig. 3). Cruzain normally localizes to the cell surface and lysosome membrane of T. cruzi amastigotes (34, 35). CPI treatment significantly reduced the amounts of cruzain appearing both on the cell membrane (Fig. 4) and in the lysosome consistent with an arrest of cruzain transport. An increase of fluorescence with Bodipi FL of the Golgi complex in treated amastigotes was similar to that described for epimastigotes and correlated with enlargement of the Golgi complex secondary to retention of unprocessed cruzain (Fig. 2; reference 21). Radiolabeled inhibitor was used to confirm entry into amastigotes and specific labeling of the target protease cruzain (Fig. 5). The absence of any host cell or animal toxicity at therapeutic doses suggests that the parasites are more susceptible to inhibitor perhaps because of the redundancy of cysteine proteases in mammalian cells versus the parasite.

We have identified peptidomimetic cysteine protease inhibitors that consistently rescued mice from acute lethal infections of T. cruzi, reduced parasitemia by up to 3 log units, and cured mice in the chronic stage of disease treated by a 21-d regimen. These results provide an important “proof of concept” for the development of cysteine protease inhibitors as chemotherapy for a number of disease entities including cancer cell invasion, inflammation, osteoporosis, and microbial infections where cysteine proteases are thought to play a key role in pathogenesis (36). Although further improvements on vinyl sulfone inhibitor leads should be forthcoming, they clearly demonstrate that animals can tolerate CPI at concentrations and dosing schedules that eliminate an intracellular parasite. The pharmacokinetics of these inhibitors were adequate to sustain therapeutic levels and the N-methyl piperazine derivative demonstrated oral bioavailability. The efficacy and lack of toxicity of CPI in treating both acute and chronic T. cruzi infections supports a call for further development of these leads.

The authors thank Drs. M. Zimmerman and R. Smith (Prototek, Dublin, CA), and J. Palmer (Axys Pharmaceuticals, San Francisco, CA) for cysteine protease inhibitors; and G. Gan (Anatomic Pathology Service, VAMC San Francisco) for processing of histological samples.

Boc

butyloxycarbonyl

Bsn

benzyl succinic acid

CPI

cysteine protease inhibitors

F

phenylalanine

FMK

fluoromethyl ketone

hF

homophenylalanine

Mu

morpholine urea

Obzl

O-benzyl

Pip

piperazine

tic

tetrahydroisoquinoline 3-carboxylic acid

VAmBzl

valine acetamidomethylbenzyl

VS

vinyl sulfone

VSφ

vinyl sulfone phenyl

Yii

diiodo tyrosine

Z

benzyloxycarbonyl

p.o.

per oral