INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The protozoan parasite Cryptosporidium parvum causes self-limiting diarrhea in immunocompetent individuals and severe and protracted diarrhea in AIDS patients. Although many antimicrobial agents have been tested in vivo against this parasite, few have been found to be consistently effective in humans (3, 26). However, a few drugs have been found to be effective in in vivo animal models (4, 8), and in vitro studies have recently identified several promising candidates (19, 32).

In designing studies to identify drug targets in Cryptosporidium, we focused on transporters of the ATP-binding cassette (ABC) superfamily that includes the multidrug resistance (MDR) transporters (13, 15). ABC transporters have been identified in several protozoa, including Plasmodium falciparum (9, 31). A gene product, CpABC, with considerable homology to the mammalian MDR-associated protein (MRP) (22), was identified in C. parvum. In addition, an antibody generated against a P. falciparum ABC transporter, PfPgh1, cross-reacts with a 190-kDa protein in C. parvum (22). Based on their known function in mammalian cells it can be proposed that ABC transporters are involved in several aspects of transport in C. parvum. ABC transporters can function as transporters of critical nutrients, such as anions and lipids (17, 29), and thus their inhibition could result in the retardation or inhibition of cell growth. Alternatively, they could be responsible for the rapid efflux of some drugs, accounting for the high rate of innate resistance in this parasite to many classes of drugs. Drug resistance in some microbial systems has been linked to MDR expression (20).

Many modifiers of ABC transporters have been reported and characterized (10, 17), and they are often called resistance modifiers. One class of resistance modifiers that has been reported to interact with MDR transporters consists of cyclosporine and cyclosporin (CS) analogs (11, 27). Originally developed as an immunosuppressive agent by Borel et al. (6), cyclosporine has subsequently been found to have broad antimicrobial, including antiprotozoal, activity (21). Recent studies report that cyclosporine and CS analogs are active against P. falciparum (2), Plasmodium vivax (16), and Toxoplasma gondii (25). Cyclosporine is a fungal metabolite and a lipophilic, cyclic undecapeptide.

The primary cellular targets of cyclosporine are the cyclophilins, low-molecular-weight proteins that have activity in the cis-trans interconversion of proline-containing peptides and hence are named peptidyl-prolyl cis-trans isomerases (PPiases). Cyclosporine binds to cyclophilins, and the complex inhibits the Ca²⁺-calmodulin-dependent phosphatase, calcineurin, resulting in the blocking of the Ca²⁺-dependent signal transduction pathway of T-cell activation. Some analogs do not bind strongly to cyclophilin or inhibit PPiase activity. Other analogs bind to cyclophilin, but the complex does not inhibit calcineurin and is therefore not immunosuppressive (1, 24, 30). A second group of cellular targets of CS analogs are the P glycoproteins/MDR transporters, although the interaction between drug and transporter has not been characterized extensively. Analogs that bind weakly to cyclophilin generally are potent reversers of MDR through their interaction with P glycoproteins/MDR transporters (12, 27, 28). Using a photoactivatable derivative of cyclosporine, it was possible to identify a specific binding between a mammalian P glycoprotein and the drug (7). PSC-833 inhibited this interaction with a higher affinity than cyclosporine (7). Thus, because of our interest in ABC transporters and the proven antiprotozoal activity of cyclosporine and CS analogs, we examined their effect on short-term (48-h) in vitro growth, and they were found to have potent anticryptosporidial activity.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Drugs. Cyclosporine and paromomycin were from Sigma. Cyclosporine and CS analogs were a gift from A. Bell, Trinity College, Ireland, originally donated by J. F. Borel, Novartis, Basel, Switzerland. Paromomycin was prepared as a 100 mM solution in phosphate-buffered saline (PBS). Cyclosporine and CS analogs were prepared as 100 mM solutions in ethanol and stored at 20°C. The four CS analogs tested were (3'-keto-MeBmt1)-CsD (SDZ PSC-833), (8'-O-Me-dihydro-MeBmt1)-CsA (SDZ 205-549), (Me-D-Ser3)-CsA (SDZ 209-313), and (O-Ac-MeBmt1)-CsA (SDZ 033-243) (where Me is methyl, Bmt1 is 4-butenyl-4-methyl threonine, CsD is valine²-cyclosporine, CsA is cyclosporine, and Ser3 is serine 3) (2).

In vitro culture of C. parvum. C. parvum was grown in Caco-2 cells in 1-cm-diameter Transwells (Costar) essentially as described by Griffiths et al. (14). Caco-2 cells were seeded at a density of 5 × 10⁵ per well and cultured for 2 days in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal calf serum with penicillin and streptomycin (1%) at 7% CO₂. KSU-1 oocysts used to infect monolayers were a gift from Steve Upton, Kansas State University, Manhattan. Purified oocysts were washed twice with PBS, incubated in 10% Clorox for 10 min on ice, and washed twice in PBS. They were added to the Caco-2 cells at a density of 5 × 10⁵ per well. Parasites were cultured for 48 h, and the percentage of host cells infected with one or more meronts was estimated to be between 40 and 60%.

Drug assays. For drug assays, oocysts were added to the Caco-2 cells for 3 h in the presence of the drug. Unexcysted oocysts and oocyst shells were removed by washing the wells twice with DMEM. The medium was replaced, the drug was added, and the parasites were cultured for an additional 48 h. Paromomycin was used to validate the drug assay method, as its anticryptosporidial effect had been previously established (18, 32). The assay to test the activity of drugs as resistance modifiers of MDR transporters is known as a chemosensitivity assay. For this study the chemosensitivity assay was performed with cyclosporine. The cultures were treated with paromomycin and cyclosporine simultaneously. The concentrations of cyclosporine typically used in the chemosensitivity assays in mammalian cells are 0.5 to 2.0 µM, below that known to cause significant cell toxicity. In this study cyclosporine was added at concentrations of 0.1, 0.5, and 1.0 µM to cultures in combination with paromomycin.

Determination of parasite numbers in drug-treated cultures. The number of parasites in drug-treated cultures was determined by an immunofluorescence assay (IFA). For one experiment, parasites were stained with a polyclonal antibody (14). For all other experiments, monoclonal antibody (MAb) 1D8 was used. MAb 1D8 was a gift from Michael Riggs, University of Arizona, and was produced according to the protocol described previously (23). It was demonstrated to react with intracellular stages of C. parvum but not unexcysted oocysts and empty oocyst shells (23a). After 48 h of incubation with the drug. Transwells were washed twice in PBS and fixed in ethanol for 10 min. The wells were washed in PBS and then incubated with polyclonal antibody (1:500 dilution) or MAb 1D8 (1:3 dilution) for 45 min. The wells were washed twice with PBS and then overlaid with fluorescein isothiocyanate-labeled goat anti-rabbit antibody (Gibco) as the polyclonal antibody or fluorescein isothiocyanate-labeled rabbit anti-mouse antibody (Gibco) for 45 min at room temperature. The wells were washed twice, and membranes were cut out of the wells and placed on slides covered with coverslips and mounting fluid (50% glycerol in PBS). The slides were viewed with a Nikon epifluorescence microscope. Twenty microscope fields were counted for each sample.

Host cell toxicity. The effects of the drugs on Caco-2 host cell viability were tested by the CellTiter 96 assay (Promega), which is a colorimetric assay to measure the number of viable cells. Concurrent with the drug experiments, drugs were added to confluent monolayers of uninfected Caco-2 cells for 51 h. The monolayers were washed once in DMEM and then assayed in the CellTiter 96 assay according to the manufacturer's directions. The assay depends on the reduction of a tetrazolium compound, MTS, to formazan, which has an absorption optimum of 490 nm. The quantity of product as measured by optical density at 490 nm (OD₄₉₀) is directly proportional to the number of living cells. MTS and formazan at the concentrations recommended were added to drug-treated and control cells for 2 h, and the OD₄₉₀ was measured directly. Drug toxicity, resulting in nonmetabolizing cells, is reflected in low OD₄₉₀ readings compared to that of the control.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Effect of cyclosporine on C. parvum invasion and growth. Inhibition of growth of C. parvum by cyclosporine in the concentration range 0.1 to 50 µM was analyzed in a bar graph (Fig. 1A), and 50% inhibitory concentrations (IC₅₀s) were determined from line graphs as shown for the 51-h incubation assay (Fig. 1B). Maximum inhibition was observed when the drug was present during the invasion period (3 h) and the 48-h culture period, for a total of 51 h (Fig. 1A). The IC₅₀ of cyclosporine present during invasion and the 48-h culture period was 1.5 µM. The IC₉₀ was 5 µM, reflecting the narrow inhibitory range. When the drug was omitted during the 3-h invasion period but present during the 48-h growth period, inhibition was less (IC₅₀, 2.5 µM), but 90% inhibition was observed with 10 µM (Fig. 1B). Cyclosporine present only during the 3-h invasion period reduced intracellular parasitemia by 70% at high (50 µM) concentrations (Fig. 1A). To examine the effects of cyclosporine on invasion, oocysts were allowed to excyst in the presence of cyclosporine at 37°C. The cyclosporine was removed by two washes, and the sporozoites were added to Caco-2 cells. Sporozoites exposed to cyclosporine for this short time showed significantly reduced intracellular growth at high concentrations (Fig. 1C): the IC₅₀ for inhibition was 35 µM. Since the drug was present only during the excystation period and was removed before the sporozoites were challenged with Caco-2 cells, it appears that cyclosporine can inhibit invasion. As can be seen from Fig. 1, standard deviations were small, reflecting the close agreement between results of separate experiments, although there was more divergence at low concentrations. This is in contrast to the results of inhibition with paromomycin (see Fig. 3). The parasite numbers shown in Fig. 1 were estimated by IFA with a polyclonal antibody in one experiment and a MAb in the second. Although the MAb gave a clearer fluorescent pattern, it was considered necessary to confirm the dramatic loss of parasites in the drug-treated cultures with a second antibody. It was possible that cyclosporine had a direct effect on the antigen recognized by MAb 1D8 and not on parasite growth per se.

View larger version (14K): [in this window] [in a new window]   FIG. 1.   Cyclosporine inhibition of C. parvum growth in in vitro cultures. The following experiments are shown. (i) C. parvum oocysts were added to Caco-2 cell monolayers for 3 h in the presence of cyclosporine (0.1 to 50 µM), unexcysted oocysts were washed out, and the parasites were cultured for a further 48 h in the presence of the drug, for a total of 51 h (51 hr). (ii) Oocysts were added to the monolayers for 3 h in the presence of the drug, the drug and unexcysted oocysts were removed by washing, and the parasites were cultured for a further 48 h without the drug (3 hr). (iii) Oocysts were added to the monolayers for 3 h without the drug, unexcysted oocysts were removed, and the parasites were cultured for 48 h with the drug (48 hr). Parasite numbers were estimated from the IFA, using the polyclonal antibody in one experiment and, in the second, MAb 1D8. Drug activity was calculated as percent inhibition of growth, and the results represent two separate experiments. The error bars indicate standard deviations. (A) Bar graphs were plotted with Microsoft Excel software. The cyclosporine used in this experiment was from Sigma. (B) Values for percent inhibition for a 51 h incubation plotted in a linear graph, with 50% inhibition shown by the dashed line. All assays were plotted similarly to determine IC50s. (C) Oocysts were excysted in the presence of cyclosporine for 1 h, the drug was removed, and the parasites were added to Caco-2 cells and cultured for 48 h without the drug. Fifty-percent inhibition is shown by the dashed line.

Effects of nonimmunosuppressive analogs of CS on C. parvum in vitro. Inhibition of C. parvum growth by the two most active CS analogs is shown in Fig. 2. The IC₅₀s of all analogs were estimated from line graphs and are summarized in Table 1. SDZ 033-243 and SDZ PSC-833 were most effective, with IC₅₀s of 0.4 and 1.0 µM, respectively, which are less than that of cyclosporine. The IC₉₀ of SDZ PSC-833 was 2.5 µM. SDZ 205-549 and SDZ 209-313 were 10-fold less effective, with IC₅₀s of 5 and 7 µM, respectively. As mentioned above, the results with duplicates were very close between experiments, as reflected in the small standard deviations, which in some instances were too small to register on the bar graphs. In a few experiments, limited by the small amounts of analogs available, analogs SDZ 033-243 and SDZ 205-549 were also found to inhibit invasion of excysted sporozoites (data not shown). The immunosuppressive and resistance modifying activities of the CS analogs (27, 30) are included in Table 1 for the purpose of discussion.

View larger version (61K): [in this window] [in a new window]   FIG. 2.   CS analogs inhibit C. parvum growth in vitro. Oocysts were added to monolayers in the presence of CS analogs for 3 h, unexcysted oocysts were removed, and the parasites were cultured for an additional 48 h in the presence of drugs. All analogs used were inhibitory, but the two most active, SDZ 033-243 and SDZ PSC-833, are shown. Graphs were also plotted for the other two analogs, SDZ 205-549 and SDZ 209-313. IC50s determined from the graphs are listed in Table 1. Error bars indicate standard deviations. The cyclosporine (CsA) used in this experiment was from Novartis and gave results almost identical to those for the cyclosporine from Sigma (Fig. 1).

Chemosensitivity assaythe effect of cyclosporine on paromomycin sensitivity. The chemosensitizing activity of cyclosporine was tested by comparing inhibition in the in vitro growth assay of paromomycin alone and paromomycin and cyclosporine. Paromomycin was tested in the range of 5 to 400 µM (Fig. 3). The IC₅₀ was estimated to be 85 µM (Fig. 3). From the values of inhibition for each experiment, standard deviations were calculated as shown in Fig. 3, with each point representing an average of results from two separate experiments. For paromomycin, the values for percent inhibition were quite variable, as reflected in the large standard deviations. The estimated IC₅₀ is comparable to that published by Woods et al. (32), verifying the accuracy of the assay, but lower than that calculated from the data published by Marshall and Flanigan (18). This variability may be a result of the relatively weak anticryptosporidial effect of the drug. Drug assays were performed with a combination of paromomycin and cyclosporine at 0.1, 0.5, and 1.0 µM. The results for cyclosporine at 0.5 µM are shown and indicate that cyclosporine had an insignificant effect on paromomycin sensitivity, lowering the IC₅₀ only from 85 to 70 µM (Fig. 3). Cyclosporine at 0.1 µM had no effect on paromomycin sensitivity (data not shown). Cyclosporine alone at 1.0 µM had a significant inhibitory effect (Fig. 1).

View larger version (44K): [in this window] [in a new window]   FIG. 3.   Paromomycin inhibition of C. parvum is not modified by cyclosporine. Parasites were cultured for 48 h in the presence of paromomycin (5 to 400 µM) with and without cyclosporine (CsA) (0.5 µM). The values are the averages of two experiments, and the error bars represent standard deviations. IC50s were estimated from linear graphs.

Appearance of drug-treated parasites. After treatment with cyclosporine and CS analogs the normally robust intracellular parasites were small and irregular in shape (Fig. 4). IFA with MAb 1D8 revealed that the intracellular stages after 48 h of culture were heterogeneous in size. The antigen recognized by MAb 1D8 has not been identified by immunoblotting, but in the larger parasites the antibody clearly stains a membrane structure consistent with the periphery of the parasite (Fig. 4A). This distribution was lost in parasites incubated with low concentrations of cyclosporine (Fig. 4B), and the antigen recognized by MAb 1D8 was localized in a small area that was not consistent with the periphery of a healthy parasite. MAb 1D8 has been shown to react only with intracellular stages and not unexcysted oocysts (23a).

View larger version (73K): [in this window] [in a new window]   FIG. 4.   IFA of drug-treated C. parvum cultured in Caco-2 cells. Untreated parasites (A) and parasites treated with cyclosporine (2.5 µM) (B) were processed for IFA with MAb 1D8 as described in Materials and Methods. In control cultures, the number of infected cells was in the range of 40 to 60%. Bar, 10 µm.

Cell toxicity. No significant toxicity, as measured in the CellTiter 96 assay, was observed for cyclosporine or CS analogs (Table 1). However, the OD₄₉₀s for analogs SDZ 033-243 and SDZ PSC-833 were always higher than those of the other three drugs, indicating that the other three, all strong cyclophilin binders, may exhibit some small toxicity at this concentration. It also should be noted that the drugs were added to confluent cells, and thus an effect on cell division would not be reflected in the results. The OD₄₉₀s represent a measure of cell viability and are given for Caco-2 cells treated with the highest concentration of cyclosporine and CS analogs used in the drug assays (50 µM). The OD₄₉₀ for control (untreated) cells was 0.83. No significant toxicity was observed for paromomycin in this assay (data not shown).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Analogs of CS were demonstrated to be selective inhibitors of the intracellular apicomplexan parasite C. parvum, cultured in Caco-2 cells. Two analogs, SDZ 033-243 and SDZ PSC-833, were particularly potent, and their IC₅₀s were calculated to be 0.4 and 1.0 µM, respectively (Table 1). The IC₉₀ of SDZ PSC-833 was 2.5 µM. After treatment, the few remaining parasites were very small in contrast to the untreated controls and the distribution of the antigen recognized by MAb 1D8 was considerably altered. Whether or not these parasites are viable after treatment will be difficult to determine, as the time of in vitro culture is limited to at most 72 h. However, we intend to investigate whether the effect of cyclosporine is reversible by exposing the parasites to the drugs for shorter time periods. No significant toxicity on Caco-2 cells was observed at concentrations up to 50 µM with the CellTiter 96 assay. The inhibitory activity of cyclosporine compares favorably with the IC₅₀s calculated for other anticryptosporidial drugs in in vitro assays (32).

Cyclosporine was originally developed as an immunosuppressive agent by Borel and colleagues (5, 6). Subsequently, analogs of CS were developed, but some were found to have little or no immunosuppressive activity (30). The two most active analogs against C. parvum identified in this study, SDZ PSC-833 and SDZ 033-243, have very low or no immunosuppressive activity (30). However, these two analogs were demonstrated by Twentyman (27) to be extremely potent reversers of MDR in mammalian cells (Table 1). A similar picture has emerged in studies of P. falciparum (2) and T. gondii (25): the analogs that were the most potent inhibitors of parasite growth were those that bind weakly to cyclophilin/PPiase and were not immunosuppressive. Cyclosporine itself, which is both immunosuppressive and a moderate reverser of drug resistance, was inhibitory to parasite growth in this study and against P. falciparum and Toxoplasma. Bell et al. (2) have shown that SDZ PSC-833 was the most inhibitory against P. falciparum, exhibiting an IC₅₀ of 0.03 µM. Silverman et al. (25) recently reported that SDZ 215-918 was the most potent inhibitor of Toxoplasma growth in vitro and also showed some activity in vivo. This drug is nonimmunosuppressive but, according to other studies, is a potent reverser of MDR, suggesting that it is antiparasitic as a result of its interaction with the MDR transporter thought to be present in Toxoplasma. There is no direct demonstration, however, that CS analogs interact with or bind to MDR transporters of mammalian cells or other apicomplexans.

Since the relationships between structure and cellular activity of the analogs found to inhibit C. parvum growth are in general agreement with those of the analogs reported to be active against P. falciparum (2), P. vivax (16), and Toxoplasma (25), it is reasonable to propose that the analogs act by a common mechanism, which may be inhibition of MDR transporters. However, CS analogs that bind strongly to PPiase also inhibit parasite growth, albeit less dramatically. Thus, it is possible that both cyclophilin and ABC transporters are the targets of this class of drugs.

In contrast to the potent growth-inhibitory activities of cyclosporine and CS analogs, cyclosporine, at the appropriate concentrations, does not modify the sensitivity of the parasite to paromomycin. This so-called chemosensitizing effect is observed with MDR cancer cells, which can be rendered drug sensitive by cyclosporine and CS analogs as well as many other drugs (10). Generally, the sensitivity to drugs will increase two- to fourfold in the presence of the chemosensitizers. The concentrations of MDR reversers required to produce this effect is lowwell below the levels where cytoxicity is observed. For cyclosporine, it is in the range of 0.5 to 2 µM (10). We performed this assay originally with cyclosporine concentrations of 0.1, 0.5, and 1.0 µM. As shown in Fig. 3, 0.5 µM concentrations had no significant effect on sensitivity to paromomycin. Even when it is significant, this effect is most likely due simply to an additive effect of the two drugs and is not a potentiation effect. When the assay was performed with 1.0 µM concentrations, we observed a significant antiparasitic effect with cyclosporine alone and could not interpret this result as being due to any chemosensitizing activity. Paromomycin was used in this assay, as it is considered only moderately inhibitory to C. parvum. Only very high doses inhibit growth by 90% (18), a fact that could reflect poor transport into the intracellular compartment, high efflux, or alternatively, an altered target, the 30S subunit of rRNA. However, although high doses of paromomycin inhibited growth significantly, the dose response was not affected by cyclosporine. Similar results were obtained with verapamil, another reverser of drug resistance (data not shown). One interpretation of this result is that cyclosporine does not increase the intracellular concentration of paromomycin by decreasing efflux, suggesting that the ABC transporter of C. parvum does not play a role in paromomycin efflux from the parasite. Alternatively, cyclosporine may not affect the function of ABC transporters in this protozoan. Several other drugs were tested in the chemosensitivity assay with similar results. Parasites were grown in the presence of the sulfur drug sulfanilamide, which (alone, in the concentration range of 5 to 500 µM) had no effect on growth. There was not a significant increase in sensitivity to this drug in the presence of cyclosporine (data not shown). Although these represent limited studies, they do suggest that efflux via an ABC transporter is not a factor in the resistance of this parasite to certain drugs, as has been postulated (22).

In summary, we have demonstrated that cyclosporine and CS analogs are very potent inhibitors of C. parvum growth in vitro. Although the current study gives little information on their antiparasitic mechanisms, similarities in the structures of the active analogs with those of analogs active against other apicomplexan parasites suggest that they may act through a common mechanism, possibly by interacting with ABC transporters and parasite cyclophilins. Although the natural substrates for parasite ABC transporters are unknown, the results of this study and others suggest that transport of those substrates is critical for parasite growth. Based on analogy with mammalian cells, it could be an anion transporter (17). By using an antibody generated against P. falciparum PfPgh, it was possible to demonstrate that an ABC transporter was expressed in sporozoites of C. parvum (22). Subsequently, we have found by IFA that the antibody reacts strongly with intracellular stages of C. parvum cultured in Caco-2 cells, suggesting that the protein is also expressed in asexual stages. Future genetic studies with the recently identified gene for a C. parvum ABC transporter will be aimed at determining if cyclosporine and CS analogs interact with the CpABC protein (22).

There is currently no fully effective drug to treat cryptosporidiosis, despite extensive efforts to develop one, although recently several anticryptosporidial drugs have been identified as effective in in vivo animal models (4, 8) and in vitro assays (32). One of the CS analogs, SDZ PSC-833, has been tested in human clinical trials as a reverser of MDR cancer cells (10). It apparently shows little toxicity at therapeutic doses, and therefore it may be possible to test the usefulness of this and other CS analogs as anticryptosporidial drugs.