Inhibitory Activities of Epidermal Growth Factor Receptor Tyrosine Kinase-Targeted Dihydroxyisoflavone and Trihydroxydeoxybenzoin Derivatives on Sarcocystis neurona, Neospora caninum, and Cryptosporidium parvum Development

Abstract

Several gene sequences of parasitic protozoa belonging to protein kinase gene families and epidermal growth factor (EGF)-like peptides, which act via binding to receptor tyrosine kinases of the EGF receptor (EGFR) family, appear to mediate host-protozoan interactions. As a clue to EGFR protein tyrosine kinase (PTK) mediation and a novel approach for identifying anticoccidial agents, activities against Sarcocystis neurona, Neospora caninum, and Cryptosporidium parvum grown in BM and HCT-8 cell cultures of 52 EGFR PTK inhibitor isoflavone analogs (dihydroxyisoflavone and trihydroxydeoxybenzoine derivatives) were investigated. Their cytotoxicities against host cells were either absent, mild, or moderate by a nitroblue tetrazolium test. At concentrations ranging from 5 to 10 μg/ml, 20 and 5 analogs, including RM-6427 and RM-6428, exhibited an in vitro inhibitory effect of ≥95% against at least one parasite or against all three, respectively. In immunosuppressed Cryptosporidium parvum-infected Mongolian gerbils orally treated with either 200 or 400 mg of agent RM-6427/kg of body weight/day for 8 days, fecal microscopic oocyst shedding was abolished in 6/10 animals (P of <0.001 versus untreated controls) and mean shedding was reduced by 90.5% (P of <0.0001) and 92.0% (P of <0.0001), respectively, higher levels of inhibition than after nitazoxanide (200 mg/kg/day for 8 days) or paromomycin (100 mg/kg/day for 8 days) treatment (55.0%, P of <0.001, and 17.5%, P of >0.05, respectively). After RM-6427 therapy (200 mg/kg/day for 8 days), the reduction in the ratio of animals with intracellular parasites was nearly significant in ileum (P = 0.067) and more marked in the biliary tract (P < 0.0013) than after nitazoxanide or paromomycin treatment (0.05 < P < 0.004). RM-6428 treatment at a regimen of 400 mg/kg/day for 12 days inhibited oocyst shedding, measured using flow cytometry from day 4 (P < 0.05) to day 12 (P < 0.02) of therapy, when 2/15 animals had no shedding (P < 0.0001) and 11/15 were free of gut and/or biliary tract parasites (P < 0.01). No mucosal alteration was microscopically observed for treated or untreated infected gerbils. To our knowledge, this report is the first to suggest that the isoflavone class of agents has the potential for anticoccidial therapy.

Apicomplexan coccidian parasites constitute a diverse phylum of considerable medical importance to humans and animals. Cryptosporidium spp., as enteric coccidia, cause diarrheal disease worldwide and significantly threaten immunocompromised individuals (11). Neospora caninum infects a wide range of mammalian species, including sheep, goats, horses, dogs, and cattle, in which infection is recognized as a major cause of abortion (13). The heteroxenous, cyst-forming Sarcocystis neurona was the first identified etiological agent of equine protozoal myeloencephalitis, a major neurological syndrome in horses in the Americas (14). In spite of numerous attempts at screening candidate therapies, a limited number of anticoccidian agents is presently available, which prompts evaluation of new candidate compounds.

On the basis of their primary sequences, several protein kinases (PK) of parasitic protozoa belong to well-characterized families known to play pivotal roles in signal transduction in other eukaryotes. Epidermal growth factor (EGF)-like peptides, a phylogenetically conserved group of mitogens, act via binding to receptor tyrosine kinases of the EGF receptor (EGFR) family, which are located at the surface of target cells (30). Ligand binding to the extracellular portion of the corresponding EGFR leads to an activation of the intracellular tyrosine kinase (TK) domain, which initiates downstream signaling pathways, of which the best characterized is the mitogen-activated protein kinase cascade (21, 24, 39).

EGF-like peptides appear to be important mediators of host-parasite interactions during protozoan infections. It has recently been shown that incubation with mammalian EGF stimulated G protein-dependent signaling mechanisms in Trypanosoma cruzi (22). It has also been suggested that the effects of mammalian EGF are mediated by parasite-determined EGFR (23). Although genes encoding conventional receptor-linked TK have not been identified yet for the genomes of Plasmodium falciparum, Leishmania major, or Trypanosoma brucei, tyrosine/histidine phosphorylation evidenced for parasitic protozoa supports protein tyrosine kinase (PTK) activity (12). Since structural differences between parasite and host PK might result in differential affinities of inhibitory molecules, PTK and more generally PK are attractive targets in the search for therapeutic agents against parasitic diseases (25).

Potential mediation of parasite-determined, not yet characterized, EGFR-like molecules may provide a clue for candidate inhibitory molecules such as flavonoid compounds. Genistein, a naturally occurring isoflavone found chiefly in soy products, exhibits micromolar potency in inhibiting EGFR protein tyrosine kinases (EGFR PTK) (1, 2, 37). The aim of the present study was to investigate the anticoccidial activities of EGFR PTK inhibitor genistein analogs as a novel approach to identify anticoccidian agents and to explore EGFR PTK-mediated regulatory mechanisms of coccidian development.

MATERIALS AND METHODS

Agents.

Fifty-two dihydroxyisoflavone and trihydroxydeoxybenzoin derivatives were synthesized by the Romark Center for Drug Discovery at University of Liverpool, Liverpool, United Kingdom (Table 1 and Fig. 1). Their inhibitory activities on control poly(Glu-Ala-Tyr) phosphorylation were tested using a purified A-431 cell membrane preparation as the human EGF TK source (9) by reference to the standard compound PD 153035 (CEREP, La Celle l'Evescault, France; data not shown). Paromomycin (PRM) was purchased from Sigma, St Louis, Mo. Nitazoxanide (NTZ), a 2-acetyloxy-N-[(5-nitro-2-thiazolyl)] benzamide, was obtained from Romark Laboratories, Tampa, Fla. For in vitro studies, test agents were solubilized in 100% dimethyl sulfoxide (DMSO) and added to culture wells/flasks at final concentrations ranging from 1 to 10 μg/ml, leading to a final DMSO concentration in culture of 0.1% (vol/vol). The same concentration of DMSO was added to cultures with no added agent. For animal gavage, powder formulations of the new synthetic agents PRM and NTZ were suspended in 5% carboxymethylcellulose water.

TABLE 1.

Nomenclature of substituted dihydroxyisoflavone and trihydroxydeoxybenzoin derivatives

Derivative			Substituenta
	Benzopyrane ring		C2				C3					C4						C6
	Carbon	Substituent	Cl	Br	F	NO2	Cl	Br	F	NO2	NH2	Cl	Br	F	NO2	NH2	NHCOCH3	Cl	Br	F	NO2
5,7-Dihydroxyisoflavone	None	None	6436	6423	6439	6429	6437	6424	6440	6411	6415	6438	6425	6441	6403	6404	6406
(3-phenyl ring)	C2	C3O2H5					6443	6427	6446	6414	6417	6444	6428	6447	6407	6412		6442	6426	6445	6413
	C2	C3O2H5						6448
	C7	MeO
	C2	C3O2H5
	C5	MeO						6449
	C7	MeO
	C2	COOH					6451	6450					6452		6408
	C2—C3	C2=C3													6419
	C2	COOH						6453
	C7	MeO
	C2	COOH
	C5	MeO						6454
	C7	MeO
2,4,6-Trihydroxydeoxybenzoin (2-phenyl ring)			6430	6420	6433	6410	6431	6421	6434	6409	6416	6432	6422	6435	6401	6402	6405

^aNumbers correspond to agent numbers, i.e., 6436 denotes RM-6436.

FIG. 1.

Structures of (A) 5,7-dihydroxyisoflavone (3-phenyl-4H-benzopyran-4-one) and (B) 2,4,6-trihydroxydeoxybenzoin (2,4,6- hydroxy-1,2-diphenyl-ethan-1-one).

Host cell lines.

BM cells originally obtained from peripheral blood monocytes of a 6-year-old Guernsey cow (a kind gift from David S. Lindsay, Virginia-Maryland Regional College of Veterinary Medicine, Va.) were grown in RPMI 1640 medium (GIBCO BRL, Gaithersburg, Md.) supplemented with 2% (vol/vol) heat-inactivated fetal calf serum (FCS), 100 U of penicillin G/ml, and 100 μg of streptomycin/ml. Human colonic epithelial HCT-8 cells (CCL 244, passages 42 to 62; obtained from the American Type Culture Collection, Rockville, Md.) were maintained in RPMI 1640 medium supplemented with 5% FCS, 1% nonessential amino acids, 100 U of penicillin G/ml, and 100 μg of streptomycin/ml in 75-cm² Falcon flasks, and the medium was changed every second to third day.

Evaluation of host cell toxicity of agents.

Flat-bottomed 96-well microtiter plates were seeded with BM or HCT-8 cells, and resulting monolayers were used to determine the cytotoxic effects of the test agents. Each agent dilution was examined in triplicate. Cell death due to toxicity was determined using a nitroblue tetrazolium chloride monohydrate reduction assay (CellTiter 96 AQ_ueous nonradioactive cell proliferation assay; Promega Corp., Madison, Wis.). Decreases in the optical density at 450 nm (OD₄₅₀) were expressed as percentages of the OD₄₅₀ value in agent-free control cultures.

Origin of parasites.

Cryptosporidium parvum oocysts (bovine genotype 2) were purified from feces obtained from calves experimentally infected with an isolate maintained by R. Mancassola and M. Naciri, Laboratoire de Pathologie Aviaire, Institut National de Recherche Agronomique, Nouzilly, France. Feces stored in a 2.5% K₂Cr₂O₇ solution for less than 3 months were layered on a discontinuous sucrose density gradient (densities, 1.045 and 1.090), and isolated oocysts were bleached before excystation and cell culture infection. Sarcocystis neurona strain SnSO-1 merozoites originated from a Southern sea otter, and Neospora caninum merozoites of the NC-1 strain were kind gifts of David S. Lindsay and were handled as previously reported (15, 27).

Assessment of the inhibitory activities of agents on Neospora caninum and Sarcocystis neurona in vitro development.

Cells were resuspended in 25-cm² culture flasks or in 6-well culture plates containing culture medium. They were incubated for 1 to 2 days. When the monolayer was >80% confluent, each culture recipient was inoculated with Sarcocystis neurona or Neospora caninum merozoites harvested from infected BM cell cultures which were detached from the plastic growth surface by using a rubber cell scraper and passed through a 30-gauge needle attached to a syringe to rupture host cells. As previously described, this procedure resulted in cell monolayer infection, which was microscopically checked (16, 26). Parasites were harvested by centrifugation (10 min, 1,800 × g, 4 to 10°C) and resuspended in RPMI 1640 medium at room temperature. Merozoites were counted using a hemacytometer, and each culture flask or well containing confluent cell monolayers was inoculated with 2.5 × 10⁵ to 5 × 10⁵ parasites. Each culture was performed in triplicate, and each set of experiments was done twice. Two hours later, medium containing free merozoites was removed and replaced by agent-containing medium or agent-free medium as a control. Quantitation of merozoite production was conducted at a time sufficient for production of large numbers of merozoites in untreated cultures, i.e., 3 days for Neospora caninum and 9 days for Sarcocystis neurona (six and three schizogonic cycles, respectively). The merozoite-containing medium was removed separately from each flask and rinsed with 2 ml of Ca²⁺- and Mg²⁺-free Hanks' balanced solution to detach adhering merozoites, which were pooled with the merozoite-containing medium. After centrifugation at 1,800 × g for 10 min at 4°C, the supernatant was discarded, the pellet volume was measured, and the number of merozoites was determined by being counted using a hemacytometer. The total number of merozoites present in each well/flask was determined by multiplying the volume of the pellet by the number of merozoites and was expressed as the mean number (±1 standard deviation) from 10 counts. The activity of each test compound was expressed as follows: inhibition of merozoite production (%) = [(mean number of merozoites in treated cultures) − (mean number of merozoites in untreated cultures)]/(mean number of merozoites in untreated cultures) × 100. The 50% inhibitory concentration (IC₅₀) was defined as the concentration of agent (in μg/ml of culture) which resulted in a mean 50% inhibition of parasite development.

Assessment of the inhibitory effects of agents on Cryptosporidium parvum in vitro development.

To prepare monolayers for Cryptosporidium parvum infection, HCT-8 cells (10⁵/ml) were seeded in 16-well tissue culture chambers (LabTek chamber slides; Nunc, Rochester, N.Y.). Cells were cultured at 37°C in a humidified, 5% CO₂ atmosphere to >80% confluence, as assessed before Cryptosporidium parvum infection. Purified oocysts were counted in a hemacytometer and permitted to excyst in a 1.5% taurocholic acid solution (Sigma) in BHK 21 medium for 90 min at 37°C in a humidified, 5% CO₂ atmosphere. Parasite suspensions (containing excysted and nonexcysted oocysts) were sieved through a 5-μm cellulose acetate filter (Sartorius, Goettingen, Germany) to remove nonexcysted oocysts, empty shells, and debris, which were microscopically absent from final sporozoite suspensions. Fifty microliters of medium containing sporozoites at a concentration of 2.5 × 10⁵/ml was added to each confluent-monolayer well. Slides were incubated for 2 h at 37°C, and medium containing free sporozoites was removed and replaced by 0.2 ml of agent-containing or agent-free coculture medium which consisted of RPMI 1640 containing ascorbic acid (35mg/liter), glucose (25 mM), insulin (100 IU/liter), HEPES buffer (15 mM), streptomycin (1 g/liter), penicillin (100,000 IU/liter), and FCS (5% [vol/vol]). Under these conditions, complete (asexual- and sexual-stage) Cryptosporidium parvum development was obtained as previously described (20, 28). Forty-eight hours later, treated or nontreated infected cells layered on tissue culture chambers were fixed in pure methanol for 10 min and rinsed with phosphate-buffered saline (PBS). Cryptosporidium parvum development was verified by indirect immunofluorescence in a moist chamber at 37°C. PBS-normal goat serum (3%)-Tween 20 (0.1%) buffer was added to each culture well for 3 h to inhibit nonspecific protein binding. Hyperimmune rat anti-Cryptosporidium parvum serum (final dilution in PBS-normal goat serum-Tween buffer of 1:400) was added to each well for 30 min. Wells were washed three times with PBS, and for 30 min fluorescein isothiocyanate-conjugated anti-rat immunoglobulin G and immunoglobulin M (heavy and light chain) goat antibodies (final dilution, 1:200; Jackson Immunoresearch, Westgrove, Pa.) were added. Wells were washed three times with PBS, and slides were mounted in buffered Mowiol (Calbiochem, La Jolla, Calif.). All developmental parasite stages were counted in 20 microscopic fields for each culture condition by using an epifluorescence immersion microscope (×1,250). Activities of the test compounds were expressed as follows: inhibition of Cryptosporidium parvum development (%) = [(mean number of intracellular parasitic forms in treated wells) − (mean number of intracellular parasitic forms in untreated wells)]/(mean number of parasitic forms in untreated wells) × 100. The IC₅₀ was defined as the concentration (in mg/liter of culture) of agent which resulted in a mean 50% inhibition of Cryptosporidium parvum development.

In vivo assessment of RM-6427 and RM-6428 in Cryptosporidium parvum-infected immunosuppressed gerbils.

An immunosuppressed-gerbil model was used as previously described (3). Male and female Mongolian gerbils (Meriones unguiculatus) weighing 22 to 27 g at the beginning of the study were individually housed in plastic cages equipped with a grill ceiling providing rodent food granules and water ad libitum. Each cage was protected by a top sterile paper wrap to comply with level II contamination requirements. Animals were handled according to the technical and ethical regulations of the French Ministry of Agriculture. Gerbils were immunosuppressed by dexamethasone (0.8 mg/animal; Qualimed, Puteaux, France) injected every second day for 10 days before oocyst ingestion and administered until the end of the experiment, i.e., day 12 postinfection. On day 0, each gerbil was gavaged with 10⁵ Cryptosporidium parvum oocysts, and animals were divided into five treatment groups as indicated in Table 2. Starting 4 h after oocyst ingestion, RM-6427, PRM, or NTZ suspension was administered twice daily by oral gavage, for 8 (RM-6427) or 12 (PRM and NTZ) consecutive days. Cryptosporidium parvum infection was assessed by measuring oocyst shedding in feces collected for 24 h from each animal on days 4, 8, and 12 postinfection. Feces were suspended in 10% (wt/vol) formalin solution and homogenized, and oocysts were counted by phase-contrast microscopic examination of smears prepared by mixing fecal suspensions with a carbol fuchsine solution. Oocyst numbers were expressed per microscopic field (MF) after counting 10 MFs (×400 magnification) or verifying the absence of oocysts in 30 MFs. Results were measured as the mean oocyst counts determined by two independent investigators.

TABLE 2.

In vivo efficacies of RM-6427, RM-6428, nitazoxanide, and paromomycin against Cryptosporidium parvum in immunosuppressed Mongolian gerbils

Agent	No. of animals	Regimen	No. of animals with no oocyst shedding (day postinfection)	% Reduction in mean oocyst sheddinga (day postinfection)	No. of animals (day 12 postinfection) with no intracellular parasites in:
					Ileum	Biliary tract
None	10	None	0		2	4
RM-6427	10	200 mg/kg/day/8 days	6 (8)b	90.5 (8)e	5f	8f
RM-6427	10	400 mg/kg/day/8 days	6 (8)b	92.0 (8)e	7i	10k
Nitazoxanide	14	200 mg/kg/day/12 days	2 (8)c	55.0 (8)b	11j	9l
Paromomycin	14	100 mg/kg/day/12 days	2 (8)c	17.6 (8)f	9f	9f
None	9	None	0		1	0
RM-6428	15	400 mg/kg/day/12 days	2 (12)d	45.5 (4)d	11g	11k
				53.6 (8)g
				67.0 (12)h

^a[(Mean oocyst count in treated animals)/(mean oocyst count in untreated animals)] × 100.

^bP < 0.001.

^cP = 0.0001.

^dP < 0.05.

^eP < 0.0001.

^fP > 0.05.

^gP < 0.01.

^hP < 0.02.

ⁱP = 0.067.

^jP = 0.011.

^kP < 0.002.

^lP < 0.005.

In another set of experiments, another agent (RM-6428) was tested in vivo. Gerbils were divided into two treatment regimen groups and received treatment for 12 consecutive days as indicated in Table 2. Cryptosporidium infection was assessed by using flow cytometry to measure oocyst shedding in feces collected for 24 h from each animal on days 4, 8, and 12 postinfection (38). Briefly, feces (stored in formalin) were vortexed for 30 s and allowed to settle for an additional 30 s. A 100 × 10⁻³-ml aliquot of the suspension was removed and added to 900 × 10⁻³ ml of PBS. The sample was vortexed and spun at 1,800 × g for 30 min to remove the formalin-containing supernatant. The pellet containing oocysts was stained with an oocyst-specific monoclonal antibody conjugated with fluorescein isothiocyanate (Crypt-a-Glo; Waterborne, New Orleans, La.). The mixture was incubated in the dark at 37°C for 30 min and then spun for 10 min at 1,800 × g. The supernatant was discarded, and the pellet was resuspended in 1 ml of sterile water. Samples were vortexed, transferred to a polystyrene tube, and stored at 4°C in the dark until analyzed on the same day with a flow cytometer (FACScalibur; Becton Dickinson Immunocytometry Systems, San Jose, Calif.). A negative control consisted of an oocyst-free stool sample processed as described above. Oocyst gating was obtained using a suspension of purified oocysts. To quantify oocyst numbers, calibrated fluorescent beads were added to each sample-containing tube before cytometry analysis (Flow-Count fluorospheres; Beckman Coulter, Fullerton, Calif.).

Histological study.

Histological examinations of distal ilea and biliary tracts were performed with animals which were killed on day 12 postinfection. Tissues were fixed in 10% formalin, cut, and embedded in paraffin. Four-micrometer sections were stained with hematoxylin-eosin-saffron and were considered infected if at least one cryptosporidial developmental form was microscopically observed within one mucosal epithelial cell.

Statistical analysis.

Comparisons between groups were performed using the Wilcoxon rank sums or Mann-Whitney tests for oocyst shedding or Fisher's exact test for intracellular parasites in ilea and biliary tracts. A P value of ≤0.05 was considered significant.

RESULTS

Evaluation of agent cytotoxicities for host cells.

In a nitroblue tetrazolium assay, percentages of the OD₄₅₀ value in agent-free control cultures of 0 to 5%, 6 to 25%, 26 to 50%, 51 to 75%, and 76 to 100% were assigned cytotoxicity scores of 0, 1, 2, and 3/4, corresponding to no toxicity or mild, moderate, and severe toxicities, respectively. At a 10-μg/ml concentration, genistein derivatives exerted no toxicity (RM-6429-31, RM-6433-35, RM-6447, and RM-6450-54), mild toxicity (RM-6415, -6427, -6442, -6445, and -6446), or moderate toxicity (RM-6403, -6411, -6424, -6425, -6428, -6442, -6445, and -6446) after a 76-h contact with BM cells. At the same concentration, they exerted no toxicity (RM-6411, -6415, -6429, -6446, -6447, -6451, -6453, and -6454) or mild toxicity (RM-6403, -6424, -6425, -6427, -6428, -6440, -6441, -6443, -6445, and -6448) after a 48-h contact with HCT-8 cells.

Inhibition of in vitro coccidian development by genistein analogs.

As shown in Table 3, 20 agents induced a maximum development inhibition (MI) of more than 95% for at least one parasite. Seventeen inhibited Neospora caninum with IC₅₀s ranging from 0.6 to 3.1 μg/ml, 13 inhibited Sarcocystis neurona with IC₅₀s ranging from 0.7 to 2.95 μg/ml, and 11 inhibited Cryptosporidium parvum with IC₅₀s ranging from 0.75 to 3.75 μg/ml. For five agents, i.e., RM-6403, RM-6425, RM-6427, RM-6428, and RM-6436, the MI was ≥95% for the three parasites.

TABLE 3.

Inhibitory activities of dihydroxyisoflavone and trihydroxydeoxybenzoin derivatives on Neospora caninum, Sarcocystis neurona, and Cryptosporidium parvum in vitro development

Agent	Result fora:
	Neospora caninum		Sarcocystis neurona		Cryptosporidium parvum
	IC50 (μg/ml)	MI (%)	IC50 (μg/ml)	MI (%)	IC50 (μg/ml)	MI (%)
RM-6403	2.1	97 ± 3.0	2.65	95 ± 1.6	1.25	98 ± 1.4
RM-6411	3.0	97.5 ± 2.1	4.25	59 ± 4.8	2.3	99.5 ± 0.3
RM-6424	1.9	96 ± 2.5	2.7	93 ± 2.3	2.4	98.8 ± 1.7
RM-6425	2.0	98 ± 3.8	2.6	95 ± 1.8	0.9	99.4 ± 0.8
RM-6426	2.5	98 ± 3.3	2.8	86 ± 2.8	3.75	96.5 ± 2.2
RM-6427	2.1	98.5 ± 2.1	1.9	99.7 ± 0.4	0.75	97 ± 2.6
RM-6428	1.8	100	2.0	95.3 ± 0.3	0.9	99 ± 1.4
RM-6430	1.9	100	0.8	96.9 ± 4.3	3.9	68.1 ± 3.9
RM-6431	2.9	100	1.3	100	6.9	52 ± 9.5
RM-6433	0.7	100	2.2	100	7.75	66.6 ± 7.8
RM-6434	0.9	100	2.0	90.8 ± 13.0	7.75	69 ± 10.2
RM-6435	0.6	100	0.7	99.6 ± 0.6	4.75	63.9 ± 6.7
RM-6436	1.6	100	0.9	100	1.8	96.9 ± 1.6
RM-6439	3.1	98.4 ± 2.3	2.9	100	5.75	77.6 ± 2.6
RM-6440	0.7	73 ± 6.3	2.75	100	0.8	98.1 ± 2.6
RM-6441	2.5	89.9 ± 5.2	2.1	100	3.4	97.1 ± 3.5
RM-6442	0.7	83.2 ± 5.3	1.5	73.3 ± 3.6	2.4	100 ± 0
RM-6443	3.0	100	2.4	100	0.85	90.1 ± 6.2
RM-6446	0.7	96.8 ± 2.4	0.55	93.9 ± 2.1	3.0	91.2 ± 3.3
RM-6448	2.1	98.4 ± 2.1	2.95	97.2 ± 1.8	2.85	83.4 ± 4.5

^aMI, maximum inhibition of Neospora caninum or Sarcocystis neurona merozoite or Cryptosporidium parvum form development. MI results are expressed as means ± standard deviations.

In vivo efficacy of agents RM-6427 and RM-6428 against Cryptosporidium parvum infection in immunosuppressed Mongolian gerbils.

Eight days after the initiation of therapy, RM-6427 at doses of 200 mg/kg of body weight/day and 400 mg/kg/day and NTZ resulted in statistically significant reductions of oocyst shedding compared to infected controls (P values of <0.0001, <0.0001, and 0.001, respectively). PRM at 200 mg/kg/day for 12 days was not significantly active (P = 0.1172). At 200 mg/kg/day, RM-6427 was more efficient than NTZ (200 mg/kg/day, P = 0.0035) and PRM (100 mg/kg/day, P = 0.0009). At both doses, RM-6427 totally inhibited oocyst shedding in 6/10 gerbils, compared with 2/14 for NTZ and PRM (Table 2). Results obtained from microscopic examinations showed that RM-6427 at the 200-mg/kg-dose level and PRM did not show statistically significant differences in the number of infected tissues from the ilea and the biliary tracts compared to control tissues. RM-6427 at the 400-mg/kg-dose level approached a statistically significant difference (P = 0.067) in the number of gerbils exhibiting infected ilea but reached significance for animals with infected biliary tracts (P = 0.0128) compared to tissues of untreated control gerbils. The number of infected tissues from ilea and biliary tracts were lower in the NTZ group than in the control group, with statistical significances of P values of 0.011 and 0.003, respectively (Table 2). For agent RM-6428 at 400 mg/kg/day for 12 days, oocyst shedding, assessed by flow cytometry, was decreased as early as 4 days after treatment was initiated (44.5%) and oocyst shedding inhibition increased from day 4 to day 12, the time at which gerbils were killed, to reach a 67% inhibition. Two out of 15 gerbils were totally free of oocysts in their feces at day 12, compared to 0/9 untreated animals (P > 0.05). Ilea and biliary tracts of gerbils treated by agent RM-6428 were found significantly less infected than controls (P values of <0.01 and <0.002, respectively [Table 2]). No histological alteration of mucosa was observed in FRM-6427-treated, FRM-6428-treated, or untreated infected gerbils.

DISCUSSION

In this work, 20/52 EGFR PTK inhibitor isoflavone analogs (38 dihydroxyisoflavone and trihydroxydeoxybenzoin and 14 dihydroxydeoxybenzoine substituted derivatives) were able to significantly and dose-dependently inhibit in vitro development of the apicomplexan parasites Sarcocystis neurona, Neospora caninum, and/or Cryptosporidium parvum. Five compounds exerted a >95% in vitro inhibition against the three parasites. In an immunosuppressed-gerbil model of cryptosporidiosis, two agents active in vitro decreased oocyst shedding and limited Cryptosporidium parvum intracellular development in the gut and in the biliary tract. Their EGFR PTK inhibitor activities (85% and 87% at a 100-mM dose for RM-6427 and RM-6428, respectively, compared with 97% for genistein [J. F. Rossignol, et al., unpublished data) are consistent with EGFR PTK being targets for isoflavone analogs.

Present results with different coccidia, host cells, and agents suggest common mechanisms of parasite development inhibition, although differential effects of the various analogs tested may also be consistent with different modes of action. Several lines of evidence suggest that isoflavone derivatives act directly on parasites. In the present study, the cell line type did not interfere with the efficacy of agents active against more than one coccidium, and for Cryptosporidium parvum, in vitro activities of RM-6427 and RM-6428 were confirmed in vivo. For Cryptosporidium parvum-infected HCT-8 cell cultures treated with effective agents, no early Cryptosporidium parvum developmental stage was seen, which suggests no interference with host cell signaling and/or cytoskeletal rearrangement (data not shown). For Sarcocystis neurona, it was previously shown that host cell cytoskeletal rearrangement was not crucial for intracellular development in vitro, since Sarcocystis neurona develops freely in host cell cytoplasm without a parasitophorous vacuole (14). Potential coccidial targets are microneme proteins with EGF-like domains which are sequentially released at initial contact and establish physical interaction with host cells. The microneme-associated protein NcMIC3, found in Neospora caninum tachyzoites and bradyzoites, contains a stretch of four consecutive EGF-like domains, and its secretion onto the surface resulted in their outward exposure, associated with increased capacity of Neospora caninum to adhere to host cells in vitro via interaction with host cell surface receptors (32, 33). In Toxoplasma gondii tachyzoites, microneme proteins exhibit adhesive (including EGF-like) domains which carry out important functions related to intracellular protein-protein interactions, and the adhesin TgMIC3, recently characterized as a microneme protein, carries five overlapping EGF-like domains (10, 19, 29). For Eimeria tenella, a transmembrane microneme protein (EtMIC4) carrying a large extracellular domain including EGF-like sequences has been described and several overlapping expressed sequence tags are predicted to encode a protein carrying an EGF-like motif and a stretch of tyrosine residues (36). Stretches of tyrosine residues have also been previously identified for the microneme protein GP900 of Cryptosporidium parvum (4).

Alternatively, direct effects on host cells may be involved. A common feature of proteins which include EGF-like motifs is their role in extracellular functions such as ligand-receptor adhesive interactions (8). It has been hypothesized that the EGF-EGFR axis, which belongs to one family of gastrointestinal growth peptides, was involved in initial attachment and cytoskeletal alterations associated with cellular invasion of microbial organisms, including apicomplexan sporozoites/merozoites (6). Prior incubation of mammalian cells with EGF was shown to significantly inhibit invasion by Trypanosoma cruzi, and administration of EGF at the apical surface of enterocytes significantly reduced Cryptosporidium parvum epithelial colonization independently of a direct anticoccidial effect, further suggesting a role for the EGF-EGFR axis in the infection process (6, 40). Apical EGF significantly reduced Cryptosporidium andersoni infection of human and bovine epithelial cells by inhibiting Cryptosporidium-induced apoptosis and disruption of zonula occludens 1, while not affecting oocyst viability (7).

Other observations suggest integrated anticoccidial effects and host cell modulation. In Theileria parva, signal transduction processes within both sporozoites and host cells were required for infection (34). Interaction of intracellular protozoa with host cells appears to occur by involving transducing signal mechanisms (6). A rapid onset of PK activities following sporozoite attachment was previously reported for Toxoplasma, and it was shown that the parasite induces phosphorylation of host cell molecules (17, 31). For malaria parasites, it was shown that protein phosphorylation within the merozoite also played an important role in the internalization step of invasion of host erythrocytes (41). Parasite signaling and cytoskeletal reorganization are likely to be involved in establishing infection, as reported for Theileria parva, in which signal transduction processes within both host cells and sporozoites are required for infection (18, 34). EGF-like peptides are possible mediators of cell-cell communication that occurs between eukaryotic parasites and hosts (30). For Toxoplasma gondii, it has been shown that EGFR PTK inhibitors interfere with both attachment and penetration of tachyzoites in macrophages with rapid onset of PK activities and that the parasite induces phosphorylation of host cell molecules (17, 31). In a recent study, the PK inhibitor genistein significantly inhibited in vitro Cryptosporidium parvum infection, which provides indirect evidence for parasite-induced responses in host cells with activation of phosphoinositide 3-kinase activity via TK growth factor receptors (18). Sporozoite attachment in the absence of further development was observed after pretreatment with PK inhibitors, suggesting that attachment is an event prefatory to host cell responses, and kinase activity was rapidly induced by sporozoite attachment, suggesting a role for signal transduction events in early parasite-host cell dynamics.

Present data are consistent with the extensive use of the EGF-EGFR axis by apicomplexan parasites and the potential capacity for EGFR inhibitors to alter parasite interactions with host cells and interfere with the infection process. They prompt further investigation to more fully define the molecular mechanisms of activity, especially in vivo, since most TK inhibitors have served so far as tools to set up in vitro assay systems and only a few have exhibited in vivo efficacy. Investigation of potential in vivo interferences with host kinases is especially needed. Information is also required on inhibitor absorption following oral administration and the active metabolites in relation to activity on biliary parasite sequestration, by reference to NTZ, which is known to be partially absorbed from the gastrointestinal tract and metabolized in blood to form active metabolites which have been identified in the bile (5, 35). To our knowledge, this is the first report suggesting that isoflavone analogs have the potential for further development as anticoccidial therapy.