5-Aminopyrazole-4-Carboxamide-Based Compounds Prevent the Growth of Cryptosporidium parvum

ABSTRACT

Cryptosporidium parvum calcium-dependent protein kinase 1 (CpCDPK1) is a promising target for drug development against cryptosporidiosis. We report a series of low-nanomolar CpCDPK1 5-aminopyrazole-4-carboxamide (AC) scaffold inhibitors that also potently inhibit C. parvum growth in vitro. Correlation between anti-CpCDPK1 and C. parvum growth inhibition, as previously reported for pyrazolopyrimidines, was not apparent. Nonetheless, lead AC compounds exhibited a substantial reduction of parasite burden in the neonatal mouse cryptosporidiosis model when dosed at 25 mg/kg.

TEXT

Cryptosporidiosis is an important global health problem that causes diarrheal diseases in humans (1, 2). Cryptosporidium parvum-associated diarrhea is self-limiting in healthy patients but can be life-threatening in immunocompromised patients and malnourished children (1, 3, 4). Furthermore, Cryptosporidium has the potential to be used for bioterrorism because oocysts can be readily obtained and dispersed in food and water supplies and are immediately infectious on passage in feces (5). Nitazoxanide, the only FDA-approved drug for cryptosporidiosis in humans (1), is not effective in immunocompromised people and only weakly active in malnourished children, thus new drugs are needed (6, 7).

Previous studies suggested that chelation or depletion of intracellular calcium in C. parvum inhibited apical organelle discharge and decreased the gliding motility of sporozoites, which blocked parasite invasion of, but not attachment to, host cells (6). Several studies explored the activity of bumped kinase inhibitors (BKIs) on C. parvum calcium-dependent protein kinase 1 (CpCDPK1). Pyrazolopyrimidine (PP-BKIs) and 5-aminopyrazole-4-carboxamide (AC-BKIs) have been shown to block the activity of CpCDPK1 (7, 8), prevent the growth and proliferation of C. parvum in vitro (9), and reduce oocyst shedding in infected SCID beige mice (10, 11). Furthermore, BKIs 1 (compound 1294, PP), 2 (compound 1517, AC), and 3 (compound 1553, PP) were effective for treatment of clinical diarrhea and reduction of oocyst excretion in the neonatal calf C. parvum challenge model (Fig. 1) (9). Based on the cited studies, additional AC analogues were screened for activity against CpCDPK1 and inhibition of C. parvum growth. The anti-Cryptosporidium efficacies of select AC-BKIs (single 25-mg/kg dose) by gastric intubation were also evaluated in infected neonatal mice.

FIG 1.

Structures of BKIs 1, 2, and 3 and structures of select CpCDPK1 inhibitors (AC-BKIs).

New AC-BKIs (8, 11, 14, 17) were synthesized and characterized following previously reported procedures (8, 12, 13) (see Text S1 in the supplemental material). Methods for solubility testing, inhibition of CpCDPK1 and C. parvum cellular growth, drug pharmacokinetic and fecal concentration analysis, plasma protein binding, and determination of efficacy in neonatal mice were performed as previously described (9).

AC-BKI analogues with potent activity on CpCDPK1 (50% inhibitory concentration [IC₅₀], <10 nM) can be structurally divided into 5 groups (group 1, 2/4/5; group 2, 6/7/8; group 3, 9/10/11; group 4, 12/13/14; and group 5, 15/16/17) (Table 1; Fig. 1). AC-BKIs 4 and 5 in group 1 are close derivatives of AC-BKI 2, with C3 substitution replaced by 2-ethoxyquinolin-6-yl and 6-ethoxynaphthyl-2-yl, respectively. Groups 2 and 3 maintained the quinoline/naphthyl variations of group 1, with changes on the N1 substitution (cyclopropylmethyl for group 2 and cyclohexyl for group 3). Group 4, compared with group 1, contains the cyclopropyloxy variation appending to the fused aromatic ring. Other potent BKIs in group 5 contain various substitutions on both ends: AC-BKI 15 contains a longer tail substitution at the quinoline ring, AC-BKI 16 contains a hydroxyl aliphatic chain at the N1 position, and AC-BKI 17 contains cyclobutyl group on both ends.

TABLE 1.

Data for CDPK1 inhibitors

Compound no.	Compound	clogPa	CpCDPK1 IC50 (nM)	C. parvum EC50 (μM)	Src IC50 (μM)	CRL-8155 EC50 (μM)	hERG EC50 (μM)
2	1517	2.61	1.4b	0.65	>30b	>30b	>30
4	1597	3.20	2.3	0.41	>10b	>40b	>30
5	1473	3.44	3.6b	1.00	7.3b	>30b	NDc
6	1598	2.33	5.1	2.16	>10b	>40b	ND
7	1635	2.93	3.7	3.35	10b	>40b	ND
8	1779	3.16	0.03	1.03	>10	>80	ND
9	1571	3.35	4.3	14.85	>10b	>40b	30.0
10	1633	3.95	2.5	6.66	>10b	>40b	14.1
11	1776	4.18	0.3	3.29	>10	30	ND
12	1586	2.71	4.9	0.58	4.75b	>40b	22.8
13	1643	3.31	2.5	0.41	>10b	>40b	>30
14	1673	3.55	4.3	0.52	>10	>40	27.0
15	1566	3.13	3.1	>20	>10b	>40b	ND
16	1608	1.34	5.5	0.71	3.5	40	>30
17	1656	3.61	2.7	1.00	>10	>40	11.9
18	Paromomycin		>2,000	96.6	>10	>40	ND

^aCalculated log P using Instant JChem.

^bData were reported previously (8, 12, 13).

^cND, not determined.

AC-BKIs were assayed for inhibition of C. parvum growth and proliferation in human HCT-8 cells, with 50% effective concentrations (EC₅₀s) ranging from submicromolar to low micromolar (Table 1). We have shown that potencies of BKIs dramatically increased when added before host cell invasion versus after invasion (7). In this study, AC-BKIs were added before host cell invasion. Nine out of 15 compounds had EC₅₀s of ≤1 μM, which were lower than those of PP-BKIs, which are uniformly of low concentration (14). Compounds within each group exhibited similar C. parvum EC₅₀s. AC-BKIs in group 1 were all quite potent, with EC₅₀s ranging from 0.41 to 1.0 μM, with a <3-fold inhibition difference. Group 2 AC-BKIs were less potent than compounds in group 1, exhibiting low micromolar activities. Group 3 AC-BKIs exhibited EC₅₀s ranging from 3.3 to 14.9 μM, whereas group 4 showed comparable but less variable cellular inhibition than group 1, with EC₅₀s ranging from 0.41 to 0.58 μM. AC-BKIs in group 5 contained various substituents, and their inhibition was significantly different, ranging from 0.71 to >20 μM. Overall, there was poor correlation between CpCDPK1 enzyme inhibition and direct phenotypic effect on C. parvum growth and proliferation (Fig. 2). This could be due to compound differences in lipophilicity, level of ionization, solubility, penetration to the intracellular Cryptosporidium parasite (15), or alternative non-CDPK1 AC-BKI C. parvum target(s) (11, 12). A previous study in T. gondii suggested that AC-BKIs act only partially on TgCDPK1 targets (13). Mitogen-activated protein kinase (UniProtKB A3FQ38_CRYPI) and cyclic GMP-dependent protein kinase (UniProtKB Q5CWD4_CRYPI) are being investigated as alternative AC-BKI targets in C. parvum.

FIG 2.

Cell inhibition EC₅₀ versus CpCDPK1 IC₅₀ plot to determine correlation. Plot and analysis showing the poor correlation between CpCDPK1 IC₅₀ and C. parvum in vitro growth and proliferation EC₅₀s. Pearson's correlation coefficient (r) was determined and added to the graph: r ranges were between 1 and −1, with 0 indicating no correlation at all, 0.1 to −0.1 also indicating no correlation, <0.5 indicating weak correlation, and 0.5 to 1 indicating strong correlation. Closer to 1 is stronger correlation and closer to 0 is weaker correlation, so r = 0.07 indicates extremely weak to no correlation.

Overall, these AC-BKIs had relatively safe profiles with high selectivity for CpCDPK1 over human tyrosine kinase (Src), no detectable growth inhibition of CRL-8155 human lymphocyte cells (12), and, most often, a lack of human Ether-à-go-go-related gene product (hERG) inhibitory activity (13). Of the nine AC-BKIs tested, five did not show 50% inhibition of hERG at ≥30 μM, two (12 and 14) had hERG EC₅₀s between 20 and 30 μM, and two (7 and 10) had EC₅₀s between 10 and 20 μM (Table 1).

All AC-BKIs that exhibited EC₅₀s of <1 μM were selected to determine pharmacokinetic profiles in adult mice (Table 2). AC-BKIs 2 and 4 showed high maximum concentrations (C_max) and areas under the curve (AUC). AC-BKIs 12, 13, and 14, all of which contain a cyclopropoxy-substituted aryl group at the C3 position, exhibited the highest C_max and AUC values and fecal concentrations of >0.4 μM. AC-BKI 16 had low C_max and AUC values but the highest fecal concentration, at 1.06 μM. The human plasma protein binding properties of these compounds were also measured, and the results ranged from 74% to 89%.

TABLE 2.

Pharmacokinetic, human plasma binding, stool level, and neonatal mouse C. parvum efficacy results

Compound no.	Solubility pH 6.5 (μM)	Cmax (μM)	AUC (μM · min)	Human plasma binding (%)	Neonatal mice (25 mg/kg) (% reduction of infection)a	Fecal concentration (μM)
2	26.0b	9.9b	2,428b	74	39b	0.45
4	83.5b	5.3b	1,694b	89	36	0.05
7	10.9b	NDd	ND	ND	8	0.83
12	83.7b	23.4b	18,450b	84	37	ND
14	96.3b	12.9	9,098	82	56	0.42
16	>100c	2.7	339	86	53	1.06
17	8.7	ND	ND	86	24	ND
Paromomycin		ND	ND	ND	18	ND

^aDetermined by histopathologic scoring of intestinal infection levels at 92 to 94 h postinfection.

^bData were reported previously (8, 12, 13).

^cData at pH 7.4.

^dND, not determined.

All AC-BKIs that showed potency in the C. parvum growth inhibition assay (EC₅₀s, ≤1 μM), except AC-BKIs 5 and 13, were further tested for efficacy against C. parvum in the neonatal mouse model (Table 2). AC-BKI 7 (EC₅₀, 3.35 μM) was included as a negative control and exhibited a low 8% reduction of infection. The positive control, paromomycin, exhibited 18% reduction of infection. Five of the seven AC-BKIs tested (2, 4, 12, 14, and 16) exhibited >30% reduction of infection compared with the untreated control. There was a statistically significant difference (t test, P < 0.05) between the controls and the AC-BKIs.

AC-BKIs 2, 4, 12, and 14 showed high C_max and AUC levels. These four also reduced infection in the neonatal mouse model by >30%. However, AC-BKI 16, which exhibited low C_max and AUC levels in adult mice, also showed a >50% reduction of infection in the neonatal mouse model. This indicates that systemic exposure in adult mice may not be the best predictor of efficacy in neonatal mice, as low exposure does not eliminate efficacy (r = −0.18). The poor correlation may be due to differences in drug exposure in adult and neonatal mice (14). AC-BKIs 4 and 16 were very active against C. parvum in vitro, but AC-BKI 4 demonstrated lower in vivo efficacy than AC-BKI 16. AC-BKI 4 had higher systemic levels and much lower stool levels than AC-BKI 16 (Table 2). The emerging consensus, based on a recent study (14), is that higher fecal levels of AC-BKI 16 may be a good proxy for elevated intraepithelial concentrations in the gastrointestinal tract, the site of drug action against C. parvum, whereas low fecal levels of AC-BKI 4 may indicate lower intraepithelial levels and low in vivo efficacy. Nevertheless, a number of AC-BKIs exhibited efficacy in neonatal mice, which is a good predictor of efficacy in the calf model, and remain promising for the treatment of cryptosporidiosis in animals and humans (16).