SAR Studies of 5-Aminopyrazole-4-carboxamide Analogues as Potent and Selective Inhibitors of Toxoplasma gondii CDPK1

Abstract

We previously discovered compounds based on a 5-aminopyrazole-4-carboxamide scaffold to be potent and selective inhibitors of CDPK1 from T. gondii. The current work, through structure–activity relationship studies, led to the discovery of compounds (34 and 35) with improved characteristics over the starting inhibitor 1 in terms of solubility, plasma exposure after oral administration in mice, or efficacy on parasite growth inhibition. Compounds 34 and 35 were further demonstrated to be more effective than 1 in a mouse infection model and markedly reduced the amount of T. gondii in the brain, spleen, and peritoneal fluid, and 35 given at 20 mg/kg eliminated T. gondii from the peritoneal fluid.

Toxoplasma gondii is an apicomplexan protozoan that infects humans and domesticated animals. Infection by T. gondii presents major health concerns in immunocompromised patients.¹ Current therapies have some shortcomings,² thus new and effective therapy for treating T. gondii infection is needed.

The life cycle of T. gondii is regulated in part by a family of calcium-dependent protein kinases (CDPKs) that have no direct homologues in humans.^3,4T. gondii CDPK1 (TgCDPK1) has been shown both genetically and chemically to be critical for parasite reproduction, specifically in the process of cell invasion and egress.^3,5−7 Furthermore, TgCDPK1 contains a glycine residue at the “gatekeeper” position, which is extremely rare for mammalian kinases.⁸⁻¹⁰ The critical role of CDPK1 in combination with its unique active site gatekeeper makes this kinase an attractive target for antiparasitic drug discovery.^{3,8,9,11−13} Three different chemical scaffolds of CDPK1 inhibitors have been reported previously: the pyrazolopyrimidine (PP) scaffold,¹⁴⁻¹⁷ the acylbenzimidazole scaffold,^18,19 and the 5-aminopyrazole-4-carboxamide (AC) scaffold.²⁰ Overexpression of CDPK1 with a methioinine replacing the glycine mutant conveys resistance to certain inhibitors based on the PP scaffold.⁸ However, it should be recalled that T. gondii is not transmitted from person to person (except in rare congenital acquisition), so that should resistant parasites arise, they would be confined to the treated individual. Preliminary structure–activity relationship (SAR) studies of the AC scaffold found a promising inhibitor (1, Table 1) which inhibited TgCDPK1 with an IC₅₀ value of 2.0 nM and exhibited 0.22 μM EC₅₀ in the T. gondii cell assay.²⁰ However, its relatively low solubility, short half-life, and relatively high oral clearance could limit its further development. In this report, we describe further SAR studies of the AC scaffold based on compound 1, which resulted in the discovery of compounds with improved aqueous solubility, potency, and/or pharmacokinetic (PK) properties.

Table 1. SAR Study of N1-Substitution.

In the previous study most variations at N1 position were carried out with Ar₁ (on C3) being fixed as 2-naphthyl group, and for analogues containing other heteroaryl groups C3, only t-butyl analogues at N1 position were made.²⁰ Therefore, we reinvestigated a larger panel of N1 substitutions, while first fixing a 7-ethoxyquinolin-3-yl group at C3 (Table 1). Compounds with aliphatic chains (2–5) or rings (6–11) at the N1 position did not improve enzyme potency over 1. Hydroxyl aliphatic chains were also introduced to decrease lipophilicity of compounds, resulting in compounds (13–16) with comparable activity relative to the hydroxyl-free compounds. Though the potency did not improve during the investigation of the N1 substitutions, some of the substituents, especially the hydroxyl aliphatic chains, improved the solubility of compounds and provided the opportunity to optimize the PK properties. Additional study of N1 substitutions was conducted with a 2-ethoxyquinolin-6-yl group at C3. All the compounds (17–22) exhibited low nanomolar potency against TgCDPK1, indicating that the quinolin-6-yl moiety was as well accommodated in the hydrophobic pocket as the quinolin-3-yl. These two Ar₁ alternatives may lead to potent compounds with different pharmacological properties.

Because the t-butyl group seems to be the optimal N1 substitution, it was fixed at the N1 position, while we investigated the C3 substituent (Table 2) with the goal to optimize PK properties. Replacement of the ethoxy group on the 7-position of the quinolin-3-yl with fluorine, chlorine, trifluoromethyl, or alkane groups all produced decreased potency (23–26). Coming back to the alkoxy substitutions, small sized chain- or ring-alkoxy groups were potent inhibitors (28, 31, 34, and 36), while the trifluoromethoxyl led to decreased activity. Similarly, the quinolin-7-yl analogues with 2-cyclopropoxy and 2-cyclobutoxy substitutions (35 and 37) were also potent for TgCDPK1 inhibition.

Table 2. SAR Study of C3-Substitution.

The consistent binding pose of the AC scaffold in the presence of chemical elaboration at Ar₁ and R₂ was confirmed by crystal structures of selected compounds bound to TgCDPK1. In detail, the amine group and the oxygen atom of the amide group make hydrogen bonds to the hinge region of TgCDPK1, projecting the Ar₁ group at the C3 position into the hydrophobic pocket adjacent to the gatekeeper residue and the R₂ group at the N1 position into the ribose-binding pocket. In particular the bound position of the Ar₁ quinoline is not significantly perturbed by the position of the ring N, but its precise orientation is constrained by the need to accommodate a distal cyclopropyl (compounds 34 and 35) or cyclobutyl (compounds 36 and 37) between TgCDPK1 residues Leu 114 and Leu 198 (Figure 1). Co-crystallizations and structure determinations were performed as previously described.^14,20 Crystal structures showing the bound state of compounds 1 and 37 have been deposited in the PDB (entry codes 4ONA and 4YJN).

Figure 1.

Superposition of crystal structures of TgCDPK1 in complex with 1 (gold) and 37 (blue).

A select group of potent inhibitors were evaluated in a variety of assays to assess aqueous solubility (see Supporting Information), selectivity over mammalian kinases, potency against parasite cells, cytotoxicity to mammalian cells (Table 3), and, for select compounds, PK properties after oral (PO) dosing in mice (Table 4).

Table 3. Comprehensive Evaluation of Top Inhibitors^a.

compd number	SRC IC50 (μM)	T. gondii EC50 (μM)	cytotoxicity (μM)	compd number	SRC IC50 (μM)	T. gondii EC50 (μM)	cytotoxicity (μM)
1	>30	0.22 ± 0.09	>30	21	ND	0.32 ± 0.02	ND
2	>30	0.72 ± 0.10	NDb	22	>10	0.21 ± 0.04	>40
6	>10	1.21 ± 0.14	>40	23	>30	2.25 ± 0.94	>40
7	>10	0.52 ± 0.003	>40	28	>30	0.40 ± 0.04	>40
8	>10	0.97 ± 0.40	>40	30	ND	1.21 ± 0.42	>40
9	>10	0.33 ± 0.07	ND	31	>10	0.69c	>40
11	>10	0.41 ± 0.0	>40	32	>10	1.47c	>40
16	>10	1.19 ± 0.14	>40	34	4.75	0.42 ± 0.10	>40
17	>10	0.20 ± 0.09	>40	35	>10	0.089 ± 0.044	>40
18	3.3–10	0.14 ± 0.003	>40	36	>10	0.48 ± 0.09	>40
19	>10	0.48 ± 0.07	>40	37	>10	0.48 ± 0.06	>40
20	>10	0.43 ± 0.01	>40

^aSelectivity at enzyme level was conducted using a human kinase (SRC) that has a small gatekeeper residue; efficacy against parasites was performed using a T. gondii growth assay,¹⁵ values shown are the average ± standard deviation from two or more experiments; and cytotoxicity to mammalian cells was measured using the CRL-8155 cell line. Experimental details are provided in the Supporting Information.

^bND: not determined.

^cValue from one experiment.

Table 4. PK Profiles for Select Inhibitors^a.

compd	Cmax (μM)	Tmax (min)	AUC (μM·min)	CL (mL/min/kg)	T1/2 (min)
1	9.9 ± 2.9	60 ± 52	2428 ± 277	11.8 ± 0.5	88 ± 19
9	4.6 ± 1.0	80 ± 35	1608 ± 182	16.4 ± 1.5	204 ± 34
17	5.3 ± 0.6	80 ± 35	1694 ± 379	16.7 ± 3.8	134 ± 28
22	2.4 ± 2.0	30 ± 17	140 ± 97	186 ± 75	<30
32	7.1 ± 0.6	140 ± 92	3713 ± 1073	7.1 ± 0.7	549 ± 131
34	23.4 ± 3.1	130 ± 17	18450 ± 470	1.4 ± 0.04	290 ± 48
35	3.2 ± 0.7	160 ± 69	2676 ± 784	10.7 ± 1.0	559 ± 209
36	4.1 ± 1.5	30 ± 0	458 ± 206	57 ± 4	42 ± 17

^aMice dosing at 10 mg/kg, po. Reported as average ± standard deviation of measurements from three mice.

As a preliminary evaluation of the selectivity of the AC analogues, the top compounds were tested for inhibition of the human kinase SRC that contains a small gatekeeper residue using reported procedures.¹⁶ As shown in Table 3, none of the top compounds except compounds 18 and 34 displayed significant inhibition of SRC activity at concentrations up to 10 μM. Though compound 34 showed an IC₅₀ of SRC inhibition at 4.75 μM, the selectivity of TgCDPK1 over SRC was more than 450-fold.

We investigated the efficacy of compounds against T. gondii cells (Table 3). With 7-ethoxyquinolin-3-yl at C3 position, the modification of the t-butyl group merely resulted in reduced potency compared to 1. However, the C3 2-ethoxyquinolin-6-yl compounds were generally more potent than their 7-ethoxyquinolin-3-yl analogues, while compounds 18, 17, and 22 exhibited comparable potency to 1. Modification of the 7-ethoxy group of quinolin-3-yl moiety resulted in compounds mostly with lower potency. Even the best compounds with this modification, the cyclopropoxy (34) and cyclobutoxy (36) substituted analogues, showed at least 2-fold less potency compared to 1. Interestingly, for the last two quinoline pairs in Table 3, while the cyclobutoxy substituted pairs (37 and 36) showed similar potencies, the cyclopropoxy substituted pairs (35 and 34) were quite different, as 35 was ∼5-fold more potent than 34. Compound 35 was the most potent inhibitor found in this study, with increased activity against parasites compared to 1. As an indicator of selectivity at cellular level, we measured the cytotoxicity of these inhibitors (Table 3) using human lymphocyte cell line (CRL-8155).¹⁶ These compounds showed no significant inhibition of cell growth at concentrations up to 30 or 40 μM.

Select compounds were subjected to PK studies in mice following oral administration at 10 mg/kg (Table 4). Compounds 9 and 17 had a higher oral clearance than compound 1 resulting in lower maximum concentration (C_max) and total exposure (area under the curve, AUC) compared to 1. This is likely due to lower bioavailability of compounds 9 and 17 as their half-lives were longer than that of compound 1. Compound 22 had greatly improved solubility but had nearly 16-fold higher oral clearance than compound 1 resulting in significantly diminished C_max and AUC in comparison to 1. Compound 32 exhibited improved PK profile though the potency is relatively poor. Other analogues with small alkyl rings on quinoline oxygen are 34 and 36 and their quinolin-6-yl counterparts 35 and 37. While the cyclobutoxy compound 36 with poor solubility did not show improvement on PK properties, the cyclopropoxy compound 34 with improved solubility did show a significantly improved PK profile. The oral clearance of compound 34 (1.4 mL/min/kg) was the lowest of the tested compounds and about 8-fold lower than that of compound 1 (11.8 mL/min/kg). This decrease in CL was associated with a 3-fold longer half-life and 2.3-fold higher C_max. The exposure of compound 34 was 8-fold higher (Figure 2) in comparison to 1 likely due to both better bioavailability and lower systemic clearance (Figure 2). The PK profile of 35 was similar to 1 in terms of AUC and oral clearance (10.7 mL/min/kg). However, 35 is still promising considering its much longer T_1/2 (559 min) compared to 1 (88 min) and 34 (290 min) and its high potency against T. gondii (EC₅₀ = 0.089 μM). The human plasma protein binding properties of these three compounds were also measured, and all three compounds showed low protein binding with unbound fractions of 57% for 1, 84% for 34, and 88% for 35. Based on the overall properties, 34 and 35 were selected for testing in a mouse model of T. gondii infection in comparison to 1. All the three compounds get into the CNS, as estimated by a single time point, comparing plasma with brain tissue concentrations, where we found a brain/plasma ratio of 15.8% for 1, 16% for 34, and 43% for 35.

Figure 2.

Comparison of oral PK curves of compounds 1, 34, and 35.

In vivo efficacy of 34 and 35 was tested using a model of acute T. gondii infection with the Type I RH strain that is uniformly fatal in mice. Treatment was started 2 days after T. gondii inoculation to allow establishment of a robust systemic infection.²¹ Compounds 1, 34, and 35 were administered via oral (po) dosing for 5 days. Compound 1 significantly reduced T. gondii infection (Figure 3A). However, 34 and 35 were much more efficacious than 1, reducing T. gondii in the peritoneal fluid and spleen to near-undetectable levels at comparable or lower does than 1 (Figure 3B,C). Compound 34, given at 50 mg/kg and 20 mg/kg, reduced the mean concentration of T. gondii in the peritoneal fluid by 99.6% and 93.4%, respectively. Compound 35, given at 20 mg/kg, decreased T. gondii below the limits of detection (<100 parasites/mL). Similarly, 34 and 35 reduced the mean number of T. gondii in the spleen to less than 1% of controls. Both compounds also demonstrated more than 5-fold reduction in the mean concentration of detectable T. gondii DNA in the brain, suggesting that either both 34 and 35 achieve brain concentrations during active infection that are sufficient to inhibit T. gondii replication or 34 and 35 prevent brain infection from progressing by inhibiting systemic toxoplasmosis. None of the treated mice showed signs of infection or toxicity such as weight loss, altered grooming, or lethargy. However, one mouse in the 35, 50 mg/kg group died on the fifth day of treatment, but the cause of death was not established. Based on the elimination of T. gondii in the peritoneal fluid by 35 and the trends of better efficacy in the brain and spleen, it appears that 35 is more potent than 34. These experiments demonstrate that 34 and 35 are highly active against fulminant experimental toxoplasmosis.

Figure 3.

Efficacy of 1, 34, and 35 evaluated by measurement of T. gondii in peritoneal fluid, spleen, and brain. Mice were analyzed 1 day after the last dose. Groups consisted of 4 mice. (A) Efficacy of compound 1, peritoneal parasite count. Due to its short half-life 1 was dosed twice daily. The higher dose reduced parasites by >10-fold compared to controls. Efficacy of compounds 34 and 35 (single daily dose): (B) peritoneal parasite count, (C) qRTPCR of spleen, and (D) qRTPCR of brain. The differences between all treatment groups and their respective controls and the difference between 34 and 35 on peritoneal T. gondii infection were statistically significant, while the differences between all other compared treatment groups were not statistically significant using a p-value of <0.05. Bars represent the mean and the standard error of the mean. PEG = polyethylene glycol; mpk = mg/kg; qRTPCR = quantitative real-time PCR. *Three mice in 35, 50 mg/kg group were analyzed.

Additional studies were performed using these two lead compounds 34 and 35 to address (1) the selectivity of AC series analogues beyond the single human kinase SRC in the preliminary study and (2) if the inhibitors acted on-target in parasite cells. For selectivity, we performed a fluorescence based binding inhibition assay as previously described²² using a panel of ∼80 diverse human kinases.²³ Compound 34 did not show any IC₅₀ below 1 μM. Compound 35 only showed IC₅₀s below 1 μM for three of the 80 kinases (0.2–0.5 μM). This indicated that the top two compounds are quite selective against TgCDPK1. A table of the binding inhibition results covering a selected group of 20 kinases is provided in the Supporting Information.

To study if the inhibitors acted on-target in parasite cells, we compared parasite growth inhibition against the parental cell line, a cell line transfected with a TgCDPK1 expression plasmid, and a cell line transfected with a similar expression plasmid bearing CDPK1 with a G128M mutation at the gatekeeper position as described previously.⁸ The results (see Supporting Information for additional figures) showed that neither 34 nor 35 produced significant shifts in parasite growth inhibition at low compound concentrations. However, these two compounds could not completely suppress parasite growth at high concentrations in the cell line containing the G128M mutant of TgCDPK1. This suggests that the two leads only acted partially on-target, and therefore, testing AC series inhibitors with weaker IC₅₀ against TgCDPK1 may still lead to potent inhibitors at the parasite level. Additional study will be required to identify other potential cellular targets of the AC series inhibitors.

In summary, we described the structural optimization of 5-aminopyrazole-4-carboxamide derivatives as TgCDPK1 inhibitors based on compound 1. We evaluated over 30 analogues as potent and selective TgCDPK1 inhibitors. Most of the compounds showed low nanomolar activity against the enzyme. These compounds also showed good efficacy against T. gondii cells, with the best compound 35 exhibiting an EC₅₀ of 89 nM. Furthermore, no compound showed significant toxicity to mammalian cells at concentrations up to 30 μM. Compounds 34 and 35 also exhibited improved solubility and good pharmacokinetic properties. Compared to compound 1, both 34 and 35 had improved efficacy in a mouse model of toxoplasmosis. These studies provide new lead compounds for further development of drugs for toxoplasmosis therapy.

Glossary

ABBREVIATIONS

AC

5-aminopyrazole-4-carboxamide

BKI

bumped kinase inhibitor

CDPK1

calcium-dependent protein kinase-1

PK

pharmacokinetic

PP

pyrazolopyrimidine

SAR

structure–activity relationship

Tg

Toxoplasma gondii

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.5b00319.

• Synthesis and characterization data of compounds, biological assays, and structural studies (PDF)

Research reported in this publication was supported by the National Institute of Allergy and Infectious Diseases and National Institute of Child Health and Human Development of the National Institutes of Health under award numbers R01AI089441, R01AI111341, and R01HD080670. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This work was also supported in part by Award #2014-06183 from the United States Department of Agriculture and by Career Development Award #BX002440 from the United States Department of Veterans Affairs Biomedical Laboratory Research and Development. Crystallography performed in support of the work benefitted from remote access to resources at the Stanford Synchrotron Radiation Lightsource supported by the U.S. Department of Energy Office of Basic Energy Sciences under Contract No. DE-AC02-76SF00515 and by the National Institutes of Health (P41GM103393).

The authors declare no competing financial interest.