Phthalazinone Inhibitors of Inosine-5′-Monophosphate Dehydrogenase from Cryptosporidium parvum

Abstract

Cryptosporidium parvum (Cp) is a potential biowarfare agent and major cause of diarrhea and malnutrition. This protozoan parasite relies on inosine 5′-monophosphate dehydrogenase (IMPDH) for the production of guanine nucleotides. A CpIMPDH-selective N-aryl-3,4-dihydro-3-methyl-4-oxo-1-phthalazineacetamide inhibitor was previously identified in a high throughput screening campaign. Herein we report a structure-activity relationship study for the phthalazinone-based series that resulted in the discovery of benzofuranamide analogs that exhibit low nanomolar inhibition of CpIMPDH. In addition, the antiparasitic activity of select analogs in a Toxoplasma gondii model of C. parvum infection is also presented.

Cryptosporidium species cause self-limiting gastrointestinal disease that can be chronic and fatal in immunosuppressed patients ¹. Oocysts from these organisms are resistant to common methods of water treatment, and therefore pose a potential biowarfare threat. The drugs currently approved for the treatment of cryptosporidiosis are largely ineffective, elevating the need for new chemotherapeutic agents to counter outbreaks. Genomic analysis of the most common human pathogens, C. parvum and C. hominis, revealed that these parasites have a streamlined purine salvage pathway that relies on IMP dehydrogenase (IMPDH) for the production of guanine nucleotides ^2–4. This enzyme catalyzes the oxidation of IMP to XMP with the production of NADH as outlined in Scheme 1⁵. Curiously, the Cryptosporidium IMPDH gene appears to have been obtained from an ε-proteobacterium via horizontal gene transfer ⁶, and is structurally distinct from the host ortholog (the C. parvum and C. hominis enzymes are identical and will be denoted as CpIMPDH). Consequently, CpIMPDH displays unique enzymatic properties compared with the host counterparts ⁷.

Scheme 1.

IMPDH catalyzed conversion of IMP to XMP via intermediate E-XMP*.

A previously reported high throughput screening (HTS) campaign identified several structurally distinct classes of selective CpIMPDH inhibitors that bind to the NAD site ^{8, 9}. Structure-activity relationship studies of triazole ¹⁰, benzimidazole ^{9, 11} and urea ¹² CpIMPDH inhibitors, exemplified by A110, C64, C97 and P96, respectively, have also now been reported (Figure 1). In addition, a co-crystal structure of the benzimidazole inhibitor C64 with CpIMPDH demonstrated a unique binding mode compared to inhibitors bound to human IMPDH ⁹. The phthalazinone-based inhibitor 1 was also identified by HTS, with moderate potency (IC₅₀ = 5.4 μM) and good selectivity (IC₅₀ > 50 μM versus both human IMPDH1 and IMPDH2) ⁸. Compound 1 (50 μM) did not display antiparasitic activity ⁸. Herein we report a structure-activity relationship study of 1, and the antiparasitic activity of select analogs in a T. gondii model of C. parvum infection.

Figure 1.

Optimized CpIMPDH inhibitors A110, C64 and C97 and HST identified CpIMPDH inhibitor 1.

Compound 1 and its analogs were prepared following a four-step sequence as depicted in Scheme 2 beginning with the Wittig olefination of phthalic anhydride, 2, in the presence of carbethoxymethylidenetriphenylphosphorane in chloroform to produce (E)-ethyl-2-(3-oxoisobenzofuran-1(3H)-ylidene)acetate, 3. Addition of hydrazines in refluxing ethanol generated the phthalazinone acetoesters 4. Hydrolysis of the esters produced phthalazinone acetic acids 5 ^{13, 14}. Finally, coupling of the acetic acid moiety with various aromatic amines in DMSO afforded 6.

Scheme 2.

General procedure for the synthesis of phthalazinone derivatives of 1. a) Ph₃PCHCO₂Et, CHCl₃, reflux, 16 h. b) NH₂NH₂ or NH₂NHMe, EtOH, reflux, 3 h. c) 3 M NaOH/THF, reflux, 2 h followed by acidification. d) R¹-NH₂, EDC, HOBt, DIPEA, DMSO, 3–5 h.

The synthesized derivatives were evaluated for CpIMPDH inhibition using an assay that monitored the production of NADH in the presence of varying inhibitor concentrations ¹⁰. In order to evaluate the effects of non-specific binding, inhibition was also determined in the presence of 0.05% fatty acid free bovine serum albumin (BSA). None of the compounds inhibited human IMPDH2 (IC₅₀ > 5 μM).

Inhibitor 1 exhibited an IC₅₀ of 1.0 μM upon resynthesis. Gratifyingly, the activity was retained in the presence of 0.05% BSA (IC₅₀ = 970 nM) as shown in Table 1. However, moving the electron donating group to the 2- or 3-position resulted in loss of activity (7 and 10). Replacing the 4-OMe with a variety of other electron donating substituents (11 – 14) was also detrimental. Likewise, replacing the phenyl with an amine (8) or alkyl group (9) or inserting a methylene between the benzene and amide nitrogen atom (15) was not tolerated. Replacement of the electron donating group at the 4-position with electron withdrawing groups improved potency. For example, the 4-chloro and 4-bromo analogs (16 and 17) demonstrated IC₅₀ values of 230 and 130 nM, respectively. However, other electron withdrawing groups (18 – 21) generally were not as well tolerated. Also, transposition of the chlorine from the 4-position to the 3-position (23) resulted in loss of activity. Di-substitution at the 3,4-positions with chlorine and bromine atoms (24 and 25) resulted in increased potency. Furthermore, other 3,4-di-substituted analogs with either an electron-donating or an electron-withdrawing group at the 3-position were also active. Interestingly, in a number of cases substituents at the 3-position that alone were not tolerated when combined with electron withdrawing groups (especially Cl or Br) at the 4-position resulted in potent CpIMPDH inhibitors (e.g. 26, 27 and 37). However, transposition of the substituents from the 3-position to the 2-position (39 – 42) resulted in poor inhibitors. In addition, several other di- and tri-substitutions (43 – 46) were also detrimental. Encouraged by the results of 3,4-disubstitution, the 2-naphthyl analog 47 was prepared and demonstrated an IC₅₀ of 63 nM. Similar increases in CpIMPDH activity by introduction of 3,4-disubstitution or incorporation of 2-naphthylene had also been observed in the structure-activity relationship studies resulting in A110 and C97 ^{10, 11}. Finally, analogs that lack methylation of the phthalazinone nitrogen atom (48 and 49) were found to be active indicating that this substituent is not necessary for CpIMPDH inhibitory activity.

Table 1. SAR of anilide and nathylide phthalazinone inhibitors.

Assays as described in ¹⁰.

Compound	R	R1	(−)-BSA (nM)	(+)-BSA (nM)
1	Me	4-OMePh	1000 ± 200	930 ± 180
7	Me	2-OMePh	>5000	ND
8	Me	NH2	>5000	ND
9	Me	t-Bu	>5000	ND
10	Me	3-OMePh	>5000	ND
11	Me	4-OEtPh	>5000	ND
12	Me	4-OCF3Ph	>5000	ND
13	Me	4-MePh	>5000	ND
14	Me	4-i-PrPh	>5000	ND
15	Me	CH2(4-OMePh)	>5000	ND
16	Me	4-ClPh	230 ± 50	280 ± 50
17	Me	4-BrPh	130 ± 30	130 ± 20
18	Me	4-FPh	>5000	ND
19	Me	4-CNPh	950 ± 90	940 ± 90
20	Me	4-CF3Ph	>5000	ND
21	Me	4-SO2MePh	>5000	ND
22	Me	3-CF3Ph	>5000	ND
23	Me	3-ClPh	>5000	ND
24	Me	3,4-di-ClPh	55 ± 6	95 ± 2
25	Me	3-Cl-4-BrPh	30 ± 1	75 ± 2
26	Me	3-CF3-4-BrPh	24 ± 2	48 ± 4
27	Me	3-CF3-4-ClPh	13 ± 1	25 ± 5
28	Me	3-CF3-4-CNPh	54 ± 5	52 ± 6
29	Me	3-Cl-4-CNPh	290 ± 40	470 ± 120
30	Me	3-CF3-4-FPh	150 ± 20	150 ± 30
31	Me	3-CF3-4-OMePh	>5000	ND
32	Me	3-CF3-4-NH2Ph	>5000	ND
33	Me	3-F-4-ClPh	200 ± 30	240 ± 30
34	Me	3-Me-4-ClPh	79 ± 9	96 ± 19
35	Me	3-Me-4-BrPh	70 ± 10	140 ± 10
36	Me	3-Me-4-CNPh	400 ± 100	460 ± 90
37	Me	3-OMe-4-ClPh	33 ± 1	40 ± 7
38	Me	3,4-di-CNPh	>2500	>2500
39	Me	2-CF3-4-ClPh	>5000	ND
40	Me	2-CF3-4-BrPh	>5000	ND
41	Me	2-F-4-BrPh	1100 ± 200	1700 ± 100
42	Me	2,4-BrPh	>5000	ND
43	Me	2,4-di-ClPh	>5000	ND
44	Me	3,5-di-ClPh	>5000	ND
45	Me	3,4,5-tri-ClPh	>5000	ND
46	Me	3,4,5-tri-FPh	>5000	ND
47	Me	2-naphthyl	63 ± 3	140 ± 20
48	H	3-CF3-4-ClPh	40 ± 1	61 ± 1
49	H	3-CF3-4-BrPh	32 ± 6	130 ± 8

All values are average of three independent determinations unless otherwise noted (* two determinations). ND, not determined.

Inspired by the results with 47, a variety of fused heterocycles were explored as replacements of the 2-naphthylene (Table 2). Although the 5-benzofuranyl analog 50 had reduced activity and the 5-benzoxazolyl analog 51 was inactive, the 2-methyl-5-benzofuranyl derivative 52 retained activity with an IC₅₀ of 64 nM. Furthermore, incorporation of an additional ring (53) resulted in enhanced potency. Unsaturation of the ring further increased activity with 54 demonstrating an IC₅₀ of 4 nM. However, the regioisomer 55 and two carbazoles 56 and 57 were not active.

Table 2. SAR of heterotricyclic anilide phthalazinone inhibitors.

Assays as described in ¹⁰.

Compound	R	(−)-BSA (nM)	(+)-BSA (nM)
50		93 ± 15	85 ± 12
51		>5000	ND
52		64 ± 11	72 ± 15
53		20 ± 5	150 ± 20
54		4 ± 1	30 ± 10
55		>5000	ND
56		>5000	ND
57		>5000	ND

All values are average of three independent determinations unless otherwise stated. ND, not determined.

Compounds with IC₅₀ values less than 30 nM were candidates for evaluation of antiparasitic activity in a Toxoplasma gondii model of C. parvum infection ¹⁵. In this model, the endogenous T. gondii IMPDH and hypoxanthine-guanine-xanthine phosphoribosyltransferase genes have been knocked out and the CpIMPDH gene inserted to create T. gondii/CpIMPDH, a model parasite that relies on CpIMPDH for the production of guanine nucleotides. Both wild-type and T. gondii/CpIMPDH were cultured in human foreskin fibroblasts immortalized with hTERT, so this assay also reports on host cell toxicity. Compounds 26, 27, 53 and 54 all displayed sub-micromolar activity against T. gondii/CpIMPDH (Table 3). However, only 27 displayed selectivity • 30 versus the wild-type strain, strongly indicating that antiparasitic activity results from the inhibition of CpIMPDH.

Table 3. Antiparasitic activity.

Assays as described in ¹⁵.

Compound	EC50a (μM)		Selectivity b
	Toxo/WT	Toxo/CpIMPDH
26	4 ± 3 *	0.3 ± 0.2	14
27	23 ± 4	0.4 ± 0.3	60
53	1 ± 1 *	0.5 ± 0.2	2
54	3 ± 1	0.4 ± 0.1	7

Values are the average and standard deviations of three independent determinations unless otherwise stated

^*denotes average and range of two determinations).

^aT. gondii RH (Toxo/WT) should be resistant to CpIMPDH inhibitors while T. gondii/CpIMPDH (Toxo/CpIMPDH) should be sensitive.

^bSelectivity = EC₅₀(Toxo/WT)/EC₅₀(Toxo/CpIMPDH).

The inhibition of CpIMPDH by compound 27 was characterized further. Whereas compound 1 is a mixed inhibitor of CpIMPDH with respect to NAD⁺ (K_is = 1.8 μM, K_ii = 7μM) ⁸, 27 is a pure noncompetitive inhibitor (K_is = K_ii = 3.4 ± 0.2 nM; Figure 2).

Figure 2.

Inhibition of CpIMPDH by 27.

Compound 27 displayed good stability in mouse liver microsomes (T_1/2= 79 min). This compound was advanced into the IL12 knockout mouse model of C. parvum infection ^16–19. Disappointingly, no antiparasitic activity was observed with once per day oral dosing of 27 (250 mg/kg) for seven days. It may be that alternative dosing regimes or modifications in formulations could improve efficacy in vivo. Additional optimization of pharmacokinetic properties may also be necessary for this compound series in order to achieve in vivo efficacy.

In conclusion, a SAR study of phthalazinone-based CpIMPDH inhibitors revealed that expansion of the aniline ring could increase inhibitory activity, while maintaining selectivity relative to the human orthologs. The phthalazinone-based CpIMPDH inhibitors described herein could serve as lead compounds for the development of effective treatments of cryptosporidiosis as well as useful molecular probes for studying Cryptosporidium parasites.