The Activity of Dispiro Peroxides Against Fasciola hepatica

Abstract

Dispiro 1,2,4-trioxanes and 1,2,4,5-tetraoxanes had superior efficacy against Fasciola hepatica than the corresponding ozonides (1,2,4-trioxolanes). For highest efficacy, spiroadamantane and carboxymethyl substructures were required. Three compounds completely cured F. hepatica-infected mice at single oral doses of 50 mg/kg and two were partially curative at single doses of 25 mg/kg.

The liver flukes Fasciola hepatica and F. gigantica are pathogenic trematodes infecting an estimated 2.4–17 million people in the Andean countries, Cuba, Western Europe, Egypt and Iran.¹ Moreover, the morbidity and mortality of fascioliasis in cattle and sheep results in considerable economic loss.² The benzimidazole triclabendazole (Fig. 1) is the drug of choice used to treat veterinary fascioliasis, but it is registered in only four countries for the treatment of human fascioliasis.^3,4 Evidence of drug resistance to triclabendazole in veterinary medicine^5,6 provides an impetus for the discovery and development of new drugs against fascioliasis.

Figure 1.

We have shown that semisynthetic artemisinins (Fig. 1) and synthetic ozonides (Fig. 2) have good efficacy against F. hepatica.^7–10 It is postulated^10,11 that such peroxidic compounds possess antiplasmodial^12,13 and flukicidal activities because both plasmodia and Fasciola spp. degrade hemoglobin to generate free heme, a likely target¹⁴ of bioactive peroxides. In an effort to identify more effective synthetic peroxides, a structurally diverse ozonide library of OZ78 (cis-1a) analogues was recently studied.¹⁰ It was found that a spiroadamantane substructure, an acidic functional group (or ester prodrug), and the peroxide bond and non-peroxide oxygen atom of the ozonide heterocycle, were all required for high efficacy against F. hepatica.¹⁰ We now report an investigation of the 1,2,4-trioxane and 1,2,4,5-tetraoxane analogs of ozonides (1,2,4-trioxolanes) 1a–5a (Fig. 2).¹⁵ Target peroxide heterocycles 5a–5c were designed on the basis of the superior pharmacokinetic profiles of 8’-aryl ozonides compared to those of 8’-alkyl ozonides.¹⁶

Figure 2.

Target dispiro peroxides

Griesbaum coozonolysis¹⁷ of oxime ether 6¹⁸ and ketoester 7 afforded ozonide ester 8 as a 2.5:1 mixture of cis and trans isomers. Chromatographic separation into the individual isomers followed by hydrolysis of the cis isomer afforded 4a in high yield. Ozonides 1a–3a and 5a were obtained as previously described.^19–21

Acid-catalyzed condensation of β–hydroperoxy alcohols 9–11 with 2-adamantanone or cyclohexanone afforded trioxane esters 12–16 (20–56% yields) (Scheme 2) which were isolated as single trans isomers after crystallization (vide infra). Hydrolysis of 12–16 afforded trioxane acids 1b–5b. β–Hydroperoxy alcohols 9–11 were formed in 52–98% yields by regioselective perhydrolysis^22,23 of the corresponding epoxides, which in turn, were formed predominantly as their cis isomers²⁴ by treatment of their keto ester precursors^20,21 with the sulfur ylid formed from trimethylsulfoxonium iodide and potassium tert-butoxide.²⁵ Even though β–hydroperoxy alcohols 9–11 were formed as mixtures of cis and trans isomers, the trans isomers predominated as demonstrated by the triphenylphosphine reduction of 9 to its corresponding 1,2-diol 19 and NMR analysis of the latter (Scheme 3). Observation of a signal at 66.5 ppm in the ¹³C NMR spectrum of 19²⁶ is consistent with a shielded axial hydroxymethyl group indicating that the hydroperoxide in 9 is equatorial. Similarly, we suggest that the signal at 62.9 ppm in the ¹³C NMR spectrum of 9 is that of a shielded axial hydroxymethyl group. By way of comparison, Li et al.²³ report hydroxymethyl group ¹³C NMR signals at 61.6 and 67.2 ppm for the isomers of β–hydroperoxy alcohol 20. Finally, trioxane acid 1d was obtained by hydrolysis of trioxane ester 18; the latter was formed in low yield by acid-catalyzed condensation of β–hydroperoxy alcohol 17²² with methyl 2-(4-oxocyclohexyl)acetate.

Scheme 2.

Reagents and conditions: (a) 2-adamantanone, CSA, CH₂Cl₂, rt, 12 h; (b) cyclohexanone, CSA, CH₂Cl₂, rt, 12 h; (c) 15% aq. KOH, EtOH/THF, 50 °C, 4 h, then AcOH; (d) methyl 2-(4-oxocyclohexyl)acetate, CSA, CH₂Cl₂, rt, 12 h.

Scheme 3.

β–Hydroperoxy alcohol stereochemistry. Assigned hydroxymethyl group ¹³C NMR signals are indicated in ppm.

Tetraoxane acids 2c and 5c were obtained by hydrolysis of their respective tetraoxane esters 23 and 24; the latter were formed in 45 and 75% yields by Re₂O₇ catalyzed condensation²⁷ of 1,1-dihydroperoxide esters 21 and 22 with 2-adamantanone. 1,1-Dihydroperoxide esters 21 and 22 were obtained in quantitative yields from the corresponding keto esters by treatment with 50% aq. H₂O₂ and I₂ catalyst.²⁸ Tetraoxane acids 1c, 3c, and 4c were synthesized as previously described.^29,30

Target compound efficacy data against F. hepatica are shown in Tables 1 and 2. At eight to thirteen weeks post-infection, rats were treated with single 25–100 mg/kg oral doses of target compounds prepared as suspensions in 7% (v/v) Tween 80 and 3% (v/v) EtOH. At day 6 post-treatment, rats were sacrificed and adult flukes were recovered from the bile ducts and livers. Target compound efficacies were evaluated by comparing the mean total worm burdens of treated and untreated control rats. Statistical significance was calculated using the Kruskal-Wallis test.

Table 1.

Worm burden reductions in adult F. hepatica harbored in rats following the administration of dispiro peroxides at single oral doses of 100 mg/kg.

Compd	X	Y	Worm Burden Reduction (%)	Curesa
Control	------	------	------	0/12
ASb	------	------	30	2/5
1ac	O	CH2COOH	100e	10/10
1b	OCH2	CH2COOH	100e	3/3
1c	OO	CH2COOH	100e	4/4
1d	CH2O	CH2COOH	0	0/3
2a	O	COOH	52	0/4
2b	OCH2	COOH	61e	1/3
2c	OO	COOH	100e	4/4
3ad	O	CH2COOH	0	0/3
3b	OCH2	CH2COOH	95e	2/3
3c	OO	CH2COOH	29	0/3
4a	O	COOH	26	0/3
4b	OCH2	COOH	100e	3/3
4c	OO	COOH	100e	3/3
5a	O	------	100e	3/3
5b	OCH2	------	100e	3/3
5c	OO	------	92e	2/4

^acures = number of rats cured/number of rats treated

^bAS = artesunate

^cdata from Keiser et al.⁸

^ddata from Zhao et al.¹⁰

^ep < 0.05 from the Kruskal-Wallis test comparing the median values of the responses between the treatment and control groups

Table 2.

Worm burden reductions in adult F. hepatica harbored in rats following the administration of selected dispiro peroxides at single oral doses of 50 and 25 mg/kg.

Compd	Dose (mg/kg)	Worm Burden Reduction (%)	Cures
Control	------	------	0/12
1aa	50	53	2/4
1bb	50	100c	4/4
1bb	25	88c	3/4
1cb	50	100c	4/4
1cb	25	71c	0/4
2c	50	48	1/4
4b	50	71c	0/4
4cb	50	61c	1/4
5a	50	19	0/4
5bb	50	100c	3/3
5bb	25	0	1/4

^adata from Zhao et al.¹⁰

^bdata from Kirchhofer et al.³¹

^cp < 0.05 from the Kruskal-Wallis test comparing the median values of the responses between the treatment and control groups

Like 1a, seven compounds were completely curative at 100 mg/kg doses (Table 1). These compounds were then tested at lower doses of 50 and 25 mg/kg (Table 2). Several trends can be seen from the combined efficacy data. First, we suggest that the complete loss of efficacy for 1,2,4-trioxane 1d compared to its active regioisomer 1b results from a different iron (II) reaction profile. Previous investigations²² with the unsubstituted (Y = H) analogs of 1d and 1b reveal that although both 1,2,4-trioxanes undergo a preferred attack of iron (II) on the less hindered peroxide oxygen atom, 1d forms a higher proportion of inactive carbonyl-containing reaction products, and, unlike 1b, does not form a secondary carbon-centered radical by β-scission of the spiroadamantane substructure. Second, for 1,2,4-trioxolane/1,2,4-trioxane/1,2,4,5-tetraoxane compound sets 1a–1c, 2a–2c, 3a–3c, 4a–4c, and 5a–5c, the trioxanes and tetraoxanes were superior to the trioxolanes. Third, the efficacies of 1a–1c compared to 2a–2c show the superiority of an 8’-carboxymethyl vs. 8’-carboxy substituent. Fourth, the efficacies of 1a–1c compared to 3a–3c and 2a–2c compared to 4a–4c show the superiority of a spiroadamantane vs. spirocyclohexane. Fifth, unlike the superior antimalarial efficacies of 8’-aryl vs. 8’-alkyl ozonides,¹⁶ the 8’-aryl 5a–5c were inferior to the corresponding 8’alkyl peroxide heterocycles 1a–1c. Finally, compounds 1b and 1c had the highest overall efficacies, but of the two, as previously determined by Kirchhofer et al.,³¹ 1c had the best efficacy against juvenile F. hepatica and is easier to synthesize than 1b. Ongoing investigations will assess if the superior efficacies of the 1,2,4-trioxanes and 1,2,4,5-tetraoxanes vs. the corresponding 1,2,4-trioxolanes are due to pharmacokinetic differences.³²

Scheme 1.

Reagents and conditions: (a) O₃, CH₂Cl₂/cyclohexane, 0 °C then sg chromatography; (b) aq. NaOH/EtOH, 60 °C, 4 h, then 1 M HCl, 0 °C.

Scheme 4.

Reagents and conditions: (a) 2-adamantanone, Re₂O₇, CH₂Cl₂, rt, 12 h; (b) 15% aq. KOH, EtOH/THF, 50 °C, 4–20 h, then AcOH.