Evaluation of an FDA approved library against laboratory models of human intestinal nematode infections

Drug repurposing is the key drug discovery strategy for human helminth infections, since anthelmintic drug discovery and development has languished [7]. Veterinary anthelmintics have obviously been the most attractive candidates for cross-over development for the treatment of human STH infections, since the regulatory standards for veterinary drugs are compatible with those for human drugs and a track-record of their use in animals exists [8]. Indeed, all of the few anthelmintics currently used for humans stem from veterinary medicine. In the present work, we applied a broader repurposing strategy, by evaluating 1,600 compounds from the FDA library, which contains approved drugs over a range of indications.

While a handful of studies employing target screening approaches have been conducted to identify new pharmacophores for the treatment of human STH infections [9], phenotypic screening of selected libraries on parasitic nematodes is basically non-existent. However, whole-organism screens have several advantages over target-based screening, since they are not constrained to single targets (which are not well characterized for helminths) [10]. Indeed, phenotypic approaches are more successful for small-molecule, first-in-class medicines than target based approaches [11]. High throughput assays with Caenorhabditis elegans often serve as a substitute for helminth phenotypic assays, because most assays for parasitic nematodes are laborious and low throughput. However, the correlation of activity against C. elegans to parasitic nematodes is not universal, with for example albendazole lacking activity against C. elegans [12].

We observed a hit rate of 3.4 % against A. ceylanicum larvae, our primary screen, with hits ranging across a large spectrum of indications. For comparison, a hit rate of ~7.6 % was observed against S. mansoni schistosomula using the same library and also encompassing a wide range of compound indications [13]. As highlighted already for the S. mansoni screen [13], it was disappointing that many of the hits were unsuitable for further testing, notably due to documented toxicity or restriction to topical use. In addition, we excluded the common anthelmintics (e.g. mebendazole, pyrantel). Selected compounds were next tested against a panel of helminths, namely A. ceylanicum adults, H. polygyrus larvae and adults, as well as T. muris larvae and adults. A few observations are offered for discussion. First, A. ceylanicum L3 were selected for the primary screen since, compared to adult stages, their use offers many advantages, mainly in terms of ethical considerations, numbers available, and ease of provision [14]. However, larval stages may not always be equally as sensitive as the target parasite stage, the adult worms. For example, the veterinary anthelmintic, monepantel, lacks activity against A. ceylanicum L3, while it is active against the adult worm [15]. Hence, larval-based assays should be validated in terms of how larval sensitivity compares to the sensitivity of adult worms. Additional file 1: Table S1 summarizes the activity of all 54 A. ceylanicum L3 active drugs (including topical and toxic drugs) against adult A. ceylanicum, T. muris L1 and adult T. muris. In general, A. ceylanicum L3 were more sensitive in our assay to the test compounds than the adult stages. Increased sensitivities of larval forms over adult worms have been reported earlier including for S. mansoni schistosomula [16]. Though a high false positive rate (larval activity does not always translate to adult activity) is not optimally cost-effective, the risk of losing interesting compounds is minimal, and in any case, larval pre-screens are still more cost-effective and more ethical than conducting the entire screen on adult stage worms. However, for T. muris this trend could not be confirmed (Additional file 1: Table S1). Adult T. muris were equally or even more sensitive than T. muris L1. Second, to our knowledge, we have for the first time compared the sensitivities of a small series of test drugs on larval and adult stages of the two hookworm species. The H. polygyrus model is widely used to study intestinal nematode infections, as it is easy to maintain in the laboratory and far more cost-effective than other laboratory hookworm rodent models [17]. The activity of compounds on A. ceylanicum L3 and H. polygyrus L3 was similar (only bitoscanate, dexlansoprazole, metformin and natamycin lacked activity against H. polygyrus L3). Similarly, the A. ceylanicum adult screen mirrored the adult H. polygyrus screen (except for dicyclomine HCl and ethopropazine, which showed activity against adult H. polygyrus). Hence, our results with a small panel of compounds support the use of H. polygyrus in the framework of anthelmintic drug discovery.

Only two compounds revealed activity in vivo, namely trichlorfon and bitoscanate. In addition, a moderate activity was observed with natamycin against H. polygyrus (two out of four mice cured), while the drug was not tolerated by T. muris infected mice. The low in vivo activity of many of the in vivo tested drugs might not be surprising, as bioavailable drugs from the FDA library of 1,600 approved drugs may not have the ideal profile for in vivo activity. Though there are exceptions (e.g. levamisole, ivermectin or albendazole) an ideal treatment for STH, particularly Trichuris spp. infections should be only poorly absorbed in order to target worms sitting in the gastrointestinal tract.

In vivo, trichlorfon revealed high activity against H. polygyrus and T. muris, while bitoscanate showed high activity against hookworm but was moderately active against T. muris. The organophosphate trichlorfon is mainly used as an ectoparasiticide [18]. However, it is also known for its antischistosomal properties [trichlorfon (metrifonate) was marketed for the treatment of S. haematobium prior to the advent of praziquantel] [19] and is applied for the control of intestinal nematode parasites of cattle and sheep. Trichlorfon is particularly used for nematodes that have developed resistance to other commonly-used anthelmintics [20]. In the past years, it was investigated for the treatment of Alzheimer’s disease. Metrifonate, at various fixed and loading doses, was associated with significant cognitive improvement compared to placebo, where the slow-release break-down byproduct, 2,2-dichlorovinyl dimethyl phosphate (DDVP), is supposed to be the active component [21, 22]. The broad activity of metrifonate is perhaps not surprising considering its mechanism of action; metrifonate is a cholinergic drug, acting as an irreversible non-selective acetylcholinesterase and butyrylcholinesterase inhibitor [21]. There is a history of active cholinergic anthelmintics including pyrantel and levamisole, which are both selective nAChR agonists, and ivermectin and moxidectin, which are modulators of glutamate-gated ion channels and nAChRs [23]. Emerging cholinergic anthelmintics, monepantel, tribendimidine and derquantel [24–26] have different nAChR subtype selectivities.

Bitoscanate was widely used prior to the advent of the benzimidazoles for the treatment of hookworm infections [27, 28]. Despite its history of use, the literature is very scarce. Bitoscanate belongs to the broad isothiocyanate class of anthelmintic compounds, which occur widely in nature, particularly in plants of the mustard (Brassicacae) family [29]. In India, clinical trials with thousands of hookworm-infected patients were conducted, which revealed a high efficacy of the drug. Conflicting results were observed with regard to Trichuris spp.; while a good trichuricidal activity was observed in some settings [30], the high efficacy was not consistently seen [28]. Mild and transient adverse events were reported in the large number of studies done with bitoscanate [28]. However, in the US bitoscanate is listed as an extremely hazardous substance (http://nj.gov/health/eoh/rtkweb/documents/fs/2172.pdf) which, when released in certain amounts in the environment, may be of immediate concern to the community. A related compound, amoscanate, was found to be effective against hookworm and schistosomes in humans, but was accompanied with severe liver toxicity in laboratory animals at high doses [31, 32]. However, another related compound, nitroscanate, is widely employed in veterinary medicine to treat roundworm, hookworm and tapeworms [33]. Currently, the isothiocyanates are researched for activity against intestinal bacteria and cancer [34, 35].

The same library was screened against S. mansoni larval stage and adult worms [13] with a very different set of hits. When compared, 15 (27.8 %) of the compounds identified as hits in this screen were also hits against S. mansoni newly-transformed schistosomula (NTS) (abamectin, chlorprothixene hydrochloride, clomiphene citrate, doramectin, eprinomectin, hexachlorophene, metformin hydrochloride, metitepine mesylate, moxidectin, natamycin, phenylmercuric acetate, selamectin, thonzonium bromide, trichlorfon and triflupromazine hydrochloride). The vast majority of these common hits were either toxic or for topical use only. However, it is also noted that the macrocyclic lactones, a group that acts on glutamate-gated ion channels [36], in general appear to be active against both S. mansoni and hookworm.