The search for new antiparasitic drugs for use in humans has accelerated in the past decade, based partly on the growing recognition that addressing these widespread infections is necessary for poverty reduction. There is a consensus that the drugs available for these pathogens are far from optimal, plagued by susceptibility to resistance, lack of activity against key species (or stages of the life cycle), lack of adequate efficacy in field-compatible delivery regimens, and reliant on single agents for control programmes [1], [2]. Expansion of programs for discovery and development of new compounds has been fueled by investment from donor organizations (such as the Bill and Melinda Gates Foundation and the Wellcome Trust) and increasing participation of the pharmaceutical industry. Some companies have established drug discovery centers for a select set of diseases. For example, the Novartis Institute in Singapore is focusing on malaria, dengue and tuberculosis and the GlaxoSmithKline (GSK) facility in Tres Cantos, Spain, is expanding its efforts to include a number of neglected diseases. We have also witnessed drug donation programmes essential for filariasis control by Merck and GSK as well as praziquantel donation by Merck Serono. This effort has been extended to sharing of proprietary and non-proprietary screening data, exemplified by the recent publication of screening results of corporate compound libraries against malaria parasites [3], [4]. In another development, Novo Nordisk transferred its entire compound library to the National Centre for Drug Screening in Shanghai to support drug discovery for neglected tropical diseases (NTDs) and related capacity building, in collaboration with The Special Programme for Research and Training in Tropical Diseases, at the World Health Organization (WHO/TDR) [5].

WHO-TDR has a long history of drug discovery and development for NTDs [1], [6], [7]. For several decades (1970–2000), it was the primary source of support for systematic antiparasitic screening programmes outside of military institutions and animal health companies [8], [9]. Since then, the creation and evolution of Product Development Partnership (PDP) organizations such as the Medicines for Malaria Venture (MMV), Drugs for Neglected Diseases initiative (DNDi) and Institute for One World Health (iOWH) to support development of promising drugs has become an important factor in bringing modern approaches to pharmaceutical research on neglected diseases [1], [10]. The development of sophisticated antiparasitic drug discovery activities in countries such as India, Brazil, South Africa and China, and increasingly in less developed countries in which these diseases are endemic, introduces new and influential contributors to the renaissance in this area [11]. Furthermore, the continuing efforts of the animal health industry in antiparasitic discovery, particularly in the area of anthelmintics, are being incorporated into human discovery programs [12]; almost all available human anthelmintics were initially developed for use in veterinary settings.

Despite these efforts, major gaps in the discovery of new chemical entities for neglected diseases remain, and apart from some repurposed drugs and few new molecules for malaria [13], [14], [15], the international community has not been able to transition novel chemical entities from discovery into development in the past 15 years. An innovation gap has therefore been defined for the various phases of the drug discovery process [16]. Against this background, it is critical to invest in the discovery of new drugs that will contribute in the medium to long term control of these diseases.

A network model for drug discovery against multiple neglected diseases has been described as a suitable mechanism to overcome these challenges [1], [16]. This approach involves a coordinated North-South network of collaborators (from both public and private sectors) which operate low- to medium-throughput primary screens against target organisms, supported by medicinal chemistry, drug metabolism and pharmacokinetics resources to kick-start hit-to-lead and lead optimization programs. Over the years part of this screening network funded by TDR has systematically optimized throughput to enable the testing of hundreds to thousands of pre-selected chemicals in diverse collections against the following: Plasmodium falciparum, Leishmania infantum, Trypanosoma brucei, Trypanosoma cruzi, Schistosoma mansoni, Onchocerca lienalis or O ochengi (counterparts in cattle of the human pathogen, Onchocerca volvulus) and Brugia malayi. Higher-throughput methods have been developed for whole organism antimalarial and anti-Chagas screens, both in industrial and non-industrial settings [3], [4], [17]–[20]. However, the throughput capacities of whole-organism screens employing the other pathogens are generally low to medium, permitting parallel testing of tens or hundreds to a few thousand compounds per batch [21].

TDR has taken advantage of the availability of multiple assays available within its network to implement coordinated screening of highly triaged small-to-medium sized libraries against multiple organisms, combined with a rapid assessment of mammalian cell toxicity. Although no in vitro test is fully predictive of in vivo activity, the TDR strategy is generally accepted as standard practice in neglected diseases screens [22], [23]. Consistent with a virtual operation, library evaluation, management and sampling are coordinated with the support of external partners and consultants, who have extensive experience in medicinal chemistry related to drug discovery. These consultants help with prioritization of compound collections to be screened, analysis of the resultant screening hits thorough chemo-informatics analysis and review of available literature on the compound or series identified. Data evaluation, communication and decisions by the drug discovery network are supported by a readily accessible, centralized database in which biological and chemical results from distant sites are uploaded in close to real-time processes for analysis and searching [24]. The database provides chemoinformatics strategies enabling substructure searches to analyze and expand structure-activity relationships (SAR).

Here we present the outcome of screening a set of 10,000 compounds against seven (7) NTD pathogens in whole-organism assays through a collaborative network model. The implications of this approach, which we refer to as an integrated multi-pathogen screening strategy, for enhancing drug discovery productivity and efficiency for neglected diseases are discussed.

Historically, antiparasitic drug discovery has relied on whole organism screens, as exemplified by recent reports of promising leads from whole-malaria parasite screens [4], [13], [18]. Although progress has been made in drug discovery built on mechanism-based screening approaches and in identifying parasite proteins which are good drug targets, clear success stories are not yet available [38], [39]. Similar but much more limited screening efforts have been devoted to less publicized NTDs, including those caused by kinetoplastids and helminths. Whole parasite screening for drug lead identification for kinetoplastids has been limited, and outside of the animal health industry, few systematic efforts have been directed at screening against helminths. TDR efforts to identify novel starting points using whole-organism screening for the more neglected diseases warrant recognition and further support.

This work underscores the need for increased efforts in integrated screens for neglected diseases. The use of specific library selection methods which exclude previously described or known scaffolds and the early filtering of active molecules provides novel mechanisms to enhance screening outcomes for NTDs. A striking outcome of the current approach is the low number of actives identified for P. falciparum. This is due to the nature of the compound set and the fact that most known actives against P. falciparum were eliminated from the library. However, one of the identified novel malaria hit series is structurally similar to a previously identified series undergoing lead optimization in collaboration with Merck Serono (data not shown). This screening strategy also highlights the critical importance and value of early cytotoxicity assessment (Figure 3 and Table S1). This single test allowed us to rapidly discriminate highly attractive series from poorly selective molecules, as most cytotoxic compounds also hit multiple pathogens. It therefore helps to remove nearly all non-selective molecules, meaning molecules active against >2 unrelated pathogens [structures # 1 to 6]. It is noteworthy that we identified only one molecule [structure # 36], with a cytotoxicity IC₅₀<5 µg/ml that was active against >2 pathogens.

The cost of resolving selectivity issues in secondary assays, or even later during clinical trials, can detract greatly from efficiency of the overall operation. We consider that lead series identified in in vitro assays, subjected to cross pathogen analysis as described in this paper, represent the most promising candidates for in vivo follow-up, minimizing resource wastage on unpromising hits. Although discovery is an expensive endeavor, expansion of chemistry or transition of a compound to development with associated ‘hit-to-lead’, lead optimization and preclinical toxicological assessment markedly escalates the investment. It is vital to advance only compounds that have a legitimate chance of success, especially with limited resources. Our approach of screening against multiple pathogens represents a path to overcoming some of the challenges of early drug discovery.

Our results also show that none of the anthelmintic hit series which showed no appreciable cytotoxicity was active on > one helminth pathogen, i.e., identified positives are selectively active in the schistosomiasis, onchocerciasis or LF screens [Table S1, structure # 27–35]. This again justifies the multiple species approach, as it helps to rapidly identify compounds that are active against multiple pathogens. Therefore, it helps to determine whether such overlapping activity is indicative of general toxicity as well as different or related mechanisms of action in different or related species, respectively. As indicated earlier, some hits against T. cruzi also showed activity against T. brucei [structure #11, 12 and Figure 4]. The activity shown against related kinetoplastid parasites (T. brucei and T cruzi) is suggestive of a specific mechanism of action. Although azoles may be active against different parasites, they probably act by different mechanisms in each [40]. The cross-Phylum activity of artemisinins is another example [41], but depending on the target product profile for the disease [1], [42], this kind of activity should be a signal for caution in pursuing a series. It should also be noted that a few non-cytotoxic hits against one or two protozoa are also active against one helminth (Table S1, structure # 30). Noteworthy are series in which some compounds display an activity against one species, whereas other compounds display activity against another species. The mechanism of action may be the same, but with a slight difference of target structure between the 2 parasites. In this case, the chemical exploration of SAR could be beneficial to research on both species.

The major benefit of the current approach centers on the synergy gained by simultaneously screening a compound collection against multiple organisms. Salient advantages include:

It should be emphasized that it has been traditionally difficult to generate novel, robust hits against leishmaniasis, leaving the current screening and drug discovery pipeline weak for this disease (www.dndi.org). It is therefore remarkable to see the interesting hit rate found in our screening campaign. The fact that we discovered different compound series, containing up to 17 similar molecules active against L. infantum, gives some degree of confidence to the results. This again points to the usefulness of careful selection of chemical libraries used for screening.

Presently, whole-organism screens for many of these parasites are not amenable to the scale of throughput possible with P. falciparum in erythrocytes [20]. However, the integrated screens against the 7 pathogens reported here demonstrate the feasibility and importance of leveraging resources across parasite screens. It can also aid the development of SAR data around identified starting points. In the context of the actives described here, expansion of a series should be accomplished by acquisition of close-in analogs such that hit series from the various screens can be further prioritized across indications. It would be interesting to evaluate the previously identified molecules from the various antimalarial screening projects using the approach described in this paper. This early evaluation could help to select the most promising series for development based on selectivity, thereby limiting the high risk inherent in transitioning molecules from early discovery to early development.

While the integrated screens presented here represent current “best practice”, many unknowns plague the screening pathways for new antiparasitic agents for human use. The dearth of safe and effective drugs for most NTDs means that it is difficult to fully validate the screening streams. Some of the most important antiparasitic drugs seem to require an interaction with the host immune system, for example, praziquantel for schistosomiasis and ivermectin for filariasis [2], a paradigm that is difficult to incorporate into a whole-organism screen in culture. We do not know how closely parasites in culture resemble parasites in a host in terms of drug target expression and the role of these targets in survival in the host. Some of the screens employ representative species for discovery, even though there are well known differences in sensitivity of other species in the same genus or clade (oxamniquine, for example, is active against S. mansoni but inactive against S. haematobium [42]. In some cases, the target parasite species of greatest interest is unavailable for screening (for example Plasmodium vivax, Wuchereria bancrofti, Onchocerca volvulus). Finally, the screens do not always employ the most important target stage of the parasite (microfilariae vs. adult filariae, for example). The screens are thus compromises between ideal and practical operations, with the central assumption inherent in this strategy: that drugs with broad-spectrum activity within a parasite group will be discovered and are of the greatest interest. On the logistics side: the network screening model requires careful plate formatting and repartitioning to minimize risks of inability to trace inter or intra screening variations.