**2.5.5 Transporters**

332 Malaria Parasites

models of other enzymes belonging to the FasII metabolism such as FabB/F confirmed that the FasII pathway was dispensable through the entire parasite development to the exception of the hepatic-blood stage transition phase. These results raised the question of the nature of the parasitic target of triclosan in blood stages. Recently five laboratories reported their failed attempts to inhibit rodent *P. berghei* and *P. chabaudi* proliferation in mice by using

Conversely, recent works focusing on the DOXP non-mevalonate isoprenoid pathway confirmed it appears as a relevant target for antimalarial drug development. Key results include the recent failure to delete the DOXP reductoisomerase gene (Odom & Van Voorhis, 2009). The DOXP metabolic pathway is active in all erythrocytic stages but problems with fosmidomycin, which has long been known to inhibit *Plasmodium* growth *in vitro* and *in*  mice (Jomaa et al., 1999), emerged. Due to its short half life that may lead to parasite relapses and/or facilitate selection of resistant parasites, analogues with improved half-lives are actively looked for against this enzyme and the other enzymes of this metabolic pathway (Seeber & Soldati-Favre, 2010). The recent discovery that fosmidomycin is little effective against *Theileria*, *Eimeria* and *Toxoplasma* also raised the question of its bioavailability (Seeber & Soldati-Favre, 2010). This led to the proposal that fosmidomycin and its more active derivative FR900098, both active against *Plasmodium* erythrocytic but not hepatocytic stages, would enter the parasite through new permeation pathways called NPPs (Baumeister et al., 2011). Recently also two triose-phosphate transporters present on the apicoplast membranes have been characterized, PfiTPT/PfAPT1 and PfoTPT/PfATP2, and are being considered as putative targets since they are believed to import from the cytosol into the apicoplast key

Malarial kinases but also host cell kinases involved in host-parasite interactions are also currently considered as promising targets (Doerig et al., 2010). The *P. falciparum* theroretical kinome was predicted to comprise 85 to 99 enzymes, depending on the stringency of the computational methods used to identify them, and phylogenetic studies have revealed interesting differences with the human kinome that may be exploited for kinase-targeted drug discovery (Leroy & Doerig, 2008). For example, *P. falciparum* possesses kinases and kinase families such as calcium-dependent protein kinases (CDPK) and the apicomplexan specific FIKK kinase families, that are absent in humans. Important advances in this field have been provided by the fact that many *P. falciparum* kinases have been successfully produced as recombinant active enzymes, allowing medium/high troughput studies, many structures are known or can be deduced by modelling (see http://www.thesgc.com) and several chemical libraries, developed for other pathologies such as cancer may be screened on *P. falciparum* putative targets and parasites. For example, PfCDPK1 that has no ortholog in humans was produced as an active recombinant enzyme and used to screen a library of ~20,000 molecules developed against kinases, from which a purine derivative called purfalcamine active to the low nM range on the enzyme was isolated (Kato et al., 2008). This molecule, able to prevent the *P. falciparum* development in culture had however a low *in vivo* activity on *P. berghei*, possibly due to poor pharmacokinetic parameters. Importantly, not only protein kinases but kinases phosphorylating other substrates such as choline kinase and pantothenate kinases could also be targeted (Leroy & Doerig, 2008). Genetic validation

similar doses of triclosan (Baschong et al., 2011).

metabolites to fuel the DOXP pathway (Lim & McFadden, 2010).

**2.5.4 Kinases** 

Transporters are integral transmembrane proteins that enable the movement of solutes across biological membranes. In *Plasmodium,* they are currently considered as highly promising targets, due to their involvement, as carrier proteins, channels or pumps, in the movement of nutrients, metabolites and ions into and out of the parasite as well as between subcellular compartments within the parasite. They thus regulate essential nutrient uptake into the parasite, homeostasis and disposal of toxic wastes. Host cell transporters modified by the presence of the parasite may also be involved in these biological functions. Of note, malarial transmembrane proteins such as PfCRT, Pgh1, PfNHE1 and possibly PfATP6 are involved in drug-resistance mechanisms.

The recent publication of the theoretical permeome of *P. falciparum* indicated that at least 2.5% of the parasite genome encode transporters, channels and pumps (Martin et al., 2009). Experimental localizations, that have been achieved by using either specific antibodies or epitope-tagged or fluorescent recombinant fusion proteins transfected into parasites, concern few proteins of the parasite surface (PfHT1 or the hexose transporter, PfENT1 or PfNT1, PfMRP, PfAQP), the mitochondrial membrane (ATP/ADP transporter), or the foodvacuole membrane (PfCRT, Pgh1, V-type H+ ATPase and ATP-driven H+ pump). Substrate specificities are also often ill-defined. Even if these could be predicted by orthology for about half of the candidates, experimental validations are still necessary to confirm or discover which they are. Extensive experimental work therefore needs to be done to further exploit such a rich source of potential targets (Staines et al., 2010). Some promising results nonetheless emerged in this field. First, the challenging functional expression of recombinant proteins, which is necessary for the biochemical testing of substrates and inhibitors, has been achieved successfully in *Xenopus laevis* oocytes for several *Plasmodium* transporters. Alternative heterologous systems have also been used such as yeast, *Dictyostelium discoideum* and mammalian cells (Martin et al., 2009). In some cases, cell-free transporter assay systems are used. Recently, recombinant expression of codon-optimized PfHT1 in yeast was performed to permit high-throughput screening of inhibitors (Blume et al., 2010). The P-type ATPases, PfATP4 and PfATP6, and the drug-resistance involved PfCRT and Pgh1 have also been expressed in such heterologous system, opening the road for functional studies (Martin et al., 2009). The number of known 3D-structures remains however extremely low, limited to PfAQP although a model has been proposed for PfHT1 based on the structure of *E. coli* permease (Staines et al., 2010).

Transfection methods have been also critical in this field, in particular to assess the role of PfCRT and Pgh1 in drug resistance, by transfecting *P. falciparum* resistant alleles into sensitive parasites and vice-versa, or by transfecting these alleles in various heterologous expression systems such as yeast (Martin et al., 2009). Gene deletion studies have been also undertaken, but so far for a limited number of candidates. While several genes could be deleted with none or limited impact on parasite asexual growth such as those encoding PfENT1, PfMRP, the *P. berghei* orthologue of the PfAQP, PfKch1 (PfK1) and PfACα, the

Advances in Antimalarial Drug Evaluation and New Targets for Antimalarials 335

pharmacokinetic profiling found most of them suitable for further development. One lead was further evaluated *in vivo* on *P. yoelii*-infected mice and displayed significant antimalarial activity. In another study, the screening of a focused library of about 12,000 natural and synthetic chemicals led to identification of spiroindolone compounds as appropriate candidates for further development (Rottmann et al., 2010). Optimization studies produced

*Plasmodium* species that infect humans are essentially unable to infect non-primate animal models. Historically, *in vivo* evaluation of antimalarial compounds began with the use of avian parasites in birds and simian parasites in monkeys. The discovery by Ignace Vincke and Marcel Lips in 1948 at Keyberg in the then Belgian Congo, of a rodent malaria parasite (*Plasmodium berghei*) which readily infected laboratory mice and rats, greatly facilitated *in vivo* drug screening management. Since then, several other species and subspecies of rodent parasites have been described and employed. With the discovery in the early 90's that the owl monkey (*Aotus trivirgatus*) is susceptible to infection with the human parasites *P. falciparum* and *P. vivax*, the simian models of malaria have regained interest. Infection with *P. falciparum* is now well characterized in both *Aotus* and *Saimiri* monkeys (Collins, 1992), and primate models, because they provide a clear prediction of drug efficacy and pharmacokinetics in humans, are a logical transition to clinical studies. However, there are obvious limitations to their use, and any primary screen dependent upon monkeys appears both wasteful in terms of animal conservation, drug consumption, and ethics. So, the first steps of *in vivo* drug screening typically begins with the use of mouse models of malaria with the rodent parasites *P. berghei*, *P. yoelii*, *P. chabaudi* and *P. vinckei*, that have been

Individual parasite species and strains have been well characterized, including duration of cycle, synchronicity, drug sensitivity and course of infection in genetically defined mouse strains. *P. berghei, P. vinckei* and few strains of *P. yoelii* and *P. chabaudi* cause lethal infections in mice, whereas *P. yoelii, P. chabaudi,* and *P. vinckei petterei* are usually cleared after the

*P. berghei* - The strains K173, ANKA, NK65 and SP11 are the most frequently used for experimental work. The affinity for immature erythrocytes (reticulocytes) varies according to the strain. ANKA invades preferentially reticulocytes, while NK65 invades indifferently reticulocytes and mature erythrocytes, i.e. normocytes. Both strains are lethal, but ANKA kills the mouse with a low parasitaemia (it rarely exceeds 20%), whereas infection with NK65 induces a continuous increase in parasitaemia until the mouse dies. The erythrocytic cycle is asynchronous and its duration is about 21 h. Schizogony lasts 50 h in hepatocytes. Susceptible mice infected with *P*. *berghei* ANKA die within two weeks after infection with severe neurological symptoms and cerebral microvascular abnormalities. These are common pathologies to both murine and human infections that make of *P. berghei* ANKA a useful experimental model of cerebral malaria. Resistant mice do not show any neurological symptom but die during the third or fourth week post-infection (p.i.) with high

initial acute parasitaemia or after a subsequent low-grade chronic parasitaemia.

NITD609, which is developed by Novartis.

**4.** *In vivo* **antimalarial drug evaluation** 

validated through the evaluation of several antimalarial drugs.

**4.1 Rodent models of malaria** 

parasitaemia.

deletion of genes encoding PfHT1, PfCRT and Pgh1 could not be achieved, suggesting that the two latter ones play essential functions in the parasite biology beside being involved in drug-resistance (Sanchez et al., 2010). Such experiments performed using rodent malaria models further indicated that some candidates, apparently dispensable for the asexual development, turned out important for the development of other parasite stages in insects or liver (Martin et al., 2009, Staines et al., 2004). Among all these candidates, PfHT1 is the only malarial transporter that has been validated both chemically and genetically (Staines et al., 2010). Gene deletion studies and D-glucose derivatives used as inhibitors confirmed the essential role of the hexose transporters for the asexual parasite development and other parasite stages (Blume et al., 2010, Slavic et al., 2011).

#### **3. Renew of phenotypic screening approaches**

Advanced antimalarial drug discovery programs revealed four general strategies to discover new drugs (Wu et al., 2011) : 1) To start from specific, validated or not, malaria targets to find new hits (e.g. haem polymerization, enzymes of the pyrimidine synthesis pathway); 2) To synthesize new analogs from known antimalarial pharmacophores (e.g. synthetic endoperoxides (Trioxolanes) or hybrid molecules having two pharmacophores (Trioxaquines)); 3) To start from drugs developed for other diseases whose similar targets are present in malaria parasites (as illustrated by the piggy-back approach undertaken against parasite protein farnesyltransferases using anti-cancer agents targeting protein prenylation developed by pharmaceutical companies); 4) To screen chemicals on whole cellbased assays to identify hits that kill parasite for further optimizations. Most of the current antimalarials were originated from this last phenotypic drug discovery approach. The revolution in molecular biology led to a switch to target-based approaches for drug discovery in pharmaceutical industries. However, this approach failed to deliver the expected results, especially for antimicrobials (Keller et al., 2011, Payne et al., 2007, Sams-Dodd, 2005). In the recent years, the access to large chemical libraries and the improvements of whole cell-based high throughput screening assays led to a renaissance of the phenotypic approach with the forward chemical genetic strategy. Chemical genetics is the study of biological systems using small molecules as tools (O'Connor et al., 2011). Forward chemical genetic uses small molecules to modulate gene-product function leading to a phenotype of interest (parasite killing for example), and the target must be further identified. In contrast, reverse chemical genetic screens specific gene products with libraries of small molecules to identify ligands, which are then tested on cells for phenotypic changes. Over the last few years, reports demonstrated the power of chemical genetic for antimalarial drug discovery. The GlaxoSmithKline (GSK)'s chemical library constituted of nearly 2 million chemical entities was tested upon drug sensitive and multidrug resistant *P. falciparum* strains. More than 8,000 compounds, clustered into 416 molecular frameworks, showed potent antiplasmodial activity. Analysis of historical GSK data suggested that the main target classes affected by these compounds are malaria kinases, proteases and G-protein coupled receptors (Gamo et al., 2010). In a similar study, a library of nearly 310,000 chemicals, designed to cover a large diversity of bioactive compounds, was screened upon drug sensitive and multidrug resistant *P. falciparum* strains (Guiguemde et al., 2010). Amongst hits, 172 were cross-validated by three laboratories using distinct assays. A reverse chemical genetic approach was undertaken with the validated set of compounds using 66 malarial target assays and identified 19 new inhibitors of 4 validated targets. Preliminary pharmacokinetic profiling found most of them suitable for further development. One lead was further evaluated *in vivo* on *P. yoelii*-infected mice and displayed significant antimalarial activity. In another study, the screening of a focused library of about 12,000 natural and synthetic chemicals led to identification of spiroindolone compounds as appropriate candidates for further development (Rottmann et al., 2010). Optimization studies produced NITD609, which is developed by Novartis.
