Today, anthelmintic drugs are less effective than in the past due to widespread of drug resistance in parasites. Although non-chemical control (Dave’s post) could extend the effectiveness of the remaining drugs, there is still an urgent need to explore novel potential drug targets to treat these infections. Sequencing the genomes of parasites is a first step to understand how the parasites live and grow, and could eventually lead to the development of new and specific medicines to help eradicate the parasites and the diseases they cause. With the completion of sequencing projects for several parasite genomes, there are efforts ongoing to make sense of the mass of information in terms of the gene products encoded and their interactions in the growth, development and survival of parasites. Published parasite genomes including Haemonchus contortus, Ascaris suum, Schistosoma mansoni, Echinococcus granulosus, Echinococcus multilocularis, Meloidogyne hapla, Meloidogyne incognita, Trichinella spiralis, Taenia solium, Hymenolepis microstoma, Brugia malayi, along with available genomes of Strongyloides ratti, Bursaphelenchus xylophilus, Panagrellus redivivus, Necator americanus, allow for pathways reconstruction, gene gain and loss, and essentially, drug targets prediction and prioritization in parasites.
One area in system biology, which aims to explain the complex relationship between genotype and phenotype, is in chokepoint analysis of metabolism modeling (chokepoints showed in Fig 1 and Fig 2). Metabolic network consists of all the molecules in system (e.g. glucose, etc), and all the metabolic enzymes (e.g. glucose-6-phosphate dehydrogenase, etc) that catalyze any metabolic reaction converting one molecule to the other one. A chokepoint reaction would be either consumes a unique substrate or produces a unique product as defined, and a chokepoint enzyme is associated with any chokepoint reaction. Theoretically, if a chokepoint enzyme does not function, the entire pathway will be blocked, which is lethal to the parasite considering accumulation of toxic or starvation of nutrients. This is also the reason that chokepoint study provides an opportunity for drug target prediction. However, we have to notice that there is still a lot of false positive and false negative in the model.
Chokepoint analysis has been performed in unicellular organism Entamoeba histolytica (Singh et al., 2007) and Plasmodium falciparum (Yeh et al., 2004), and chokepoint enzymes got enrichment in existing drug database. In a single cell, molecules and enzymatic reactions are dense and simple, which means pathways are easier to be blocked if chokepoint enzymes turn off in the system. However, for multicellular species, such as nematode, there is still no strong evidence that the chokepoints are ideal potential drug targets. Pathways are more complex in multicellular organisms, because the same reaction may participate in different tissues and organs, and there are more isoenzymes that allow for alternative pathways when the defined chokepoint reaction breaks off. In this case, chokepoint analysis in nematode is not as efficient as it is in protozon.
Recently, ten nematode species were analyzed with prioritized chokepoints listed and the top drug-like compounds were experimentally tested (Taylor et al., 2013). The criteria they chose included expression in certain stages and tissues, less identity with mammals, participation in more than one pathways or nucleic acid metabolism, and catalysis of hydrolase reactions. The last three were drawn from their analysis, because the enrichment test showed that chokepoint enzymes with those properties were more likely to be drug targets. The first two are reasonable since drugs are going to target expressed genes. Some papers suggested that predicted targets should have less identity with the host (mostly are humans). However, most chokepoint enzymes were high conservative, and many existing enzymatic drug targets in fungi, bacteria, and protozon also have high similarity with human. Isoenzymes (multiple proteins are assigned into the same enzyme) were excluded when prioritizing drug targets in some paper as well, considering multi-gene knockdown was not easily performed. However, there is still no strong evidence that isoenzymes are less likely to be a target.
Seven drug-like compounds were tested, and three got a phynotype of a reduction in motility. It is possible that the worms would still be alive, develop, and reproduce after given those drugs, so it is hard to say if those drugs are really effective. The author also mentioned that high throughput RNAi experiments might fail to generate a phenotype since single gene knockdown would have no effect due to genetic redundancy and RNAi resistance for genes or cell type. Additionally, although there are some enzymatic drugs for human, bacteria, fungi, protozon and insect, those drugs for nematodes are still quite few, which can not provide a full comparison with the chokepoint prediction. Without a golden standard, validation comes to a hard part of this analysis.
Genomes include all the information in the organism. Metabolism modeling begins with enzyme prediction, so it greatly depends on the quality and completeness of the genome annotation. When genome sequencing and assembly become more complete, the prediction will be more accurate. There are still many uncertain factors at this moment. The method always yielded a long list of chokepoints that almost occupied half of the total number of enzymes, and the prioritization criteria are still to be determined. Although there are inevitable false positive and false negative results, the advantage of high throughput data analysis is that it could avoid lots of experiments, which saves money and time. It is usually difficult to validate computational prediction, but chokepoint analysis based on genome annotation is still worth doing because it is the first step to predict potential drug targets; otherwise we have no clue to develop new drugs. And that is what the genome tells us.