Horizontally transferred genes as new targets for drug development in Cryptosporidium parvum


Boris Striepen
Center for Tropical and Emerging Global Diseases & Department of Cellular Biology, University of Georgia, Athens, GA 30602, e-mail: striepen@cb.uga.edu


Cryptosporidium parvum has emerged as one of the most troublesome waterborne infections in the industrialized world. A number of large outbreaks have occurred in the U.S. with the largest in Milwaukee causing 403,000 cases of acute gastrointestinal diseases. Cryptosporidiosis is also an important factor in severe diarrheal disease in children around the world who lack access to contamination free drinking water. Crytosporidiosis is a gastrointestinal disease, characterized by watery diarrhea, abdominal cramps, nausea and fever. The disease is self-limiting with symptoms usually subsiding after 2 to 3 weeks. In contrast, immunosuppressed patients suffer from prolonged chronic disease resulting in severe dehydration and weight loss which can become live-threatening. It is these patients for which effective antimicrobial therapy is most urgently needed.

A wide variety of antimicrobial agents has been tested in vitro, in animal models and in clinical trials. Despite considerable effort no fully effective therapy has been established yet. However, two drugs have emerged which show a consistent albeit modest benefit in placebo controlled studies: paromomycin and even more promising nitazoxaninde, which has now received FDA approval for the treatment of immunocompetent children. The resistance of C. parvum against drugs, which are highly effective against related apicomplexan parasites has puzzled and frustrated researchers and clinicians alike. Two general models can be developed to explain this resistance. The first model is the extracytoplasmatic model. It argues that the peculiar extracytoplasmatic subcellular localization of the parasite within its host cell severely limits access of drugs to the parasite. In addition efflux pumps could further rid the parasite of toxic compounds protecting susceptible target enzymes. The second model, the metabolism model, proposes that the metabolism of Cryptosporidium differs from other Apicomlexa much more than initially appreciated and that drugs active against other apicomplexa fail because their targets are absent or divergent in C. parvum. Obviously, these models are not mutually exclusive and both mechanisms could act synergistically.
The metabolic hypothesis put forward in this talk explains C. parvum’s resistance to typical anti-apicomplexan drugs as a reflection of its phylogenetic and metabolic uniqueness. One of the most striking differences between C. parvum and other apicomplexan pathogens is the absence of fully functional mitochondria and plastids. An experimental study by Zhu and colleagues had predicted the absence of an apicoplast in C. parvum based on PCR and hybridization experiments, which failed to detect the presence of sequences conserved among all plastid genomes. Indeed the completed and now published C. parvum and hominis genome sequences do neither contain the organellar genome nor the extensive set of nuclear encoded plastid targeted genes described for P. falciparum and T. gondii. The secondary endosymbiosis that let to the presence of the apicoplast is generally viewed as and early event in the evolution of Apicomplexa and Alveolata. The observation that several genes in the C. parvum genome show strong phylogentic relationships to plants and algae support this view, and suggest that C. parvum is derived from a lineage which once harbored an algal endosymbiont which was lost later. The lack of a plastid has important metabolic and pharmacological consequences as it e.g. explains C. parvum’s resistance to macrolide antibiotics like clindamycin which specifically target protein synthesis in the plastid and which are quite effective in T. gondii.
C. parvum is a challenging experimental system, and the lack of continuous culture models and transfection technology has posed limitations on the molecular analysis of this pathogen. However, C. parvum has one of the Apicomplexa’s most accessible genome. The genome is small and introns are rare making gene prediction straightforward. The analysis of the genome sequence has just begun, but it already uncovered a surprising number of metabolic differences between C. parvum and other apicomplexans.
Our work has been focused on nucleotide biosynthesis. Nucleotide biosynthesis has been a main stay of antiprotozoal treatment. Antifolates are highly active against T. gondii and P. falciparum. C. parvum on the other hand is resistant to pyrimethamine/sulfadiazine. DHFR-TS is one of the few potential C. parvum drug targets that has been studied in detail. Structural as well as kinetic analysis recently published by other investigators suggests that the Cryptosporidium enzyme is quite different from previously characterized fused enzymes from kinetoplastids and apicomplexans. In addition to differences in DHRF-TS the general pattern of pyrimidine nucleotide synthesis and salvage could equally modulate the efficiency of antifolates. Recent genomic and experimental work has uncovered a surprising diversity of pyrimidine biosynthetic pathways within the Apicomplexa. P. falciparum is entirely dependent on de novo synthesis of pyrimidines making DHRF-TS an essential enzyme. T. gondii posses the ability to salvage uracil using uracil-phosphoribosyltransferase. However as recently reported by Fox and Bzik, this salvage pathway is not sufficient to sustain the parasite in the absence of de novo synthesis . C. parvum finally, has lost all six genes for the enzymes in this pathway indicating that it is unable to synthesize pyrimidines de novo. The parasite depends entirely on salvage, and three pyrimidine salvage enzymes have been identified, two of them uridine kinase-uracil phosphoribosyltransferase (UK-UPRT) and thymidine kinase (TK) are not found in any other apicomplexan. Phylogenetic analysis suggests that both enzymes were obtained from other organisms by horizontal gene transfer, UK-UPRT from an algal endosymbiont which has since been lost, and TK from a proteobacterium. The presence of TK in C. parvum provides an additional and potentially alternative source of dTMP for this parasite and could reduce its sensitivity to inhibition of DHFR and the subsequent starvation of the thymidylate synthase reaction. Equally the finding of UK-UPRT explains the difference in susceptibility to cytosine arabinoside between C. parvum and T. gondii. While T. gondii is highly resistant to this prodrug, which has to be activated by cytosine or uridine kinase, C. parvum was surprisingly susceptible in a large drug screen conducted by Woods and Upton.
Both UK-UPRT and TK could be new targets to pursue for C. parvum. Especially TK might hold promise based on the successful exploitation of this target in the therapy of Herpes viruses. A large variety of compounds subverting the viral TK has been generated, and the relationship between enzyme structure and drug sensitivity and resistance is well understood. The divergent phylogenetic origin of the parasite enzyme from a proteobacterium might allow for the identification of compounds with selective specificity for the parasite versus the human enzyme. But will nucleoside analogs known to subvert this enzyme be able to reach their target within the parasites? Experiments using 5-deoxybromo-uridine in C. parvum infected tissue cultures indeed suggest that this class of compounds gains access to the parasite.
Horizontal gene transfers into the C. parvum nucleotide metabolism are not limited to the pyrimidine pathway. C. parvum has obtained its inosine 5’-monophosphate dehydrogenase (IMPDH), an enzyme central to the purine salvage pathway, from an e-proteobacterium whereas P. falciparum and T. gondii harbor enzymes of clear eukaryotic phylogeny. As TK, IMPDH is a well established target of antiviral and immunosuppressive therapy. Furthermore prokaryotic and eukaryotic IMPDH differ in structure and mechanism which should facilitate the identification of C. parvum specific inhibitors. Kinetic analysis of recombinant C. parvum IMPDH has shown pronounced differences between parasite and human enzyme, most importantly a 1000 fold difference in susceptibility to mycophenolic acid. This protein has also been crystallized recently and the results from the ongoing structural study should greatly enhance the ability to design parasite specific inhibitors.
Genomic and experimental studies show a highly streamlined salvage pathway for C. parvum which relies on adenosine as sole source of purine (earlier biochemical studies on crude parasite lysates had also predicted adenine, hypoxanthine, xanthine and guanine salvage, however the genes for these enzymes seem not to be present in the genome. IMPDH is at the center of this streamlined pathway and an essential enzyme of the multi-step conversion of AMP to GMP. Treatment of infected tissue cultures with the IMPDH inhibitors mycophenolic acid and ribavirin consequently results in dose dependent inhibition of C. parvum development.
One of these drugs, ribavirin, has also been tested in a neonatal mouse model of cryptosporidiosis and treatment with 50 mg/kg for one week resulted in a 90% reduction of parasite load when compared to untreated controls. Interestingly, in these experiments the drug was injected into the peritoneum, rather than given orally suggesting that uptake from the intestinal lumen through the apical membrane of the host cell might not be necessary for this compound. Ribavirin, like the aforementioned drug targeting TK, is a nucleoside analog. This class of drugs does not freely diffuse across membranes, but subverts the nucleoside transporters of their target cell to get access. But where is the parasite nucleoside transporter localized? Several nucleoside transporters have been characterized in related intracellular parasites and most seem to localized over the entire surface of the parasite. A potential nucleoside transporter with similarity to the T. gondii adenosine transporter has been identified in C. parvum and could provide an important molecular reagent to further address nutrient and drug transport in C. parvum.

A series of recent genomic, biochemical and cell biological studies has produced considerable support for the metabolism hypothesis for C. parvum’s drug resistance. C. parvum’s metabolism differs dramatically from its better studied cousins P. falciparum and T. gondii. In all cases where the target of a widely used anti-apicomplexan drug has been characterized in molecular detail, it was either absent in C. parvum (e.g. clindamycin and atovaquone) or the enzyme was highly divergent and resistant (pyrimethamine). Limited drug access might still remain as an important challenge to treatment, but the data warrants a fresh look at a new and metabolically more appropriate set of drugs and targets. The divergence of C. parvum from the generic eukaryotic metabolism might after all present an Achilles heel, and the presence of numerous bacterial enzymes provides an exciting set of candidate targets for parasite specific inhibition. Transgenic models might provide urgently needed assay systems to validate targets identified by genome mining and screen compounds.

Acknowledgements: The abstract format prevented me from citing published work by colleagues appropriately for which I apologize. Work in the author’s laboratory is currently funded by grants from the National Institutes of Health (AI 48475 & AI 55268), the American Heart Association, and Merck Research Laboratories. I would like to thank Liz Hedstrom, Jessica Kissinger and Jan Mead for many discussions.