A screening pipeline for antiparasitic agents targeting cryptosporidium inosine monophosphate dehydrogenase

Abstract

Background: The protozoan parasite Cryptosporidium parvum is responsible for significant disease burden among children in developing countries. In addition Cryptosporidiosis can result in chronic and life-threatening enteritis in AIDS patients, and the currently available drugs lack efficacy in treating these severe conditions. The discovery and development of novel anti-cryptosporidial therapeutics has been hampered by the poor experimental tractability of this pathogen. While the genome sequencing effort has identified several intriguing new targets including a unique inosine monophosphate dehydrogenase (IMPDH), pursuing these targets and testing inhibitors has been frustratingly difficult.

Methodology and principal findings: Here we have developed a pipeline of tools to accelerate the in vivo screening of inhibitors of C. parvum IMPDH. We have genetically engineered the related parasite Toxoplasma gondii to serve as a model of C. parvum infection as the first screen. This assay provides crucial target validation and a large signal window that is currently not possible in assays involving C. parvum. To further develop compounds that pass this first filter, we established a fluorescence-based assay of host cell proliferation, and a C. parvum growth assay that utilizes automated high-content imaging analysis for enhanced throughput.

Conclusions and significance: We have used these assays to evaluate C. parvum IMPDH inhibitors emerging from our ongoing medicinal chemistry effort and have identified a subset of 1,2,3-triazole ethers that exhibit excellent in vivo selectivity in the T. gondii model and improved anti-cryptosporidial activity.

Figure 1. Schematic of purine salvage in T. gondii and C. parvum.

A, Genetic studies have shown that the salvage of adenosine via adenosine kinase is the predominant route to GMP for T. gondii and IMPDH catalyses the rate limiting step of this pathway. However in the absence of adenosine kinase, TgHXGPRT allows for the salvage of adenosine, adenine and guanosine such that the activity of TgHXGPRT is sufficient for parasite proliferation . Several transporters for the uptake of nucleobases and nucleotides have been characterised in T. gondii –; B) Unlike T. gondii and other Apicomplexa, C. parvum lacks HXGPRT , and is dependent on the salvage of adenosine and thus the activity of CpIMPDH. A single adenosine transporter has been identified in the genome of C. parvum , . The T. gondii pathways shown in grey in A highlight the genes disrupted in the parasite clones used in this study, TgHXGRT in a previous studies (HXGPRT;[41]) and TgIMPDH in this study. Hyp, hypoxanthine; Xan, xanthine; Gua, guanine; Guo, guanosine; Ade, adenine; Ado, adenosine; Ino, inosine; AMP, adenosine monophosphate; IMP, inosine monophosphate; XMP, xanthosine monophosphate; GMP, guanosine monophosphate; HXGPRT, hypoxanthine xanthine gunanine phosphoribosyltransferase; IMPDH, IMP dehydrogenase, 1, adenine deaminase; 2, adenosine deaminase; 3, purine nucleoside phosphorylase; 4, adenosine kinase; 5, AMP deaminase; 6, adenoylsuccinate synthase and adenoylsuccinate lyase; 7, GMP synthase , , .

Figure 2. Disruption of TgIMPDH to generate a T. gondii parasite dependent on CpIMPDH.

A shows a Southern blot confirming the replacement of the native TgIMPDH coding sequence. Genomic parasite DNA or purified cosmid DNA was digested with XhoI, subjected to electrophoresis and blotting, and hybridized with a probe specific to the region shown in B. In A, lane 1 shows the wild-type cosmid TOXOU05, lane 2 the knock-out cosmid TOXOU05CATΔIMPDHYFPYFP, lane 3 the parent parasite T. gondii RHΔHX-CpIMPDH-5MX, lane 4 the T. gondii-CpIMPDH-ΔHXGPRT-ΔTgIMPDH knock-out parasite, and lane 5 a merodiploid transformant that retained the native TgIMPDH locus. Note that T. gondii-CpIMPDH-ΔHXGPRT-ΔTgIMPDH is hereafter referred to as T. gondii-CpIMPDH.

Figure 3. Validation of the T. gondii-CpIMPDH reporter parasite.

Schematic representation of the routes to GMP for the wild-type T. gondii, T. gondii-ΔHXGPRT, and T. gondii-CpIMPDH are shown in A, D & G respectively. B, E & H show parasite growth in the presence of 0 µM and 7.8 µM MPA for wild-type T. gondii, T. gondii-ΔHXGPRT, and T. gondii-CpIMPDH respectively. C, F, and I show parasite growth curves in the presence of 0 µM and 7.8 µM MPA, with the addition of 0.33 mM xanthine to the culture media, for wild-type T. gondii, T. gondii-ΔHXGPRT, and T. gondii-CpIMPDH respectively. Data are representative of 2 independent experiments. Abbreviations as in Figure 1.

Figure 4. Overview and validation of the high content imaging C. parvum growth assay.

A, schematic representation of differential labelling of parasite and host. B, detail of an exemplary micrograph obtained through the screening routine. Numbers indicate object identifies after segmentation analysis. C shows a 2-fold titration of C. parvum oocysts where the top concentration was 5.9×10⁵ oocysts per well. For D, the ratio of the number of FITC-VVL labelled parasites to DAPI labelled host cell nuclei was used to standardize each well. Percent C. parvum growth (solid line) was normalized to parasites receiving DMSO alone. The paromomycin EC₅₀ for C. parvum growth was 97 µM. Paromomycin in addition to reducing parasite number also reduces the average size of the parasite (dashed line). Mean parasite area was measured for each treatment in triplicate. The percent area was then calculated respective to the mean area of parasites receiving DMSO alone. Data shows the mean of two independent experiments set up with triplicate wells in a 96-well format.

Figure 5. Compound structures and summary of activities.

N.A., not applicable; N.D., not determined; a. Selectivity = EC₅₀(T. gondii-wild-type)/EC₅₀(T. gondii-CpIMPDH); b. Highest concentration tested; c. Synthesis described previous study ; d. Lowest concentration tested; e. Determined in earlier study ; f. Determined using qPCR as described in . Note that the structure and activities of additional amide derivatives of compound A are shown in Figure S5.

Figure 6. Identification of derivatives with high potency and selectivity in the T. gondii-CpIMPDH model.

A shows EC₅₀ values for a selection of inhibitors assayed in the T.gondii-CpIMPDH parasite model and demonstrates a range in inhibitor selectivity and potency. Compounds were assayed in triplicate and growth inhibition was calculated on a day during the exponential phase of growth, by normalisation to wells receiving DMSO alone. The EC₅₀ calculation was performed as described in Figure S2. Compounds A82, A89, A90, A92, A102, A103, A105, and A110 were selected for rescreening and the mean values for at least 2 replicate experiments are shown. The inhibitors were then tested for inhibition of C. parvum (B) and host cell growth (C). For B percent C. parvum growth was determined using the high-content imaging assay as detailed in Figure 4, with the inhibitor at 12.5 µM and 25 µM. A subset of compounds was selected for re-screening and the mean over at least 2 replicate experiments is shown for compounds A90, A92, A98, A103, A105, A109 and A110. C shows percent host cell growth assayed using the fluorescent HCT-8 cell line with inhibitors at 12.5 µM and 25 µM. GFP expressing HCT-8 cells were seeded at 4000 cells per well into 96-well plates and triplicate wells were spiked with test compound and fluorescence was measured daily for 7 days. Percent growth inhibition was calculated on a day during the exponential phase of growth, by normalisation to wells receiving DMSO alone. Inhibitors A89, A90 and A92 were selected for re-screening and the mean over at least 2 replicate experiments is shown.

Figure 7. Correlation between CpIMPH enzyme inhibition and potency and selectivity in the T. gondii-CpIMPDH model.

A shows a relatively weak (r = 0.58) and statistically insignificant (p-value = 0.3) correlation between the triazole compound IC₅₀ values for the CpIMPDH enzyme inhibition and the EC₅₀ for proliferation of the T. gondii-CpIMPDH parasite. However, a strong, positive correlation exists between the potency of CpIMPDH enzyme inhibition when assayed in the presence of BSA and inhibition of T. gondii-CpIMPDH proliferation (r = 0.94, p<0.0001; B). C, shows that selectivity in the T. gondii model, determined by the relative inhibition of the T. gondii-CpIMPDH parasite over wild-type T. gondii, also correlates well with the potency of enzyme inhibition in the presence of BSA (r = −0.92, p<0.0001).

Figure 8. Compounds A103 and A110 are potent inhibitors of C. parvum growth.

C. parvum growth was determined using the HCI assay. The ratio of the number of FITC-VVL labelled C. parvum parasites to DAPI labelled HCT-8 host cell nuclei was used to standardize each well and percent C. parvum growth was normalised to parasites receiving DMSO alone. A and B show compounds A103 and A110 respectively (EC₅₀<0.8 µM). Data shows the mean of two independent experiments with triplicate wells.