A high-throughput phenotypic screen identifies clofazimine as a potential treatment for cryptosporidiosis

Abstract

Cryptosporidiosis has emerged as a leading cause of non-viral diarrhea in children under five years of age in the developing world, yet the current standard of care to treat Cryptosporidium infections, nitazoxanide, demonstrates limited and immune-dependent efficacy. Given the lack of treatments with universal efficacy, drug discovery efforts against cryptosporidiosis are necessary to find therapeutics more efficacious than the standard of care. To date, cryptosporidiosis drug discovery efforts have been limited to a few targeted mechanisms in the parasite and whole cell phenotypic screens against small, focused collections of compounds. Using a previous screen as a basis, we initiated the largest known drug discovery effort to identify novel anticryptosporidial agents. A high-content imaging assay for inhibitors of Cryptosporidium parvum proliferation within a human intestinal epithelial cell line was miniaturized and automated to enable high-throughput phenotypic screening against a large, diverse library of small molecules. A screen of 78,942 compounds identified 12 anticryptosporidial hits with sub-micromolar activity, including clofazimine, an FDA-approved drug for the treatment of leprosy, which demonstrated potent and selective in vitro activity (EC50 = 15 nM) against C. parvum. Clofazimine also displayed activity against C. hominis-the other most clinically-relevant species of Cryptosporidium. Importantly, clofazimine is known to accumulate within epithelial cells of the small intestine, the primary site of Cryptosporidium infection. In a mouse model of acute cryptosporidiosis, a once daily dosage regimen for three consecutive days or a single high dose resulted in reduction of oocyst shedding below the limit detectable by flow cytometry. Recently, a target product profile (TPP) for an anticryptosporidial compound was proposed by Huston et al. and highlights the need for a short dosing regimen (< 7 days) and formulations for children < 2 years. Clofazimine has a long history of use and has demonstrated a good safety profile for a disease that requires chronic dosing for a period of time ranging 3-36 months. These results, taken with clofazimine's status as an FDA-approved drug with over four decades of use for the treatment of leprosy, support the continued investigation of clofazimine both as a new chemical tool for understanding cryptosporidium biology and a potential new treatment of cryptosporidiosis.

Conflict of interest statement

I have read the journal's policy and the authors of this manuscript have the following competing interests: Authors MSL, FCB, TMW, PGS, and CWM have filed a provisional patent related to this work (U.S. Provisional Application No. 62/397,793).

Figures

Fig 1. Screening images and software analysis.

Representative images and software analysis from 1536-well High Content Imaging on Cellomics CellInsight CX5. (Left) Images of two wells either mock-treated with dimethyl sulfoxide (DMSO) or 10 μM nitazoxanide (NTZ). A merged image of the two fluorescent channels are artificially colored to show contrast: DAPI in cyan and FITC in red. Well area for each full-sized image is 802,511.39 μm2. (Right) Zoomed-in images from the inset white squares in the left panel (area is 24,318.53 μm2) are shown. The image overlays are applied by the imaging software to assess the assay metrics: HCT-8 cell count and Cryptosporidium spot count. First, the host cell nuclei are identified and counted based on DAPI signal; cell debris and other particles are rejected based on a size filter (orange). Next, a region of interest, or “cell area” is drawn around each host cell nuclei to encompass where Cryptosporidium parasites may be located. Finally, the software identifies and counts “spots” within the “cell area” based on signal from the FITC-conjugated Vicia villosa lectin (red).

Fig 2. High-throughput screening results of 78,942 compounds shows inherent variability.

Scatter plot of normalized activity all compounds screened at 1.88 μM (Black = Bioactives; Gray = GHCDL). A strict cut-off of 70% inhibition of C. parvum proliferation (dotted red line) was applied to yield 812 primary hits. Neutral controls (0.125% DMSO; Blue = Bioactives; Lavender = GHCDL) and inhibitor controls (0.5 μM FDU; Red = Bioactives; Light red = GHCDL) are also shown to demonstrate the inherent variability in the assay.

Fig 3. Screening hits show activity against two species of Cryptosporidium.

Twelve filtered hits and multiple controls were tested against C. parvum and C. hominis for cross-species confirmation. Data points are log-transformed EC50 values. Controls and compounds of interest are denoted with a square symbols and labeled. The correlation is shown by the trendline slope (solid line, m = 0.7445) compared to a perfect linear correlation (dotted line, m = 1).

Fig 4. In vitro characterization of CFZ.

(A) Chemical structure of CFZ. (B) CFZ inhibited C. parvum proliferation by ≥ 70% at every point of the asexual life cycle. The first asexual life cycle after infection was evenly divided into six 3-h blocks, and labeled as hours post infection (hpi). Infected cells were treated by one of four compounds at the EC99 for 3 h followed by drug washout, and then allowed to continue growing until 48 hpi, when they were fixed, stained, imaged, and analyzed for C. parvum proliferation. EC99 values: NTZ = 8 μM; FDU = 100 nM; CFZ = 30 nM; BKI-1294 = 2 μM. Data shown are the mean ± SEM of two independent experiments.

Fig 5. Pharmacokinetic properties of CFZ.

(A) Plasma concentration of CFZ or BKI-1294 in mice dosed with 20 mg/kg compound. CFZ was formulated in either corn oil (black) or MC-Tween (gray); BKI-1294 was formulated in 7% Tween 80, 3% ethanol, and 90% water (white). Data shown are mean ± SEM (n = 3). (B) Unchanged CFZ or BKI-1294 recovered in the feces of mice dosed in (A). Recovery was measured each day for three days. Data shown are mean ± SEM (n = 3).

Fig 6. CFZ is efficacious in a mouse model of acute cryptosporidiosis.

Fecal oocyst recovery from IFN-γ-/- mice commenced three days after oral delivery of C. parvum oocysts. Line graph data are weight-adjusted mean oocyst counts ± SD (n = 4); inset bar graphs are mean % recovery relative to mock-treated control mice ± SEM (n = 4). (A) Mice were infected with 104 oocysts then treated orally with 10 mg/kg BKI-1294 (light gray) or CFZ (dark gray) on days 4, 5, and 6 p.i. Dotted line is the reliable limit of detection. A two-way ANOVA was conducted to determine significance between mice treated with compound vs mice treated with vehicle: * p < 0.05; ** p < 0.01. (B) Mice were infected with 106 oocysts then treated orally with a single dose of 100 mg/kg CFZ on day 4. Multiple Student’s t-tests were used to determine significance between vehicle-treated and CFZ-treated mice: * p < 0.05; ** p < 0.01.