Analysis of vascular disruption in zebrafish embryos as an endpoint to predict developmental toxicity.

Inhibition of angiogenesis is an important mode of action for the teratogenic effect of chemicals and drugs. There is a gap in the availability of simple, experimental screening models for the detection of angiogenesis inhibition. The zebrafish embryo represents an alternative test system which offers the complexity of developmental differentiation of an entire organism while allowing for small-scale and high-throughput screening. Here we present a novel automated imaging-based method to detect the inhibition of angiogenesis in early life stage zebrafish. Video subtraction was used to identify the location and number of functional intersegmental vessels according to the detection of moving blood cells. By exposing embryos to multiple tyrosine kinase inhibitors including SU4312, SU5416, Sorafenib, or PTK787, we confirmed that this method can detect concentration-dependent inhibition of angiogenesis. Parallel assessment of arterial and venal aorta ruled out a potential bias by impaired heart or blood cell development. In contrast, the histone deacetylase inhibitor valproic acid did not affect ISV formation supporting the specificity of the angiogenic effects. The new test method showed higher sensitivity, i.e. lower effect concentrations, relative to a fluorescent reporter gene strain (Tg(KDR:EGFP)) exposed to the same tyrosine kinase inhibitors indicating that functional effects due to altered tubulogenesis or blood transport can be detected before structural changes of the endothelium are visible by fluorescence imaging. Comparison of exposure windows indicated higher specificity for angiogenesis when exposure started at later embryonic stages (24 h post-fertilization). One of the test compounds was showing particularly high specificity for angiogenesis effects (SU4312) and was, therefore, suggested as a model compound for the identification of molecular markers of angiogenic disruption. Our findings establish video imaging in wild-type strains as viable, non-invasive, high-throughput method for the detection of chemical-induced angiogenic disruption in zebrafish embryos.

PatchSorter: a high throughput deep learning digital pathology tool for object labeling.

The discovery of patterns associated with diagnosis, prognosis, and therapy response in digital pathology images often requires intractable labeling of large quantities of histological objects. Here we release an open-source labeling tool, PatchSorter, which integrates deep learning with an intuitive web interface. Using >100,000 objects, we demonstrate a >7x improvement in labels per second over unaided labeling, with minimal impact on labeling accuracy, thus enabling high-throughput labeling of large datasets.