At Science Night last week we invited Emily Ricq, a PhD student in Chemical Neurobiology who works on my floor, to come give us an introduction to high-throughput screening. She made us this awesome handout introducing the subject. We invited her to talk so that we could begin to get our heads around the field and how to design a new sort of assay for antiprion therapeutics that does not rely on PrP-res as a proxy for disease-causing misfolded PrP. Here are a few key points from the discussion:
- Since pathways of pathogenesis in prion diseases are as yet poorly understood, a phenotypic assay is probably more realistic than a target-based assay.
- Morphology-based phenotypic assays are already in wide use (including at the Broad Institute) and so are a realistic readout for an assay.
- The choice of cell culture model for screening will be one of the more difficult parts of the assay development. We discussed the possibility of having a diverse environment of different types of neurons and glia in a well, as a recapitulation of in vivo conditions. This type of screening is actually done for embryonic mouse tissue. 3 pregnant mice can yield ~50 96-well plates worth of embryonic brain tissue, enough for what you might call medium-throughput screening. But doing this with human adult induced pluripotent stem cells is still sci-fi as of 2012: the people in attendance who have worked in tissue culture agreed that right now the only protocol we have for making neurons from iPSC is iPSC > NPC > radial glia > glutamatergic neurons; there don’t yet exist protocols for differentiating iPSC into other types of neurons or into a diverse milleu of neurons and glia. People are working on it and we expect progress in the next few years.
- We also discussed how to induce a disease state in the cell culture model. Genetic prion diseases in humans incubate for 40 or 60 years, so how do you induce disease in culture in a time frame relevant for high throughput screening, such as 96 hours or even 2 weeks? Evidence from iatrogenic infection suggests that cross-linking aldehydes reinforce the stability of misfolded PrP, pointing to the possible existence of pro-prion compounds that might be able to induce disease more quickly in cell culture than by just waiting for spontaneous disease. Alternately the cells could be infected, say, with brain homogenate from mice with the same mutation. In either case, the assay would not quite recreate the conditions of spontaneous disease generation in a genetic model, but there’s plenty of other reasons that no assay is perfect.
We also discussed the growing list of therapeutics that have shown some efficacy in mouse models. It was noted that the proposed compounds are very chemically diverse and purported to act through different pathways, which (though this should be double-checked) suggests low inter-drug interactions and possibly additive effects. One goal for a mouse therapeutic trial might be to test a cocktail of these hits to look for a larger effect size than from one drug alone.