I'm trying to assess the utility of various in vitro assays/models, and I'm looking for *general thoughts/opinions* here. Elsewhere, I'm looking at the specifics of modelling the brain, modelling nanmomaterial-blood interactions, modelling cell transplantation etc.
I suspect each aspect of each model would need to be tested, and then the data compared to whatever in vivo data is available (if any; it may be incomplete, or rely on surrogate outcomes/biomarkers).
E.g. for a white blood cell (WBC) in vitro model: are WBC responses to immunosuppressive agents in vitro, identical to responses in vivo? You'd presumably have to measure the levels of lots of cytokines, in vitro and in vivo, and some may correlate better than others. (also morphological, motility, toxicological assays, etc.)
I presume that you can only conclude that (this specific) in vitro WBC assay is a good model of (this specific) cytokine release in response to (this specific) immunosuppressive agent.
Strictly, you shouldn't extrapolate from this to assume that, say, apoptosis or chemotaxis in this model would also be representative of in vivo, but are there any standardised approaches to this kind of vitro/vivo comparison?
Have any rules for generalisation/extrapolation emerged? (e.g. if the in vitro secretome of a model is representative of in vivo, can we expect proliferation or apoptosis to also be representative of in vivo? Or is every element distinct?)
I'm sure most models work well for some aspects, but not others (e.g. rodent models vs human). Is there any database where failed/predictive applications of models are reported/collected?
Is there any database comparing (specific) cell lines with primary cells and/or in vivo data, for various assays?
I appreciate that this is broad and woolly, but I thank you for any comments on this.