High content screening methods or automated microscopy based assays, are a more recent development in drug discovery. This technology is rapidly becoming a mainstream tool in profiling compound activities. One clear harbinger of this uptake in high content screening is the numerous different vendors which have brought their platforms to market enabling this work to proceed.
There are many advantages of using automated microscopy assays, compared to conventional assays. One generally assumed advantage is removal of compounds that cause optical interference due to the type of data obtained by a high content readout and the methods used in the assay (wash steps for example)
This assumption was tested in the following publication:
In this article the authors screened 315,000 compounds with a high content assay using the IN cell Analyser 3000 platform, where they were looking for modulators of micro RNA biogenesis pathway using HeLa S3 expressing green fluorescent protein. Active compounds would lead to an increase in expression of the fluorescence signal.
Using a hit threshold of 20% signal increase for the screen, they were able to obtain a hit rate of 0.36% (1130 compounds). The authors were able to retest 836 of these primary screen hits, in both single concentration and concentration response curves with both the original cell line and the parental cell line which did not express the green fluorescent protein. This is where the project hit some trouble.
Around 22% of hit compounds reconfirmed with the GFP cell line in both the single concentration and concentration response curves, which is not unreasonable. Disappointingly, roughly the same numbers repeated in the cell line which did not contain the GFP expression system. Effectivity all the active compounds were not specific and would suggest that they are all false positives.
These identified false positive compounds were grouped together by structure into four main classes and were listed in the publication by the authors. This could be a useful tool if you want to compare other compounds which you might be working against this list. It should be remembered that (currently) these compounds are only false positives in this assay, with its specific fluorescent wavelengths used. Just because the key compound in your project is a member of one of these classes, it does not mean you need to stop working on them, but it would suggest further investigation would be warranted.
If possible, it would have been interesting to see if the results were repeated with a standard conventional fluorescence assay compared to the automated microscopy method whether the team would have achieved the same results?
So in summary, automated microscopy assay can suffer from optical interference caused by compounds, they are not immune.
However, overall the authors should be commended for releasing this publication; it highlights that all assay formats (including automated microscopy assays) will have a degree of false positive compounds and that you have to use all available methods to ensure your data output is as confident as possible.
Blog written by Gareth Williams