Introduction: Fluorescence microscopy has seen a spectacular expansion in the biomedical sciences due to the advent of new probes, new contrast techniques and the development of super-resolution that allows imaging biological structure and function beyond the diffraction limit. By far, most current research microscopes are 'low-throughput' instruments and - in this respect - do not differ much from their counterparts a 100 years ago. In a context of costly clinical trials and depleted pipelines, compound design and candidate selection could benefit from microscopy-based 'screening by imaging' approaches.
Areas covered: The article identifies some of the obvious and sometimes not-so-evident bottlenecks for going 'high-throughput' with imaging-based protocols. The article focuses on recent (2007 - 2010) studies, with an emphasis on live-cell screening with subcellular resolution and takes both instrumentation and experimental design into consideration. Given that most screens from pharmaceutical high-throughput microscopy are never disclosed, this review is necessarily biased toward published (academic) work.
Expert opinion: Despite high expectations on imaging-based drug-discovery strategies, both robotic subcellular imaging itself and data mining still present major challenges that will need to be overcome to make high-throughput microscopy as versatile and as powerful as high-end research microscopes. Going high throughput while keeping the high information content of fluorescence microscopic imaging will need concerted developments of hardware, image analysis and segmentation software, but probably most important is the design of live-cell assays that generate interpretable data. Also the physiological and disease relevance of the biological models amenable to high-throughput microscopies must be critically evaluated.