Biomedical Engineering Reference
In-Depth Information
expected phenotype can be identified during manual image review, and the analysis
algorithm can be adjusted to measure additional parameters characterizing the unex-
pected effect. For example, in an image-based screen aiming to identify perturbagens
involved in ciliogenesis by quantifying the percentage of ciliated cells, manual image
review identified a phenotype with increased cilia length [26]. Subsequently, the size
of the cilia detected was added to the analysis algorithm, and the project succeeded
in identifying perturbagens for a decrease in number of ciliated cells as well as a
stabilized cilia phenotype. This exemplifies the potential of image-based screening
assays to identify unexpected phenotypes or to distinguish subtle differences in phe-
notypes that may provide additional insight into cellular profiling of DOS compound
libraries.
Extraction and accurate quantification of multiple biologically relevant effects
allows for rapid characterization of compounds, making image-based high-content
assays an ideal choice for screening of DOS-based libraries. However, image-
based assays have inherently lower throughput than that of cellular plate reader-
based assays, due to longer image acquisition times (typically, 15 min to 2 h per 384-
well plate, depending on the assay design). Specialized (and expensive) high-content
screening equipment and infrastructure are required to prepare, image, and analyze
image-based assay plates [15]. Additionally, advanced knowledge in image analysis,
population statistics, and data mining is paramount to being able to mine and correctly
interpret the rich biological information resulting from these imaging assays.
As the image-based high-content screening field matured over the last decade,
new image-based screening modalities emerged to better characterize the biologi-
cal systems. The recent increase in image acquisition speeds and improvement in
image handling, analysis, and storage allowed for observation of live cell phenotypic
changes over time. Temporal phenotypic changes can be evaluated for short periods
(several seconds) at fast imaging rates (10 to 100 Hz) in kinetic image cytometry,
where each well is imaged for the entire duration before moving to the next well.
Examples of kinetic image cytometry assays are assays measuring action potentials
(calcium waves) of cardiomyocytes or ion channel activity. Time-lapse imaging per-
forms multiple imaging runs of an entire well plate over long periods of time (hours to
days): for example, to evaluate subtle changes in cell cycle and mitosis [27,28]. Live
cell imaging requires high-content screening systems with environmental control
to maintain incubation conditions during imaging and an image-data infrastructure
to deal with terabytes of data per experiment. The same technological advances
that allowed for live cell imaging also enabled investigating more complex biolog-
ical systems by evaluating the phenotypic changes in three-dimensional biological
model systems, such as tumor spheroids, or small organisms such as
Caenorhabditis
elegans
. Specimen preparation for such assays required sophisticated assay design
and is often the limiting factor as to throughput for these types of assays. However,
they mimic the true biological system or disease better than do two-dimensional cell
monolayers, and an increase in the use of these assays should be expected as plate
preparation methodologies and imaging techniques improve.
The advances in technologies described above have also helped to push some func-
tional cell assays to the medium throughput required for screening compound libraries
