Phenotypic Drug Discovery in 3D : Testing Drug Efficacy by Rapid Size Profiling Over Time with Tumor Microtissues

Full text

GEN Genetic Engineering &
Biotechnology News
Vol. 34, No. 16, September 15, 2014
© Mary Ann Liebert, Inc.
pp. 18–19
DOI: 10.1089/gen.34.16.08
Tutorials Drug Discovery Tutorial
Phenotypic Drug Discovery in 3D
Testing Drug Efficacy by Rapid Size Profiling Over Time with Tumor
Microtissues
Randy Strube, Ph.D., Andreia F. Fernandes, Markus Furter, Jens M. Kelm,
Ph.D., and David A. Fluri, Ph.D.
t is commonly accepted that tumor sensitivity or resistance to
chemotherapeutic agents is not only genetically determined, but also
driven by the tumor microenvironment. Metabolic gradients and the extracellular
matrix can influence nutrient availability to various segments of a solid tumor, and
also limit the ability to administer an effective dose of chemotherapeutic agent.
These biological and therapeutic gradients result in the formation of
phenotypically different tumor cell subpopulations exposed to a variety of
therapeutic to subtherapeutic drug doses in vivo.
Tumor size is the most frequently used in vivo endpoint when assessing
antitumor efficacy in animal xenograft models, whereas proliferation is the more
typically evaluated growth endpoint in vitro using two-dimensional (2D)
monolayer cultures. Such 2D in vitro assays frequently fail to correlate with in
vivo observations, owing to the inability of 2D cultures to recapitulate the native
tumor microenvironment described above. Three-dimensional (3D) tumor
microtissues, or multicellular tumor spheroids, are considered a more
representative, organotypic model for assessment of tumor growth. They contain
layers of cells that exhibit more in vivo-like size- and gradient-dependent
proliferation and viability profiles.

Randy Strube, Ph.D. (randy.strube@insphero.com), is director of global
marketing, Andreia F. Fernandes and Markus Furter are scientists, Jens M.
Kelm, Ph.D., is CSO, and David A. Fluri, Ph.D., is senior scientist at InSphero.
Website: www.insphero.com. Third-party trademarks: CellTiter-Glo® is a
registered trademark of Promega.
I

For example, more proliferative cells tend to cover the outer oxygen and
nutrient-rich layers of the spheroid, whereas the nutrient-restricted inner cells
tend to display a more quiescent or even necrotic phenotype, as the spheroid
diameter increases. Additionally, the ability to monitor spheroid growth in terms of
size provides a similar unit of measure with which to make a direct in vitro to in
vivo comparison.
Although the benefits of 3D culture models have been widely
acknowledged, their widespread implementation in drug discovery screening
efforts has been slowed by the lack of appropriate companion assays and
instrumentation to maximize their value. Imaging spheroids to monitor their
growth using conventional or even automated microscopes is a slow, low-
throughput process. High-content imaging systems overcome some of these
throughput issues, but are comparatively expensive, and still suffer from slow
image capture/processing, and image analysis software that has not been
optimized for 3D spheroids.
Although adaptable to higher-throughput plate readers, biochemical
assays that monitor viability can effectively correlate to spheroid size, but are
often lytic or otherwise destructive in nature, requiring multiple replicates to be
run for longitudinal studies.
For this study, we aimed to improve the throughput of monitoring tumor
spheroid growth in response to anticancer drugs in vitro using the Dainippon
SCREEN Cell3iMager, a rapid, robust brightfield imager for measuring spheroid
size and morphology.
Rapid Assessment of Spheroid Size and Morphology
The Cell3iMager (Figure 1) facilitates analysis of spheroids by fast, parallel
scanning in a brightfield, allowing determination of spheroid count per well,
diameter, area, pseudo volume, and loss of circularity in individual spheroids.
Image capture is rapid, requiring less than 1 minute per plate at 2,400 dpi (10.6
µm/pixel). The 4-plate stage can accommodate 6-well to 384-well plates, with
resolution up to 9,600 dpi (2.6 µm/pixel). The reagent-free, label-free system
allows faster sample processing (30 × 384-well plates/hour vs 2.5 plates/hour
with conventional systems), and convenient measurement of spheroid growth
over time in a nondestructive way.

The software provided with the Cell3iMager offers multiple analysis
options, and provides flexibility to customize scanning recipes for automated
removal of debris (e.g., dust, fibers) and compensation for undesired well effects
such as shadowing.
Drug Sensitivity Testing
Drug sensitivity of tumor microtissues derived from the colon cancer cell
line HCT116 to two clinically relevant cytostatic drugs, gemcitabine and
docetaxel, was tested in a time course study. HCT116 spheroids formed in
GravityPLUS™ hanging drop plates were transferred to GravityTRAP™ assay
plates after aggregation. Both compounds were tested at three different
concentrations, re-dosed with compound dissolved in fresh medium at day 3, and
then monitored for growth over a 7-day incubation period on the Cell3iM-ager.
Exemplary scan images at day 0, day 4, and day 7 are shown (Figure 2).
Stitched well-images are automatically generated by the scanner software, and
extracted information is analyzed according to customizable analysis recipes.
The Cell3iMager software outputs different characteristic measurement
parameters such as the spheroid area, pseudo volume, count, and circularity of
measured microtissues. Automated analysis is fast, with turnover times in the
range of 30 seconds per 96-well GravityTRAP plate.
The scan overviews shown in Figure 2 provide qualitative information on
size and morphology changes for the different treatment groups. Quantitative
data is captured by the measurement software and presented either as graphical
output (histograms, line plots, etc.) or as raw data in a standard comma delimited
format (.csv).

For the studied HCT116 tissues, a size increase is observed for the
control and lowest compound treated groups, visualized by plotting the spheroid
area in µm2 over time (Figure 3, left). For gemcitabine, growth is inhibited at
concentrations of 4 nM and higher. At the maximum concentration tested (20
nM), tissue sizes start decreasing after day 5, indicating not only growth arrest,
but also actual cell death (Figure 3, top left).
Docetaxel sensitivity of the tested HCT116 spheroids is lower compared to
gemcitabine. Microtissues treated with concentrations of 4 nM still increase in
size for the entire time course although at reduced growth rates compared to the
control group (Figure 3, lower left). At the highest concentration, tissue growth
plateaus after day 3, indicating strongly reduced proliferation rates of the cancer
cells.

The scanner software allows easy and quick generation of quantitative
data for treated microtissues. In order to compare results acquired by growth
profiling to biochemical readouts, viability of the HCT116 tissues was quantified
at the end of the study utilizing a luminescence-based ATP assay (CellTiter-Glo®
2.0 Assay, Promega). As expected, ATP content of microtissues decreases with
increasing compound concentrations for both compounds, gemcitabine and
docetaxel (Figure 3, bars). Overall, the decrease in viability exhibits good
correlation with the measured tissue sizes at day 7 (Figure 3, lines).
Owing to its nonlytic properties and the fact that microtissue biology is only
minimally impacted by repeated measurements, the here described size profiling
assay using the Cell3iMager provides a powerful solution to complement or
replace existing endpoints for efficacy testing with 3D cultures. GEN