Table of Contents
Materials and Methods:
Cell Titer Glo Assay:
T47D breast cancer cell harvesting and NMR spectroscopy:
NMR data analysis:
Mesoscale AMPK activation:
Quantitative Reverse Transcription PCR:
Statistical analysis and hit evaluation:
Title: 3D high-content screening for the identification of compounds that target cells in dormant tumor spheroid regions
Authors: Carsten Wenzel1, Björn Riefke1, Stephan Gründemann1, Alice Krebs1, Sven Christian1, Florian Prinz1, Marc Osterland1, Sven Golfier1, Sebastian Räse1, Nariman Ansari2, Milan Esner3, Marc Bickle3, Francesco Pampaloni2, Christian Mattheyer2, Ernst H. Stelzer2, Karsten Parczyk1, Stefan Prechtl1 and Patrick Steigemann1
Affiliations: 1 Bayer Pharma AG, Global Drug Discovery, Berlin, Germany; 2 Physical Biology Group, Buchmann Institute for Molecular Life Sciences (BMLS), Goethe University Frankfurt, Germany; 3 Max Planck Institute of Molecular Cell Biology and Genetics, High-Throughput Technology Development Studio (TDS), Dresden, Germany
Conflict of Interests: All authors stated with 1 are employees of Bayer Pharma AG.
This work was supported by the German Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung, BMBF grant 13N11115 (ProMEBS)).
Keywords: 3D cell culture; multicellular tumor spheroids; tumor dormancy; 3D high-content screening; advanced light microscopy
Total number of tables and figures: 6 figures + 1 table (5 supplemental figures + 1 supplemental table, 5 movies)
illustration not visible in this excerpt
Cancer cells in poorly vascularized tumor regions need to adapt to an unfavorable metabolic microenvironment. As distance from supplying blood vessels increases, oxygen and nutrient concentrations decrease and cancer cells react by stopping cell cycle progression and becoming dormant. As cytostatic drugs mainly target proliferating cells, cancer cell dormancy is considered as a major resistance mechanism to this class of anti-cancer drugs. Therefore, substances that target cancer cells in poorly vascularized tumor regions have the potential to enhance cytostatic-based chemotherapy of solid tumors.
With three-dimensional growth conditions, multicellular tumor spheroids (MCTS) reproduce several parameters of the tumor microenvironment, including oxygen and nutrient gradients as well as the development of dormant tumor regions.
We here report the setup of a 3D cell culture compatible high-content screening system and the identification of nine substances from two commercially available drug libraries that specifically target cells in inner MCTS core regions, while cells in outer MCTS regions or in 2D cell culture remain unaffected. We elucidated the mode of action of the identified compounds as inhibitors of the respiratory chain and show that induction of cell death in inner MCTS core regions critically depends on extracellular glucose concentrations. Finally, combinational treatment with cytostatics showed increased induction of cell death in MCTS. The data presented here shows for the first time a high-content based screening setup on 3D tumor spheroids for the identification of substances that specifically induce cell death in inner tumor spheroid core regions. This validates the approach to use 3D cell culture screening systems to identify substances that would not be detectable by 2D based screening in otherwise similar culture conditions.
One of the main properties of cancer cells is sustained proliferative growth. Accordingly, the cell cycle is a major target for chemotherapy. Cytostatic drugs show strong anti-cancer efficacy in conventional in vitro assays; however, findings from 2D cell culture based experiments can only be partially translated to experimental outcomes in vivo and resistance to chemotherapy is still a frequent cause for treatment failure in patients with advanced and inoperable cancer. Several factors confer resistance to standard treatment regimens including, but not limited to, pharmacokinetic properties, genetic heterogeneity, drug clearance by cancer cells (1–4). As commonly used cytostatics mainly target proliferating cells, tumor cell dormancy could be a factor for a limited response to these compounds (5,6).
Tumor cell dormancy is influenced by regional differences in oxygen and nutrient supply within the neoplastic tissue, depending on the amount and quality of (neo-) vascularization (i.e. the distance from supplying blood vessels). As tumor growth requires high amounts of energy and nutrients, tumor cell proliferation is therefore mainly restricted to regions adjacent to blood vessels and human tumor tissue can show relatively low proliferative indices in poorly perfused areas (3,5,7,8). Cancer tissue can therefore be subdivided, depending on vascularization, into well-supplied, proliferating tumor cell regions in the vicinity of blood vessels and mostly dormant cells in poorly vascularized tumor regions.
Dormant cancer cells could potentially lead to disease relapse after cytostatic-based chemotherapy. Therefore, targeting this cell population could be of interest to enhance cytostatic-based chemotherapy (6).
Despite the potential role of dormant cells in limiting the effectiveness of cytostatic-based chemotherapy, few efforts have been made to specifically target this tumor cell population (6,9,10). This could be due to the fact, at least in part, that there is a lack of appropriate screening-compatible in vitro models that are able to simulate the metabolic microenvironment in tumors.
Recently, 3D cancer cell culture models have gained interest, as they have the potential to mimic the complex three dimensional organization of tumor tissue in vivo. Similar to native tumor tissue, cells cultured as multicellular tumor spheroids (MCTS) show strong proliferation gradients that reflect distribution gradients of oxygen, nutrients and energy, as well as the accumulation of metabolites from outer to inner spheroid regions (3,4,11–13). However, conventional 3D-based methods are not able to identify localized phenotypes in 3D models. Therefore we set up a high throughput, high-content microscopy compatible 3D MCTS assay on 384-well microtiter plates to identify substances that specifically target dormant cells in MCTS core regions. As a proof of principle, we screened two small compound libraries and identified nine hits that specifically target cells in inner tumor spheroid regions, while cells in outer regions or cultured under 2D cell culture conditions remain unaffected. We identified all hits as being inhibitors of the respiratory chain and further characterized their mode of action in MCTS. Finally, we show additive effects in combination therapy with selected compounds when combined with cytostatics in vitro.
Materials and Methods:
Spheroid generation was carried out using a modified version of the liquid overlay cultivation technique described previously (14). For the generation of imaging-compatible 3D tumor spheroids, 10 µl of a heated 1.5% w/v agarose (in DMEM without phenol red and fetal bovine serum (FBS)) solution was dispended by liquid dispensers (Multidrop Combi, Thermo Scientific) into sterile 384-well clear bottom imaging plates. To prevent premature gelation of the agarose suspension, the Multidrop and dispensing cassette was heated by infrared lamps. For tumor spheroid seeding, a single cell suspension was seeded into agarose-coated (1.5% w/v) 384-well clear bottom plates in 40 µl RPMI1640 containing 10% (v/v) FBS supplemented with 1% Penicillin/Streptomycin (and 0.01 µg/ml Insulin for T47D cells (Gibco)) using a liquid dispenser. Cell lines seeding number was optimized to obtain spheroids with an approximate diameter of 400 µm on day 4 and were seeded in following density: 2000 cells per well (c/w) for T47D, 5000 c/w for DLD1, 2000 c/w for DU145, and 1000 c/w for primary colon cancer cells. For schematic overview please see Supplementary Fig. S2.
The plates were incubated under standard cell culture conditions at 37°C and 5 % CO2 in humidified incubators for 4 days to allow formation of reproducible spheroids of defined size and morphology. In general approximately 50% of all tested cell lines are capable of spheroid formation in these conditions. As described by others (15) spheroid formation can be facilitated by addition of low percentage of reconstituted basement membrane preparation if cells are not capable of forming compact spheroids (e.g. BD Matrigel). Drugs (Enzo Life Sciences Screen-Well® FDA Approved Drug library (640 compounds) and Screen-Well® ICCB Known Bioactives library (480 compounds)) were added in 20 µl culture medium for additional 3 days.
Prior to imaging, spheroids were stained for 24 h by adding Hoechst 33342 (1 mg/ml, Life Technologies) as counterstain for all nuclei and Sytox Green, as stain for dead cells (2 mM, Life Technologies) at a final dilution of 1:10000 each.
One image per spheroid and wavelength, focused on the spheroid center were captured by Molecular Devices Micro widefield system with a 2X objective. Quantification of inner core cell death was done with MetaXpress software (Molecular Devices) using custom written image analysis routines. Briefly, spheroid borders were detected on Hoechst channel and masks were generated, scaled down and transferred on to the Sytox Green channel to quantify cell death in inner spheroid regions. All images were captured as 12-bit tiff files and no non-linear corrections have been applied.
3D image acquisition was done on a custom build mDSLM microscopy system as described earlier (16). Briefly samples were dehydrated in an ascending ethanol series (50%, 70%, 85%, 99%) for 5 min each. Then spheroids were transferred to benzyl alcohol/benzyl benzoate (1:1, v/v), transferred in a glass capillary and imaged using 2,5X illumination objective and 10X detection objective. 142 z-planes with 2,58 µm spacing were imaged for T47D spheroid in movie 1, 137 z-planes were imaged for antimycin A (100 nM) treated MCTS. 3D reconstruction and movie generation was done with Imaris software (Bitplane).
For 2D toxicity assessment T47D cells were seeded at 2250 cells per well (c/w) in 40 µl on 384-well plates and were allowed to attach for 24 h. After 3 days drug incubation, cell viability was determined. Hoechst (1 mg/ml) and Sytox Green (2 mM) were used for staining at a final concentration of 1:10000. Cell death index was calculated by counting all cells (as detected by Hoechst staining) divided by the number of Sytox Green positive dead cells.
Cell Titer Glo Assay:
Viability for in vitro combination studies in MCTS were measured with Cell Titer Glo Assay (Promega). To support reagent penetration, lysis and ATP recovery from MCTS an equal volume of reagent was added to sample and shaken for 15 min at 450 rpm. Luminescence readout was done after 30 min incubation at room temperature.
Prior to harvesting, spheroids were fixed for 24 h in 4% PFA. Then spheroids were transferred to 50 mL tubes (Falcon), washed twice in ice-cold DPBS and equilibrated in 30% sucrose (w/v) DPBS solution for 1 h. Then spheroids were transferred to cryomolds and covered in Tissue-Tek OCT compound. After 30 min of equilibration cryomolds were frozen by incubation in a mixture of dry ice and 2-Methylbutane (Sigma Aldrich). Prepared samples were cut into 5 µm sections by cryostat, mounted on SuperFrost Plus slides (Menzel-Glaser) and then rehydrated in DPBS for 20 min. After 1 h in blocking and permeabilization solution (1% BSA, 0.1%Triton, 0.1% TWEEN-20) the primary antibody was incubated over night at 4°C. Incubation and staining with Hypoxyprobe-1 kit (Hypoxyprobe-1, Chemicon), mouse monoclonal IgG1 labeled with FITC and cell labeling with Click-iT EdU imaging kit (Alexa Fluor 555 azide, Life Technologies) were done according to the respective manufacturers instructions. After staining, slides were mounted in Slowfade Gold (Life Technologies) and imaged on AxioInvert 500 (Carl Zeiss) with 10X air objective and attached camera.
T47D breast cancer cell harvesting and NMR spectroscopy:
For the extraction of metabolites and sample preparation for 1H NMR spectroscopy the protocol from (17) was adapted to T47D breast cancer cells. In brief, cells were washed, methanol quenched and transferred for subsequent extraction. Spectra of extracted aqueous metabolite phase were acquired in 3 mm NMR tubes at 600.13 MHz and 300 K using a Bruker AVANCE III spectrometer equipped with a TCI-Cryo-Probe and a sample jet system (Bruker BioSpin). The residual water signal was suppressed by a 1D-NOESY presaturation pulse sequence. Typically, a total of 512 transients each of 64 k data points was acquired with an acquisition time of 2.65 s, an interpulse delay of 4 s, a spectral width of 20 ppm and a pulse width of 8.2 µs at 5 dB (90°). The free induction decay (FID) was multiplied by a 0.3 Hz exponential line-broadening factor to improve the signal-to-noise ratio prior to Fourier transformation. Phase correction and referencing was performed using Topspin 2.1 (Bruker BioSpin). For the baseline correction ACD Software Suite 12 (ACD/Labs) was used. The TSP signal was set to 0.00 ppm.
NMR data analysis:
Multivariate analysis of the integrated bucket data was performed using SIMCA-P software (version 13.0, Umetrics AB) applying pareto scaling. Unsupervised principal components analysis (PCA) and supervised models (PLS-DA, OPLS-DA) were used to extract the main drivers, or spectral regions of the spectra responsible for group separation.
- ISBN (eBook)
- ISBN (Buch)
- 690 KB