Download PDF

Prathyushakrishna Macha, Oksana Sirenko, Jay Hoying,

Angeline Lim | Molecular Devices, LLC

Introduction

With the onset of extensive 3D biology studies, we have come to a crossroads of transitioning from 2D biologybased research to more accurate and reliable 3D models. Relying on more informative and complex organ-like structures developed from stem cells or tissue samples seems inevitable; these organoid models are paving the way to next-generation, animal-free research models.

Organoids hold great promise in applications including disease modeling, drug discovery and development, personalized medicine, cell biology processes, and genetic disease research. In cancer research, there is specific interest in patient-derived organoids (PDOs) created using cells from patient-derived samples (biopsies, surgical resections).

Tumor heterogeneity and genetic variations within tumors challenge drug treatment strategies and necessitate several lines of treatment, which can further lead to the development of a resistant population. While there are several traditional in vitro assays for drug response studies, we need a better combination of 3D models, biology handling systems, and assays to come up with efficient ways to create and manage true representatives of in vivo complexity. Ways that can help ease pain points such as timely handling of complex 3D models, scaling up workflows, automated monitoring of assay steps, improved data acquisition, and deeper analysis of therapeutic drug discoveries.

We selected colorectal cancer (CRC) organoids as our model for drug response studies using known anti-cancer compounds to probe the utility of robotics-based assay development.

The BioAssemblyBot® 400 (BAB400) (refer to Methods, Table 1 and [1]) was used to automate CRC dome assays to enable automation and the uniform distribution of patient-derived colorectal cancer organoids. The seeded organoids were maintained and treated over a period of six days, during which an image-based end-point assay was performed to study drug responses.

Matrigel dome 3D expansion and assay development using the BAB400

The BAB400 pipette hand was used to dispense organoids in Matrigel® domes. Then organoid cultures were monitored over a period (here, six days) using imaging for time-course observation and endpoint assays.

Methods

BAB400

The BAB400 is a cGMP-certified, multipurpose, robotic laboratory automation platform used to build, manage, and assay 3D cellular and tissue models. Its pneumatic and mechanical dispensing, 3D printing, and plate handling capabilities enable it to build and manage a variety of different 3D models—including organoids— efficiently and without damage to the tissue constructs. The intuitive Human Machine Interface (HMI) allows for effective experimental design and implementation, recreating protocols developed by hand or de novo. The BioApps™ Creator and Player enable easy-to-make and easy-to-run automation plans entailing start-tofinish 3D model fabrication, culturing, media additions/ exchanges/collections, and imaging. The BAB400 increases efficiency, reproducibility, and scalability while standardizing culture conditions and removing manual errors. In combination with peripheral instrumentation, tasks such as cell seeding, media exchange, drug treatment, and organoid imaging can be automated to provide relief from redundant tasks to better focus on model development and testing strategies. For example, integration with a high-content imager, such as the ImageXpress® Confocal HT.ai High-Content Imaging System, enables automated image capture of multiple organoids in a multi-well plate format, providing automated monitoring of organoid growth, morphology, and assay parameters over time.

Imaging

Transmitted light (TL) or fluorescent images were acquired on the ImageXpress Micro Confocal High-Content Imaging System using MetaXpress® High-Content Image Acquisition & Analysis Software (both from Molecular Devices). For CRC organoids, Z-stack images were acquired with 4X or 10X objectives using confocal mode. MetaXpress software was used for all analyses.

Figure 1. Drug screening assay workflow showing Matrigel dome dispensing with CRC organoids.

Maintenance and general assay setup of organoids for drug screening

Colorectal cancer organoids were maintained as per their proprietary protocol.

A vial of frozen CRC organoids (Molecular Devices) was thawed in a 37°C water bath. With a few ice crystals left in the vial, the contents were aseptically transferred to the cell culture hood. We used a proprietary method to culture and plate the organoids into multi-well plates for assay. The assay seeding density of organoids depended on the assay requirement. We used 80% GF-reduced Matrigel (Corning, Cat. No. 356231) and a final seeding density of 200 organoids/ 10 μL. Using the BAB400, we dispensed 7 μL into each well of a 96 multi-plate. During the entire process Matrigel was kept on ice and pipetting was slow to avoid bubbles and minimize organoid fragmentation. The Matrigel domes containing organoids were gelled for 10 minutes on the stage set to 37 °C followed by the addition of 100 μL media (DMEM/F12 Gibco, Cat. No. 11039, + 10% FBS, B27-Invitrogen [Thermo Fisher Scientific], ROCK inhibitor [Tocris Bioscience], n-Acetyl cysteine [Sigma-Aldrich]) to each well. Media was switched on day two to add select drugs for drug screening assays in treatment media (DMEM/F12 Gibco, Cat. No. 11039, + 10% FBS, n-Acetyl Cysteine, B27 – Invitrogen).

Compound treatment

Several anti-cancer compounds (Tocris), including cisplatin (100 μM), romidepsin (10 μM), trametinib (40 μM), and 5FU (5 mM) at their starting concentrations and a 5-fold dilution series to study the dose-response. The endpoint assay was done on day six.

Staining method for live imaging

Here we used one of the two methods for staining for live and dead.

Method 1 – The organoids were stained with a live and dead kit that has acridine orange and propidium Iodide (TheWell Bioscience) with Hoescht 33342 (Thermo Fisher Scientific) for 30 minutes. The propidium staining postfixing and permeabilization is significantly lost and was imaged followed by post-fixation dyes.

Method 2 – MitoTracker™ Orange (Thermo Fisher Scientific) at a final concentration of 100 nM was used to stain the mitochondria with an overnight incubation of organoid domes. Next, these organoid domes were stained with calcein AM (Thermo Fisher Scientific), and Hoechst 33342 with a final concentration of 2 μM, and 2 μg/mL, respectively, at 37˚C in an incubator for 30 minutes. Finally, the organoids were imaged using a confocal high-content imager (Molecular Devices).

In case you need to stain for dead cells with ethidium homodimer, MitoTracker overnight staining can be removed from the steps to add ethidium homodimer III at a concentration of 1 μM (Biotium) along with the live and nuclei stains.

Figure 2. BAB400 uses the pipette hand (left) to dispense domes containing organoids (right) into select wells of the micro-well plate.

Table 1. Average time comparison chart of the BAB400 vs. manual execution of assay steps.

Staining method for fixed imaging

To perform fixed organoid imaging for better long-term preservation and imaging, these organoid domes were fixed with 4 % paraformaldehyde for 30 minutes. They were then stained with phalloidin green (Thermo Fisher Scientific, final conc. 5 µM), and Hoechst 33342 (final conc. 1 µg/mL) in the presence of a permeabilization agent (0.1 percent TritonX-100, Thermo Fisher Scientific), also you can add MitoTracker orange along with the above-mentioned dyes in case of the first live staining method.

Results

Figure 3. Uniformly seeded CRC organoids (~150 organoids/well) treated after 48h of seeding for four days with romidepsin (10 μM) in 5-fold dilutions (columns A1 – A4), and cisplatin (100 μM) in 5-fold dilutions (columns A5 – A8). The green panel indicates the cells stained with acridine orange for live imaging, and the blue panel below indicates the cells stained with Hoechst 33342 for nuclei count.

Figure 4. Uniformly seeded CRC organoids (~150 organoids/well) treated for four days with trametinib (40 μM) in 5-fold dilutions (columns A1 – A4), and 5FU (5 mM) in 5-fold dilutions (columns A5 – A8). The green panel indicates the cells stained with acridine orange for live imaging, and the blue panel below indicates the cells stained with Hoechst 33342 for nuclei count.

Figure 5. Images of CRC organoids on day seven in transmitted light (A.) and stained (B.) with live dead markers – acridine orange (green), propidium iodide (red), and Hoechst (blue). In the untreated fixed control samples (C, and D) the green represents phalloidin green (a marker for actin), blue represents Hoechst (a marker for nuclei), and the red represents MitoTracker orange (a marker for mitochondria). Similarly, the treated (E – cisplatin, F – romidepsin) and stained with phalloidin (green) and Hoechst (blue) were observed.

Figure 6. CRC organoids’ control and treated confocal images, 4X (top panel) and their representation of cell scoring analysis masks (bottom panel) generated by MetaXpress software (graph) A. Untreated B. Trametinib (40 μM) C. 5FU (5 mM) D. Staurosporine (10 μM) E. Graphical representation of the live cell numbers generated from the analysis (live staining Method 1)

Figure 7. CRC organoids’ confocal images, 10X (top panel) and their representative cell scoring analysis masks generated by MetaXpress software (bottom panel) A. Romidepsin (10 μM) B. Cisplatin (100 μM) C. Untreated controls. D. Graphical representation of the organoid area numbers generated from the analysis (live staining Method 1)

Figure 8. CRC organoids’ confocal images, 10X (top panel) and their representative cell scoring analysis masks generated by MetaXpress software (bottom panel) A. 5FU(5 mM) B. Trametinib (40 μM) C. Untreated controls. D. Graphical representation of the organoid area numbers generated from the analysis (live staining Method 1).

Results and summary

Organoid seeding uniformity and drug response

CRC organoids in Matrigel domes were plated in multiple 96-well plates. They were observed and treated with a known cancer drug panel for four days starting on day two after plating. Prior to the drug treatment, the organoids were uniformly distributed in the domes, located in the same position of each well facilitating automated imaging and organoid tracking over time. Drug effectiveness was assessed on day six using live/dead assay methods for live/dead counts and organoids area. The organoids were treated with the following – Plate 1 (Figure 3): romidepsin, cisplatin, Plate 2 (Figure 4): 5FU, and trametinib (for concentrations and dilutions, refer to Methods). The Z’-values for the drugs were as follows: romidepsin (10 μM)–0.58, cisplatin (100 μM)–.48, 5FU (5 mM)–0.63, and trametinib (40 μM)–0.69.

Dose-response analysis of individual drug effects on patient-derived colorectal organoids

Drugs had both cytostatic and cytotoxic effects. A decrease in both organoid sizes and cell viability was observed as a result of drug treatment. The organoids in Figures 3 and 4 show overall drug effects in the 96-well seeded organoids. Some of the organoids showed growth-arrested responses and stayed in the same size range as day two due to the growth suppression/cytostatic effects. Whereas most of the higher concentration-treated organoids disintegrated into smaller fragments and lost their morphology, and had higher dead cells due to cytotoxicity effects.

Staurosporine (Figure 6 D) was the most cytotoxic, followed by trametinib (Figure 6 B, 8 B). Cisplatin was the least cytotoxic in the treatments carried out. Though cisplatin was the least cytotoxic among the drugs, it led to morphological changes in the organoids compared to the controls. The organoids treated with cisplatin appeared hollow inside and significantly smaller in size (Figure 7 B). The cytotoxic effects on the organoid shape and size are shown in Figures 6, 7, and 8 with their corresponding graphical representation. Using 5-FU in higher concentrations (Fig. 6 C and 8 A) showed considerable cytotoxicity effects, as did romidepsin (Fig. 5 F and 7 A).

Summary

We successfully used the BAB400 lab automation platform to build, treat, and assay CRC organoids in Matrigel domes in a 96-well format (Figures 2, 3 and 4). Workflows involved dispensing domes into plates and culturing these plates. Organoids were monitored with additional incubation and compound treatments subsequently outside of the BAB400 workflow. After compound treatment, organoids were stained with viability dyes and imaged using the Image Xpress Micro Confocal system and its appropriate filters. Then cell scoring analysis was used to evaluate compound effects. Compounds used had an expected impact on organoid growth: treated organoids were greatly reduced in size and changed in morphology, and viable cell numbers decreased in comparison to control samples. Z’-values between treated and untreated samples were greater than 0.5. This value was obtained using the following parameters: means (µ) and the standard deviations (σ) of the positive and negative controls. An assay and instrument can be considered excellent when the Z´-value is greater than 0.5, while a Z’-value below 0.5 could be considered marginal—not just for the model system but also for the efficacy of the assay. The assays we conducted using the BAB400 and CRC model systems in Matrigel were above 0.5 and can be considered an efficient assay for medium high-throughput compound testing applications.

References

1. Automated dispensing, monitoring, and assay development of hydrogel-based cellular models

Download PDF