Download PDF

Oksana Sirenko, Astrid Michlmayr, Emilie Keidel, Marco Lindner, Jeff McMillan, Angeline Lim, Zhisong Tong, Bruce Gonzaga, Felix Spira | Molecular Devices

Introduction

Attrition in the therapeutic pipeline can often be attributed to the lack of translational efficacy from the pre-clinical phase to the clinic. Organoids show great promise as a game-changer in disease modeling and drug screening since they better resemble tissue structure and functionality and show more predictive responses to drugs. However, challenges associated with the practical adoption of organoids, such as assay complexity, reproducibility, and the ability to scale up have limited their widespread adoption as a primary screening method in drug discovery.

To alleviate the bottlenecks that come with labor-intensive manual protocols, we developed the CellXpress.ai™ Automated Cell Culture System. This powerhouse workstation automates the entire organoid culture process for prolonged complex workflows. The CellXpress.ai system provides media exchange, plating, passaging, organoid monitoring, endpoint assay execution, and performs complex image analysis. Here we present results from the automation of several commonly used organoid protocols, including the culture of 3D organoids in matrix domes. Healthy intestinal organoids were cultured, passaged, and expanded in Matrigel® domes (24-well). Organoids were cultured with automated media exchange and monitored by machine learning-assisted imaging every 24 hours. After 5–6 days, organoids were automatically collected, purified from Matrigel, and dispersed then mixed with fresh Matrigel and re-plated. Organoids then self-organized and developed complex crypt structures. Organoids were monitored in transmitted light and machine learning-based image analysis was used to determine organoid number, size (by area), and optical density. For endpoint assay (96-well), organoids were stained for viability markers and monitored for concentration and time-dependent effects of compounds on healthy intestinal organoids (toxicity evaluation), or patient-derived colorectal cancer organoids (drug screening).

Automated cell culture processes powered by imaging and machine learning-controlled decision-making have great potential to bring 3D biology to another level, allowing for increased throughput and reproducibility for drug discovery and disease modeling applications.

Methods

Instrumentation

The powerful new CellXpress.ai automated cell culture system automates the entire cell culture process with an integrated incubator, liquid handler, and image-based decision-making. This hands-off automated solution manages demanding feeding and passaging schedules by monitoring the development of cell cultures with periodic imaging and analysis, and leverages machine learning to initiate passaging, endpoint assay, or troubleshooting steps.

Fluorescent (FL) images were acquired on the ImageXpress® Confocal HT.ai High-Content Imaging System (Molecular Devices) using MetaXpress® High- Content Image Acquisition and Analysis Software. For intestinal organoids, Z-stack images were acquired with the 4X or 10X objectives using confocal mode.MetaXpress or IN Carta® Image Analysis Software was used for all analysis.

Cell culture protocols

We’ve developed a protocol for 3D intestinal organoids* that were derived from primary mouse intestinal cells using established methods (STEMCELL Technologies). Cells were cultured and differentiated according to the STEMCELL Technologies protocol. IntestiCult™ Organoid Growth Medium (STEMCELL Technologies) was utilized for cell culture. Cells were seeded in 50% growth-factor reduced Matrigel or Cultrex (Corning) domes in a 24-well plate format and were fed every second day with fresh media for 7–10 days. Intestinal organoids were then passaged- dissociated and re-plated into fresh Matrigel domes.

Figure 1. The CellXpress.ai cell culture system components and functionality

Organoid cell culture automation

Organoid culture in Matrigel domes was performed according to the basic STEMCELL Technologies recommended protocol for mouse intestinal organoids. Organoids were grown and passaged in 24-well plate format with single dome 40–50μl, 50–60% Matrigel or Cultrex.

Organoid plating was started from organoids suspension in Matrigel which was placed into a pre-chilled 96-well deep well block. The suspension was pipetted by 4 pipette tips and dispensed into 24-well plate, by 4 tips at a time. Plating into a 96-well format was also tested.

Organoid feeding was done by media removal/addition of fresh media, 4 wells at a time.

Organoid imaging/monitoring was done in transmitted light using CellXpress.ai automated cell culture system with 2x or 4x magnification. Image analysis was performed using a machine learning-based protocol. The analysis evaluated organoid number, mean and total area, density, and several other measurements. Passaging organoids was set for every 4 days, or as user-directed, or depending on automatic decision-making based on one or more selected measurements (e.g. total area or organoid density).

Organoid passaging was done using a combination of pipetting steps and external centrifugation steps to optimize for the mouse organoids workflow. Modification of the flow rates, pipetting steps and repeats, centrifugation speed, etc. can be customized by changing appropriate “fine-tuning” steps. For the passaging process, media was removed and Matrigel domes were incubated with Gentle cell dissociation reagent, then rigorous pipetting was done to break Matrigel domes. The mix was harvested into a 96 deep-well block. Then optional pooling of two wells into one was performed, followed by centrifugation on an external centrifuge with a speed of 400g for 5 min. The block was then placed back, the media was removed, and organoids were washed once. After a second centrifugation, most of supernatant was aspirated, and then organoids were broken with rigorous pipetting using smaller tips. Then fresh Matrigel was added to the appropriate volume, mixed, and seeded into a new plate.

Organoid staining/imaging for endpoint assay was done using FL imaging with the ImageXpress Confocal HT.ai instrument.

Figure 2. Schematic diagram of automated organoid culture and passaging protocol

Automated organoid workflow

Figure 3. Steps of the organoid culture and passaging protocol

The workflow steps that include seeding, media exchange, monitoring and passaging of intestinal organoids are programmed by using pre-set “phases” of the protocol.

First phase is seeding which programs step of placing Matrigel organoids into 24 well plates.

The next phase of the protocol includes feeding/ monitoring/passaging steps with timing and parameters defined by user.

Figure 4. Steps of the organoid culture and passaging software protocol

Results

We have performed continuous culture of intestinal organoids for more than a month. Automated seeding of organoid domes allowed consistent size and accurate centering of domes in 24-well wells, making imaging and image analysis easier and more accurate. Organoids grow as expected during automated culture, forming protrusions consistent with typical intestinal organoids morphology. Distribution and number of organoids was consistent across the wells. Images were taken daily using 2X objective and image analysis was done using pre-trained machine-learning model. The image analysis allowed to find organoids, also determine number of organoids, mean and total area, organoid density, roundness, granularity, and other morphological criteria. The software allows to review organoid domes during the culture, also provides analysis on the fly and time course plots representing various measurements. Data also can be exported into Excel datasheet for additional analysis. Figure 6 shows changes of organoids area sum and number of organoids over the time.

Figure 5. A. Representative images (4X) of organoid culture taken at different time points of the continuous culture. B. Graphical representation of organoid analysis: organoid count and skewness over time. Skewness includes a combination of optical density, granularity, and other optical parameters. The value is useful in defining the time for organoid passaging. C. Imaging of a single well over time showing the “cell journey”. D. Tiled 4X images of organoid domes from a 24-well plate.

Figure 6. Presentation of organoid count and organoid total area (area sum) over time. Averages and STDEV were calculated from 24 wells of the plate.

Summary

• Organoid technologies are considered a game-changer in disease modeling and drug screening, since they better resemble tissue structure and functionality, and show more predictive responses to drugs.
• Challenges associated with the practical adoption of organoids, such as assay complexity, lack of reproducibility, and the ability to scale up screening have limited their widespread adoption as a primary screening method in drug discovery.
• We demonstrate here how researchers can alleviate the bottlenecks that come with labor-intensive, complex protocols to increase productivity for drug screening.
• A powerful new solution, CellXpress.ai Automated Cell Culture System, allows labs to automate the entire cell culture process—from assay set-up to screening and data analysis—with machine learning-powered workflows to make assays more reliable and reproducible.
• Automated cell culture processes powered by imaging and machine learning-controlled decision-making have great potential to bring 3D biology to another level by increasing throughput and reproducibility, and enabling a variety of high-throughput drug discovery and precision medicine applications.

*HUB Organoid Technology used herein was used under license from HUB Organoids. To use HUB Organoid Technology for commercial purposes, please contact bd@huborganoids.nl for a commercial use license.

Download PDF