Download PDF

Oksana Sirenko | Scientist | Molecular Devices
Krishna Macha | Scientist | Molecular Devices

Introduction

Many drugs fail in the later stages of the drug development pipeline due to unacceptable toxicity effects. Drug failures in clinical trials are, in part, a result of insufficient predictive models used to screen drug candidates. Three dimensional (3D) organoids show promise to increase predictivity of in vitro assays suggesting the value of using organoids in pre-clinical testing.

The most common side effects of anti-cancer drugs are their toxicity to the intestine, which often limits the dose that can be administered to treat patients. In vitro assay using 3D organoids can help to evaluate the toxic effects of anti-cancer compounds in vitro and provide essential information in the process of drug development. Using automated high-content imaging increases throughput and extends the scope of information about toxicity effects, especially as they relate to complex 3D biology models. In this study we demonstrate how toxic effects can be evaluated and quantitated in intestinal organoids using high-content imaging.

In this application note, we present a method developed to evaluate toxicity using 3D mouse intestinal organoids cultured in Matrigel domes where concentration-dependent phenotypic effects of 10 compounds were tested.

Methods

Organoid culture

Primary Mouse Intestinal organoids from StemCell Technologies were cultured in Matrigel domes using IntestiCult media according to manufacturers recommended protocols. During organoid culture, automated media exchanges and monitoring by transmitted-light imaging were done every 24h. Organoids self-organized and developed complex crypt structures as expected for intestinal organoid phenotypes. For toxicity evaluation assays, organoid domes were seeded into 96-well plates (Ibidi plates), with 50% Matrigel domes, 15 µL per dome. Each dome contained approximately 60 organoids. Organoids were plated manually or with the CellXpress.ai™ Automated Cell Culture System. Compounds were added to organoids after 48h in culture. Compound treatments were prepared using a 7-point, 4-fold serial dilution starting from a 200 µM concentration, except for staurosporine, which started at 20 µM. Each dilution step reduced the concentration to 25% of the previous level. In addition, the controls were treated with 0.1% DMSO. Organoids were cultured with compounds for 3 days. After compound treatments, organoids were stained with Hoechst and MitoTracker orange, then fixed with 4% paraformaldehyde and additionally stained with Alexa488 Phalloidin in the presence of 0.05% of TritonX. All Dyes were from Thermo Scientific.

Organoid imaging

Organoids were imaged using the ImageXpress® HCS.ai High-Content Screening System with confocal option (60 µm pinhole) in three fluorescent channels DAPI, FITC, TRITC, and at 10X magnification. Next, 3x3 sites per well were taken at 10X magnification to cover the entire dome area. Additional images were taken using 4X magnification. Four tiled images were used to cover the organoid dome area. Z-stacks of 16 images were taken at 8 µm interval, covering approximately a Z-range of 120 μm. Maximum projection 2D images were used for analysis. For volumetric analysis, 3D Z-stacks of images were used.

Image analysis

Image analysis was done using IN Carta® Image Analysis Software. Within the IN Carta software, the Custom Module Editor (CME) was used to create a multi-step analysis protocol that defined organoids as blobs using the DAPI channel (nuclear stain) in projection images. Analysis was applied to maximum projection images. Numbers of organoids, average organoid area, and average organoid fluorescent intensities for DAPI (Hoechst stain), FITC (Alexa488 Phalloidin) and TRITC (MitoTracker) were measured. First, a Gaussian filter was used to blur the Hoechst signal to facilitate segmentation of organoids. Next, nuclei were segmented and used to define cells. Cells were then scored as positive or negative depending on the signal intensities for Phalloidin (actin cytoskeleton) or MitoTracker (mitochondria). Thresholds for positive and negative cell scoring were set empirically using control (untreated) samples and samples treated with toxic compounds (e.g. ailuropodine). Cells that were scored positive for actin or mitochondria were defined as having an intact cytoskeleton or intact mitochondria, respectively. Positive and negative cells per organoid were counted and the average area and average intensity of positive cells were measured. Average organoid volumes were evaluated by 3D CME analysis using a custom module which defined organoids in 3D as volumetric objects. After analysis, concentration dependencies for different readouts were plotted as 4-parametric curve fit (for 0.12–200 µM concentration range) to calculate EC50s for compound toxicity effects. SoftMax® Pro Software was used for the curve fit and calculation of EC50s.

Results

Ten compounds were tested for toxic effects. Cisapride was used as negative control, staurosporine was used as a positive control. Compounds were tested in 4X dilutions, 0–200 µM concentration range, in duplicates or triplicates. The toxicity evaluation assay was performed in 96-well plate format. Organoid cultures were set up using the CellXpress.ai system (see application note titled, Automated testing for compound toxicity effects using healthy intestinal organoids) but can also be set up manually. After compound treatment for 3 days, organoid domes were fixed, stained and imaged as described in Materials and Methods. Organoids were imaged using the HCS.ai system as described in the Methods section.

Figure 1 shows maximum projection images of cultured intestinal organoids stained with Hoechst, MitoTracker, and Phalloidin. Organoids typically vary in size and complexity, but all intact organoids have strong actin signal showing cytoskeleton (in green) and MitoTracker signals showing intact mitochondria (in orange). Intestinal crypts, typical for the intestinal organoid phenotype were also observed.

Image analysis determined the number of organoids in the dome and measured the organoid size (area) and fluorescent intensities with different markers. In addition, it identified individual cells and counted intact or damaged cells in the organoids. To quantify the intact cells, we scored cells as positive, or intact, if they had a high signal for actin or MitoTracker. In contrast, cells with low staining for actin or mitochondria were scored as damaged or dead cells. The thresholds between positive and negative cells were determined empirically by comparing positive and negative sample wells. Then the analysis was applied to the entire plate, including wells treated with different concentrations of compounds. The 4X magnification better captured phenotypes of whole organoids, while the improved nuclear and cellular resolution at 10X magnification better quantitated the numbers and percentages of positive and negative cells for different markers. Organoid density was important for results accuracy: greater density allows for better statistics in quantitation, however, if seeded too densely, organoids would overlap in the image and result in inaccurate segmentation (detection) of organoids. In our studies we had 60.3±18.2 organoids per well.

Organoids treated with compounds demonstrated significant changes in phenotype. Figure 2 shows images of organoids taken with 10X magnification. The organoid shapes morphed to a more rounded or collapsed phenotypes. Increased concentration of compounds resulted in decreased Phalloidin or MitoTracker stains.

Figure 1. Intestinal organoids (untreated control). Organoids were stained in Matrigel domes with Hoechst nuclear stain, and Alexa-488 Phalloidin, as described in Materials and Methods. Confocal Z-Stacks (16 planes, 8µm interval) in DAPI, FITC, TRITC channels were taken with the HCS.ai imaging system at 10X magnification. Maximum projection composite images for intestinal organoids are shown (Hoechst-blue, Phalloidin- green).

We then selected the following readouts for evaluation of phenotypic changes: number of cells with intact actin (actin positive cells) per organoid; number of actin negative cells per organoid (damaged cells); number of cells with intact mitochondria per organoid; total area of actin positive cells; average nuclear intensity; total area of actin positive cells; average volume of organoids. Cellbased analysis and quantitation of cell counts averaged per organoid and per organoid and well was the most efficient in quantitation of toxic effects. Such selection of readouts allowed us to quantitate different aspects of toxicity: disintegration of cytoskeleton/cell death, or collapsing cytoskeleton (by area); mitochondria integrity, DNA integrity (by Hoechst stain), inhibition of growth/ collapsing organoids by evaluation of volume. A CME rule was created that finds organoids using Hoechst stain, then defines individual cells and scored them as positive or negative using actin and mitochondria stains. Figure 3A shows several steps of CME analysis that use nuclear stain (DAPI channel) to define cell nuclei and the actin staining (FITC channel) to score cells as positive or negative. Figure 3B shows analysis masks for actinpositive and -negative cells in organoids. Image analysis demonstrates clear differences between treated and untreated samples as well as concentration-dependent changes in numbers of intact (live) cells or effected (negative) cells, cells with intact mitochondria, live cell area, and nuclear intensity.

Figure 2. Phenotypic changes caused by selected cytotoxic drugs. Confocal images of organoids treated with selected anti-cancer drugs. Samples in the picture were treated with control (0.1% DMSO), mitomycin (10 μM), trametinib (10 μM), cisplatin (10 μM), staurosporine (1 μM), and taxol (10 μM), respectively. Confocal images were taken with 10X magnification using Z-stack of 16 images 8 µm apart. Maximum projection composite images shown. Hoechst – blue, Phalloidin – green.

Figure 3. A. CME analysis masks show steps of finding cells with intact cytoskeleton, i.e. positive for actin staining or damaged cells, i.e. negative (weak) for actin staining. B. Analysis masks shown for untreated organoids and organoids treated with trametinib. Masks: Blue – organoids; Yellow – cells with weak actin stain; Dark blue – cells with intact actin; Pink – cytoplasm of cells with intact actin. Maximum projection images were used for analysis.

The bar graphs (Figure 4) show concentration-dependent changes for a subset of tested compounds. They also show decreases in average numbers of intact cells per organoid (positive for actin stain) and decreased numbers of damaged or dead cells (with decreased actin stain). Panel A shows the number of cells with intact cytoskeletons. Panel B shows the number of cells per organoid that have decreased actin staining. This phenotype was consistent with damaged or dead cells. Note the decrease of this number with very high concentrations of compounds, indicating that cells apparently fell apart and were not detected anymore by nuclear stain. In addition, a decrease in nuclear stain was observed for several DNA-intercalating agents, especially with doxorubicin and cytarabine. Panel C shows a dosedependent decrease in nuclear intensity. We also detected numbers of cells per organoid with intact mitochondria. There was a decrease in mitochondria-positive cells with increasing concentrations of drugs. We also evaluated average organoid volumes using nuclear staining. Interestingly, average volumes decreased with several compounds, suggesting that organoid growth was also inhibited with those compounds.

After analysis, numeric data for appropriate readouts were imported into the curve-fit software (SoftMaxPro) and effective concentrations for toxicity effects were determined. (Prism or any other software can be used for EC50 calculation). EC50s presented in Table 1.

The main observations can be summarized as the following:

All tested compounds, with the exception of cisapride, had toxic effects on intestinal organoids as measured by several morphological changes. Compound effects were most prominent when observing cytoskeleton integrity measured by actin staining. Mitochondria potential decreased with increased concentrations of compounds, but effective concentrations were typically higher than for cytoskeleton integrity, suggesting that mitochondria damage was not a primary mechanism for cytotoxicity. In contrast, DNA-intercalating compounds doxorubicin, mitomycin, cisplatin, and cytarabine decreased nuclear intensity which is consistent with expected mechanisms of action for those compounds. Decrease of average organoid volume was observed with most anti-cancer compounds, most notably with staurosporine, doxorubicin, mitomycin, cytarabine, and trametinib and was consistent with inhibition of cell proliferation that apparently has limited organoid growth.

Observation showed that actively proliferating, healthy intestinal microtissues were susceptible to the toxic effects caused by anti-cancer drugs and, as such, can be used for in vitro evaluation of anti-cancer drugs side effects.

Figure 4. The bar graph shows decreases in the numbers of actin-positive (intact) cells per organoids, averaged per well, and changes in other phenotypic measurements with increased concentrations of tested compounds. First bar of each series indicates control.

Table 1. EC50s for compound effects across different readouts (nd – not determined).

Discussion

High-content imaging measures and quantitates multiple effects of compounds reflecting various phenotypic changes: organoid size, marker intensities, count of intact or damaged cells, nuclear intensity, or mitochondria signal. This method is suitable for in vitro evaluation of toxic effects of drugs on healthy intestines. Depending on the experimental design, various additional markers can be used that would address other specific effects on cellular subtypes in organoids. This approach allows for the evaluation of not only cell death phenotypes, but also the effects on mitochondria, nuclei, or cell proliferation. A multi-parametric approach, as well as simultaneous analysis for different read-outs, can be further enhanced by extending the number of analytical readouts and using statistical methods for compound clustering to find similar effects and further insights into the mechanism of action.

Conclusion

We developed a method for compound screening using intestinal organoids in a 96-well plate format using highcontent imaging with the HCS.ai system and IN Carta image analysis software. The workflow shows how complex organoid models can be used for compound testing and toxicity assessment studies in automated protocols combining process automation and high-content imaging. The methods are suitable for toxicity assessment studies and allow evaluation of multiple phenotypic changes in complex organoids.

Download PDF