Evident LogoOlympus Logo
Application Notes

Optimizing Image Acquisition for High-Throughput 3D Analysis of Cancer Spheroids


Summary

In this application note, we determined the optimal image acquisition settings of a confocal microscope to obtain high-speed images with the quality required for 3D analysis of cancer spheroids. High-throughput drug efficacy evaluation was performed using a 3D InSight™ microtissues plate from InSphero.

・Drug evaluation workflow

Drug evaluation workflow

・Optimization of imaging conditions

Optimization of imaging conditions
 

Benefits

  • Optimize the FV3000RS microscope’s acquisition settings to obtain high-speed images with the quality required for NoviSight™ 3D analysis.
  • NoviSight software can analyze objects at low image quality since it uses signal information from the recognized object based on the nuclei signal with a high signal-to-noise ratio.
     

Introduction

Assessing drug performance using three-dimensional cancer spheroids is important because they reflect the complicated in vivo microenvironment of the cancer. This enables researchers to evaluate a drug’s effectiveness under conditions that more closely resemble a tumor’s natural environment.

However, acquiring optical slices of a large number of samples can be time-consuming. To speed up this workflow, we determined the optimal image acquisition settings of the FLUOVIEW™ FV3000RS confocal microscope to obtain high-speed images with the quality required for NoviSight™ 3D analysis. 

By optimizing the acquisition conditions, including the lens correction collar, step size in the depth direction, and average times,  we captured optical sections of 252 samples of 3D InSight™ microtissues in an Akura™ 384-well plate (InSphero) in 56 minutes and 23 seconds. NoviSight software enabled us to accurately analyze multiple sample images acquired at high-speed by the FV3000RS confocal microscope.
 

Methods

Sample preparation

3D InSight™ tumor microtissues in an Akura™ 384-well plate were provided by InSphero. The green-fluorescent-protein-tagged cell line HCT-116 (human colorectal carcinoma) was aggregated together with NIH3T3-RFP fibroblasts. On each reference compound, the 7-day treatment (with dosing at day 0 and redosing at day 4) was performed. Figure 1 shows the plate layout and the list of test compounds. After drug treatment, we washed the samples three times with a 1x phosphate-buffered saline (PBS) solution and fixed them with a 4%paraformaldehyde solution overnight at 4 °C (39.2 °F). Then we washed the samples with a 1x PBS solution, stained them with 1 μM TO-PRO-3 (Thermo Fisher Scientific) in a 0.1% TritonX-100 solution, and incubated them with the SCALEVIEW-S4 clearing reagent overnight at 37 °C (98.6 °F).

Fig. 1 Plate layout and list of test compounds

Fig. 1 Plate layout and list of test compounds

Imaging and analysis

To avoid RFP and TO-PRO-3 crosstalk in this experiment, the images were taken in sequence mode (which takes twice as long). 
In the analysis, the fluorescent signals from all cells (TO-PRO-3) enabled us to recognize the nuclei. All cells were classified as either an HCT-116 cell or a NIH3T3 cell based on the GFP and red fluorescent protein (RFP) signals.
 

Results

Optimizing the objective lens correction collar

First, the objective lens’ correction collar (CC) was optimized to obtain 3D images with high resolution and high contrast. A UCPLFLN20X objective with a long working distance and large numerical aperture (NA) was used for imaging. The position with the best contrast was adjusted by changing the dial little by little beginning at the lower end, and we found that CC = 0.17 was the optimal condition (Fig. 2). As a result, we adjusted the CC to 0.17 to observe an Akura™ 384-well plate with a UCPLFLN20X objective.

CC: lower end

CC: 0

CC: 0.17

CC: 1

Fig. 2 Optimizing the objective lens correction collar (Scale bar=100 μm)

Optimizing the Z-axis step size

Next, we examined the effect of the Z-axis step size on object recognition and analysis in NoviSight™ software. The results show that recognition accuracy decreased in proportion to the step size (Fig. 3). The minimum step size (1.28 μm), which is half of the depth resolution, was used as a reference. After classification, each cell number showed no significant difference between the minimum step size (1.28 μm) and 3 μm (Student’s t-test). The nucleus size is approximately 10 μm, so about three images can be taken for each nucleus using this step size.

Fig. 3 Optimizing the step size

Fig. 3 Optimizing the step size

Optimizing the average times (resonant scan mode)

The image quality of the FV3000RS microscope’s resonant scan depends on the number of averaging times. We optimized the number of averaging times to obtain the image quality required for NoviSight analysis. NoviSight software uses the fluorescence signal to recognize the object for analysis, so it is unnecessary to use a high-resolution image. In this experiment, nuclei with a high a signal-to-noise ratio are used for object recognition so that it is sufficient to only obtain intensity information in other channels (GFP/RFP). When we compared the recognition accuracy by the galvano-scan with the averaged image of the resonant scan, the recognition accuracy was almost the same as the control, even when the averaging was not performed (Student’s t-test) (Fig. 4). In this experiment, we proved that the image quality was sufficient for NoviSight analysis without the averaging.

Fig. 4 Optimizing the average times

Fig. 4 Optimizing the average times

Imaging under optimized conditions

Optimizing the above conditions (Fig. 5A) enabled us to capture 252 samples of 3D InSight™ microtissues in Akura™ 384-well plates (InSphero) in 56 min 23 sec (Fig. 5B). The results show that each drug inhibits spheroid growth in a dose-response manner. In addition, treatment with high concentrations of Sunitinib, a molecular-targeted drug that inhibits vascular endothelial growth factor (VEGF), significantly reduced the intensity of the green fluorescent protein (GFP), indicating that it affects only cancer cells (GFP-tagged HCT-116).

(A)

Optimized settings

Optimal conditions

Objective lens correction collar

CC: 0.17

Step size

3.0 µm

Average times (resonant scan) 

None (1)

(B)

Fig. 5 Optimal conditions and imaging results (Scale bar=100 μm)

Fig. 5 Optimal conditions and imaging results (Scale bar=100 μm)

Analysis

After cells are detected (Fig. 6A), it’s easy to confirm the total number of cells in each spheroid using a heat map on the NoviSight™ software (Fig. 6B). After classification, the percentage of relative cell numbers was calculated and plotted (Fig. 6C). The analysis results showed that Sunitinib affected the growth of cancer cells (HCT-116, IC50=1.46 μM). It was possible to sufficiently carry out 3D drug evaluation with high-speed imaging using the FV3000RS system under the optimized conditions.

(A)

Fig. 6A

(B)

Fig. 6B

(C)

Fig. 6C

Fig. 6 Three-dimensional efficacy drug evaluation

Conclusion

252 samples of 3D InSight™ microtissues in Akura™ 384-well plates (InSphero) were imaged using the FV3000RS confocal microscope under conditions optimized for high-throughput imaging. This study demonstrates that imaging can be completed within one hour and NoviSight 3D analysis can be performed. Adjusting the combination of dyes and the sample size may further improve the throughput.

Author

Hiroya Ishihara, Biological Evaluation Technology 2, Research and Development
Takashi Sugiyama, Biological Evaluation Technology 2, Research and Development
 

Products related to this application

3D Cell Analysis Software

NoviSight

NoviSight 3D cell analysis software provides statistical data for spheroids and 3D objects in microplate-based experiments. Use it to quantify cell activity in 3D, easily capture rare cell events, obtain accurate cell counts, and improve detection sensitivity. NoviSight software works with a range of imaging techniques, including point-scan confocal imaging, two-photon imaging, spinning disk confocal imaging, and super resolution live cell imaging.

  • Fast 3D image recognition from whole structures to subcellular features
  • Accurate statistical analysis
  • Equipped with a variety of ready-to-use default assays or easily design your own

was successfully added to your bookmarks

View BookmarksClose

Maximum Compare Limit of 5 Items

Please adjust your selection to be no more than 5 items to compare at once

Sorry, this page is not
available in your country.

Sorry, this page is not available in your country