Maximising efficiency in 3D cell culture model Imaging with Celigo and CQ1
von Michael Schell
The synergistic use of Celigo and CQ1 enhances research efficiency by combining rapid initial screenings with detailed high-resolution imaging. This approach is beneficial for core facilities and ambitious labs aiming to push their research boundaries.
Cellular imaging has come a long way over the last few decades, and various dedicated instrumental approaches make it easier than ever to visualize and analyse cellular events and behaviour.
In our recent blog we have addressed the general requirements for successful imaging and analysis of (mostly) 3D objects such as spheroids and organoids in a quite generic way. In a nutshell it was a speedy overview, high throughput and reasonable amounts of data vs. subcellular resolution and tons of data. In this blog we want to dig a little deeper into this matter and look at real hardware examples for these different focus areas and how they synergize.
In Cenibra´s portfolio you can find two instruments which look very similar at first sight, and many boxes can be ticked for both. As a matter of fact, however, the technological basis and the application focus of each of these instruments is quite unique, and the overlap in applications actually rather small. In a modern cell biology lab with an aspiration to efficiently and effectively elucidate cell biology, both have their place and provide information in a quite synergistic fashion.
Nexcelom´s Celigo and Yokogawa´s CQ1 offer complementary strengths which, when combined, enhance research productivity and workflow efficiency.
In a nutshell, the Celigo helps to analyse cell-based assays and detects trends in size, morphology and intensities in a very speedy and precise way, while Yokogawa´s CQ1 with its confocal scanning unit reveals the fine subcellular details of cellular processes like vesicular traffic or the inside of organoids and colonies.
Figure 1: Prototypical image comparison of Celigo and CQ1. On the left a whole well image is represented, which can be crucial in stem cell differentiation processes, where Celigo can be used to monitor the differentiation protocol and efficiency. On the right, a tiled 40x magnified image of human iPSC-derived neurons is represented. Critical are crisp and clear images of the fine aggregates that may develop in neural disorders diseases and can be quantified this way.
As mentioned they look quite similar at first sight. Both come as a fully integrated system in a reasonably small box, both are run by an external PC, offer bright field imaging and 4 standard wavelengths fluorescence excitation and serve various plate and sample vessel formats from tissue slides to 1536well plates.
The Celigo is built around a dedicated f-theta-lense and galvanometric mirror system, which provides a homogeneous illumination and imaging of an entire microtiter plate well (regardless of density). The magnification is fixed and sufficient to discriminate individual mammalian cells (or recognize individual ones, e.g. for monoclonality tests). This way the system can work pretty fast, typically a few minutes per plate. It also means that an incubation unit is not needed, as plates can go back to the incubator soon enough after imaging that damage on the cells is not to worry much about and an internal hotbox is not needed. Which, no surprise, is positively reflected in the price point of the system as well.
Obviously there are sacrifices to be made on the resolution compared to a confocal system, but for many assays it´s by far good enough and the speed gains are valuable.
On the other hand the CQ1´s core is Yokogawa´s microlense enhanced double Nipkow spinning disc. This ingenious piece of technology was once invented by Yokogawa with a different intention, but then discovered to be highly useful in high content cell imaging. It provides excellent image quality with minimum phototoxicity, and in combination with the CQ1´s large camera chip surprising speed for a one-camera-system. With the subcellular resolution of this approach with it´s z-stacking capabilities, imaging campaigns on 3D objects obviously deliver a different kind of information. (We have already illustrated the strengths of the CQ1 in one of the last blog articles, please check it out for more detail.
Quite obviously the data evaluation software packages of both instruments are tailored towards the respective “focus talents”, just to extract the best possible information or real data from what initially appears as just pretty images.
A question we hear a lot in the field when we present these technologies, is this: can the CQ1 not do everything the Celigo does? The fact that it is roughly twice the price might hint to that, but it is only partly true. While the Celigo is optimized for speed and whole well imaging at a well suited magnification for mammalian cells, the confocal CQ1 focuses on resolution for a more tightly defined space within a well, in both XY and Z dimension. And yes, you can probably do many Celigo experiments with the CQ1 and get it to look at the whole well, but it takes a lot longer, leaves you with tons of (not necessarily productive) data, and it simply blocks the precious device up from other, better fitting tasks for a long time in the process.
But lets go into some real world example of where Celigo and CQ1 “talents” can be best exploited and combined.
A great Celigo assay
First, let´s look at the Celigo. Very recently we have been working with the AKURA plate system provided from Insphero, Switzerland. The aim of the experiment was to quantify microtissue in a quality ccontrol workflow. Microtissue is a 3D cell aggregate that grows in the cavity of the AKURA plate and the task of the Celigo was to image the plate as quickly as possible and to automatically analyze the number of microtissues per well. In this setup the main criteria were acquisition speed and analysis precision. The Celigo nicely imaged the 96 well plate in under 2 minutes and the 384 well plate in under 4 minutes. Also, the data amount for such an experiment is reasonably low with around 100 MB. The result is a clear thumbnail image of the plate showing you the number of microtissues in each well as you may see in the below presented figure.
Figure 2: Thumbnail of an AKURA 96 well plate imaged with the Celigo. Individual microtissues are present in the well and can be analyzed automatically.
Specifically for these process related QC scans Celigo would be the option of choice. Not only would the CQ1 take roughly twice as long for the same task, plus extra analysis time, it also lacks the CFR compliance the Celigo offers for assays in regulated environments. In a more generic sense the fundamental question Celigo can quickly answer for this type of assays is this: how many objects are there and what is their general behaviour (e.g. in terms of viability or shape).
Similar assays can be thought of in a research environment, where, subsequent to such an initial screen, individual wells with “hits” (according to set criteria) can be looked at on the confocal screening microscope CQ1.
Typical CQ1 assays
Here, the CQ1´s spinning-disc confocal technology effectively blocks light from different z-planes from interfering with others and hence invites to look at many of these z-levels for optimum resolution of bigger objects. It´s software makes it very convenient to create big and clear z-stacks of your complete 3D cell aggregate, in addition objects can easily be segmented as diffused light is not disturbing. However, as the speed of the imaging process is still amazing it is also tempting to just collect as much as possible from the images. So, the fundamental questions answered by the CQ1 (and somewhat different from Celigo) are: How do cells interact and what are they doing at a certain point in time?
Figure 3: A whole universe in one image. Typical images of 3D objects on CQ1.
Left and middle image: Spheroid screening of co-cultured cells, grown in 96 well U-bottom plates. Images were taken with Yokogawa CQ1 microscope with 10x magnification. A z-stack in a range of 250 um with 60 planes was taken and the maximum intensity projection is shown here. Individual cells and cell fractions can be counted with the CellPathfinder software as shown on the left image. On the right a 40x magnified image in a different well plate is presented.
Comparing fig. 2 and fig. 3, the differences between Celigo and CQ1 images are obvious. In the former, rough effects can be quickly visualized and quantified on a large number of conditions. In the latter, the individual nucleus or individual cells are of importance and the aim is to potentially understand the mode-of-action of those high resolved cells that interact in the model. As the research is complex, the complexity of cell assays may also vary a lot from co-culture assays to assays where bacteria are growing on the organoid or to killing assays over time.
Figure 4: Spheroid screening of co-cultured cells, grown in 96 well U-bottom plates. Images were taken with Yokogawa CQ1 microscope with 10x magnification. A z-stack in a range of 250 um with 60 planes was taken and the maximum intensity projection is shown here. Individual cells and cell fractions can be counted with the CellPathfinder software as shown on the right and are localized in their X and Y position. Yellow dots represent tumor cells.
Great resolution, but also a lot of data
Such an assay obviously can create a huge amount of images and data. Just a quick calculation as an example. The spheroids were embedded in Matrigel in a 96 U-bottom shaped well plate. Because of the plate choice (remember, this is important to know in advance and U-bottom plates tend to have a thick bottom), we can use the 10x or the 20x objective. Before we did the imaging run, we quickly created a map from the whole well so we exactly knew the localization of the spheroid and by using our automated imaging approaches ACE (automated imaging explorer) you may even automize the “finding” process completely (this topic will be addressed in a coming blog article on image analysis). Importantly, this run covered a z-distance of 250 um that was divided into 60 slices with a spacing between of 4.2 um in 4 colors, red for tumor cells, blue, green and magenta for other cell types and a time lapse assay was created to stably image this setup under controlled gas- and temperature conditions for 24 hours. Considering the data (one time point can be already 20 GB) and speed we focused in this assay on an exclusive number of wells (for the sake of speed). But just to give you an idea, you may just calculate, (60x4) x the number of wells you are looking at and a full plate under these conditions mean we are collecting around 23000 images.
As we collected a lot of z-information we can also analyze the number of cells and their volume. However, addressing the volume of cells needs more calculation and preparation in advance as there are many physical parameters to be considered such as the wavelength, the camera and the numerical aperture of the objective.
Speeding up the process and minimizing harddrive consumption with the smart use of both instruments
Looking at these enourmous amounts of data if makes sense to exploit the synergies between Celigo and CQ1. The use of Celigo as a workhorse for first screening rounds of multiple plates and the subsequent “deep dive” into promising wells and conditions with CQ1 makes a lot of sense, minimizes the amount of raw data needed to be stored and processed, and the staged approach frees up the higher resolution instrument for tasaks and experiments that truly require such capacity.
Let´s get back to our example from fig. 2. You have already scanned your Akura plates with the Celigo, identified wells of interest, and can now progress and image just and only those with confocal resolution on the CQ1. The results which are prototypically shown in fig. 5.
Figure 5: Thumbnail of an AKURA 96 well plate imaged with the CQ1 in confocal resolution. Individual microtissues are present in the wells and can be analyzed for e.g. nuclei numbers automatically by CellPathfinder. Imaging was done with a 10x air objective.
And finally a quick warning: even though instruments and software try to make it as easy as possible, 3D analysis is not trivial and requires many thoughts in advance. If you want to learn more about 3D analysis, please check out future episodes of this blog. Or just contact us.
In coming blog articles, we will provide insights into extracting quantitative data from microscopy images efficiently and explore advanced sample analysis techniques, including automation and deep learning in image analysis.
References:
- https://www.yokogawa.com/eu/blog/life-innovation/organoids-paving-the-way/
- Vinci M, Gowan S, Eccles SA, et al., Advances in establishment and analysis of three-dimensional tumor spheroid-based functional assays for target validation and drug evaluation. BMC Biol. 2012 Mar 22; 10:29.