Six Things to Consider Before Adopting 3D Cell Culture Models into Your Drug Discovery Pipeline
In an effort to improve the translational gap between in vitro and in vivo studies, researchers are shifting their focus towards advanced 3D cell culture models (e.g. spheroids, organoids, microtissues) that more accurately mimic the in vivo microenvironment compared to traditional 2D monolayer models. However, the field of 3D cell culture models is relatively nascent as is its application in large scale drug discovery efforts. Many companies are just beginning the process of evaluating and implementing assays that leverage the improved predictive capability of these models. The first thing that any researcher will notice about this space is that there is an incredible amount of diversity in the biology of these models, how they are cultured, how these models are characterized and ultimately the advantages and disadvantages of each model. It is also interesting to note that because of how new the field is there is not yet industry consensus about the best types of models for specific research questions and applications.
To begin adopting 3D cell cultures into your research workflow, the first thing that you need to understand is that these models are not a replacement for 2D cell culture models. While, 2D cell culture models are purported to not mimic the in vivo microenvironment as well as 3D cell culture models, these models still have their place in the drug discovery workflow as they are inexpensive and ultra-high-throughput. 3D cell culture models by their nature will always be significantly more expensive than less selective 2D cell culture models and thus it is suggested that 3D models be used to supplement and enhance 2D models and not replace them.
To best develop a plan for 3D cell culture adoption, the first step is to begin at the end and to determine the type of data that you need to acquire from your model and what your research question is.
1. Characterization Method:
One of the challenges with 3D cell culture models is that currently the techniques used for characterization were originally developed for 2D cell culture models and these tools often provide limited or even misleading insights about these models.
The simplest assays that a researcher can implement using 3D cell culture models are dissolution-based assays (e.g. total ATP) in which a fluorescent readout gives a singular data-point for each well in a plate. While these assays are ultra-high-throughput, inexpensive and do not require special equipment, they only provide one data point for each model and do not leverage the spatial information that makes 3D cell culture models intrinsically more valuable.
Size and Number of 3D Cell Culture Models:
Within some 3D cell culture systems, multiple models can be generated per well and in simple assays the total number of 3D cell culture models is used as a singular data point per well. Similarly, simple image analysis algorithms can be used to measure the diameter of 3D cell culture models and to use this as a singular end point as well.
Some questions will require that image-based end-points are acquired and these models can be extracted from a well plate and processed using conventional histology (e.g. H&E and IHC). However, one of the challenges with this approach is that removing these models from a well plate and performing histological processing is a tedious and cumbersome process that is ultra-low-throughput and expensive.
High content imagers can be used to acquire image-based end points from 3D cell culture models in a high-throughput setting. However, the major limitation with wide-field imaging in a 3D cell culture model context is that these imagers are designed for 2D constructs and thus have a tendency to poorly represent the complex nature of 3D cell culture models. The best analogy for describing this limitation is to consider the Earth as a 3D cell culture model that you want to study and simply taking a picture of the earth from the moon. While you would be able to characterize half of the surface of the Earth with this approach, you would know nothing about the majority of the earth, i.e. its interior. This is especially problematic with 3D cell culture models as the periphery is most exposed to a compound of interest and exhibits vastly different characteristics than the interior of the model.
Confocal microscopy allows for optical sectioning throughout tissues and for 3D data sets to be acquired which enables complete 3D cell culture characterization. Through confocal microscopy, every single cell within a 3D cell culture model can be imaged and characterized for multiple biomarkers through using multiple illumination channels. A major limitation with confocal microscopy in this context is that if you are acquiring multiple images from a single well, the imaging and processing is going to take significantly longer than widefield assays and thus your throughput is much lower with this improved data density. The other problem is that without the use of a tissue clearing technique like Visikol® HISTO-M™, optical attenuation limits confocal microscopy to only characterizing the periphery of these models instead of their entirety.
The reason that it is so important to intimately understand your desired endpoint is that each type of characterization assay requires different requirements for automation, well plates, labeling and data processing. There is a balance between acquiring lots of data and maintaining throughput where improved throughput will ultimately reduce data generation and thus it is important to understand the minimum amount of data per well required to answer your research question.
2. Cell Culturing Technique:
To generate 3D cell culture models there are a wide range of technologies available, but these techniques can be primarily characterized into scaffold-based techniques and non-scaffold techniques. Scaffold techniques such as Matrigel allow for cells to form 3D constructs within a matrix that mimics the in vivo extracellular matrix. For the non-scaffold approach, there are a wide-range of techniques such as the hanging drop approach, the n3D “bioprinting” approach and ultra-low-attachment (ULA) U bottom plates. Each one of these techniques has its own advantages and disadvantages in regards to cost, ease-of-use and performance and therefore there is not a specific cell culturing technique that is ideal for all applications. For example, while Corning ULA plates make cell culturing very easy and inexpensive, the round bottom of the plates can cause imaging artifacts with LED-based imaging systems for confocal imaging and are incompatible with some objectives such as the inverted watering dipping objectives on the Perkin Elmer Opera Phenix.
3. Model Type and Relevancy:
For some types of cells it is relatively easy to get them to aggregate and form 3D cell cultures and for others it can be more challenging. However, a major question that must be addressed is whether or not the biology you have chosen to generate in 3D in vitro assays is actually indicative of your in vivo target microenvironment. Therefore, much thought should be put into the types of cells that you use for your assay and the overall composition of your 3D cell culture model. Additionally, a validation plan needs to be developed to demonstrate that the model you have created replicates the functionality of an in vivo system. Some assays will benefit from highly complex 3D cell culture models with multiple cell types with well characterized functionality data while for other applications very simple 3D cell culture models might be able to give you the type on data that you want to collect. Ultimately, the model which you develop and its utility will be dependent upon your specific research question.
4. Throughput and Cost
Along the way of identifying an ideal model, cell culturing technique and characterization approach, an eye should always be kept on throughput and cost. The main rationale for this is that the throughput will determine the type of automation equipment that is required as well as the overall cost of the study. For example, if your research question requires a 500 um model with three cell types and the colocalization of four different markers in all cells with 5,000 compounds, then this assay will require dramatically different equipment than a similar assay with only an ATP total fluorescence readout. The reason for this is that the first assay would require the use of a confocal imager and would likely require one with multiple detectors for simultaneous marker imaging for accelerated throughput while the second assay would only require a simple plate reader and a robot to pipette and quickly load plates.
5. Data Processing
There is a common misconception in the drug discovery context that more data is always better as more data comes at the expense of reducing throughput or increasing study cost. It is important to always consider the minimum amount of data required to answer a research question when considering imaging-based end-points (e.g. sectioning, wide-field microscopy, confocal microscopy) as doubling imaging magnification from 10X to 20X can result in an 8 fold increase in data as well as a >8 fold increase in image acquisition time.
6. Internalize or Outsource
The last consideration is whether you should internalize 3D cell culture model generation and their analysis in-house or if you should outsource this process or components of it to third parties. There are many companies such as InSphero, Stemonix, TARA Biosystems and Mattek that have already developed validated 3D cell culture models that you can purchase and there are several companies such as Visikol that already offer full-service 3D cell culture assay services. Like all outsourcing, this choice comes down to resources and competencies where 3D cell culture assays require a deep set of diverse competencies and resources that you may not have access to.