Methods to replace animal models

Today there are many research methodologies that allow research experimentation without the need to perform animal experiments or to do testing of substances without use of animals or animal products.

These methodologies are often more precise, efficient and ethical. Furthermore, they are often more relevant for humans. Replacement can occur through better use of available knowledge or through new methodologies, for example:

  • knowledge of the chemical structure of a molecule – gives an understanding of its effects on an organism
  • drug impact directly on the immune cells – predicts immune response in a tissue
  • understanding of the order of events triggered by a chemical – allows to predict impact on the whole organism.

The methodologies are simple or advanced cell culture models (in vitro) and computational models (in silico).

Cell culture models (in vitro)

Cell culture models are models based on the cultivation of cells outside of an organism. They are an important tool for replacement of animal experimentation. Cells are often cultured in the presence of animal-derived products, however, they can now be cultured completely animal-free.

How to culture cells using animal-free components is described on the page Animal-free cell culture.

Based on the method of cell cultivation cultures can be subdivided into 2D cultures (e.g. monolayer cell cultures, suspension cultures), 3D cultures (eg. spheroids, organoids) or microphysiological cultures (e.g. organ-on-a-chip). All these cultures re-create cell physiological conditions with different level of detail and complexity. One should remember, though, that all of them have their own limitations. It is important to find a balance and use a model that is simple enough to answer the question of interest.

Cells grown in a culture can be of human or animal origin and by type they can be isolated from various adult tissues (somatic cells), embryonic tissues (embryonic cells) or reprogrammed from adult to embryonic tissue (iPS cells). Different cell types can further be grown together – defined as co-culture. We focus on the human cell cultures as alternatives to animal models.

Two-dimensional cell culture (2D model)

Two-dimensional culture is a culture system that allows to grow cells in monolayers on a substrate of interest, which can be glass, plastic or plastic coated by filaments from extracellular matrix e.g. fibronectin. Many cell types can be efficiently cultured and co-cultured in a 2D system. Cell suspension culture is a subtype of 2D culture for cells that do not adhere to the surface and prefer to be cultivated in a suspension.

2D cell culture is one of the world’s most widely used models and owes a big part of its popularity to the ease of culture and availability of hundreds of different immortalised cell lines.

Advantages

Main advantage of the 2D culture is its simplicity that allows for relatively straight-forward cell propagation, manipulation, treatments, observations, sorting and quantification. 2D culture is a base for many established cell-based assays that predict an effect of a drug or a chemical on a certain cell type (e.g skin). Due to the fact that cells are grown in monolayers, they can easily be studied in a microscope and subjected to automated microscopy e.g. for drug development (e.g. cytotoxicity, genetic toxicity studies) or studied with other optical techniques.

Disadvantages

The planar growth of 2D cultures does not recreate to the full extent native conditions in which cells grow in a tissue, which is three-dimensional. Another important aspect is a lack of flow – cells grow in the same growth media in a cell culture dish and it is usually exchanged every few days. Some cell types do not grow well in this scaled-down environment, for example liver cells.

How to get started

Established cell lines for cell culture can be bought from ATCC American Type Culture Collection (a non-profit global resource centre) or can be isolated from patient tissues (primary cultures, with proper ethical approval and consent from the patient). The cells can be immortalised for prolonged use (this will allow to culture cells under longer time).

Three-dimensional culture (3D model)

Three-dimensional culture is a culture system that allows layered growth of cells, often with the help of a scaffold, to better mimic tissue organization. Scaffolds are structural supports that often consist of proteins from extracellular matrix (e.g. fibronectin) or hydrogels. Components of the scaffolds can be completely animal-free (like alginate, recombinant proteins) or of animal-origin (with batch variation).

More information about animal-free 3D scaffolds is available on the page Animal-free cell culture.

Many cell types can be efficiently cultured and co-cultured in a 3D system. In 3D environment cells usually display a different morphology than in the classic 2D culture: they tend to aggregate more as a result of increased cell-to-cell contact, whereas in the 2D culture systems cells usually are more stretched and have less contact points.

The 3D system is a great tool to mimic a more complex, organ-like, environment. It takes advantage of cells spontaneous self-organisation and differentiation capacity that leads to assembly of spheroid structures and even mini organs, organoids (e.g. from embryonic stem cells, iPS cells, or somatic cells).

Another subtype of a 3D culture model is advanced organotypic models (or assembloids, reconstructed tissue models) – these are models of tissues assembled through differentiation or co-culture of two or more cell types or organoids that can further be assembled to mimic organotypic physiology, such as presence of specialized cell-to-surface contacts e.g. cell-to-air.

Organoids are mini-organ structures formed by self-assembled and self-differentiated embryonic or iPS cells with accurate microanatomy and functionality. Organoids can be established for many organs e.g. liver, lung, brain, intestine, stomach, kidney and heart.

Moreover, organoids can be established straight from the patient tissues – for example tumour organoids. Depending on tissue type, organoids can be used in many applications. They can for example be used for drug testing, developmental studies, and toxicity assessment on various human organs, as well as screening of human developmental and perinatal toxicity.

Spheroids are self-assembled structures that arise when culturing or co-culturing cells in 3D environment, with less complex structure than organoids. They can be used for toxicity, drug screening and to monitor cell responses in a 3D environment, where cells have different exposure to toxins and nutrients, dependent on their position in the three-dimensional tissue.

Advanced organotypic models can be established by culture or co-culture of different type of cells assembled to mimic tissue layers or tissue interactions of a human organ, e.g. skin or lung, or a process, e.g. infection. Examples of such models are advanced lung models, infection models and so on.

Advantages

A main advantage of a 3D culture is that cells experience an environment that allows growth and cell-cell interactions in all directions, which is not possible in an 2D cultures. The 3D structure facilitates self-organisation and differentiation of cells into small organ-like or tissue-like structures. These structures have the potential to replace animal testing e.g. when it comes to study development and drug toxicity.

Since they are derived from human cells, they may also produce more human relevant data than what can be acquired from animal models. The 3D-structures can also be effectively used for cells that do not function well in 2D environment e.g. liver cells.

Disadvantages

Likewise, the complexity of these cultures can also be a disadvantage. To cultivate a more complex structures like organoids one has to follow complex and long protocols, with addition of expensive growth factors, and the process of differentiation often lasts for weeks and can lead to a high variation. In addition, cells growing in clump-like spheroid structures are more difficult to study than 2D models using standard microscopy settings and may require clearing protocols for efficient imaging.

Another important aspect is that spheroids/organoids can only grow to 1-2 mm in size after which efficient exchange of nutrients and gases gets limited to the core of the cell structure. This is because these structures lack vascularization. Use of a bioreactor that effectively stirs the organoids may help in efficient exchange with the culture media.

How to get started

Choice of a 3D cell model is heavily dependent on the research question and which organ/tissue you want them to represent. This can include both disease models (e.g. infection disease model, 3D cancer model) and healthy tissue models (e.g. advanced lung model, reconstructed epidermis model, soft tissue model).

There are many different protocols developed by researchers and companies on how to establish 3D cell models. This may range from relatively easy 3D spheroid models to more complex multicellular organoid models. These protocols may need to be adapted to your particular need, and type of cell starting material you have (cancer cells, primary cells, iPSCs or embryonic stem cells). In general, low attachment culture plates facilitate 3D growth and organization, and often some type of hydrogel may be needed to simulate the extracellular matrix scaffold.

There are many non-animal hydrogels available, such as alginate, originating from algae and many new are under development. Mouse-derived Matrigel is a popular hydrogel, but has the disadvantage that mouse-specific factors in the gel may interfere with your cells.

More information about animal-free 3D hydrogels and scaffolds is available on the page Animal-free cell culture.

A drawback with 3D cultures is that they easily aggregate and lack vascularization. Therefore, it is needed to keep them in continuous movement which also more efficiently allows exchange of nutrients and gasses. To this end, a plate shaker, or better a spinning bioreactor, is needed to grow healthy and uniformed sized organoids.

Microphysiological systems

Microphysiological systems – also known as organ on a chip – are basically miniature microfluidic devices for cell culture, where cells can be grown on membranes in a 3D environment that even closer mimics human tissue/organ physiology.

These devices recreate the microenvironment with the biomechanical forces experienced by organs, which include the organ’s interaction with a regulated flow of media or gasses, or both. Access to the flow is provided through a network of very small channels, often lined with for example endothelial cells, where fluid or air circulates (microfluidic).

Moreover, certain forces can be applied to the membranes on which cells are growing (e.g. expansion-shrinking) stimulating mechanical movement that is a part of natural environment of many cells, especially lung cells. Many cell types can be successfully cultured and co-cultured on organ-on-a-chip devices, for example lung, brain, liver, kidney, heart cells giving rise to the specific names lung-on-a-chip, liver-on-a-chip, brain-on-a-chip etc.

Organ-on-a-chip are great tools for recreating different tissue microenvironments such as the blood-brain-barrier, skin or the lung to study their function and response to drugs in a more controlled and precise manner than using conventional cell culturing or animal models. Thus, these devices can successfully circumvent the need for animal testing, by creating an optimal and controlled environment for accurate testing of substances on human tissue samples.

A series of devices that represent different organs can be joined into a network, creating a mini-organism connected via a microfluidic flow. Alternatively, different cell types from different organs can be cultured on a single chip, emulating an organism (body-on-a-chip). This has a great potential to emulate the human organism and study impact of chemicals, pharmaceutics or disease progression on several organs/tissues at the same time.

Advantages

Main advantage is a presence of the flow of nutrients through different tissue or organ compartments that better recreate the microenvironment and physiology of organs. Complex interactions such as cell-air, blood-brain can be studied relatively easy. Furthermore, many different cell types can be grown in this environment, event those that do not thrive in the simpler two-dimensional culture. The device can be equipped with many micro-sensors that allows to monitor, measure and control several different parameters automatically.

Disadvantages

Challenge lies in the design of these devices, since they need to reproduce the complexity of the tissue mimicking composition through co-culture of cell types of interest and exposure to flow. Since you introduce several new parameters, this also introduces new level of complexity when using and troubleshooting this kind of system. When it comes to combination of devices to imitate organ system, the difficulty lies in knowing how to link them and in which order to mimic a system of interest. In addition, these devices are not available for high throughput screenings yet.

How to get started

Researcher from the Wyss Institute were first with developing organ-on-a-chip models originate. Today these models are developed in many labs and you can find those in Europe through the European Organ-on-Chip Society. There are also several organ-on-a-chip devices that are commercially available. The field of microphysiological models is actively growing and more and more tissue-specific models are getting available.

Are there any validated cell culture models?

There are several established cell culture models that allow replacing animal experiments with a much simpler, accurate and straight-forward technique, and with maybe greater human relevance. These models are validated to a high standard of reproducibility and accuracy by the European Commission's Joint Research Centre (JRC) and EU Reference Laboratory for alternatives to animal testing (EURL ECVAM) and are a part of OECD guidelines, which make them applicable for regulatory use.

2D and 3D models

EpiSkinTM (SM), LabCyte EPI-MODEL24 SIT and EpiDerm SIT reconstructed human epidermis models predict the skin irritation potential of chemicals and are regulatory approved as a full replacement for the Draize rabbit skin irritation test under the EC test method regulation (method B.46) and OECD test guideline No.439.

More examples can be found at the EURL ECVAM list of validated methods.

Microphysiological models

To promote the use of organ-on-chip devices for regulatory applications, EURL ECVAM created a catalogue of resources for developers and end-users to support validation and qualification of these new technologies. It is called EURL ECVAM Validation resources.

Computational models (in silico)

Computational models are accurate tools based on mathematical equations, machine learning and AI that allow to test substances, predict their behavior and effect on the human/animal organism.

In combination with existing data for e.g. toxicity or sensitization of substances, computational models can be used to predict the effect of new substances on the human/animal organism, without performing any new testing on humans or animals.

Computational methods can be subdivided into two categories: dynamical model (or also known as mechanistic models) and statistical models, also known as machine learning, deep learning, AI. There are also hybrid models that combine elements from both categories.

Many of computational models are successful examples of novel and more accurate methods that can be used in toxicology for among others risk assessment. They are in many cases more accurate than standard animal trials, not to mention more cost-efficient and ethically acceptable.

How to create a computational model

There are two ways you can create a computational model – you can start from a known or a hypothesized interaction (mechanism) and build your model around it with the help of ordinary differential equations. If you don’t have a mechanism, but a lot of data, you can analyze it with the help of mathematical statistical algorithms.

If you start building your model from mechanism, you will have a mechanistic model, also referred as dynamical model. If you start with a dataset, you will build a statistical model, known as machine learning, deep learning, AI.

In case you need to combine both dynamical and statistical elements within the same model you will have a hybrid model.

How to get started with computational models

Good news is that you don’t always need to start from scratch when you want to use computational models in research. There are useful models that are ready-to-use or that you can build upon. Here we explain established models that you can use.

If you are interested in detailed dynamical models for biological systems or processes, go to the European Bioinformatics Institute BioModels (EMBL) database. Models here are usually published also in scientific papers, and many of them have been manually curated by staff at EMBL.

If you need to predict a function of a new or a known molecule – you should use QSAR models, this may replace your need for testing in an animal organism.

If you need to predict the adversity of function of your molecule – you should use the adverse outcome pathway methodology (AOPs).

In case you are interested in what your body does to the molecule (pharmacokinetic) or what the molecule does to your body (pharmacodynamic) - you should use PKPD (pharmacokinetic-pharmacodynamic) models.

QSAR model (Quantitative Structure-Activity Relationship)

The QSAR model is a model built to predict the function of a specific molecule. This model can help you find connection between molecular features of your molecule and the functional outcome it can have on the organism.

For example, features like chemical structure, molecular weight or distribution of the electrons can predict related functions like mutagenicity or skin sensitization. You can look for molecular functions you want to avoid or those you want to seek.

AOP (Adverse Outcome Pathways)

AOPs are not computational models, instead a systematic way of gathering knowledge about a molecule/group of molecules and the specific adverse outcome. AOP can help you determine a link between your molecular initiating event and an adverse outcome. For example, how a substance of interest can generate a sequence of molecular and cellular that lead to a toxic effect.

Moreover, AOPs can serve as a base for quantitative AOPs, i.e. computer models that can be useful for some application. OECD has initiated a knowledge base around AOPs to provide a platform to share, develop, and discuss AOPs.

PKPD models (Pharmacokinetic-Pharmacodynamics)

PKPD models are dynamical models commonly used in drug development projects to design trials and to analyze outcoming data. PKPD models can help you describe the pharmacokinetic and pharmacodynamic properties of a drug (molecule). For example, how fast the drug is cleared out from the circulation, a binding site of the drug, drug-caused intracellular changes and changes in secretion patterns.

These properties of the drug will affect both the safety and the wanted outcome of the drug. FDA provides guidelines for the use of these models for regulatory purposes.

Are there any established or validated computational models?

We give you some examples of approved models that replace animal testing. More examples can be found at the OECD and JRC websites.

Dynamical model of glucose-insulin system (artificial pancreas) for preclinical testing of closed-loop control strategies (Kovatchev et al 2008). Accepted and approved by the US Food and Drug Administration FDA.

GARD™; statistical model of Genomic Allergen Rapid Detection assay to evaluate skin sensitization (Johansson et al 2016, Johansson et al 2019). It is included in the OECD Test Guideline Program, TPG no. 4.160.

Contact with users

If you want to come in contact with someone that works with any kind of replacement method, please contact us at the 3R Center and we will connect you with the right expert or experts in our network.

Revision date: 2024-10-02

To top