Has it ever occurred to you how we can study diseases without endangering human lives? This question is also at the heart of modern biomedical science. Disease modeling provides the answer to this deceptively simple question. It involves recreating illnesses under controlled conditions—whether in cells, animals, or computers, in order to understand what drives disease and how we might intervene in a better way.
In the past, disease research used to rely on observational studies in patients and relatively crude animal models, which offered limited understanding of the mechanisms behind the diseases. However, the advent of molecular biology, advances in gene-editing technologies, and the explosive growth in computational power have revolutionized the field. This evolution has led to a paradigm shift towards the creation of more sophisticated, human-relevant, and often predictive models that can bridge the gap between basic scientific discovery and clinical application.
From mechanisms to medicines
Disease models allow researchers to tweak a gene, remove a protein, or simulate a cellular interaction. This helps not only to investigate the root causes of illness, but also to identify new therapeutic targets—proteins or pathways—for drugs to target1,2. Disease models also enable the screening of hundreds or thousands of compounds, which makes them essential in the early stages of drug development for identifying promising candidates while eliminating ineffective or toxic ones3. The utility of disease models is not limited only to drug discovery. They also help to identify biomarkers and test novel approaches such as gene editing or stem cell therapies4-6. Computational models are highly valuable for simulating disease outbreaks and preparing more effective responses.
Types of disease models
In vitro models
In vitro models, meaning those developed outside a living organism, are often the first line of inquiry in disease research. Traditionally, 2D cultures are the primary type of in vitro model, that utilize flat sheets of cells grown in petri dishes7. The simplicity and scalability of such cultures make them useful in early drug testing, although they fall short when it comes to replicating the complex interactions of real tissues.
To address this, researchers now use 3D models such as organoids and spheroids. Organoids are miniaturized and simplified versions of organs grown from stem cells that mimic some of the functions of the human body8. Spheroids, on the other hand, are simple clusters of cells used for cancer research9.
3D bioprinting has taken it one step ahead with its ability to construct tissue-like structures using layer-by-layer deposition of cells and biomaterials, producing an architecture that more closely resembles functional organs10.
This progress brings ethical questions too. As models become more human-like—particularly brain organoids—new guidelines are needed to address issues of consciousness, consent, and moral status11.
The rapid advancement in this space has been possible particularly due to the use of induced pluripotent stem cells (iPSCs)3. They help to address the ethical concerns associated with the use of embryonic stem cells, as they are basically adult cells, reprogrammed into a stem-cell-like state and then induced to develop into any desired cell type.
The use of iPSCs has also enabled the development of highly personalized disease modeling, as for example, it allows researchers to collect cells from a person with Parkinson’s disease, turn them into neurons, and study what’s going wrong inside those neurons. Such “disease-in-a-dish” approach offers unparalleled insight into the genetic and cellular mechanisms of any disease12.
In vivo models
Despite all the advancements in the field of in vitro models, they may never replicate the complexity of a living organism. Mice and rats are the most widely used in vivo models due to their genetic similarity to humans and can thus be bred to naturally develop human-like diseases. Also, due to the ease with which their genomes can be manipulated, these animals can be genetically engineered to carry specific mutations.
Transgenic mice, for example, may be made to overexpress a disease-related protein, which helps scientists understand its role13. Knockout models, where a gene is entirely deleted, enable us to investigate what happens when certain proteins are missing14. Conditional inducible models even allow researchers to control when and where these genetic changes take place within the body15.
However, rodent models do not have identical physiology to ours and so results often fail to translate perfectly to human outcomes. To bridge this gap, researchers also use “humanized” mice that carry elements of the human immune system or express human genes16.
Nevertheless, ethical concerns loom large in animal research17. There has been an ongoing debate about how to balance scientific progress with animal welfare, although guidelines promoting replacement, reduction, and refinement (the 3Rs) are in place to address some of the concerns.
Other animals like zebrafish and frogs offer unique advantages18,19. For example: zebrafish embryos are transparent, and this makes them excellent for observing development and drug responses in real time.
Smaller organisms such as the nematode C. elegans and the fruit fly Drosophila melanogaster are also invaluable for genetic studies, despite their simplicity20,21. These creatures share many basic biological pathways with humans, and their short lifespans and genetic manipulability make them ideal for studying aging, neurodegeneration, and other complex traits.
In silico models
In silico models are the various computational models that have emerged as powerful tools to manage the explosive growth of biological data. They use various algorithms and mathematical frameworks to analyze the large biological datasets, in order to simulate disease processes. Artificial intelligence has made these models more efficient, allowing them to sift through complex datasets to predict disease risk or uncover new drug targets22.
There are various kinds of in silico models:
- Systems biology models track how changes in molecular pathways affect cell behavior23.
- Agent-based models simulate interactions between individual cells. This enables researchers to study tumor growth or immune responses in silico24,25.
- Pharmacokinetic and pharmacodynamic models help predict how a drug will behave in the body and what effects it will have over time26.
- Epidemiological models, like the widely known SIR (Susceptible-Infected-Recovered) framework, help predict the spread of infectious diseases and evaluate how successful will the impact of interventions like vaccines or lockdowns be27.
However, these models are only as good as the data they rely on. Poor-quality or incomplete datasets can lead to misleading predictions28. Computational models also suffer from challenges like model transparency and generalizability across populations29,30. Therefore, most simulations must eventually be validated through experimental or clinical research.
Building a reliable model
The process of building a model starts with a clear research question, which determines the selection of the appropriate platform: cellular, animal, or computational. The model is then developed, rigorously validated, and used to generate data. The model is then refined iteratively using constant feedback from the experimental results.
The following are the key criteria to determine the utility and reliability of a disease model:
- Face validity: The model should resemble the human disease in symptoms31.
- Construct validity: The underlying mechanisms that cause the disease in the model should be similar to those in the human condition32.
- Predictive validity: A good model should respond to treatments in ways that match human outcomes33.
- Reproducibility and robustness: The model should consistently produce similar results across different laboratories and under varying conditions23.
Ultimately, a reliable model should enable confident extrapolation of findings from the model to human disease for successful clinical application.
While acute diseases are easier to model and observe, chronic conditions like Alzheimer’s, diabetes, and autoimmune disorders are particularly difficult to replicate in the lab34. This is because their long-term nature requires extended study periods, as well as the involvement of multiple organs in such conditions and the environmental influences require intricate model designs, which are both costly and time-consuming.
Conclusion
Disease modeling is invaluable in modern medicine. These models let us study illnesses in controlled settings and bridge the gap between discovery and clinical application. Each model type—in vitro, in vivo, and in silico—offers unique advantages and presents specific challenges. However, future progress depends on integrating these diverse approaches. Only by combining methods, balancing simplicity with complexity, and leveraging both data and biological insights can we advance our understanding and improve our ability to tackle human disease.