Pancreatic cancer, a formidable foe with one of the lowest survival rates among major cancers, is now facing an innovative arsenal of research tools. At the forefront of this battle are pancreatic cancer models, which are revolutionizing our understanding and treatment approaches.
A Breakthrough in Cancer Modeling
Christian Pilarsky's journey is a testament to the power of perseverance in science. After three years of refining his technique, he successfully created 3D pancreatic tumor replicas derived from patients' cells. This achievement brought a unique joy, akin to the birth of a child, as he and his team could now grow organoids from a significant portion of the patients they worked with.
The Power of Models
These robust models are not just scientific curiosities; they are pivotal in the quest to improve pancreatic cancer treatments and detect the disease early. By understanding the aggressive nature of this cancer and its resistance to therapies, researchers aim to develop more effective strategies. As Ben Stanger, a gastroenterologist, puts it, these models offer a way to study tumor biology without subjecting patients to clinical research.
A Growing Arsenal of Tools
Organoids are just one piece of the puzzle. Researchers are exploring various experimental systems, including animal models engineered to develop pancreatic cancer and those injected with human tumor cells. Additionally, computer models and artificial intelligence are being employed to analyze patient records, aiming to predict disease development.
Global Support and Initiatives
Both European and US governments are backing these efforts. The PRECODE initiative, spanning nine European countries, has established 13 centers of excellence for organoid research, leading to the creation of numerous new pancreatic cancer organoids. In the US, the National Institutes of Health (NIH) has announced plans for a dedicated center focused on organoid-based modeling.
A Front-Row View of Progress
David Tuveson, a cancer biologist, has witnessed the field's progress since 2015 when his team developed the first pancreatic cancer organoids. This breakthrough addressed the limitations of studying pancreatic cancer cells in a dish, which couldn't replicate the complex, 3D nature of tumors in the body. Organoids, with their 3D structure, preserved the architecture, heterogeneity, and drug responses of individual tumors.
Organoids as a Testing Ground
In 2018, Tuveson's team created a diverse library of organoids, representing much of the disease's complexity. By testing various drugs on these models and comparing results with patient data, they found that organoids responded to chemotherapy similarly to the patients they were derived from. This solidified organoids' role as a testing ground for treatments, as Naomi Walsh, a cancer researcher, notes: "It was that research that changed the field in terms of how we can use these models."
Unraveling the Deadliness of Pancreatic Cancer
Organoids are also helping researchers understand why pancreatic cancer is so deadly. In a study published in March, Vincenzo Corbo and his team used patient-derived organoids to demonstrate that circular fragments of DNA outside chromosomes, known as extrachromosomal DNA (ecDNA), can carry a specific tumor-promoting gene. Variations in the expression and number of copies of this gene can cause significant differences in cancer cell shape and response to their environment. This research suggests potential therapeutic strategies, either by targeting the weakness caused by high levels of ecDNA or directly targeting the cancer-driving gene in ecDNA.
Mimicking the Tumor Microenvironment
While organoids have provided valuable insights, scientists are now adding elements to the network of proteins and carbohydrates in which organoids are cultured to mimic the tumor's microenvironment. Specialized devices called organs-on-a-chip are also being used to grow cells in an environment that mimics the functional and physiological conditions of human tissues, such as the pancreatic ductal network.
The Accessibility of Organoids
Creating lab-grown pancreatic cancer organoids is time-consuming, but researchers now have the option to purchase them ready-made. Karla Queiroz, a cancer biologist, notes that researchers can choose from an increasing number of customizable commercial options. The ATCC, in partnership with the Human Cancer Models Initiative, has developed over 50 pancreatic cancer organoids, with more in development, each accompanied by clinical and demographic information and the full genomic characterization of the donor.
The Advantages of Mouse Models
While organoids offer unique advantages, mouse models still have their benefits. For instance, patient-derived organoids are made from already-formed tumors, meaning the steps leading up to tumor formation have already occurred. As Stanger explains, "The advantage in mice is you can really study progression, which gives you the opportunity to intervene before the tumor develops."
KPC Mice and Experimental Drugs
One extensively used animal model is the KPC mouse, genetically engineered with mutations implicated in pancreatic cancer. In 2024, researchers, including Stanger, used KPC mice to test an experimental drug targeting proteins known to drive pancreatic cancer growth. The positive results from this work provided preclinical validation for the drug, which is now being tested in humans.
Modeling Disease Progression in Mice
Pilarsky and his colleagues have used KPC mice to investigate the effect of blocking the gene encoding CDK7, which controls cell growth and division. When they blocked this gene with a drug and administered chemotherapy to the mice, they found that pancreatic cancer cells were damaged, stopped dividing, and were more likely to die.
Another approach is to implant mice with human pancreatic cancer cells, allowing for a closer mimicry of human disease. In 2020, Tuveson's team developed a specialized version of this model by delivering pancreatic cancer organoids instead of fully formed human tumors. This allows the cells to grow and evolve over time, better mimicking tumor progression in humans. The team also delivered the cells directly into pancreatic ducts, enabling the tumors to grow in a natural environment.
Challenges and Workarounds
There are challenges with various mouse models. Administering human pancreatic cancer cells to mice requires specialized training, and studying the immune system's influence on tumors is not currently possible due to the need to suppress the mouse's immune system. Researchers are working on solutions, including engineering mice with human-like immune systems.
For KPC mice, the constant breeding required to maintain a colony with two cancer-causing mutations can be challenging. One workaround is to engineer mouse embryonic stem cells to carry the desired mutations, reducing the number of breeding steps.
The Role of AI in Personalized Care
As animal and organoid models advance, AI is quickly becoming an integral part of pancreatic cancer research. AI models offer the advantage of real-time predictions, capturing the disease's complexity. As Søren Brunak, a disease systems biologist, puts it, "AI algorithms can model highly different routes to the disease, and I, therefore, see them as complementary to more generic wet-lab models."
AI researchers use retrospective data from thousands of individuals to train and test models, which can then predict who is likely to develop cancer, reducing the need for more invasive follow-up tests. Azadeh Tabari, a radiologist studying AI, emphasizes that AI does not replace medical expertise but can confirm suspicions and ensure timely patient care.
AI excels at analyzing large sets of information, such as genetic and protein data or medical images, to uncover patterns that are hard for humans to spot. It can also find subtle connections between symptoms in electronic medical records. This is particularly useful for pancreatic cancer, where telltale signs often appear too late.
In 2023, Brunak and his colleagues used AI to analyze medical records from Denmark and the US, predicting individuals at high risk of developing pancreatic cancer up to three years before their diagnosis. They are now mining clinical records for leads on symptoms linked to disease development.
Other groups are using AI to diagnose people earlier. Tabari's team trained a deep-learning model on clinical data from individuals with precancerous pancreatic lesions, achieving around 80% accuracy in predicting cancer development. The next step is to test the model in real-time and track individuals to understand its effectiveness.
One of the challenges with AI is the availability of high-quality data. In 2020, a pilot study to build a digital twin for pancreatic cancer was hindered by data sparsity. As Matthew McCoy, who led the project, notes, this led to "a lot of uncertainty in the model predictions."
With better data availability and regulatory approval, AI tools could provide a holistic view of a person's health, combining various data sources to aid diagnosis and treatment prediction. This approach, as Lucchesi suggests, is the future of personalized medicine.
The Future of Pancreatic Cancer Research
Tuveson believes that improvements in modeling and therapies will continue to go hand in hand. As treatments allow patients to live longer, models will need to evolve to capture emerging resistance and relapse mechanisms. Pilarsky is optimistic, believing that with better model systems, precision medicine will become a reality, offering successful treatment for patients.
This innovative approach to pancreatic cancer research, combining various modeling techniques and AI, offers hope for improved outcomes and a brighter future for patients.
