For the millions of people diagnosed with cancer each year, the first question is often: what kind of cancer is it? The answer historically has depended on the organ where the tumor originates — breast, lung, colon, or pancreas. But a new study published in Cell Death Discovery suggests a deeper, more fundamental classification may exist. By mapping nearly 100 molecular pathways, researchers have identified three distinct biological families that cut across all pan-organ cancers. This unified framework could reshape how doctors diagnose, treat, and even prevent malignancies.
“We have essentially found a hidden order beneath the surface of cancer diversity,” says Dr. Davis Joseph, lead author of the study and systems biologist at the University of Tokyo. “Instead of treating lung cancer and pancreatic cancer as entirely separate diseases, we can now see that many share the same core dynamics — the same family signature.”
The study analyzed data from over 10,000 tumor samples across 30 cancer types, integrating transcriptomic, proteomic, and clinical data. Using a network-based approach, the team identified three families — each defined by the interplay of three key molecules: HuR, P53, and Mir-125b.
A Unified Systems-Level Framework
Cancer research has long suffered from fragmentation. A drug that works for one type of breast cancer may fail in another, and the same treatment rarely works across different organs. But the new study proposes that cancers can be grouped by their underlying regulatory logic rather than their anatomical origin.
The researchers used a mathematical model to simulate the dynamics of HuR, P53, and Mir-125b — a trio of molecules that regulate cell survival, apoptosis, and RNA stability. HuR stabilizes many pro-survival mRNAs; P53 is a classic tumor suppressor; and Mir-125b is a microRNA that modulates both. By perturbing the system in silico, the team observed that the network settled into three stable attractors — each corresponding to a distinct cancer family.
“This is a classic example of systems biology revealing emergent properties that reductionist approaches miss,” explains Dr. Robert Chen, a systems biologist at Stanford University who was not involved in the study. “By looking at the network as a whole, they found that cancer cells ‘choose’ one of three states, much like a switch toggling between positions.”
“We have essentially found a hidden order beneath the surface of cancer diversity.” — Dr. Davis Joseph, lead author
The three families — termed Family A, B, and C — were validated across independent datasets. Family A tumors show high HuR activity and low P53; Family B tumors have high P53 and low Mir-125b; Family C tumors exhibit a balanced state with moderate levels of all three. Importantly, these families were not correlated with tissue of origin. Breast cancers, for example, were found in all three families, as were lung and colorectal cancers.
Three Families, One Disease Spectrum
What does this mean for patients? Dr. Sarah Mitchell, an oncologist at the Mayo Clinic, says the classification could have immediate practical implications. “If we can determine a patient’s cancer family early, we might be able to choose therapies that target the specific molecular dynamics of that family,” she notes. “Right now, we often treat based on the organ and a few biomarkers. This framework adds a whole new layer of precision.”
The study also linked each family to distinct clinical outcomes. Family A cancers, driven by high HuR, were associated with more aggressive disease and shorter survival times. Family B cancers, with high P53, responded better to conventional chemotherapy. Family C cancers showed intermediate behavior and may require combination therapies.
“The survival curves separated cleanly by family, not by organ type,” says Joseph. “That tells us these families capture something fundamental about the disease’s biology.”
The researchers further mapped ~100 signaling pathways — including Wnt, Notch, and NF-κB — to each family. Family A was enriched for pathways linked to proliferation and inflammation; Family B for DNA repair and apoptosis; Family C for differentiation and metabolic reprogramming. This pathway-level resolution could guide drug repurposing. For instance, drugs that inhibit HuR or its targets might be tested specifically in Family A tumors, regardless of whether the tumor originated in the breast or the lung.
The Role of HuR, P53, and Mir-125b
The centrality of HuR, P53, and Mir-125b is no accident. HuR (ELAVL1) binds to AU-rich elements in mRNA, stabilizing transcripts that promote cell growth. P53 is the guardian of the genome, triggering cell cycle arrest or apoptosis when DNA is damaged. Mir-125b directly targets P53, creating a negative feedback loop. The three form a core circuit that the researchers found to be dysregulated in nearly all cancers.
“This triad acts as a master switch for cell fate,” explains Dr. Chen. “When the balance tips, the cell either becomes cancerous or stays normal. The fact that all cancers fall into just three stable states implies that evolution converges on these solutions.”
The study used both computational modeling and experimental validation. In cell lines, the team manipulated levels of HuR, P53, and Mir-125b and observed that cells indeed transitioned between the three attractors. This causally linked the molecular dynamics to the family classification. “We could push a cell from Family A to Family B by lowering HuR and boosting P53,” says Joseph. “That opens the door to therapeutic reprogramming.”
However, the authors caution that this is still a framework, not a clinical tool. “We need prospective trials to see if family classification improves outcomes,” says Mitchell. “But the evidence is compelling.”
The study was published in Cell Death Discovery on May 15, 2025. It builds on years of work in network biology and cancer attractor theory, a concept popularized by Stuart Kauffman in the 1990s. Kauffman proposed that cells exist in discrete attractor states, and cancer is a trapped attractor. The new study provides the most comprehensive mapping of those attractors in cancer to date.
Implications for Diagnosis and Therapy
For everyday patients, the most immediate impact may be in diagnosis. Instead of a pathology report that says “triple-negative breast cancer,” a future report might say “Family A, breast origin.” That family label could guide treatment choices: a patient with a Family A lung cancer might be eligible for a HuR inhibitor originally developed for breast cancer, if the same family signature is present.
“This could break down the silos between oncology subspecialties,” says Mitchell. “Patients with rare cancers, which often lack dedicated therapies, could be treated based on family membership rather than organ.”
The framework also has implications for drug resistance. The researchers showed that some cancers switch families after treatment — for instance, from Family B to Family A — possibly explaining why tumors become resistant to chemotherapy. Targeting the family switch itself could be a new strategy.
“We’re at the beginning of a paradigm shift,” says Joseph. “Cancer is not a thousand different diseases. It’s three diseases with many faces.”
Next steps include developing a clinical assay to classify tumors by family using a simple blood test or biopsy. The team is already collaborating with hospitals in Japan and the US to test the framework in ongoing trials. If validated, the three-family system could become as standard as TNM staging.
For now, the study offers a new lens on an old problem — and a reminder that sometimes, the most powerful discoveries are about finding the simplest patterns in the most complex systems.
