New research from Memorial Sloan Kettering Cancer Center (MSK) reveals how pancreatic cancer cells organize themselves to fuel tumor growth; finds an antibody-drug conjugate shows efficacy in rare pediatric sarcoma; develops an AI tool could accelerate identification of disease-causing genetic variants; and develops quantum nanosensors for cancer blood tests.
How pancreatic cancer cells organize themselves to fuel tumor growth
Pancreatic tumors contain two distinct, interdependent subpopulations of cells that work together to drive tumor growth and spread, new research from an MSK-led team shows.
Scientists have long known that pancreatic ductal adenocarcinoma contains a diversity of cancer cell types, but how these cells interact and organize within tumors has remained unclear. A new study — led by Stefan Torborg and Jung Yun Kim, PhD in the lab of senior author Tuomas Tammela, MD, PhD at MSK’s Sloan Kettering Institute — reveals a critical cellular ecosystem underpinning these aggressive tumors.
Using genetically engineered mouse models, the researchers showed that pancreatic tumors contain two key cancer cell populations: WNT-secreting cells (WNT-S) that produce critical growth signals, and WNT-responding cells (WNT-R) that receive these signals.
Not only do the two cell populations reside in close proximity and depend on each other for survival, the two types of cells are dynamically related: The more transient WNT-R cells emerge from basal cancer cells in response to WNT made by the WNT-S cells, and eventually differentiate into WNT-S themselves creating a self-reinforcing cycle. When the researchers eliminated WNT-S cells or blocked their signaling with a drug, the WNT-R population collapsed, dramatically slowing tumor growth and metastasis.
The team identified Porcupine, an enzyme essential for WNT secretion, as a key player in this process. They also discovered that a subset of WNT-S cells express the Notch signaling molecule DLL1, which supports WNT-R cells. The study showed that both WNT and Notch signaling pathways work in concert — and that blocking either pathway suppressed tumor growth. Analysis of samples from patients with pancreatic cancer confirmed similar patterns exist in people, too, with high Porcupine expression linked to poor survival outcomes.
“This work reveals that different cancer cell states must cooperate for pancreatic cancer to grow and spread,” Dr. Tammela says. “Understanding these dependencies opens new therapeutic opportunities for disrupting the tumor ecosystem.”
Read more in Developmental Cell.
Antibody-drug conjugate shows efficacy in rare pediatric sarcoma
Desmoplastic small round cell tumor (DSRCT) is an aggressive cancer that primarily affects teenagers and young adults. Because there are currently no good treatments for this very rare sarcoma, doctors and scientists are looking for better options, including targeted therapies.
Research has revealed that DSRCT often has extra expression of HER2, a protein that is associated primarily with breast cancer but is also found in a subset of other solid tumors. A number of drugs that target HER2 are approved for treating these other cancers. One of these is trastuzumab deruxtecan (also known as T-DXd or Enhertu®). T-DXd is a type of drug called an antibody-drug conjugate: It consists of an antibody designed to seek out the HER2 protein linked to a payload of chemotherapy that it then delivers to the tumor.
Because of the urgent need for better DSRCT treatments, doctors from MSK Kids, led by Emily Slotkin, MD, decided to investigate giving T-DXd “off-label” — meaning the drug was used outside its formal FDA-approved indications, but in a manner consistent with how it has been used in other cancers and where its safety profile was already well established.
Doctors gave T-DXd to 19 MSK Kids patients with DSRCT. None of them had their tumors grow, and more than half of those with disease that could be measured (9 of 17 patients) had their tumors shrink measurably. There were no unexpected side effects. Importantly, the researchers also found that some tests of DSRCT may underestimate the levels of HER2, potentially missing treatment opportunities for some patients.
“The efficacy of T-DXd in DSRCT patients in this initial off-label experience is very striking and extremely promising,” Dr. Slotkin says. “It will be critical to follow up these results with a formal clinical trial, and our hope is that this approach might offer patients with this devastating disease a new and better treatment option.” Read more in JCO Oncology Advances.
New AI tool could accelerate identification of disease-causing genetic variants
A new AI tool developed by an international team of scientists could accelerate identification of disease-causing genetic variants by learning from evolution.
The computational model they developed, Orthrus, was trained on mature RNA sequences from 10 species (including humans, mice, chickens, zebrafish, fruit flies, nematodes, and others) and corresponding genes across more than 400 mammalian species. The team tested whether the model could predict RNA properties like stability, protein location, and ribosome binding.
While DNA stores genetic information, RNA molecules carry instructions from DNA to create proteins and regulate which genes are turned on or off. Different versions of RNA from the same gene (called isoforms) can have dramatically different — even opposing — effects on health and disease. Understanding how RNA sequence variations affect function is critical for developing therapeutics and identifying disease causes, but traditional laboratory experiments have been limited by their expense and the challenge of testing variants at scale.
Orthrus, on the other hand, was able to successfully predict multiple RNA properties and, remarkably, could identify functionally distinct isoforms from their sequence alone. For example, it correctly clustered the BCL2L1 gene’s isoforms that have opposite effects on cell death, and distinguished OAS1 isoforms with different antiviral activities. The model outperformed approaches trained only on DNA or single RNA sequences.
Because Orthrus learns to recognize functionally divergent isoforms entirely from sequence patterns, it could speed identification of disease-causing variants and predict how genetic changes affect cellular function — opening new possibilities for understanding genetic disease and developing new medicines.
The work was led by graduate students Philip Fradkin, of University of Toronto; Ruian (Ian) Shi, of University of Toronto and MSK; and Taykhoom Dalal of MSK. It was jointly overseen by senior authors Bo Wang, PhD, Leo J. Lee, PhD, and Brendan Frey, PhD of the University of Toronto; and Quaid Morris, PhD, of MSK’s Sloan Kettering Institute.
Read more in Nature Methods.
Building quantum nanosensors for cancer blood tests
MSK researchers discovered a method to greatly expand the capabilities of quantum nanosensors, tiny sensors that use quantum mechanical properties to detect minute molecular changes. The laboratory of MSK biomedical engineer Daniel Heller, PhD, developed a new type of chemistry that can modify carbon nanotubes — rod-like particles nearly 100,000 times smaller than the width of a human hair — with a wide range of different types of molecules that create “defects” on their surface. This allows the nanotubes to trap a tiny energy-transferring packet.
The resulting quantum nanosensors can be used to develop powerful sensor arrays that can detect subtle biological changes through measurable optical signals. This new approach greatly expands the library of sensors available to further improve the technology that the lab has been building into potential blood tests for early cancer detection, most recently for brain cancers.
“We now have the ability to build larger, more diverse sensor arrays that improve sensitivity and specificity for detecting cancer and other diseases at its earliest stages,” says lead author Stanislav Piletsky, PhD, a postdoctoral research in the Heller Lab at the Sloan Kettering Institute.
Read more in Nature Synthesis.