New research from Memorial Sloan Kettering Cancer Center (MSK) provides structural insights into how cancer cells thwart targeted RAS therapies; uncovers promising combination therapies for a rare childhood brain tumor; uses organoids to provide important clues about what drives pancreatic cancer; takes aim at appendix cancer with new laboratory models; and a develops a new precision tool for targeting cancer’s energy factory.
Cancer cells can develop clever workarounds to evade promising new “molecular glue” drugs that target RAS mutations. Understanding these resistance mechanisms points to potential combination treatment strategies, new MSK research finds.
RAS mutations drive approximately one-third of all human cancers, and new drugs called tri-complex inhibitors — including daraxonrasib — have shown promise in clinical trials. These drugs work by forming a three-way complex between RAS, the drug, and a helper protein called cyclophilin A (CYPA), which can block cancer signals.
The new MSK study investigated why some patients’ tumors stop responding to these therapies. The research was led by Ben Sang, PhD , Ling Feng Ye, MD, PhD , Zheng Fu, PhD , and Yasin Pourfarjam, PhD in the lab of senior author Piro Lito, MD, PhD , in MSK’s Human Oncology and Pathogenesis Program .
Using patient samples from 40 people treated with daraxonrasib, cell models, and sophisticated techniques including X-ray crystallography, the researchers identified three distinct types of resistance mechanisms, all of which disrupt the function of the drug.
One mechanism involves secondary mutations in RAS that weaken the drug’s ability to bind to the protein. A second mechanism involves mutations in the BRAF gene that enable RAF proteins (which bind to RAS) to form pairs that block the drug from working effectively. A third potential mechanism involves mutations in CYPA itself, though these were identified primarily in laboratory models.
“Understanding the mechanistic basis of resistance led us to identify strategies to potentially overcome it,” Dr. Lito says. “This could be accomplished by inhibitors that can directly target secondary RAS mutations or by combination therapies to target other resistance alterations.”
The findings have implications beyond daraxonrasib alone, as they apply to an expanding class of similar drugs being developed to target RAS-driven cancers.
Read more in Cell .
Most tumors contain multiple coexisting cell populations, a phenomenon known as intratumor heterogeneity. Because these populations differ in biology and drug sensitivity, treatments that eliminate one often spare others, enabling resistance and relapse. Overcoming this heterogeneity is one of the central challenges in cancer therapy.
A team led by MSK’s Jovana Pavisic, MD , Andrea Califano, PhD , of Columbia University, and former Columbia graduate student Ester Calvo Fernandez developed a computational framework to systematically identify drug combinations that target multiple tumor cell states at once. They applied this approach to diffuse midline glioma (DMG), a devastating pediatric brain tumor with few effective treatments and mutations that are difficult to drug and give rise to diverse tumor cell states.
By analyzing individual cells from DMG tumors, the scientists identified seven coexisting cell states that are the same across tumors. Each state was controlled by a set of “master regulator” proteins, which are key molecular drivers of cell identity and survival. The team profiled the effects of 372 cancer drugs in DMG cells and used algorithms to predict drugs that could switch off the master regulators of each state. In mouse models of DMG, eight out of nine drugs selectively reduced cancer cells as predicted.
Notably, several drugs targeting minority cell states showed limited effects on tumor growth as single agents, but combinations targeting complementary cell states were more effective than individual drugs. For example, avapritinib with ruxolitinib nearly tripled survival in mouse models; similar benefits were seen for avapritinib with larotrectinib. Together, these combinations target all DMG cell states. All three drugs are already used in patients, including children, making these realistic candidates for clinical use in the near term.
“This work establishes a tumor- and mutation-agnostic framework for designing combination therapies by co-targeting tumor cell states that are the same across patients, enabling more broadly effective treatments for heterogeneous cancers,” says Dr. Pavisic, a pediatric hematologist-oncologist at MSK Kids . “Using DMG as a proof of concept, the approach identified combinations we can now use to treat patients. More broadly, we have developed a systematic, rational strategy applicable to any heterogeneous tumor.”
Read more in Nature Genetics.
Because pancreatic cancer is usually diagnosed after it has spread, doctors and scientists don’t know much about the molecular changes that play a role in early stages of the disease. This makes it difficult to develop ways to detect and block the growth of this deadly cancer when it’s most likely to be treatable.
For several years, the lab of Sloan Kettering Institute developmental biologist Danwei Huangfu, PhD , has been developing organoids (tiny 3-D tissue cultures that mimic the structure, function, and complexity of human organs) as a tool to study pancreatic cancer. In a new study, led by Xian Zhang, PhD , a senior scientist in the Huangfu lab, the team made these organoids by guiding human pluripotent stem cells into pancreatic progenitor cells, a cell state that resembles those seen during inflammation and appears especially likely to develop into cancer.
The researchers then introduced combinations of key cancer-associated genetic changes, including activation of KRAS and loss of genes such as CDKN2A , TP53 , and SMAD4. The goals was to model how alterations accumulate as the disease progresses.
This approach — which looked at both genetic and epigenetic changes (those that control gene activity) — revealed an important difference between human cancer and traditional mouse models. In human cells, multiple genetic changes, such as KRAS activation together with loss of CDKN2A and TP53, are required to trigger tumor formation. In contrast, mouse models can develop similar lesions with fewer alterations, such as KRAS activation with TP53 loss alone.
The organoids already are providing clues about how to stop the formation of pancreatic cancer in people who have a high risk of developing the disease.
“This research revealed that a protein called TET1, which normally acts as a molecular gatekeeper that maintains healthy pancreatic cells, gets suppressed when these genes are mutated,” Dr. Zhang says. “Restoring TET1 activity may offer a new way to prevent pancreatic cancer.”
Read more in Developmental Cell .
Cancer of the appendix (appendiceal cancer) that spreads throughout the abdomen is particularly aggressive and difficult to treat, but new MSK-developed laboratory models are helping researchers to understand why — and to identify new potential therapies.
A study from the lab of physician-scientist Karuna Ganesh, MD, PhD , describes the development of a unique collection of three-dimensional, lab-grown mini-organs (organoids) for both primary appendiceal cancers and their matched peritoneal metastases. The research was led by Ahmed Mahmoud, PhD , a recent graduate of the Weill Cornell Pharmacology Program and a Howard Huges Medical Institute Gilliam Fellow.
The organoids allowed the team to identify actionable, cancer-causing mutations that couldn’t be determined from the original tumor tissue samples. The team was also able to compare differences between primary tumors and the metastatic tumors. For example, researchers observed increased resistance to chemotherapy after the cancer had spread.
Importantly, the organoid platform allowed the researchers to identify two promising therapeutic strategies: Inhibiting the RAS pathway and WNT pathway both effectively targeted the tumors in the lab models — suggesting new potential paths to help patients with this extremely rare and difficult-to-treat disease.
“Our organoid biobank is the first of its kind and provides an essential resource for studying appendiceal cancer,” Dr. Ganesh says. “It allows testing new treatment approaches for this rare cancer that hasn’t received as much scientific attention as other gastrointestinal cancers.”
MSK scientists have developed a highly selective molecule that blocks a key enzyme that cancer cells use to fuel their growth — while establishing a new framework for ensuring small molecules hit their intended targets in cells with precision.
A major research challenge is confirming that covalent inhibitors — drug-like molecules that permanently bind to proteins — hit only their intended targets. Standard methods can identify binding partners but often can’t distinguish between off-target proteins where the drug binds a large proportion of the total pool (risking toxicity) versus those where only a small fraction is bound (less concerning).
A team overseen by MSK chemical biologist Heeseon An, PhD , developed a solution that combines traditional methods with a new “scavenging proteomics” approach that analyzes what’s left behind, providing a more complete picture of how the molecules engage proteins throughout the cell. The effort was co-led by postdoctoral fellow Liang Sun, PhD , and former postdoc Sang Ah Yi, PhD (now an assistant professor at Sung Kyun Kwan University in South Korea).
The team applied this framework to develop a small molecule called CNP7, which targets HMGCS1, the first enzyme in the mevalonate pathway — a cellular assembly line that produces cholesterol and other metabolites essential for cell growth. Cancer cells boost this pathway to fuel their proliferation.
While existing statins disrupt this pathway by inhibiting HMGCR, a downstream enzyme of HMGCS1, they are reversible inhibitors, and their ability to engage HMGCR efficiently declines over time.
CNP7, on the other hand, works like a molecular key in a lock, binding a specific amino acid in HMGCS1’s active site and shutting down the enzyme with high selectivity and durability. Using cryo-electron microscopy and other methods, the team confirmed CNP7 engages the catalytic cysteine residue of HMGCS1, capturing atomic-level images of how it works.
At the functional level, CNP7 treatment shut down the mevalonate pathway and caused cancer cells to stop growing. Meanwhile, different cancer cell lines showed varying sensitivity to CNP7. These patterns were distinct from those of statins, suggesting that targeting HMGCS1 specifically — rather than the downstream enzymes such as HMGCR — may be more effective in certain cancer types.
“This work provides a precision tool for studying cancer metabolism, a new potential therapeutic strategy, and a rigorous framework for better understanding the off-target binding of small molecule inhibitors,” Dr. An says.
Read more in the Journal of the American Chemical Society .