RAS is one of the most frequently mutated genes in human cancers, present in about 20 percent of all cases. Its protein product acts as a master regulator of cell proliferation by initiating multiple downstream signalling cascades. Oncogenic mutations lock RAS protein in an active, GTP-bound state, relentlessly promoting cell division and tumorigenesis. Despite being a key cancer driver, directly targeting RAS has long eluded drug developers due to its high affinity for GTP/GDP and the smooth surfaces devoid of good binding pockets.

Instead, the research teams focused on a critical effector of RAS: the phosphoinositide 3-kinase enzyme PI3K, which propagates signals essential for cell growth and survival. However, indiscriminate inhibition of PI3K has posed significant clinical challenges because this enzyme also participates in vital functions like insulin signalling. Inhibitors that block PI3K broadly often incur metabolic toxicities such as hyperglycemia, limiting their therapeutic window.

To solve this conundrum, scientists employed a combination of sophisticated chemical biology methods and selective compound screening to identify molecules capable of covalently binding near the RAS-binding domain of PI3Kα isoform. These small molecules irreversibly attach to specific amino acid residues at the PI3K surface, effectively occluding the RAS binding site. Remarkably, this selectivity preserves PI3K’s ability to engage with other interaction partners, such as those in the insulin signalling axis, thereby reducing systemic side effects.

A bespoke biochemical assay developed at the Crick Institute enabled the verification that these covalent inhibitors disrupted the PI3K-RAS interaction with high specificity. Structural and functional characterizations confirmed that the compounds prevent the pathogenic activation loop driven by mutant RAS without compromising normal enzyme activity necessary for homeostasis. This targeted mechanism represents a major leap forward in precision oncology.

The in vivo efficacy of one leading compound was judiciously evaluated in mouse models engineered to develop RAS-mutated lung tumors. Treatment led to significant arrest of tumor progression without detectable increases in blood glucose levels. This outcome underscores the concept that uncoupling RAS-dependent oncogenic signalling from PI3K can suppress tumors effectively while sparing healthy physiology, a milestone in mitigating the therapy-limiting toxicities observed with previous PI3K inhibitors.

Further investigations demonstrated that combining the PI3K-RAS interaction blocker with other drugs targeting parallel nodes within the RAS pathway resulted in synergistic and durable tumor control. The combination therapies enhanced suppression of tumor growth beyond the capability of single agents, providing a compelling rationale for multi-modal treatment regimens leveraging pathway redundancies to overcome cancer resistance mechanisms.

The scope of the drug’s utility expanded unexpectedly when researchers explored its effects against HER2-driven tumors, commonly found in breast cancer and characterized by overexpression of the HER2 receptor tyrosine kinase. Since HER2 also signals via PI3K, but operates independently of RAS, the inhibitor nonetheless blocked PI3K-driven tumor growth in these models. This intriguing discovery implies the drugs could serve as versatile therapeutics across a wider spectrum of cancers harboring mutations in either RAS or HER2 oncogenes.

Following these promising preclinical results, the lead compound has entered Phase 1 clinical trials designed to assess safety, tolerability, and preliminary efficacy in patients with tumors driven by RAS or HER2 mutations. The trial will also investigate whether administering the drug in combination with other agents targeting RAS-associated pathways enhances therapeutic outcomes. The initiation of this clinical evaluation represents a significant translational achievement stemming from deep mechanistic insights into protein-protein interactions and covalent drug design.

Julian Downward, Principal Group Leader at the Francis Crick Institute, highlighted the perseverance required to address one of oncology’s most challenging targets: “Our journey to disrupt RAS-driven signalling without harmful side effects reflects decades of fundamental biology research and innovative chemistry. The ability to selectively prevent RAS from binding PI3K while preserving other cellular functions exemplifies how nuanced targeting can unlock new treatment avenues.”

Matt Patricelli, Chief Scientific Officer at Vividion Therapeutics, emphasized the transformative potential of this discovery for drug development: “These covalent inhibitors open a fresh paradigm for targeting oncogenic signalling complexes. By precisely blocking pathological protein interactions rather than entire enzymes, we have created molecules that can thwart tumor growth while maintaining normal cellular processes. Seeing this science advance into the clinic is truly rewarding.”

This breakthrough exemplifies the power of combining chemical biology, structural insights, and rigorous preclinical validation to overcome long-standing barriers in drug discovery. Should clinical trials validate safety and efficacy in humans, these compounds offer hope for improved therapies that can more effectively combat cancers driven by RAS and HER2 mutations without the burden of debilitating side effects. The approach also lays the groundwork for the design of next-generation molecular glues and inhibitors that selectively modulate oncogenic signalling pathways with unprecedented precision.

The Francis Crick Institute continues its mission to translate fundamental scientific insights into impactful medical advances that can save and improve lives. This collaboration with Vividion Therapeutics underscores the synergy between academic research and industry innovation, fostering rapid development of targeted cancer therapies. As this drug candidate progresses through clinical evaluation, it positions itself at the forefront of precision oncology focused on exploiting vulnerabilities in cancer cell signalling networks.

Subject of Research: Targeted disruption of the RAS-PI3K interaction to inhibit tumor growth in cancers driven by RAS and HER2 mutations.

Article Title: Covalent inhibitors of the PI3Kα RAS binding domain impair tumor growth driven by RAS and HER2

News Publication Date: 9 October 2025

Web References: http://dx.doi.org/10.1126/science.adv2684

References: Klebba, J. et al. (2025). Covalent inhibitors of the PI3Kα RAS binding domain impair tumor growth driven by RAS and HER2. Science. 10.1126/science.adv2684.

Keywords: Drug discovery, Tumor cells

The research team meticulously analyzed cerebrospinal fluid obtained from patients before they commenced CAR-T therapy, leading to the discovery of several proteins that serve as biomarkers for predicting the development of ICANS. This research opens new avenues for enhancing patient safety, as early identification of high-risk individuals may allow for timely intervention or preventative measures that could mitigate the risk of severe neurological events.

The significance of this discovery cannot be overstated, especially given the increasing popularity of CAR-T-cell therapy, which has shown remarkable success in treating refractory blood cancers. However, patients undergoing this therapy are at risk of encountering serious and potentially fatal side effects, including inflammatory responses within the central nervous system that can lead to impaired consciousness, seizures, or even brain hemorrhages.

Dr. Yuya Kunisaki, a key figure in this study, emphasized the urgency of finding reliable predictive measures for ICANS, particularly considering the high incidence rate of this syndrome following CAR-T therapy—estimated to be around 64%. Kakuni, along with his colleagues, has identified specific proteins that exhibit distinct levels in patients who develop ICANS in comparison to those who do not, thereby positioning these proteins as viable biomarkers for the prediction of ICANS.

The researchers began their study with a sample of cerebrospinal fluid from 29 patients diagnosed with B-cell non-Hodgkin’s lymphoma. Through rigorous proteomic analysis, they identified a total of 864 different proteins present within the cerebrospinal fluid samples. Ultimately, their investigation narrowed this number down to 46 proteins that displayed significant differences in expression levels between patients who developed ICANS and those who remained unaffected.

Among these proteins, two emerged as particularly noteworthy: C1RL, which was found to be elevated in patients who developed ICANS, and FUCA2, which demonstrated decreased levels in the same group. By analyzing the ratio of these two proteins, the researchers established a predictive equation that achieved remarkably high accuracy, yielding a score of 0.95 in ROC curve analysis—an impressive validation of its predictive potential.

To confirm the efficacy of this predictive model, the researchers conducted a follow-up analysis with a second group of 10 patients undergoing CAR-T therapy. The results corroborated their earlier findings as the biomarker ratio accurately identified the risk of all patients for developing ICANS, thereby highlighting its robustness as a predictive tool.

Despite these promising findings, the researchers caution that their results are still preliminary due to the limited size of their sample population. As co-first author Dr. Tomoko Nomiyama pointed out, further studies with a larger cohort of patients are necessary to fully validate these observations and to strengthen the evidence for the biomarker ratios they have identified.

If subsequent studies confirm these preliminary results, the implications for clinical practice could be transformative. The identification of patients at high risk for ICANS through a simple cerebrospinal fluid analysis permits a paradigm shift in patient management—a proactive approach toward treatment that could include early interventions or preventive therapies aimed at reducing the likelihood of ICANS manifestations altogether.

Additionally, the researchers are exploring the feasibility of identifying similar predictive biomarkers in less invasive samples, such as blood serum. Given that the collection of cerebrospinal fluid is typically invasive and not routinely performed prior to CAR-T therapy in most hospitals, the ability to glean this vital information from standard blood tests would vastly improve accessibility and practicality for predicting ICANS.

This research not only underscores the potential for personalized medicine in the realm of cancer treatment but also exemplifies the promising intersection between proteomics and clinical oncology. By marrying advanced proteomic analysis with clinical insights, the ability to tailor cancer treatment to individual patient profiles is becoming increasingly realistic, paving the way for safer and more effective therapeutic strategies.

As the team continues to expand their research, they are committed to confirming the reliability of their biomarkers across a broader spectrum of blood cancers. The hope is that these findings could eventually lead to the establishment of universal predictive tests that could become standard practice in oncology settings, allowing for more informed treatment decisions and improved patient outcomes.

Ultimately, the strides made by this research team may redefine the landscape of cancer immunotherapy. With ongoing advancements in biomarker identification and validation, clinicians will be better equipped to preemptively address the risks associated with therapies like CAR-T-cell treatment, facilitating a more patient-centric approach to cancer care.

Subject of Research: People
Article Title: Cerebrospinal Fluid Proteomics Exerts Predictive Potential for ICANS in CAR-T Therapy
News Publication Date: 11-Mar-2025
Web References: Kyushu University
References: Nomiyama T., Setoyama D., Yamanaka I., et al. “Cerebrospinal Fluid Proteomics Exerts Predictive Potential for ICANS in CAR-T Therapy.” Leukemia.
Image Credits: Daiki Setoyama, Kyushu University

Keywords: ICANS, CAR-T therapy, cerebrospinal fluid, biomarkers, predictive model, proteomics, cancer immunotherapy, central nervous system, personalized medicine, Kyushu University.