Liquid Droplets Inside Cancer Cells Explain Why a Leukemia Drug Works So Slowly

One of the most celebrated triumphs of modern cancer medicine is a drug that should, in theory, shut down the engine of chronic myeloid leukemia with surgical precision. Yet clinicians have long observed something puzzling: even when tyrosine kinase inhibitors such as imatinib bind their target effectively, leukemic cells do not surrender immediately. Signaling persists, cell death is delayed, and a stubborn population of cells survives long enough to seed relapse. A new study published in Cell Research by Gen Li, Jun Wu, Zhijun He, Junhua Zhao and colleagues, with Peng Jiang of Tsinghua University as senior author, offers a startlingly physical explanation for this delay. The culprit, the researchers report, is not a genetic mutation or a bypass signaling pathway, but a biophysical phenomenon known as phase separation, in which the BCR-ABL1 oncoprotein congregates into liquid-like droplets that act as a barrier between the drug and its target.

BCR-ABL1 is the fusion protein born from the Philadelphia chromosome, the hallmark genetic abnormality of chronic myeloid leukemia and a subset of B-cell acute lymphoblastic leukemia. First described in landmark reviews by Goldman and Melo, the fusion fuses the BCR gene on chromosome 22 with the ABL1 tyrosine kinase gene on chromosome 9, producing a constitutively active kinase that drives uncontrolled proliferation and survival of white blood cells. Tyrosine kinase inhibitors were designed to slip into the ATP-binding pocket of ABL1 and freeze the enzyme in an inactive state. In structural studies of the kinase domain, including work by Cowan-Jacob and colleagues, imatinib and related compounds achieve exactly that. And yet, in patients, the kinetics of treatment response are markedly slower than direct enzyme inhibition would predict, an effect sometimes described as target sluggishness.

The new research reframes this sluggishness as an emergent property of how BCR-ABL1 organizes itself inside the cell. Rather than floating freely through the cytoplasm as isolated molecules, the team found that BCR-ABL1 molecules condense into dense, membraneless assemblies reminiscent of liquid droplets. These condensates form through multivalent, weak interactions among intrinsically disordered regions of the protein, the same class of physical chemistry that governs the formation of cellular structures such as nucleoli, stress granules and P bodies. The study connects this condensate behavior directly to therapeutic response: when BCR-ABL1 resides within these droplets, the local molecular environment becomes a physical barricade that slows the entry and action of tyrosine kinase inhibitors.

The concept of phase separation has transformed cell biology over the past decade. In a widely cited 2018 paper in Cell, Qamar and colleagues demonstrated that low-complexity protein domains can undergo liquid-liquid phase separation, creating compartments whose material properties dictate how molecules exchange with the surrounding cytoplasm. The new study applies this framework to cancer signaling for the first time in the context of targeted therapy. The authors showed that disrupting the conditions that promote condensate formation made BCR-ABL1 more accessible to drugs, while conditions that stabilized the droplets exaggerated the sluggish response. The droplet, in effect, functions as a microscopic shelter: drug molecules can reach the droplet surface, but penetrating the dense interior to reach every kinase molecule takes far longer than engaging freely diffusing protein.

Technically, the researchers combined protein biochemistry with cellular assays and clinical material. They purified BCR-ABL1 protein and observed its condensation behavior in solution, finding that the protein spontaneously demixes from the aqueous phase to form spherical droplets that fuse with one another and exchange internal contents, hallmarks of a liquid state. In cells, they visualized BCR-ABL1 condensates and correlated their abundance with the speed and completeness of kinase inhibition after tyrosine kinase inhibitor treatment. Crucially, the team collected bone marrow and blood samples from patients with BCR-ABL1-positive leukemia through collaborations with clinicians at Zhejiang Cancer Hospital and the First Hospital of China Medical University, allowing them to test whether condensate behavior in patient-derived cells tracked with treatment response.

The clinical implications of this reframing are substantial. Resistance to tyrosine kinase inhibitors has traditionally been attributed to kinase domain mutations, most famously the T315I substitution that abolishes imatinib binding, or to the persistence of leukemic stem cells that are intrinsically insensitive to the drugs. Studies such as those by Braun and colleagues and by Schneider and colleagues in Nature Cancer have catalogued the biology of these persistent cells, which survive initial therapy and fuel relapse. The phase separation model adds an entirely orthogonal mechanism: a cell can carry a completely drug-sensitive kinase and still mount a delayed response simply because its target protein is packaged inside droplets that physically exclude or retard drug penetration. This innate, non-genetic sluggishness could explain why a measurable fraction of cells in every treated patient survives the earliest hours and days of therapy without carrying any resistance mutation at all.

The finding also resonates with earlier structural and biochemical work on the ABL1 kinase. Structures of ABL1 bound to imatinib, dasatinib and nilotinib published by Tokarski and colleagues revealed exactly how these compounds lock the kinase in its inactive conformation, and kinetic studies showed rapid association rates in purified systems. The paradox between fast in vitro inhibition and slow cellular response now finds a candidate resolution: the purified enzyme in a test tube has no condensate, no barrier and no sluggishness, while the same enzyme inside a leukemic cell is wrapped in a liquid compartment that throttles drug access. Zhao and colleagues’ early structural characterization of the BCR-ABL1 complex, and Smith and colleagues’ dissection of its signaling architecture, provided the molecular map; phase separation now supplies the cellular geography that shapes how drugs navigate that map.

From a therapeutic standpoint, the study suggests that modulating condensate properties could become a strategy to sensitize leukemic cells to existing drugs. If the physical barrier created by BCR-ABL1 condensates is a principal cause of sluggish drug response, then agents that dissolve or destabilize the droplets, or that alter the material properties of the condensate so that small molecules diffuse through it freely, could accelerate and deepen the effect of tyrosine kinase inhibitors. Conversely, the work raises a caution for drug development: potency measured against purified kinase may systematically overestimate how quickly a compound will work in a cell whose target is phase-separated. Screening platforms that incorporate condensate-relevant conditions could help identify compounds that retain efficacy against droplet-sequestered targets, particularly for B-cell acute lymphoblastic leukemia, where early response kinetics strongly influence long-term outcome, as population studies by Qin and colleagues and reports by Ravandi and Molica have documented.

The broader significance extends beyond a single kinase or a single disease. Cancer biologists have increasingly recognized that many oncogenic proteins contain the disordered, multivalent regions that drive phase separation, and that signaling complexes such as those assembled by fusion oncoproteins, including the EML4-ALK and NUP98 fusions studied by Dixon and colleagues in engineered systems, may exploit condensation to amplify and sustain their signals. The BCR-ABL1 study demonstrates that this same organizational principle can also serve as a defensive architecture against therapy. Pendergast and colleagues’ classic 1991 work showed that BCR sequences activate ABL1 tyrosine kinase; three decades later, the new findings suggest that those same BCR-derived regions may coil the fusion protein into droplets that protect the activated kinase from the drugs designed to silence it.

For patients with chronic myeloid leukemia, tyrosine kinase inhibitors have converted a uniformly fatal disease into a manageable chronic condition, and a minority of patients now attempt treatment-free remission under close monitoring. Yet discontinuation fails in a substantial fraction, and persistent cells endure for years. By exposing the physical mechanism behind innate sluggishness, this research opens a new front in the effort to eliminate residual disease: rather than only designing better inhibitors, oncologists may one day prescribe drugs that strip away the droplet shield itself. The image of a cancer protein hiding inside a liquid droplet is a vivid one, and it captures a larger truth about modern biology. Cancer is not only a disease of genes and pathways but of physical organization, and conquering it may require manipulating not just what proteins do, but where and how they gather inside the cell.

Subject of Research: Phase-separated BCR-ABL1 condensates that delay the response of leukemic cells to tyrosine kinase inhibitor therapy

Article Title: Phase separation drives the innate sluggishness of BCR-ABL1 in response to targeted therapy

Article References: Li, G., Wu, J., He, Z., Zhao, J., Chen, H., Zhou, J., Zhang, Q., Wang, Z., Li, Q., & Jiang, P. (2026). Phase separation drives the innate sluggishness of BCR-ABL1 in response to targeted therapy. Cell Research. https://doi.org/10.1038/s41422-026-01286-w

Image Credits: AI Generated

DOI: 10.1038/s41422-026-01286-w

Keywords: BCR-ABL1, phase separation, chronic myeloid leukemia, tyrosine kinase inhibitors, condensates, imatinib, targeted therapy, leukemia, drug resistance, liquid-liquid phase separation, oncology, cell biology