Two Cancer Drugs Self-Assemble Into Glowing Nanoparticles That Block Metastasis

Metastasis, the process by which cancer cells spread from a primary tumor to distant organs, remains the leading cause of death among cancer patients, and decades of research have yet to produce reliably effective antimetastatic therapies. A new study published in Materials Today Bio by Rujing Wang, Lan Zou, Dandan Mi, Chuan Wang, Zijian Song, Tiantian Liu, Mengnan Zhao, Dong Wang and Sanjun Shi offers an unusually creative approach: rather than designing a new drug, the team simply mixed two existing ones and let chemistry do the rest. The result is a self-assembled nanoparticle that glows on its own, tracks itself through living cells and tissues, and suppresses cancer spread through a mechanism that no single drug could achieve alone.

The two ingredients are sorafenib, a multikinase inhibitor used clinically against liver and other cancers, and curcumin, the polyphenolic compound extracted from turmeric rhizome that has long attracted interest for its anticancer properties. Both drugs share a frustrating limitation: patients frequently develop resistance to sorafenib after prolonged treatment, while curcumin suffers from poor solubility and weak fluorescence. The researchers dissolved the two molecules in dimethyl sulfoxide and injected the solution dropwise into water. As the proportion of water rose, the molecules spontaneously clustered into spherical nanoparticles roughly 286 nanometers in diameter, with a narrow size distribution and a strongly negative surface charge consistent with good biocompatibility.

What makes these particles remarkable is that neither drug fluoresces meaningfully on its own. Free curcumin actually loses its weak emission when it aggregates, a phenomenon known as aggregation-caused quenching, and sorafenib is similarly dark. Yet when the two co-assemble, the resulting nanoparticles emit bright light, a behavior the authors describe as a 0 + 0 > 2 fluorescence effect and term pharmaceutical dots-induced emission. Spectroscopic analysis revealed why: the pyridine ring of sorafenib stacks against the aromatic rings of curcumin through pi-pi interactions, while hydrogen bonds form between curcumin’s ketonic groups and sorafenib’s nitrogen atoms. Molecular dynamics simulations running over 50 nanoseconds showed the drugs first clustering separately and then merging into a stable core-shell architecture, with fluorine and chlorine atoms from sorafenib concentrated in the core and curcumin woven throughout.

The practical advantages of this carrier-free design are substantial. Traditional nanocarriers often achieve drug loadings below 10 percent because the carrier material dilutes the payload, but the self-assembled particles reached encapsulation efficiencies above 98 percent for both drugs, with sorafenib loading at 78 percent and curcumin at 14 percent. The particles also exhibited a large Stokes shift of 130 nanometers, which minimizes self-absorption and background autofluorescence, and their emission position stayed fixed across excitation wavelengths, a property that simplifies high-precision bioimaging. Stability tests in serum showed no change in particle size or fluorescence over extended periods, and in tumor-bearing mice the nanoparticles produced strong detectable signals in the lungs where a simple drug mixture produced none.

But the imaging capability turned out to be only the opening act. When the researchers treated breast cancer 4T1 cells and human liver cancer MHCC-97H cells with the nanoparticles, transmission electron microscopy revealed something unexpected: lysosomes and mitochondria, the cell’s digestive and energy-producing organelles, migrated toward the nucleus and clustered tightly around it. In untreated cells, mitochondria sat an average of 242 nanometers from one another and more than 225 nanometers from the nucleus. After treatment, those distances collapsed to roughly 17, 22 and 45 nanometers respectively, close enough to constitute direct interorganelle contact sites. Structured illumination microscopy and confocal imaging of the organelle marker proteins LAMP1, TOMM20 and Lamin B1 confirmed the same tripartite lysosome-mitochondrion-nucleus contacts in both cell lines and later in metastatic lung tissue.

To understand what these contacts accomplish, the team turned to proteomics. Quantitative mass spectrometry identified 2,244 differentially expressed proteins after nanoparticle treatment, with gene ontology analysis showing that the altered proteins concentrated in the nucleus, cytoplasm and mitochondrion. Upregulated proteins enriched in mitochondrial functions such as oxidative phosphorylation and NADH dehydrogenase activity, while downregulated proteins clustered around nuclear processes including DNA replication, cell cycle transitions and epigenetic regulation. Cross-referencing the differentially expressed proteins against survival genes from breast cancer patient data pointed to a single nuclear suspect: histone deacetylase 2, or HDAC2, an epigenetic enzyme whose overexpression is associated with poor outcomes and metastatic progression in breast cancer patients.

The mechanistic picture that emerged links organelle geometry to gene regulation. Mitochondria are the primary generators of reactive oxygen species, and lysosomes can amplify mitochondrial ROS production. By herding both organelles against the nuclear membrane, the nanoparticles appear to deliver a concentrated burst of ROS directly into the nucleus. Flow cytometry of isolated nuclear fractions confirmed elevated nuclear ROS after treatment, and the signal scaled with both dose and time. That nuclear ROS, the researchers found, suppresses HDAC2. When they scavenged ROS with N-acetylcysteine, HDAC2 expression rebounded; when the nanoparticles were present, ROS climbed and HDAC2 fell again. Conversely, blocking HDAC2 directly with the approved inhibitor vorinostat reproduced the antimigratory effect, confirming HDAC2 as a functional node rather than a bystander.

The downstream consequence is the collapse of epithelial-mesenchymal transition, the cellular program that allows cancer cells to detach, invade and colonize new tissue. Treated cells showed restored E-cadherin, the adhesion molecule lost during metastatic conversion, and reduced Vimentin, the motility marker that rises as cells turn invasive. Wound healing assays told the story in numbers: untreated cells closed a scratch almost completely within 24 hours, free sorafenib slowed closure to about 52 percent, the drug mixture to 34 percent, but the nanoparticles cut it to roughly 11 percent. Transwell invasion assays, western blots and immunofluorescence all corroborated the shift, and the nanoparticles also preserved sorafenib’s original function, suppressing phosphorylation of VEGFR and ERK while additionally sensitizing cells to mitochondrial ferroptosis through effects on GPX4 and DHODH.

In vivo, the strategy proved itself in mouse models of breast cancer lung metastasis. When 4T1 cells pretreated with the nanoparticles were injected into mice, lung weights and histopathology showed dramatically fewer metastatic lesions than with either drug alone. In a therapeutic model using luciferase-labeled 4T1 cells, intravenous dosing every two days for five doses produced significant reductions in tumor burden, metastatic nodule counts and bioluminescent signal compared with paclitaxel, curcumin or sorafenib, and paclitaxel itself has been reported to worsen metastasis in some models. Immunohistochemistry of metastatic lungs confirmed the same molecular signature seen in culture: high E-cadherin, low Vimentin, suppressed HDAC2, and dense perinuclear clusters of lysosomes and mitochondria. Body weights remained stable throughout, and histological and serum chemistry analyses of major organs showed no pathological damage, with liver and kidney markers staying within normal ranges.

The study’s broader significance lies in its demonstration that organelle positioning, not just molecular targeting, can be a therapeutic lever. Cancer cells depend on coordinated communication among lysosomes, mitochondria and the nucleus, yet these tripartite contacts had remained largely unexplored as drug targets. By pairing a kinase inhibitor with a dietary polyphenol in a carrier-free assembly, the researchers achieved simultaneous multi-target action, real-time imaging and a previously uncharacterized nuclear ROS-HDAC2-EMT axis, all without the cytotoxic residues that chemical fluorescent labeling would introduce. The work remains at the preclinical stage, and the specific drug pair was chosen partly for its assembly chemistry, but the concept of converting drug molecules themselves into self-reporting, organelle-orchestrating therapeutics opens a genuinely new direction for antimetastatic drug design, one where the medicine and the microscope share the same molecule.

Subject of Research: A carrier-free supramolecular nanoassembly of sorafenib and curcumin that inhibits tumor metastasis through lysosome-mitochondrion-nucleus contacts and a nuclear ROS-HDAC2-EMT signaling axis

Article Title: A label-free supramolecular nanoassembly triggers multi-organelle crosstalk-dependent metastasis inhibition via nuclear ROS-HDAC2 axis

Article References: Wang, R., Zou, L., Mi, D., Wang, C., Song, Z., Liu, T., Zhao, M., Wang, D., & Shi, S. (2026). A label-free supramolecular nanoassembly triggers multi-organelle crosstalk-dependent metastasis inhibition via nuclear ROS-HDAC2 axis. Materials Today Bio, 41, Article 103681. https://doi.org/10.1016/j.mtbio.2026.103681

Image Credits: AI Generated

DOI: 10.1016/j.mtbio.2026.103681

Keywords: sorafenib, curcumin, self-assembly, nanoparticles, metastasis, lysosome-mitochondria contacts, nuclear ROS, HDAC2, epithelial-mesenchymal transition, pharmaceutical dots, fluorescence imaging, breast cancer