Fluid viscosity drives mechanical memory and invasion in glioblastoma

Most cancer studies focus on chemical signals or stiff tumor surroundings, but the stickiness of the fluid itself has received little attention. For glioblastoma, the invasion front is about eight times more viscous than the necrotic core, creating a rising resistance that migrating cells must overcome. Standard closed microfluidic systems poorly mimic this condition: they restrict oxygen and nutrients, alter cell behavior through wall friction, and make long-term observation difficult. Based on these challenges, there is a clear need to study how sustained exposure to high viscosity remodels glioblastoma cells without interference from additional physical constraints.

Researchers at Chongqing General Hospital and Chongqing University, China, publish (DOI: 10.1038/s41378-026-01241-0) this work on April 13, 2026, in Microsystems & Nanoengineering. They developed a two‑layer open microfluidic membrane with a detachable cap and a micropillar array. The design precisely controls when migration starts, allows real‑time imaging of nuclear deformation, and supports long‑term culture for up to one month, revealing how viscosity‑driven mechanical stress transforms glioblastoma invasion.

The team cultured two human glioblastoma cell lines, U‑251 and LN‑229, for one month in a viscous medium matching the tumor's invasive periphery (7.1 cP). When placed on the chip, viscosity‑adapted cells migrated farther and faster than control cells, even though the thicker fluid normally slows movement. Microscopy showed that these cells became smaller and more deformable, helping them slide through narrow valleys between micropillars. Inside those confined valleys, nuclei were visibly squeezed and the mechanosensitive protein YAP accumulated in the nucleus—a known sign of mechanical activation. Strikingly, the two cell lines responded very differently at the molecular level. U‑251 cells underwent a mesenchymal‑like reprogramming, turning on invasion‑related genes such as CD44, FN1, and MMP9. LN‑229 cells changed their shape and migration similarly but showed almost no lasting gene‑expression shift. Western blots confirmed that the protein changes persisted even after cells were returned to normal‑viscosity medium, indicating a stable, not temporary, adaptation.

The authors said that they were surprised to see viscosity alone act as a lasting instructor rather than just a physical hurdle. They explained that the open‑chip design finally allows them to separate fluid resistance from wall confinement, two forces that are usually mixed together in closed systems. They added that watching cells change their nucleus and migration strategy after weeks in thick fluid was a clear sign that mechanical memory exists in these tumor cells. For them, the most striking result was how one cell line rewrote its gene program while another stayed largely unchanged, despite looking similar under the microscope.

The open microfluidic platform can be placed in standard multi‑well plates, making it compatible with routine cell‑culture and imaging workflows. Because it allows direct access for staining and long‑term live imaging without clogging, the chip offers a practical way to screen drugs that target mechanosensitive pathways. For glioblastoma, the findings suggest that high viscosity may actively select for more invasive cells, so therapies aimed at YAP signaling or cytoskeletal remodeling could be tested under more realistic physical conditions. More broadly, the device can be adapted to study other cancers where viscosity gradients exist, helping to identify patients whose tumors might rely on mechanical adaptation to spread.

Source:

Chinese Academy of Sciences

Journal reference: