How Cancer Cells Bend, Fold and Drift in Blood Flow: Shape and Stiffness Hold the Key

Every metastatic journey begins with a moment of extreme mechanical stress. When a tumor cell slips into the bloodstream and becomes a circulating tumor cell, it is squeezed, stretched and dragged by the shear forces of blood plasma moving through vessels far narrower than itself. A new computational study published in the Annals of Biomedical Engineering has now revealed, with unprecedented detail, that the way a cancer cell deforms under this assault is governed first by its shape and then fine-tuned by the stiffness of its membrane and nucleus. The findings offer a mechanistic framework that could reshape how scientists interpret circulating tumor cell transport, design diagnostic devices and even think about blocking metastasis.

The research team, led by Meraj Ahmed, Lahcen Akerkouch and Trung Bao Le at North Dakota State University, together with colleagues at the University of North Dakota, tackled a problem that has long frustrated modelers. Most previous simulations represented cancer cells as idealized spheres or capsules, an assumption that bears little resemblance to the irregular, protrusion-covered geometries seen under the microscope. To break free of that limitation, the team cultured breast cancer cells of the MDA-MB-231 line on glass coverslips, fixed and stained them, and then imaged individual cells with a Leica Stellaris confocal microscope at an axial step size of half a micrometer. From these image stacks, four distinct three-dimensional cell geometries, labeled M-1 through M-4, were reconstructed and converted into triangulated surface meshes that preserved the experimentally observed irregularity.

The simulation framework itself is a hybrid of two computational worlds. The cell membrane and the nuclear envelope were modeled using dissipative particle dynamics, a mesoscopic technique in which the membranes are represented as networks of nonlinear springs obeying a Helmholtz free energy functional. That energy accounts for in-plane stretching through a Wormlike Chain-Power formulation, bending resistance, and the conservation of surface area and enclosed volume, with an additional short-range repulsive potential preventing the membrane and nucleus from unphysically overlapping. The surrounding blood plasma, by contrast, was treated as a continuous incompressible Newtonian fluid governed by the three-dimensional unsteady Navier-Stokes equations. The two descriptions were stitched together with a sharp-interface curvilinear immersed boundary method, so that forces generated by the deforming membranes are transferred to the fluid while the local fluid velocity is interpolated back to the membrane nodes. Each of the four geometries was then simulated under three stiffness combinations, in which the Young’s moduli of the membrane and nucleus were varied from 1.26 to 2.52 times ten to the minus five newtons per meter, yielding twelve cases in total.

Before any flow was applied, the reconstructed cells had to be brought to mechanical equilibrium. Because scanned geometries cannot generally be represented by equilateral triangles, the team ran shape relaxation simulations for 900 milliseconds under force-free conditions, temporarily removed the scanned nuclei, and then reintroduced them as spheres inside the relaxed cell volumes. Only then were the cells placed in a rectangular microchannel with plasma viscosity of 1.2 times ten to the minus six square meters per second and a bulk velocity of one millimeter per second, ramped up over half a millisecond. The simulations, run on 64 CPUs at North Dakota State University’s computing cluster with a physical time step of one microsecond, captured the first ten milliseconds of deformation in exquisite detail.

The first striking result is how fast everything happens. All four cell models exhibited a rapid deformation response within the first one to two milliseconds of flow onset, followed by shape evolution that depended strongly on both morphology and stiffness. To classify the ever-changing shapes, the researchers introduced a scheme based on two classical indices, sphericity and aspect ratio, partitioning the shape space into five categories: round or compact, amoeboid ellipsoidal, lobed, elongated, and streamer. Model M-1, which began highly elongated with an aspect ratio above ten, remained streamer-like throughout. Model M-2, the most compact, stayed amoeboid ellipsoidal with remarkable stability, its aspect ratio hovering between roughly 4.4 and 5.3 and its sphericity near 0.8. The transition-prone models M-3 and M-4 told a more dramatic story, switching between compact, elongated, lobed and streamer states as time progressed and as stiffness changed.

Stiffness, it turns out, does not act uniformly. Membrane stiffening dominated elongation and the loss of compactness, while nuclear stiffening mainly modulated the timing and amplitude of deformation excursions and the degree of partial recovery. The most vivid example came from model M-3, which was compact ellipsoidal for all stiffness cases at five milliseconds but, by ten milliseconds, had transitioned to a streamer state only under the stiffest membrane and nucleus combination, a stiffness-induced switch from compact deformation to tether-dominated elongation. Model M-4 underwent a clear temporal transition, starting elongated, briefly becoming streamer-like, and then converging to a folded, lobed configuration by ten milliseconds regardless of stiffness, indicating that lobing and folding, rather than continued stretching, dominated its loss of compactness.

Because the framework resolves the extracellular flow as well as the cell, the researchers could watch how these shape changes reorganized the surrounding fluid. When membrane stiffening did not substantially change the cell’s shape class, the wake vortices were merely stretched and pushed closer to the membrane as near-surface velocity gradients sharpened. But when stiffness drove a genuine shape transition, as in M-3’s shift from compact to streamer, the flow responded dramatically: separation and reattachment points moved, recirculation zones shifted, and some vortices were effectively washed away. Traction maps, which combine pressure and viscous shear on the membrane surface, told a complementary story. The compact M-2 showed the smoothest and most symmetric loading with weak stiffness sensitivity, while the deformation-prone models concentrated traction into hot spots at geometric constrictions, high-curvature protrusions, neck and tether regions, and surface folds, with membrane stiffening sharpening these gradients into smaller, more intense patches.

The net hydrodynamic force on the membrane followed a common pattern across all models: a start-up rise within the first half millisecond followed by relaxation. Beyond that shared transient, morphology took over. The compact M-2 settled at the lowest and steadiest force levels, around 14 to 16 piconewtons, with the three stiffness curves nearly overlapping. M-1 stabilized near 18 to 20 piconewtons with mild unsteadiness. The deformation-prone models showed far richer behavior, including a compliant-case overshoot in M-3 reaching about 30 piconewtons at six to seven milliseconds, and late-time values up to 25 to 27 piconewtons in M-4 near ten milliseconds. Cross-sectional velocity and vorticity fields added the final piece, revealing a pronounced asymmetry about one lateral axis that produces a dominant hydrodynamic imbalance and drives lateral migration, with stiffness modulating the intensity of these features without changing their direction.

The clinical implications ripple outward in several directions. Highly deformable cells that adopt elongated or lobed configurations can pass through narrow capillaries more easily and are more likely to drift toward vessel walls, raising the probability of the adhesion and extravasation events that seed distant metastases, whereas stiffer cells tend to remain in the flow core. Cell deformability is increasingly recognized as a biophysical biomarker of malignancy, and the deformation-driven lateral migration described here could enable label-free microfluidic sorting of circulating tumor cells. Therapeutically, interventions that alter membrane or nuclear stiffness, such as cytoskeletal drugs or lamin-targeted strategies, might shift deformation pathways and reduce metastatic potential by limiting wall-ward transport. The authors caution that their simulations used periodic rather than no-slip lateral boundaries and did not include red blood cell collisions or the cell-free layer, effects they plan to incorporate in future work. They also note that the trends identified, with membrane stiffness controlling elongation and compactness and nuclear stiffness governing recovery and internal reorganization, are expected to hold qualitatively even in curved vessels and at higher velocities, even though exact transition thresholds may shift. For now, the study provides something the field has lacked: a mechanistic bridge linking what a cancer cell looks like, how stiff it is, and where the blood flow will carry it.

Subject of Research: Cancer cell deformation under shear flow as governed by cell morphology and membrane and nuclear elasticity

Article Title: Impacts of Morphology and Elasticity on Cancer Cell Deformation in Shear Flows

Article References: Ahmed, M., Akerkouch, L., Vanyo, A., Haage, A., & Le, T. B. (2026). Impacts of Morphology and Elasticity on Cancer Cell Deformation in Shear Flows. Annals of Biomedical Engineering. https://doi.org/10.1007/s10439-026-04363-8

Image Credits: AI Generated

DOI: 10.1007/s10439-026-04363-8

Keywords: circulating tumor cells, cancer cell deformation, shear flow, dissipative particle dynamics, immersed boundary method, membrane stiffness, nuclear stiffness, metastasis, mechanobiology, microfluidics, fluid-structure interaction, cell mechanics