Swirling motion of breast cancer cells radially aligns collagen fibers to enable collective invasion
bioRxiv. 2025. DOI 10.1101/2025.01.31.635980. PMID 39974994. PMCID PMC11838510.
How to cite
AMA
Saraswathibhatla A, Rabbi MF, Varma S, Srivastava V, Ilina O, Alyafei NHK, et al. Swirling motion of breast cancer cells radially aligns collagen fibers to enable collective invasion. bioRxiv. 2025. doi:10.1101/2025.01.31.635980APA
Saraswathibhatla, A., Rabbi, M. F., Varma, S., Srivastava, V., Ilina, O., Alyafei, N. H. K., et al. (2025). Swirling motion of breast cancer cells radially aligns collagen fibers to enable collective invasion. bioRxiv. https://doi.org/10.1101/2025.01.31.635980BibTeX
@article{saraswathibhatla2025swirling,
title = {Swirling motion of breast cancer cells radially aligns collagen fibers to enable collective invasion},
author = {Saraswathibhatla, A. and Rabbi, M. F. and Varma, S. and Srivastava, V. and Ilina, O. and Alyafei, N. H. K. and others},
journal = {bioRxiv},
year = {2025},
doi = {10.1101/2025.01.31.635980},
pmid = {39974994}
}In invasive breast cancer, the collagen fibers just outside a tumor line up pointing outward, a pattern called TACS-3 that gives invading cells tracks to follow. How the cancer cells themselves build that alignment has stayed unclear.
Working in engineered collagen–alginate matrices with tumor spheroids, live imaging and a computational fiber model, the team found the cells do not pull the matrix outward. They swirl sideways along the tumor edge, and that tangential motion generates shear forces that squeeze the surrounding collagen into radial alignment through negative normal stress, an under-appreciated property of collagen networks. An intact basement membrane blocks the effect; once it is breached, the swirling aligns collagen and collective invasion follows.
Key findings
- Cancer cells swirl sideways, yet the collagen still aligns outward. Time-lapse imaging of tumor spheroids on days 1, 3 and 5 showed cells migrating tangentially to the tumor–matrix interface in a non-coherent swirling motion, with no net inward or outward movement, even as collagen aligned radially before invasion (n≥10 spheroids; N=3).
- The alignment comes from shear-induced contractile stress, not cells pulling outward. Shear rheology of invasive breast-cancer tissue and engineered matrices produced contractile negative normal stresses (n=4 patients), and a fiber computational model reproduced radial alignment from purely tangential forces, with power-law deformation exponents of 2.2 tangential versus 1.2 radial.
- In patient breast tissue, collagen alignment tracked basement-membrane breach. Multiplex immunofluorescence of normal, DCIS and invasive tissue-microarray cores found collagen more radially aligned in invasive than in DCIS regions (n≥3 patients), and tumor-cell clusters that had lost an intact laminin-332 basement membrane showed lower circularity and more radially aligned collagen (n=6 clusters; N=3 patients).
Orion in the methods
“Paraffin-embedded tissue microarray (TMA) sections (Stanford TA 417, 419 and 424 – normal, DCIS and IBC regions respectively) were stained for Immunofluorescence (IF). Primary antibodies used for IF staining include laminin-332 (Santa Cruz Biotech catalogue no. sc-28330) conjugated with ArgoFluor676, E-cadherin (Rarecyte, conjugated with ArgoFluor546) and collagen-I (Abcam ab138492 Anti-Collagen I antibody [EPR7785]) conjugated with ArgoFluor706. Followed IF staining protocol recommended by Rarecyte. The slides were imaged using Rarecyte (Orion Imager).”
— Saraswathibhatla et al., bioRxiv (2025), Methods, “Immunostaining and imaging”
Why it matters for Orion users
If you are weighing Orion for a tissue study, look at the specific job its spatial arm did here. The mechanism was worked out in engineered matrices and models, but the claim only becomes clinical when it holds in real tumors, and that is where the multiplex immunofluorescence came in. Orion imaged paraffin-embedded breast tissue-microarray cores — normal, DCIS and invasive — reading a basement-membrane marker (laminin-332), an epithelial marker (E-cadherin) and collagen-I together in the same intact section. Because those channels are co-registered, the authors could score basement-membrane integrity and collagen architecture around the very same tumor-cell clusters, then segment the collagen fibers to quantify radial alignment per patient. That is the value of reading several proteins in place rather than across serial slides: structure and its context stay on the same cells. If your question is where one marker sits relative to another in patient tissue, this is the readout Orion is built to give you, on archival tissue you already have.
