Extracellular matrix mediates circulating tumor cell clustering in triple-negative breast cancer metastasis
Nature Communications. 2026;17(1):1352. DOI 10.1038/s41467-026-69007-w. PMID 41651843. PMCID PMC12881594.
How to cite
AMA
Bobkov GOM, Patel KJ, Lege BM, Zheng R, Shaulsky G, Ellis MJ, et al. Extracellular matrix mediates circulating tumor cell clustering in triple-negative breast cancer metastasis. Nat Commun. 2026;17(1):1352. doi:10.1038/s41467-026-69007-wAPA
Bobkov, G. O. M., Patel, K. J., Lege, B. M., Zheng, R., Shaulsky, G., Ellis, M. J., & Cheng, C. (2026). Extracellular matrix mediates circulating tumor cell clustering in triple-negative breast cancer metastasis. Nature Communications, 17(1), 1352. https://doi.org/10.1038/s41467-026-69007-wBibTeX
@article{bobkov2026extracellular,
title = {Extracellular matrix mediates circulating tumor cell clustering in triple-negative breast cancer metastasis},
author = {Bobkov, Georg O. M. and Patel, Khushali J. and Lege, Bree M. and Zheng, Rong and Shaulsky, Gad and Ellis, Matthew J. and Cheng, Chonghui},
journal = {Nature Communications},
volume = {17},
number = {1},
pages = {1352},
year = {2026},
doi = {10.1038/s41467-026-69007-w},
pmid = {41651843}
}Clusters of circulating tumor cells seed new tumors far more efficiently than single cells, yet how they hold together has been unclear in triple-negative breast cancer. The usual molecular glue, epithelial adherens-junction proteins, is largely missing in these aggressive tumors, so something else must bind their circulating cells.
Working in mouse models and patient blood, this team found the binding agent outside the cell: hyaluronan, a matrix sugar, brings tumor cells together by engaging its receptor CD44. Coated cells extend protrusions, pull close, and lock in with desmosomes; the same coat also pulls in non-tumor cells, including immune cells, to build mixed clusters.
Stripping away hyaluronan left cells unable to cluster and far less able to survive the shear of blood flow. That points at the matrix, not the tumor’s own junctions, as what holds these clusters together.
Key findings
- Hyaluronan is required for TNBC cells to cluster, and only large HA restores it. Degrading or knocking down hyaluronan left triple-negative breast cancer cells unable to form stable clusters; adding back high-molecular-weight HA (750–1000 kDa) rescued clustering dose-dependently, while low-molecular-weight HA (130–150 kDa) did not.
- HA-mediated clustering protects circulating tumor cells from the shear of blood flow. Under fluid shear that models flow in arteries and veins, only 20–25% of HA-depleted cells were alive at 24 hours, compared with 65% of controls.
- Blocking HA synthesis cut lung metastasis more than 5-fold in mice, without slowing primary tumors. Knocking down the HA synthase HAS2 produced a more than 5-fold decrease in lung metastasis, while primary tumor growth was not significantly changed.
CyteFinder in the methods
“PBMCs were resuspended in Transfer Fluid (RareCyte) and spread onto slides. After fixation with 10% NBF for 20 min, immunofluorescent staining was performed as described above. CTCs were detected using a cocktail of four antibodies consisting of Pan-Cytokeratin, Cytokeratin 19, EpCAM, and EGFR. Slides were scanned using a CyteFinder® instrument (RareCyte) to identify CTCs.”
— Bobkov et al., Nature Communications (2026), Methods, “Patient CTC samples”
Why it matters for CyteFinder users
If you work on circulating tumor cells, notice what this study asked its imaging step to do. The central claim is about cluster composition: hyaluronan pulls non-tumor cells, including immune cells, into mixed clusters alongside the tumor cells. You cannot read that composition if the workflow selects for an epithelial marker first, because that selection would discard the very partners the paper set out to count. Here the authors took the opposite route. They spread whole blood pellets and patient PBMCs onto slides and scanned them on CyteFinder, imaging the deposited cells rather than selecting for a marker up front. That no-enrichment, slide-based approach is what let them find rare tumor cells against a large white-blood-cell background and read out both the tumor cells and their non-tumor neighbors in one field. For your own work the trade is concrete: you keep heterotypic clusters intact and every deposited cell in view, at the cost of scanning that whole population rather than a pre-selected subset. When the biology lives in what surrounds the tumor cell, that is the trade you want.
