Notre Dame team prints capillaries under 10 micrometres, easing bioprinting's vascular limit

A team led by Yanliang Zhang of the University of Notre Dame has printed vascular networks whose smallest channels measure fewer than 10 micrometres across, narrower than the finest human hair, in work published in Nature Chemical Engineering ^[1][2][17]. That figure matters because reproducing the body's vascular networks, and capillaries in particular, has been the main obstacle to building useful bioprinted tissue ^[3].

Capillaries carry oxygen and nutrients to every living cell, which is why the authors treat them as an essential component of any tissue model worth having ^[4]. Zhang's framing of the problem is specific: vessels vary in size, and getting the smallest ones right without losing scalability and structural integrity has remained one of the hardest tasks in current bioprinting ^[5].

The method combines two printing techniques. A soft, gel-like scaffold that stands in for tissue is laid down by extrusion, a common pressure-based method that dispenses material layer by layer ^[6]. Threads of gelatin are then deposited inside that matrix by aerosol jet printing and later removed, leaving channels embedded in the gel ^[7]. Aerosol jet printing uses aerodynamic focusing with a sheath flow, which lets the team set channel sizes from hundreds of micrometres down to several micrometres, matching the varied architecture of natural vessels ^[8].

Because small changes in the ink flow rate and the sheath gas flow rate change the finished channel size, the researchers built a machine learning framework to find the right combination of parameters for a given size, instead of searching by trial and error ^[9][10]. The approach produced stable one-, two- and three-dimensional vascular structures ^[11]. Selected channels were then seeded with endothelial cells, the cells that line the body's blood and lymphatic vessels ^[12][13]. According to the researchers, the cells rapidly attached and spread along the inner walls into single-cell layers, replicating the barrier function of real human capillaries ^[14].

The work was done in collaboration with Y. Shrike Zhang of Harvard Medical School and Brigham and Women's Hospital ^[15]. Yanliang Zhang is an affiliate of Notre Dame's Berthiaume Institute for Precision Health and NDnano ^[16].

The context the authors cite is the transplant queue: more than 100,000 people in the United States are waiting for an organ, with a candidate added every 10 minutes ^[18]. Whole printed organs remain far from that. What the paper demonstrates is capillary-scale resolution and a living endothelial lining in a lab structure, not a functioning organ or an implant ^[2][12]. For organ-on-chip and tissue-model roadmaps, the near-term consequence is narrower and more useful: a way to build a hierarchical vascular network down to capillary width, plus a parameter-tuning method that does not depend on manual iteration.

What to watch is whether these networks perfuse under flow and hold their barrier function over time, and whether the machine learning parameters transfer across different inks, cell types and larger tissue volumes.