The deepest moat in engineered tissue remains vasculature. Existing printed vascular models mostly fall into two categories: a hollow tube lined with endothelium, or a cluster of microvessels — neither has the native “three-layer wall,” and neither plugs into a hierarchical vascular system.

In 2026, a team at the University of Würzburg published a study in Advanced Science: induced pluripotent stem cell-derived mesodermal progenitor cells were mixed into an optimized bioink, and bio 3D printing produced a centimeter-scale “mother vessel” in one step. Within a week, the cells differentiated and aligned themselves, growing intima, media, and adventitia — three distinct layers. Assembled with prevascularized organoids, the microvascular networks completed the connection on their own.

*All values quoted from the original paper: wall thickness from Abstract (~150 μm median within first 10 days) and quantification (≈150 μm at day 13); coverage and perfusion from Results.

PART 01 Why Vasculature Is the Ceiling Problem of Tissue Engineering

A native blood vessel is not a “plastic tube.” Except for capillaries, nearly all vessels have three wall layers: the innermost intima, a dense endothelial lining pressed against the lumen; the middle media, smooth muscle cells arranged circumferentially for contraction and relaxation; and the outermost adventitia, loose connective tissue hosting fibroblasts and stem/progenitor cells. The paper emphasizes that most bio 3D-printed vascular models replicate neither this multilayered wall nor the in vivo hierarchy — large vessels branching into small vessels branching into capillaries — so their translational value is naturally limited.

The Würzburg team took a “developmental biology route”: during embryogenesis, blood vessels arise from mesodermal progenitor cells that spontaneously differentiate and self-assemble. The team had previously derived mesodermal progenitor cells (hiMPCs) from human induced pluripotent stem cells (hiPSCs) — capable of differentiating into all cell types required for a vascular wall, and still able to self-organize into vessel-like structures after printing. But the bottleneck was clear: self-organization alone cannot produce a “main channel” with an inner diameter over 1 mm, and perfusion and surgical anastomosis require exactly such large-diameter conduits.

The paper’s approach is therefore “each covers the other’s weakness”: bio 3D printing pre-defines the millimeter-scale “mother vessel” skeleton, and self-organization is left to the cells — the mother vessel serves as the main inlet/outlet channel, printed first in a support bath, then surrounding microvessels grow out and connect to it. One print, two scales solved together.

PART 02 Four-Ink Screening: Why FGXC Wins

The team first set four exams for candidate inks: printability, thermal stability, mechanical strength, and ability to support vascular morphogenesis. Alginate and fibrin were eliminated first — cells could not complete wall construction inside them. The protagonist became gelatin methacryloyl (GelMA), but porcine GelMA alone could not support a large mother vessel: poor thermal stability, difficult rheology control, insufficient mechanical strength. The team then added ingredients like a cocktail: collagen I for cell adhesion; cold-water fish GelMA with a lower gelation temperature to improve thermal handling; xanthan gum to tune viscosity and impart shear-thinning; and fibrillar collagen I as rebar to reinforce the matrix.

Four formulations were tested from simple to complex, and the differences were stark: GelMA and GelMA + collagen I showed sharp thermal transitions near 30 °C, with unstable flow behavior; whereas FGX (fish + porcine GelMA + xanthan gum) and FGXC (FGX plus fibrillar collagen) remained stable across 15–45 °C, showed clear shear-thinning, and could be continuously extruded at low pressure (<15 kPa). Printability index (Pr, closer to 1 = more regular): GelMA only 1.35±0.22, GelMA + collagen I 0.92±0.07, both with pore distortion or fiber breakage; FGX 0.89±0.05 and FGXC 0.87±0.03 firmly in the top Class A.

Mechanically, the xanthan gum + fibrillar collagen combination pushed FGXC’s compressive modulus to 13.4±2.6 kPa — the other three hovered at 4–5 kPa — placing FGXC squarely in the mechanical range of soft connective tissue. Under cryo-SEM, FGXC showed a dense, uniform fibrous network, corroborating its high printability index and high stiffness.

PART 03 Printing Tubes in a Support Bath: Soft Ink Can Still Stand

Before crosslinking, FGXC is actually quite “soft” — printing a tube directly in air would cause the uncrosslinked portion to collapse under its own weight. The team’s solution was embedded printing: first prepare a support bath from hyaluronic acid (0.2% w/v) + xanthan gum (1% w/v) — this bath holds the ink like jelly at rest, locally “melts” under the shear of the printing needle, and self-heals immediately after the needle passes. Rheological tests confirmed this self-healing and shear-thinning behavior: the support bath not only provides temporary support during printing but also allows the vessel to be “washed out” intact after crosslinking.

PART 04 Three-Layer Vascular Wall in One Week: Cells “Lined Up” on Their Own

Here is the main event. Day-7 histological sections showed that the lumen surface of the printed tube was already covered by a three-layer vascular wall about 150 μm thick: the innermost layer is CD31-positive endothelial cells lining the lumen as a continuous, intact intima; 3–4 layers inward are αSMA-positive cells arranged circumferentially, forming a media-like smooth muscle ring; further out is loose connective tissue, corresponding to an early adventitia. Even more remarkably, these endothelial cells also expressed eNOS (endothelial nitric oxide synthase) — an important marker of functional endothelial identity; type III collagen deposition and elastic fibers were also detected in the wall, indicating early basement membrane-like structures and extracellular matrix remodeling were already underway.

The adventitia-like layer also harbors a “cell reservoir”: the paper detected CD150-, CD44-, CD45-, CD34-positive hematopoietic and vascular progenitor cells residing within — highly consistent with the native adventitia acting as an “endogenous stem cell niche.” Even more unexpectedly, IBA1-positive macrophage-like cells appeared in the constructs — yet the initial cell suspension contained no macrophages whatsoever. In other words, hiMPCs not only differentiated into vascular wall cells but also spontaneously generated immune cell types — a recapitulation of embryonic vascular development’s “one progenitor does it all,” and something most “differentiate-then-print” engineered vascular models cannot achieve. Tissue-cleared 3D reconstruction further showed microvascular networks penetrating the wall, with angiogenic sprouts extending from the lumen all the way to the construct periphery.

PART 05 The Two-Week Timeline: From Seeding and Layering to “Use It or Lose It”

Extending the timeline to 13 days, the developmental rhythm becomes clearly discernible.

• Day 4: CD31-positive cells form scattered clusters; focal lining appears on the lumen surface — endothelial differentiation just beginning;
• Day 7: endothelial cells elongate and align circumferentially along the lumen; cord-like vascular structures appear peripherally and interconnect into a branching, sprouting capillary network; the three-layer wall takes shape;
• Day 11: media becomes more prominent; wall continues to thicken;
• Day 13: lumen visibly narrows; wall remodels — under static culture, the printed tube begins to collapse.

Quantitative data draw this curve in finer detail: wall thickness rose from a few microns initially to about 150 μm at day 13 (all time points p<0.0001 vs day 4); lumen narrowed continuously from 3.5 mm post-printing, almost collapsing to a slit by day 13; surrounding mesodermal tissue first thickened, then stabilized from day 9. Microvascular networks within the mesoderm evolved in parallel — endothelial branch length increased significantly, CD31-positive area rose from day 3 to day 7 then declined by day 10, which the paper interprets as network maturation and remodeling. Two experiments best demonstrate "function online": endothelial coverage climbed continuously from day 3 to day 10, reaching nearly 90%; in monocyte adhesion assays, endothelium stimulated with TNF-α inflammatory challenge bound significantly more monocytes than unstimulated controls — the native vascular behavior of “recruiting leukocytes during inflammation” was reproduced on this printed tube.

PART 06 Mother Vessel × Organoids: Assembling Large Tissue Like Building Blocks

With the “main channel” in place, the next step is connecting the “neighborhoods.” The team used an agarose honeycomb microwell mold (159 wells, ~700 μm each) to standardize production of prevascularized mesodermal organoids — uniform in size (~300 μm), each containing a branching CD31-positive microvascular network internally. Approximately 750 organoids were then co-cultured with a mother vessel cultured for 4 days for 3 days: organoids not only attached to the printed tube’s outer surface but gradually fused with each other and with the vessel wall, forming a continuous tissue coating around the mother vessel.

Vascular connection was also occurring: adjacent organoids’ microvascular networks fused with each other and showed spatial continuity with the vascular network inside the mother vessel’s mesodermal tissue — macroscopic conduits and microscopic networks “shook hands” within this printing system. Immunostaining showed widespread distribution of CD31 endothelium and αSMA pericytes in organoid-derived tissue, type IV collagen deposition along the vessel wall indicating early basement membrane formation, and IBA1-positive macrophage-like cells also present. The final stress test was perfusion: the team 3D-printed a custom perfusion chamber (22×22×8 mm), clamped the day-7 mother vessel into the chamber, and perfused with a peristaltic pump at 1 mL/min pulsatile flow — no leakage was detected along the entire vessel, only minimal seepage at the clamped ends, directly verifying the wall’s density and robustness.

PART 07 What This Platform Really Breaks Through, and What Remains Owed

The conventional approach to engineered vascular models is to pre-differentiate endothelial cells, smooth muscle cells, and fibroblasts separately, then assemble and print — long workflow, many formulations. This platform’s difference lies in single progenitor source, one-step printing: hiMPCs in the ink differentiate into endothelium, pericytes, adventitial cells, and even macrophage-like cells on their own — the developmental program has been transplanted directly into the printed tube. The “mother vessel” is positioned clearly: not to replace microvessels, but to serve as the main inlet/outlet channel between large tissue and the host vascular system (or perfusion device), connecting macroscopic perfusion with microscopic self-assembled networks into a hierarchical system.

The paper is equally unambiguous about its limitations:

Closing Thoughts

📄 Paper Information

· END ·