Hyaluronsäure steuert, ob amputierte Fingerkuppen regenerieren oder vernarben

Mammalian tissue regeneration is exceptionally rare, but the distal tip of the mouse third phalanx (P3) is a well-established exception: it can fully regrow after amputation, including bone, nail, and soft tissue. More proximal amputations through the second phalanx (P2) fail entirely, ending in scarring. This paper, published in Science in April 2026 by Mui et al. from the Cambridge Stem Cell Institute, set out to understand what drives these divergent outcomes — and found a compelling answer in the extracellular matrix (ECM) and its mechanical properties.

Using second harmonic generation microscopy and transmission electron microscopy, the authors first showed that non-regenerative P2 wounds contain dense, architecturally organized collagen fibers resembling fibrotic tissue, while regenerative P3 blastemas are largely devoid of fibrillar collagen. Single-cell RNA sequencing — integrating new datasets with published blastema and non-regenerative wound data — revealed seven major cell populations. Fibroblasts (marked by Pdgfrα expression) carried the highest ECM activity scores and divided into eight sub-types. Fibroblast 1 (PDGFR α+CD34+CD31−) and Fibroblast 2 (SULF2+) sub-types were enriched specifically in non-regenerative wounds and expressed collagen-modeling factors tenascin X (TNX) and thrombospondin-4 (THBS4), respectively. By contrast, regenerative blastemata were dominated by osteo-lineage cells (OLs) expressing HA-binding genes including Acan, Col2a1, Hapln1, Sox6, and Sox9.

Atomic force microscopy confirmed that non-regenerative wound tissue was significantly stiffer and less fluid than the blastema. Immunostaining showed that HA, HAPLN1, and aggrecan (ACAN) were all substantially more abundant in the blastema, with OLs co-localizing with pericellular HA coats reminiscent of a glycocalyx. Importantly, HA enrichment in regenerative digits only emerged by 10 days post-amputation (DPA), not at earlier timepoints, suggesting that injury-induced ECM remodeling — not just pre-existing cellular identity — shapes regenerative capacity.

To test causality, the team depleted HA in two ways. Serial hyaluronidase injections into regenerating P3 digits at 14 DPA markedly reduced digit length and area, diminished RUNX2+ OLs, and depleted pericellular HA. Dietary supplementation with 4-methylumbelliferone (4-MU), which reduces HA biosynthesis by depleting UDP-glucuronic acid, produced similar results: 4-MU digits at 28 DPA had significantly smaller bone volume, surface area, and length compared to controls, and showed fibrosis-like collagen deposition. Blastema marker ARSI was globally downregulated, Ki67+ proliferating osteoblasts were substantially reduced, and SP7+ osteoblast numbers declined. Histolysis — a normal step in digit regeneration — was also delayed, and no new bone formation was detected by microCT at 14 DPA in 4-MU animals.

The therapeutic arm of the study is particularly compelling. The authors applied HAPLN1 — a link protein that stabilizes HA–proteoglycan complexes in the ECM — to non-regenerative P2 amputations. HAPLN1 treatment softened the wound tissue (reducing stiffness as measured by AFM), reduced scarring and fibrosis-like collagen deposition, and significantly enhanced bone repair at the stump. This demonstrates that manipulating ECM mechanics via HA stabilization can shift a non-regenerative wound toward a more regenerative outcome, even in an adult mammal that would otherwise scar. Together, the findings establish ECM composition and mechanics as master regulators of mammalian regeneration and open a new therapeutic axis — ECM-targeted interventions — for promoting tissue repair in aging and injury.