Flexible ensheathment of axons enables myelination of complex CNS networks
Article Date: 01 April 2026
Article URL: https://www.nature.com/articles/s41586-026-10312-1
Article Image: (none provided)
Summary
This Nature paper revises how oligodendrocytes build myelin in the central nervous system. Using longitudinal in vivo imaging in zebrafish and mice, volumetric electron microscopy across zebrafish, mouse and human samples, and human iPSC-derived myelinoids, the authors show that myelin wrapping is not always a smooth, synchronous process. Instead, individual oligodendrocyte processes wrap axons asynchronously and at variable rates. When a process encounters axon branch points or nodes of Ranvier it can extend beyond them and form separate compacted myelin segments linked by a thin cytoplasmic connection called a paranodal bridge. These bridged chains expand the territory an oligodendrocyte process can myelinate, are frequent on highly branched axons (notably PV interneurons), appear in zebrafish, mouse and human tissue, and are more vulnerable to degeneration in ageing cortex.
Key Points
- Ensheathment by single oligodendrocyte processes is often asynchronous along an axon: wrapping can start, pause or proceed at different speeds at multiple sites.
- When processes extend past axon branch points they can produce two (or more) separately compacted sheaths connected by a thin cytoplasmic paranodal bridge.
- Bridged sheath chains are common across species (zebrafish, mouse and human) and can significantly increase the effective myelinated length supplied by a single oligodendrocyte process.
- PV interneuron axons, which are highly branched, show a higher incidence of paranodal bridges than average cortical axons, linking bridge formation to axon morphology rather than oligodendrocyte subtype.
- Bridged sheaths are more susceptible to degeneration in aged mouse cortex, suggesting physiological consequences for maintenance and remyelination with age.
- Bridge formation is largely independent of axonal vesicular exocytosis (blocking neurotransmitter release did not change bridge frequency in zebrafish), suggesting the mechanism is structural and opportunistic rather than strictly activity-driven.
Content summary
The authors combined time-lapse imaging in larval zebrafish and two-photon imaging in mouse cortex with serial electron microscopy reconstructions. In zebrafish they observed many nascent sheaths with unwrapped gaps (bridges) that sometimes persist for weeks. Lifeact imaging showed these gaps represent genuine non-wrapping segments rather than imaging artefacts. Electron microscopy confirmed partially encircled axon segments and outer-tongue continuity across nodes consistent with paranodal bridges. In mouse visual cortex roughly half of nascent sheaths were irregularly wrapped and >30% of nodes in human cortex showed immunohistochemical evidence for paranodal bridges. In organoid cultures derived from human iPSCs, many oligodendrocytes also produced bridged sheaths. The authors propose a revised model: rather than uniform concentric wrapping from a single initiation site, oligodendrocyte processes can initiate wrapping at multiple sites and complete wrapping asynchronously, allowing efficient myelination across branch points without needing extra oligodendrogenesis.
Context and relevance
This work challenges the canonical, synchronous ‘liquid croissant’ or carpet-rolling models of CNS myelination and reconciles longstanding ultrastructural observations of gaps, alternating wrap directions and filopodial features seen in older anatomical studies. It explains how limited oligodendrocyte resources and heavily branched cortical axons (notably PV interneurons) can still be effectively myelinated. The finding that bridged sheaths are comparatively fragile with ageing is important for thinking about age-related myelin loss and circuit dysfunction in diseases that affect interneurons (for example, multiple sclerosis, schizophrenia, Alzheimer’s disease).
Why should I read this?
Short version: if you care about how the brain wires fast and reliably — and why myelin sometimes fails — this paper gives you a practical re-think. It’s full of beautiful in vivo movies and EM reconstructions and explains a neat trick oligodendrocytes use to myelinate messy, branched axons without making more cells. If you work on myelin, interneurons, remyelination or ageing brains, skip the slog and jump to the figures — they’ve done the heavy lifting for you.
Author style
Punchy: the authors redefine a core cellular mechanism. They don’t just add nuance — they propose a new, conserved, functionally relevant behaviour (asynchronous, multisite ensheathment and paranodal bridging) that solves a real spatial problem in cortical neuropil. The paper is a must-read if you want to follow where myelin biology is headed.