Our Take
This is direct microscopy evidence overturning a foundational assumption about protein structure inside living cells, not a computational inference or in vitro artifact—but the export mechanism remains theoretical.
Why it matters
Cell biologists have puzzled for decades how 400-nanometer collagen molecules fit through 60–90 nm transport vesicles. If the liquid-extrusion model holds in living tissue, it reframes how cells handle their most abundant protein and opens therapeutic angles on liver fibrosis, lung scarring, and tumor immunity evasion.
Do this week
Structural biologists: flag collagen condensate dynamics as a new parameter in your ER-stress or secretion pathway models; the team's mouse work (in progress) will settle whether this holds beyond hepatic stellate cells.
Collagen Exists as Liquid Droplets Inside Cells
Scientists at the Centre for Genomic Regulation in Barcelona published direct evidence that collagen, the protein constituting roughly one-third of human body mass, exists inside cells as a liquid-like condensate rather than the long, rigid rod-like structure described in textbooks for more than 50 years.
Using high-resolution live-cell imaging of human hepatic stellate cells (liver collagen-producing cells), the team observed procollagen 1 (PC1, the precursor to type 1 collagen, which makes up 90% of total body collagen) gathering into small droplets that merge, split, and exchange material with their surroundings. These behaviors are hallmarks of phase-separated condensates, compartments of concentrated proteins that disassociate from their surroundings like oil droplets in water.
Lead researcher Vivek Malhotra, ICREA research professor at CRG, explained the implication: "Inside a cell, collagens are not rigid molecules as one had assumed. They are, in fact, very pliable, taking a liquid condensate form much like oil in a drop of water." The findings appear in the Journal of Cell Biology.
This Solves a 50-Year Transport Puzzle and Suggests a New Export Pathway
Cell biologists have long faced a geometric impossibility: purified collagen molecules measure up to 400 nanometers in length, but the vesicles that transport proteins from the endoplasmic reticulum to the cell surface are only 60–90 nanometers in diameter. How such large molecules escape the cell has remained unresolved despite decades of study.
The liquid-condensate model offers an answer. Rather than traveling as rigid fibers through conventional vesicular transport (the Nobel Prize-winning pathway elucidated in the 1980s and 1990s), collagen may move to the ER exit site as a pliable droplet and then extrude through the membrane by capillary action, like water wetting and flowing through a nozzle.
The study also clarifies the function of TANGO1, a protein discovered by Malhotra's lab roughly 20 years ago and known to be essential for collagen export. When the researchers depleted TANGO1, collagen condensates still formed but no longer positioned themselves at the exit sites where cargo leaves the cell. Secretion dropped correspondingly. TANGO1 appears to act as a physical mooring point rather than a conventional cargo receptor.
This mechanism has direct implications for pathologies driven by excess collagen secretion. In liver, lung, and skin fibrosis, tissues harden because cells flood the extracellular matrix with collagen. In cancer, dense collagen matrices act as a protective shell, helping tumors evade chemotherapy and immune attack. Targeting either TANGO1 or the condensate itself could offer new therapeutic leverage in these conditions.
The Model Remains Unproven in Living Tissue
The findings rest on direct microscopy, which is stronger evidence than computational modeling or cell-free assays. However, the team acknowledges the liquid-extrusion hypothesis is still a model. The researchers are already designing experiments to directly visualize the export mechanism and plan to collaborate with external partners to develop a mouse model to confirm the findings in living tissue.
Until in vivo validation arrives, the therapeutic implications (degrading TANGO1 or dissolving condensates to block collagen secretion) remain speculative. The next 12 to 18 months of experimental work will determine whether this represents a fundamental rewiring of how cell biologists understand ER transport or a phenomenon specific to high-collagen-production systems.