Collagen exists as liquid droplets inside living cells, study reveals

Collagen, the protein that builds skin, bones, tendons and organs, exists inside cells as a liquid-like droplet rather than the long, rigid rod seen in textbooks over the last half century, according to a new study from the Centre for Genomic Regulation (CRG) in Barcelona.

The finding, published in the Journal of Cell Biology, is the first direct observation of how the most abundant protein in the human body, which accounts for around a third of total protein mass, exists naturally inside living cells.

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."

Vivek Malhotra, ICREA Research Professor, senior author of the study, CRG, Barcelona

The liquid-like state may serve a protective function. Collagen's job, once outside the cell, is to assemble into the rigid fibres that hold tissues together. The same process inside the cell would be catastrophic. "This is another way by which cells ensure that collagens probably never become fibrous inside the cell," says Malhotra. "Because if it were to become fibrous, it would kill the cell."

The finding has implications for how the body exports its primary structural building block from production sites inside cells. The researchers suggest cells avoid using conventional receptors or vesicles, the route established by work carried out in the 1980s and 1990s and recognised with a Nobel prize in 2013.

Instead, they propose a "liquid extrusion" hypothesis, whereby collagens move from their site of synthesis to the next compartment of the secretory pathway through capillary action. The theory has important implications for wound healing, fibrosis and cancer.

A sixty-year puzzle in cell biology

Collagen is built inside a cellular compartment called the endoplasmic reticulum (ER). The study specifically looked at a precursor form inside cells called procollagen 1, which matures into type 1 collagen. Type 1 collagen is the most common type of collagen, consisting of around 90% of the body's total collagen.

Under a microscope, purified collagen looks like a long, rigid, rod that can be up to 400 nanometres long. However, vesicles, the sacs which transport proteins out from their site of synthesis to the cell's exterior, are only 60 to 90 nanometres in diameter.

Since collagen's structure was first described more than half a century ago, the field of cell biology has asked how such large molecules can be transported out of cells. The new answer is that, inside the cell, collagen is not yet a rod. The canonical picture of the protein describes collagen only after it has left cells and assembled into the fibres that hold tissues together.

Using high-resolution live-cell imaging of human hepatic stellate cells, the liver cells that produce collagen and drive scarring in liver fibrosis, the team showed that collagen inside the cell gathers into small droplets that merge, split and exchange material with their surroundings.

These are all signatures of a condensate, compartments of proteins that become so concentrated they disassociate from their surroundings, like droplets of oil in water.

According to Soumya Bhattacharyya, first author of the study, most of cell biology has focused on condensates in the nucleus and on stress granules in the cytosol. "We're just beginning to understand condensates inside the endoplasmic reticulum.," says Bhattacharyya.

The discovery: "I had no idea where it would lead to"

The findings emerged from microscopy images taken by Dr. Soumya Bhattacharyya, a postdoctoral researcher in Vivek Malhotra's lab, in May 2024. Bhattacharyya was using the liver cell system as a tool to study what happens when collagen production is increased in fibrotic cells.

"I had no idea what it would lead to. But when we took the samples, what struck me were these bright spherical structures you can't miss." recalls Bhattacharyya.

The initial reaction in the laboratory to a finding that challenges cell biology dogma was sceptical. "I thought it must be an artefact," says Malhotra.

In the months that followed, the team had to settle whether the protein clumping they observed inside the endoplasmic reticulum was junk. Cells have an elaborate system for detecting badly folded proteins and either refold them or mark them for destruction, centred on a chaperone called BiP.

If the collagen droplets were heaps of misfolded protein, the researchers would detect high levels of BiP. The droplets contained, instead, a mixture of helper proteins including chaperones that specifically recognise properly folded collagen.

The role of TANGO1

The study also clarifies the function of TANGO1, a protein discovered by the Malhotra lab roughly two decades ago and known to be required for collagen export. When the researchers depleted TANGO1, the collagen droplets still formed but were no longer positioned at the ER exit sites where cargo leaves the compartment. Collagen secretion dropped accordingly.

The discovery suggests TANGO1 acts as a mooring point that holds the droplet at the export site rather than as a conventional cargo receptor. The authors propose that collagen then leaves the cell by a physical process called wetting, in which the liquid droplet attaches to and flows through the exit site.

Malhotra offers two possible physical mechanisms for this transfer. "Imagine you have a rubber ball with a nozzle, filled with liquid. You squeeze it, you force the liquid to come out of this little orifice. Is that the mechanism? Or is the liquid rising by capillary forces, just like nutrients flow up against gravity in plants by capillary action?".

The proposed liquid extrusion mechanism remains a model, but the next experiments in order to obtain direct visualisation of the export mechanism are already underway. The team also plan to develop a mouse model, in collaboration with external partners, to confirm the findings in living tissue.

Implications for fibrosis and cancer

If the model is confirmed, the work has implications for several pathological conditions in which excess collagen secretion plays a central role, including liver, lung and skin fibrosis, as well as for targeting the dense matrix that tumors use to shield themselves from chemotherapy and the immune system.

"One of the major problems in cancer is that the cells secrete so many collagens and other proteins out into the extra cellular matrix that they hide in a shell made of these components and become chemo- and immuno-refractory, meaning they are not seen by the chemical therapeutics or by the immune system," Malhotra says.

"People are trying to find ways to break this tissue cement and our study could help inform those strategies," he adds.

The proposed collagen secretion model suggests either degrading TANGO1 to prevent cargo from being captured at the exit site or dissolving the condensate itself to prevent the cargo from being properly organised in the first place, could be new strategies worth exploring.

The study was led by Soumya Bhattacharyya and Vivek Malhotra at the CRG, with contributions from Jose Wojnacki and Nathalie Brouwers and technical support from the CRG Advanced Light Microscopy Unit and CRG/UPF Flow Cytometry Unit. The study was funded by the Spanish Ministry of Science and Innovation, the Generalitat de Catalunya and the European Research Council.