Astrocytes Refill Brain Lesions by Moving New Nuclei Through Their Own Branches

The swellings appeared inside the branches. Researchers at the University of Zurich were tracking individual star-shaped brain cells in living mice, week after week, through a glass window set into the skull, and in cells sitting at the rim of a small wound they kept seeing thickened spots partway along the long arms that reach out from the cell body. The spots sat at different distances from the body in different cells, as though something were parked in transit. Thirty days in, the team washed a dye that stains cell nuclei across the exposed brain, went back to the same cells, and found that each swelling held a nucleus.

Those arms belong to astrocytes, the most numerous support cells in the brain. Astrocytes feed neurons, help regulate local blood flow and clear spent neurotransmitters, and they do it in a strikingly orderly way: each cell claims its own patch of tissue and keeps its branches almost entirely out of its neighbors’ patches, so the cortex is covered by a quilt of barely overlapping territories.

A Human Antibody Emptied a Patch of Cortex

To find out what happens when part of that quilt goes missing, the team needed a way to delete astrocytes and nothing else. They used an antibody reconstructed from a patient with neuromyelitis optica spectrum disorder, a rare autoimmune disease in which the immune system attacks a water channel called aquaporin-4 that sits mainly on astrocyte surfaces. Injected into the somatosensory cortex of a mouse along with complement, the blood proteins the antibody needs to finish the job, it carved out a cylinder of tissue with no astrocytes left in it.

Neurons inside that cylinder survived. Blood vessels running through it kept their diameter and their flow, which matters because it rules out the simplest competing explanation for everything that follows. This was not a wound made by cutting, crushing or starving the tissue of blood.

Then the hole filled in. Over the following weeks the astrocyte-free area shrank on the microscope screen, the number of astrocyte cell bodies inside it climbed and leveled off, and no glial border formed, none of the dense wall of glial cells that normally walls off a brain injury from the surviving tissue.

What refilled was the cell network, not necessarily the function. Nothing in this study tested whether the rebuilt patch does what the original patch did, whether the neurons inside it work better, or whether a mouse with a repopulated lesion behaves any differently from one without.

Six in Ten Rim Cells Divided

The cells doing the refilling were not new arrivals. By labeling scattered astrocytes in a second color and following them one at a time, the team traced every new cell inside the lesion back to a surviving cell on the rim, and found that roughly 60 percent of those rim cells divided at least once in the two months after injury, against about 1 percent in control tissue. Some divided more than once, producing up to four daughters. The dividing cells were not confined to those touching blood vessels, which is where astrocyte division after other kinds of brain injury tends to concentrate.

The Nucleus Moved and the Cell Stayed Put

Division is not the surprising part. What the swellings revealed is where the daughters went. To catch the movement, the team switched to awake mice, trained over about a week to hold still under the objective, and imaged every three hours for six straight days. The pictures show a new nucleus traveling along the inside of the mother cell’s thickest arm in fits and starts: a burst of movement, then a pause that could run nine hours, then another burst. Across seven nuclei followed this way, top speeds averaged about 3.5 micrometers an hour and reached 5.6 at the fastest, a little over a hundredth of a millimeter a day.

While it moves, the nucleus stops being round. Squeezed through a channel narrower than itself, it stretches into an ellipse, and once it arrives it rounds out again. Biologists already have a name for this, nucleokinesis, and they have watched it in the developing brain, where newborn neurons advance in the same start-stop rhythm, hauling the largest and stiffest thing they own through packed tissue toward their final position. Seeing it in a mature astrocyte in an adult brain is what is new here.

Mother and Daughter Shared One Body

An obvious alternative remained: perhaps the daughter separates at once and crawls along its mother’s arm, the way progenitor cells travel along glial fibers in a developing brain. To settle it, the team located one mother and daughter pair under the light microscope, then returned to the same tissue with an electron microscope and reconstructed both cells slice by slice. The two nuclei, tens of micrometers apart, sat in one continuous cytoplasm with no membrane between them. For some days after dividing its chromosomes, the astrocyte is a single cell carrying two nuclei, one of them being walked out into vacant territory. They do separate eventually. When mother cells were killed with a laser months later, six of seven daughters survived.

Stroke Lesions Did Not Behave This Way

The team then made stroke lesions of similar size in the same region, clotting the local vessels with a light-activated dye. Astrocytes at the rim divided just as readily, but almost none of the daughters traveled. Only about 18 percent moved at all, and every one of those stayed within 30 micrometers, roughly one astrocyte’s own territory, while in the antibody lesions nuclei went as far as 113. A glial border formed and the lesion never refilled. Dividing, in other words, is not the bottleneck. One obvious candidate for the missing signal failed too: mice lacking both proteins that form the junctions between neighboring astrocytes refilled their lesions as well as normal littermates.

The team also examined brain tissue from seven people affected by the same autoimmune disease. Around the lesion edges they found astrocytes with two, three and four nuclei, drawn into the same elongated shapes, some carrying a marker of active division. “The findings of our study reveal a previously unknown ability of the adult brain to repair itself,” says Bruno Weber, who led the group. But fixed tissue is a snapshot. It shows the shapes without showing the movement, and what those cells were doing for the patients is unknown.

Whether any of this starts at all seems to be a question of how much is gone. Killing astrocytes one nucleus at a time with a laser, the team found that removing three to five cells drew almost no response from the neighbors, and removing a single flat layer made them stretch toward the gap without dividing. Only when the killing went deeper than about 100 micrometers, through several layers of territory at once, did the nuclei start to move. What the surviving cells are measuring, and how, is not in these data.

FAQ