Some tissues in the human body regenerate surprisingly quickly. For example, after a minor eye injury, the cornea can heal in just one day. However, the brain is not one of these organs. Most brain neurons are formed at an early age and persist throughout life, with almost no renewal. That is why stroke, head injury, or other brain lesions often lead to long-term and severe consequences.

Researchers have long considered stem cell therapy as a possible way to restore lost functions, but so far these attempts have encountered serious obstacles.

One of the main problems is the environment of the damaged brain itself. In the developing brain, stem cells receive specific signals that help them develop into neurons and integrate into neural circuits. But after a stroke, the situation is different: a cavity filled with inflammatory molecules forms at the site of the lesion, and around it is dense scar tissue. Such an environment is not only hostile to the transplanted cells, but also hinders the regeneration of neural connections.

Researchers from the Sanford Burnham Prebys Medical Discovery Institute and the National University of Singapore School of Medicine have presented their work, where they have managed to overcome these limitations, in the journal Cell Stem Cell.

These researchers have developed a therapy based on human stem cells, which is supplemented with a special mixture of small molecules and structural proteins, which promotes cell survival. When transplanted into the brains of stroke-modeled mice, the cells not only survived in an unfavorable environment, but also grew to fill in the damaged area.

But the key question was not simply the survival of the new neurons, but their ability to integrate into existing neural networks. Using three-dimensional reconstructions, the researchers showed that the projections of the transplanted neurons formed structures that resembled the natural pathways connecting the cerebral cortex to the spinal cord. This meant that the cells were not creating chaotic connections, but rather rebuilding lost neural “bridges.”

Additional analysis, using genetic labeling and gene expression sequencing, showed that different types of transplanted neurons have their own “internal programs” that determine exactly where to direct their projections.

Machine learning enabled to identify several subtypes of neurons, each of which established connections with specific parts of the brain. And changing the work of individual regulatory proteins led to significant deviations in the direction of growth of these connections.

According to the researchers, this discovery is of fundamental importance, as it shows that with the right selection of cells, they already “know” how to restore lost circuits. This work opens up prospects for creating more precise and predictable cell therapies for the treatment of stroke and other severe neurological diseases.