Nearly a third of zebrafish, the benchmark species for spinal cord regeneration, remain paralyzed after injury. Scientists identified hoxb5a as the gene that determines whether a fish recovers

Nearly a third of zebrafish, the benchmark species for spinal cord regeneration, remain paralyzed after injury. Scientists identified hoxb5a as the gene that determines whether a fish recovers

Researchers at the Morgridge Institute tested spinal cord recovery in 254 adult zebrafish with complete transection, reporting their findings in a preprint published September 14. In 93 fish (28%), swimming never returned over six weeks, even though tissue at the injury site bridged in every animal. The difference lay in whether nerve fibers regrew far enough to reach their original targets. Single-cell profiling linked the outcome to the transcriptional program activated at the lesion, and knocking out the gene hoxb5a showed that it determines the result.

Zebrafish, like lampreys and salamanders, are among the few vertebrates whose spinal cord regenerates after a complete transection: the tissue bridges, nerve fibers grow through the lesion, and the fish swims again. In mammals, including humans, such a transection is irreversible. Kenneth Poss's lab, the first to demonstrate heart regeneration in zebrafish, came to this work through routine: while preparing a batch of fish for gene-activity profiling, some animals failed to recover swimming after injury. The authors tested this systematically in a large cohort and found that the outcome did not depend on the batch of animals, the sex of the fish, or which surgeon performed the transection.

The authors traced nerve fiber trajectories directly. In recovered fish, axons entered the injury zone almost completely and extended beyond it to reach their former targets. In paralyzed fish, axons penetrated only about a third of the way and barely advanced further, even though the tissue bridge formed equally well in all animals.

To find the cause, the authors sequenced RNA from more than 60,000 spinal cord nuclei. All fish activated the same set of cell types at the injury site. The difference was in which program those cells turned on. Fibroblasts, the cells that remodel connective tissue after injury, activated growth and axon-guidance genes in recovered fish, while in paralyzed fish they produced excess matrix fibers that blocked axon extension. A similar pattern appeared in immune cells: recovered fish had more active T-cells at the lesion site, while paralyzed fish had noticeably fewer.

Gene regulatory network reconstruction, mapping which genes control these programs, pointed to hoxb5a as the most connected Hox gene in fibroblasts of recovered fish. The Hox family patterns the body axis during embryonic development, and in adult cells it persists as a memory of that original patterning. The authors generated a stable line of hoxb5a knockout mutants: the fraction of paralyzed fish rose nearly fivefold, to 63% compared with 13% at week six, and the fraction of axons extending beyond the lesion dropped from 70% to 14%. The authors propose that the gene acts not on the nerve itself but remodels the matrix and signaling environment around fibroblasts, through which axons must grow.

Of seven candidate genes, only hoxb5a produced a consistent effect; the other six showed no phenotype. The authors attribute this to the frequent functional redundancy within the Hox family: loss of one gene is masked by another filling the same role. The question remains open, and it is possible that the outcome depends not on a single gene but on a combination of them.

Even in a species with one of the highest known capacities for spinal cord regeneration, the outcome hinges on a specific gene that can be identified and potentially modified. In mice with the same complete transection, partial movement has been restored by a different approach: microrobots made of nerve cells and magnetic nanoparticles used a magnetic field to deliver cells to the injury and trigger impulses externally, because the mouse spinal cord cannot regenerate on its own. In zebrafish, the same goal is closer at hand: the hoxb5a switch already exists inside the tissue.