Hidden progenitor in the newborn ferret cerebellum may supply granule neurons

A newly described cell type in the developing cerebellar cortex of newborn ferrets may act as a second source of the brain’s most abundant neurons, according to a perspective article published in Frontiers in Neuroscience. The cells, which the author calls internal granular progenitors, or IGPs, sit in the internal granular layer of the cerebellar cortex, express stem-cell markers, and display features of immature Bergmann glia, a specialized support cell of the cerebellum. The proposal, based on the author’s earlier experiments in ferret neonates, suggests that granule neurons can arise from more than one origin during cerebellar development.

The cerebellum is an ancient brain structure conserved across mammalian species. Classically, it regulates muscle tone, limb coordination, and equilibrium, but it also contributes to non-motor functions such as language, cognition, and emotion. These non-motor functions are mediated largely by the cerebellar hemispheres, sometimes called the neocerebellum, which show gross anatomical and functional asymmetry and maintain lateralized connections with other brain regions. Despite decades of work on cerebellar circuitry, the histological and developmental differences between the regions responsible for motor and non-motor functions remain unclear, and that gap motivated the new synthesis.

Structurally, the cerebellar cortex is organized into three layers: the molecular layer, the Purkinje cell layer, and the internal granular layer. Granule neurons packed densely in the internal granular layer receive inputs from the spinal cord, brainstem precerebellar nuclei, and vestibular organs through mossy fibers, and relay excitatory signals to Purkinje cells via parallel fibers. That excitatory transmission is often amplified by unipolar brush cells and dampened by inhibitory interneurons, including basket and stellate cells of the molecular layer and Golgi neurons. Purkinje cells, in turn, inhibit deep cerebellar neurons, which send excitatory outputs to targets such as the thalamus and red nucleus.

During embryonic development, cerebellar neurons come from two distinct germinal zones. One is the ventricular zone in the roof of the fourth ventricle, marked by expression of the transcription factor Ptf1a, which generates Purkinje cells and multiple classes of inhibitory interneurons. The other is the upper rhombic lip, which produces excitatory neurons, including deep cerebellar neurons and unipolar brush cells, and also gives rise to external granular precursors. These precursors cover the developing cerebellar surface to form the external granular layer, where they proliferate, differentiate into granule neurons, and migrate inward along Bergmann glial processes before settling in the internal granular layer. This two-zone scheme has long framed how researchers think about cerebellar neurogenesis.

The new candidate progenitor emerged from work in ferrets, carnivores that have become increasingly popular in neurodevelopmental research because of their folded, gyrencephalic cerebral cortex. In the developing cerebellar cortex of ferret neonates, the author identified cells expressing proliferating cell nuclear antigen, a marker of dividing cells, together with BrdU labeling after a two-hour pulse, located in the internal granular layer adjacent to the Purkinje cell layer. Notably, these cells were also immunopositive for Nestin, an intermediate filament protein selectively expressed in neural stem cells. The author designated them internal granular progenitors.

Further immunostaining sharpened the picture. IGPs were positive for Sox2, a transcription factor characteristic of neural progenitors and a persistent feature of Bergmann glia, and they extended GFAP-immunopositive processes perpendicularly through the Purkinje cell layer into the molecular layer. They also expressed the glial markers BLBP and S100β. Together, these morphological and immunohistochemical traits give the cells an immature Bergmann glia-like appearance. Cerebellar histogenesis in ferrets, which builds the three-layered cortex, occurs during early postnatal life and follows a trajectory similar to that of rodents; IGPs were detected on postnatal day 7 in ferrets, a stage corresponding to postnatal day 4 in mice.

The key evidence for a neuronal fate came from lineage tracking. BrdU labeling revealed that IGPs transitioned into granule neurons, Bergmann glia, and/or astrocytes through differentiative division in normal ferrets, with Nestin immunostaining attenuating by postnatal day 20. The author therefore proposes that IGPs function as another source of granule neurons distinct from the external granular precursors. Supporting this possibility, a separate mouse study reported expression of the transcription factor Meis1 in cells with Bergmann glia-like morphology around the Purkinje cell layer in neonates; Meis1 promotes cell cycle exit, granule neuron differentiation, migration toward the internal granular layer, and the formation of parallel fibers through activation of Pax6.

The relationship of IGPs to other recently described progenitors remains unresolved. Bergmann glia-like progenitors transiently appear in the cerebellum of mouse neonates and share the Bergmann glia-like morphology, with the potential to differentiate into molecular layer interneurons, Bergmann glia, and astrocytes, but their ability to generate granule neurons is unknown. Nestin-expressing progenitors, also found in mouse neonatal cerebellum, primarily produce astrocytes and interneurons but switch to generating granule neurons through Hedgehog signaling after cerebellar injury. A rare Sox2-positive progenitor population in the outer external granular layer can also differentiate into granule neurons after injury, yet IGPs are distinguished by their Bergmann glia-like morphology, their cortical location, and their activity under physiological conditions. Whether IGPs are identical to any of these populations is unclear, and the author notes that granule neuron differentiation is enhanced in IGPs by inhibiting histone deacetylase 1 and 2 with valproic acid, whereas in Nestin-expressing progenitors it is driven by injury-activated Hedgehog signaling.

The evolutionary angle is central to the author’s hypothesis. Cerebellar expansion across mammals is linked to the complexity of lobulation and to the expansion of the cerebral cortex, and volume lateralization and lobulation are more pronounced in the order of mice, ferrets, and macaque monkeys. Granule neurons are the most numerous neurons in the brain, estimated at roughly 300 million in rodents and 45 billion in humans, so cerebellar expansion may have depended primarily on increasing granule neuron numbers. In ferret neonates, valproic acid stimulates the differentiative division of IGPs toward granule neurons while reducing the proliferation of external granular precursors, suggesting epigenetic regulation of the balance between granule neurons arising from the two origins. The author speculates that granule neurogenesis from IGPs may contribute to evolutionary expansion, structural and functional asymmetry, and lobulation complexity relevant to the acquisition of non-motor cerebellar functions, while emphasizing that the existence and number of IGPs in mice, monkeys, and humans are unknown and must be addressed in future studies.

The framework also touches on neurodevelopmental disease. Neonatal exposure to valproic acid has been reported to induce autism spectrum disorder-like social behavioral deficits in ferrets, and the author argues that excess production of IGP-derived granule neurons could disrupt the balance between excitatory mossy fiber input and inhibitory interneuron modulation that regulates Purkinje cell activity, representing developmental alterations relevant to such disorders. Because IGPs may be a subpopulation of immature Bergmann glia, expressing Sox2 and extending GFAP-positive radial processes through the molecular layer to the external granular layer despite their Nestin expression, their lineage may also connect to adult cerebellar neural progenitors, since a Sox2-expressing Bergmann glial lineage lacking mature glial markers has been proposed to serve as adult neural progenitors. The author concludes that further research on IGPs could illuminate cerebellar cortical histogenesis, the differences between motor and non-motor cerebellar regions, and how abnormal IGP development relates to neurodevelopmental conditions, while cautioning that these ideas remain hypotheses to be tested in species beyond the ferret.

Subject of Research: Internal granular progenitors in the developing ferret cerebellar cortex as a proposed second source of granule neurons

Article Title: A progenitor with immature Bergmann glia-like characteristics in the developing cerebellar cortex

Article References: Sawada, K. (2026). A progenitor with immature Bergmann glia-like characteristics in the developing cerebellar cortex. Frontiers in Neuroscience, 20, 1963374. https://doi.org/10.3389/fnins.2026.1963374

Image Credits: AI Generated

DOI: 10.3389/fnins.2026.1963374

Keywords: cerebellum, internal granular progenitors, Bergmann glia, granule neurons, ferret, Nestin, Sox2, neurodevelopment, valproic acid, autism spectrum disorder, cerebellar histogenesis, evolution