In lab, human radial glia take cues from glucose and thalamus

At UCLA, Claudia Nguyen and her colleagues fused stem cell-derived thalamic and cortical organoids—lab-grown brain models—into assembloids, then examined how the tissues communicated. In those models, thalamic projections made direct physical contact with radial glia, the stem cells that build much of the developing cortex. The contact prompted the cells to produce more excitatory neurons, especially the upper-layer neurons most expanded in the human brain. A parallel study from Aparna Bhaduri's and Heather Christofk's labs found a second instruction system: metabolism.

Before birth, radial glia make billions of decisions about which brain cells to create, and when. They generate many of the neurons and support cells in the cerebral cortex, the region associated with thought, memory and language. Their importance also makes them a point of interest in neurodevelopmental and neuropsychiatric disorders, as well as cancer, because cells resembling radial glia can reemerge in brain tumors after largely disappearing before birth.

The study published in Cell, led by co-first authors Jessenya Mil and Jose Soto, used donated human tissue and stem cell-derived brain organoids to map metabolism in the developing cortex. The researchers found that radial glia rely heavily on the pentose phosphate pathway, a process that uses glucose to generate building blocks for rapidly dividing cells. When glucose availability was reduced or the pathway was disrupted, the stem cells changed their output, producing more inhibitory neurons and other later-arising cell types. The result places metabolism inside the decision-making process, rather than treating it as background activity.

The Science study, led by Claudia Nguyen, addressed why thalamic projections reach the cortex before they form their final connections with specific neurons. Human assembloids showed that the early projections connect directly to radial glia. The team traced that contact to NRXN1, a gene known for helping neurons build connections; mutations in the gene have previously been linked to autism spectrum disorder. In assembloids made from patient-derived cells carrying an NRXN1 mutation, thalamic signals behaved differently and shifted the balance between radial glia and the neurons they produced. The team said the physical contact very likely does not exist in rodents.

Concretely, the work gives researchers two new levers for testing how the human cortex develops: changing nutrient-processing pathways and altering early signals from another brain region. The metabolic atlas could help investigate how maternal nutrition and metabolic disorders influence development, while the assembloids offer a way to examine how disrupted signaling may affect the cortex. The evidence remains laboratory evidence—from donated tissue, organoids and assembloids—not a clinical intervention or a result observed in a living patient.