Deleting Meis2 in already-born neurons shifts mouse prefrontal cortex toward a motor identity

Most accounts of how the cortex gets carved into areas put the decision early, in dividing progenitors, under opposing gradients of transcription factors and morphogen signalling centres. Ingrowing thalamocortical axons arrive later to instruct them ^[8]. The authors say the same framework has explained sensorimotor areas far better than it explains the sensorimotor-to-association axis ^[8]. Their manipulation came after the cells' final division, in excitatory neurons that were already born and already in place ^[3].

The loop they propose is short and it closes on itself: RA to MEIS2 to ALDH1A3 to RA ^[5]. ALDH1A3 is the retinoic acid synthesising enzyme. When Meis2 was deleted, the excitatory neurons expressing it were substantially reduced in the developing medial prefrontal cortex, along with RA signalling itself ^[4]. So the transcription factor sustains the cells that make the signal that keeps the transcription factor on. Of the three arrows in the loop, the deletion experiment bears directly on two ^[1]; the RA to MEIS2 arrow rests on the human regulatory network and on earlier RA work ^[1][6].

The human contribution here is expression-based. The group had previously found broad gene expression gradients across fetal frontal and temporal cortex marking out prospective association regions including the PFC, most prominent at mid-fetal stages ^[9]. RA-associated genes sat inside that prefrontal gradient, enriched relative to motor cortex ^[10]. The human tissue gives a spatial correlation. The causal test was in mouse ^[3].

The abstract calls the respecification partial ^[3]. Prospective prefrontal territory acquired motor-like molecular and connectivity features, falling short of motor cortex. The abstract puts no number on the reduction in the ALDH1A3-expressing population. The authors themselves stop short on the clinical end. They write that their findings link a classic morphogen to transcriptional identity, circuit formation and function, and "potentially to neuropsychiatric disorders" ^[16].

That caution is worth keeping. MEIS2 carries links to intellectual disability and autism spectrum disorder ^[2], and human variants in components of the RA pathway have been linked to neurodevelopmental disorders and schizophrenia ^[12]. What the loop offers those genetics is something to count: if it holds in humans, reduced MEIS2 dosage should shrink the ALDH1A3-expressing neuron population in medial prefrontal cortex. Nothing in the abstract measures behaviour, and the causal evidence is mouse only. In primates, the PFC expands prominently relative to motor and sensory cortices ^[13].