Much has been written about the role of the floor plate in neural tube patterning, but the function of its dorsal counterpart, the roof plate, has remained relatively obscure. However, after the finding that it is required for the specification of dorsal spinal cord neurons, the roof plate is finally beginning to emerge from the shadows to establish itself as an important signalling centre. Now, Monuki et al. show that it also controls the growth of the cerebral cortex, and might even be a source of cortical neurons.

Monuki et al. ablated the roof plate by expressing the diphtheria toxin A chain under the control of the roof-plate-specific Gdf7 promoter. This ablation had profound effects on the development of the dorsal forebrain — the neural tube failed to close and the amount of cortical tissue was markedly reduced. It also affected the expression of Lhx2, a LIM homeodomain transcription factor that is regulated by roof plate signals in the spinal cord and has previously been implicated in forebrain patterning. Normally, Lhx2 is expressed in a gradient in the dorsal telencephalon. Following the line of the epithelium starting from the roof plate, there is an Lhx2-negative zone around the midline, then the expression rises abruptly to its highest level in the prospective hippocampus and fades out towards the lateral edge of the cortical neuroepithelium. After roof plate ablation, what remained of the cortex expressed Lhx2 at a reduced level, and the expression was uniform rather than graded.

Which roof plate signals regulate the expression of Lhx2? In the spinal cord, bone morphogenetic proteins (Bmps) are important for roof plate signalling, so Monuki et al. looked at the effects of implanting Bmp2- or Bmp4-soaked beads into embryonic forebrain explants. Lhx2 expression was suppressed in the vicinity of the bead, but at a distance its expression was strongly enhanced, returning to endogenous levels further away from the bead. This neatly mirrored the endogenous gradient of expression relative to the roof plate, indicating that Bmps could indeed be responsible for generating this gradient.

The authors also showed that targeted disruption of Lhx2 leads to loss of cortical neuroepithelium and overgrowth of the roof-plate-derived choroid plexus and cortical hem. This indicates a dual role for Lhx2 — restricting the growth of midline tissues and promoting cortical progenitor cell fate. Intriguingly, although there were virtually no cortical plate neurons in the Lhx2 mutant brain, preplate neurons were generated, indicating that these might be derived from roof plate tissue, a theory that the authors corroborated by fate-mapping of Gdf7-expressing cells.

So, this study shows that the roof plate not only controls the expansion of the cerebral cortex through the bimodal regulation of Lhx2 expression, but might also make a cellular contribution. This adds significantly to the roof plate's expanding functional repertoire, and we eagerly await further insights into its role.