Drawing edges

Listen to audio

The concepts of compartment and boundary between compartments emerged from work carried out in Drosophila. Boundaries have several functions: they act as obstacles to cell migration, as edges to be followed for axon migration, and as signaling centers. These signaling centers participate in the patterning of embryonic tissues by regulating the secretion of oncogenes and morphogens, sometimes in the same molecular form (Wnt, FGF, etc.).

This segmentation of vertebrates was rediscovered by Lumsden and Keynes in 1989 in the case of the rhombencephalon. I say "rediscovered" because of the pioneering nature of Étienne Geoffroy Saint-Hilaire's proposal in 1822. But to speak of rediscovery is a bit of an exaggeration. If, in both cases, the first observation is anatomical, what won the day in 1989 was the genetic correlate of Hox gene expression and the orthology between fly and vertebrate Hox genes, which made it possible to think of this orthology at the anatomical level too.

It was therefore the correlation between genetics and anatomy, or rather a form of congruence between genetic orthology and anatomical orthology, that in 1989 enabled us to extend to vertebrates a concept of segmentation that had previously been reserved for certain invertebrates, including arthropods.

The forebrain, which gives rise to the telencephalon, diencephalon and mesencephalon, also features temporary swellings and constrictions of the neuroepithelium, suggesting the existence of "compartments". This anatomical configuration, along with the expression pattern of numerous genes, has led to the emergence of a neuromeric conception of the forebrain that designates a form of anteroposterior (AP) compartmentalization. DV compartmentalization is also present.

The prosumeric model proposed by Puelles, Rubenstein and colleagues in the 1990s has had a real influence on the way we think about the development and evolution of the cortex. This model proposed a division of the brain - along the AP axis - into 6 segments or prosomeres (metameres of the prosencephalon): P1 to P3 for the diencephalon and P4-P6 for the hypothalamus and telencephalon. Remarkably, these AP and DV boundaries almost always lie on developmental gene expression dividing lines.

At this point, we introduced the hypothesis born of our laboratory's work that certain transcription factors are themselves morphogens, i.e. molecules capable of transducing a signal, and that edges are formed where they "meet" within the neuroepithelium. This idea was developed later.

In the meantime, we have provided some information on the development of the cortex. First, which neurons make up the cortex? 80% are excitatory glutamatergic pyramidal neurons, generated in the ventricular (subventricular) zone. The remaining 20% are inhibitory GABAergic neurons, which are generated in the ventral, ganglionic, medial, lateral (and caudal) regions of the telencephalon and migrate tangentially to join the dorsal telencephalon.

The telencephalon is formed from the prosencephalon, at the front of this structure, which also forms the diencephalon and mesencephalon at the rear. The telencephalon is formed around E8.5 with the expression of the Foxg1 transcription factor. As soon as Foxg1 is expressed, the dorsal telencephalon divides into sub-regions, starting with an anterior, lateral region that gives rise to the neocortex, and a posterior, medial region that gives rise to the hippocampus, the "cortical hem" and the choroid plexus.

The ventral telencephalon divides into a median region (MGE) and two lateral and caudal regions (LGE and CGE), the ganglionic eminences. These regions give rise to the basal ganglia and associated limbic structures (amygdala and accumbens). MGE also produces GABA neurons with somatostatin or parvalbumin (PV) and NPY interneurons in the basal ganglia and cortex (after migration). The CGE produces calretinin and VIP interneurons. The LGE produces olfactory bulb interneurons plus neurons from inhibitory projections in the striatum and limbic areas.

In this diagram, we have introduced another interaction between Lhx2 and Lhx5. Lhx5 is expressed in the cortical hem and its loss leads to the loss of the choroid plexus. Lhx2 is expressed in the ventricular zone, along a gradient, and is strongly repressed by BMP2/4 (dorsal), which excludes it from the cortical hem. Loss of Lhx2 leads to expansion of the hem and plexus at the expense of the neocortex. This is yet another example of the antagonisms that increasingly appear to represent a general law.