Recent advances in understanding neocortical development

Progenitor cell diversity

Work in the last decade has identified a diversity of NSPC types and their lineage relationships. Following the seminal discovery of apical radial glial cells (aRGCs) as the primary type of cortical progenitor cell^1,2, intermediate progenitor cells (IPCs) were identified^3–5. IPCs are transit-amplifying progenitors expressing the transcription factor Tbr2 and producing the majority of excitatory neurons in the mouse and rat neocortex^3,4,6,7. Most IPCs cluster and undergo mitosis in a layer basal from the ventricular zone (VZ), the subventricular zone (SVZ), although a subset reside in the VZ and undergo apical mitosis, named short neural precursors⁸. Further studies in primates led to the discovery of an expanded and specialized SVZ, subdivided in inner and outer domains (ISVZ and OSVZ, respectively)⁹. These layers were later identified in other species with an expanded neocortex, like ferret, cat, and sheep¹⁰ and even in New World monkeys and the Amazonian rodent agouti^11,12. Progenitor cell lineage tracings in human, macaque, ferret, and mouse led to the discovery of additional types of basal progenitor cells in the developing neocortex, which are bound by complex lineage relationships. These include various types of basal radial glial cells (bRGCs)^{10,11,13–15} and subapical progenitor cells¹⁶. The relative abundance and proliferative capacity of these progenitor cell types are much greater in species with a neocortex that is large and folded (carnivores and primates) than in those where it is small and smooth (mouse and rat). This has been proposed to contribute to the evolutionary increase in complexity, or complexification, of the neocortex^16–18.

Pioneer transcriptomic studies of the developing cerebral cortex using bulk tissue, identified differences in gene expression between germinal layers, cortical areas, developmental stages, and mammalian species^15,19–23. This illustrated the profound diversity of transcriptional landscapes in cortical development at all levels, including folding and across phylogeny, and set the conceptual foundations for the next technological leap: transcriptomics of single cells. The advent of single-cell transcriptomic analyses has revolutionized our approach to studying cortical development by providing a global and unbiased picture of cell diversity with unprecedented resolution^24–26. This technology has enabled identification of multiple sets of transcriptionally distinct progenitor cell classes in the cortical primordium, generating excitatory neurons^27–29, and in the basal ganglia, generating inhibitory interneurons^30,31.

Single-cell analyses are also beginning to shed light on long-standing hypotheses about the heterogeneity of cortical progenitor cells and the dynamics of their lineage and fate potential during development^32–36. Early experimental studies of these fundamental questions indicated that the fate potential of cortical progenitors is temporally restricted, such that early progenitors can produce neurons for all cortical layers but late progenitors can produce neurons for superficial layers only^34,37. Such late fate restriction would be cell-autonomous as late cortical progenitors continued producing superficial-layer neurons even when transplanted into the new cellular environment of a young host cortex^34,37. The identification of a subset of RGCs that expressed Cux2 and that were fate-restricted to produce upper-layer neurons further supported this model³⁵. However, such fate-restricted progenitors have not been identified in single-cell transcriptomic studies^28,33,38. Rather, some of these studies support the existence of epigenetically regulated temporal molecular birthmarks in RGCs, which act in their daughter neurons as seeds for neuronal diversity. It is proposed that these conserved differentiation programs may then be integrated with environmental (non-cell-autonomous) cues to ultimately define the identity of the neuronal progeny^28,39,40. Nevertheless, controversial points of view on these and related issues remain because of differences in single-cell data processing, analysis, and interpretation^30,31,41.

Modes of neurogenesis and influence on cortex size

Cortical excitatory neurons may be generated directly from the primary progenitor cells (aRGCs) or indirectly via secondary basal progenitors such as IPCs and bRGCs. Indirect neurogenesis is considered a milestone of mammalian cortical evolution, enabling a phenomenal increase in the numbers of neurons produced, particularly those destined to superficial layers, and leading to cortical expansion^42,43. Recent studies reveal that indirect neurogenesis is the major mode of producing deep as well as superficial cortical layers in mouse^40,44 and that it also exists in non-mammals, albeit at lower levels and mostly in birds^44,45. At the molecular level, regulation of the balance between direct and indirect neurogenesis is phylogenetically conserved across amniotes, from snakes to birds and mammals, including humans, where high Robo1/2 signaling promotes expression of the Notch ligands Jag1/2 while repressing the canonical ligand Dll1⁴⁴. Blockade of this signaling system drives indirect neurogenesis, increasing neuron numbers and cortex size. Intriguingly, indirect neurogenesis also correlates with hyperpolarization of apical progenitor cells, which represses Wnt signaling⁴⁰. Changes in the mode of neurogenesis are responsible for the reduced cortex size in developmental brain disease, like microcephaly induced by Zika virus infection. In this case, activation of the unfolded protein response drives direct neurogenesis, leading to premature and limited neuron production and to a small cortex^46–48.

Regulation of progenitor cell proliferation

Cerebral cortex size depends on the mode of neurogenesis and also on the proliferative activity of cortical progenitor cells. Multiple mechanisms regulating progenitor cell proliferation have recently been uncovered. Regulation of gene transcription by epigenetic mechanisms has emerged as a key factor where histone deacetylases and methyltransferases regulate the generation and position of IPCs, neuron migration, and cortical lamination^49,50. Similarly, regulation of chromatin accessibility and other mechanisms related to non-coding genomic regions critically determines levels and patterns of gene expression in the developing cortex, defining neuron production, cortex size, and area identity^51,52.

One of the most exciting findings has been the identification of mRNA species transported within the long basal process of aRGCs and locally translated at their basal endfeet, next to the pial membranes^53–55. This includes proteins regulating the cell cycle, like Ccnd2, and lengthening of G₁, S, and M phases of the cell cycle, which depletes progenitor cells causing premature neurogenesis or apoptosis^56–58.

IPCs are usually depicted as round cells extending few short processes^3,59, and bRGCs are usually depicted as having a single smooth and unbranched basal process^10,60–62. Recent observations demonstrate that the degree of process branching and elaboration in both IPCs and bRGCs is greater in ferret and human (which have a large and folded cortex) than in mouse (which has a small and smooth cortex) and this is linked to a greater proliferation rate of these cortical progenitor cells^63,64. Basal progenitor process growth and proliferation are related to the membrane-bound protein PALMD, enabling these cells to receive pro-proliferative signals related to integrin function^60,64,65. Indeed, modulation of cortical progenitor cell proliferation by cell-extrinsic signals has now been demonstrated to occur from multiple sources, including growing axons and migrating neurons^63,66,67, extracellular neurotransmitters and ions^68,69, and Notch, Wnt, Fgf, and Shh signals⁷⁰.

Once IPCs and neurons are born from aRGCs, they must first delaminate and migrate away from the VZ toward the SVZ. This critical process is regulated by several signaling mechanisms, including Robo⁷¹, YAP⁷², Insm1 via the apical adherens protein Plekha7⁷³, and the centrosome-associated protein Akna⁷⁴.

Delamination

New advances in understanding cortical development have identified the microtubule-associated protein Lszts as a master regulator of cellular dynamics, promoting the delamination of neurons and bRGCs born from aRGCs by altering apical junction organization⁷⁵.

Multipolar phase transition

Once newborn neurons delaminate from the VZ, they enter the SVZ and undergo a transition phase displaying multipolar morphology⁷⁶. During this process, NeuroD1 expressed by multipolar cells represses Prdm16, which regulates mitochondrial reactive oxygen species, and this signaling axis is crucial for the regulation of the multipolar phase migration⁷⁷. Dbnl, a protein interacting with F-actin, is another key regulator of neuronal multipolar morphology, polarity, and migration by regulating the levels of N-cadherin⁷⁸.

Bipolar locomotion

Cortical neurons resume radial migration by re-acquiring polarity and extending a single leading process as they exit the SVZ. This leading process establishes intimate adhesive interactions with the basal process of RGCs for guidance in their migratory displacement. The long-held concept that the leading process of these radially migrating neurons is single and unbranched^79,80 has been challenged by new observations, demonstrating frequent branching and thus more complex migratory behaviors than previously reported⁸¹. In fact, leading process branching is much more frequent in the developing cortex of ferret than mouse and is related to the tangential displacement of radially migrating neurons. Thus, this seems directly related to the maintenance of the radial organization of the cerebral cortex during the tangential expansion and folding of the neocortex^81,82. However, an excessive tangential displacement of cortical neurons is deleterious, and the horizontal tiling of the developing cerebral cortex, or regular distribution of its radial units, must be actively maintained. The microtubule stability regulator protein Memo1 plays key functions in the maintenance of RGC structure and cortical tiling by repressing the hyperbranching of the basal process of RGCs and the excessive dispersion of radially migrating neurons⁸³. On the other hand, it was recently shown that the chromatin-modifying enzyme Prdm16 is also necessary for transcriptional silencing in RGCs and to promote the migration of late-born cortical neurons and cortical layering⁴⁹.