An innovative approach to analysing the functions and gene-expression profiles of neural stem cells in developing human and mouse brains sheds light on the differences — and similarities — between the two species. See Letter p.264
The mammalian cerebral neocortex is a thin mantle of neurons on the surface of the brain's hemispheres that is largely responsible for higher-order mental functions. In different mammalian species, the cortex can vary by more than 10,000-fold in mass1, and such variations have been associated with species-specific differences in cognition and behaviour1,2,3,4. Alterations in the proliferation and differentiation of radial glia — non-neuronal progenitor cells5,6,7 that give rise to most of the approximately 16 billion neurons in the human cerebral cortex1 — are thought to play a crucial part in determining the size of the neocortex. Unfortunately, identifying the genes that regulate development of the radial glia in humans has proved exceedingly difficult. On page 264 of this issue, Lui et al.8 use a sophisticated approach to identify and characterize the role of a secreted protein, PDGFD, in mediating the proliferation of radial glia in humans.
Radial glia serve both as progenitors and as migratory guides for the daughter cells they give rise to, and so provide for the production and placement of appropriate cell numbers and types in the developing neocortex5,6,7,9. Changes to division of radial glia, the length of their cell cycle or the rate at which they proliferate may therefore have profound effects on the size, composition and functional repertoire of the neocortex2,3,4,5,6,7. These cells can be efficiently studied experimentally in mice and many other model organisms, but ethical and practical barriers hinder similar tests in humans. As such, genetic studies have identified many genes that both control proliferation in mice and regulate the size of the human neocortex2,10,11, but the identities of genes involved in the development of radial glia only in humans or in non-standard laboratory animals have remained elusive.
A complete list of the genes expressed within a sample or cell type, together with information about the abundance of transcripts produced from each of these genes, is called a transcriptome. When a transcriptome is generated from a complex biological tissue such as the neocortex, it is an amalgamation of the transcriptomes of the millions or billions of cells that make up that tissue. However, cells of the same type and developmental stage tend to have more-similar transcriptomes than those of different cell types and, furthermore, cells of a given type are generally not distributed evenly throughout a tissue. Thus, it stands to reason that the expression levels of all of the genes expressed by a given cell type will rise and fall in unison as the abundance of that cell type varies in different regions of the tissue under analysis. This provides the basis for identifying the transcriptome of a single cell type, even when starting with a mixed population of cells12,13.
Following this reasoning, the authors generated transcriptomes from 87 cross-sections of tissue taken from a single human fetal neocortex. From these data, they identified 55 groups of genes, called modules, in which gene expression rose and fell together, reflecting differences in the cellular composition of each cross-section. Six of these modules contained genes that are already known to be expressed by radial glia in mice, suggesting that these modules each represented a portion of the transcriptome of human radial glia. Importantly, these six modules contained several genes that the researchers found to be expressed in human, but not murine, radial glia. Lui and colleagues chose to focus on just one of these genes, PDGFD.
This gene encodes a secreted growth factor that has not previously been implicated in neocortical development14. The authors report that inhibiting PDGFD protein signalling in slices taken from the human neocortex reduced the proliferation of radial glia. Conversely, introducing abnormally high levels of either PDGFD or a permanently activated form of its receptor protein, PDGFRβ, into the developing mouse neocortex resulted in an increase in the proportion of radial glia undergoing cell division. Combined, these results suggest that, in normal conditions, PDGFD–PDGFRβ signalling drives the proliferation of radial glia in humans, but not in mice (Fig. 1).
Nearly all murine radial glia reside in an area of the developing brain called the ventricular zone. In some species, however, additional progenitor populations take up residence adjacent to the ventricular zone, in another proliferative region called the subventricular zone (SVZ)3,4,5,6,7. These cell types, including outer radial glia, have been implicated in the expansion of the neocortex of humans and other species3,4,7. Lui and co-workers found that, in mice expressing the activated form of PDGFRβ, cells resembling radial glia were distributed not only throughout the ventricular zone, but also throughout the SVZ. PDGFD-induced signalling may thus help to explain the establishment and proliferation of SVZ progenitors and, therefore, the expansion of the neocortex that has occurred in some mammals.
What particular type of progenitor cell Lui et al. observed in the SVZ remains unclear. Although these cells expressed a molecular marker for radial glia, they were not clearly identified as any one specific progenitor-cell population. Whether the effects of PDGFD–PDGFRβ signalling extend to all types of neural progenitor is also yet to be investigated. Furthermore, because the authors compared radial glia only between human and mouse, we do not know whether the role of PDGFD is unique to humans or common to the development of radial glia in other species with an expanded neocortex. Finally, possible associations between mutations in genes encoding members of the PDGF pathway and human neurodevelopmental abnormalities, in particular microcephaly (an abnormally small head circumference)2,10,11, should now be pursued.
Perhaps most intriguing is the relatively limited number of genes whose expression seems to differ between human and murine developing radial glia. The authors developed a method to identify genes expressed in human radial glia and, as they had hoped, subsequently found a molecular pathway not shared by humans and mice. But the majority of the gene-expression data collected in this study emphasizes the remarkable commonalities between two transcriptional programs that diverged roughly 75 million years ago15.