Evolution of The Neocortex: A Perspective From Developmental Biology

Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

REVIEWS

Evolution of the neocortex:


a perspective from
developmental biology
Pasko Rakic
Abstract | The enlargement and species-specific elaboration of the cerebral
neocortex during evolution holds the secret to the mental abilities of humans;
however, the genetic origin and cellular mechanisms that generated the distinct
evolutionary advancements are not well understood. This article describes how
novelties that make us human may have been introduced during evolution, based on
findings in the embryonic cerebral cortex in different mammalian species. The data on the
differences in gene expression, new molecular pathways and novel cellular interactions that
have led to these evolutionary advances may also provide insight into the pathogenesis and
therapies for human-specific neuropsychiatric disorders.

The neocortex, as the name implies, is the newest addi- the developing human brain have revealed quantitative
tion to our brain and is considered to be the crowning and qualitative differences in the gene composition and
achievement of evolution and the biological substrate expression between the neocortex of rodents and primates,
of human mental prowess. Although its origin can and we have only begun to explore the developmental and
be traced to reptiles1,2 in the Carboniferous Period, it evolutionary origin of these differences.
first appears as a uniform, six-layered sheet consisting As the excavated skulls of our common predecessors
of radially deployed neurons in small mammals that are empty, perhaps the only way to understand cortical
emerged during the transition from the Triassic period evolution at the cellular level is by comparison of the
to the Jurassic period. Increases in size and complexity of differential gene expression and developmental events
the cerebral cortex have culminated in the modern that occur during the embryonic development of living
human that separated from rodents between 90 and 100 mammals7. This approach, known as evo–devo8 (BOX 1)
million years ago and from the Old World monkeys, such reflects the realization that developmental neurobiol-
as macaques, 25 million years before the present time3,4 ogy can explain how the genetic information contained
(FIG.1). Given that the cerebral neocortex is the centre of within progenitor cells regulates the number, phenotype,
extraordinary human cognitive abilities, it is surprising migration and allocation of neurons in the developing
how little has been done to research its emergence. It brain, where they establish species-specific circuits.
seems that we are sometimes so seduced by similarities For obvious ethical and logistic reasons, experimen-
between species that we neglect the differences. tal manipulation is not possible in humans and thus,
There are several possible explanations for this ten- comparative analysis of the developmental events in
dency. First, the mouse is the best and most economical other mammalian species is used to reveal the origin of
experimental model system for studying the cerebral evolutionary novelties in the cerebral cortex.
cortex 5. Second, the concept, aptly advocated by Charles In this Review, I provide selective examples of what
Darwin, that the biological world is unified, and that evolutionary concepts can be deduced from evo–devo
Department of Neurobiology “there is no fundamental difference between man and studies that have been carried out in embryonic brains
and Kavli Institute for the higher mammals in their mental faculties”6 is still of rodents (mostly Mus musculus), non-human pri-
Neuroscience, Yale University engrained in our mind and we hope that human traits mates (mostly Macaca mulatta) and humans (Homo
School of Medicine, can be deduced from this commonality. Third, research sapiens). Although there is a large body of data on
New Haven, Connecticut
06510, USA.
on the development of the human brain has been mainly other mammalian species, I focus on only these three
e-mail: [email protected] descriptive. However, the recent application of the most species because of space limitations. These three pres-
doi:10.1038/nrn2719 advanced methods of molecular and cell biology to study ently living species effectively illustrate different levels

724 | O CTObeR 2009 | VOlume 10 www.nature.com/reviews/neuro

© 2009 Macmillan Publishers Limited. All rights reserved


f o c u S o n c n S E V o lo
REVu ItEI W
onS

a Mouse b Macaque c Human share extrinsic connectivity 9–10. However, this apparent
cytoarchitectural uniformity exhibits different degrees
of variability in each area depending on their function.
This compartmentalization is particularly pronounced
in the large and convoluted human neocortex, with more
than 50 distinct cytoarchitectonic areas11.
Second, a remarkable feature of cortical development
in all mammalian species is that none of its constitu-
ent neurons is generated within the cortex itself. Rather,
they are generated in the proliferative transient embryonic
zones (such as the ventricular zone (VZ) and subven-
tricular zone (SVZ)), which are situated near the surface
of the cerebral lateral ventricles, and acquire their proper
areal and laminar positions through long-distance radial
and tangential migration across the intermediate zone
(reviewed in REFS 12–17). Neurons migrating radially
to the increasingly distant cortex follow the shafts of a
transient population of radial glial cells18–20. A detailed
description of the complex migratory process and its
underlying complex and multiple molecular mechanisms
cannot be described in this short Review. FIG. 2 conveys
the basic concepts and further details can be viewed
d Mus musculus in the Rakic laboratory animated video of a migrating
Macaca mulatta cortical neuron and animated video of radial migration.
Third, not only do the postmitotic neurons migrate
Homo sapiens
to the overlaying cortex, but also each generation
bypasses the previous one, a phenomenon known as the
–100 mya –25 mya 0 mya
‘inside-out gradient of neurogenesis’ (REFS 21,22) (see the
Figure 1 | Broad comparison of the brain of a mouse, a macaque monkey and a
animated video of radial migration). The sole excep-
human, and the phylogenetic divergence of these species. In the top panels of a, b
and c the cerebral hemispheres of a mouse, a macaque monkey Nature Reviews
and | Neuroscience
a human brain,
tion to the inside-out sequence of neurogenesis occurs
respectively, are drawn to approximately the same scale to convey the overall difference in the layer I of primates, in which, contrary to rodents,
that exits in the size and elaboration of the cerebral cortex in these three species. The neurogenesis continues throughout the entire period
prefrontal cortex, which has no counterpart in mouse, is shaded in blue in the macaque of corticogenesis23. The biological significance of the
and the human. In the bottom panels of a, b and c, diagrams of cerebral sections are inside-out sequence of settling of neurons in the cerebral
shown for these same species to illustrate that there has been a relative small increase in cortex is not clear 24, but if disturbed, by either genetic
the thickness of the cortex (dark purple outline) compared with the large increase in or environmental factors, neurons display abnormal
surface area (1: 100: 1000 X in mouse, macaque monkey and human, respectively). In part cortical function25. Only recently have methodological
d, the timescale of phylogenetic divergence of mouse (Mus musculus), macaque monkey advances given us an opportunity to study mechanisms
(Macaca mulatta) and human (Homo sapiens) is also shown based on the available DNA
involved in neuronal migration at the molecular and cel-
sequencing data (reviewed in REFS 3,4). mya, million years ago.
lular level (reviewed in REFS 14,26–28). In most mam-
mals, including humans, corticogenesis is completed
of neocortical expansion and elaboration (FIG. 1a,b,c), before or around the time of birth29.
and this is often taken, implicitly and erroneously, as an
indication of the lineage continuity between them with Human ascent through cortical expansion
Transient embryonic layers
the human species at the top; however, comparative ana- The first step in the evolutionary ascent of the human
Layers identified in the
embryonic brain, such as the tomical and DNA sequencing data clearly indicate that cerebral cortex is its enlargement, which occurs mainly
proliferative ventricular and these three species belong to different branches on the by expansion of the surface area without a compara-
subventricular zones (VZ/SVZ) phylogenetic tree (FIG. 1d). ble increase in its thickness (FIG. 1). The modest, only
or migratory intermediate zone about twofold, difference in cortical thickness between
that lack direct counterparts
in the adult brain, as defined
Peculiarities of the cerebral cortex rodents and primates is in part due to the enlargement of
by the Boulder Committee. The evolution of the cerebral neocortex cannot be fully neurons and neuropile in the extracellular space.
appreciated without understanding the way in which it is An explanation of the cellular mechanism that
Neuropile built. There are several features of cortical development underlies the enormous surface expansion of the cortex
The tissue situated between
that differ from the development of other organs of the without a comparable increase in its thickness has been
neuronal cell bodies,
composed of a complex body and even from the rest of the brain. offered by the radial unit hypothesis12 (FIG. 2). According
network of neuronal and glial First, the cerebral cortex in all mammalian spe- to this model, an increase in the number of neural stem
processes including dendrites, cies, including humans, is a cellular sheet composed cells by symmetrical divisions before the onset of neu-
dendritic spines, axonal of projection (or pyramidal) and local circuit neurons rogenesis would result in an exponential increase in
terminals and synapses, used
often to measure
(or interneurons) deployed in horizontal layers, inter- the number of founder cells that give rise to radial cor-
connectedness of a given sected by vertical (or radial) columns that are stere- tical columns (FIG. 2a). At later stages, the number of
structure. otypically interconnected in the vertical dimension and neurons increases linearly within each column, mostly

NATuRe ReVIeWS | NeuroscieNce VOlume 10 | O CTObeR 2009 | 725

© 2009 Macmillan Publishers Limited. All rights reserved


REVIEWS

by asymmetrical divisions of neural progenitors30. For Similarly, an even more dramatic increase in cor-
example, less than 7 extra rounds of cell divisions in tical surface occurs with an experimentally induced
the VZ at the early embryonic stages can account for increase in neuronal production. It is counterintuitive
the 1,000-fold difference in total cortical surface area that such a large increase in neocortical size is achieved
between mice and humans (FIGS 1,2a). The fact that the by a substantial prolongation, rather than reduction
forebrain primordium in humans is much bigger than in in duration of the cell cycle36–37. It is the changes in
mice and monkeys, even at embryonic stages before the pool size of the neural stem cells and the rate of the
onset of cortical neurogenesis19,31,32, supports this model. exit from the cell cycle that have a crucial role. For
Furthermore, this model predicts that neurons gener- example, the β-catenin-mediated signalling pathway
ated from the founder cells will follow adjacent shafts acts on the decision of the neuronal precursor cells to
of radial glial cells and thus, form more columns, which exit the cell cycle38 and, hence, has an influence on the
results in an expanded laminar rather than nuclear number of postmitotic neurons. Thus, transgenic mice
shaped structure. However, the radial unit hypothesis that express a stabilized form of β-catenin transgene in
does not imply homogeneity of radial units across the neural precursors develop an increased number of pre-
cortex within or between species. On the contrary, there cursor (founder) cells in the VZ, which in turn lead to a
is a large body of evidence to support that radial col- larger number of radial columns and consequently the
umns differ across the cerebral hemisphere and change formation of elaborate cerebral convolutions, contrast-
considerably in size and composition during cortical ing the smooth (lissencephalic) hemispheric surface of
evolution (BOX. 2). their wild type counterparts38,39 (FIG. 3b). Despite severe
additional brain abnormalities in these transgenic mice,
Experimental tests of the radial unit hypothesis. The many of the overproduced neurons migrate radially, set-
radial unit model has been tested by either diminish- tle into apparently normal laminar and radial positions
ing the rate of programmed cell death (apoptosis) or by and generate a much larger surface of the neocortex of
increasing the rate of neural stem cell proliferation in relative uniform thickness. by contrast, overexpression
the VZ. mice deficient in caspase 3 and 9, two impor- of cyclin-dependent kinase (CDK) inhibitor p27, which
tant components of programmed cell death (apop- acts on neuronal progenitors at later developmental
tosis) — an essential process during development 33,34 stages, decreases the production of neurons destined for
— have a larger number of neural stem (founder) cells the more superficial layers within the radial columns
and radial glial cells. These cells, in turn, produce more without increasing its surface40.
than the normal number of radial columns and, hence, Although, it is unlikely that mutation in caspase 9,
result in a larger cortical surface that forms incipient β-catenins, or CDK genes were involved in neocortical
convolutions35 (FIG. 3a). evolution, as mice that are deficient for these particu-
lar genes are not viable, these experiments can serve as
an example of the evo–devo approach to draw theories
on how a single gene mutation, that may have occurred
Box 1 | Evo–devo approach to reconstruct neocortical evolution
in our common, long-extinct ancestors, could have
The use of advanced methods to study embryonic brain development in living species increased the number of proliferative founder cells in
provides insights into possible cellular and molecular mechanisms that have enabled the VZ, which triggered a cascade of events that led to
the evolutionary expansion and elaboration of organs in common ancestors. The new an increase in the number of radial units and cortical
subfield of developmental biology called evo–devo, was created to elucidate how new
surface expansion, including formation of convolutions
features may have been created by gene mutation and, if inherited, propagated during
without an increase in cortical thickness.
evolution by natural selection7,8,139. The contemporary evo–devo approach should be
distinguished from Haeckle’s law of ontogenetic recapitulation of phylogeny that arose
in the ninteenth century (for historical reviews see REFS 1,2,7). The evo–devo approach Development of convolutions. It has been recognized as
embraces the concept of modularity or functional cellular units, such as segments, common sense that the increase in the initially smooth
somites, limb buds and cortical columns. Changes in these modules, if inheritable, often cortical surface must have led to its folding (FIG. 1).
result in lethality; but, if useful can enhance the survival and propagation of the species. However, understanding how these convolutions
The present Review is, by necessity, based on selected data on cortical development formed involves rather complicated logistical problems:
in humans, non-human primates and rodents, which are relevant for understanding one of them being that the pattern of convolutions is
neocortical evolution. For example, experimental inhibition or overexpression of highly reproducible within each species with relatively
specific genes in the embryo allows us to test their effect on development of the
small individual variations. To account for this stere-
cerebral cortex and to draw conclusions on their involvement in increasing
otypic pattern, a role of the long cortico-cortical axonal
evolutionary complexity35,38,74. In addition, it is possible to explore the timing of
expression of new genes and their modification directly in the embryonic human tracts in this process has been proposed 41–43. Indeed,
cortex128. Based on the available data we can propose models of how gene expression the formation of the convolutions in the human cer-
within the neural stem cells in the proliferative embryonic zones control the number of ebrum is associated with the formation of the volu-
columnar units, regulate neuronal migration and allocation into proper laminar and minous intermediate zone that eventually transforms
radial locations in the cortical plate, where they promote differentiation into into subcortical white matter, crisscrossed by the large
species-specific phenotypes, establish new patterns of synaptic connections and number of distinct fascicles of cortico–cortical connec-
determine particular neurotransmitters and receptors within the cytoarchitectonic tions. It has been speculated that the tension created by
areas subserving specific functions140. Thus, the evo–devo approach represents a these fascicles is responsible for the stereotyped shape
scientific way to unravel the events that led to the evolution of the neocortex, which
and orientation of the convolutions in gyrencephalic
we will never be able to verify in real time.
brains43.

726 | O CTObeR 2009 | VOlume 10 www.nature.com/reviews/neuro

© 2009 Macmillan Publishers Limited. All rights reserved


f o c u S o n c n S E V o lo
REVu ItEI W
onS

a b
Cortical
II column

Postmitotic
Cortical cell death ( ) CC 11 MZ
layers 1 2 3 4 5 6 7 8 9 10 E100

VI
CP

E40
Asymmetrical SP
division Pro-apoptotic
factors CC
RG
MN
Inhibitory factors TR
IZ

Symmetrical
Anti-apoptotic factors
division
Premitotic VZ
cell death ( ) Neural stem cell pool ( ) NB
Cortical founder pool ( ) MA VZ

1 2 3 4 5 6 7 8 9 10
Figure 2 | radial unit lineage model of cortical neurogenesis. a | Based on the radial unit hypothesis12,30, the model
illustrates how changes in the mode and the rates of cell proliferation and/or programmed cell death within the neural stem
cell pool (blue circles) in the ventricular zone (VZ) that divide symmetrically at early embryonic stages causes
Nature an exponential
Reviews | Neuroscience
increase in the number of radial columns, which, in turn, results in surface expansion of the cerebral cortex without changes in
its thickness. By contrast, similar changes in proliferation kinetics occurring in the founder cells (red circles), which divide
asymmetrically, cause a linear increase in the number of neurons within radial columns without a change in the cortical surface
area. b | The model of radial neuronal migration that underlies columnar organization based on REFS 12,30. The cohorts of
neurons generated in the VZ traverse the intermediate zone (IZ) and subplate zone (SP) containing ‘waiting’ afferents from
several sources (cortico-cortical connections (CC), thalamic radiation (TR), nucleus basalis (NB), monoamine subcortical
centers (MA)) and finally pass through the earlier generated deep layers before settling in at the interface between the cortical
plate (CP) and marginal zone (MZ). The timing of neurogenesis (E40–E100) refers to the embryonic age in the macaque
monkey12,22,32. The positional information of the neurons in the VZ and corresponding protomap within the SP and CP is
preserved during cortical expansion by transient radial glial scaffolding. Further details can be viewed in the Rakic laboratory
animated video of radial migration. RG, radial glia cell; MN, migrating neuron. Part a is modified with permission from REF. 147
 (2005) Macmillan Publishers Ltd. All rights reserved. Part b is reproduced, with permission, from REF. 140  (2007) Elsevier.

Evolutionary adaptation of radial glial cells. The tran- mature glial markers such as the glial fibrillary acidic
sient population of radial glial cells in the developing protein (GFAP) at the onset of corticogenesis, whereas
cerebral cortex serves both as neural stem cells and as a in rodents they are expressed only after birth48,50–55.
guide for the migration of their offspring and adjacent Therefore, even basic cell types such as the radial glial,
neuronal progenitors44–46. However, formation of con- have diversified during evolution to accommodate the
volutions requires long, curvilinear migratory pathways need to develop cerebral convolutions.
and durable radial glial shafts that serve as guides for
neuronal migration to the distant cerebral cortex at later Multiplication and elaboration of cortical maps
fetal stages. Thus, in the primate forebrain, many radial During evolution, functionally distinct cortical areas
glial cells transiently stop dividing while their fibres have expanded at individual rates and new areas have
span the entire cerebral wall, retaining their endfeet at been added, forming an elaborate mosaic56. For example,
the ventricular and pial surfaces47. A single radial glial broca and Wernicke language-related areas or the pre-
fibre in primates may serve as a guide for cohorts of up frontal granular cortex have not been, either anatomi-
to 30 simultaneously migrating neurons, including some cally or functionally, identified in rodents4,10,57 (FIG. 1).
of the interneurons that originate from the expanded Numerous ideas about the developmental mechanisms
SVZ48,49 and from the larger pool of progenitors with that could explain how this elaborate cortical parcella-
short radial processes in the VZ46,50,51. The greater stabil- tion occurred in evolution have been discussed. early
ity and longevity of the radial glial scaffolding may be an debates about whether structural differences between
Cortical parcellation evolutionary adaptation that enables proper allocation of areas are induced in the initially homogeneous popula-
Regionalization of the cerebral
neocortex into areas with
neurons to the expanded and convoluted cerebral cor- tion of cortical neurons by the patterned activity of the
distinct structural and tex32,49. Indeed, the radial glial in both human and non- inputs arriving from the periphery through the thala-
functional attributes. human primates differentiate precociously and express mus (the tabula rasa hypothesis) or, whether the cortical

NATuRe ReVIeWS | NeuroscieNce VOlume 10 | O CTObeR 2009 | 727

© 2009 Macmillan Publishers Limited. All rights reserved


REVIEWS

Patterning center progenitors themselves are targets of evolution (the pro- factors that are involved in neuronal production and
Group of cells in the embryonic tomap hypothesis) have been recently reviewed28,58,59 and phenotypic and areal specification65–69. However, this
brain that secrete molecules are not repeated here. In brief, the consensus that has FGF study shows dramatically that a change in expres-
(morphogens) that initiate emerged is that the species-specific diversity of the neo- sion of a single factor can result in a duplication of an
differential expression of
transcription factors that
cortical areas originates from differential gene expres- entire functional area.
specify formation of the sion in the neural stem cells in the embryonic VZ and
cortical areas. this information is transferred radially through migrat- Decisions in the proliferative centers. Numerous recent
ing postmitotic neurons to the overlying cortical plate. studies have demonstrated that the subdivisions and area
Thus, each area attracts appropriate inputs, rather than enlargements of the cortical areas in embryonic mice
being specified by them. Activity-dependent mecha- could be traced back to the proliferative VZ and SVZ
nisms have a significant role at subsequent stages that (reviewed in REFS 28,58,65). The generation of transgenic
refine the already formed synaptic connections12,60–62. mice in which the number of neurons and radial units is
increased serves as an example of how a larger-than-nor-
Duplication through morphogenetic factors. The evo- mal number of founder cells can create a cortex with an
lution of the cerebral cortex involves both an increase increased surface area and number of neurons that even-
in the number of new cytoarchitectonic areas, as well tually form new connections. For example, the expres-
as duplication of sensory representations63. It is gener- sion of FGF8 in the anterior parts of the telencephalon
ally accepted that the regionalization of the neocortex suppresses the posteriorly expressed chick ovalbumin
is initiated by patterning centres that secrete signalling that normally functions upstream of COuP transcrip-
molecules, such as the fibroblast growth factors (FGFs), tion factor 1 (COuP-TF1) and homeobox (emX2) in
wingless (WNTs) and bone morphogenetic protein the telencephalic neuroepithelial cells65,70–73, and this
(bmP), that regulate the position and size of cortical suppression results in a decrease in the anterior but an
areas (reviewed in REFS 28,64,65). For example, the increase in the posterior expression of these molecules.
mouse forebrain has an anterior patterning centre at The development of frontal and sensory subdivisions
the commissural plate that secretes FGF8 and FGF17. of the cortex is dependent on the graded action of FGF8
Perturbations of the patterning centre and expression (REF. 73) and FGF17 (REFS 74,75) and provides an elegant
of FGF8 can profoundly change the somatosensory example of how one area, such as the frontal cortex, can
map of the mouse cerebral cortex: overexpression of expand differentially and independently of the growth
FGF8 in the embryonic forebrain results in forma- rate of other areas75,76 (FIG. 4c). Furthermore, the size of an
tion of an additional barrel field, with a mirror image area can be regulated by the change in FGF17 expression
topographical representation of mouse whiskers64. The at early embryonic stages independent of the synaptic
cellular mechanism of this remarkable duplication is input from the periphery through the thalamus74 (FIG. 4).
unknown, but it is instructive that FGF8 expression The fact that it is now possible to enlarge and/or
is coordinated by distinct sets of transcription factors, duplicate distinct cytoarchitectonic areas in the neocor-
namely zinc-finger and sonic hedgehog transcription tex by genetic manipulation opens an unprecedented
opportunity to explore how these maps develop in each
individual, as well as how they may have been introduced
Box 2 | Evolutionary elaboration of radial columns during neocortical evolution. As a next step, it is impor-
tant to search for additional genes and morphoregulatory
During evolution, the neocortical surface expands by the addition of radial columns23,30
molecules that may be involved in cortical specification;
(FIG. 2), but the composition of the columns also undergoes changes. The notion of the
homogeneity of the columns141 has been abandoned in favour of their heterogeneity, and, to develop rodent and possibly non-human primate
both in different functional areas of an individual, as well as across species142,143. In all models, of cortical dysgenesis that mimic specific genetic
mammals, functional columns consist of an intermix of projection (excitatory, or acquired cortical disorders in human77. In the spirit of
glutaminergic) neurons and inhibitory (GABAergic) interneurons that are evo–devo reasoning, these studies illustrate how a single
stereotypically interconnected in the vertical dimension144 but receive diverse extrinsic mutation could have a sudden and profound effect dur-
input9,10,140. To achieve proper cellular composition, several neuronal clones move ing evolution on the pattern of cortical parcellation and
between their parental radial glial cells within the subventricular zone (SVZ) and the presents some insights into how it could have occurred
intermediate zone (IZ) of the embryonic cerebral wall145. The disturbance of this at the genetic and cellular level78.
intermixing, essential for proper cortical operations, may underlie disorders of selected
cognitive functions143. However, although this lateral shift of migrating neurons in the
More neurons, more connections. Obviously, an increase
SVZ–IZ was recognized 35 years ago based on reconstruction from serial electron
microscopic sections145, the molecular mechanisms that regulate this intermix of in the number of neurons, radial columns and cytoar-
neuronal clones has been elusive. chitectonic areas is not sufficient to explain the mul-
A recent study using a combination of the most advanced molecular technology titude of functional advances that evolved to generate
showed that, during radial neuronal migration, the lateral intermixing of neurons that the human cerebral cortex and which must also involve
belong to different clones within a given functional column depends on the expression the elaboration of neuronal connections2,62,79. There is a
levels of A-type Eph receptor tyrosine kinase and their ligands, ephrin-As146. large amount of comparative anatomical data showing
Eph/Ephrin-dependent intermixing in the SVZ and IZ may be only one of several that, in spite of considerable inter-individual variabil-
mechanisms for allocating proper composition of neuronal phenotypes that have been ity, the basic pattern of neuronal connection is species-
operating during mammalian evolution. However, it can serve as an example of how specific and highly reproducible. It is generally agreed
developmental studies in mice may provide an insight into mechanisms of evolutionary
that functional usefulness of evolutionarily new cortical
elaboration of columnar composition that may have happened in our extinct ancestors.
connections must be validated through the process of

728 | O CTObeR 2009 | VOlume 10 www.nature.com/reviews/neuro

© 2009 Macmillan Publishers Limited. All rights reserved


f o c u S o n c n S E V o lo
REVu ItEI W
onS

a Wild type Caspase 9 knockout between the formation of stereotypic, species-specific


connections that are established by surface mediated
attractions and innate, activity-dependent refinements,
from the experience-dependent synaptic plasticity that
shapes connections in cooperation with the outside
environment in each individual. These later-occurring,
experience-dependent changes are not inheritable and,
thus, cannot contribute to evolutionary advances, which
b Wild type ∆90β-catenin–GFP
require gene mutation in germinal cells.

New cell types and migratory pathways


Apart from an increase in the number of neurons and
their allocation into distinct cytoarchitectonic areas, the
evolution of the cerebral cortex has also been accompa-
nied by the introduction of new cell types that migrate
through old or new routes to appropriate, often distant
locations. During 100 million years of separation from a
common ancestor, the mouse and human cerebral cortex
have acquired a number of different features, only some of
which are briefly described in the following subsections.
c Lisencephalic cerebrum Convoluted cerebrum
The subpial granular layer. The subpial granular layer is
a prominent and voluminous transient cellular layer
situated below the pial surface of the fetal human cere-
brum that was first described by O. Ranke more than
a century ago. However, there are only a handful of
studies on its developmental history and function in
human and non-human primates19,23,87,88 and it has not
Increased number
yet been observed in rodents17. It appears around the
Founder cells
of founder cells eleventh gestational week in the human marginal zone,
peaks in size during mid-gestation and disappears by
Figure 3 | enlargement of cortical surface by decrease in programmed cell death
the time of birth17,31,89. Analysis of proliferative activity
or increase in proliferation. a | Brain sections of a wild type mouse
Nature and that
Reviews of a
| Neuroscience
in the developing cerebrum of cats and primates indi-
littermate in which both copies of caspase 9 were deleted are shown. Knockout of
caspase 9 leads to expansion of the progenitor pool, which in turn results in increased cates that the subpial granular layer produces neurons
number of radial columns and consequently a convoluted cortical surface. that may descend to the underlying cortical plate23,87.
c | Mid-coronal section through the forebrain (stained with cresyl violet) of an embryonic based on the uneven size and high incorporation of tri-
day 15.5 wild type mouse and of a transgenic mouse that expresses a Δ90β-catenin–GFP tiated thymidine in the subpial granular layer, which is
(green fluorescent protein) fusion protein in neural precursors. Expression of the fusion located above the developing cortical plate, it may con-
protein results in an increase of the precursor population and of the number of cortical tribute to the wealth of interneurons in certain cortical
columns. c | Diagrams showing how an increased number of progenitor cells can be areas, such as the primary visual cortex, which has an
distributed through radial glial scaffolding to form a sheet, rather than a lump of extraordinary large layer IV compared with rodents23.
postmitotic neurons. Parts a and c are reproduced, with permission, from REF. 35  (1999)
Thus, better characterization of the subpopulations of
Oxford University Press. Part b is reproduced, with permission, from REF. 38  (2002)
interneurons that originate in the subpial granular layer
American Association for the Advancement of Science.
may elucidate their function as well as their possible
involvement in human cortical disorders that do not
natural selection and, if proven to be advantageous, may occur or cannot be mimicked in mice.
contribute to the survival and propagation of the spe-
cies. However, as neurons are generated before they form Migration from the ganglionic eminence to the
axonal connections, an increase in neuronal number, thalamus. A human-specific migratory pathway formed
diversification of their types and attainment of their during mid-gestation by neurons streaming from the
positions will precede the establishment of new, useful ganglionic eminence of the ventral telencephalon
connections in cortical evolution. to the dorsal thalamus of the diencephalon, known
As reviewed in the next section, the new classes of as the corpus gangliothalamicum (CGT), was discovered
neurons allow the formation of new patterns of connec- four decades ago90. This voluminous, easily identifiable
tions. In this complex process, innate neuronal activity stream of migrating bipolar neurons in humans could
is essential for the validation and refinement of synap- not be found in any other species examined to date,
tic connections and leads to their selective stabilization including rodents, carnivores and New and Old World
and/or elimination through competitive interactions80–86. primates31,91. more recently, cell labelling in organotypic
This is a large research field that cannot be considered slice cultures of human embryonic brain tissue enabled
in sufficient detail in this Review. However, it is impor- tracing the migration of these cells directly from the
tant to mention that there should be a clear distinction ganglionic eminence to the dorsal thalamic association

NATuRe ReVIeWS | NeuroscieNce VOlume 10 | O CTObeR 2009 | 729

© 2009 Macmillan Publishers Limited. All rights reserved


REVIEWS

Homotypic–neurophilic nuclei92. These migratory cells, which rely on homotypic– derived almost exclusively from rodents as these are more
guidance neurophilic guidance and express distal-less homeobox 1 amenable to more advanced methods. Thus, tracing cell
Mode of neuronal migration (DlX1) and DlX2 proteins, which guide their tangential lineages using retroviral vectors in mice first showed that
along the surface of other migration, eventually form GAbAergic neurons in the interneurons and projection neurons originate from dif-
neurons that depends on
membrane-bound adhesion
dorsal thalamus, in particular, the pulvinar and medi- ferent progenitors96. Next, ingenious experiments in mice
molecules present on both odorsal nuclei, which are anatomically related to the demonstrated that the majority, if not all, of the cortical
migrating and guiding neurons association neocortex involved in higher cognitive func- interneurons in this species originate from the gangli-
as opposed to heterotypic tions, including symbolic reasoning or language92. Thus, onic eminence in the ventral pallium and then migrate
gliophilic migration that is
it seems that the human thalamus, to accommodate the tangentially to the dorsal telencephalon14,97–100 (FIG. 5).
guided by the shafts of radial
glial cells.
increased axonal input from the expanding association The suggestion that, in addition to those derived
neocortex, recruits an additional complement of neurons from the ganglionic eminence, a large number of
from the nearby, still mitotically active ganglionic emi- interneurons in humans originate from the volumi-
nence in the ventral telencephalon after the diencephalic nous SVZ was initially based on classical histological
proliferative VZ becomes exhausted92. observations31. However, more recently, several studies
that have succeeded to explore the origin of interneu-
Two origins of neocortical interneurons — a working rons in primates with more advanced methods, have
hypothesis. It was Ramón y Cajal93 who suggested that uniformly suggested that cortical GAbAergic neurons
extraordinary human mental abilities are closely related originate from the ganglionic eminence and the VZ
to the increase in the number and diversity of cortical and SVZ101–106 (FIG. 5b). This stands in contrast to the
interneurons, which he classified using the Golgi silver exclusive ganglionic eminence origin of interneurons
impregnation method. However, since then, many other in rodents. Importantly, fate-mapping experiments in
subclasses of primate-specific interneurons have been human mid-fetal slice cultures suggest that the percent-
characterized94,95. In spite of this, our knowledge of the age of cortical GAbAergic neurons that arise from the
origin and development of the interneurons has been dorsal pallium VZ and SVZ might be higher than those
arising from the ventral progenitors101. Furthermore, it
has been shown that proliferative cells (as identified by
a b bromodeoxyuridine incorporation) in slices of the dor-
Rostral Caudal
Cx FGF8 sal human VZ and SVZ during mid-gestation express
LGE M the ventral transcription factors (DlX1, DlX2 and
S NKX2.1)105. Finally, bipolar morphology and orientation
CP MGE
LT FGF18 of DlX-expressing cells also indicate their local origin
OC and radial migration89. These studies collectively sup-
HT port the idea that in primates, in addition to interneu-
FGF8 Eye
FGF17
rons born in the ganglionic eminence of the ventral
telencephalon, there is a population that originates in
the VZ and SVZ of the dorsal telencephalon (FIG. 5b).
Interneurons derived from the ganglionic eminence and
FGF15
the VZ and SVZ are GAbAergic, but their morphology
and molecular characteristics indicate that their func-
tion might be different and that the latter subclass may be
c Wild type Fgf17–/– Fgf8neo/neo Fgf15+/+ Fgf15–/– involved in human-specific higher cortical functions.
Support of the dual origin of interneurons in humans
comes from analysis of a congenital disorder named
holoprosencephaly (HPe) in which the ventral forebrain
structures, including the medial ganglionic eminence, in
some cases fail to develop. examination of the distribu-
Figure 4 | control of arealization of the frontal cortex by FGF expression.
Nature(which
a | Diagram of the forebrain depicting the rostral patterning centre Reviews Neuroscience
is |also known as tion of various classes of cortical GAbAergic interneu-
the commissural plate (CP)) of fibroblast growth factor 8 (FGF8). b | Coronal sections of the rons showed that the neocortex and hippocampus of
telencephalon of an embryonic day 12.5 mouse showing the expression of FGF isoforms in HPe specimens with severe ventral forebrain hypopla-
the rostral patterning centre. c | Coronal sections of the telencephalon of an embryonic sia lack specific subtypes of cortical GAbAergic neurons
day 12.5 mouse showing the regulation of COUP-TF1 (COUP transcription factor 1) that in mice are known to derive from the ganglionic
expression by different FGF isoforms. COUP-TF1 expression is increased in FGF17–/– null eminence106. by contrast, other subtypes of interneu-
mutant mice and FGF8 mutant mice (hypomorphic neo/neo mutant), which indicates that rons seem to be little or not affected by ventral midline
these factors repress COUP-TF1 expression in the wild type background. By contrast, defects, which suggests that they probably originate in
COUP-TF1 expression is decreased in FGF15–/– null mutants, which suggests that FGF15 the preserved dorsal pallium VZ and SVZ. The detailed
promotes COUP-TF1 expression. Cx, cortex; HT, hypothalamus; LGE, lateral ganglionic
analysis of the expression of various cell-specific mark-
eminence; LT, lamina terminalis; M, midbrain; MGE, medial ganglionic eminence; OC,
optic chiasm; S, septum. Part a is modified, with permission, from REF. 148  (1973) ers in the fetal and infant brains of HPe specimens with
American Association for the Advancement of Science. The panels in part b are courtesy mild and severe ventral forebrain defects revealed that
of J. L. Rubenstein and reproduced, with permission, from REF. 75  (2008) Wiley-Liss. The subclasses of GAbAergic cells, distinguished by co-
panels in part c are courtesy of J. L. Rubenstein and reproduced, with permission, from staining of nitric oxide synthase 1 (NOS1), neuropep-
REF. 76  (2008) Biomed central. tide Y (NPY), somatostatin (SST) and parvalbumin

730 | O CTObeR 2009 | VOlume 10 www.nature.com/reviews/neuro

© 2009 Macmillan Publishers Limited. All rights reserved


f o c u S o n c n S E V o lo
REVu ItEI W
onS

a Projection b The finding of evolutionarily novel origins that pro-


neuron duce different subtypes of interneurons in human and
non-human primates does not negate the significance
Interneuron
of the tangential migration of cortical GAbAergic cells
from the ganglionic eminence that were discovered in
CALR the mouse. In fact, studies in human101,102,106 and the
VZ/SVZ DF DF
cynomolgus monkey 103 indicate that the ganglionic
eminence produces a significant portion of interneu-
LGE rons that migrate to the neocortex through the tan-
VZ/SVZ
NOS1, NPY, SST, gential route. A better understanding of the underlying
CALR PVALB mechanisms that generate the greater diversity of inhibi-
NPY tory interneurons in human and non-human primates
SST MGE could uncover a possible role of these cells in human
VF PVALB GE psychiatric disorders95,111.

VF Predecessor neurons. The most recently discovered type


of neurons in human, the predecessor cell, has so far not
been observed in other species32. These large, bipolar
cells are the earliest generated neurons in the human
forebrain that emerge under the pial surface of the ven-
trolateral cerebral wall at the end of the first gestational
Figure 5 | rodent and human fetal forebrains at the peak of corticogenesis. month, before onset of neurogenesis in the VZ and SVZ
Diagrams of cross sections of half of a rodent (a) and a human (b) fetal
Nature forebrain
Reviews are
| Neuroscience of the dorsal telencephalon. They express Tu20 and
shown to scale (a zoom in of the rodent section is also provided for legibility). In rodents, T-box brain gene 1 (TbR1), have long horizontal proc-
the main source of interneurons is the ganglionic eminence (GE) — which is comprised esses and form an extensive network over the forebrain
of the lateral ganglionic eminence (LGE) and the medial ganglionic eminence (MGE) — of primordium32. my research team has proposed that
the ventral telencephalon. These neurons migrate tangentially to the neocortex in the predecessor cells may be a transient population, involved
dorsal telencephalon. By contrast, interneurons in the human forebrain originate both in
in determining the number of functional radial units in
the GE as well as locally in the ventricular and subventricular zones (VZ/SVZ) of the
dorsal telencephalon subjacent to the neocortex. Interneurons originating from the
the cerebral cortex. However, it is important to estab-
cortical VZ/SVZ and GE express a different set of markers, namely nitric oxide synthase 1 lish whether these cells exist in non-human primates to
(NOS1), neuropeptide Y (NPY), somatostatin (SST) and parvalbumin (PVALB) when experimentally determine their lineage, pattern of gene
originating from the GE, and calretinin (CALR) when originating from the VZ/SVZ. expression, function and fate.
DF, dorsal forebrain; VF, ventral forebrain. Data to generate part a was derived from The selected examples of human-specific develop-
REF. 149; data to generate part b was derived from REFS 101,106. mental features and cellular events described earlier are by
no means exclusive. Such evolutionary novelties are
probably the result of gene mutations and/or the intro-
(PVAlb) are consistently absent in both fetal and infant duction of new genes and/or the addition of the devel-
stages. by contrast, the number of calretinin-positive opmental mechanisms and cell–cell interactions that act
GAbAergic cells seem to be little or not affected in the during the progenitor’s exit from the mitotic cycle. They
HPe cases. Furthermore, the absence of cells express- generate a different outcome depending on the evolu-
ing NOS1, NPY, SST and PVAlb correlated with the tionary context. Recent studies indicate that even sub-
dramatic depletion of the thyroid transcription factor 1 types of the pyramidal cells destined for layer V, which
(NKZX-1/TITF1)-positive medial ganglionic eminence are the main efferent system that connects the cortex
progenitors, which are the main source of these sub- with subcortical structures, are determined at the time
types of interneurons in mice. Although not conclusive, of their last division in the VZ112–114. These experimental
the results strongly suggest that cells expressing NOS1, studies in mice indicate that evolutionary novelties,
NPY, SST and PVAlb originate in the ganglionic emi- including introduction of new neuronal types, may start
nence and migrate tangentially into the cortex, whereas with a mutation that regulates the fate of neuronal stem
the calretinin-positive cells could have arisen from the cells in the proliferative zones.
dorsal neuroepithelium of the lateral ventricular wall
and migrated to the cortex radially together with gluta- Small is big: genetic differences
matergic projection neurons. Admittedly, this study One frequently repeated statement, which has almost
cannot eliminate the possibility that the calretinin-posi- become a cliché, is the emphasis on how small the
Association areas
tive cells could have arisen from less affected ganglionic genetic differences are between humans and the rest
Areas of the neocortex that are
particularly large in the human eminence progenitors. However, it is noteworthy that of the animal kingdom. Indeed, only about 1% of the
cortex (for example, prefrontal the neurons expressing NOS1, NPY and SST are most genes are human specific and yet, their effect on tim-
granular cortex or numerous in the derivatives of the marginal zone and ing, sequence and level of gene expression has obviously
language-related Broca and subplate zone in monkeys107, whereas the calretinin- great functional significance. For example, the human
Wernicke areas) are considered
as analysers for integration of
expressing interneurons populate cortical layers II–VI, brain is characterized by expansion in size and complex-
information from various which include the double bouquet cells that are absent ity of the association areas of the neocortex, most notably
sensory and motor areas. from the rodent cortex 108,109,110 (reviewed in REF. 95). the prefrontal cortex, but also the perisylvian areas that

NATuRe ReVIeWS | NeuroscieNce VOlume 10 | O CTObeR 2009 | 731

© 2009 Macmillan Publishers Limited. All rights reserved


REVIEWS

are related to speech and language processing 115. During autism or schizophrenia. Although microarray analy-
human development these areas have a several-fold sis and other techniques have been available for more
larger subplate zone that consists of multipolar neurons than a decade and have been applied in mice118,119, so far
and incoming ‘waiting’ thalamic axons116 and mature at there have been only few genome-wide expression stud-
a slower pace than in other areas117. Furthermore, some, ies of the human brain. The most comprehensive studies
if not most, of the specific functional prefrontal cortex in terms of brain regions were done on aged brains120,121,
subdivisions do not exist in non-primate species and or on comparisons between diseased and healthy
their development, therefore, cannot be modelled easily brains122,123–124.
in mice4,10. Yet, it is precisely these areas that are probably
most significant for the unique human mental abilities, Region specific differential gene expression. A recent
as well as neuropsychiatric disorders such as dyslexia, study examined 13 brain regions from both cerebral
hemispheres of the fetal human brain125. using the most
advanced generation of microarrays, over one million
a known and predicted exons in nine distinct neocorti-
cal areas were profiled, including subdivisions of the
1–3,4 prefrontal cortex, as well as the temporal and parietal
association areas of the perisylvian cortex (FIG. 6a).
32 40 The data revealed a huge number of transcriptional
9
differences, including genes that were both differen-
44,45
41 tially expressed and alternatively spliced. For example,
13 22 although the greatest distinction within the neocortex
13 is between the prefrontal and non-frontal areas, there
are also expression patterns common to the perisylvian
Prefrontal cortex
areas (prospective broca’s, Geschwind’s and Wernicke’s
Perisylvian areas
area), which, although distributed across three lobes,
are functionally linked by their involvement in speech
and language125. The functions of these genes commonly
b Prefrontal cortex Perisylvian areas enriched in perisylvian areas might therefore hold some
PART1 CDH13 clues to the evolution of structure and connectivity that
ASAM HTR7
SLN LRP1B allowed for the development of language in our species.
AKR1B1 FARP1 In addition, the prospective broca’s area (and its right
FAM59A LPHN2
SDK1 AMIGO2 hemisphere homologue) is, on the basis of gene expres-
CACNA1A OPRK1
CBLN2 ZNF8048 sion patterns, more similar to the neighbouring motor
WBSCR17
C12orf38
ENC1
NGEF
cortex than to the prefrontal cortex 125. Interestingly, no
EXTL3 PTPNS statistically significant left–right transcriptional asym-
CPNE8
SLC5A8 LRRC2 metry at the population level was identified, suggesting
KCNJ6 PLXNA4B
DSCR1L1 CSMD2 that such asymmetries most likely occur earlier during
C20orf102 CDH3
SLITRK5
fetal gestation as previously suggested126; but, also not-
H2AFY2
CTNNBL1 UNC5C ing that more sensitive technologies, as well as greater
CYP20A1 FLJ20366
CA8 SYT13 sample sizes, may yet uncover such asymmetry at the
MEGF11
SULF1 CNTNAP5
late mid-fetal stage.
ABCA1 DKFZP564O0823 The application of weighted gene co-expression net-
FKBP9 CENTA1
ZIC3 KIAA0513 work analysis to microarray-derived expression data
PACSIN2 THY1 has previously provided unique insights into the genetic
MARCH3 TMEM28
mechanisms of human brain evolution and disease127.
Figure 6 | Gene expression patterns of neocortical
Applying this analysis to data from the developing
areas of the human fetal cerebral hemispheres.
a | The neocortical areas investigated included four distinct human brain led to the identification of networks of
Nature Reviews | Neuroscience genes that are linked both to brain regions, including
subdivisions of the prefrontal cortex (blue), namely orbital,
medial, dorsolateral and ventrolateral (Brodmann’s areas: the prefrontal cortex, and to biological processes such
9,13,32,44,45), and perisylvian areas (1–3,4,40,41,42) as alternative splicing 125. Genes with the highest degree
(yellow). In addition to the ventrolateral prefrontal cortex, of interaction within such networks, known as ‘hubs’,
which encompasses the prospective Broca’s area (44/45), may represent critical transcription or other regula-
Network of genes two other speech and language-related areas were tory factors, and are prime candidates for functional
A collection of genes that are assayed: parietal and temporal association cortices, which characterization.
co-regulated or interact with encompass the prospective Wernicke’s–Geschwind’s areas
each other. Several recent studies indicate that tissue- and cell
(22/39). b | Examples of different clusters of correlated
type-specific enhancers have crucial roles in neuronal
gene expression patterns in the perisylvian and cortical
Enhancers
areas. Selected highly correlated clusters of genes with the specification and development 128. For example, the tran-
A short region of DNA to which scription factor SOX5 (SRY (sex determining region Y)
proteins including transcription most restricted expression patterns are listed at the right
factors can bind to enhance side of columns. Red is higher level of expression, blue is -box 5) regulates connections of the neurons in the sub-
transcription levels of genes in lower expression. Part b is modified with permission from plate zone and cortical layer 6. Thus, subsets of such
gene clusters. REF. 125  (2009) Cell Press. transcription factors and their enhancers, may have

732 | O CTObeR 2009 | VOlume 10 www.nature.com/reviews/neuro

© 2009 Macmillan Publishers Limited. All rights reserved


f o c u S o n c n S E V o lo
REVu ItEI W
onS

led to the emergence of new gene expression patterns abilities, such as language, are also involved in disorders,
and neocortical areal phenotypes through recent evolu- such as autism, for which there is no accepted mouse
tionary modification129. These spatially comprehensive model137. Future analyses contrasting fetal human brain
expression data in the fetal neocortex have opened the gene networks with comparable data from mice (as has
door to the systematic investigation of the relationship been done for human and chimpanzee adult brains138
between recently evolved enhancers and transcriptional will provide new insights into the complex transcrip-
specificity. Genes that are differentially expressed in the tional and molecular underpinnings of cortical evolu-
developing human neocortex were associated with puta- tion and development of human specific neuronal and
tive recently evolved enhancers, providing a short list of synaptic organization.
candidate genes and regulatory elements for functional
characterization. The number of such associations was Conclusion
statistically disproportionate relative to the genome- The evo–devo approach might eventually elucidate the
wide frequency of such putative enhancers, providing advent of the human in the universe, achieved through
further evidence that the evolution of regulatory factors expansion and elaboration of the cerebral cortex. based
may have had an important role in the evolution of the on the recent findings from the descriptive and experi-
human cortex 125. mental embryonic studies of various mammalian species,
The recent data in humans125 have uncovered an we can develop realistic models of how gene mutations
order of magnitude of greater transcriptional differ- that influenced the fate of cell progenitors during evolu-
ences between neocortical areas than has been obtained tion affect neuron number, regulate their migration into
in comparable studies in rodents130,131. For example, the proper regions, promote formation of new phenotypes
gene CNTNAP2 (contactin associated protein-like 2), that establish new areas with new connections which
previously studied for its role in autism and specific lan- express particular sets of neurotransmitters and recep-
guage impairment132–135 is selectively and highly enriched tors. Neglecting or minimizing evolutionary differences
in the orbital prefrontal cortex, an area involved in regu- between the species may be among the reasons for the
lation of social behaviour in humans and has no com- failures of many clinical trials that were based exclusively
parable analogue in rodents124. The mouse homologue on the highly promising findings in rodents. However,
Cntnap2 has not been found to be expressed in any areal the evo–devo approach for the study of corticoneuro-
pattern or gradient in the mouse brain at any stage of genesis is not only important to the compelling problems
development 136 (m.b. Johnson and N. Sestan, personal of congenital disorders of higher cortical functions in
communication). This is an example of how unique humans, but also provides a hint about how we may have
structures and gene expression patterns that give rise to evolved to be masters of our destiny.

1. Striedter, G. F. Principles of Brain Evolution (Sinauer, (protomap hypothesis). The article also proposes This is the first and the most comprehensive review
Sunderland, Massachusetts, 2005). the radial unit hypothesis of cortical development of the modes and patterns of neuronal migration in
2. Northcutt, R. G. Evolution of the telencephalon in and evolution that has recently been supported by the embryonic and fetal human brain with a special
non-mammals. Ann. Rev. Neurosci. 4, 301–350 genetic and cell biological methods (see also emphasis on the cerebral and cerebellar cortices
(1981). references 38, 39, 65, 74 and 75). (see reference 33).
3. Murphy, W.J, Pevzner, P. A. & O’Brian, S. J. 13. Rakic, P. Principles of neuronal cell migration. 20. Kriegstein, A. R. & Noctor, S. C. Patterns of neuronal
Mammalian phylogenomic comes of age. Trends Experientia 46, 882–891 (1990). migration in the embryonic cortex. Trends Neurosci.
Genet. 20, 631–639 (2004). 14. Marin, O. & Rubenstein, J. L. A long, remarkable 27, 392–399 (2004).
A concise and informative review of the DNA journey: tangential migration in the telencephalon. 21. Angevine, J. B. Jr & Sidman, R. L. Autoradiographic
sequencing-based time-scale of phylogenetic Nature Rev. Neurosci. 2, 780–790 (2001). study of cell migration during histogenesis of
divergence of various mammalian species. A comprehensive and highly informative review on cerebral cortex in the mouse. Nature 192, 766–768
4. Preuss, T. M. in The Cognitive Neuroscience IV. (ed. the pattern of neuronal migration to the cerebral (1961).
Gazzaniga, M. S.) (The MIT Press, Cambridge, cortex with a particular emphasis on the tangential 22. Rakic, P. Neurons in the monkey visual cortex:
Massachusetts, 2009). migration of GABAergic interneurons from the systematic relation between time of origin and
5. Goffinet, A. M. & Rakic, P. (eds) Mouse Brain ganglionic eminence of the ventral telencephalon eventual disposition. Science l83, 425–427 (1974).
Development. (Springer-Verlag, Berlin; New York, (see references 98 and 99). 23. Zecevic, N. & Rakic, P. Development of layer I neurons
2000). 15. Molnár Z. et al. Comparative aspects of cerebral cortical in the primate cerebral cortex. J. Neurosci. 21,
6. Darwin, C. Descent of Man (J. Murray, London, UK, development. Eur. J. Neurosci. 23, 921–934 (2006). 5607–5619 (2001).
1871). 16. Anderson, S. A., Marín, O., Horn, C., Jennings, K. & 24. Aboitiz, F., Montiel, J. & López, J. An hypothesis on
7. Gould, S. J. Ontogeny and Phylogeny (Harvard Rubenstein, J. L. Distinct cortical migrations from the the early evolution of the development of the
University Press, Cambridge, Massachusetts, 1977). medial and lateral ganglionic eminences. Development isocortex. Brain Res. Bull. 57, 481–483 (2002).
8. Carroll, S. B. Evo-Devo and an expanding evolutionary 128, 353–363 (2001). 25. Caviness, V. S. Jr & Rakic, P. Mechanisms of cortical
synthesis: A genetic theory of morphological evolution. 17. Bystron, I., Blakemore, C. & Rakic, P. Development of development: a view from mutations in mice. Ann. Rev.
Cell 134, 25–36 (2008). human cerebral cortex. Boulder Committee revisited. Neurosci. 1, 297–326 (1978).
9. Mountcastle, V. B. The evolution of ideas concerning Nature Rev. Neurosci. 9, 110–122 (2008). A review of the early evidence of the data from
the function of the neocortex. Cereb. Cortex 5, A historical review of the discoveries of the spontaneous mutation in mice showing that the
289–295 (1995). transient embryonic zones and with an update of basic neuronal phenotype reflects the time of
10. Goldman-Rakic, P. S. in Handbook of Physiology, The their nomenclature. neuron origin irrespective of their subsequent
Nervous System, Higher Functions of the Brain Vol. V., 18. Rakic, P. Mode of cell migration to the superficial laminar positions. This finding led to the conclusion
Part 1, Ch. 9 (ed. F. Plum) 373–417 (Bethesda, Md layers of fetal monkey neocortex. J. Comp. Neurol. that neurons attract appropriate thalamic input
Am. Physiol. Soc., Section I, Vol. V., Part 1 1987). 145, 6l-84 (1972). rather than being initially equipotent and specified
11. Brodmann, K. Beiträge zur histologischen This paper presents evidence from a combination by the type of input as previously assumed (see
Lokalisierung der Grosshirnrinde. Dritte Mitteilung: of the Golgi method and serial electron microscopy references 74 and 137).
Die Rindenfelder niederer Affen. J. Psychol. Neurol. 9, that postmitotic neurons in the large and convoluted 26. Gleeson, J. G. & Walsh, C. A. Neuronal migration
177–226 (1905) (in German). primate cerebrum follow the increasingly elongated disorders: from genetic diseases to developmental
12. Rakic, P. Specification of cerebral cortical areas. and curvilinear shafts of the radial glial cells, some mechanisms. Trends Neurosci. 23, 352–359 (2000).
Science 241, 170–176 (1988). of which do not divide while serving transiently as 27. Hatten, M. E. New directions in neuronal migration.
A review of the initial evidence that phenotype, and migratory guides (see reference 47). Science 297, 1660–1663 (2002).
laminar and areal position of cortical neurons are 19. Sidman, R. L. & Rakic, P. Neuronal migration with 28. Rakic, P., Ayoub, A. E., Breunig, J. J. & Dominguez,
specified at the proliferative zones and are only special reference to developing human brain: a review. M. H. Decision by division: Making cortical maps.
later influenced by the incoming axonal input Brain Res. 62, 1–35 (1973). Trends Neurosci. 32, 291–301 (2009).

NATuRe ReVIeWS | NeuroscieNce VOlume 10 | O CTObeR 2009 | 733

© 2009 Macmillan Publishers Limited. All rights reserved


REVIEWS

29. Rakic, P. Pre and post-developmental neurogenesis in 50. Levitt, P., Cooper, M. L. & Rakic, P. Coexistence of 70. Bishop, K. M., Goudreau, G. & O’Leary, D. D.
primates. Clinical Neurosci. Res. 2, 29–39 (2002). neuronal and glial precursor cells in the cerebral Regulation of area identity in the mammalian
30. Rakic, P. A small step for the cell — a giant leap for ventricular zone of the fetal monkey: an ultrastructural neocortex by Emx2 and Pax6. Science 288, 344–349
mankind: a hypothesis of neocortical expansion during immunoperoxidase analysis. J. Neurosci. 1, 27–39 (2000).
evolution. Trends Neurosci. 18, 383–388 (1995). (1981). 71. Garel, S. et al. Molecular regionalization of the
A review of the mechanisms of neocortical The first evidence that neuronal and glial cell lines neocortex is disrupted in Fgf8 hypomorphic mutants.
expansion with a suggestion on how a single or only can be distinguished at the initial stages of corti- Development 130, 1903–1914 (2003).
few genes can shift the timing from symmetric to coneurogenesis in primates, the finding that has 72. Storm, E. E. et al. Dose-dependent functions of Fgf8 in
asymmetric mode of cell division in the embryonic been confirmed in human post-mortem material regulating telencephalic patterning centers.
VZ and can suddenly and exponentially change the (see references 54 and 55). Development 133, 1831–1844 (2006).
surface area of the neocortex. 51. Levitt, P., Cooper, M. L. & Rakic, P. Early divergence 73. Armentano, M. et al. COUP-TFI regulates the balance
31. Sidman, R. L. & Rakic, P. in Histology and and changing proportions of neuronal and glial of cortical patterning between frontal/motor and
Histopathology of the Nervous System. (eds. precursor cells in the primate cerebral ventricular sensory areas. Nature Neurosci. 10, 1277–1286
Haymaker W. & Adams, R. D.) 3–145 (C. C. Thomas, zone. Dev. Biology 96, 472–484 (1983). (2007).
1982). 52. Levitt, P. & Rakic, P. Immunoperoxidase localization of 74. Cholfin, J. A. & Rubenstein, J. L. Patterning of frontal
32. Bystron, I., Rakic, P., Molnar, Z. & Blakemore, C. The glial fibrillary acid protein in radial glial cells and cortex subdivisions by Fgf17. Proc. Natl Acad. Sci.
first neurons of the human cerebral cortex. Nature astrocytes of the developing rhesus monkey brain. USA 104, 7652–7657 (2007).
Neurosci. 9, 880–885 (2006). Comp. Neurol. l93, 8l5–840 (1980). The experimental evidence that prospective
The authors use the newest immunocytochemical 53. Kadhim, H. J., Gadisseux, J.-F. & Evrard, P. cytoarchitectonic areas are specified in the
methods on fresh tissues from the early stages of Topographical and cytological evolution of the glial proliferative zones, as predicted by the protomap
the embryonic human telencephalon to discover a phase during prenatal development of the human hypothesis (see reference 12). The authors also
previously unrecognized cell class, termed brain: Histochemical and electron microscopic study. show that the manipulation of cell proliferation
‘predecessor neuron’. Both the evolutionary J. Neuropath. Exp. Neurol. 47, 166–188 (1988). rate in the ventricular zone can independently
implication and medical significance of this finding 54. Zecevic, N. Specific characteristic of radial glia in the change the size of a selected cortical area (see
in the human cortical primordium are discussed. human fetal telencephalon. Glia 48, 27–35 (2004). reference 64).
33. Kuida, K. et al. Reduced apoptosis and 55. Howard, B. M. et al. Radial glia cells in the 75. Cholfin, J. A. & Rubenstein, J. L. Frontal cortex
cytochrome c-mediated caspase activation in mice developing human brain. Neuroscientist 14, 459–473 subdivision patterning is coordinately regulated by
lacking Caspase-9. Cell 94, 325–333 (1998). (2008). Fgf8, Fgf17, and Emx2. J. Comp. Neurol. 509,
34. Kuida, K. et al. Decreased apoptosis in the brain and 56. Felleman, D. J. & Van Essen, D. C. Distributed 144–155 (2008).
premature lethality in CPP32-deficient mice. Nature hierarchical processing in the primate cerebral cortex. 76. Borello, U., Cobos, I., Long, J. E., Murre, C. &
384, 368–372 (1996). Cereb. Cortex 1, 1–47 (1991). Rubenstein, J. L. R. FGF15 promotes neurogenesis
35. Haydar, T. F., Kuan, C.-Y., Flavell, R. A. & Rakic, P. The 57. Preuss, T. M. Do rats have prefrontal cortex? The and opposes FGF8 function during neocortical
role of cell death in regulating the size and shape of Rose-Woolsey-Akert program reconsidered. J. Cogn. development. Neural Dev. 3, 17 (2008).
the mammalian forebrain. Cereb. Cortex 9, 621–626 Neurosci. 7, 1–24 (1995). 77. Rubenstein, J. L. R. & Rakic, P. Genetic control of
(1999). 58. O’Leary, D. D. M. & Borngasser, D. Cortical ventricular cortical development. Cereb. Cortex 9, 521–552
36. Kornack, D. R. & Rakic, P. Changes in cell cycle zone progenitors and their progeny maintain spatial (1999).
kinetics during the development and evolution of relationships and radial patterning during preplate 78. Rakic, P. Neurocreationalism: making new cortical
primate neocortex. Proc. Natl Acad. Sci. USA 95, development indicating an early protomap. Cereb. maps. Science 294, 1011–1012 (2001).
1242–1246 (1998). Cortex 16 (Suppl. 1), i46–i56 (2006). 79. Krubitzer, L & Kaas, J. The evolution of the
37. Dehay, C. & Kennedy, H. Cell-cycle control and cortical This article reviews the evidence that the initial neocortex in mammals: how is phenotypic diversity
development. Nature Rev. Neurosci. 8, 438–450 neuronal phenotypes for the prospective generated? Curr. Opin. Neurobiol. 15, 444–453
(2007). species-specific pattern and size of (2005).
38. Chenn, A. & Walsh, C. A. Regulation of cerebral cytoarchitectonic areas are indicated early in the 80. Changeux, J. P. & Danchin, A. Selective stabilisation of
cortical size by control of cell cycle exit in neural proliferative zones (see the protomap hypotheses developing synapses as a mechanism for the
precursors. Science 297, 365–369 (2002). in references 13 and 65) as well as recent specification of neuronal networks. Nature 264,
39. Chenn, A. & Walsh, C. A. Increased neuronal experimental evidence (see references 74 and 75). 705–712 (1976).
production, enlarged forebrains and cytoarchitectural 59. Lukaszewicz, A. C. et al., The concerted modulation of 81. Rakic, P. & Riley, K. P. Overproduction and elimination
distortions in beta-catenin overexpressing transgenic proliferation and migration contributes to the of retinal axons in the fetal rhesus monkey. Science
mice. Cereb. Cortex 13, 599–606 (2003). specification of the cytoarchitecture and dimensions of 209, 1441–1444 (1983).
40. Tarui T. et al. Overexpression of p27(Kip1), probability cortical areas. Cereb. Cortex 16 (Suppl. 1), i26–i34 82. Rakic, P. & Riley, K. P. 1983 Regulation of axon
of cell cycle exit, and laminar destination of neocortical (2006). numbers in the primate optic nerve by prenatal
neurons. Cereb. Cortex, 15, 1343–1355 (2005). 60. Rakic, P. Prenatal genesis of connections subserving binocular competition. Nature 305, 135–137
41. Richman, D. P., Steward, R. M., Hutchinson, J. W. & ocular dominance in the rhesus monkey. Nature 26l, (1983).
Caviness, V. S. Jr Mechanical model of brain 467–471 (1976). 83. Huttenlocher, P. R., de Courten, C., Gare, L. J. & Van
convolutional development. Science 189, 18–21 61. Rakic, P. Development of visual centers in the primate der Loos, H. Synaptogenesis in human visual
(1975). brain depends on binocular competition before birth. cortex—evidence for synapse elimination during
42. Goldman-Rakic, P. S. & Rakic, P. in Cerebral Science 2l4, 928–993l (1981). normal development. Neurosci. Lett. 33, 247–252
Dominance, The Biological Foundation (eds 62. Shatz, C. J. Impulse activity and the patterning of (1982).
Geschwind, N. & Galaburda, A. M.) 179–192 connections during CNS development. Neuron 5, 84. Rakic, P., Bourgeois, J.-P., Eckenhoff, M. E., Zecevic,
(Harvard University Press, Cambridge, MA, 1984). 745–756 (1990). N., & Goldman-Rakic, P. S. Concurrent overproduction
43. Van Essen, D. C. A tension-based theory of 63. Kaas, J. H. & Preuss, T. M. eds. in The Evolution of of synapses in diverse regions of the primate cerebral
morphogenesis and compact wiring in the central Primate Nervous Systems Volume 4 (Elsevier, Oxford, cortex. Science 232, 232–235 (1986).
nervous system. Nature 385, 313–318 (1997). UK, 2007). 85. LaMantia, A. S. & Rakic, P. Axon overproduction and
44. Noctor, S. C., Flint, A. C., Weissman, T. A., 64. Fukuchi-Shimogori, T. & Grove, E. A. Neocortex elimination in the corpus callosum of the developing
Dammerman, R. S. & Kriegstein, A. R. Neurons patterning by the secreted signaling molecule FGF8. rhesus monkey. J. Neurosci. 10, 2156–2175
derived from radial glial cells establish radial units in Science 294, 1071–1074 (2001). (1990).
neocortex. Nature 409, 714–720 (2001). The authors use in utero microelectroporation- 86. Shatz, C. J. Form from function in visual system
The experimental evidence, obtained from the live mediated gene transfer to introduce an extra development. Harvey Lect. 93, 17–34 (1997–1998).
images of retrovirally labelled cells in the slices of source of FGF8 into the occipital pole which results 87. Brun, A. The subpial granular layer of the foetal
the embryonic mouse cerebral wall, that radial glial in an extra somato-sensory (barrel) cerebral cortex in man. Acta. Pathol. Microbiol.
cells can produce neurons that directly or after cytoarchitectonic field with a nearly perfect Scand. Suppl. 179, 1–98 (1965).
additional mitotic divisions, migrate and form duplication of the topographic map. 88. Gadisseux, J-F, Goffinet, A. M. Lyon, G. & Evrard, P.
radial columns in the overlaying cortical plate. 65. Mallamaci, A. & Stoykova, A. Gene networks The human transient subpial granular layer: An
45. Fishell, G. & Kriegstein, A. R. Neurons from radial glia: controlling early cerebral cortex arealization. Eur. optical, immunohistochemical, and ultrastructural
the consequences of asymmetric inheritance. Curr. J. Neurosci. 23, 847–856 (2006). analysis. J. Comp. Neurol. 324, 94–114 (2004).
Opin. Neurobiol. 13, 34–41 (2003). 66. Sahara, S., Kawakami, Y., Izpisua Belmonte, J. C. & The detailed study of the subpial granular layers in
Review of the evidence that radial glial cells can O’Leary, D. D. Sp8 exhibits reciprocal induction with the developing human cerebral cortex that, in spite
produce neurons (see reference 44). Fgf8 but has an opposing effect on anterior-posterior of its large size in human (see reference 87) is
46. Gal, J. S. et al. Molecular and morphological cortical area patterning. Neural Dev. 2, 10 (2007). neglected in the literature because of its absence in
heterogeneity of neural precursors in the mouse 67. Rash, G. & Grove, E. A. Patterning the dorsal rodents.
neocortical proliferative zones. J. Neurosci. 26, telencephalon: a role for sonic hedgehog? J. Neurosci. 89. Rakic, P. & Zecevic, N. Emerging complexity of layer I
1045–1056 (2006). 27, 11595–11603 (2007). in human cerebral cortex. Cereb. Cortex 13,
47. Schmechel, D. E. & Rakic, P. Arrested proliferation of 68. Crossley, P. H., Martinez, S., Ohkubo, Y. & Rubenstein, 1072–1083 (2003).
radial glial cells during midgestation in rhesus monkey. J. L. Coordinate expression of Fgf8, Otx2, Bmp4, and 90. Rakic, P. & Sidman, R. L. Telencephalic origin of
Nature 227, 303–305 (1979). Shh in the rostral prosencephalon during development pulvinar neurons in fetal human brain. Z. Anat.
48. Schmechel, D. E. & Rakic, P. A Golgi study of radial of the telencephalic and optic vesicles. Neuroscience Entwickl.-Gersch. 129, 53–82 (1969).
glial cells in developing monkey telencephalon: 108, 183–206 (2001). 91. Letinic, K. & Kostovic, I. Transient fetal structure,
Morphogenesis and transformation into astrocytes. 69. Breunig, J. J. et al. Primary cilia regulate hippocampal the gangliothalamic body, connects telencephalic
Anat. Embryol. 156, 115–152 (1979). neurogenesis by mediating sonic hedgehog signaling. germinal zone with all thalamic regions in the
49. Rakic, P. Elusive radial glial cells: Historical and Proc. Natl Acad. Sci. USA 105, 13127–13132 developing human brain. J. Comp. Neurol. 384,
evolutionary perspective. Glia 43, 19–32 (2003). (2008). 373–395 (1997).

734 | O CTObeR 2009 | VOlume 10 www.nature.com/reviews/neuro

© 2009 Macmillan Publishers Limited. All rights reserved


f o c u S o n c n S E V o lo
REVu ItEI W
onS

92. Letinic, K. & Rakic, P. Telencephalic origin of human parvalbumin in visual areas of the occipital and 131. Muhlfriedel, S. et al. Novel genes differentially
thalamic GABAergic neurons. Nature Neurosci. 4, temporal lobes of the macaque monkey. J. Comp. expressed in cortical regions during late neurogenesis.
931–936 (2001). Neurol. 412, 515–526 (1999). Europ. J. Neurosci. 26, 33–50 (2007).
The use of contemporary methods, including 110. Hendry, S. H. C. et al. Two classes of cortical GABA 132. Arking, D. E. et al. A common genetic variant in the
retroviral labelling in the slices of embryonic human neurons defined by differential calcium binding protein neurexin superfamily member CNTNAP2 increases
brain tissue, to follow the migratory pathway of a immunoreactivities. Exp. Brain Res. 76, 467–472 familial risk of Autism. Am. J. Hum. Genet. 82,
class of thalamic interneurons from their origin in (1989). 160–164 (2008).
the telencephalic ganglionic eminence to the 111. Lewis, D. A. & Levitt, P. Schizophrenia as disorder of 133. Alarcón, M. et al. Linkage, association and gene-
diencephalons, to settle in the thalamic association neurodevelopment. Ann. Rev. Neurosci. 25, 409–432 expression analyses identify CNTNAP2 as an autism-
nuclei. This uniquely human migratory stream was (2002). susceptibility gene. Am. J. Hum. Genet. 82, 150–159
initially observed by classical histological methods 112. Chen, J. G. et al. Zfp312 is required for subcortical (2008).
(see reference 90). axonal projections and dendritic morphology of deep- 134. Bakkaloglu, B. et al. Molecular cytogenetic analysis
93. Ramón y Cajal, S. Textura del sistema nervioso del layer pyramidal neurons of the cerebral cortex. Proc. and resequencing of contactin associated protein-like
hombre y vertebrados. Vol. 2, (Moya, Madrid, Spain, Natl Acad. Sci. USA 102, 17792–17797 (2005). 2 in autism spectrum disorders. Am. J. Hum. Genet.
1899) (in Spanish). 113. Chen, B., Schaevitz,.R. & McConnell, S. K. Fezl 82, 165–173 (2008).
94. DeFelipe, J. Cortical interneurons: from Cajal to 2001. regulates the differentiation and axon targeting of layer 135. Vernes, S. C. et al. A functional genetic link between
Progr. Brain Res. 136, 215–238 (2002). 5 subcortical projection neurons in cerebral cortex. distinct developmental language disorders. N. Engl.
95. Jones, E. G.. The origins of cortical interneurons: Proc. Natl Acad. Sci. USA 102, 17184–17189 (2005). J. Med. 359, 2337–2345 (2008).
mouse versus monkey and human cerebral cortex. 114. Molyneaux, B. J. et al. Fezl is required for the birth 136. Abrahams, B. S. et al. Genome-wide analyses of
Cereb. Cortex 19, 1953–1956. and specification of corticospinal motor neurons. human perisylvian cerebral cortical patterning. Proc.
96. Parnavelas, J. G., Barfield, J. A., Franke, E. & Luskin, Neuron 47, 817–831 (2005). Natl Acad. Sci. USA 104, 17849–17854 (2007).
M. B. Separate progenitor cells give rise to pyramidal 115. Rilling, J. K. & Insel, T. R. The primate neocortex in 137. Levitt, P. Developmental neurobiology and clinical
and nonpyramidal neurons in the rat telencephalon. comparative perspective using magnetic resonance disorders: lost in translation? Neuron 46, 407–412
Cereb. Cortex 1, 463–491 (1991). imaging. J. Hum. Evol. 37, 191–223 (1999). (2005).
97. de Carlos, J. A., López-Mascaraque, L. & Valverde, F. 116. Kostovic, I. & Rakic, P. Developmental history of the 138. King, M. C. & Wilson, A. C. Evolution at two levels in
Dynamics of cell migration from the lateral ganglionic transient subplate zone in the visual and humans and chimpanzees. Science 188, 107–116
eminence in the rat. J. Neurosci. 16, 6146–6156 somatosensory cortex of the macaque monkey and (1975).
(1996). human brain. J. Comp. Neurol. 297, 441–470 (1990). 139. Carroll, S. B. Evolution at two levels: on genes and
98. Anderson, S., Mione, M., Yun, K. & Rubenstein, J. L. R. The detailed temporal and regional description of form. PLoS Biol. 3, e245 (2005).
Differential origins of neocortical projection and local the subplate zone in developing human and 140. Rakic, P. The radial edifice of cortical architecture:
circuit neurons: Role of Dlx genes in neocortical non-human primate cerebrum which shows its from neuronal silhouettes to genetic engineering.
interneurogenesis. Cereb. Cortex 9, 646–654 enlargements subjacent to the prospective Special Issue on: Centenery of Neuroscience
(1999). association areas (see reference 117). Discovery: Reflecting on the Nobel Prize to Golgi
99. Lavdas, A. A., Grigoriou, M., Pachnis, V. & Parnavelas, 117. Kostovic, I. Structural and histochemical and Cajal in 1906. Brain Res. Rev. 55, 204–219
J. G. The medial ganglionic eminence gives rise to a reorganization of the human prefrontal neocortex (2007).
population of early neurons in the developing during perinatal and postnatal life. Prog. Brain Res. 141. Rockel, A. J., Hiorns, R. W. & Powell T. P. S. The basic
cerebral cortex. J. Neurosci. 19, 7881–7888 85, 223–239 (1990). uniformity in structure of the neocortex. Brain 103,
(1999). 118. Zapala, A. et al. Adult mouse brain gene expression 221–244 (1980).
A study of the origin and tangential migration of patterns bear an embryologic imprint. Proc. Natl 142. Herculano-Housel. S., Collins, C. E., Wang, P. & Kaas J.
early generated GABAergic interneurons from the Acad. Sci. USA 102, 10357–10362 (2005). The basic non-uniformity of the cerebral cortex. Proc.
medial portion of the ganglionic eminence of the 119. Semeralul, M. O. et al. Microarray analysis of the Natl Acad. Sci. USA 105, 12593–12598 (2008).
ventral telencephalon to the cerebral cortex (see developing cortex. J. Neurobiol. 66, 1646–1658 143. Rakic, P. Confusing cortical columns Proc. Natl Acad.
also references 14 and 98 for a comprehensive (2006). Sci. USA 105, 12099–12100 (2008).
review on this subject). 120. Roth, R. B. et al. Gene expression analyses reveal 144. Yu, Y. C., Bultje, R. S., Wang, X. & Shi, S. H. Specific
100. Batista-Brito, R., Machold, R., Klein, C. & Fishell, G. molecular relationships among 20 regions of the synapses develop preferentially among sister
Gene expression in cortical interneuron precursors is human CNS. Neurogenetics 7, 67–80 (2006). excitatory neurons in the neocortex. Nature 458,
prescient of their mature function. Cereb. Cortex 18, 121. Haroutunian, V., Katsel, P., Dracheva, S. & Davis, K. L. 501–504 (2009).
2306–2317 (2008). The human homolog of the QKI gene affected in the 145. Rakic, P., Stensaas, L. J., Sayre, E. P. & Sidman, R. L.
101. Letinic, K., Zoncu, R. & Rakic, P. Origin of GABAegic severe dysmyelination “quaking” mouse phenotype: Computer-aided three-dimensional reconstruction and
neurons in the human neocortex. Nature 417, downregulated in multiple brain regions in quantitative analysis of cells from serial
645–649 (2002). schizophrenia. Am. J. Psychiatry 163, 1834–1837 electronmicroscopic montages of fetal monkey brain.
The evidence obtained by using retroviral labelling (2006). Nature 250, 3l–34 (1974).
in the supravital slices of the embryonic human 122. Mirnics, K. et al. Molecular characterization of 146. Torii, M., Hashimoto-Torii, K., Levitt, P. & Rakic, P.
forebrain, showing that a large proportion of schizophrenia viewed by microarray analysis of gene Integration of neuronal clones in the radial cortical
interneurons in the primates are generated in the expression in prefrontal cortex. Neuron 28, 53–67 columns by EphA/ephrin-A signaling. Nature (2009)
local SVZ and migrate radially to the suprajacent (2000). (in the press).
cortex (see references 102, 103 and 106). The 123. Ryan, M. M. et al. Gene expression analysis of bipolar 147. Rakic, P. Less is more: progenitor death and cortical
evolutionary and medical implication for disorder reveals downregulation of the ubiquitin cycle size. Nature Neurosci. 8, 981–982 (2005).
human-specific psychiatric disorders is discussed and alterations in synaptic genes. Mol. Psychiatry 11, 148. Sur, M. & Rubenstein, J. L. R. Patterning and
(see reference 111). 965–978 (2006). plasticity of the cerebral cortex. Science 310,
102. Petanjek, Z., Dujmović, A., Kostović, I. & Esclapez, M. 124. Mao, R. et al. Primary and secondary transcriptional 805–810 (1973).
Distinct origin of GABA-ergic neurons in forebrain of effects in the developing human Down syndrome brain 149. Wonders C. P. & Anderson S. A. The origin and
man, nonhuman primates and lower mammals. Coll. and heart. Genome Biol. 6, R107 (2005). specification of cortical interneurons. Nature Rev.
Antropol. 32, (Suppl 1), 9–17 (2008). 125. Johnson, M. B. et al. Functional and evolutionary Neurosci. 7, 687–696 (2006).
103. Petanjek, Z., Berger, B. & Esclapez, M. Origins of insights into human brain development through
cortical GABAergic neurons in the cynomolgus global transcriptome analysis. Neuron 62, 494–509 Acknowledgements
monkey. Cereb. Cortex 19, 249–262 (2009). (2009). I am grateful to the former and present members of my labo-
104. Rakic, S. & Zecevic N. Emerging complexity of cortical A whole-genome, exon-level expression analysis of ratory whose skills, wisdom, hard work and insightful discus-
layer I in humans. Cereb. Cortex, 13, 1072–1083 fetal human cortex that reveals a large number of sions made this article possible. I am also grateful to the U.S.
(2003). human specific gene expression, alternative Public Health Service and private philanthropic organizations
105. Zecevic, N., Chen, Y. & Filipovic, R. Contributions of splicing patterns and co-expression networks in the that provided funding over the past four decades including
cortical subventricular zone to the development of the prefrontal and parietal association neocortex. NINDS, NEI, NIMH, NIDA, March of Dimes, NARSAD, NAAR
human cerebral cortex. J. Comp. Neurol. 491, 126. Sun, T. et al. Early asymmetry of gene transcription in and MOD Foundations and the Kavli Institute for
109–122 (2005). embryonic human left and right cerebral cortex. Neuroscience at Yale.
106. Fertuzinhos, S. et al. Selective depletion of Science 308, 1794–1798 (2005).
molecularly defined cortical interneurons in human 127. Oldham, M. C., Horvath, S. & Geschwind, D. H.
holoprosencephaly with severe striatal hypoplasia. Conservation and evolution of gene coexpression DATABASES
Cereb. Cortex 19, 2196–2207 (2009). networks in human and chimpanzee brains. Proc. Natl uniProtKB: http://www.uniprot.org
107. Hendry, S. H. C., Jones, E. G. & Emson, P. C. Acad. Sci. USA 103, 17973–17978 (2006). calretinin | caspase 9 | COUP-TF1 | FGF8 | NPY | NOS1 | SST |
Morphology, distribution and synaptic relations of 128. Kwan, K. Y. et al. SOX5 postmitotically regulates PVALB
somatostatin and neuropeptide Y-immunoreactive migration, postmigratory differentiation, and projections
neurons in rat and monkey neocortex. J. Neurosci. 4, of subplate and deep-layer neocortical neurons. Proc. FURTHER INFORMATION
2497–2517 (1984). Natl Acad. Sci. USA 105, 16021–16026 (2008). Pasko Rakic’s homepage: http://rakiclab.med.yale.edu/
108. Ramón y Cajal, S. Histologie du Système Nerveux de 129. Prabhakar, S. et al. Human-specific gain of function in Video of a migrating cortical neuron: http://rakiclab.med.
l’Homme et des Vertébrés, Volume II. Translated by a developmental enhancer. Science 321, 1346–1350 yale.edu/MigratingCorticalNeuron.html
L. Azoulay (Maloine, Paris, France, 1911). (2008). Video of radial migration: http://rakiclab.med.yale.edu/
109. DeFelipe, J. González-Albo, M. C., Del Río, M. R. & 130. Kudo, L. C. et al. Genetic analysis of anterior posterior RadialMigration.html
Elston, G. N. Distribution and patterns of connectivity expression gradients in the developing mammalian All liNks Are Active iN the oNliNe pdF
of interneurons containing calbindin, calretinin, and forebrain. Cereb. Cortex 17, 2108–2122 (2007).

NATuRe ReVIeWS | NeuroscieNce VOlume 10 | O CTObeR 2009 | 735

© 2009 Macmillan Publishers Limited. All rights reserved

You might also like