GABAergic interneuron migration and the evolution of the neocortex

Authors

  • Daisuke H. Tanaka,

    Corresponding author
    • Department of Anatomy, Keio University School of Medicine, Shinjuku-ku, Tokyo, Japan
    Search for more papers by this author
    • Present address: Institut de Recherches en Biologie Humaine et Moléculaire (IRIBHM), Université Libre de Bruxelles (ULB), Brussels, B-1070, Belgium.
  • Kazunori Nakajima

    Corresponding author
    • Department of Anatomy, Keio University School of Medicine, Shinjuku-ku, Tokyo, Japan
    Search for more papers by this author

Authors to whom all correspondence should be addressed.

Email: emaniatahon@gmail.com; kazunori@z6.keio.jp

Abstract

A neocortex is present in all mammals but is not present in other classes of vertebrates, and the neocortex is extremely elaborate in humans. Changes in excitatory projection neurons and their progenitors within the developing dorsal pallium in the most recent common ancestor of mammals are thought to have been involved in the evolution of the neocortex. Our recent findings suggest that changes in the migratory ability of inhibitory interneurons derived from outside the neocortex may also have been involved in the evolution of the neocortex. In this article we review the literature on the migratory profile of inhibitory interneurons in several different species and the literature on comparisons between the intrinsic migratory ability of interneurons derived from different species. Finally, we propose a hypothesis about the mammalian-specific evolution of the migratory ability of interneurons and its potential contribution to the establishment of a functional neocortex.

Introduction

The neocortex is located in the dorsal part of the telencephalon, which is at the dorsal part of the rostralmost forebrain subdivision (Puelles 2001; Medina & Abellan 2009). It has been found to be present in all mammals examined thus far, but it is extremely elaborate in humans. The neocortex governs the highest order brain functions, and impairment of the functions of the neocortex can result in cognitive disorders, paucity of motivation for social interactions, and personality changes.

The neocortex has a six-layered cytoarchitecture and consists of two major types of neurons, excitatory glutamatergic projection neurons and inhibitory gamma-aminobutyric-acid (GABA)-containing interneurons (Molnar et al. 2006; Medina & Abellan 2009). The excitatory projection neurons and inhibitory interneurons are connected to each other and form highly organized networks that provide the neuronal basis for the functions of the neocortex.

In this article, we briefly summarize how the neocortex is formed ontogenetically and phylogenetically, and then we review the recent reports on migratory profile/ability of inhibitory neurons in several species, and, finally, we discuss their possible contribution to the evolution of the neocortex.

Ontogenetic origins of the neocortex in rodents

During rodent development, most excitatory projection neurons are generated on the apical side of the neocortical primordium, more specifically, in the ventricular zone (VZ) and subventricular zone (SVZ), and migrate radially toward the basal side of the neocortical primordium (Rakic 1972; Nadarajah et al. 2001; Tabata & Nakajima 2003). They stop migrating just below the marginal zone (MZ) and form the cortical plate (CP; Rakic 1988). Since late-born neurons pass through the early-born neurons that have formed the first part of the CP, the CP ultimately forms in a birth date-dependent inside-out (apical-basal) manner (Angevine & Sidman 1961). On the other hand, most inhibitory GABAergic interneurons are generated in the subpallium, more specifically, in the VZ/SVZ of the medial ganglionic eminence (MGE), caudal ganglionic eminence (CGE), and preoptic area (POa), and migrate tangentially toward the neocortex (Gelman & Marin 2010).

Phylogenetic origin and evolution of the neocortex

Although the neocortex is found only in mammals, since phylogenetically it originates from the dorsal pallium of the ancestor of amniotes (mammals and sauropsids), a homologous structure has been found in the sauropsidian dorsal pallium (avian hyperpallium and turtle/lizard dorsal cortex; Puelles 2001; Medina & Abellan 2009). The cellular mechanism of the development of the sauropsidian dorsal pallium is different from that of the neocortex in several ways: the developing sauropsidian dorsal pallium lacks a SVZ, and its CP is poorly organized and forms in a birthdate-dependent outside-in (basal-apical) manner (Molnar et al. 2006; Cheung et al. 2007). It has been proposed that a mammalian-specific evolutionary change in excitatory projection neurons and their progenitors and a mammalian-specific modulation to them from surrounding tissues was crucial to the establishment of the neocortex (Molnar et al. 2006; Cheung et al. 2007; Nomura et al. 2008, 2009; Sessa et al. 2010; Molnar 2011; Puelles 2011). On the other hand, the tangential migration by inhibitory interneurons from the subpallium to the pallium is highly conserved in all gnathostomes that have been examined thus far, including birds, reptiles, amphibians, and fishes (Medina & Abellan 2009). Thus, the phylogenetic source of the tangential migration by interneurons from the subpallium to the dorsal pallium is likely to be the common ancestor of gnathostomes. This raises the intriguing questions of whether and to what extent mammalian-specific evolution of another important component of the neocortex generated in the subpallium, inhibitory interneurons, contributes to the establishment of a functional neocortex.

Migratory pathways of GABAergic interneurons destined for the neocortex in rodents

In mice, GABAergic interneurons generated in the subpallium mainly enter the neocortex through the neocortical SVZ and intermediate zone (IZ), and after entering the cortex they migrate through several cortical zones, including the SVZ/IZ and the MZ, and ultimately settle in the CP (Kriegstein & Noctor 2004; Yozu et al. 2005; Tanaka et al. 2006, 2009; Yokota et al. 2007; Gelman et al. 2009; Rubin et al. 2010; Yamasaki et al. 2010; Inada et al. 2011; Miyoshi & Fishell 2011). Failure of interneuron migration during development leads to abnormal interneuron distribution (Liodis et al. 2007; Li et al. 2008; Lopez-Bendito et al. 2008; Tanaka et al. 2010), which results in alterations of the inhibitory tone in the postnatal brain (Li et al. 2008). Thus, proper migration of GABAergic interneurons during development is essential for the establishment of a functional neocortex.

Origins and migratory pathways of GABAergic neurons destined for the dorsal pallium in other species

The migratory pathways of neocortical GABAergic interneurons observed in rodents appear to be largely conserved in mammals as a whole. GABAergic interneurons are densely distributed in the MZ and SVZ/IZ of the developing neocortex of opossums (Puzzolo & Mallamaci 2010), marmosets (Fig. 1), cynomolgus monkeys (Petanjek et al. 2009), and humans (Zecevic & Milosevic 1997; Letinic et al. 2002; Zecevic et al. 2011) as well as of rodents, suggesting that most GABAergic interneurons migrate within the neocortex through these zones in mammals. The origins of GABAergic interneurons in mammals, however, may differ with the species, because interneurons appear to be generated within the neocortex in addition to the ganglionic eminences in cynomolgus monkeys and humans (Letinic et al. 2002; Petanjek et al., 2008; Jakovcevski et al. 2011; Yu & Zecevic 2011; but also see Hansen et al. 2010). The origins of the GABAergic neurons in opossums and marmosets have not been investigated.

Figure 1.

Distribution of GABA-positive cells in coronal sections of the marmoset pallium at E80 (A) and E91 (B). Open arrowheads indicate two major migratory streams in the marginal zone (MZ) and the subventricular zone (SVZ)/intermediate zone (IZ). An additional stream is visible between these two streams in A. D, dorsal;GE, ganglionic eminence; L, lateral; Neo, neocortex. Scale bar: 300 μm.

The origins and migratory pathways of pallial GABAergic neurons have been studied in non-mammalian species. In the chicken, approximately 90% of pallial GABAergic neurons originate from the subpallium and approximately 75% originate from the pallidum (Cobos et al. 2001; but also see Tuorto et al. 2003), namely the MGE. They migrate tangentially from the subpallium to the dorsal pallium, but, interestingly, they do not form a clear migratory stream and are evenly distributed in the mantle zone during their migration (Cobos et al. 2001; Figs 2A–D, 3). Similar behaviors have been observed in Xenopus, in which MGE cells migrate tangentially to the pallium without forming a clear migratory stream in the pallial mantle zone (Moreno et al. 2008) (Fig. 3). The origins of GABAergic neurons in the turtle dorsal pallium appear to be diverse, i.e., from the MGE, lateral ganglionic eminence, and around the pallial-subpallal boundary (Metin et al. 2007). Most turtle GABAergic neurons appear to migrate tangentially without forming a clear migratory stream in the pallium (Metin et al. 2007; Figs 2E,F, 3), although transient accumulation of GABA-positive cells in the MZ and at the border between the CP and the VZ has been noted (Blanton & Kriegstein 1991). Clear migratory streams of GABAergic neurons in the pallial superficial layer and periventricular region of the pallium have been observed in sharks (Carrera et al. 2008) (Fig. 3). Many of the GABAergic neurons in the shark pallium appear to originate from the subpallium (Carrera et al. 2008), but the subdivision within the subpallium from which pallial GABAergic neurons originate remains unclear. A migratory stream of GABAergic neurons in the periventricular region of the pallium has also been observed in lampreys (Melendez-Ferro et al. 2002; Martinez-de-la-Torre et al. 2011) (Fig. 3), and the GABAergic neurons appear to originate from the subpallium (Martinez-de-la-Torre et al. 2011). Since Sonic hedgehog and Nkx2.1, which are essential for MGE specification (Sussel et al. 1999; Fuccillo et al. 2004), are not expressed in lampreys (Murakami et al. 2005), pallial GABAergic neurons in lampreys are likely to originate from subpallial subdivisions other than the MGE.

Figure 2.

Distribution of GABA-positive cells in coronal sections of the chicken pallium (A–D) at E5.5 (A), E6.5 (B), E7.5 (C), and E10.5 (D) and of the turtle pallium (E, F) at E14 (E) and E16 (F). D, dorsal; DP, dorsal pallium; DVR, dorsal ventricular ridge; HyP, hyperpallium; L, lateral. Scale bars: 200 μm.

Changes in the migratory ability of MGE-derived GABAergic interneurons and their potential contribution to neocortical evolution

The migratory behaviors and pathways of subpallium-derived GABAergic neurons within the pallium are determined by the intrinsic migratory ability of subpallium-derived GABAergic neurons and the environment found along the route. The environment may vary with the species, because the morphology and cytoarchitecture as well as the developmental mechanisms of the telencephalon differ from species to species. On the other hand, it remains unclear whether there are interspecies differences in the intrinsic migratory ability of subpallium-derived GABAergic neurons. A pioneering study was conducted by Metin et al. to answer this question. They performed interspecies transplantation in vitro and showed that turtle MGE cells transplanted into the mouse MGE region migrated tangentially and entered the mouse neocortex and vice versa (Metin et al. 2007), suggesting that the migratory ability of MGE cells to enter the neocortical primordium had already been established in the most recent common ancestor of amniotes. It has remained unclear, however, whether sauropsidian MGE cells are able to follow the migratory pathways that mammalian MGE cells pass through in vivo and whether they are able to reach the neocortical CP, final destination of most mammalian MGE cells.

To answer these questions, we recently performed interspecies transplantation of chicken, turtle, and marmoset MGE cells into the mouse MGE in utero and examined the distribution of the chicken, turtle, and marmoset MGE cells within the mouse neocortex at different times after transplantation (Fig. 4; Tanaka et al. 2011). While a majority of the chicken MGE cells and turtle MGE cells were found in the SVZ/IZ (Fig. 5A,B) and had failed to enter the CP, especially layers 2–4, the marmoset MGE cells and mouse MGE cells were able to enter the CP (Tanaka et al. 2011). In view of the fact that no equivalent of the supragranular cells of mammals, neocortical layer 2–4 neurons, seems to exist in the dorsal pallium of sauropsids (Medina & Reiner 2000; Aboitiz et al. 2003; Molnar 2011), both the neocortical layer 2–4 neurons in the dorsal pallium and the intrinsic ability of MGE cells to enter the layers 2–4 had been established in the most recent common ancestor of mammals. At 5 days after transplantation a majority (~55%) of the chicken MGE cells within the mouse neocortex were GABA-positive (Fig. 5C–F), and at 29 days after transplantation approximately 80% of the chicken MGE cells expressed either parvalbumin (PV) or somatostatin (SST; Tanaka et al. 2011), suggesting that most of the chicken MGE cells had differentiated into GABAergic interneurons within the mouse neocortex. These results support the hypothesis that the intrinsic ability of MGE cells to generate both PV-positive cells and SST-positive cells in the mammalian telencephalon that was observed in rodents (Wonders & Anderson 2006) had been already established in the most recent common ancestor of amniotes. We also found that at 5 days after transplantation approximately 22% of the chicken MGE cells were positive for platelet-derived growth factor α receptor (PDDFRα; = 50 mCherry-positive chicken cells, two brains), suggesting that a substantial number of tangentially migrating chicken MGE cells within the mouse neocortex were glial progenitors, although we did not detect any morphological evidence of glia-like cells in postnatal brains (Fig. 6). In the postnatal brains the chicken MGE cells were preferentially distributed outside the neocortical gray matter, including in the piriform cortex and the hippocampus (Tanaka et al. 2011; Figs 6, 7). Thus, while the intrinsic ability of MGE cells to migrate tangentially within the neocortical SVZ/IZ had already been established in the most recent common ancestor of amniotes, the intrinsic ability of MGE cells to enter the neocortical CP, especially layers 2–4, had been established in the most recent common ancestor of mammals.

Figure 3.

Major migratory pathways of GABAergic neurons within the developing dorsal pallium of different species. In the species highlighted in yellow the GABAergic neurons derived from the subpallium preferentially migrate within two zones in the developing dorsal pallium, i.e., the superficial zone and periventricular zone, whereas in the species highlighted with light blue they do not form clear migratory streams and instead disperse in the mantle region. In the species highlighted in light green the GABAergic neurons derived from the subpallium preferentially migrate within the periventricular zone in the developing dorsal pallium. The filled circle indicates the presumptive phylogenetic origin of the tangential migration of GABAergic neurons from the subpallium to the pallium. Mya, million years ago.

Figure 4.

Schema of the experimental design of the interspecies transplantations (Tanaka et al. 2011). The medial ganglionic eminence (MGE) was dissected from E13.5 mouse embryos expressing green fluorescent protein (GFP) and from E6.5 chicken, E16 turtle, and E86–93 marmoset embryos, and the chicken, turtle, and marmoset MGE cells were labeled with mCherry by electroporation. Pooled donor cells were then injected into the MGE of E13.5 host embryos in utero. The host brains were analyzed at later stages.

Figure 5.

Chicken medial ganglionic eminence (MGE) cells within the neocortex 5 days (E18.5) after transplantation into the mouse MGE. (A) Distribution of chicken MGE cells (magenta) within the neocortex. Neurofilament (green) have been counterstained to show the intermediate zone (IZ). (B) Enlarged view of the boxed region in A. Some chicken cells (magenta) showed migratory morphology. (C) Distribution of chicken MGE cells (magenta), mouse MGE cells (green), and GABA-positive cells (purple) within the neocortex. (D–F) Enlarged view of the boxed region in C. F is a merged view of D (magenta) and E (purple). The results showed that 55% of the mCherry-positive cells within the neocortex were positive for GABA (= 72 mCherry-positive cells, two brains). D, dorsal; Hip, hippocampus; IZ, intermediate zone; L, lateral; M, medial; Neo, neocortex; SVZ, subventricular zone; VZ, ventricular zone. Scale bars: 200 μm in A, 25 μm in B, 100 μm in C, and 10 μm in D–F.

Figure 6.

Distribution (A) and morphology (B–H) of chicken medial ganglionic eminence (MGE) cells (magenta) and mouse MGE cells (green) in coronal sections of the telencephalon 29 days (P23) after transplantation into the mouse MGE. (B–H) show the morphology of chicken MGE cells labeled with b–h, respectively, in A. AuD, dorsal area of secondary auditory cortex; C, caudal; D, dorsal; ec, external capsule; M, medial; Pir, piriform cortex; Or, oriens layer of the hippocampus; R, rostral; VEn, ventral endopiriform nucleus. Scale bars: 1 mm in A, 200 μm in B–H.

Figure 7.

Quantification of the distribution of chicken medial ganglionic eminence (MGE) cells (magenta) and mouse MGE cells (green) in the telencephalon 29 days (P23) after transplantation into the mouse MGE. The percentages of mCherry-positive cells (magenta) and green fluorescent protein (GFP)-positive cells (green) distributed in different regions are shown (mean ± SEM; n = 3 brains, 43 slices, 1012 mCherry-positive cells, 1790 GFP-positive cells). Regions are ranked in descending order of the percentage of mCherry-positive cells minus the percentage of GFP-positive cells. For the full spellings of the abbreviations of the names of the brain regions, see Franklin & Paxinos (2008).

Excitatory glutamatergic projection neurons are the largest cell population in the neocrtex in terms of number of cells. Since the size and cytoarchitecture of the dorsal pallium vary considerably from species to species, the evolution of the projection neurons of the dorsal pallium and their progenitors was obviously crucial to the evolution of the structure of the neocortex. It is now clear that the evolution of a minor but functionally indispensable component of the neocortex, inhibitory GABAergic interneurons, was important to the evolution of neocortical function. Identification of the molecular and genetic mechanisms underlying the changes in the migratory ability of GABAergic interneurons during evolution will provide a novel view of neocortical evolution and may provide some insights into neurological diseases in humans.

Acknowledgments

This work was supported by the Strategic Research Program for Brain Sciences (“Understanding of molecular and environmental bases for brain health”), the Grant-in-Aid for Scientific Research, the Global COE program of the Ministry of Education, Culture, Sports, Science, and Technology of Japan, and the Promotion and Mutual Aid Corporation for Private Schools of Japan.

Ancillary