The neurotransmitter γ-aminobutyric acid (GABA) plays an important role in the motor thalamic nuclei. This report analyzes the distribution of the GABA-producing enzyme glutamic acid decarboxylase isoform 65 (GAD65), stained with monoclonal antibody, in human and rhesus monkey thalami and compares it with staining patterns of some widely used cytoskeletal and calcium binding protein markers. GAD65 immunoreactivity distinctly labeled two systems: fibers and terminals of basal ganglia thalamic afferents and local circuit neurons, revealing fine features of GABAergic circuitry in the human thalamus. Gross distribution patterns of GAD65 were identical in human and rhesus monkey thalami. The area displaying specific staining of large-caliber beaded fibers coincided with nigro- and pallidothalamic afferent territories previously identified in monkeys with anterograde tracers. Accordingly, a similarly stained region in the human thalamus was considered basal ganglia territory. Except for cytoarchitecture, no specific markers differentiating between the nigro- and pallidothalamic projection zones within this territory were found. GAD65 staining in the cerebellar afferent territory reflected organization of its local circuit neuron network, distinguishing it from adjacent nuclei. Specific GAD65 staining pattern and negative calcium binding protein immunoreactivity identify the cerebellar afferent territory in humans. It is subdivided further into ventral and dorsal regions based on the cytoskeletal protein SMI31 staining pattern. The nuclear outlines revised according to the results are compared with those of Hassler (Schaltenbrand G and Bailey P [1959] Einfuhrung in die stereotaktishen Operationen mit einem Atlas des menschlichen Gehirns, vol 3. Stuttgart: Thieme) and discussed in light of the ongoing controversy regarding delineations of the motor thalamic nuclei in humans.

Earlier results on molecularly coded progenitor domains in the chicken pretectum revealed an anteroposterior subdivision of the pretectum in precommissural (PcP), juxtacommissural (JcP), and commissural (CoP) histogenetic areas, each specified differentially (Ferran et al. [2007] J Comp Neurol 505:379-403). Here we examined the nuclei derived from these areas with regard to characteristic gene expression patterns and gradual histogenesis (eventually, migration patterns). We sought a genoarchitectonic schema of the avian pretectum within the prosomeric model of the vertebrate forebrain (Puelles and Rubenstein [2003] Trends Neurosci 26:469-476; Puelles et al. [2007] San Diego: Academic Press). Transcription-factor gene markers were used to selectively map derivatives of the three pretectal histogenetic domains: Pax7 and Pax6 (CoP); FoxP1 and Six3 (JcP); and FoxP2, Ebf1, and Bhlhb4 (PcP). The combination of this genoarchitectonic information with additional data on Lim1, Tal2, and Nbea mRNA expression and other chemoarchitectonic results allowed unambiguous characterization of some 30 pretectal nuclei. Apart from grouping them as derivatives of the three early anteroposterior domains, we also assigned them to postulated dorsoventral subdomains (Ferran et al. [2007]). Several previously unknown neuronal populations were detected, thus expanding the list of pretectal structures, and we corrected some apparently confused concepts in the earlier literature. The composite gene expression map represents a substantial advance in anatomical and embryological knowledge of the avian pretectum. Many nuclear primordia can be recognized long before the mature differentiated state of the pretectum is achieved. This study provides fundamental notions for ultimate scientific study of the specification and regionalization processes building up this brain area, both in birds and other vertebrates.

We describe the expression pattern of cSix3, a chick homologue of the murine Six3. cSix3 transcripts are expressed from presomitic stages in the most anterior portion of the neural plate. As the neural tube folds and the optic vesicles evaginate, cSix3 is expressed in the optic vesicle and the rostroventral forebrain. At later stages, cSix3 is found in most of the structures derived from the anterior neural plate, i.e. olfactory epithelium, septum, adenohypophysis, hypothalamus and preoptic areas. During eye development, cSix3 expression is first found in the entire optic vesicle and the overlying ectoderm but soon becomes restricted to the prospective neural retina and to the lens placode. In the developing neural retina, cSix3 is expressed in the entire undifferentiated neuroepithelium but is rapidly downregulated, first in the postmitotic photoreceptors and later in the majority of retinal ganglion cells.

In a previous work, mapping early tyrosine hydroxylase (TH) expressing primordia in human embryos, the tegmental origin of the substantia nigra (SN) and ventral tegmental area (VTA) was located across several neuromeric domains: prosomeres 1-3, midbrain, and isthmus (Puelles and Verney, [1998] J. Comp. Neurol. 394:283-308). The present study examines in detail the architecture of the neural wall along this tegmental continuum in 6-7 week human embryos, to better define the development of the SN and VTA. TH-immunoreactive (TH-IR) structures were mapped relative to longitudinal subdivisions (floor plate, basal plate, alar plate), as well as to radially superposed strata of the neural wall (periventricular, intermediate, and superficial strata). These morphologic entities were delineated at each relevant segmental level by using Nissl-stained sections and immunocytochemical mapping of calbindin, calretinin, and GABA in adjacent sagittal or frontal sections. A numerous and varied neuronal population originates in the floor plate area, and some of its derivatives become related through lateral tangential migration with other neuronal populations born in distinct medial and lateral portions of the basal plate and in a transition zone at the border with the alar plate. Some structural differences characterize each segmental domain within this common schema. The TH-IR neuroblasts arise predominantly within the ventricular zone of the floor plate and, more sparsely, within the adjacent medial part of the basal plate. They first migrate radially from the ventricular zone to the pia and then apparently move laterally and slightly rostralward, crossing the superficial stratum of the basal plate. Several GABA-IR cell populations are present in this region. One of them, which might represent the anlage of the SN pars reticulata, is generated in the lateral part of the basal plate.

The pretectal region of the brain is visualized as a dorsal region of prosomere 1 in the caudal diencephalon, including derivatives from both the roof and alar plates. Its neuronal derivatives in the adult brain are known as pretectal nuclei. The literature is inconsistent about the precise anteroposterior delimitation of this region and on the number of specific histogenetic domains and subdomains that it contains. We performed a cross-correlated gene-expression map of this brain area in chicken embryos, with the aim of identifying differently fated pretectal domains on the basis of combinatorial gene expression patterns. We examined in detail Pax3, Pax6, Pax7, Tcf4, Meis1, Meis2, Nkx2.2, Lim1, Dmbx1, Dbx1, Six3, FoxP2, Zic1, Ebf1, and Shh mRNA expression, as well as PAX3 and PAX7 immunoreaction, between stages HH11 and HH28. The patterns analyzed serve to fix the cephalic and caudal boundaries of the pretectum and to define three molecularly distinct anteroposterior pretectal domains (precommissural, juxtacommissural, and commissural) and several dorsoventral subdomains. These molecular specification patterns are established step by step between stages HH10 and HH18, largely before neurogenesis begins. This set of gene-architectonic data constitutes a useful scaffold for correlations with fate maps and other experimental embryologic results and may serve as well for inquiries on homologies in this part of the brain.

A polyclonal antibody against the Drosophila distal-less (DLL) protein, cross-reactive with cognate vertebrate proteins, was employed to map DLL-like expression in the midlarval lamprey forebrain. This work aimed to characterize in detail the separate diencephalic and telencephalic DLL expression domains, in order to test our previous modified definition of the lamprey prethalamus [Pombal and Puelles (1999) J Comp Neurol 414:391-422], adapt our earlier schema of prosomeric subdivisions in the lamprey forebrain to more recent versions of this model [Pombal et al. (2009) Brain Behav Evol 74:7-19] and reexamine the pallio-subpallial regionalization of the lamprey telencephalon. We observed a large-scale conservation of the topologic distribution of the DLL protein, in consonance with patterns of Dlx expression present in other vertebrates studied. Moreover, evidence was obtained of substantial numbers of DLL-positive neurons in the olfactory bulb and the cerebral hemispheres, in a pattern consistent with possible tangential migration out of the subpallium into the overlying pallium, as occurs in mammals, birds, frogs and teleost fishes.

Combining gene expression data with morphological information has revolutionized developmental neuroanatomy in the last decade. Visualization and interpretation of complex images have been crucial to these advances in our understanding of mechanisms underlying early brain development, as most developmental processes are spatially oriented, in topologically invariant patterns that become overtly distorted during brain morphogenesis. It has also become clear that more powerful methodologies are needed to accommodate the increasing volume of data available and the increasingly sophisticated analyses that are required, for example analyzing anatomy and multiple gene expression patterns at individual developmental stages, or identifying and analyzing homologous structures through time and/or between species. Three-dimensional models have long been recognized as a valuable way of providing a visual interpretation and overview of complex morphological data. We have used a recently developed method, optical projection tomography, to generate digital three-dimensional models of early human brain development. These models can be used both as frameworks, onto which normal or experimental gene expression data can be mapped, and as objects, within which topological morphological relationships can be investigated in silico. Gene expression patterns and selected morphological structures or boundaries can then be visualized individually or in different combinations in order to study their respective morphogenetic significance. Here, we review briefly the optical projection tomography method, placing it in the context of other methods used to generate developmental three dimensional models, and show the definition of some CNS anatomical domains within a Carnegie stage 19 human model. We also map the telencephalic EMX1 and PAX6 gene expression patterns to this model, corroborating for the first time the existence of a ventral pallium primordium in the telencephalon of human embryos, a distinct claustroamygdaloid histogenetic area comparable to the recently defined mouse primordium given that name [Puelles L, Kuwana E, Puelles E, Bulfone A, Shimamura K, Keleher J, Smiga S, Rubenstein JLR (2000) Pallial and subpallial derivatives in the embryonic chick and mouse telencephalon, traced by the expression of the genes Dlx-2, Emx-1, Nkx-2.1, Pax-6, and Tbr-1. J Comp Neurol 424:409-438; Puelles L, Martínez S, Martínez-de-la-Torre M, Rubenstein JLR (2004) Gene maps and related histogenetic domains in the forebrain and midbrain. In: The rat nervous system, 3rd ed (Paxinos G, ed), pp 3-25. San Diego: Academic Press].

The extent of neurotensin (NT) colocalization in the different dopamine (DA) terminal fields of the rat cerebral cortex has been investigated and compared to previous data obtained in man (Gaspar et al., J. Comp. Neurol., 279 (1989) 249-271). Both innervations were revealed with single- or double-labeling immunocytochemical methods. Tyrosine hydroxylase (TH) was used as a specific marker of DA fibers after lesioning the noradrenergic system either with 6-hydroxydopamine (6-OHDA) at birth or DSP4 in adulthood. Three classes of afferents were observed which had a different regional and laminar distribution. First, a dense meshwork of finely dotted NT-positive varicosities occupied restricted areas of the limbic system: the granular retrosplenial and the deep entorhinal cortices and the subicular complex. These NT projections contained no double-labeled fibers and did not correspond to a mixed NT/TH pathway. Secondly, the mixed NT/DA projections identified previously in the prefrontal cortex (Studler et al., Neuropeptides, 11 (1988) 95-100), extended in fact rostrocaudally in layer VI of the whole cerebral cortex and formed small cluster-like groupings in layers II-III of the medial and lateral entorhinal cortex. In all these areas, the mixed NT/TH projections constituted approximately half of the DA terminals. Finally, the DA projections to the superficial layers of the anterior cingulate, motor, retrosplenial and visual cortices, were not colocalized with NT. The DA innervation of layers I-III of the rat anterior cingulate cortex displays striking similarities with that observed in the cingulate, primary motor, premotor and supplementary motor cortices in man: highest regional and laminar density of DA afferents and lack of colocalization with NT. It might thus represent a valuable model for understanding the pharmacology of the DA system besides the mixed DA/NT pathway which does not seem to have a counterpart in the human cerebral cortex. By contrast, that part of the NT innervation of the limbic system which is not colocalized with DA in rat, appears to represent the major fraction of the cortical NT innervation in man.

The avian lateral septal organ (LSO) is a telencephalic circumventricular specialization with liquor-contacting neurons (Kuenzel and van Tienhoven [1982] J. Comp. Neurol. 206:293-313). We studied the topological position of the chicken LSO relative to molecular borders defined previously within the telencephalic subpallium (Puelles et al. [2000] J. Comp. Neurol. 424:409-438). Differential expression of Dlx5 and Nkx2.1 homeobox genes, or the Shh gene encoding a secreted morphogen, allows distinction of striatal, pallidal, and preoptic subpallial sectors. The chicken LSO complex was characterized chemoarchitectonically from embryonic to posthatching stages, by using immunohistochemistry for calbindin, tyrosine hydroxylase, NKX2.1, and BEN proteins and in situ hybridization for Nkx2.1, Nkx2.2, Nkx6.1, Shh, and Dlx5 mRNA. Medial and lateral parts of LSO appear, respectively, at the striatal part of the septum and adjacent bottom of the lateral ventricle (accumbens), in lateral continuity with another circumventricular organ that forms along a thin subregion of the entire striatum, abutting the molecular striatopallidal boundary; we called this the "striatopallidal organ" (SPO). The SPO displays associated distal periventricular cells, which are lacking in the LSO. Moreover, the SPO is continuous caudomedially with a thin, linear ependymal specialization found around the extended amygdala and preoptic areas. This differs from SPO and LSO in some molecular aspects. We tentatively identified this structure as being composed of an "extended amygdala organ" (EAO) and a "preoptohypothalamic organ" (PHO). The position of LSO, SPO, EAO, and PHO within a linear Dlx5-expressing ventricular domain that surrounds the Nkx2.1-expressing pallidopreoptic domain provides an unexpected insight into possible common and differential causal mechanisms underlying their formation.

For a better insight into general and derived traits of developmental aspects of catecholaminergic (CA) systems in amniotes, we have studied the development of these systems in the brain of a lizard, Gallotia galloti, with tyrosine hydroxylase (TH)- and dopamine (DA) immunohistochemical techniques. Two main groups of TH-immunoreactive (THi) perikarya appear very early in development: one group in the midbrain which gives rise to the future ventral tegmental area, substantia nigra and retrorubral cell groups, and another group in the tuberomammillary hypothalamus. Somewhat later in development, TH/DA-immunoreactive cells are observed in the thalamus, rostrodorsal hypothalamus and spinal cord, and, with another delay, in the suprachiasmatic nucleus, the periventricular organ, and the pretectal posterodorsal nucleus. CA cell groups that appear rather late in development include the cells in the olfactory bulb, the locus coeruleus and the caudal brainstem. As expected, the development of immunoreactive fibers stays behind that of the cell bodies, but reaches the adult-like pattern just prior to hatching. The present study revealed considerable variation in the relation between the state of cytodifferentiation and first expression of TH/DA immunoreactivity between CA cell groups. Catecholamine cells in the midbrain and tuberomammillary hypothalamus are still migrating, immature (absence of dendrites) and express only TH immunoreactivity at the time of first detection. Cells which appear at later developmental stages lie already further away from the ventricle, possess two or more dendritic processes, and generally express both TH- and DA immunoreactivity.

Tyrosine hydroxylase (TH) immunoreactive (IR) central catecholaminergic neurons have been observed in human CNS from 4.5 gestational weeks (g.w.) on [Verney, C., Zecevic, N. and Puelles, L. Eur. J. Neurosci., Suppl. 8 (1995) 7044; Zecevic, N. and Verney, C., J. Comp, Neurol., 351 (1995) 509-535]. We describe here a discrete TH-IR cell population localized in the rostral nasal region during embryonic development. Tyrosine hydroxylase-IR cells spread from the olfactory placode towards the basal and medial telencephalon. They follow the same migration path as the gonadotropin-releasing hormone (GnRH)-IR hypothalamic neurons. Tyrosine hydroxylase-IR neurons are first detected at 4.5 g.w., while GnRH-IR cells are visualized later at 5.5 g.w. Double immunocytochemical labeling reveals the presence of three neuronal populations comigrating along the developing vomeronasal-nervus terminalis complex. These populations express either one or both TH and GnRH phenotypes depending on their position in the migration route. At 6 g.w., most of the neurons express TH immunoreactivity as they leave the vomeronasal organ whereas most of the GnRH-IR neurons are detected closer to the CNS and in the CNS itself. These results emphasize the early phenotypic heterogeneity of the different migrating neuronal populations generated in the olfactory placode in humans. At later stages, very few TH-IR neurons are detected in the anterior forebrain suggesting a transient expression of TH immunoreactivity within these neuronal populations.

It is generally believed that the spinal cord and hindbrain consist of a motor basal plate and a sensory alar plate. We now have molecular markers for these territories. The relationship of migrating branchiomotor neurons to molecularly defined alar and basal domains was examined in the chicken embryo by mapping the expression of cadherin-7 and cadherin-6B, in comparison to genetic markers for ventrodorsal patterning (Otp, Pax6, Pax7, Nkx2.2, and Shh) and motoneuron subpopulations (Phox2b and Isl1). We show cadherin-7 is expressed in a complete radial domain occupying a lateral region of the hindbrain basal plate. The cadherin-7 domain abuts the medial border of Pax7 expression; this common limit defines, or at least approximates, the basal/alar boundary. The hindbrain branchiomotor neurons originate in the medial part of the basal plate, close to the floor plate. Their cadherin-7-positive axons grow into the alar plate and exit the hindbrain close to the corresponding afferent nerve root. The cadherin-7-positive neuronal cell bodies later translocate laterally, following this axonal trajectory, thereby passing through the cadherin-7-positive basal plate domain. Finally, the cell bodies traverse the molecularly defined basal/alar boundary and move into positions within the alar plate. After the migration has ended, the branchiomotor neurons switch expression from cadherin-7 to cadherin-6B. These findings demonstrate that a specific subset of primary motor neurons, the branchiomotor neurons, migrate into the alar plate of the chicken embryo. Consequently, the century-old concept that all primary motor neurons come to reside in the basal plate should be revised.

An unknown chicken gene selected from a published substractive hybridization screen (GenBank Accession No. ; [Christiansen, J.H., Coles, E.G., Robinson, V., Pasini, A., Wilkinson, D.G., 2001. Screening from a subtracted embryonic chick hindbrain cDNA library: identification of genes expressed during hindbrain, midbrain and cranial neural crest development. Mech. Dev. 102, 119-133.]) was deemed of interest because of its dynamic pattern of expression across the forebrain and midbrain regions. A 528bp fragment cloned from early chick embryo cDNA and used for in situ hybridization corresponded to part of the 3' untranslated region of the chicken gene Leucine-rich repeat neuronal protein 1 (Lrrn1). The expression of this gene, mapped in the embryonic chick brain between stages HH10 and HH26, apparently preconfigures the zona limitans thalami site before overt formation of this boundary structure. Apart of colateral expression in the forebrain, midbrain and hindbrain basal plate, the most significant expression of Lrrn1 was found early on across the entire alar plate of midbrain and forebrain (HH10). This unitary domain soon divides at HH14 into a rostral part, across alar secondary prosencephalon and prospective alar prosomere 3 (prethalamus; caudal limit at the prospective zona limitans), and a caudal part in alar prosomere 1 (pretectum) and midbrain. The rostral forebrain domain later downregulates gradually most extratelencephalic signal of Lrrn1, but the rostral shell of zona limitans retains expression longer. Expression in the caudal alar domain also changes by downregulation within its pretectal subdomain. Caudally, the midbrain domain ends at the isthmo-mesencephalic junction throughout the studied period. Embryonic Lrrn1 signal also appears in the somites and in the otic vesicle.