The prosencephalic morphogenetic process has not yet been fully characterized because of important anatomical transformations occurring mainly in the optic and telencephalic epithelial domains that have not yet been studied in detail. At neural plate stages (Hamburger and Hamilton [HH] 4–8 in chick and embryonic day [E]7–8.5 in mouse embryos), the neuroepithelium of the anterior neural plate is an almost flat structure which, after dorsal folding, develops into the prosencephalic vesicle at the neural tube stage (HH9 in chick and E9.5 in mouse embryos). Later, four spherical vesicles protrude from the secondary prosencephalon. While the optic field evaginates bilaterally to develop the optic vesicles (between stages HH10–14 in chick and E9.5–11.5 in mouse embryos), which then produces the retina and the optic nerve, the telencephalic field, bulges dorsally to generate the telencephalic hemispheres (between stages HH13–20 in chick and E10.5–12.5 in mouse embryos).
The anterior neuroepithelium undergoes a progressive process of regionalization which is initiated at the neural plate stage by the region-specific expression of different genes. A number of descriptive studies have summarized and discussed the accumulated knowledge of forebrain ontogeny and regional development based on morphological observations and neurogenetic patterns (Altman and Bayer,1986; Puelles et al.,1987; Easter et al.,1993). The identification of regulatory genes specifically expressed in restricted subdomains of the neuroepithelium has been essential to understand the mechanisms by which genetic information encodes complex positional information during vertebrate neural development. Indeed, the prosomeric model, based on the analysis of topological interrelationships between gene expression patterns during neural plate and neural tube stages, is the currently accepted paradigm to classify brain regions during development (Bulfone et al.,1993; Puelles and Rubenstein,1993,2003; Rubenstein et al.,1994; Rubenstein and Puelles,1994; Puelles,1995,2001; Shimamura et al.,1995,1997). According to this model, the secondary prosencephalon (which includes the telencephalon and the hypothalamus) is a proto-segment that is not subdivided into transverse units or prosomers. Instead, this area seems to present patterning properties that are partially independent of the rest of the neural tube (Puelles and Rubenstein,2003).
By comparing patterns of molecular expression in many vertebrate species, it was possible to identify two main components of the telencephalic neuroepithelium, the pallium and the subpallium (Smith-Fernandez et al.,1998; Puelles et al.,2000). The pallial domain gives rise in mammals to the cerebral cortex and the claustroamygdaloid complex (and corresponding parts in birds), whereas the subpallial domain develops into characteristic nuclear areas: the striatum and pallidum, as well as entopeduncular and preoptic areas. Information concerning the cell lineage relationships between these molecularly defined neuroepithelial domains and the mature neural formations in the mantle layer of the telencephalon is incomplete (Balaban et al.,1988; Smith-Fernandez et al.,1998; Cobos et al.,2001b). As a consequence, the topology of telencephalic prospective subdomains at early neural tube stages and their morphogenetic transformations during subsequent development are only vaguely known at present.
Previous fate map studies have thrown much light on our understanding of the organization of the neural plate in different vertebrate species (axolotl: Jacobson,1959; Xenopus laevis: Eagleson and Harris,1990; zebrafish: Woo and Fraser,1995; chick: Couly and Le Douarin,1985,1987; Smith-Fernandez et al.,1998; Cobos et al.,2001b; Fernandez-Garre et al.,2002). The map proposed by Cobos et al. (2001b) was based on experimental grafting data and describes the location of specific neuroepithelial domains at the neural plate stage, which later develop into different telencephalic areas, classified according to the prosomeric model (Rubenstein et al.,1994; Puelles1995,2001). Our goal in this work was to analyze the evolution of the topological neighborhood relations between internal prosencephalic regions, from neural plate to neural tube stages, and to evaluate related growth patterns. Thus, we needed to compare the shape, size, and relations of prospective prosencephalic regions in neural plate and neural tube fate maps. To this end we used quail–chick chimeric embryos and analyzed the positional progression of the grafted neuroepithelium at different developmental stages.
The resulting fate map demonstrates that the topological relationships between the neuroepithelial regions are consistent during neurulation and maintained afterwards during later development, irrespective of their changing topography as the neural tube forms. The resulting distribution of the telencephalic neuroepithelial subdomains in relation to the secondary organizers acting upon the telencephalon, the anterior neural ridge (ANR), and the insertion locus of the choroid plexus, or cortical hem, help to better understand the scaffold of positional information which controls prosencephalic growth and regionalization.
Commissural plate; pallial commissure
Dorsal pallial lamina
Lateral pallial lamina
Lateral septal nucleus, ventral part
Olfactory tubercle, pallidal part
Olfactory tubercle, striatal part
Ventral pallial lamina
MATERIALS AND METHODS
We use the anatomical nomenclature proposed by Puelles (2001), recently revised in a chick brain atlas (Puelles et al.,2007), in order to facilitate comparative analysis with the neural plate fate map (Cobos et al.,2001b). The most relevant points requiring mention is that the telencephalic pallium is subdivided into ventral pallium (nidopallium in Reiner et al.,2004), lateral pallium (mesopallium in Reiner et al.,2004), dorsal pallium (hiperpallium in Reiner et al.,2004), and medial pallium or hippocampus.
Quail (Coturnix coturnix japonica) and chick (Gallus gallus) fertilized eggs used for this study were obtained from commercial sources. In order to reach the adequate stages for the experiments, chick eggs were incubated at 37°C in a forced air incubator for 36 hours, whereas quail eggs were incubated for 31 hours, as quail develops faster. Chick embryos were used at stage HH10 according to Hamburger and Hamilton (1951); the quail embryos used showed equivalent morphological characteristics.
Experiments were performed under relatively sterile conditions, and quail embryos were always used as donors. A grid with concentric circles was inserted in one ocular of the operating microscope in order to normalize graft dimensions and establish relative distances between recognizable structures and presumptive territories. At the working magnification used during microsurgery (40×), the difference in radial distance between any two adjacent concentric circles in the grid was 40 μm. For surgery, the grid center was positioned on the anterior end of the notochord, while the vertical axis was superposed to the ventral midline (Fig. 1).
The egg shell was opened using a thin surgical blade under sterile conditions. First, embryos were counterstained in ovo with Indian ink, diluted 1:5 in Tyrode's solution supplemented with antibiotics, which was injected under the blastoderm using a glass micropipette. Afterwards, the vitelline membrane was slit open over the anterior pole of the embryo with a tungsten needle and the selected neuroepithelial segment was excised from the host. An equivalent piece of tissue from the quail donor was cut with a tungsten needle (using the same procedure as in the chick), transferred to the host using a glass micropipette, and grafted into the chick embryo. The piece of donor tissue was inserted into the space prepared previously in the host, maintaining the original rostrocaudal and dorsoventral orientation. Emplacement with respect to the ocular frame was annotated for every transplant. Rectangular or wedge-shaped pieces of neuroepithelium were usually cut using the grid as reference, although occasionally other shapes were also used.
After the operation the eggs were sealed with a piece of tape and incubated without tilting until they reached the stages chosen for histological analysis. Utmost care was taken to minimize distortion of the chimeric neural tube. Chimeric embryos with morphological alterations were rejected.
Histological analysis of chimeras
The majority of the embryos were incubated until stage HH25 (4 days after graft) or HH35 (8 days after graft), although several embryos were analyzed at other intermediate or posterior stages. HH25 chimeras were fixed overnight in 4% paraformaldehyde in phosphate-buffered saline solution (PBS; 0.1 M, pH 7.4). Afterwards, the brain was isolated and rinsed with PBS, dehydrated through a series of ascending methanol concentrations, and stored in 100% methanol at −20°C before being processed for immunostaining and/or in situ hybridization.
For HH35 chimeras the head was fixed overnight in Clarke's fixative. Afterwards, it was stored in 70% ethanol at room temperature. Subsequently, the heads were dehydrated progressively in ethanol and embedded in paraffin. Finally, 14-μm-thick sagittal serial sections were obtained and mounted in four parallel series.
In some cases the embryos were collected at the indicated stages and fixed overnight in 4% paraformaldehyde at 4°C, cryoprotected in 30% sucrose, then embedded in 4% gelatin, and finally frozen in cryoblocks using liquid-nitrogen cooled isopentane (Prolabo, Barcelona, Spain). Finally, 20 μm sagittal or horizontal serial sections were obtained using a cryotome.
In situ hybridization.
Several chimeric embryos fixed at stage HH25 were processed in toto by in situ hybridization using a digoxigenin-labeled mRNA probe to detect Dlx-2 transcripts, as indicated by Shimamura et al. (1994). For in situ hybridization we synthesized antisense digoxigenin-labeled riboprobe for Dlx-2. Dlx-2 riboprobe was provided by Rubenstein's laboratory (the Dlx-2 DNA sequence used was 5′TGTTTTTGTTCTCTTTTTGTTCAGTCCCCAACATCAGAATG- AGGTCTTCCGCAAAGGCACCTATCGACTTTAAAAAAATAAA- CGGGGGAATGATCCGCGGTCCCCTAGAAGATGCCGCCG- CCGCCGCCGCCGCCGCCGCCACCGCCGCCGTGGGGTGC- TGAGGGAGGGGGGCGGCGGGNGGCGCGGGGCCGGGCG- GGTACCACGAGTAGCTGCCGAGGAAGGCCGCGGCGGG- GCCGGGGCTGGCGGGGCCGCCCGCCGCGCGCGGGT- GCGGGTGCGGGGCGAAGTCCCAGGCGGGCGG- CGAGGCAGGGGGCGGCGGCGAGGCGCTGCGGC- GCGGGGGCGGCTCGGCGGGCAGCTCGCCGCTC- TTCCACATCTTCTTGAACTTGGAGCGGCGGTTC- TGGAACCAGATCTTCACCTGGGTCTGGGTGAGGC- CGAGGGAGGCGGCCAGCTCGGCGCGCTCGGGCA- GCGCCAGGTACTGCGTCTTCTGGAAGCGGCGCT- GCAGGGCCGCCAGCTGGAAGCTGGAGTAGATG- GTGCGCGGCTTGCGGACCTTCTTGGGCTTCCCCTCGTGCC 3′). The RNA-probe was detected with an alkaline-phosphatase-coupled antibody against digoxigenin (Roche Diagnostics, Mannheim, Germany), and NBT/BCIP was used as a chromogenic substrate for the alkaline phosphatase (Boehringer, Mannheim, Germany). After hybridization the embryos were rinsed with PBT (phosphate-buffered saline solution; 0.1 M, pH 7.4, containing 0.1% Tween-20) and processed for immunostaining.
One of the serial sections was stained with cresyl violet and another was immunostained with monoclonal anti-quail antibody, which reacts with quail cell nuclei (QCPN), using the standard biotin-avidin procedure. For quail cell nuclei (QCPN), a monoclonal mouse antibody (Hybridoma Bank, Iowa City, IA; 1:10) raised against quail wing bud ZPA (zone of polarizing activity) was used. This antibody, developed by Bruce M. and Jean A. Carlson, stained the perinucleolar heterochromatin in all quail cells from stages HH20 through HH40 according to previous reports (Cobos et al.,2001a,b). In some cases parallel serial sections were counterstained with cresyl violet to detect anatomical landmarks. No staining was seen when the antibody was used to stain tissue from control chick embryos without quail tissue grafted. After immunostaining the slides were mounted in Eukitt (O. Kindler, Freiburg, Germany). HH25 specimens were immunostained with QCPN antibody in toto using the same protocol.
In all cases selected sections and whole-mount embryos were photographed with a digital camera (Leica DC500). Contrast and brightness of the photomicrographs were adjusted in Adobe PhotoShop, Macintosh or PC version, CS2 (Adobe Systems, San Jose, CA).
Fate map of chick embryo at stage HH10
Quail–chick transplants were used to analyze the localization of the prospective pallial and subpallial domains of the telencephalon at the neural tube stage. Differently shaped quail grafts were transplanted into dorsal regions of the prosencephalic vesicle of chick embryos (Fig. 1), which correspond to the prospective telencephalon according to previous fate mapping data (Smith-Fernandez et al.,1998; Garcia-Lopez et al.,2004).
Anterior and anterolateral prosencephalic grafts in the neural tube were found to map to the prospective subpallial structures (n = 70). In the early-fixed (HH25) chimeric embryos with anterior grafts (case AQ243 is representative of 11 similar grafts; Fig. 2a), quail cells gave rise to a ventromedial sector of the telencephalic vesicle extending roughly between the olfactory bulb and the ventral subpallial eminence (Puelles et al.,2000). According to the molecular regionalization reported by Puelles et al. (2000), the transplanted areas largely correspond to the subpallial septum and the rostromedial areas of the pallidum and striatum (Fig. 2a). These grafts did not include laterocaudal part of the subpallium and barely impinged upon the pallium. Analyzed at later developmental stages (HH 35; n = 18), indeed, the grafted areas gave rise to the septum and adjacent pallidal and striatal regions (Fig. 2b,c). The grafted tissue did not extend to tissue located ventral to the anterior commissure, which according to the neural plate fate map (Cobos et al.,2001b) includes the precursors of the preoptic area and lamina terminalis. In case AQ184 (Fig. 2b,c), a small part of the ventral pallium (olfactory bulb) was also transplanted together with the subpallial precursor tissue (Fig. 2b,c). Moreover, we noted that whenever prospective ventral pallium was included in the graft, quail cells were also detected in the superficial subpallial mantle layer down to the preoptic area (Fig. 2c).
Grafts that were located slightly caudal to the anterior contour of the prosencephalic vesicle, as seen in case AQ517, representative of five similar grafts (Fig. 2f–n) did not exhibit quail cells in the most anterior subpallium (pallidum). Rather, quail cells were restricted to the anterior striatum and the striatal septum (Fig. 2f,g). Comparative analysis of AQ184 and AQ517 chimeric brains reveals that the striatopallidal boundary projects on the HH10 midline approximately at the sixth circle, that is, 240 μm rostral relative the grid center. The mutual positional relationships of the striatum and pallidum observed at neural plate stages, where the pallidum lies medial to the striatum (Cobos et al.,2001b), were therefore maintained during neurulation, but after neural tube closure the mediolateral topographic distribution they had at the anterior pole of the neural plate becomes ventrodorsal in the closed neural tube.
Several quail–chick transplants involved a more lateral area of the anterior part of the prosencephalon, separated from the midline (AQ368, AQ245, AQ186, and AQ159 are representative cases of 25 similar grafts; Fig. 2i–p). Chimeric embryos fixed at stage HH25 (n = 13) showed a preferential growth of transplanted epithelium in a ventrodorsal direction, transforming the square shape of the transplant into a somewhat triangular form. While the anterior border of the graft (#4, Fig. 2i,m) fell outside of the evaginated telencephalon and remained narrow, the evaginated posterior (#2, Fig. 2i,m) became broader, suggesting a heterogeneous distribution of epithelial growth in the process of telencephalic morphogenesis. The lateral end of the graft was transformed into a ventral narrow band ending close to the optic recess of the third ventricle (Fig. 2i) or inside the optic stalk, when the retinal field was included in the graft (Fig. 2m).
At HH25 the expression pattern of Dlx2 in the subpallium allowed us to localize the pallio-subpallial boundary (psp), and then to detect how the medial border of the AQ368 graft crossed this limit (#1, Fig. 2i). The posterior border of the graft (#2 in Fig. 2i) was parallel to this limit and consequently, mapped to the direction of the prospective pallio-subpallial boundary in the prosencephalic vesicle.
These chimeric grafts, when analyzed at HH35 (late fixation stage; n = 12), gave rise to both the lateral part of the striatum (Fig. 2j) and the pallidum (Fig. 2j,k). The grafted striatal and pallidal regions continued laterally toward the amygdala nuclear complex (Fig. 3a–f). Graft AQ186 also included part of the ventral pallium (Fig. 2j,k). In all grafts which extended more than 240 μm from the midline into the lateral edge of the prosencephalic vesicle, quail cells were detected in the nasal retina (n = 16) (data not shown). The entopeduncular and preoptic areas, which later give rise to the peduncular subpallium, occupied a concentric epithelial band between 240 and 280 μm, 6 and 7 circles lateral to the midline (Fig. 2m–o). In the chick neural plate, the retinal field was located concentrically to the subpallium (Cobos et al.,2001b). In our present results, growth and dorsal movement of the anterior alar plate to form the roof of the prosencephalon, together with the evagination of the optic vesicle, resulted in the presumptive retina moving from its central position in the neural plate to the lateral pole of the prosencephalic vesicle (see below). In the anterior midline, the presumptive territory of the subpallial septum was exclusively grafted in the case AQ517, showing that this septal region extends up to two grid circumferences (80 μm) from the anterior pole of the dorsal midline (Fig. 2a–h).
Analysis of chimeric brains revealed that pallial at HH10 corresponded to primordia included alar prosencephalic epithelium lying between the anterior five concentric circles (i.e., a radius of 200 μm from the central point) and the lateral six frame circles from the central reference point of the frame (240 μm radius from the central point) (n = 63; Fig. 4). According to these data, the prospective epithelium of the pallium is found caudally to the subpallium and medial to the retina, occupying the central part of the dorsal prosencephalon. Nested and overlapping transplants, which expanded from the tip of the neural tube to the center of the grid, were analyzed to locate the presumptive epithelium of the different pallial subdomains (Fig. 4).
Cases AQ255 (as a representative case of nine chimeras fixed at HH25; Fig. 4a) and AQ156 (as a representative case of 15 chimeras fixed at HH35; Fig. 4b,c) are examples of rectangular-shaped grafts in the anteromedial prosencephalon. The transplanted epithelium was located in the anterior forebrain, close to the midline (Fig. 4a). At later developmental stages, the graft developed into the following regions: the subpallial septum, medial zones of the subpallium including the unevaginated subpallium (formed by the preoptic and entopeduncular areas), pallidal and striatal regions and the rostralmost pallial domains: the anterior parts of the ventral pallium, including the olfactory bulb, the lateral pallium, the dorsal pallium, and the ventral pallium (Fig. 4b,c, and data not shown). At the level of the dorsal midline the graft developed into the anteromedial pallium and pallial septum as well as the pallial commissure (data not shown). Remarkably, no part of the choroidal plexus of the lateral ventricle was labeled. The posterior limit of the graft separated anterior from the posterior parts of the pallium. The lateral edge of the graft extended to a line separating anterior from posterior parts of the subpallium.
Cases AQ304 (as a representative of 12 chimeras fixed at HH25; Fig. 4e) and AQ138 (as a representative of eight chimeras fixed at HH35; Fig. 4f,g) involved grafts that were more caudal along the dorsal midline than in the case of AQ156. AQ138 presented quail derivatives distributed from the ventral to the medial pallium. However, no transplanted derivatives were observed in the subpallium, indicating that the psp boundary is located at the dorsal midline between the fifth and sixth circumferences. In this chimera, the anterior border of the graft was almost parallel to the pallial-subpallial boundary. Hence, we can place this median limit at 220 μm anterior to the center of the grid. The neuroepithelium between the 4th and 5th concentric circles, in both AQ138 and AQ156, represented the presumptive territory of the anterior and middle areas of the medial, dorsal, lateral, and ventral pallium, excluding caudal and amygdaloid parts. The grafted regions of the roof plate corresponded to the pallial septum and pallial commissure, including the anterior part of medial pallium, but no choroid plexus was produced by these types of grafts (Fig. 4b,c,f,g).
In even more caudal transplants, quail cells contributed to various caudal pallial regions in the HH25 fixed chimeras (n = 8; see AQ232 and AQ253 as representative cases; Fig. 4i,m), as well as in the HH35 fixed chimeras (n = 11; see chimeras AQ187 and AQ320 as representative cases; Fig. 4j–l,n–p). In the case AQ187 neuroepithelial regions included the dorsal and medial pallium in their lateral part (excluding the hippocampus proper, dentatus gyrus primordium, and lateral fimbria portions), whereas in the roof plate, the donor epithelium included the caudal septum and medial part of the fimbria. Case AQ320 (Fig. 4n–p) labeled the caudal dorsal pallium and caudomedial parts of the medial pallium (Fig. 4n), which included at HH35 the hippocampus proper, the dentate gyrus primordium, and the neighboring fimbria. In both sets of the data, part of the choroidal plexus of the lateral ventricle was labeled. In chimera AQ320, the rostral diencephalic territory was also formed by quail cells, together with the choroid plexus in its dorsal midline (Fig. 4n,o).
According to these results, the dorsal pallium primordium is located roughly along the 4th concentric circumference of the grid, while the presumptive epithelium of the medial pallium was mapped largely between the 3rd and 4th circumferences. When the grafts extended caudally behind the third circumference, exactly closer than 100 μm to the reference central point, the grafted epithelium contained choroid plexus (Fig. 4o,p). Moreover, since the prosencephalic choroid plexus showed a quail origin only when prospective medial pallium was grafted, we mapped the roof plate corresponding to the medial pallium at the caudal part of the telencephalic commissural plate, where axons running through the fimbria cross the midline to form the hippocampal commissure. The insertion of the choroid plexus would thus occur at the caudally oriented rim of the medial pallium, in a manner alike the IV ventricle choroid plexus attaches to the cerebellar rhombic lip.
The presumptive pallial subdomains therefore appear to be located essentially transverse to the longitudinal axis. Whereas the ventral pallium was the most anterior region bordering on the subpallium, lateral, and dorsal pallial regions were located at more posterior levels, preceding the medial pallium. The latter, however, is both caudal and dorsal, topologically, since its longitudinal border with the choroidal tela has to be interpreted at longitudinal border between alar and roof plate tissues. Moreover, the medial pallium and ventral pallium form a continuous ring (the pallial ring) around a central island formed by the lateral and the dorsal pallium. The virtual line separating ventral pallium from medial pallium would be topologically longitudinal. Caudal to the medial pallium appears the future prethalamus (anterior diencephalic prosomeric region or p3), represented dorsally by the primordium of the prethalamic eminence. The velum transversum presumably limits at the roof of the prosomer 3 (p3) from the prosomer 2 (p2), and the paraphysis lies in p3.
The presumptive territory of the pallium covered the dorsal part of the secondary prosencephalon and extends ≈240 μm in the mediolateral axis and 130 μm in the anteroposterior axis in a dorsal view of the embryo. Thus, the territory of the prospective pallium reported by Cobos et al. (2001b) in the neural plate has almost tripled its size in the mediolateral dimension after neural tube closure, whereas the rostrocaudal dimension maintained a similar size. Therefore, the presumptive pallial neuroepithelium preferentially extends first mediolaterally and later ventrodorsally, when the neural plate folds to form the neural tube (Fig. 5a–c).
Fate map of the chick telencephalon at the neural tube stage
According to the present fate map, which is restricted to the portion of dorsal forebrain that is visible and accessible for grafting at stage HH10 from a dorsal viewpoint (Fig. 5g), the presumptive subpallium derives from the anteriormost region of the prosencephalic vesicle, and the striatum maps caudal to the pallidum. Posterior to the subpallium, we localized the region that will give rise to the pallium, which is subdivided into ventral, lateral, dorsal, and medial pallium. The limits of these neuroepithelial subdomains are roughly parallel to the prospective pallio-subpallial limit (psp) and are located at successive transverse intervals along the rostrocaudal axis (Fig. 5g). With the exception of the anterior and posterior poles of the pallial ring, intrapallial boundaries are orthogonal relative to the rostral brain roof and course toward the locus of the retinal anlage.
Transplants of lateral parts of the prosencephalic vesicle showed that the presumptive pallium and subpallium extend in ventral topological direction, first into the amygdala complex (Fig. 3; amygdala defined according to Puelles et al.,2007) and then into the peduncular and optic extratelencephalic structures (Fig. 2i,m,o). Among the chimeric brains studied, some subpallial grafts entered into the optic cup crossing through the preoptic area and the optic stalk (Fig. 2m,o), whereas pallial transplants extending into the optic cup tended to cross through the supraopto-paraventricular region. The fate map of the chick neural plate (Fig. 5a; Cobos et al.,2001b) described the limits of presumptive pallial as being transverse to the rostrocaudal axis, represented by the lateral edge of the neural plate (the prospective dorsal midline), and they progressively converged more ventrally upon the retinal field. After neural tube closure, combined morphogenetic phenomena related to formation of the optic peduncle, evagination of the retinal domain, evagination and expansion of the lateral (topologically dorsal) region of the pallium and subpallium, lead to apparently changed relationships of these domains by HH25, and notably of the supraopto-paraventricular area (Fig. 5g,h), although in fact the initial neighborhood relationship are conserved.
The present fate map bears distinctly upon the septum and amygdale areas, whose primordial loci had not been resolved before. The presumptive territory of the septum constitutes a longitudinal region at the dorsal end of the alar plate, extending as well into the roof plate, along the whole axial dimension of the telencephalon. All the transverse subdomains of subpallium, as well as the medial pallium, end dorsal and ventrally in contact with septal counterparts and amygdala complexes, respectively, a pattern that explains the anatomical complexity of these regions. The axial bending that occurs after HH10 brings this continuum of subpallioseptal domains to an apparently ventral topographic position (Fig. 5g,h). On the other hand, the amygdala forms another longitudinal region of the telencephalon, but in this case it lies at the topological ventral end of the telencephalon (Fig. 5g). Like the septum, this domain consists of pallial and subpallial portions (Puelles et al.,2000). The amygdala complex was mapped using consecutive lateral grafts; subpallial (anterior) amygdala was grafted together with the subpallial derivatives (Fig. 3a,c), whereas pallial (posterior) amygdala regions were grafted when pallial neuroepithelium was grafted (Fig. 3d–i).
At stage HH10, the medialmost subpallial regions of the pallidum, entopeduncular, and preoptic areas are not detected in the grafts performed in this work, because these structures were located underneath the operative field, and are thus hidden for observation and manipulation purposes under the prosencephalic vesicle.
Evaluating the precision of the fate map
Since one of the uses of fate maps is their application in predictive experiments, we explored the accuracy of our fate map by grafting specific domains. Several grafts were designed to check the specific pallio-subpallial boundary (psp) either at the caudal end of a subpallial graft (e.g., case AQ517; Fig. 6a–e), or at the rostral end of a pallial graft (e.g., case AQ654; Fig. 6f–k). Case AQ517 labeled essentially subpallial areas (Fig. 6d,e). The graft stopped precisely at the psp boundary (Fig. 6b,c) and did not invade the pallial region. The anterior limit of the pallial graft AQ654 also stopped at the psp limit close to the midline complementarily to AQ517 (Fig. 6g–i,k; compare Fig. 6a,f). The caudal limit of this pallial graft lay between the dorsal and medial pallium, close to the dorsal pallial lamina (dpl; Fig. 6i,j).
The grafted epithelium in the AQ693 graft was designed to cross longitudinally the caudal pallial domains with an edge at the circumferential boundary between ventral and lateral pallial presumptive domains (Fig. 6l). As expected, quail derived territories were detected across the medial, dorsal, and lateral pallium, while the end of the graft corresponded to the lamina between the lateral and the ventral pallium (vpl; ventral pallial lamina; Fig. 6m–p). In lateral sagittal sections, quail cells were found distinctly in lateral and dorsal parts of the pallium, but not in the ventral pallium (Fig. 6m). In medial sections, transplant derivatives were found in the medial pallium and in the posterior pallial septum (not shown).
We had observed that the medial pallium was grafted only when caudal grafts were performed. Case AQ657 was designed to graft the medial pallium (Fig. 6q), and indeed showed the majority of quail cells in this region and the fimbria (Fig. 6r–u). Parts of the choroid plexus were grafted at the dorsal midline of cases AQ693 and AQ657 (Fig. 6l,q), at the back of the telencephalic commissural plate and the fimbria (Fig. 6t,u).
Cell proliferation analysis by BrdU labeling
We observed that the size of the presumptive neuroepithelial domains in the alar secondary prosencephalon underwent an important increase in transversal dimension during neurulation (Fig. 5a). We speculated that this might be due to either generation of a heterogeneous pattern of mitotic activity prosencephalic alar plate, or to irregular distribution of cell divisions in the neuroepithelium, favoring the transverse plane. We therefore studied the spatial pattern of neuroepithelial proliferation in the prosencephalon in order to determine whether differential growth effect, occurring during neural tube closure, is related to differential proliferation.
We analyzed cell proliferation in chick embryos by injecting bromodeoxyuridine (BrdU) directly into the neural tube. The brains of these embryos were then examined for anti-BrdU-labeling 1–4 hours after injection. After these periods of incubation, there were always fewer BrdU-labeled neuroepithelial cells in the ventral part of the neural tube compared to the dorsal region; for instance, after 1 hour of incubation, we have counted 15 ± 3 cells per section in the basal plate versus 25 ± 6 cells per section in an equivalent territory of the alar plate, Fig. 7a). Embryos analyzed 2–4 hours after BrdU injection showed less clear differences. These results suggest that at the neural tube stage the alar region of the prosencephalic vesicle has a higher proliferation rate than its basal counterpart in agreement to previous studies (Hamburger,1948; Corliss and Robertson,1963, review), and that the lateral growth of the telencephalic primordium may be due principally to the higher proliferation rate of the alar plate neuroepithelial cells.
We did not detect substantial differences in the distribution of BrdU positive cells in the different parts of the prosencephalic alar plate. Thus, the heterogeneous growth (transverse enlargement: from 80μm in the neural plate to 240 μm in the neural tube in the mediolateral axis. Anteroposterior dimension is conserved during neurulation (Fig. 5a) of the telencephalic primordium may well be due to a predominance of mediolateral cell intercalation in the alar prosencephalic neuroepithelium, as opposed to dorsoventral intercalation, asymmetry which is favored toward the transverse plate.
According to previously published chick fate maps (Couly and Le Douarin,1985,1987; Balaban et al.,1988; Smith-Fernandez et al.,1998; Cobos et al.,2001b; Fernandez-Garre et al.,2002), the telencephalic primordium is located in the anterior portion of the dorsal prosencephalon. Smith-Fernandez et al. (1998) described three major transverse divisions in the telencephalon: the striatum, the dorsal ventricular ridge (DVR), and the dorsal and medial pallium. Their DVR encompasses our lateral and ventral pallium. Our results are consistent with the previously suggested topographic arrangement of the main telencephalic primordia, but add novel anatomical information bearing upon current concepts of telencephalic regionalization in the context of the revised prosomeric model (Puelles and Rubenstein,2003), as well as a more detailed analysis of internal topological relations among pallial and subpallial telencephalic structures.
Prospective forebrain roof
The prevailing concept about the primary alar plate longitudinal domain of the neural plate is that it crosses the midline at the domains that later produce the septal bed of the anterior commissure, lamina terminalis, and the optic chiasma (Shimamura et al.,1995; Rubenstein et al.,1998). The prospective telencephalon represents a dorsal subdomain of the forebrain alar plate, which includes at its rostralmost median part of the anterior commissure and the lamina terminalis with the associated preoptic area, but excludes chiasmatic and caudally related peduncular areas of the forebrain alar plate (Puelles and Rubenstein,2003; Puelles et al., 2008). The prospective roof plate found on top of the telencephalon field is also considered a part of it. Obviously, there are no basal parts of the telencephalon within this concept.
A topological peculiarity that pertains to such fundamental morphological analysis of the telencephalon is that neurulation produces via fusion of the anterior neural ridge at the midline, caudally to the anterior commissure, two distinct domains of the telencephalic midline: one of them is primary and transversal (i.e., the midline across the lamina terminalis, up to anterior commissure) and the other secondary and longitudinal, and represents the definitive roof plate. These portions were originally orthogonal to each other in neural plate, but become co-aligned at the midline (90–180° change; Fig. 5a) in the neural tube. This topological change is the first cause of the apparent change in the position of the telencephalic domains when they are mapped from the neural plate into the closed tube (Cobos et al.,2001b). The second cause of the apparent change is represented by the axial bending of the neural tube at the cephalic flexure.
We noted that at stage HH10 (rostral neuropore still open) the presumptive epithelium of the lamina terminalis as well as the anterior preoptic and entopeduncular areas, were already ventrally positioned (topographically) and thus were not visible from a dorsal viewpoint upon the neural tube (Figs. 1, 5g). These results suggest that the morphogenetic movements producing the cephalic flexure have already commenced during early neurulation. This process produces a ventral rotation of the forebrain, which progressively hides the anterior pole of the alar forebrain epithelium under the evaginating eye vesicles (Fig. 5a–d). With the progression of the cephalic flexures, this ventral rotation will shift anterior telencephalic structures under topologically more caudal ones; for instance, the lamina terminalis, anterior commissure, and subpallium become placed under the pallial derivatives (Fig. 5g). It is important to remark that dorsal growth of the subpallial and pallial epithelium (toward the roof plate) could be misinterpreted as rostral growth if this early rotation is not taken into account.
The telencephalic roof plate begins rostrally in the subpallial septum, starting at the anterior commissure and follows into the caudal septum, which contains the pallial and hippocampal commissure (there is no corpus callosum in birds) and the anterior tip of the choroid plexus. Several genes are specifically expressed in this territory: Fgf8 (Crossley and Martin,1995; Crossley et al.,1996; Shimamura and Rubenstein,1997), Noggin (Shimamura et al.,1995), BMP (Pera and Kessel.,1997; Furuta et al.,1997; Golden et al.,1999; Streit and Stern,1999), Wnt and Lmbx (Lindwall et al.,2007). Moreover, the telencephalic roof plate is held to be a dorsal organizing center for the neighboring alar forebrain (Shimamura and Rubenstein,1997; Rubenstein et al.,1998; Cheng et al.,2006; Storm et al.,2006). The overall transverse topology of the different telencephalic regions (see below) is consistent with a shared dorsoventral character of the roof plate histogenetic influence upon these territories. Indeed, this inductive effect would be essentially similar to the influence of the roof plate over the alar plate of the spinal cord and the rhombencephalon (reviewed by Chizhikov and Millen,2005).
Choroid roof tissue derived from grafted cells was found (attached to the fimbria) exclusively when the caudal pole of the pallium and the anterior diencephalon (prethalamus; Puelles and Rubenstein,2003) were included in the grafted epithelium. In contrast, slightly more rostral grafts exclusively containing medial pallium areas mapped into the septal midline roof along the fimbria and the hippocampal commissure, with very few choroid components. These results strongly suggest that the roof plate of the secondary prosencephalon scarcely has choroid derivatives, which are mainly located in the caudally adjacent diencephalic roof. The limit between the telencephalon and the diencephalon is thus located at the roof plate just behind the insertion of the choroid tela in the prospective fimbria, laterally, and the hippocampal commissure in the midline (Fig. 4). A further “medial” organizer region, the cortical hem, has been postulated at the fimbria, where Wnt signaling seems fundamental for the normal development of the medial pallium (hippocampus) (Shimogori et al.,2004). Therefore, the growth of the telencephalic vesicle and the rhomboidal expansion of the roof plate domain, caudal to the telencephalon proper, results in a peculiar “cerebellar-rombic-lip-like” cortical hem organizer. Curiously, mitogenic effects from these two areas produce peculiarly extended periods of neurogenesis in hippocampus and cerebellum. This suggests that Wnt signaling at the insertion of these choroid roof plate territories may represent an analogous morphogenetic mechanism at the cerebellar rhombic lip and the hippocampal fimbria.
Prospective domains of the forebrain alar plate
The anterior forebrain alar plate contains the telencephalic field dorsally and the alar hypothalamus ventrally (Puelles and Rubenstein,2003). The latter contains the eye field and surrounding hypothalamic areas, which later expand when the eye evaginates. Consistently with the results of Garcia-Lopez et al. (2004) on the diencephalon, the eye field at stage HH10 is displaced laterally into the tip of the “optic vesicle.” Note that this movement coincides with an axial rotation of the telencephalic and retinal fields, through which the primary dorsal counterpart of the optic vesicle becomes nasal retina and the ventral epithelium becomes the temporal retina. This occurs between stages HH9 to HH11, when cell fate seems to be set up along the naso-temporal axis (Thanos et al.,1996). Due to the rotation of the prosencephalic vesicle, the chiasmatic area ventral to the lamina terminalis is shifted to an apparently caudal position.
The presumptive subpallial territory was located in the anterior region of the nonevaginated telencephalon at HH10 consistently with other fate maps (Smith-Fernandez et. al.,1998; Cobos et al.,2001b; Fernandez-Garre et al.,2002). Cobos et al. (2001b) placed the subpallial primordium at stage HH8 in a mediorostral area of the prosencephalic alar plate primordium, medial to the pallium. The telencephalic primordium of the neural plate progressively acquires a dorsal paramedian position during neurulation to neural tube closure. Subsequent morphogenetic axial bending of the neural tube will push the telencephalic subpallium into an even less visible underlying position, topped by the enlarging pallium (Fig. 5e–g). The rostral halves of the neural plate edge closer and eventually meet and fuse in the dorsal midline of the tube to give rise to the septal domain and the telencephalic commissural roof plate, which ends rostrally in the anterior neuropore.
The pallial telencephalon contains various primary subregions, which are homologous to the corresponding pallial areas in mammals: medial, dorsal, lateral, and ventral pallium (Puelles et al.,2000). Although these regions were interpreted as topologically transverse in our previous works on the neural plate fate map (Cobos et al.,2001a,b), the limits separating them were drawn only tentatively. In this work, we have precisely located the intrapallial boundaries separating these domains, and also improved the interpretation of the heterogeneous dorsal and ventral telencephalic domains forming the septum and the amygdala, respectively (Fig. 5g). The boundaries between the pallial regions were characterized as roughly parallel transverse limits. This major conclusion on the topological transverse character of the major pallial domains, jointly with the same conclusion for subpallial domains is highly relevant for the interpretation of telencephalic patterning in terms of anteroventral versus dorsoventral signaling mechanisms. Nevertheless, the interpretation of pallial domains as transverse entities needs to be qualified by the fact that there is factual support for the idea that the medial and the ventral pallium contact each other at the retrobulbar and amygdaloid meeting points, enclosing the dorsal and the lateral pallium as an island (Altman and Bayer,1991; Puelles,2001; Puelles et al., 2008). Our results are also consistent with this concept, which can be conciliated with the other interpretation by incorporating the nontransverse parts of the medial and the ventral pallium as parts of the septum and the amygdala.
We have identified three principal coexisting events underlying prosencephalic morphogenesis during neurulation: ventral rotation of the prosencephalon, telencephalic and optic evagination, and preferential neuroepithelial growth in the lateral direction. These morphogenetic events represent the cellular basis of cephalic flexure development (axial rotation), and optic and telencephalic vesiculation with the preferential dorsal expansion of telencephalic domains.
In conclusion, our fate map is able to predict how the presumptive epithelial progenitor domains of the diverse telencephalic pallial and subpallial regions are distributed in the prosencephalic vesicle, well before differential molecular characteristics or structurally visible landmarks appear among them. Therefore, this map constitutes a helpful tool for the experimental analysis of molecular and cellular mechanisms underlying regionalization in the chick telencephalon.
We thank Luis Puelles for discussion and suggestions on this topic as well as our technicians M. Rodenas, C. Redondo, A. Torregrosa, and O. Bahamonde for technical assistance in this study. We also thank Jonathan Jones for revision of the English and the URGASA farm (Lleida) for kindly providing quail eggs used for this work.