Development 127, 457-467 (2000) 457 Printed in Great Britain © The Company of Biologists Limited 2000 DEV9689

A local Wnt-3a signal is required for development of the mammalian hippocampus

Scott M. K. Lee1,*, Shubha Tole2,‡, Elizabeth Grove2 and Andrew P. McMahon1,¤ 1Department of Molecular and Cellular Biology, The Biolabs, Harvard University, 16 Divinity Avenue, Cambridge, MA 02138, USA 2Department of Neurobiology, Pharmacology and Physiology, Committees on Developmental Biology and Neurobiology, and Pritzker School of Medicine, University of Chicago, Chicago, IL 60637, USA *Current address: Nina Ireland Laboratory of Developmental Neurobiology, Center for Neurobiology and Psychiatry, Department of Psychiatry and Programs in Neuroscience, Developmental Biology and Biomedical Sciences, 401 Parnassus Avenue, University of California at San Francisco, CA 94143-0984, USA ‡Current address: Department of Biological Sciences, Tata Institute of Fundamental Research, Mumbai-400,005, India ¤Author for correspondence (e-mail: [email protected])

Accepted 27 October 1999; published on WWW 12 January 2000

SUMMARY

The mechanisms that regulate patterning and growth of the appear to be specified normally, but then underproliferate. developing cerebral cortex remain unclear. Suggesting a By mid-gestation, the hippocampus is missing or role for Wnt signaling in these processes, multiple Wnt represented by tiny populations of residual hippocampal genes are expressed in selective patterns in the embryonic cells. Thus, Wnt-3a signaling is crucial for the normal cortex. We have examined the role of Wnt-3a signaling at growth of the hippocampus. We suggest that the the caudomedial margin of the developing cerebral cortex, coordination of growth with patterning may be a general the site of hippocampal development. We show that Wnt- role for Wnts during vertebrate development. 3a acts locally to regulate the expansion of the caudomedial cortex, from which the hippocampus develops. In mice Key words: Wnt, Hippocampus, Pattern formation, Proliferation, lacking Wnt-3a, caudomedial cortical progenitor cells Cortical patterning

INTRODUCTION cerebral cortex is divided into allocortex and neocortex, a number of studies are beginning to provide clues about how The embryonic dorsal telencephalon of vertebrates contains the specific areas are specified within the neocortex. A number of cerebral cortex, which is the seat of higher cognitive functions lines of evidence reveal an instructive role for thalamic inputs in mammals. The cerebral cortex is subdivided into many in specifying the functional properties of neocortical areas histologically and functionally distinct regions. The broadest (reviewed in O’Leary et al., 1994). For example, if fetal visual subdivisions are between the six-layered isocortex (neocortex) cortex is transplanted into neonatal somatosensory cortex, the and the non-six-layered allocortices, which include the transplanted tissue gives rise to a somatosensory cortex with archicortex and the paleocortex, as well as transitional cortices many of its normal properties (Schlaggar and O’Leary, 1991). (Zilles and Wree, 1995). The neocortex and the allocortex are A corollary of these studies is that positional information must further subdivided into unique areas based on cytoarchitectural be present in the cortical primordia prior to the ingrowth of and connectional specialization. thalamic axons in order to guide them to their appropriate The hippocampus is an archicortical structure located at the target areas. The existence of pathways regulating the caudomedial edge of the neocortex (Amaral and Witter, 1995). establishment of such positional information can be inferred In the adult, the longitudinal axis of the hippocampus forms a directly from a number of studies examining the mechanisms ‘C’-shape. From its rostral pole, located just dorsal and that specify the differential expression of genes in different posterior to the septal nuclei, the hippocampus curves over and cortical areas (Barbe and Levitt, 1991; Arimatsu et al., 1992; behind the diencephalon to the incipient temporal lobe, its Ferri and Levitt, 1993; Barth and Stanfield, 1994; Cohen- caudal pole. The transverse axis of the hippocampus is divided Tannoudji et al., 1994; Bulfone et al., 1995; Ebrahimi-Gaillard into distinct fields. From proximal to distal they are the dentate and Roger, 1996; Levitt et al., 1997). While these studies all gyrus (DG) and the CA3 and CA1 fields of Ammon’s horn (the point to the existence of positional information within the early CA fields are often referred to as the hippocampus proper). cortical primordia, the molecular mechanisms responsible for Lateral to the hippocampus is the subiculum and the establishing this information remain obscure. pre/parasubiculum, transitional cortices that lie between the Developmentally the hippocampus arises at the caudomedial hippocampus and the adjacent neocortex. edge of the continuous dorsal telencephalic neuroepithelium While little is known about the mechanisms by which the adjacent to the telencephalic roof (Stanfield and Cowan, 458 S. M. K. Lee and others

1979a,b, 1988; Bayer, 1980a,b). A strip of cortical progenitor 3a acts at 10.5 dpc to regulate the proliferative expansion of cells along this edge, the cortical hem (or hem), expresses caudomedial cortical progenitor cells. The loss of this many members of the Bone Morphogenetic Protein (BMP) and ‘hippocampal’ progenitor pool leads to a failure of normal Wnt families of inductive signaling factors (Furuta et al., 1997; hippocampal development in Wnt-3a mutants. Grove et al., 1998). A number of studies have revealed roles for BMP and Wnt signaling in the regulation of dorsal patterning at more caudal regions of the vertebrate neuraxis MATERIALS AND METHODS (Dickinson et al., 1994; Liem et al., 1995, 1997; Arkell and Beddington, 1997; Ikeya et al., 1997; Lee et al., 1998;). Thus Embryo collection, histology and RNA in situ analysis the cortical hem is a candidate source of information regulating The Wnt-3a mouse line was maintained on an outbred Swiss Webster the induction and/or patterning of hippocampal development at background and genotyping was performed as described (Takada et the caudomedial margin of the continuous cerebral cortical al., 1994; Yoshikawa et al., 1997). Embryos from heterozygous neuroepithelium. intercrosses were dissected into PBS and fixed in either 4% Wnt-3a expression marks the cortical hem from 9.75 dpc paraformaldehyde (for RNA in situ analysis and TUNEL) or in (Roelink and Nusse, 1991; Parr et al., 1993; Grove et al., 1998) Bouin’s fixative (for histological analysis and bromodeoxyuridine (BrdU) detection). For analysis of tissue sections, embryos were either and is the only Wnt gene expressed exclusively in the cortical sectioned at 5-7 µm on a conventional microtome or 40 µm on a hem at this time, suggesting a role for Wnt-3a signaling in the freezing microtome. Sections for general histology were stained induction and/or patterning of the hippocampus. In Wnt-3a with haematoxylin and eosin. All analyses were performed using mutants, medial hippocampal fields are absent and lateral previously published protocols: whole-mount RNA in situ hippocampal fields are severely reduced. We show that Wnt- hybridization according to Parr et al. (1993) with modifications described in Knecht et al. (1995), section in situ hybridization with 35S-labeled probes according to Wilkinson et al. (1987), section in situ hybridization with digoxigenin (DIG)-labeled probes was performed according to Tole et al. (1997), and TUNEL analysis according to Gavrieli et al. (1992). BrdU detection and cell counting Pregnant females were injected intraperitoneally with 50 µg BrdU/g body weight and killed 30 minutes later. The uterine horns were immediately removed into ice-cold PBS. Embryos were dissected and processed as described above (note that exclusively 5 µm

Fig. 1. Expression of Wnt genes during early cerebral cortical development. (A-L) Whole-mount in situ hybridizations using digoxigenin (DIG)-labeled probes. (M-R) In situ hybridizations to tissue sections using 35S-labeled probes. At 8.5 dpc (A-C), prior to cephalic neural tube closure, Wnt-7b and Wnt-8b are expressed in similar domains along the dorsocaudal edge of the telencephalon (B,C) with the expression of Wnt-7b extending further rostrally than Wnt-8b (black arrowheads mark the telencephalon/diencephalon boundary). By 9.5 dpc (D-I), after cephalic neural tube closure, the domain of Wnt-8b expression (E,H) comes to lie at the caudomedial edge of the dorsal telencephalon. The expression of Wnt-7b (F,I) is broader than Wnt-8b, and includes much of the dorsal telencephalon. At both 8.5 dpc (A) and 9.5 dpc (D,G), Wnt-3a is expressed in the dorsal CNS with an anterior limit at the prosomere 2/ prosomere 3 (p2/p3) boundary in the anterior diencephalon. Wnt-3a expression initiates in the dorsal telencephalon at 9.75 dpc and at 10.5 dpc (J-L) Wnt-3a (J) and Wnt-8b (K) are expressed in nested domains at the caudomedial edge of the dorsal telencephalon. Wnt-8b expression extends further rostrocaudally and laterally than the domain of Wnt- 3a expression, which is limited to the presumptive cortical hem. (M- R) At 12.5 dpc, six Wnt genes are expressed in the dorsal cerebral cortex. Wnt-2b, Wnt-3a, Wnt-5a and Wnt-7b are expressed adjacent to the choroid plexus (CP) in the cortical hem (white arrowheads in M-P). Wnt-8b is expressed in a much broader domain along the medial wall of the telencephalic vesicle. A second site of Wnt-8b expression is the eminentia thalami, ventral to the CP. In all cases, Wnt gene expression is extremely low or undetectable in the CP. Wnt-7a is expressed broadly in the cortical neuroepithelium, but is excluded from the cortical hem. Between 10.5 dpc and 12.5 dpc Wnt- 7b expression in the cortical neuroepithelium turns off and a second phase of Wnt-7b expression is initiated in the cortical mantle layer (P). Abbreviations: CP, choroid plexus. Wnt signaling in hippocampal development 459

Fig. 2. Histological analysis of cerebral cortical development in 18.5 dpc Wnt-3a mutants. Sections were stained with haematoxylin and eosin. (A) In coronal sections through the rostral hippocampus of an 18.5 dpc wild-type brain, the hippocampus is located beneath the neocortex and bulges into the lateral ventricle from its medial edge. (B) In contrast, there is no morphological sign of hippocampal development in Wnt-3a mutant brains. At higher magnification, a dense pyramidal cell layer indicative of the CA fields and the subiculum is easily identifiable in the wild type (C), but not the Wnt-3a mutant (D). The most medial cortical tissue present in the mutant appears to be the cingulate and retrosplenial fields of the neocortex (D). (E,F) In horizontal sections, the subiculum, CA fields and dentate gyrus (DG) are identifiable in wild-type brains but absent in Wnt-3a mutants. Note that the pre/parasubiculum and entorhinal cortex lie between the caudal hippocampus and the neocortex in the wild type (E). In Wnt-3a mutants, the entorhinal cortex and a reduced pre/parasubiculum appears to be the most medial cortical structure present. (G) In coronal sections, immediately rostral to the rostral end of the hippocampus, the hippocampal commissure crosses the midline, ventral to the corpus callosum and dorsal to the anterior commissure. In Wnt-3a mutants, the hippocampal commisure is entirely absent and the axons that would normally cross the midline to form the corpus callosum fail to cross and instead form ectopic Probst bundles medially on either side of the midline. (H) In contrast the anterior commisure forms normally. Abbreviations: AC, anterior commisure; CA1, hippocampal field CA1; CA3, hippocampal field CA3; CC, corpus callosum; CP, choroid plexus; DG, dentate gyrus; Cing/Rsp, cingulate/retrosplenal neocortical fields; Ent, Entorhinal cortex; F, Fimbria; HC, hippocampal commisure; Ncx, neocortex; Pb, Probst bundle; pS, pre/parasubiclular complex; S, subiculum; Th, thalamus. sections were used for BrdU analysis). BrdU was detected during dorsal telencephalic development, we performed a immunohistochemically as described in Miller and Nowakowski comprehensive analysis of the expression of Wnt family (1988) and Dickinson et al. (1994) with the following modifications: members throughout the period of embryonic dorsal embryos were fixed in Bouin’s solution, PBS-based buffers were telencephalic development in the mouse, 8.0-18.5 days replaced with TBS-based buffers, and the proteinase K digestion postcoitum (dpc). Of 16 Wnt genes analyzed, six were found step was omitted. Labeled sections were counterstained with haematoxylin. 10.5 dpc embryonic heads were oriented so that to be expressed during this process; Wnt-2b, Wnt-3a, Wnt-5a, coronal sections would be cut through the dorsal telencephalon. Wnt-7a, Wnt-7b and Wnt-8b (Grove et al., 1998 and this study). Analysis of labeled sections was performed by first determining the Expression of Wnt-7b and Wnt-8b first appears in the dorsal 200 most medial nuclei of the cortical neuroepithelium on either side telencephalon at the 5- and 3-somite stages, respectively (S. M. of the roof epithelium (thus each section has two populations with 200 K. L. and A. P. M., unpublished data). By the 12-somite stage, nuclei each). Next the number of nuclei that had incorporated BrdU Wnt-7b and Wnt-8b are expressed in similar patterns, in the was determined (the mitotic index). To control for BrdU incorporation dorsoposterior region of the future dorsal telencephalon (Fig. differences between embryos, a population of 200 lateral cortical 1B,C). Note that the domain of Wnt-7b expression extends nuclei was also counted. Two sections from each of six wild-type and further anteriorly than that of Wnt-8b, while the expression of six Wnt-3a mutant embryos were analyzed in this manner, giving a both Wnts broadens to extend more medially (future ventrally) total of 24 data points in each case. The statistical significance of the data was determined using the Student’s t-test. at the posterior border of the telencephalon with the diencephalon (black arrowheads in Fig. 1B,C). At this time, Wnt-3a is expressed dorsally in the spinal cord, hindbrain, midbrain and diencephalon to an anterior boundary at the RESULTS presumptive prosomere 2 (p2) / prosomere 3 (p3) boundary (black arrowhead in Fig. 1A). Analysis of Wnt gene expression during dorsal At 9.5 dpc, Wnt-7b and Wnt-8b are expressed in nested telencephalic development domains in the dorsal telencephalon. Wnt-8b is expressed in a A number of Wnt genes are expressed during development of wedge of cells along the caudomedial edge of the dorsal the dorsal telencephalon in various vertebrate species (Roelink telencephalic vesicle (Fig. 1E,H) while Wnt-7b is broadly and Nusse, 1991; Parr et al., 1993; Cui et al., 1995; Hollyday expressed throughout the entire dorsal telencephalon (Fig. et al., 1995; Kelly et al., 1995; Grove et al., 1998; Lako et al., 1F,I). Expression of Wnt-3a still maintains an anterior limit at 1998). As a first step to addressing the role of Wnt signaling the p2/p3 boundary (Fig. 1D,G). Wnt-3a expression initiates in 460 S. M. K. Lee and others the dorsal telencephalon at 9.75 dpc (Roelink and Nusse, 1991; and the fimbria/fornix are all easily identifiable (Fig. 2A,C). In Parr et al., 1993). By 10.5 dpc, the expression domains of all contrast, none of these structures are identifiable in Wnt-3a three Wnt genes are nested with respect to the caudomedial mutants (Fig. 2B,D). The most medial cortical structure in the edge of the dorsal telencephalon. Wnt-3a is expressed in a mutant brain contains a broad cortical plate, characteristic of narrow band of cells immediately abutting the caudomedial neocortex and abuts the choroid plexus of the lateral ventricle edge (Fig. 1J). This domain of Wnt-3a expression is nested (Fig. 2D). Note that the choroid plexus of the lateral ventricles within a broader domain of Wnt-8b expression (Fig. 1K). Note always forms in Wnt-3a mutants but may be reduced in size that the domain of Wnt-8b expression extends further (Fig. 2A,B,E,F). In horizontal sections through the caudal rostrocaudally and mediolaterally than the domain of Wnt-3a end of the hippocampal formation of a control brain, the expression (compare Fig. 1J,K). Wnt-7b is expressed hippocampus and adjacent pre/parasubiculum and entorhinal throughout the entire cortical neuroepithelium by this time cortex are identifiable (Fig. 2E). (Note that the (Fig. 1L). These expression domains implicate Wnt signaling pre/parasubiculum and entorhinal cortex are only present at in the regulation of growth and patterning of the dorsal more caudal levels of the hippocampal longitudinal axis). In telencephalic neuroepithelium prior to neural differentiation, Wnt-3a mutant brains, similarly to more rostral levels, no which begins at 10 dpc in the dorsal telencephalon. hippocampal structures are identifiable (Fig. 2F). The adjacent At 12.5 dpc, six Wnt genes display restricted expression in multilayered entorhinal cortex, however, appears to be present the dorsal telencephalon. Expression of Wnt-2b, Wnt-3a and and possibly some subicular fields (Fig. 2F). Wnt-5a is restricted to a longitudinal band of cells at the Immediately adjacent to the rostral pole of the hippocampus, junction of the choroid plexus (CP) and the cortical the fimbria becomes the fornix and the commissural neuroepithelium, the cortical hem (Grove et al., 1998) (Fig. 1M- component of the fornix forms the hippocampal commissure O). Interestingly, the longitudinal axis of the cortical hem (Amaral and Witter, 1995) (Fig. 2G). The corpus callosum predicts the location of hippocampal development, at the crosses the midline immediately dorsal to the hippocampal caudomedial edge of the neocortex. Wnt-7a, in contrast, is commisure (Fig. 2G). In Wnt-3a mutants, both the expressed throughout the dorsal and lateral cerebral cortex but hippocampal commisure and the corpus callosum fail to form is specifically excluded from medial regions (Grove et al., 1998) (Fig. 2H), while a small ectopic bundle of axons (a Probst (Fig. 1R). Wnt-8b is expressed in a broad graded fashion along bundle) is present between the retrosplenial cortex and the the caudomedial wall of the dorsal cerebral cortex in a domain choroid plexus, which is likely formed by misrouted colossal that includes the cortical hem (Fig. 1Q). Finally, Wnt-7b axons (Fig. 2D). This suggests either that Wnt-3a signaling is expression in the ventricular zone is restricted to the cortical directly required for corpus callosum formation or that the hem by this time (Fig. 1P). Interestingly, as Wnt-7b expression hippocampus or hippocampal commissure provide a growth is turned off by ventricular zone cells expression is initiated in substratum and/or directional cues required for colossal axons the postmitotic cells which are forming the cortical mantle layer to cross the midline. (preplate in neocortical regions) (Fig. 1P; S. M. K. L. and A. P. The morphological absence of hippocampal development in M., unpublished data), suggesting a role for Wnt-7b in cortical Wnt-3a mutants could result from a failure to generate differentiation. Between 14.5 dpc and birth, as cortical hem hippocampal precursor cells or from the failure of hippocampal cells appear to differentiate into the ependymal cells that line precursors to develop into a hippocampus once they are the ventricular surface of the fimbria, they continue to express generated. To address whether hippocampal field-specific cell Wnt-2b, Wnt-3a, Wnt-5a, Wnt-7b and Wnt-8b (S. M. K. L. and types are formed in Wnt-3a mutants, and to define the defects A. P. M., unpublished data), suggesting the possibility that Wnt in the medial cortex with a cellular resolution, we used a panel signaling marks the future path of this axon tract. of region-specific molecular markers. Brains from mutant and In summary, expression of multiple Wnt genes marks the control mice were first analyzed at 15.5 dpc, the earliest age at dorsocaudal edge of the dorsal telencephalon from 8.5 dpc. which hippocampal subfields can be readily identified with Following closure of the rostral neuropore, this domain is molecular probes. In controls at this stage embryonic CA3 is identifiable as the caudomedial edge of the dorsal marked by KA1 expression (Fig. 3A) (Wisden and Seeburg, telencephalon, the cortical hem, and predicts the future site of 1993; Bahn et al., 1994; Tole et al., 1997) while CA1 and the hippocampal development at the margin of the cerebral cortex. DG can be identified by patches of Steel expression (Fig. 3C) (Motro et al., 1991; Tole et al., 1997). However, in the brains Hippocampal development requires Wnt-3a of 15.5 dpc Wnt-3a mutants, no CA1, CA3 or DG cells can be The expression of multiple Wnt genes in the cortical hem, identified by either marker (Fig. 3B,D). In both control and adjacent to the site of hippocampal development, suggests the mutant brains, a dense band of neocortical cells expresses Steel involvement of Wnt signaling in the regulation of hippocampal (Fig. 3C,D). Interestingly, in the mutant, this band of development. As Wnt-3a is the earliest Wnt gene to be neocortical Steel expression extends almost to the edge of the expressed exclusively in the cortical hem, between 9.75 and cortex, a further indication that little or no hippocampus is 11.5 dpc, we began to address this issue by analyzing cerebral present. cortical development in Wnt-3a mutant mice (Takada et al., Brains were next analyzed at 18.5 dpc. At this age, more 1994). Histological sections through 18.5 dpc Wnt-3a mutant hippocampal molecular markers are expressed and each brains were examined to assess hippocampal development in hippocampal subfield is larger. For both reasons, any residual the absence of Wnt-3a. In coronal sections through the rostral hippocampal cells in the mutant brain should be easier to hippocampal formation of wild-type brains the pyramidal cell identify. In control brains, CA3 is identifiable by KA1 layer of the subiculum and the CA1 and CA3 fields, as well as expression, CA1 and the DG express Steel, and the subiculum the dentate granular cells which are beginning to form the DG expresses BIG-1 (Fig. 3E,G,I) (Motro et al., 1991; Wisden and Wnt signaling in hippocampal development 461

Seeburg, 1993; Bahn et al., 1994; Yoshihari et al., 1994; Tole progenitor cell populations in the medial wall: Msx-1 marks et al., 1997). In some mutant brains, a small population of CA3 the choroid plexus, the cortical hem is Wnt-3a-positive and Bf- cells is now detectable by KA1 expression (Fig. 3F), but no DG 1-negative, and the domain of graded Wnt-8b expression marks cells are ever detectable by Steel expression (Fig. 3H). The a broad domain of caudomedial wall neuroepithelium. In Wnt- subiculum and CA1 remain undetectable at most rostrocaudal 3a mutants, the choroid plexus expresses Msx-1 despite its levels through the mutant brain (Fig. 3H,J). However, very abnormal morphogenesis (Fig. 4H). Within the caudally, at a level at which CA1 and the subiculum normally neuroepithelium, complementary Wnt-3a and Bf-1 expression appear at their largest in coronal sections, small patches of marks the cortical hem (Fig. 4D,F), which is somewhat reduced Steel-positive CA1 and Big1-positive subicular cells can be in size relative to wild type (compare with Fig. 4C and E). identified (data not shown). Finally, a small pre/parasubicular (Note that Wnt-3a expression can be detected in Wnt-3a−/− complex is evident in the mutant by expression of Steel and embryos using a probe that hybridizes to sequences outside of NT3 (data not shown). By contrast with the gross shrinkage the disrupted region.) Thus the cortical hem and choroid plexus or absence of all hippocampal fields in the mutant, NT3 are formed and maintained in the absence of Wnt-3a signaling. expression, which marks the adjacent cingulate and In contrast, the graded medial wall expression of Wnt-8b is retrosplenial cortices (Freidman et al., 1991; S. T. and E. G., absent and only the cortical hem expression remains (Fig. 4J). unpublished observations), reveals that these neocortical areas Thus a large domain of caudomedial cerebral cortical neural are strikingly close to normal in size (compare Fig. 3K and L), progenitor cells, abutting the cortical hem, is absent by 12.5 given the overall slight shrinkage of the mutant cerebral cortex. dpc. This implies that Wnt-3a acts prior to this stage and is In summary, both morphological and molecular marker required for either the specification or maintenance/expansion analysis reveal a severe reduction in hippocampal development of this progenitor cell pool. in Wnt-3a mutants. Interestingly, the severity of the defects are To determine the exact onset of the loss of caudomedial graded along both the mediolateral and longitudinal axes of the cortical progenitors in Wnt-3a mutants, we performed a archicortex. At the medial edge of the hippocampal formation histological analysis of dorsal telencephalic development the DG is entirely absent at all longitudinal levels. In contrast, between 9.75 dpc (the time of initial Wnt-3a expression in CA3, CA1 and the subiculum are absent rostrally and severely the dorsal telencephalon) and 11.5 dpc. Up to 10.5 dpc, reduced caudally. At the lateral edge of the hippocampal development of the mutant dorsal telencephalon is formation the pre/parasubicular complex, which is only present indistinguishable from wild type (data not shown). However, caudally in wild-type mice, is reduced but present in Wnt-3a by 10.75 dpc, there is a dramatic shortening of the caudomedial mutants. This severe reduction in hippocampal development cortical neuroepithelium in Wnt-3a mutants relative to controls suggests that an early defect within the caudomedial cortical (compare Fig. 5B,F with A,E). neuroepithelium from which the hippocampus develops may In summary, the caudomedial cortical neuroepithelium of underlie the mutant phenotype. Wnt-3a mutants fails to undergo a proper morphological elongation and this results in a significant reduction in the size The caudomedial cerebral cortical wall is shortened of the caudomedial cortical wall. While the choroid plexus and in Wnt-3a mutants the cortical hem neuroepithelial populations are specified In 12.5 dpc wild-type brains, the caudomedial wall of the and maintained, a large population of caudomedial neural cerebral cortex forms a characteristic S-shaped curve, which progenitors, which normally expresses Wnt-8b, is missing by presages the future morphology of the hippocampal formation 12.5 dpc, implying a requirement for Wnt-3a signaling to either (Fig. 4A). In contrast, there is a severe shortening of the specify or maintain/expand this neuroepithelial cell population. caudomedial wall archicortical neuroepithelium in Wnt-3a mutant brains (Fig. 4B). As a result, the choroid plexus (CP), Wnt-3a is required to maintain neural progenitor cell which forms adjacent to the archicortical neuroepithelium and proliferation in the caudomedial cortical loops into the lateral ventricle at this time, is pulled taut in the neuroepithelium mutant (Fig. 4A,B). Wnt-3a could act to maintain/expand neural progenitor cells The shortening of the medial telencephalic wall in Wnt-3a via at least three distinct mechanisms: as a mitogenic factor mutants suggests either that specific medial neural progenitor required to stimulate proliferation, as a survival factor required cell populations are entirely missing or alternatively that these to inhibit apoptosis or as an antidifferentiation factor required populations are present but significantly reduced in size. In to maintain neural progenitors in a proliferative, order to distinguish between these alternatives, we analyzed the undifferentiated state. We addressed each of these possibilities expression of a panel of genes whose expression marks specific in turn. domains of the medial wall neuroepithelium and the choroid To ask whether Wnt-3a is required for neural progenitor cell plexus at 12.5 dpc. The cortical hem is marked by the proliferation, we directly assayed cell proliferation in the expression of Wnt-3a (Fig. 4C) (Roelink and Nusse, 1991; telencephalic neuroepithelium at 10.75 dpc, at the onset of Grove et al., 1998) and the absence of Bf-1 expression (Furuta the caudomedial neuroepithelial defect, by monitoring the et al., 1997) (Fig. 4E). Msx-1 expression marks the choroid incorporation of bromodeoxyuridine (BrdU) into S-phase plexus (Furuta et al., 1997) (Fig. 4G), which forms ventral to nuclei. To quantify the proportion of neuroepithelial cells that the cortical hem at the medial border of the cerebral cortical were proliferating, we first defined the 200 neuroepithelial cell neuroepithelium, and also the very most ventral portion of the nuclei that were closest to the roof (this is approximately the cortical hem itself. Finally, Wnt-8b is expressed in a broad population of cells that lies between the bars in Fig. 5E,F and domain of the medial wall neuroepithelium (Fig. 4I). Thus the corresponds approximately to the domain of Wnt-3a expression of these markers unambiguously defines three expression at this time). Note that, in the mutant, this 462 S. M. K. Lee and others

Fig. 4. Expression of markers differentially expressed in the telencephalic medial wall at 12.5 dpc. (A) At 12.5 dpc, the cortical hem is located dorsal to the choroid plexus and adjacent to the presumptive archicortex. (B) In Wnt-3a mutants, the medial wall neuroepithelium is truncated and the choroid plexus is pulled taut. In wild-type brains the cortical hem is marked by the expression of Wnt-3a (white arrowhead in C) and the absence of Bf-1 expression (white arrowhead in E), while the adjacent choroid plexus is marked by Msx-1 expression (G). In Wnt-3a mutants the morphologically abnormal choroid plexus expresses Msx-1 (H), and is flanked by a slightly reduced population of cortical hem cells which is Wnt-3a+ and Bf-1- (white arrowheads in D,F). Wnt-8b is normally expressed highly in the cortical hem (white arrowheads in I,J) and in a ventral to dorsally decreasing gradient along the medial wall (I); however, in Wnt-3a mutants, only the cortical Fig. 3. Expression of hippocampal field-specific markers in coronal hem portion of Wnt-8b sections through wild-type and Wnt-3a mutant brains. (A-D) At 15.5 expression remains (J). dpc, CA3 is identifiable by KA1 expression in the wild type (A), but Wnt-8b expression in the not the Wnt-3a mutant (B). CA1 and the dentate gyrus (DG) can be diencephalon (I) is normal identified by patches of Steel expression in the wild type (C) but not in Wnt-3a mutants (J), the Wnt-3a mutant (D). A dense line of Steel expression (C,D) marks consistent with the normal neocortex in wild type and Wnt-3a mutants alike. (E-L) At 18.5 dpc, development of this region in wild-type brains the neocortex, CA1, and the DG are identifiable in the mutant. by Steel expression (G) while the subiculum and CA3 are marked by Abbreviations: Acx, Big1 (I) and KA1 (E) expression, respectively. In Wnt-3a mutants, a Archicortex; CH, cortical small residual patch of CA3 cells marked by KA1 expression is hem; CP, choroid plexus; present (F) but the CA1 and DG domains of Steel expression (H) and MW, medial wall. the subicular domain of Big1 expression (J) are absent. Neocortical domains of Steel (G) and Big1 (I) expression appear medially shifted in Wnt-3a mutants (H,J) as a result of the hippocampal loss. NT3 labeled cells in the medial (morphologically abnormal) wall of expression marks the cingulate and retrosplenial neocortical fields in the dorsal telencephalon of Wnt-3a mutants as compared to both wild-type (K) and Wnt-3a mutant (L) brains. Abbreviations: wild-type littermates (Student’s t-test, P<0.0001) (Fig. 5C-F). CA1, hippocampal field CA1; CA3, hippocampal field CA3; DG, In contrast, there was no significant difference in the proportion dentate gyrus; Cing/Rsp, cingulate/retrosplenal neocortical fields; Ncx, neocortex; Sub, subiculum. of labeled cells when a lateral control region (approximately the region between the bars in Fig. 5G,H) was compared between Wnt-3a mutant and wild-type littermates (Student’s t- population is within the morphologically abnormal region of test, P=0.41). Thus Wnt-3a plays a critical role in maintaining the caudomedial neuroepithelium. We next determined the neural progenitor cell proliferation in the caudomedial dorsal number of nuclei, out of these 200, that had incorporated BrdU telencephalon. (the mitotic index). Calculations were based on six wild-type We next asked whether Wnt-3a is required for neural and six Wnt-3a mutant embryos, scoring two sections per progenitor cell survival. While we routinely observed embryo. We observed a 26% reduction in the proportion of characteristically high levels of cell death in the telencephalic Wnt signaling in hippocampal development 463

Fig. 6. Analysis of neural differentiation in Wnt-3a mutants. 10.5 dpc wild-type (A,C) or Wnt-3a mutant (B,D) embryos were stained as whole mounts using probes against delta-1 (dl-1) (A,B) or class III Β-tubulin (Β-tub) (C,D). In all panels dorsal views of the cortex are shown and rostral is oriented down. dl-1 is expressed transiently by differentiating neurons as they leave the cell cycle, and expression appears in a punctate pattern throughout the cortex of both wild-type and Wnt-3a mutant embryos. Β-tub marks differentiated neurons and is expressed in the cortical mantle zone of both wild-type and Wnt-3a mutant embryos. For both markers there is a slight reduction in the level of expression at the caudomedial margin of the cortex in Wnt- 3a mutants (arrows in B,D).

Fig. 7. Analysis of regionally expressed markers in the dorsal telencephalon of 10.5- Fig. 5. Bromodeoxyuridine (BrdU) incorporation in Wnt-3a mutant 10.75 dpc Wnt-3a mutants. brains at 10.5 dpc. Sections were stained with haematoxylin and (A,C,E,G) Wild-type or eosin (A,B) or processed for immunohistochemical detection of (B,D,F,H) Wnt-3a mutant BrdU and counterstained with haematoxylin (C,D). Nuclei that have embryos at 10.5 dpc (A-D) or 10.75 dpc (E-H) were incorporated BrdU are brown while those that have not are light stained as whole mounts. In purple. C and D are near adjacent sections to those in A and B. E-H all panels, dorsal views of are high-power views of the regions indicated in C and D. In coronal the cortex are shown and sections through the telencephalon of wild-type embryos (A) the roof rostral is oriented down. (arrowheads in A and B) is flanked by the elongating caudomedial Wnt-3a is expressed in the cortical walls of the paired telencephalic vesicles. In Wnt-3a mutants presumptive cortical hem at (B), the roof epithelium is morphologically abnormal and the the caudomedial edge of caudomedial cortical walls are shortened (B). Note that in the the cerebral cortical morphologically abnormal medial cortical neuroepithelium of Wnt- neuroepithelium (A) in a 3a mutants there is a marked reduction in the proportion of nuclei domain which is nested which have incorporated BrdU relative to the similar region in within a broader domain of controls (E,F). In contrast, the proportion of nuclei incorporating Wnt-8b expression (C). BrdU in lateral cortical regions is indistinguishable between Wnt-3a Wnt-8b expression extends mutants and controls (G,H). further rostrocaudally and mediolaterally than Wnt-3a expression (compare C with A). In Wnt-3a roof of both control and mutant embryos, we never observed mutants, Wnt-3a is expressed normally in the any significant ectopic telencephalic cell death in the medial presumptive cortical hem neuroepithelium of mutants (data not shown). Therefore, (B). In contrast, there is a ectopic cell death does not contribute to the loss of significant reduction in the lateral extent of the Wnt-8b expression neuroepithelial cells in Wnt-3a mutants. domain (D), such that much of the expression beyond the cortical hem A third mechanism by which the loss of Wnt-3a signaling is lost (compare D with C). Bmp4 is expressed along the medial and could lead to a loss of neural progenitor cells is if Wnt-3a is dorsocaudal margin of the cortical neuroepithelium in both wild-type normally required to maintain neural progenitors in an (E) and Wnt-3a mutant (F) embryos. Finally, Bf-1 is expressed throughout the cortical neuroepithelium but is specifically excluded undifferentiated state. In this scenario, the absence of Wnt-3a from a caudomedial domain which approximately corresponds to the would lead to the premature differentiation of neural domain of Wnt-3a expression in both wild-type (G) and Wnt-3a mutant progenitors into neurons, which would have the effect of (H) embryos. 464 S. M. K. Lee and others prematurely depleting the progenitor cell pool and result in too dorsolateral cortex, BMP-coated beads induce ectopic few neurons being generated. In order to investigate this apoptosis and Msx-1 expression, while inhibiting Bf-1 possibility, we analyzed the expression of the neurogenic gene expression, suggesting a role for BMP signaling in the delta-1 (dl-1), which is transiently expressed by neural induction of roof/choroid plexus and perhaps cortical hem cell precursor cells soon after they withdraw from the cell cycle fates and/or the inhibition of dorsolateral cortical cell fates (Bettenhausen et al., 1995), and the neural-specific class III Β- (Furuta et al., 1997). To address the possibility that some of tubulin (Β-tub), a marker of differentiating neurons (Burgoyne the defects in Wnt-3a mutants are mediated by a reduction in et al., 1988; Grove et al., 1998). Both genes are expressed in a BMP signaling, we assayed Bmp4 and Bf-1 expression in the reticular pattern throughout the mantle zone of the dorsal dorsal telencephalon of 10.75 dpc mutant and control embryos. telencephalon at 10.75 dpc (Fig. 6A,C). In Wnt-3a mutants, the In both mutant and control embryos, BMP-4 expression is cortical expression of both dl-1 and Β-tub is normal or perhaps limited to the telencephalic roof and midline head mesenchyme slightly reduced at the caudomedial edge (arrows in Fig. 6B,D), (Fig. 7E,F) and Bf-1 is expressed throughout the cortex but is where the hippocampal primordium lies. This result argues excluded from a caudomedial domain that includes the roof against a requirement for Wnt-3a to suppress neuronal and the presumptive cortical hem (arrows in Fig. 7G,H). differentiation. Further, in Wnt-3a mutants cell death is normal and Msx-1 In summary, while we failed to observe an increase in cell expression in the choroid plexus is unaltered (data not shown death or any premature neuronal differentiation in Wnt-3a and Fig. 4H). These observations indicate that some patterning mutants, we did observe a significant reduction in neural events that are presumed to be downstream of BMP signaling progenitor cell proliferation specifically in the region of Wnt- occur normally in Wnt-3a mutants. 3a signaling. Thus, the loss of neural progenitor cells in the caudomedial cortical neuroepithelium of Wnt-3a mutants is at least partially the result of a decreased level of proliferation in DISCUSSION these cells. The patterning of the vertebrate brain occurs through the Wnt-3a is required for the expansion of the stepwise subdivision of the neuroepithelium into an orthogonal caudomedial cortical wall, the site of the grid of regional domains, which are the primordia of adult hippocampal primordium anatomical structures (Puelles and Rubenstein, 1993; The observed reduction in caudomedial cortical progenitor cell Rubenstein et al., 1994; Shimamura et al., 1995; Lumsden and proliferation at 10.75 dpc and the subsequent truncation of the Krumlauf, 1996; Rubenstein et al., 1998). Within each regional caudomedial neuroepithelium by 11.5 dpc implies that Wnt-3a domain, a specific and highly complex array of neural cell driven proliferation is necessary for the expansion of a pool types are generated. The processes of regionalization and cell of caudomedial progenitor cells, which normally form the type specification appear to be mediated by the action of caudomedial cortical wall. To investigate this issue, we inductive signaling factors (Lumsden and Krumlauf, 1996; analyzed the expression of Wnt-3a and Wnt-8b between 10.5 Tanabe and Jessell, 1996; Rubenstein et al., 1998). It is thus and 10.75 dpc in Wnt-3a mutant and control embryos. Between the local interaction of these inductive factors with neural 10.5 and 10.75 dpc, the nested expression of Wnt-3a and Wnt- progenitor cells that determines the spatial relationship 8b in the caudomedial cortical neuroepithelium defines two between different anatomical structures and the types of neural adjacent populations of progenitor cells in controls (Fig. cells contained within each structure of the adult brain. An 7A,C). A domain of cells abutting the caudomedial cortical intriguing problem is that of how growth is controlled in edge expresses both Wnt-3a and Wnt-8b, while more lateral coordination with these primary patterning processes in order cells express only Wnt-8b. The Wnt-8b-expressing sector thus to generate anatomical structures of the appropriate size made contains a broad region of caudomedial cortex, which, by up of the appropriate numbers of each cell type. Our data argue its position, is highly likely to include the hippocampal that Wnt-3a is required for the expansion of a caudomedial primordium. In Wnt-3a mutants, nested domains of Wnt-3a and cortical progenitor pool that generates the hippocampus. We Wnt-8b expression are indistinguishable from controls up to suggest that it is the coordination of Wnt-3a-mediated 10.5 dpc (data not shown). However, beginning at 10.75 dpc, proliferation with hippocampal cell fate specification that is there is a marked reduction in the lateral extent of the Wnt-8b responsible for the generation of a normal hippocampus. expression domain (Fig. 7D), while the medial domain of Wnt- 3a and Wnt-8b co-expression remains intact (Fig. 7B,D). These A local Wnt-3a signal is required for the expansion observations suggest that, while the cortical hem is maintained of a population of hippocampal progenitor cells in in Wnt-3a mutants, the hippocampal primordium is greatly the caudomedial wall of the cerebral cortex reduced by 10.75 dpc. In Wnt-3a mutants, the initiation of dorsal telencephalic development, as indicated by the establishment of Bmp-4 BMP-mediated dorsal patterning events do not expression in the roof, and the specification of caudomedial require Wnt-3a cortical cell fates, as indicated by the establishment of The Wnts are not the only secreted factors with precise complementary Wnt-3a and Bf-1 expression domains, occurs domains of expression within this area of the brain. Beginning normally. However, our data indicate that Wnt-3a is required at 8.0 dpc multiple BMP-encoding genes are expressed in the to promote the proliferation of cortical progenitor cells at the dorsal telencephalon in domains that variably include the caudomedial margin of the cerebral cortical neuroepithelium. roof/choroid plexus and the cortical hem (Furuta et al., 1997; In Wnt-3a mutants, this region shows a marked decrease in Grove et al., 1998). In cultured explants of 10.5 dpc proliferation, which most likely accounts for the failure in Wnt signaling in hippocampal development 465 expansion of the caudomedial wall of the cerebral cortex. known as DWnt-1) in the wing pouch is required for local Interestingly, the shortened caudomedial wall of mutants proliferation in the wing hinge (Neumann and Cohen, 1996), maintains a domain of Wnt-3a-expressing cells along its while local wg expression in the Malpighian tubule anlage is margin but is missing a large domain of rostrolaterally adjacent required for proper elongation of the Malpighian tubules progenitors, which normally express Wnt-8b. This implies that (Skaer and Martinez Arias, 1992). Conversely, the ectopic the domain of Wnt-3a-expressing cells, which require Wnt-3a expression of wg in either of these tissues leads to local for their proliferation at 10.75 dpc, normally grows to generate overgrowth but does not affect primary patterning (Skaer and the caudomedial wall of the cortical neuroepithelium and that Martinez Arias, 1992; Neumann and Cohen, 1996). Thus the over time caudomedial cortical progenitor cells, which initially local regulation of proliferation may be a general role for Wnt express Wnt-3a, expand to occupy more rostrolateral positions signaling during pattern formation. and turn off Wnt-3a transcription. The initiation of dorsal telencephalic Wnt-3a expression at Implications for the mechanism of hippocampal field 9.75 dpc and the subsequent initiation of hippocampal neuronal specification birth at 10.5 dpc (Angevine, 1965; Stanfield and Cowan, 1988) Cell lineage studies have indicated that the field identity of suggests that Wnt-3a signaling is associated with the hippocampal neurons is not determined by cell lineage, arguing production of hippocampal neurons. The presence of small that hippocampal fields are not generated by restricted pools of residual clusters of CA1, CA3 and subicular cells at the progenitor cells that are permanently specified or committed to caudomedial margin of the Wnt-3a mutant cortex suggests that produce neurons of a single field (Grove et al., 1992; Walsh hippocampal cells are specified in the mutant but that the and Cepko, 1992). Clues as to the timing of CA field failure of Wnt-3a driven hippocampal progenitor cell specification come from the observation that molecular expansion results in a depletion of the reduced progenitor cell markers of CA1 and CA3 (SCIP and KA1, respectively) are pool once hippocampal neuronal birth begins. Thus, it may be first expressed in the presumptive CA1 and CA3 (Pca1 and the coordination of the production of a locally acting mitogenic Pca3) fields at 15.5 dpc (Tole et al., 1997). In slice cultures of signal, Wnt-3a, with the primary patterning processes Pca1 or Pca3 isolated at 17.5 dpc, later aspects of mature CA responsible for inducing hippocampal development and field identity develop autonomously (Tole et al., 1997). The specifying hippocampal cell types that determines the size of origin of these field-specifying signals is suggested by the the hippocampus. observation that the expression of SCIP in Pca1 and KA1 in The local regulation of proliferation may be a general role Pca3 initiates in presumptive CA pyramidal cells at the for Wnt signaling during vertebrate CNS development. Studies subicular (lateral) and dentate (medial) poles of the of dorsal hindbrain and spinal cord development have presumptive CA region, respectively. Over a period of 3-4 implicated BMP and Wnt signaling in the regulation of dorsal days, fronts of SCIP and KA1 expression gradually sweep CNS patterning (Dickinson et al., 1994; Liem et al., 1995, across the hippocampal primordia until the expression of these 1997; Arkell and Beddington, 1997; Ikeya et al., 1997; Lee et two markers complementarily defines the Pca1 and Pca3 fields al., 1998). This process is initiated by BMP signaling from the in their entirety (Tole et al., 1997). Together, these observations non-neural ectoderm, which induces neural crest formation imply that signals from the medial and lateral poles of the from early dorsolateral neural progenitors concomitantly with hippocampus are responsible for specifying CA field identity the induction of roof plate (Liem et al., 1995). BMP signaling and that these signals act on presumptive CA neurons prior to from the roof plate subsequently acts to specify the 15.5 dpc. differentiation of dorsal neuronal cell types from dorsolateral The expression of multiple Wnt genes in the cortical hem neural progenitors (Liem et al., 1997). Wnt-1 and Wnt-3a are (Grove et al., 1998 and this study), adjacent to the medial edge expressed in the dorsal CNS from 8.0 dpc (Wilkinson et al., of the developing hippocampus, together with the known 1987; Roelink and Nusse, 1991; Hollyday et al., 1995) and ability of Wnt signals to specify cell fate choices acting as both their expression can be ectopically induced in intermediate local inducers and long-range gradient morphogens in a neural explants following recombination with non-neural number of developmental contexts (eg.: Vincent and Lawrence, ectoderm (Dickenson et al., 1995), suggesting that Wnt-1/3a 1994; Zecca et al., 1996; Neumann and Cohen, 1997; Dorskey signaling acts downstream of BMP signaling in the dorsal et al., 1998; and see Cadigan and Nusse, 1997 and Moon et al., CNS. Ectopic expression and loss-of-function studies have 1997 for recent reviews), suggests the possibility that cortical identified a critical role for Wnt-1/3a signaling in the dorsal hem-derived Wnt signaling is involved in the specification hindbrain and spinal cord. Ectopic expression of Wnt-1 of medial hippocampal fields. In Wnt-3a mutants, the throughout the DV axis of the spinal cord leads to overgrowth hippocampal progenitor cell pool is dramatically reduced in of neural progenitors but does not affect the expression of size and this results in a severe reduction in hippocampal molecular markers of DV pattern (Dickinson et al., 1994). In development. However, small populations of cells expressing complementary studies, the analysis of mouse embryos mutant markers of CA1, CA3 and the subiculum can still be specified. for both Wnt-1 and Wnt-3a reveals a severe reduction in This is consistent with the idea that these fields are normally dorsolateral neural progenitors and a subsequent failure to specified by other signals originating from the subicular and generate many neural crest derivatives (Ikeya et al., 1997). dentate poles of the hippocampus (see above) (Tole et al., Together with our findings, these studies suggest that the 1997), and suggests that these signals are still active in Wnt-3a expansion of dorsal neural progenitor cells by Wnt-1/3a mutants. The dentate gyrus normally develops adjacent to the signaling is a common role for Wnt signaling during vertebrate cortical hem domain of Wnt-3a expression and we observed a CNS development. complete absence of dentate gyrus development in Wnt-3a In Drosophila, the local expression of wingless (wg) (also mutants. This is consistent with the possibility that Wnt-3a 466 S. M. K. Lee and others signaling normally specifies dentate gyrus cell fates. A a mitogenic effect of Wnt-1 in the developing mammalian central nervous definitive interpretation of this result is, however, difficult in system. Development 120, 1453-1471. the context of the dramatic reduction in overall hippocampal Dorskey, R. I., Moon, R. T. and Raible, D. W. (1998). Control of neural crest cell fate by the Wnt signalling pathway. Nature 396, 370-373. development and the fact that the timing of specification of Ebrahimi-Gaillard, A. and Roger, M. (1996). Development of spinal cord dentate gyrus cell fates is unknown. Experiments in which projections from neocortical transplants heterotypically placed in the Wnt-3a is ectopically expressed may help to clarify this issue. neocortex of newborn hosts in highly dependent on the embryonc locus of An additional implication of this observation is that the dentate origin of the graft. J. Comp. Neurol. 365, 129-140. Ferri, R. T. and Levitt, P. (1993). Cerebral cortical progenitors are fated to gyrus is not the source of signals responsible for specifying the produce region-specific neuronal populations. Cerebral Cortex 3, 187-198. identity of adjacent CA3 cells. Finally, expression of Wnt-2b, Freidman, W. J., Ernfors, P. and Patterson, H. (1991). Transient and Wnt-5a, Wnt-7b and Wnt-8b is maintained in the cortical hem persistent expression of NT-3/HDNF mRNA in the rat brain during postnatal of Wnt-3a mutants. Thus, it is possible that Wnt family development. J. Neurosci. 11, 1577-1584. members have redundant roles in the specification of Furuta, Y., Piston, D. W. and Hogan, B. L. M. (1997). Bone morphogenetic proteins (BMPs) as regulators of dorsal forebrain development. hippocampal fields. Development 124, 2203-2212. Gavrieli, Y., Sherman, Y. and Ben-Sasson, S. A. (1992). Identification of We thank Bianca Klumpar for histology and Pao-Tien Chuang for programmed cell death in situ via specific labeling of nuclear DNA valuable discussions and comments on an earlier version of the fragmentation. J. Cell Biol. 119, 493-501. manuscript. This work was supported by a grants from the NIH to A. Grove, E. A., Kirkwood, T. B. L. and Price, J. (1992). Neuronal precursor P. M and E. G. cells in the rat hippocampal formation contribute to more than one cytoarchitectonic area. Neuron 8, 217-229. Grove, E. A., Tole, S., Limon, J., Yip, Y. and Ragsdale, C. W. (1998). The hem of the embryonic cerebral cortex is defined by the expression of REFERENCES multiple Wnt genes and is compromised in Gli3-deficient mice. Development 125, 2315-2325. Amaral, D. G. and Witter, M. P. (1995). Hippocampal Formation. In The Rat Hollyday, M., McMahon, J. A. and McMahon, A. P. (1995). Wnt expression Nervous System, 2nd edition. (ed. G. Paxinos). pp. 443-493. San Diego, CA: patterns in chick embryo nervous system. Mech. Dev. 52, 9-25. Academic Press. Ikeya, M., Lee, S. M. K., Johnson, J. E., McMahon, A. P. and Takada, S. Angevine, J. B. (1965). Time of neuron origin in the hippocampal region. Exp. (1997). Wnt signalling required for expansion of neural crest and CNS Neurol. Supplement 2, 1-70. progenitors. Nature 389, 966-970. Arimatsu, Y., Miyamoto, M., Nihonmatsu, I., Hirata, K., Uratani, Y., Kelly, G. M., Greenstein, P., Erezyilmaz, D. F. and Moon, R. T. (1995). Hatanaka, Y. and Takaguchi-Hayashi, K. (1992). Early regional Zebrafish wnt8 and wnt8b share a common activity but are involved in specification for a molecular neuronal phenotype in the rat neocortex. Proc. distinct developmental pathways. Development 121, 1787-1799. Natl. Acad. Sci. USA 89, 8879-8883. Knecht, A. K., Good, P. J., Dawid, I. B. and Harland, R. M. (1995). Dorsal- Arkell, R. and Beddington, R. S. P. (1997). BMP-7 influences pattern and ventral patterning and differentiation of noggin-induced neural tissue in the growth of the developing hindbrain of mouse embryos. Development 124, absence of mesoderm. Development 121, 1927-1936. 1-12. Lako, M., Lindsay, S., Bullen, P., Wilson, D. I., Robson, S. C. and Strachen, Bahn, S., Volk, B. and Wisden, W. (1994). Kainate receptor gene expression T. (1998). A novel mammalian wnt gene, WNT8B, shows brain-restricted in the developing rat brain. J. Neurosci. 14, 5525-5547. expression in early development, with sharply delimited expression Barbe, M. F. and Levitt, P. (1991). The early commitment of fetal neurons boundaries in the developing forebrain. Human Molecular Genetics 7, 813- to the limbic cortex. J. Neurosci. 11, 519-533. 822. Barth, T. M. and Stanfield, B. B. (1994). Homotopic, but not heterotopic, Lee, K. J., Mendelsohn, M. and Jessell, T. M. (1998). Neuronal patterning fetal cortical transplants can result in functional sparing following neonatal by BMPs: a requirement for GDF7 in the generation of a discrete class of damage to the frontal cortex. Cerebral Cortex 4, 271-278. commissural interneurons in the mouse spinal cord. Genes Dev. 12, 3394- Bayer, S. A. (1980a). Development of the hippocampal region in the rat. II. 3407. Morphongenesis during embryonic and early postnatal life. J. Comp. Neurol. Levitt, P., Barbe, M. F. and Eagleson, K. L. (1997) Patterning and 190, 115-134. specification of the cerebral cortex. Annu. Rev. Neurosci. 20, 1-24. Bayer, S. A. (1980b). Development of the hippocampal region in the rat. I. Liem, K. F., Tremml, G. and Jessell, T. M. (1997). A role for the roof plate Neurogenesis examined with 3H-thymidine autoradiography. J. Comp. and its resident TGFB-related proteins in neuronal patterning in the dorsal Neurol. 190, 87-114. spinal cord. Cell 91, 127-138. Bettenhausen, B., Hrabe de Angelis, M., Simon, D., Guenet, J. L. and Liem, K. F., Tremml, G., Roelink, H. and Jessell, T. M. (1995). Dorsal Gossler, A. (1995). Transient and restricted expression during mouse differentiation of neural plate cells induced by BMP-mediated signals from embryogenesis of Dll1, a murine gene closely related to Drosophila Delta. epidermal ectoderm. Cell 82, 969-979. Development 121, 2407-2418. Lumsden, A. and Krumlauf, R. (1996). Patterning the vertebrate neuraxis. Bulfone, A., Martinez, S. S., Shimamura, K., Peterson, A., Puelles, L. and Science 274, 1109-1114. Rubenstein, J. L. R. (1995). T-Brain-1: a homolog of brachyury whose Miller, M. W. and Nowakowski, R. S. (1988). Use of bromodeoxyuridine expression defines molecularly distinct domains within the cerebral cortex. immunohistochemistry to examine the proliferation, migration, and time of Neuron 15, 63-78. origin of cells in the central nervous system. Brain Research 457, 44-52. Burgoyne, R. D., Cambray-Deakin, M. A., Lewis, S. A., Sarkar, S. and Moon, R., Brown, J. and Torres, M. (1997). WNTs modulate cell fate and Cowan, N. J. (1988). Differential distribution of beta-tubulin isotypes in behavior during vertebrate development. Trends in Genetics 13, 1571-1582. cerebrllum. EMBO J. 7, 2311-2319. Motro, B., van der Kooy, D., Rossant, J., Teith, A. and Berstein, A. (1991). Cadigan, K. M. and Nusse, R. (1997). Wnt signaling: a common theme in Contiguous patterns of c-kit and steel expression: analysis of muations at animal development. Genes Dev. 11, 3286-3305. the W and Sl loci. Development 113, 1207-1221. Cohen-Tannoudji, M., Babinet, C. and Wassef, M. (1994). Early Neumann, C. J. and Cohen, S. M. (1996). Distinct mitogenic and cell fate determination of a mouse somatosensory cortex marker. Nature 368, 460- specification functions of wingless in different regions of the wing. 463. Development 122, 1781-1789. Cui, Y., Brown, J. D., Moon, R. T. and Christian, J. L. (1995). Xwnt-8b: a Neumann, C. J. and Cohen, S. M. (1997). Long-range action of wingless maternally expressed Xenopus Wnt gene with a potential role in establishing organizes the dorsal-ventral axis of the drosophila wing. Development 124, the dorsoventral axis. Development 121, 2177-2186. 871-880. Dickinson, M. E., Selleck, M. A. J., McMahon, A. P. and Bronner-Fraser, O’Leary, D. D. M., Schlaggar, B. L. and Tuttle, R. (1994). Specification of M. (1995). Dorsalization of the neural tube by the non-neural ectoderm. neocortical areas and thalamocortical connections. Annu. Rev. Neurosci. 17, Development 121, 2099-2106. 419-439. Dickinson, M. E., Krumlauf, R. and McMahon, A. P. (1994). Evidence for Parr, B. P., Shea, M. J., Vassileva, G. and McMahon, A. P. (1993). Mouse Wnt signaling in hippocampal development 467

Wnt genes exhibit discrete domains of expression in the early embryonic Takada, S., Stark, K. L., Shea, M. J., Vassileva, G., McMahon, J. A. and CNS and limb buds. Development 119, 247-261. McMahon, A. P. (1994). Wnt-3a regulates somite and tailbud formation in Puelles, L. and Rubenstein, J. L. R. (1993). Expression patterns of homeobox the mouse embryo. Genes Dev. 8, 174-189. and other putative regulatory genes in the embryonic mouse forebrain Tanabe, Y. and Jessell, T. M. (1996). Diversity and pattern in the developing suggest a neuromeric organization. Trends in Neurosci. 16, 472-479. spinal cord. Science 274, 1115-1123. Roelink, H. and Nusse, R. (1991). Expression of two members of the Wnt Tole, S., Christian, C. and Grove, E. A. (1997). Early specification and family during mouse development Ð resrticted temporal and spatial patterns autonomous development of cortical fields in the mouse hippocampus. in the developing neural tube. Genes Dev. 5, 381-388. Development 124, 4959-4970. Rubenstein, J. L. R., Martinez, S., Shimamura, K. and Puelles, L. (1994). Vincent, J. and Lawrence, P. A. (1994). Drosophila wingless sustains The embryonic vertebrate forebrain: the prosomeric model. Science 266, engrailed expression only in adjoining cells: evidence from mosaic embryos. 578-580. Cell 77, 909-915. Rubenstein, J. L. R., Shimamura, K., Martinez, S. and Puelles, L. (1998). Walsh, C. and Cepko, C. L. (1992). Widespread dispersion of neuronal clones Regionalization of the prosencepalic neural plate. Annu. Rev. Neurosci. 21, across functional regions of the cerebral cortex. Science 255, 434-440. 445-477. Wilkinson, D. G., Bailes, J. A. and McMahon, A. P. (1987). Expression of Schlaggar, B. L. and O’Leary, D. D. (1991). Potential of visual cortex to the proto-oncogene int-1 is restricted to specific neural cells in the develop an array of functional units unique to somatosensory cortex. Science developing mouse embryo. Cell 50, 79-88. 252, 1556-1560. Wisden, W. and Seeburg, P. H. (1993). A complex mosaic of high-affinity Shimamura, K., Hartigan, D. J., Martinez, S., Puelles, L. and Rubenstein, dainate receptors in rat brain. J. Neurosci. 13, 3582-3598. J. L. L. (1995). Longitudinal organization of the anterior neural plate and Yoshihari, Y., Kawasaki, M., Tani, A., Tamada, A., Nagata, S., neural tube. Development 121, 3923-3933. Kagamiyama, H. and Mori, K. (1994). BIG-1: a new TAG-1/F3-related Skaer, H. and Martinez Arias, A. (1992). The wingless product is required member of the immunoglobulin superfamily with neurite outgrowth- for cell proliferation in Malpighian tube anlage of Drosophila melanogaster. promoting activity. Neuron 13, 415-426. Development 116, 745-754. Yoshikawa, Y., Fujimori, T., McMahon, P. P. and Takada, S. (1997). Stanfield, B. B. and Cowan, W. M. (1979a). The development of the Evidence that absence of Wnt-3a signaling promotes neuralization hippocampus and dentate gyrus in normal and reeler mice. J. Comp. Neurol. instead of paraxial mesoderm development in the mouse. Dev. Biol. 183, 185, 423-460. 234-242. Stanfield, B. B. and Cowan, W. M. (1979b). The morphology of the Zecca, M., Basler, K. and Struhl, G. (1996). Direct and long-range action of hippocampus and dentate gyrus in normal and reeler mice. J. Comp. Neurol. a wingless morphogen gradient. Cell 87, 833-844. 185, 393-422. Zilles, K. and Wree, A. (1995). Cortex: Areal and Laminar Structure. In The Stanfield, B. B. and Cowan, W. M. (1988). The development of the Rat Nervous System, 2nd ed., (ed. G. Paxinos). pp. 649-685. San Diego, CA: hippocampal region. Cerebral Cortex 7, 91-131. Academic Press.