R ESEARCH A RTICLES Cadherin-8 (Cdh8), in P6 brain whole mounts. At this age, differential expression of Neocortex Patterning by the Cdh6 and Cdh8 distinguishes a frontal do- main composed of cingulate, prefrontal, and Secreted Signaling Molecule FGF8 motor areas; a parietal domain that contains somatosensory areas; and an occipital domain Tomomi Fukuchi-Shimogori and Elizabeth A. Grove* that includes visual areas (5, 6, 25) (Fig. 3, A to C). In Fgf8-electroporated P6 left hemi- A classic model proposes that the mammalian neocortex is divided into areas spheres (n ϭ 12), the frontal domain is ex- early in neurogenesis, but the molecular mechanisms that generate the area panded at the apparent expense of parietal map have been elusive. Here we provide evidence that FGF8 regulates devel- and occipital domains, which are shrunken opment of the map from a source in the anterior telencephalon. Using elec- and shifted back (Fig. 3, B and D). Thus, troporation-mediated gene transfer in mouse embryos, we show that aug- consistent with the hypothesis that FGF8 reg- menting the endogenous anterior FGF8 signal shifts area boundaries posteri- ulates pattern along the A/P axis, augmenting orly, reducing the signal shifts them anteriorly, and introducing a posterior the endogenous FGF8 source results in an source of FGF8 elicits partial area duplications, revealed by ectopic somato- expansion of an anterior neocortical domain sensory barrel fields. These findings support a role for FGF signaling in specifying with a concomitant shifting and shrinkage of positional identity in the neocortex. more posterior areas. Area boundary shifts are not due to a The mammalian cerebral cortex is divided direct gene misexpression to select sites in a simple growth effect. FGF8 can regulate into anatomically and functionally distinct single cerebral hemisphere, we adapted the cell proliferation in vivo and shows trans- areas, forming a species-specific area map method of microelectroporation (19, 20) for forming potential in vitro (15, 26–28). How- across the cortical sheet (1). Identifying the gene transfer in mice in utero (Fig. 1A). Mice ever, the expansion of anterior neocortex seen mechanisms that generate the map is thus key are born normally and can be analyzed at any with FGF8 overexpression is not a simple to understanding the development of cortical age, making this method a useful adjunct to growth effect. Although the anterior domain function and may clarify how different maps the generation of genetically engineered expands, more posterior domains contract, so are generated in different species. In a classic mice, which often do not survive past birth. that Fgf8-electroporated hemispheres do not model, an area “protomap” is set up in the Expanding the anterior FGF8 source show gross overall increases in A/P length proliferative cell layer of the neocortex (2). shifts cortical area boundaries posterior- compared with control hemispheres (Fig. 2, Recently, it has been proposed that the pro- ly. FGF8 was initially overexpressed in the A to F, and Fig. 3, B and D). By contrast, tomap could be specified by signaling pro- anterior cortical primordium, just posterior to anterior overexpression of another growth teins secreted from nearby signaling centers, the endogenous source (Fig. 1C). We predict- factor, WNT3A, implicated in hippocampal a patterning strategy used elsewhere in the ed that augmenting the endogenous FGF8 cell proliferation (29), expands the frontal embryo (3–7). Candidate sources have been signal in this way would distort the area map Cdh8-expressing cortical domain by causing identified of proteins implicated in vertebrate along the A/P axis. Embryos were electropo- a marked overgrowth at the frontal pole of the and invertebrate embryonic patterning, in- rated at embryonic day 11.5 (E11.5)—early hemisphere (n ϭ 5) (Fig. 3G). Anterior over- cluding members of the fibroblast growth in neocortical neurogenesis, before neocorti- expression of FGF8 appears, instead, to shift factor (FGF), Wingless-Int (WNT), and bone cal area identity is determined (21–23)—and the position of areas within the hemisphere. morphogenetic protein (BMP) families (8– analyzed postnatally. At postnatal day 0 (P0), Reducing the endogenous FGF8 sig- 11). In this study we sought direct evidence several neocortical gene-expression patterns nal shifts cortical area boundaries ante- that such a patterning strategy is used to indicate emerging area boundaries along the riorly. To test whether an endogenous FGF generate the neocortical area map. A/P axis, although true cytoarchitectonic signal coordinates the area map, we ex- Patterning roles for the FGF family mem- boundaries are not yet visible. EphrinA5 en- pressed a soluble form of FGFR3c ber FGF8 have been reported for the first codes an Eph ligand and is expressed most (sFGFR3) close to the anterior FGF8 branchial arch, the midbrain, and the initial strongly in presumptive somatosensory cor- source. FGFR3c is a high-affinity FGF8 formation of the telencephalon (10, 12–16). tex; sFrp2 encodes secreted frizzled related receptor isoform, and the soluble form is Indicating that FGF8 could also be a regula- protein 2 and is expressed anterior to Eph- expected to sequester endogenous FGF8, tor of anterior/posterior (A/P) neocortical pat- rinA5; and Rzr-beta encodes an orphan nu- and potentially other FGF family members, tern, FGF8 is expressed close to the anterior clear receptor and is expressed in both do- blocking their ability to activate endoge- pole of the neocortical primordium (4, 8, 17) mains (4, 5, 24) (Fig. 2, A to C). As predict- nous receptors (14, 30, 31). sFGFR3-elec- (Fig. 1B), and the primordium itself shows ed, in Fgf8-electroporated hemispheres (n ϭ troporated hemispheres show no gross de- A/P graded expression of genes encoding 6), these expression domains shift in a coor- crease in overall size, but the frontal Cdh8- FGF receptors FGFR1, 2, and 3 (18). To test dinated manner toward the posterior pole of expressing cortical domain shrinks (n ϭ 7) this hypothesis, we analyzed the effects on the cortex (Fig. 2, D to F). By P6 in the (Fig. 3F) and other gene expression do- the area map of augmenting the anterior mouse, a true area boundary, defined by cy- mains shift anteriorly (32). These observa- FGF8 source in the embryonic mouse cere- toarchitecture (1), appears between primary tions and others cited below indicate that an brum, sequestering endogenous FGF8 with a somatosensory and motor areas (Fig. 2G). In endogenous FGF signal regulates neocorti- soluble FGF receptor construct, or introduc- Fgf8-electroporated hemispheres analyzed at cal pattern and that it is still active at E11.5. ing a second, posterior source of FGF8. To P6 (n ϭ 10), this area boundary, also distin- A signal feature of neocortical sensory guished by transitions in gene expression (4), and motor areas is that they contain topo- is shifted posteriorly (Fig. 2, J to L). graphic, functional representations of the Department of Neurobiology, Pharmacology and Physiology, University of Chicago, Chicago, IL 60637, For a global view of the changes in the body. We examined the effects of augment- USA. area map caused by anterior FGF8 overex- ing or reducing FGF8 signaling on this fea- *To whom correspondence should be addressed. E- pression, we examined expression of the type ture of area identity, focusing on the modular mail: [email protected] II classic cadherins, Cadherin-6 (Cdh6) and organization of primary somatosensory cor- www.sciencemag.org SCIENCE VOL 294 2 NOVEMBER 2001 1071 R ESEARCH A RTICLES tex (S1). In rodent S1, an array of barrels mation from a single whisker (33–35). The shifts the fields posteriorly and compresses reflects the pattern of whiskers on the ani- barrel fields are normally located in a central them (n ϭ 9) (Fig. 4, D and E), whereas mal’s snout, each barrel innervated by position along the A/P axis of the neocortex electroporation of sFGFR3 shifts them ante- thalamocortical axons carrying sensory infor- (Fig. 4A). Anterior electroporation of Fgf8 riorly (n ϭ 12) (Fig. 4, G and H). Suggesting that FGF signaling regulates neocortical pat- Fig. 1. (A) In utero electroporation-mediated terning not only at a broad scale, but also at a gene transfer used to modify FGF8 signaling in fine scale, the latter manipulation also skews mouse cortical primordium. Laparotomies were the outline of both barrel subfields and indi- performed at E11.5 (38), and embyros were vidual barrels, elongating them along their visualized through the uterus with a fiber optic A/P axis (compare Fig. 4, B and H). light source. Plasmid DNA (0.5 to 1.0 ␮g/␮l) (39) was mixed with 1% fast green (Sigma) and Introducing a second FGF8 source re- injected into the left cerebral ventricle of each sults in duplicate somatosensory barrel embryo through a glass capillary. A fine tung- fields. These effects on the cortical map sten negative electrode and a platinum positive could reflect a role for FGF8 in modulating electrode were inserted into the left and right the relative size of neocortical areas along the hemispheres, respectively, and a series of three A/P axis, or a more fundamental role in spec- square-wave current pulses (7 to 10 V, 100 ms) were delivered, resulting in gene transfection ifying area identity itself. To distinguish be- into the medial wall of the left hemisphere. The tween these possibilities, we introduced a surgical incision was closed and embryos were new source of FGF8 into the posterior corti- allowed to develop in utero, with 50 to 60% cal primordium, i.e., at the opposite pole from survival beyond birth (40). (B) An untreated the endogenous source. Posterior electropo- E10.5 forebrain viewed from the dorsal side, ration of Fgf8 elicits a partial duplication of anterior to the top, processed for whole-mount S1: New whisker barrels appear ectopically in situ hybridization (11).