Tbr1 regulates regional and laminar identity ofpostmitotic neurons in developing neocortexFrancesco Bedognia,b, Rebecca D. Hodgea,b, Gina E. Elsena, Branden R. Nelsona, Ray A. M. Dazaa, Richard P. Beyerc,Theo K. Bammlerc, John L. R. Rubensteind, and Robert F. Hevnera,b,e,1

aCenter for Integrative Brain Research, Seattle Children’s Research Institute, Seattle, WA 98101; bDepartment of Neurological Surgery, University ofWashington School of Medicine, Seattle, WA 98104; cDepartment of Environmental and Occupational Health Sciences, University of Washington School ofPublic Health, Seattle, WA 98195; dNina Ireland Laboratory of Developmental Neurobiology, Department of Psychiatry, University of California, San Francisco,CA 94143; and eDepartment of Pathology, University of Washington School of Medicine, Seattle, WA 98195

Edited* by Pasko Rakic, Yale University, New Haven, CT, and approved May 28, 2010 (received for review February 22, 2010)

Areas and layers of the cerebral cortex are specified by geneticprograms that are initiated in progenitor cells and then, imple-mented in postmitotic neurons. Here, we report that Tbr1, a tran-scription factor expressed in postmitotic projection neurons, exertspositive and negative control over both regional (areal) and lami-nar identity. Tbr1 null mice exhibited profound defects of frontalcortex and layer 6 differentiation, as indicated by down-regulationof gene-expression markers such as Bcl6 and Cdh9. Conversely,genes that implement caudal cortex and layer 5 identity, such asBhlhb5 and Fezf2, were up-regulated in Tbr1 mutants. Tbr1 imple-ments frontal identity in part by direct promoter binding and acti-vation of Auts2, a frontal cortex gene implicated in autism. Tbr1regulates laminar identity in part by downstream activation ormaintenance of Sox5, an important transcription factor controllingneuronal migration and corticofugal axon projections. Similar toSox5 mutants, Tbr1 mutants exhibit ectopic axon projections tothe hypothalamus and cerebral peduncle. Together, our findingsshow that Tbr1 coordinately regulates regional and laminar iden-tity of postmitotic cortical neurons.

arealization | Auts2 | microarray

The mammalian cerebral neocortex has a conserved modularorganization comprisedof areasparcellating the cortical surface

and layers stratifying the cortical thickness (1–4). The developmentof neocortical areas and layers is coordinated by specialized pro-grams of neurogenesis and neuronal subtype specification (5–7).Areal and laminar identity are initially specified in cortical pro-genitor cells and then, implemented through subsequent processesof fate acquisition in postmitotic neurons (3, 4, 8). Some tran-scription factors (TFs) have been implicated in the regulation ofboth areal and laminar identity. For example, Pax6, a TF expressedin progenitor cells, promotes both rostral identity (9) and upperlayer neurogenesis (10). Similarly, Bhlhb5, a TF expressed in post-mitotic neurons, is required for the acquisition of caudal motorand sensory areal identity and layer 5 corticospinal motor neuron(CSMN) identity (8).In the present study, we investigated whether Tbr1, a T-box TF

expressed in postmitotic projection neurons (11), might regulateareal and laminar identity of cortical neurons. Previously, we haveshown that Tbr1 is necessary for the differentiation of preplateand layer 6 neurons (11).Whereas Tbr1 is expressed at the highestlevels in developing frontal cortex (12–14), we hypothesized thatTbr1 might implement frontal identity. This hypothesis accordswith evidence that Tbr1 is expressed downstream of Pax6 througha TF cascade from Pax6+ radial progenitors to Tbr1+ postmitoticprojection neurons (15–18). In addition, we examined whetherTbr1 might suppress alternative regional and laminar fates.To test our hypothesis, we profiled changes of regional and

laminar identity in Tbr1 null mutant neocortex (11, 19) using gene-expression markers. We found that markers of frontal and layer 6differentiation were markedly down-regulated, whereas markersof caudal cortex and layer 5 were significantly increased in Tbr1mutants. Furthermore, the expression of caudal markers was shif-ted rostrally, and early-born neurons shifted from layer 6 to layer 5

identity. Finally, we found that Tbr1 implements frontal identity inpart by transcriptional activation of Auts2, a frontal marker gene(20) linked to autism (21) and mental retardation (22).Our findings show that Tbr1 modulates the balance of cortical

areas and layers by regulating gene expression in postmitoticneurons. Perturbations of regional and laminar identity may beimportant factors in neurodevelopmental diseases.

ResultsPrevious studies have shown that Tbr1 exhibits high rostral and lowcaudal expression in developing neocortex (12–14). This suggeststhat Tbr1 may contribute to the implementation of frontal cortexidentity. Alternatively, Tbr1 expression could simply reflect therostrocaudal gradient of neurogenesis (23).To further characterize Tbr1 expression along the rostrocaudal

axis, we compared Tbr1 with Bhlhb5, a caudal marker (8) in de-veloping mouse cortex (Fig. S1). These experiments revealedopposing gradients of Tbr1 (high rostral) and Bhlhb5 (high cau-dal), first discernible in the cortical plate (CP) on embryonic days(E) 13.5–14.5 (Fig. S1 A–C). Tbr1 and Bhlhb5 also exhibitedcomplementary laminar patterns, most obvious from E16.5 topostnatal day (P) 0.5 (Fig. S1 F–U). In frontal cortex, Tbr1 washighly expressed in all layers, whereas Bhlhb5 was virtually absent.In more caudal regions, Tbr1 was highly expressed in layer 6, sub-plate (SP), and Cajal-Retzius (C-R) neurons, whereas Bhlhb5 washighly expressed in layers 2–5 as noted previously (8). Tbr1 wasnot absent from layers 2–5 but was expressed at much lower levelsthan in early-born neurons.To determine whether Tbr1 is required for the acquisition of

frontal identity, we studied cortical regionalization in Tbr1 nullmice, which die shortly after birth (11, 19). Because anatomicallandmarks of cortical areas (e.g., somatosensory barrels) are notdeveloped on P0.5, we assessed regional identity using molecularexpression patterns (12, 13). Previous studies have describedseveral regional markers in embryonic and neonatal cortex, suchas Bcl6 rostrally andOdz3 caudally (14, 24, 25). To find additionalregional markers, we mined online databases (26–28), previouspublications, and our microarray data (Dataset S1 and DatasetS2). This approach yielded 20 rostral and 12 caudal markers inE14.5 cortex and 28 rostral and 30 caudal markers in P0.5 cortex(Tables S1 and S2). Regional markers were assembled into panelsfor gene-set analysis (GSA), a statistical method for testing the

Author contributions: F.B., B.R.N., and R.F.H. designed research; F.B., R.D.H., G.E.E., B.R.N.,R.A.M.D., R.P.B., and T.K.B. performed research; J.L.R.R. contributed new reagents/analytic tools; F.B., R.D.H., G.E.E., B.R.N., R.P.B., T.K.B., and R.F.H. analyzed data; andR.F.H. wrote the paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.

Data deposition: Microarray datasets have been deposited in the NCBI GEO databaseunder accession number GSE22371.1To whom correspondence should be addressed. E-mail: rhevner@uw.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1002285107/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1002285107 PNAS | July 20, 2010 | vol. 107 | no. 29 | 13129–13134

NEU

ROSC

IENCE

Dow

nloa

ded

by g

uest

on

June

4, 2

020

significance of coordinate changes in the expression of multiplegenes (29, 30).Regional marker analysis revealed profound defects of frontal

differentiation in Tbr1 mutant neocortex (Fig. 1). Early in CPdevelopment (E14.5), frontal markers Auts2, Bcl6, and Rorb allhad reduced expression, shown anatomically by in situ hybridiza-tion (ISH) and quantitatively bymicroarray profiling (Fig. 1A–D).Among eight markers of rostral identity (excluding Tbr1) in E14.5CP and intermediate zone (IZ), seven had decreased expression inTbr1 null cortex by microarray (Fig. 2A). The reduction of rostralmarkers in Tbr1 null E14.5 CP/IZ was highly significant (GSA P <0.001). Notably, these defects of frontal gene expression precededthe onset of increased apoptosis in frontal cortex (Discussion).Interestingly, the only rostral marker with increased expression inTbr1 null E14.5 CP/IZ was Spry2 (Fig. 2A). Because Spry2 is reg-ulated by FGF signaling (31, 32), this could suggest that FGFsignaling was increased in Tbr1mutant cortex. This interpretationwas supported by expression data on some other FGF signaling-related molecules, including Fgf15, Fgf17, Spry1, and Etv1 (Fig. S2,Dataset S1, and Dataset S2). These results suggest that Tbr1 mayactivate genes that suppress FGF signaling, among other possibili-ties (Discussion).Later inCPdevelopment (P0.5), frontal differentiation remained

severely defective in Tbr1 mutants. Auts2, Bcl6, Rorb, and Etv5 allshowed decreased expression on P0.5 (Fig. 1 E–L). Among 28markers of rostral identity in P0.5 CP and IZ, 27 showed decreasedexpression in Tbr1 null frontal cortex by microarray (Fig. 2B).

Frontal markers were reduced in all layers of cortex, includingBhlhb2 in layers 2–3,Rorb in layer 4,Etv5 in layers 4–6, andWscd1 inlayer 6 and SP. Many frontal markers were reduced in parietal andoccipital cortex as well, consistent with graded expression (Fig. 2 Cand D). The decrease of rostral markers was highly significantin Tbr1 null frontal (GSA P < 0.002) and parietal (GSA P= 0.001)cortex. The only rostral marker to increase in P0.5 Tbr1 null frontalcortex was Sfrp2 (Fig. 2B), possibly regulated through the Fgf15–Pax6–Sfrp2 pathway (33, 34).To determine if Tbr1 regulates caudal identity, we studied caudal

markers in Tbr1mutants. On E14.5, caudal markers Bhlhb5, Crym,andNhlh1 all showed up-regulation (Fig. 3 A–D). Significantly, theexpression boundaries of these caudal genes were shifted rostrallyin Tbr1 null cortex, indicating that they were expressed ectopically.The rostral shift of Bhlhb5 expression in E14.5 cortex (Fig. 3A) wasespecially significant, because Bhlhb5 mediates the postmitoticacquisition of caudal identity (8). In a panel of seven caudal CP/IZmarkers, six had increased expression by microarray (Fig. 2A). Theoverall increase of E14.5 caudal gene expression almost reachedstatistical significance (GSAP=0.06). The only caudalmarker thatexhibited a (slight) decrease of expression inE14.5Tbr1 null cortex,Odz1, is expressed mainly in SP (25), and SP differentiation is se-verely defective in Tbr1 mutants (11). This observation illustratesone way in which regional and laminar identity may interact.Caudal CP/IZ markers remained elevated in P0.5 Tbr1 mutant

cortex. Crym, Epha6, Pcdh8, Tshz2, and Odz3 all showed in-creased expression with rostral shifts (Fig. 3 E–N). Lmo4, which

Fig. 1. Down-regulation of rostralmarker genes inTbr1mutant cortex. (A–D) E14.5.Auts2 (A and A′), Bcl6 (B andB′), and Rorb (C and C′) mRNA were reduced in Tbr1 nullrostral cortex (A′–C′) compared with controls (A–C).Microarray results (D) confirmed down-regulation ofthese genes, although only Auts2 and Bcl6 reached sta-tistical significance. *P< 0.05. (E–L) P0.5.Auts2 (E and E′),Bcl6 (G andG′), Rorb (I and I′), and Etv5 (K and K′) mRNAwere reduced in Tbr1 null frontal cortex (E′,G′, I′, and K′)compared with controls (E, G, I, and K). Microarray pro-filing (F, H, J, and L) showed significant reductions ofthese genes mainly in frontal (FR) and parietal (PAR)regions of Tbr1 mutant cortex. Microarray results areexpressed as log2 of the ratio between knockout (KO)and wild-type (WT) normalized mRNA levels. OCCIP, oc-cipital cortex. [Scale bars (A–K′): 200 μm.]

13130 | www.pnas.org/cgi/doi/10.1073/pnas.1002285107 Bedogni et al.

Dow

nloa

ded

by g

uest

on

June

4, 2

020

normally has expression domains in caudal and rostral but notparietal cortex (Fig. 3O), likewise showed a marked rostral shift(Fig. 3 O′ and P). In a panel of 30 caudal markers in P0.5 CP/IZ,26 showed increased expression in Tbr1 null frontal cortex (Fig.2B). The up-regulation of caudal markers was highly significant(GSA P < 0.001). Many caudal markers were also increased inparietal and occipital regions (Fig. 2 C and D). Bhlhb5 was in-creased in P0.5 Tbr1 mutant prefrontal cortex but did not seemelevated in other regions (Fig. 3 Q and Q′). Bhlhb5 expressionnormally declines in neonatal mice and becomes restricted toprimary sensory-input areas (8). This could reflect dependenceon thalamic innervation, which is defective in Tbr1 mutants (11).The decline of rostral identity and rise of caudal identity in Tbr1

null cortex could have several explanations. First, Tbr1 couldregulate regional identity autonomously in postmitotic neurons;our data seem consistent with this idea. Second, Tbr1 might affectregional identity nonautonomously (e.g., by feedback to progen-itor cells). However, regional TF gradients in the ventricular zone(VZ) and subventricular zone (SVZ) were not significantly al-tered in Tbr1 null cortex (Fig. S3). Third, Tbr1 might regulate thegenesis or survival of neurons in different cortical regions. Neu-ronal production seems intact in Tbr1mutants (11). To investigatewhether apoptosis is increased in Tbr1 mutants, we tested for ex-pression of activated caspase-3 protein, a sensitive marker of ap-optosis (35). Apoptotic cells were rare in normal cortex but weremarkedly increased in Tbr1 mutant rostral cortex beginning onE16.5 (Fig. S4). The death of frontal neurons undoubtedly ac-counted for some changes in regional marker expression on P0.5,

but because apoptosis was not increased onE14.5, cell death couldnot explain defects of rostral identity at this age. Additionally,apoptosis could not explain the rostral shift of caudalmarkers (Fig.3). Together, our results suggest that Tbr1 is necessary for thedirect implementation of rostral identity in postmitotic neurons.We next analyzed laminar identity in Tbr1mutants using panels

of layer-specific markers in E14.5 and P0.5 neocortex (Tables S3and S4). On E14.5, we assayed the differentiation of early-bornneuron types, including C-R, SP, and CP neurons. Microarray

Fig. 2. Microarray analysis of rostral and caudal marker genes in Tbr1 nullcortex. (A) E14.5. Most rostral markers decreased (red) and caudal markersincreased (green) in Tbr1 null cortex. GSA P values indicate the probability thatsets of genes (rostral or caudal) increased or decreased by chance. (Inset) Tbr1ISH, E14.5 (28). (B–D) P0.5. Rostral and caudal markers were analyzed in rostral(B), parietal (C), and caudal (D) regions of Tbr1 KO cortex versus controls.(B Inset) Tbr1 immunohistochemistry, E18.5. n, number of genes in the set.

Fig. 3. Up-regulation of caudal marker genes in Tbr1 mutant cortex. (A–D)E14.5. Immunofluorescence for Bhlhb5 (A and A′) and ISH for Crym (B and B′)and Nhlh1 (C and C′) showed increased rostral expression of these genes inTbr1 null cortex (A′–C′) compared with controls (A–C). (D) Microarray resultsshowed that Crym and Nhlh1 expression changes were statistically signifi-cant. (E–R) P0.5. Crym (E and E′), Epha6 (G and G′), Pcdh8 (I and I′), Tshz2(K and K′), Odz3 (M and M′), Lmo4 (O and O′), and Bhlhb5 (Q and Q′) mRNAwere increased in Tbr1 null cortex. The Lmo4 low-expression domain (O,between arrows) was shifted rostrally in Tbr1 null cortex (O′). Bhlhb5 ex-pression was disorganized and increased rostrally in the malformed Tbr1null cortex (Q′) compared with control (Q). Microarray results are shown inF, H, J, L, N, P, and R. [Scale bars (A–Q′): 200 μm.]

Bedogni et al. PNAS | July 20, 2010 | vol. 107 | no. 29 | 13131

NEU

ROSC

IENCE

Dow

nloa

ded

by g

uest

on

June

4, 2

020

indicated that C-R neuron markers Reln (Reelin) and Calb2(calretinin) were significantly decreased in E14.5 Tbr1 null neo-cortex, confirming previous ISH results (11). Also, the E14.5microarray data showed decreased expression of SP markers Kitl,Odz1, and Mef2c and CP markers Cdh13, Cnr1, Bcl11b (Ctip2),Sox5, and Zeb2 (SIP1). Reductions of Bcl11b, Mef2c, and Sox5were further validated by ISH (Fig. S5). These findings supportedthe conclusion that Tbr1 is required for the differentiation ofpreplate and layer 6 identity (11).In P0.5 cortex, we analyzed markers for all layers. Microarray

indicated severe defects of C-R, SP, and layer 6 differentiation inTbr1 null P0.5 cortex (Fig. 4). For example, Wnt7b, a layer 6/SPmarker, and Reln, a C-R marker, were down-regulated on mi-croarray, supporting previous ISH results (11). Additional ISHexperiments validated the down-regulation of SP marker Ctgfand layer 6 markers Sox5 and Tle4 (Fig. 5 A–I). Reduction ofSox5 was noteworthy, because Sox5 suppresses aspects of layer 5identity, including Fezf2 expression (36, 37). Interestingly, layer6/SP markers restricted to caudal cortex, such as Nr4a2 (Nurr1)and Ngfr, did not decrease as much as other layer 6/SP markers(Fig. 4). The relative sparing of caudal layer 6/SP markers mayreflect an interaction of regional and laminar fate determination,balancing up-regulation of caudal identity with down-regulationof layer 6 and SP identity.Microarray also showed increased expression of some layer 2–5

markers in P0.5 Tbr1 null cortex (Fig. 4). Layer 5 markers were up-regulated most consistently, including subcerebral projection neu-ron (SCPN) and corticospinal motor neuron (CSMN) markers(Fig. 4). For example, Fezf2, a determinant of SCPN identity (38–40), and Etv1 (Er81), an established layer 5 marker (Tables S3 andS4), were significantly increased in all regions (Fig. 4). Up-regula-tion of Fezf2, Er81, and Crim1 (a CSMN marker) was con-firmed by ISH (Fig. 5 J–R). Interesting exceptions to the generallyincreased expression of upper-layer genes included Rorb, a layer

4 marker, and Pcdh20, a layer 2–4 marker (Fig. 4). Because thesegenes are also P0.5 rostral markers (Tables S1 and S2), their down-regulation was consistent with impaired acquisition of frontalidentity in Tbr1 mutant cortex.Increased expression of layer 5 markers, together with de-

creased expression of Sox5, suggested that some early-born cellsswitched from layer 6 to layer 5 identity in Tbr1 null mutants. Tomore definitively test for changes in laminar fate, we used twoapproaches: BrdU birthdating followed by analysis of laminarfate markers and axon tracing of corticofugal fibers (Fig. S6).Cells born on E12.5 (BrdU+) had decreased layer 6 fates (Tle4+)and increased layer 5 fates (Er81+, Ctip2+) in all regions (Fig. S6A–L), including occipital cortex where apoptosis was not a factor(Fig. 4). The Sox5+ fate index did not change, probably becauseSox5 is expressed not only in layer 6 but also at lower levels inlayer 5 (36), and our cell counts did not distinguish between highand low Sox5 immunoreactivity.To examine whether cortical axon projections shifted from

layer 6 corticothalamic to layer 5 subcerebral or ectopic pathwaysin Tbr1 mutants, we studied projections from frontal cortex. Wefocused on frontal cortex, because parietal and occipital regionshave severe generalized defects of axon growth (11). Theseexperiments revealed exuberant subcerebral projections in Tbr1mutants, whereas projections to thalamus were reduced but notabsent (Fig. S6M andN). These findings fit well with the increasedCtip2 fate index in Tbr1 null cortex (Fig. S6L), because Ctip2promotes subcerebral axon projections (41). Ectopic projectionsto hypothalamus were also noted (Fig. S6N). Together, the exu-berant subcerebral projections and increased layer 5 markerssuggested that some early-born neurons switched from layer 6 tolayer 5 identity in Tbr1 mutants.Most T-box transcription factors function as transcriptional

activators (42). To test whether Tbr1 might directly activate layer6 or frontal marker genes, we used Tbr1 overexpression and ChIPassays. Tbr1, along with GFP reporter in the same cells, wasoverexpressed by plasmid electroporation from the lateral ven-tricles in E13.5 cortex followed by cortical slice culture for 1 d (Fig.S7). Electroporated cells were probed for up-regulation of candi-date Tbr1 target genes, includingAuts2 (frontal cortex), Sox5 (layer6), Tle4 (layer 6), and FOG2 (layer 6). As controls, we tested forinduction of genes predicted not to be targets of Tbr1, includingBhlhb5 (caudal cortex/layers 2–5) and Ctip2 (layer 5). Tbr1 over-expression consistently induced ectopic Auts2 (Fig. 6 A–H) but didnot induce Sox5, Tle4, FOG2, Bhlhb5, or Ctip2. These data sug-gested that Tbr1 might bind and activate Auts2 in vivo.To determine if Tbr1 binds the Auts2 promoter, we used Tbr1

antibodies for ChIP of E14.5 cortex. Sequence analysis identifiedsix potential Tbr1 binding sites near the Auts2 transcriptionalstart site (Fig. 6I). One of these candidate Tbr1 binding sites washighly enriched (>50-fold) in chromatin fromTbr1 ChIP (Fig. 6K).This Tbr1 binding site was located in a region of transcriptionallyactive open chromatin adjacent to the Auts2 transcriptional startsite, as shown by acetylated histone ChIP (Fig. 6J). Whereas Tbr1is required for cortical Auts2 expression (Figs. 1 and 2), inducesAuts2 expression ectopically in cortex (Fig. 6 A–H), and binds theAuts2 promoter (Fig. 6K), we conclude that Auts2 is a directtranscriptional target of Tbr1 in developing neocortex.

DiscussionOur findings indicate that Tbr1 is required for the implementationof regional and laminar identity in postmitotic neurons. UnlikeBhlhb5, which mediates the acquisition of caudal and layer 5fates (8), Tbr1 also suppresses alternative identities—specifically,caudal and layer 5 fates (Figs. 2–5). Thus, Tbr1 probably functionsupstream of Bhlhb5 in a transcriptional network to implementcortical-neuron subtype identity.In the last decade, much has been learned about factors that

promote frontal identity of progenitor cells, such as FGF8 andPax6 (3, 5, 13). Tbr1 is a TF found to promote frontal identity inpostmitotic neurons, presumably implementing fate that is ini-tially specified in radial progenitor cells. Interestingly, Tbr1+

Fig. 4. Microarray profiling of laminar marker genes in P0.5 Tbr1 mutantmice. Layer 6, SP, and C-R cell markers were down-regulated in all regions ofTbr1 null cortex. Layer 5 markers were up-regulated in all regions, includingmarkers of subcerebral projection neurons (SCPNs) and corticospinal motorneurons (CSMNs). Markers of layers 2 and 3 were also mostly up-regulated inTbr1 null cortex. Callosal projection neurons (CPN) were not significantlyelevated overall, as determined by GSA. (Inset) Tbr1 immunohistochemistry(P0.5) showing laminar expression in parietal cortex.

13132 | www.pnas.org/cgi/doi/10.1073/pnas.1002285107 Bedogni et al.

Dow

nloa

ded

by g

uest

on

June

4, 2

020

neurons are produced both directly from Pax6+ radial progenitorsand indirectly from intermediate neuronal progenitors (INPs)that express neither Pax6 nor Tbr1 (15). Therefore, we predictthat INPs must express other TFs that maintain regional identity.One candidate could be Tbr2 (Eomes), a T-box factor related toTbr1 that is specifically expressed in INPs (15) and moreover, isregulated by Pax6 (16, 17).Our data revealed an unexpected interaction between Tbr1 and

FGF signaling. Increased FGF signaling was suggested by ele-vated levels of Fgf17, Spry1, Spry2, and other molecular reportersof FGF signaling (Fig. S2, Dataset S1, and Dataset S2). DifferentFGFs have diverse effects on telencephalic development (13, 34,43–45), and the impact of these FGF signaling changes is unclear.Considering that FGFs regulate mainly progenitor cells and Tbr1controls postmitotic neurons, our data could suggest that Tbr1activates a feedback loop to dampen FGF signaling. Furtherstudies will be necessary to investigate this and other possibilities.Laminar identity, like regional identity, is specified in progeni-

tors and transmitted into postmitotic neurons (4, 8). Among TFsknown to implement laminar fate in postmitotic neurons, Tbr1 andSox5 are important for layer 6 and SP (36, 37). Our present resultssuggest that they are part of the same transcriptional network (Fig.S6O). Sox5 expression was decreased in Tbr1 null cortex on E14.5(Fig. S5) and P0.5 (Figs. 4 and 5 D–F). Phenotypic similaritiesbetween Sox5- and Tbr1-deficient mice also support this link. In-activation of Tbr1, like Sox5 (36, 37), led to increased Fezf2 ex-pression (Figs. 4 and 5) along with ectopic corticofugal projectionsfrom frontal cortex into the hypothalamus and cerebral peduncle(Fig. S6). Outside frontal cortex, axon growth and guidance are sodefective in Tbr1 mutants that most cortical efferents do not growbeyond the internal capsule (11). Whereas layer 5 SCPN markersFezf2 and Ctip2 were up-regulated in P0.5 Tbr1 mutant cortex,callosal projection neuron (CPN) markers such as Satb2 (4) in-creased little or not at all (Fig. 4). This indicated that early-born

Tbr1 null neurons switched to layer 5 SCPN, not CPN identity,perhaps explaining callosal agenesis in Tbr1 mutants (11). Imple-mentation of laminar identity, like regional identity, may also re-quire expression of specific TFs in INPs.The present study and others (1, 5, 8) show that regional and

laminar fates are regulated coordinately by overlapping transcrip-tional programs involving someof the sameTFs. For individual TFssuch as Tbr1, such dual roles may be difficult to reconcile. Forexample, although Tbr1 promotes frontal cortex and layer 6 iden-tity, how does Tbr1 regulate layer 6 identity in caudal cortex? Someinsights come from studying markers with distinct expression pat-terns, such as Ngfr, a marker of caudal layer 6. Ngfr was down-regulated in Tbr1mutant occipital cortex (Fig. 4) in contrast to theup-regulation of most other caudal markers (Fig. 2). This suggeststhat Ngfr was more sensitive to the effects of Tbr1 on laminar thanon regional identity. Other informative examples include Rorb(rostral layer 4) and Pcdh20 (rostral layers 2/3). These genes weredown-regulated inTbr1mutant cortex (Fig. 2), althoughmost otherupper layer markers were up-regulated (Fig. 4). Thus, Rorb andPcdh20 were more sensitive to Tbr1 regulation of regional thanlaminar identity. Because Tbr1 does not induce rostral identity incaudal layer 6 or layer 6 identity in rostral upper layers, additionalregulators must be postulated. These could include combinatorialinteractions with other TFs, epigenetic regulation of chromatin, orposttranslational modifications of Tbr1.Finally, the present study identified Auts2, a frontal cortex

marker gene (20) linked to autism (21) andmental retardation (22),as a direct target of Tbr1 binding and activation (Fig. 6). Otherpotential targets of Tbr1 activation proposed in previous studiesinclude Reln, Grin1, and Grin2b (46–48). Our previous (11) andcurrent results (Fig. 4) support Reln as a direct target of Tbr1.However, by microarray, Grin1 and Grin2b were not reduced inTbr1mutant cortex on E14.5 or P0.5, except for amodest reductionof Grin2b (log2FC = −0.17) in P0.5 frontal but not somatosensoryor occipital cortex. In future studies, we hope to evaluate theseand other candidate targets of Tbr1 through additional ChIP andTbr1 overexpression assays.

MethodsDetailed Methods. Additional procedural details are available in SI Methods.

Fig. 5. Expression of subplate, layer 6, and layer 5 marker genes in Tbr1 nullcortex. (A–I) SP and layer 6 markers Ctgf (A–C), Sox5 (D–F), and Tle4 (G–I) haddecreased expression in Tbr1 null cortex by ISH and microarray. (J–R) Layer 5markers Fezf2 (J–L), Er81 (M–O), and Crim1 (P–R) had increased expression.[Scale bars (A–Q): 200 μm.]

Fig. 6. Tbr1 binds and activates the Auts2 gene in neocortex in vivo. (A–H)Tbr1 expression plasmid (Tbr1–ires GFP), but not GFP only (ires–GFP), droveectopic expression of Auts2 (red) in electroporated GFP+ cells (green) incontrol (A, B, E, and F) and Tbr1 null (C, D, G, and H) VZ. The electroporatedregion from A, C, E, and G is shown at higher magnification in B, D, F, and H.[Scale bars (A and C): 50; (B and D): 30 μm; (E and G): 100 μm; (F and H):25 μm.] (I) Sequence analysis of the Auts2 promoter and proximal codingregion identified six candidate Tbr1 binding sites (red T). (J) ChIP of acety-lated histones from E14.5 forebrain showed that the Auts2 transcriptionalstart site was highly enriched in open chromatin. (K) Tbr1 ChIP showed Tbr1binding to a candidate site in the region of open chromatin (J) adjacent tothe Auts2 transcriptional start site.

Bedogni et al. PNAS | July 20, 2010 | vol. 107 | no. 29 | 13133

NEU

ROSC

IENCE

Dow

nloa

ded

by g

uest

on

June

4, 2

020

Animals and Tissue Processing. Tbr1 null mice (11, 19) were maintained ona CD1 background and genotyped by PCR. The plug date was designatedE0.5. ICC and ISH were done as described (15, 49) and repeated in at leastthree brains. Immunocytochemistry antibodies and ISH probes are listed inSI Methods.

Axon Tracing with DiI. Axon tracingwasdoneasdescribed (11)usingP0.5brains(n = 3 Tbr1 mutant; n = 3 littermate controls) fixed by perfusion with 4%paraformaldehyde. DiI crystals were injected into frontal cortex. After storagein fixative for 8 wk, brains were sectioned coronally at 100 μm, counterstainedwith DAPI, mounted forfluorescencemicroscopy, and photographed digitally.

RNA Isolation and Microarrays. E14.5 and P0.5 brains were removed, andneocortexes were immediately dissected. RNA samples from control (n ≥ 3)and Tbr1 null (n ≥ 3) embryos were hybridized on Affymetrix mouse genearrays (U430 2.0 or ST 1.0) in the Microarray Core of the University ofWashington Center for Ecogenetics and Environmental Health. Data wereanalyzed statistically by GSA as described in SI Methods.

Identification of Regional and Laminar Markers. Regional and laminar markerswere assembled by mining previous studies, online databases (26–28), andresults in the present study. Genes were organized in two age ranges corre-

sponding to embryonic CP differentiation (E13.5–E15.5) and perinatal CPdifferentiation (E18.5–P4).

Ex Utero Electroporation. E13.5 embryoswere harvested into cold (4 °C) buffer.Plasmids encoding GFP only (ires–GFP) or Tbr1 and GFP (Tbr1–ires–GFP) wereinjected into the ventricles and electroporated with paddle electrodes acrossthe cerebrum. The brain was sliced coronally (400 μm), cultured for 24 h, andfixed with cold buffered 4% paraformaldehyde for cryosectioning and ICC.Colocalization of GFP and TFs was assessed by confocal microscopy.

ChIP. Chromatin was precipitated from E14.5 forebrain and further processedas described in SI Methods.

ACKNOWLEDGMENTS. This work was supported by National Institutes ofHealth Grant R01NS050248 (to R.F.H.). Microarray experiments were sup-ported in part by University of Washington Center for Ecogenetics andEnvironmental Health National Institute of Environmental Health ServicesGrant P30ES07033. This research was facilitated by resources provided byNational Institute of Child Health and Human Development Grant P30HD02274. F.B. was supported by a grant from Universita degli Studi diMilano. R.D.H. was supported by fellowships from the Heart and StrokeFoundation of Canada and the American Heart Association.

1. Rakic P, Ayoub AE, Breunig JJ, Dominguez MH (2009) Decision by division: Makingcortical maps. Trends Neurosci 32:291–301.

2. Molyneaux BJ, Arlotta P, Menezes JR, Macklis JD (2007) Neuronal subtype specifi-cation in the cerebral cortex. Nat Rev Neurosci 8:427–437.

3. O’Leary DD, Sahara S (2008) Genetic regulation of arealization of the neocortex. CurrOpin Neurobiol 18:90–100.

4. Leone DP, Srinivasan K, Chen B, Alcamo E, McConnell SK (2008) The determination ofprojection neuron identity in the developing cerebral cortex. CurrOpinNeurobiol 18:28–35.

5. Rash BG, Grove EA (2006) Area and layer patterning in the developing cerebralcortex. Curr Opin Neurobiol 16:25–34.

6. Polleux F, Dehay C, Kennedy H (1997a) The timetable of laminar neurogenesis contributesto the specification of cortical areas in mouse isocortex. J Comp Neurol 385:95–116.

7. Polleux F, Dehay C, Moraillon B, Kennedy H (1997b) Regulation of neuroblast cell-cycle kinetics plays a crucial role in the generation of unique features of neocorticalareas. J Neurosci 17:7763–7783.

8. Joshi PS, et al. (2008) Bhlhb5 regulates the postmitotic acquisition of area identities inlayers II-V of the developing neocortex. Neuron 60:258–272.

9. Bishop KM, Goudreau G, O’Leary DD (2000) Regulation of area identity in themammalian neocortex by Emx2 and Pax6. Science 288:344–349.

10. Schuurmans C, et al. (2004) Sequential phases of cortical specification involveneurogenin-dependent and -independent pathways. EMBO J 23:2892–2902.

11. Hevner RF, et al. (2001) Tbr1 regulates differentiation of the preplate and layer 6.Neuron 29:309–311.

12. Miyashita-Lin EM, Hevner R, Wassarman KM, Martinez S, Rubenstein JL (1999) Earlyneocortical regionalization in the absence of thalamic innervation. Science 285:906–909.

13. Fukuchi-Shimogori T, Grove EA (2001) Neocortex patterning by the secreted signalingmolecule FGF8. Science 294:1071–1074.

14. Leamey CA, et al. (2008) Differential gene expression between sensory neocorticalareas: Potential roles for Ten_m3 and Bcl6 in patterning visual and somatosensorypathways. Cereb Cortex 18:53–66.

15. Englund C, et al. (2005) Pax6, Tbr2, and Tbr1 are expressed sequentially by radial glia,intermediate progenitor cells, and postmitotic neurons in developing neocortex.J Neurosci 25:247–251.

16. Holm PC, et al. (2007) Loss- and gain-of-function analyses reveal targets of Pax6 in thedeveloping mouse telencephalon. Mol Cell Neurosci 34:99–119.

17. Sansom SN, et al. (2009) The level of the transcription factor Pax6 is essential forcontrolling the balance between neural stem cell self-renewal and neurogenesis. PLoSGenet 5:e1000511.

18. Hevner RF (2006) From radial glia to pyramidal-projection neuron: Transcriptionfactor cascades in cerebral cortex development. Mol Neurobiol 33:33–50.

19. Bulfone A, et al. (1998) An olfactory sensory map develops in the absence of normalprojection neurons or GABAergic interneurons. Neuron 21:1273–1282.

20. Bedogni F, et al. (2010) Autism susceptibility candidate 2 (Auts2) encodes a nuclearprotein expressed in developing brain regions implicated in autism neuropathology.Gene Expr Patterns 10:9–15.

21. Sultana R, et al. (2002) Identification of a novel gene on chromosome 7q11.2interrupted by a translocation breakpoint in a pair of autistic twins. Genomics 80:129–134.

22. Kalscheuer VM, et al. (2007) Mutations in autism susceptibility candidate 2 (AUTS2) inpatients with mental retardation. Hum Genet 121:501–509.

23. Caviness VS, Jr, Takahashi T, Nowakowski RS (1995) Numbers, time and neocorticalneuronogenesis: A general developmental and evolutionary model. Trends Neurosci18:379–383.

24. Funatsu N, Inoue T, Nakamura S (2004) Gene expression analysis of the late embryonicmouse cerebral cortex using DNA microarray: Identification of several region- andlayer-specific genes. Cereb Cortex 14:1031–1044.

25. Li H, Bishop KM, O’Leary DD (2006) Potential target genes of EMX2 include Odz/Ten-Mand other gene families with implications for cortical patterning. Mol Cell Neurosci 33:136–149.

26. Lein ES, et al. (2007) Genome-wide atlas of gene expression in the adult mouse brain.Nature 445:168–176.

27. Magdaleno S, et al. (2006) BGEM: An in situ hybridization database of gene ex-pression in the embryonic and adult mouse nervous system. PLoS Biol 4:e86.

28. Alvarez-Bolado G, Eichele G, et al. (2006) Analysing the developing brain transcriptomewith the GenePaint platform. J Physiol 575:347–352.

29. Nam D, Kim SY (2008) Gene-set approach for expression pattern analysis. BriefBioinform 9:189–197.

30. Subramanian A, et al. (2005) Gene set enrichment analysis: A knowledge-basedapproach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA102:15545–15550.

31. Kim HJ, Bar-Sagi D (2004) Modulation of signalling by Sprouty: A developing story.Nat Rev Mol Cell Biol 5:441–450.

32. Minowada G, et al. (1999) Vertebrate Sprouty genes are induced by FGF signaling andcan cause chondrodysplasia when overexpressed. Development 126:4465–4475.

33. Kim AS, Anderson SA, Rubenstein JL, Lowenstein DH, Pleasure SJ (2001) Pax-6 regulatesexpression of SFRP-2 and Wnt-7b in the developing CNS. J Neurosci 21:RC132.

34. Borello U, et al. (2008) FGF15 promotes neurogenesis and opposes FGF8 functionduring neocortical development. Neural Develop 3:17:1–18.

35. Narasimhaiah R, Tuchman A, Lin SL, Naegele JR (2005) Oxidative damage anddefective DNA repair is linked to apoptosis of migrating neurons and progenitorsduring cerebral cortex development in Ku70-deficient mice. Cereb Cortex 15:696–707.

36. Kwan KY, et al. (2008) SOX5 postmitotically regulates migration, postmigratorydifferentiation, and projections of subplate and deep-layer neocortical neurons. ProcNatl Acad Sci USA 105:16021–16026.

37. Lai T, et al. (2008) SOX5 controls the sequential generation of distinct corticofugalneuron subtypes. Neuron 57:232–247.

38. Chen B, Schaevitz LR, McConnell SK (2005) Fezl regulates the differentiation and axontargeting of layer 5 subcortical projection neurons in cerebral cortex. Proc Natl AcadSci USA 102:17184–17189.

39. Chen JG, Rasin MR, Kwan KY, Sestan N (2005) Zfp312 is required for subcorticalaxonal projections and dendritic morphology of deep-layer pyramidal neurons of thecerebral cortex. Proc Natl Acad Sci USA 102:17792–17797.

40. Molyneaux BJ, Arlotta P, Hirata T, Hibi M, Macklis JD (2005) Fezl is required for thebirth and specification of corticospinal motor neurons. Neuron 47:817–831.

41. Arlotta P, et al. (2005) Neuronal subtype-specific genes that control corticospinalmotor neuron development in vivo. Neuron 45:207–221.

42. Naiche LA, Harrelson Z, Kelly RG, Papaioannou VE (2005) T-box genes in vertebratedevelopment. Annu Rev Genet 39:219–239.

43. Sansom SN, et al. (2005) Genomic characterisation of a Fgf-regulated gradient-basedneocortical protomap. Development 132:3947–3961.

44. Iwata T, Hevner RF (2009) Fibroblast growth factor signaling in development of thecerebral cortex. Dev Growth Differ 51:299–323.

45. Sahara S, O’Leary DD (2009) Fgf10 regulates transition period of cortical stem celldifferentiation to radial glia controlling generation of neurons and basal progenitors.Neuron 63:48–62.

46. Hsueh YP, Wang TF, Yang FC, Sheng M (2000) Nuclear translocation and transcriptionregulation by themembrane-associated guanylate kinase CASK/LIN-2.Nature 404:298–302.

47. Wang GS, et al. (2004) Transcriptional modification by a CASK-interacting nucleosomeassembly protein. Neuron 42:113–128.

48. Wang TF, et al. (2004) Identification of Tbr-1/CASK complex target genes in neurons.J Neurochem 91:1483–1492.

49. Fink AJ, et al. (2006) Development of the deep cerebellar nuclei: Transcription factorsand cell migration from the rhombic lip. J Neurosci 26:3066–3076.

13134 | www.pnas.org/cgi/doi/10.1073/pnas.1002285107 Bedogni et al.

Dow

nloa

ded

by g

uest

on

June

4, 2

020