Molecular heterogeneity along the dorsal–ventral axis of the murine hippocampal CA1 field: a microarray analysis of gene expression

MOA

EEa

b

Uc

d

e

1f

1

Ahaarapahaod0tsrggwwegadataittr

*EAtsmdSp

Neuroscience 137 (2006) 177–186

0d

OLECULAR HETEROGENEITY ALONG THE DORSAL–VENTRAL AXISF THE MURINE HIPPOCAMPAL CA1 FIELD: A MICROARRAY

Kd

TfbnrMtvlimsa2cErta1

hoaDhgpocprtshteRptccts

. D. LEONARDO,a,b,d* J. W. RICHARDSON-JONES,c

. SIBILLE,f A. KOTTMANe AND R. HENa,b,c,d

Center for Neurobiology and Behavior, Columbia University, USA

Department of Psychiatry, Columbia University, 1051 Riverside Drive,nit 87, New York, NY 10032, USA

Department of Pharmacology, Columbia University, USA

New York State Psychiatric Institute, New York, NY 10032, USA

Columbia Genome Center, Russ Berrie Medical Science Pavilion,150 St. Nicholas Avenue, New York, NY 10032, USA

Department of Psychiatry, University of Pittsburgh, Pittsburgh, PA5213, USA

bstract—There has been increasing interest in functionaleterogeneity along the septotemporal, dorsal–ventral (D–V)xis of the hippocampus. Although anatomical connectivitynd lesion studies point to discrete roles for these sub-egions, the contribution of differential gene expressioncross this axis has not been systematically studied. Here weresent findings from an Affymetrix microarray screen aimedt identifying genes in the CA1 region of the adult murineippocampus that show significant differential expressionlong the D–V axis. Our results indicate that the vast majorityf monitored genes (>90%) had tissue expression levels thatiffered by less than 20% between regions, while less than.1% of genes had expression levels that varied more thanhree-fold by sub-region. Only 23 probes showed a CA1 dor-oventral signal intensity ratio greater than three: 18 en-iched dorsally and five enriched ventrally. Probes with thereatest difference in expression levels represent a range ofenes with known functions in patterning and signaling, asell as genes without known function. Selective screeningith digoxigenin-labeled in situ hybridization confirms thexistence of CA1 sub-regionalized expression, with someenes exhibiting a graded expression pattern across the D–Vxis, and others restricted to a discrete region. Our findingsemonstrate that there are gene expression differencescross the D–V axis of the adult murine hippocampus withinraditionally recognized cytoarchitecturally defined bound-ries. Combined with the previously recognized differencesn connectivity and results from lesion studies, our data fur-her confirm the existence of functional heterogeneity alonghe D–V axis. © 2005 IBRO. Published by Elsevier Ltd. Allights reserved.

Correspondence to: Tel: �1-212-543-5266; fax: �1-212-543-5074.-mail address: [email protected] (E. D. Leonardo).bbreviations: BCIP/NBT, bromo-chloro-indolyl phosphate/nitroblue

etrazolium; CA1, hippocampal subregion CA1; CA2, hippocampalubregion CA2; CA3, hippocampal subregion CA3; cDNA, comple-entary DNA; cRNA, complementary RNA; DG, dentate gyrus; DIG,igoxigenin; D–V, dorsal–ventral; LPH, lactase-phlorizin hydrolase;

mOCS-2, suppresor of cytokine signaling 2; VIP, vasoactive intestinalolypeptide.

306-4522/06$30.00�0.00 © 2005 IBRO. Published by Elsevier Ltd. All rights reseroi:10.1016/j.neuroscience.2005.08.082

177

ey words: microarray, hippocampus, CA1, septotemporal,orsal–ventral, neuroanatomy.

raditionally, functional studies of the hippocampus haveocused on its likely role in memory formation (Eichen-aum, 1997; Squire et al., 2004). However, there are sig-ificant data that point to a role for the hippocampus inegulating anxiety as well (Gray and McNaughton, 2000).ore recently, lesion studies in rats have suggested that

hese two functions may be distributed across the dorsal–entral (D–V) axis (Bannerman et al., 2004). Specifically,

esions of the dorsal hippocampus consistently result inmpaired spatial learning without affecting anxiety-related

easures, while ventral lesions affect anxiety-related mea-ures without affecting spatial navigation (Bannerman etl., 1999, 2002; Kjelstrup et al., 2002; McHugh et al.,004). Results of these studies point to a functional disso-iation along the D–V axis of the hippocampus in rodents.merging imaging data in primates point to differential

esponses along the analogous anterior–posterior axis ofhe hippocampus, suggesting that such functional dissoci-tion extends beyond the rodent model (Colombo et al.,998; Strange and Dolan, 1999, 2001).

The functional dissociation seen in lesion studies of theippocampus correlates with, and was foreshadowed by,bserved differences in afferent and efferent connectionslong the septotemporal axis (Amaral and Witter, 1989).espite this, traditional studies of hippocampal physiologyave emphasized the intrinsic tri-synaptic circuit (dentateyrus (DG)¡CA3¡CA1) that exists in the transverselane and repeats itself in the longitudinal plane. Its highlyrganized, repetitive nature makes this circuit easily ac-essible, but lends itself to a unitary view of the hippocam-us (Anderson et al., 1971). Dorsal and ventral regionseceive distinct, separable inputs and send their projec-ions to distinct targets. For example, visual, auditory andensory association cortices project primarily to dorsalippocampus via the lateral entorhinal cortex, while olfac-ory input is distributed evenly across the D–V axis (Koselt al., 1981; Witter et al., 1989; Burwell and Amaral, 1998).egarding efferent connections, the ventral hippocampusrojects directly to the prefrontal cortex, the amygdala, and

he shell of the nucleus accumbens, while the dorsal hippo-ampus does not (Pitkanen et al., 2000). In addition, hippo-ampal projections to the hypothalamus via the septum areopographically mapped, such that the dorsal CA1 andubiculum project to dorsal lateral septum and then to the

ved.

pSlstsrhgispacav

aaritdsco(bpdsa

A

AUg1UwsiMdadRwmTm1�n

R

T(as

FcpCstpurple (CA2), blue (CA3), green (DG), yellow (subiculum).

E. D. Leonardo et al. / Neuroscience 137 (2006) 177–186178

rimarily to rostral and ventral lateral septum (Risold andwanson, 1996). The septum in turn projects to hypotha-

amic medial zone nuclei involved in the expression ofocial behavior (Risold and Swanson, 1996). In summary,he dorsal hippocampus receives highly processed sen-ory information from neocortical areas consistent with aole in spatial learning/context acquisition while the ventralippocampus receives significantly more inputs from re-ions that could impart affective valence to stimuli. Phys-

ological studies citing higher densities and higher spatialelectivity of place cells in the dorsal versus ventral hip-ocampus are consistent with this interpretation (Jung etl., 1994). Combined, the evidence from both anatomi-al, behavioral and physiological studies now stronglyrgues for functional heterogeneity across the dorso-entral axis.

In contrast to the anatomical and functional differenceslong the D–V axis, the possibility of molecular differenti-tion along this axis has not been well studied. There areecognized cytoarchitectural and connectivity differencesn the classically defined structures of CA1, CA2, CA3 andhe DG (Fig. 1C, D) (Ramón y Cajal, 1953), and molecularistinctions among these regions have been recently de-cribed as well, demonstrating good correlation betweenytoarchitecturally defined boundaries and the boundariesf gene expression as confirmed by in situ hybridizationZhao et al., 2001; Lein et al., 2004). However, there haveeen no systematic analyses of differences in gene ex-ression within sub-regions of the hippocampus. Here weescribe a DNA microarray study looking at gene expres-ion differences in the CA1 region of the hippocampuscross the D–V axis.

EXPERIMENTAL PROCEDURES

nimals and tissue

ll procedures were approved by the Institutional Animal Care andse Committee, and conform to the National Institutes of Healthuidelines on the ethical use of animals. Eight-to-10-week-old29SVEV male mice were obtained from Taconic (Germantown, NY,SA). For micro-dissection and DNA microarray analysis, brainsere extracted and immediately placed on ice. Extracted brains wereliced coronally on a Series 1000 Vibratome (Intracel, Shepreth, UK)

n 400 �m-thick sections and placed in RNAlater (Sigma, St. Louis,O, USA). Primary and secondary visual cortex (cortex overlyingorsal CA1 in the coronal plane), as well as dorsal and ventral CA1reas, was dissected out as shown in Fig. 1. The tissue was handissected under a microscope using a fine surgical scalpel (partS-6272; Roboz, Gaithersburg, MD, USA). Sub-dissected samplesere pooled to yield sufficient RNA to be labeled per array experi-ent. For each region, material from six to seven mice was pooled.hree independent pools were processed per brain area (total�20ice) on three separate arrays. For in situ hybridization, brains from0-to-12-week-old mice were fresh-frozen on dry ice and stored at80 °C until sectioning. Every effort was made to minimize theumber of mice used and their suffering.

NA preparation and microarray hybridization

otal RNA was isolated from each pool of tissue using TrizolInvitrogen, Carlsbad, CA, USA). RNA purity and integrity weressessed by optical densitometry, gel electrophoresis, and sub-

ig. 1. Schematic of tissue microdissection. (a, b) Illustration of theoronal plane of section used to obtain dorsal and ventral CA1 sam-les. The shaded area represents the tissue that was isolated fromA1 for dorsal (a) and ventral (b) samples. (a) The crosshatched areahows the region that was used for the cortex sample. (c, d) Depictshe traditional hippocampal boundaries at these levels. Pink (CA1),

ascAibc1tuflao(ncctsc2t(a(

M

Dsgasepw

D

UsDccewsU(ufMeoTCfdsBtAsj

panm

ivoiv

A

OptutsesgotGpv

pt

ttrR(m

FreGs

E. D. Leonardo et al. / Neuroscience 137 (2006) 177–186 179

ccording to the Affymetrix protocol (http://www.affymetrix.com/upport/). In brief, 6 �g of total RNA was reverse-transcribed andonverted into double-stranded complementary DNA (cDNA).biotinylated complementary RNA (cRNA) was then transcribed

n vitro using an RNA polymerase T7 promoter site, which hadeen introduced during the reverse-transcription of RNA intoDNA. After fragmentation into pieces of 50–200 bases long,5 �g of labeled cRNA sample was hybridized onto oligonucleo-ide MOE430A microarrays (Affymetrix, Palo Alto, CA, USA),sing standard protocols with the Affymetrix microarray oven anduidics station. A total of 22,690 annotated genes and well-char-cterized expressed sequence tag (EST) clusters is representedn MOE430A arrays. Quality control parameters on all nine arraysthree cortex, three dorsal CA1, three ventral CA1) were as follow:oise (RawQ) less than 2.26, background signal less than 34,onsistent number of genes called as “present” across arrays,onsistent scale factors, actin and GAPDH 5=/3= signal ratios lesshan 1.66, and consistent detection of BioB and BioC hybridizationpiked controls. Details of this procedure and microarray qualityontrol parameters are described elsewhere (Galfalvy et al.,003). For data analysis, probe set signal intensities were ex-racted with the robust multi-array average (RMA) algorithmIrizarry et al., 2003). This algorithm is a part of the R package affynd can be downloaded from the Bioconductor project websitehttp://www.bioconductor.org).

icroarray data analysis

ata were imported into a Microsoft Excel spreadsheet whereignal averages and standard deviations were calculated for eachene across the triplicate arrays. As these studies were explor-tory and sample number precluded systematic statistical analy-is, ratios of signal expression were calculated to seek genes withnrichment in expression levels in dorsal and ventral hippocam-us, relative to the cortical samples. Results for selected genesere confirmed by in situ hybridization.

igoxigenin (DIG)-labeled in situ hybridization

sing published reference sequences, we designed primers corre-ponding to the candidate genes and amplified 600–800 base-pairNA fragments from mouse brain cDNA. Amplified fragments wereloned into the pCRII-TOPO TA vector (Invitrogen), and sequence-onfirmed plasmids were linearized and used as templates to gen-rate sense and anti-sense DIG-labeled riboprobes. The exceptionsere the probe template for the gene Wolframin, which had beenub-cloned in to the pBluescript SK vector (Stratagene, La Jolla, CA,SA). Riboprobes were transcribed using the DIG RNA Labeling Kit

Roche, Manheim, Germany) in accordance with kit instructions,sing T7, Sp6, or T3 polymerases. In situ hybridization was per-ormed as previously described (Schaeren-Wiemers and Gerfin-oser, 1993), with hybridization temperature increased as needed tonsure low background. For cryo-sectioning, brains were mountedn optimal cutting temperature medium (TissueTek; Sakura Finetek,orrance, CA, USA). Fourteen-micron sections were cut on a LeicaM3050 S cryostat (Bannockburn, IL, USA) and mounted on Super-

rost slides (Fisher, Fairlawn, NJ, USA). Two sections containingorsal CA1 and two containing ventral CA1 were mounted on eachlide to ensure similarity in processing, with dorsal CA1 defined asregma �1.70 mm to �2.30 mm and ventral CA1 containing sec-

ions as Bregma �3.16 mm to �3.50 mm. DIG was detected withnti-DIG AP Fab Fragments (Roche), and visualized with BCIP/NBTolution (Roche). As a reference, slides containing immediately ad-

acent sections were Nissl stained according to standard protocols.Alkaline-phosphatase signals were allowed to develop for a

eriod of six hours up to three days, depending on signal intensitynd background. BCIP/NBT solution was changed every 24 h aseeded. Slides were coverslipped with DAKO Glycergel mounting

af

maging system (QImaging, Burnaby, Canada) and a Zeiss Axio-ert 200 (Oberkochen, Germany) fitted with a 2.5� and a 10�bjective. Images were assembled, using Adobe Photoshop. Final

mages were contrast enhanced and color matched for easieriewing.

RESULTS

rray data

ur aim was to identify genes that were differentially ex-ressed along the dorsoventral axis of the CA1 region ofhe hippocampus enriched in the hippocampus relative tonderlying cortex. An overview of the data reveals that ofhe 22,626 probes on the array, 6764 (29.9%) were “ab-ent” (i.e. did not have expression above background lev-ls). A number of genes showed enrichment in the corticalamples over the CA1 regions, with 233 genes showingreater than two-fold enrichment in the cortical samplever dorsal CA1 and 209 genes showing greater thanwo-fold enrichment in cortical sample over ventral CA1.enes that were more than two-fold enriched in hippocam-us numbered 158 in dorsal hippocampus and 139 inentral hippocampus.

These numbers are consistent with other studies re-orting regionally enriched signals on microarrays be-ween brain regions (Zirlinger et al., 2001).

Within CA1, we identified 41 genes with greater thanwo-fold enrichment in the dorsal CA1 samples relative tohe ventral CA1 samples and 67 genes with two-fold en-ichment in ventral relative to dorsal CA1 samples (Fig. 2).esults showed remarkable consistency between arrays

Supplemental Fig. 1). Genes with highest “fold enrich-ent” across the D–V axis are shown in Table 1. Expres-

ig. 2. Normalized dorsal CA1 versus ventral CA1 cDNA microarrayesults on log-log plots. Genes that show a greater than two-foldnrichment in the dorsal CA1 sample are shown in green (41 total).enes that show a greater than two-fold enrichment in the ventral CA1ample are shown in red (67 total). The remaining 22582 probe sets

rom three separate samples.

sp

V

OapHmesithpf

rs

bOimpamHtevt

T

An

D

V

a

b

c

d

e

f

g

E. D. Leonardo et al. / Neuroscience 137 (2006) 177–186180

ion levels for all 22,626 probe sets can be found in Sup-lemental Table 1.

alidation of microarray results

ther studies have demonstrated the usefulness and reli-bility of microarrays in identifying genes that are ex-ressed in a region–specific manner (Lein et al., 2004).owever, microarray results need to be validated by othereans, both to confirm their accuracy, and to providexpression resolution at the cellular level within a givenub-region. We used our array data as a starting point fordentifying genes differentially expressed in CA1 acrosshe D–V axis. We sorted the array data looking for theighest dorsal to ventral and ventral to dorsal ratios anderformed in situ hybridization on the top candidate genes

able 1. Genes enriched in dorsal and ventral CA1

ccessionumber

Description

orsally enrichedC78577 EST weakly similar to Klotho lactose-phlorizin hBG141874 Transthyretinb,c

NM_013467 Aldehyde dehydrogenase family 1, subfamily 1NM_011716 Wolframina

BB075797 EphA7: Ephrin receptor A7a

X79084 EphA7BB292785 EphA3: Ephrin receptor A3BC026153 EphA7BQ175355 Klothob (Tsujikawa, et al. 2003)BB831146 CCAAT/enhancer binding protein (C/EBP)AV152953 Transthyretinb

BC013499 Hypothetical protein MGC19163entrally enrichedNM_007833 Decorina

BC024515 Hypothetical protein LOC22564 (Gastrin-releas(Shumyatsky, et al. 2002)

AV218841 Neuromatina

NM_007706 Suppressor of cytokine signaling 2 (SOCS-2)f (NM_030699 Netrin G-1 (Laminet-1)g (Yin, et al. 2002)NM_138649 Brain&Kidney proteina

AW0448713 EST’sNM_011594 Tissue inhibitor of metalloproteinase 2CR1601 Tissue inhibitor of metalloproteinase 2BB244736 Expressed sequence AI527257AF144628 Slit homolog 2 (Drosophila)NM_026214 RIKEN cDNA 2210017A09 geneBE653037 Coatomer protein complex, subunit �2BF168458 Tissue inhibitor of metalloproteinase 2BB31657A Coatomer protein complex, subunit �2NM_009049 Regulated endocrine-specific protein 18BB426294 RIKEN cDNA 2810454P21 gene

Confirmed here by DIG-labeled in situ hybridization.Expressed in choroid plexus—probable dissection contamination.Confirmed by GENSAT databank.Expressed in meningeal channels around the hippocampus.Confirmed by literature.Expressed in CA3—probable dissection contamination.Expressed in thalamus and hippocampal interneurons.

or which expression data were not published. Good cor- t

espondence was seen between genechip results and initu data (see below).

A number of the genes with high fold enrichment haveeen described in the literature, as indicated in Table 1.ne known expression pattern confirmed our array find-

ngs, while others suggested some contamination in ouricrodissection. Specifically, gastrin-releasing peptide ap-ears to be expressed in the ventral CA1 region, but isbsent in CA1 dorsally (Shumyatsky et al., 2002), in agree-ent with the D–V signal enrichment we found of 6.86.owever, two genes with apparent dorsal enrichment,

ransthyretin and klotho, actually revealed the higher pres-nce of choroid plexus RNA material in our dorsal versusentral samples, reflecting the difficulty of removing thisissue during the dissection of the dorsal samples. Addi-

Raw signal Fold enrichment

DorsalCA1

VentralCA1

Dorsal/ventral

Dorsal/cortex

a (LPH) 521�43 36�2.9 14.45 16.98140�100 27�4.4 5.14 �6.80

er, et al. 2002) 396�12 102�31 3.87 1.473217�521 888�133 3.62 6.99382�82 106�5.4 3.60 6.79

1043�54 303�36 3.44 5.94130�25 42�5.9 3.06 4.10160�46 56�8.1 2.89 4.20215�50 75�8.5 2.87 4.11161�20 62�2.6 2.60 4.1933�22 13�1.3 2.54 �16.971�5 28�1.6 2.53 2.24

129�18 1446�136 11.24 7.17e)e 105�26 723�74 6.86 2.79

351�23 2335�451 6.66 6.02, et al. 2000) 55�8 226�60 4.12 1.14

86�20 337�41 3.91 1.68209�2.3 811�47 3.89 4.35243�15 939�118 3.87 3.84194�10 727�5 3.75 1.79

1007�47 3638�146 3.61 1.7727�5 90�27 3.36 1.37

125�7.2 417�68 3.34 1.29107�19 357�101 3.32 2.87213�37 703�171 3.30 �2.35789�34 2562�215 3.25 1.79182�56 576�120 3.16 �2.15113�24 354�60 3.14 �1.13239�26 750�68 3.13 2.04

ydrolase

Ad (Wagn

ing peptid

Polizzotto

ionally, the gene SOCS-2, which showed ventral enrich-

mrsafn

V

WDhpatrpT

sCa

spq(cdesn

D

saiAparsm

FWW

E. D. Leonardo et al. / Neuroscience 137 (2006) 177–186 181

ent in our screen, is expressed exclusively in the CA3egion within the hippocampus (Polizzotto et al., 2000),uggesting possible contamination of ventral CA1 withdjacent CA3 tissue. Lastly, aldehyde dehydrogenaseamily 1, subfamily 1a is expressed in the meningeal chan-els around the hippocampus (Wagner et al., 2002).

alidation by in situ hybridization

e sought to confirm differences in expression across the–V axis for the top three genes in each group that did notave published expression data. All showed significant ex-ression differences between dorsal and ventral CA1 areasnd were significantly enriched in hippocampus versus cor-ex. Specifically, LPH, Wolframin, and the EphA7 were en-iched in dorsal CA1 (Fig. 3), whereas Decorin, Brain&Kidneyrotein, and Neuronatin were enriched in ventral CA1 (Fig. 4).hese patterns are discussed in detail below.

While the genes that we verified by in situ primarilyhowed differences attributable to pyramidal neurons inA1, other cell types demonstrating mRNA expressioncross the D–V axis of the hippocampus were also repre-

ig. 3. Expression patterns of select dorsally enriched genes by in situ

ented in the array data. For instance, vasoactive intestinalolypeptide (VIP)-positive interneurons are found more fre-uently in ventral hippocampus than in dorsal hippocampusJinno and Kosaka, 2003). Despite being only a small per-entage of the total cell type within the dissected tissue, thisifference was detected by our microarray as a 1.78-foldnrichment of VIP signal in the ventral hippocampus. Thisuggests the existence of additional molecular heteroge-eity along the D–V axis beyond genes verified here.

orsally enriched in CA1

Wolframin (Fig. 3a and g). This gene is highly re-tricted to CA1 within the hippocampus and is expressed ingraded fashion, with strong expression dorsally dimin-

shing completely in the ventral third of CA1 (Fig. 5c).dditional, weaker, expression outside of the hippocam-us is seen in the piriform cortex and the amygdala (Fig. 3nd data not shown). Mutations in this gene in humans areesponsible for Wolfram syndrome, an autosomal reces-ive disease characterized by diabetes insipidus, diabetesellitus, optic atrophy, and deafness. In addition, people

tion. (a–c) Shows DIG-labeled in situ hybridization for the genes LPH,

ws(gs

iCasbkcrR

dashamo

Voppad

V

Capekfs

iaied

FN hippocamh ons conta

E. D. Leonardo et al. / Neuroscience 137 (2006) 177–186182

ho carry the gene but do not have the disorder haveignificantly increased risk for psychiatric hospitalizationSwift et al., 1998). More recently, polymorphisms in thisene have been associated with increased risk of depres-ion (Koido et al., 2005).

LPH (Fig. 3b, h). Robust expression is seen dorsallyn CA1 and DG; ventrally it is not expressed in CA1, orA3, and gradually diminishes in the DG, and is indeedbsent in the most ventral cells of the DG in the ventrallice (Fig. 6c). While the specific function for LPH in therain has not been identified, LPH purified from the mon-ey gut has been shown to hydrolyze both lactose anderebrosides, suggesting that LPH may be modifying ce-ebrosides in the hippocampus (Ramaswamy andadhakrishnan, 1975).

EphA7 (Fig. 3c and i). Strong expression is seenorsally in CA1, CA2, CA3, and the DG. The expression inll regions appears graded, with the diminishing expres-ion in more ventral areas (Figs. 5a, 6a). Outside theippocampus, it is expressed in the piriform cortex. Eph’sre extensively involved in regionalization and topographicapping of axonal projections during development and are

ig. 4. Expression patterns of select ventrally enriched genes by in sieuronatin, and Brain&Kidney protein in sections containing dorsalybridization for Decorin, Neuronatin and Brain&Kidney protein in secti

ften found in graded expression patterns (Flanagan and a

anderhaeghen, 1998). Eph expression in the adult brainccurs primarily at sites that undergo ongoing remodeling/atterning (Yamaguchi and Pasquale, 2004), raising theossibility that presence of EphA7 in a graded pattern indulthood may contribute to the maintenance of regionalifferences.

entrally Enriched in CA1

Decorin (Fig. 4a and g). Dorsally it is not detected inA1, whereas strong expression is seen in ventral CA1nd subiculum. Outside of the hippocampus, weaker ex-ression is seen in the endopiriform nucleus as well as thentorhinal cortex. Decorin is a small proteoglycan that isnown to bind and inhibit the action of transforming growthactor-�, which in turn may help regulate neurotrophinignaling (Sometani et al., 2001).

Neuronatin (Fig. 4b and h). Dorsally it is not detectedn CA1, although weak expression is seen in both the DGnd CA3 regions. Ventrally it shows scattered expression

n CA1, expression in CA3, and in the subiculum. Weakxpression outside the hippocampus is seen in the amyg-ala, the habenular nucleus, regions of the hypothalamus,

zation. (a–c) DIG-labeled in situ hybridizations for the genes Decorin,pus. (d–f) Adjacent Nissl sections for reference. (g–i) Show in situining ventral hippocampus. (j–l) Adjacent Nissl sections for reference.

tu hybridi

s well as the periventricular medial and lateral magnocel-

lh(

esCueBbs

cdtApg(p

sNohdegrfsetc

Ndtda

Fo al but no

E. D. Leonardo et al. / Neuroscience 137 (2006) 177–186 183

ular areas (Fig. 4 and data not shown). Neuronatin is aighly imprinted gene with currently unknown functionJohn et al., 2001).

Brain&Kidney protein (Fig. 4c and i). It is very weaklyxpressed in the most dorsal regions of CA1 (data nothown) and the adjacent CA2 region. Ventral regions ofA1 show uniform expression as does the ventral subic-lum. In CA3 it is expressed only ventrally while in the DGxpression is seen across the D–V axis. The function ofrain&Kidney protein is unknown, although it appears toe localized to the trans-Golgi network (Fukuda and Miko-hiba, 2001).

In addition to gene expression differences within oneytoarchitecturally-defined region, some genes also showifferential patterns of D–V regulation throughout the major

ripartite hippocampal sub-regions (CA1, CA3, and DG).mong the genes that we identified and confirmed, severalatterns are apparent. First, EphA7 is expressed in araded pattern in all three major regions, CA1, CA3, DGFig. 3c, 3i and Figs. 5a, 6a). One gene, Brain&Kidney

Eph

A7

Wol

fram

in

c

ig. 5. (a) EphA7 in situ at the dorsal ventral boundaries of expression.f expression. (d) Adjacent Nissl section. Note expression in the dors

rotein, shows co-varying expression across two major g

ubregions and static expression in the third. Another,euronatin, also co-varies in a graded fashion across twout of three major sub-regions in which it is expressed;owever, in this case the regions are regulated in oppositeirections, with expression in dorsal but not ventral DG andxpression in ventral but not dorsal CA1. Although theseenes are but a small sample of those that showed en-iched D–V ratios across CA1, these patterns likely existor other genes as well. Interestingly, all three genes thathowed enriched expression in ventral CA1 also showedxpression in the ventral subiculum, raising the possibilityhat these two structures share a cellular population or aommon origin.

DISCUSSION

umerous groups have successfully used microarrays toefine molecular markers of previously recognized ana-omical structures, for example in sub-regions of the amyg-ala (Zirlinger et al., 2001) and the hippocampus (Lein etl., 2004). In general, those studies have concluded that

b

d

ent Nissl section. (c) Wolframin in situ at the dorsal ventral boundariest ventralmost region of CA1 for both genes (double-headed arrows).

a

(b) Adjac

ene expression patterns respect regional boundaries that

hweC

wttstratrnussgai

lat

nrwttmtmsua

AiCa

F( ventral

E. D. Leonardo et al. / Neuroscience 137 (2006) 177–186184

ave previously been established cytoarchitecturally. Heree present data from a screen for genes that are differentiallyxpressed within one cytoarchitecturally-defined structure,A1, across its longitudinal (dorsal-to-ventral) axis.

We have identified and confirmed at least six genesith differential expression across the D–V axis of CA1 in

he adult mouse. In addition, our screen produced geneshat showed differential regulation across the D–V axis inub-regions of the hippocampus other than CA1. Fromhese results we can conclude that: 1) differential geneegulation occurs within traditionally defined sub-regionscross the D–V axis, and 2) expression across this axis of

he hippocampus can be dependently or independentlyegulated among sub-regions, i.e. the hippocampus doesot necessarily vary dorsally-to-ventrally as a concertednit. These findings not only contradict the archetype ofub-region homogeneity within the hippocampus but alsouggest the possibility that more subtle functional hetero-eneity exists beyond the D–V distinction defined bynatomical connectivity and gross lesion studies. The

LPH

Eph

A7

c

ig. 6. (a) EphA7 in situ showing fading CA3 expression (double-heb) Adjacent Nissl section. (c) LPH in situ showing fading signal in the

dentification of differential regulation of genes across the d

ongitudinal axis among subregions may elucidate yetnother layer of complexity in the septotemporal differen-iation of the hippocampus.

Identification of these differentially expressed genesot only provides molecular markers for new putative neu-onal subpopulations, but also provides a molecular handleith which to study new populations of neurons. The func-

ions of genes described here may themselves contributeo neurochemical and/or cellular properties, which in turnay help to define new subclasses of neurons. Moreover,

aking advantage of transgenic technology using the pro-oters for these genes offers the possibility of mapping

pecific functions to these subpopulations, increasing ournderstanding of the relationships between connectivitynd function in the brain.

cknowledgments—We would like to thank Lauren Tanabe fornitial work generating in situ probes, Vladan Miljkovic at theolumbia Microarray Project for performing array hybridizations,nd Paul Pavlidis for performing RMA extraction of Affymetrix

b

d

w) and loss of expression in the ventralmost dentate (single arrow).DG (double-headed arrow). (d) Adjacent Nissl section.

a

aded arro

ata. This work was supported by an APA/Wyeth Pharmaceuticals

MYMR

A

A

B

B

B

B

C

E

E

F

F

G

G

I

J

J

J

K

K

K

L

M

P

P

R

R

R

S

S

S

S

S

S

S

T

W

W

Y

Y

E. D. Leonardo et al. / Neuroscience 137 (2006) 177–186 185

.D./Ph.D. Psychiatric Research Fellowship and a NARSADoung Investigator Award to E.D.L., grants P01 MH060970,H068542-1 and a NARSAD Distinguished Investigator Award to.H. and grant MH067721 to E.S.

REFERENCES

maral DG, Witter MP (1989) The three-dimensional organization ofthe hippocampal formation: a review of anatomical data. Neuro-science 31:571–591.

nderson P, Bliss TV, Skrede KK (1971) Lamellar organization ofhippocampal pathways. Exp Brain Res 13:222–238.

annerman DM, Deacon RM, Offen S, Friswell J, Grubb M, RawlinsJN (2002) Double dissociation of function within the hippocampus:spatial memory and hyponeophagia. Behav Neurosci 116:884–901.

annerman DM, Rawlins JN, McHugh SB, Deacon RM, Yee BK, BastT, Zhang WN, Pothuizen HH, Feldon J (2004) Regional dissocia-tions within the hippocampus: memory and anxiety. NeurosciBiobehav Rev 28:273–283.

annerman DM, Yee BK, Good MA, Heupel MJ, Iversen SD, RawlinsJN (1999) Double dissociation of function within the hippocampus:a comparison of dorsal, ventral, and complete hippocampal cyto-toxic lesions. Behav Neurosci 113:1170–1188.

urwell RD, Amaral DG (1998) Perirhinal and postrhinal cortices of therat: interconnectivity and connections with the entorhinal cortex.J Comp Neurol 391:293–321.

olombo M, Fernandez T, Nakamura K, Gross CG (1998) Functionaldifferentiation along the anterior–posterior axis of the hippocampusin monkeys. J Neurophysiol 80:1002–1005.

ichenbaum H (1997) Declarative memory: insights from cognitiveneurobiology. Annu Rev Psychol 48:547–572.

isen MB, Spellman PT, Brown PO, Botstein D (1998) Cluster analysisdisplay of genome-wide expression patterns. Proc Natl Acad SciU S A. 95:14863–14868.

lanagan JG, Vanderhaeghen P (1998) The ephrins and Eph re-ceptors in neural development. Annu Rev Neurosci 21:309 –345.

ukuda M, Mikoshiba K (2001) The N-terminal cysteine cluster isessential for membrane targeting of B/K protein. Biochem J360:441–448.

alfalvy HC, Erraji-Benchekroun L, Smyrniotopoulos P, Pavlidis P,Ellis SP, Mann JJ, Sibille E, Arango V (2003) Sex genes forgenomic analysis in human brain: internal controls for comparisonof probe level data extraction. BMC Bioinformatics 4:37.

ray JA, McNaughton N (2000) The neuropsychology of anxiety: anenquiry into the functions of the septo-hippocampal system.Oxford: Oxford University Press.

rizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ,Scherf U, Speed TP (2003) Exploration, normalization, and sum-maries of high density oligonucleotide array probe level data. Bio-statistics 4:249–264.

inno S, Kosaka T (2003) Patterns of expression of neuropeptides inGABAergic nonprincipal neurons in the mouse hippocampus:Quantitative analysis with optical disector. J Comp Neurol 461:333–349.

ohn RM, Aparicio SA, Ainscough JF, Arney KL, Khosla S, HawkerK, Hilton KJ, Barton SC, Surani MA (2001) Imprinted expressionof neuronatin from modified BAC transgenes reveals regulationby distinct and distant enhancers. Dev Biol 236:387–399.

ung MW, Wiener SI, McNaughton BL (1994) Comparison of spatialfiring characteristics of units in dorsal and ventral hippocampus ofthe rat. J Neurosci 14:7347–7356.

jelstrup KG, Tuvnes FA, Steffenach HA, Murison R, Moser EI, MoserMB (2002) Reduced fear expression after lesions of the ventralhippocampus. Proc Natl Acad Sci USA 99:10825–10830.

oido K, Koks S, Nikopensius T, Maron E, Altmae S, Heinaste E,

palu A, Vasar E (2005) Polymorphisms in wolframin (WFS1) geneare possibly related to increased risk for mood disorders. IntJ Neuropsychopharmacol 8:235–244.

osel KC, Van Hoesen GW, West JR (1981) Olfactory bulb projectionsto the parahippocampal area of the rat. J Comp Neurol 198:467–482.

ein ES, Zhao X, Gage FH (2004) Defining a molecular atlas of thehippocampus using DNA microarrays and high-throughput in situhybridization. J Neurosci 24:3879–3889.

cHugh SB, Deacon RM, Rawlins JN, Bannerman DM (2004) Amyg-dala and ventral hippocampus contribute differentially to mecha-nisms of fear and anxiety. Behav Neurosci 118:63–78.

itkanen A, Pikkarainen M, Nurminen N, Ylinen A (2000) Reciprocalconnections between the amygdala and the hippocampal forma-tion, perirhinal cortex, and postrhinal cortex in rat. A review. Ann NY Acad Sci 911:369–391.

olizzotto MN, Bartlett PF, Turnley AM (2000) Expression of “sup-pressor of cytokine signalling” (SOCS) genes in the developingand adult mouse nervous system. J Comp Neurol 423:348 –358.

amaswamy S, Radhakrishnan AN (1975) Lactase-phlorizin hydro-lase complex from monkey small intestine. Purification, propertiesand evidence for two catalytic sites. Biochim Biophys Acta403:446–455.

amón y Cajal S (1953) Histologie du systáeme nerveux de l’homme& des vertâebrâes. Madrid: Consejo Superior de InvestigacionesCientâificas, Instituto Ramâon y Cajal.

isold PY, Swanson LW (1996) Structural evidence for functionaldomains in the rat hippocampus. Science 272:1484–1486.

chaeren-Wiemers N, Gerfin-Moser A (1993) A single protocol todetect transcripts of various types and expression levels in neuraltissue and cultured cells: in situ hybridization using digoxigenin-labelled cRNA probes. Histochemistry 100:431–440.

humyatsky GP, Tsvetkov E, Malleret G, Vronskaya S, Hatton M,Hampton L, Battey JF, Dulac C, Kandel ER, Bolshakov VY (2002)Identification of a signaling network in lateral nucleus of amygdalaimportant for inhibiting memory specifically related to learned fear.Cell 111:905–918.

ometani A, Kataoka H, Nitta A, Fukumitsu H, Nomoto H, FurukawaS (2001) Transforming growth factor-beta1 enhances expres-sion of brain-derived neurotrophic factor and its receptor, TrkB,in neurons cultured from rat cerebral cortex. J Neurosci Res66:369 –376.

quire LR, Stark CE, Clark RE (2004) The medial temporal lobe. AnnuRev Neurosci 27:279–306.

trange B, Dolan R (1999) Functional segregation within the humanhippocampus. Mol Psychiatry 4:508–511.

trange BA, Dolan RJ (2001) Adaptive anterior hippocampal re-sponses to oddball stimuli. Hippocampus 11:690–698.

wift RG, Polymeropoulos MH, Torres R, Swift M (1998) Predisposi-tion of Wolfram syndrome heterozygotes to psychiatric illness. MolPsychiatry 3:86–91.

sujikawa H, Kurotaki Y, Fujimori T, Fukada K, Nabeshima Y (2003)Klotho, a gene related to a syndrome resembling human prematureaging, functions in a negative regulatory circuit of vitamin D endo-crine system. Mol Endeocrinol. 17:2393–2403.

agner E, Luo T, Drager UC (2002) Retinoic acid synthesis in thepostnatal mouse brain marks distinct developmental stages andfunctional systems. Cereb Cortex 12:1244–1253.

itter MP, Groenewegen HJ, Lopes da Silva FH, Lohman AH (1989)Functional organization of the extrinsic and intrinsic circuitry of theparahippocampal region. Prog Neurobiol 33:161–253.

amaguchi Y, Pasquale EB (2004) Eph receptors in the adult brain.Curr Opin Neurobiol 14:288–296.

in Y, Miner JH, Sanes JR (2002) Laminets: laminin- and netrin-related genes expressed in distinct neuronal subsets. Mol Cell

Z

ZS

Si

E. D. Leonardo et al. / Neuroscience 137 (2006) 177–186186

hao X, Lein ES, He A, Smith SC, Aston C, Gage FH (2001) Tran-scriptional profiling reveals strict boundaries between hippocampalsubregions. J Comp Neurol 441:187–196.

irlinger M, Kreiman G, Anderson DJ (2001) Amygdala-enrichedgenes identified by microarray technology are restricted to spe-cific amygdaloid subnuclei. Proc Natl Acad Sci USA 98:

APPENDIX

upplementary data

upplementary data associated with this article can be found,n the online version, at doi: 10.1016/j.neuroscience.2005.

(Accepted 30 August 2005)(Available online 23 November 2005)