developmentally regulated expression of thy-1 in structures of the mouse sensory-motor system

Developmentally Regulated Expressionof Thy-1 in Structures of the Mouse

Sensory-Motor System

JENNIFER Z. BARLOW AND GEORGE W. HUNTLEY*

Fishberg Research Center for Neurobiology and Program in Cell Adhesion, The MountSinai School of Medicine, New York, New York 10029-6574

ABSTRACTThy-1 is a cell-surface molecule of the immunoglobulin superfamily which is expressed at

high levels in the mature nervous system. Thy-1 has been implicated in regulating axonaloutgrowth and synaptic function, but little is known regarding its cellular localization andexpression in the central nervous system (CNS) during development or in adulthood. In thisstudy, Thy-1 gene expression and protein localization were examined in sensory-motor andrelated areas of the adult and postnatally developing mouse CNS. Thy-1 mRNA expressionwas restricted to neurons; immunoreactivity was densely distributed throughout the neuropilof all regions examined, often delineated the neuronal plasmalemma, and labeled axons inwhite matter tracts of the brain and spinal cord. In adulthood, immunolabeling was region-ally widespread and was present relatively homogeneously throughout all cell-dense layers ofsensory-motor cortex, throughout most thalamic nuclei, globus pallidus, and spinal cord.Developmentally, however, Thy-1 expression and localization exhibited a spatially and tem-porally staggered sequence leading to the adult pattern. In sensory-motor cortex, Thy-1expression in layer V preceded expression in other layers; in the barrel field, labeling of barrelsepta preceeded a gradually increasing intensity of immunolabeling of barrel centers; in thethalamus, Thy-1 exhibited a differential onset and temporal pattern of expression acrossdifferent nuclei associated with motor, sensory, or limbic systems; in the caudate nucleus,Thy-1 expression was greatest during the first postnatal week of life before declining duringsubsequent development. Taken together, the adult distribution and developmental patternsleading to it form a unique profile in comparison with other structurally related glycosyl-phosphatidylinositol (GPI)-anchored neural cell adhesion molecules. The pattern and timing ofThy-1 expression across layers and nuclei during early postnatal development are more complexthan previously recognized, thus perhaps reflecting varied roles for Thy-1 in aspects of structuralor functional maturation which proceed independently of the timing of neurogenesis, migration,and dendritic and axonal growth. J. Comp. Neurol. 421:215–233, 2000. © 2000 Wiley-Liss, Inc.

Indexing terms: axon guidance; synaptogenesis; immunoglobulin superfamily; spinal cord;

corticospinal tract; thalamocortical connections; barrel cortex

Development of function in the central nervous systemdepends on the establishment of highly specific intercon-nections between neurons which can be distributed acrosswidely separated brain regions. The formation of neuralconnections arises through cell-cell, cell-matrix, or se-creted molecular cues which allow growing axons to nav-igate different cellular environments and to recognize ap-propriate target regions and cells (Tessier-Lavigne andGoodman, 1996). The neuronal cell surface expresses dif-ferent types and amounts of cell adhesion molecules(CAMs) and related signaling molecules; the interactionsbetween these and their substrates activate multiple sig-

nal transduction cascades which not only regulate devel-opmental processes such as growth, target recognition,and synapse formation, but are also important for main-

Grant sponsor: National Institutes of Health, U.S. Public Health Service;Grant number: NS34659.

*Correspondence to: Dr. G.W. Huntley, Fishberg Research Center forNeurobiology and Program in Cell Adhesion, Box 1065, The Mount SinaiSchool of Medicine, 1425 Madison Avenue, New York, NY 10029-6574.E-mail: [email protected]

Received 2 June 1999; Revised 20 January 2000; Accepted 21 January2000

THE JOURNAL OF COMPARATIVE NEUROLOGY 421:215–233 (2000)

© 2000 WILEY-LISS, INC.

taining and stabilizing functional connectivity once estab-lished (Lander, 1989; Furley et al., 1990; Yoshihara et al.,1994; Pimenta et al., 1995; Tuttle et al., 1995; Hortsch,1996; Tessier-Lavigne and Goodman, 1996; Walsh andDoherty, 1997; Benson and Tanaka, 1998).

Thy-1 is a glycosyl-phosphatidylinositol (GPI)-linked,cell-surface molecule of the immunoglobulin superfamilyfound in many tissues, but most abundantly in brain andthymus (Morris, 1985). The onset of Thy-1 expression inbrain is largely postnatal, reaching its highest level inadulthood (Morris and Barber, 1983), where it may con-stitute as much as 2.5–7% of the total surface protein(Beech et al., 1983) of both axonal and somatodendriticcompartments (Morris et al., 1985; Xue et al., 1990; Dottiet al., 1991). Previous studies using in vitro assays providestrong evidence to suggest that Thy-1 functions to inhibitprocess outgrowth, thereby representing a possible molec-ular mechanism for stabilizing neural connectivity onceformed (Morris, 1985; Xue et al., 1990, 1991; Mahan-thappa and Patterson, 1992a,b; Tiveron et al., 1992). Incontrast, results from Thy-1 knockout mice show thatsuch animals develop, and maintain, apparently normalneuroanatomical features (Nosten-Bertrand et al., 1996;unpublished observations). However, changes in gammaaminobutyric acid (GABA)ergic inhibitory synaptic func-tion in hippocampus have been described in Thy-1 knock-out mice (Hollrigel et al., 1998), suggesting that Thy-1may also have a role in regulating channel properties orother membrane constituents of cell-signaling complexes.Although the precise function of Thy-1 remains to be de-termined, the abundance of the protein, its steadily risinglevels of expression during development, and its potentialrole in modulating synaptic signaling in adulthood allsuggest the molecule has important and varied roles dur-ing establishment of brain structure and function.

One strategy for understanding when and in which cir-cuits Thy-1 may be operative is to examine anatomicallythe temporal, spatial, and cellular features of Thy-1 ex-pression and localization during development and in ma-turity. To date, anatomical studies of Thy-1 in the nervoussystem have focused on broad regional surveys of immu-noreactivity (Reif and Allen, 1964; Barclay and Hyden,1978; Morris et al., 1983), although more comprehensivecellular analyses of Thy-1 gene and protein expressionhave been reported for cerebellum and hippocampus (Mor-ris et al., 1985; Bolin and Rouse, 1986; Xue et al., 1991;

Xue and Morris, 1992). In the present study, we haveexamined Thy-1 gene and protein expression during de-velopment and at adulthood in several functionally re-lated structures of the mouse sensory-motor system, asystem for which there are considerable data on the tim-ing of pathway and neuronal development. In addition,because the distribution of other CAMs which are struc-turally related to Thy-1, such as the limbic system-associated molecule (LAMP) or the olfactory system-related molecule (OBCAM), appears relatively restrictedto specific neural systems (Levitt, 1984; Zhukareva andLevitt, 1995; Hachisuka et al., 1996), we have extendedthe analysis of the thalamus to include nuclei related tolimbic and other sensory systems in order to compare theextent to which Thy-1 localization is also distributedacross different neural systems.

MATERIALS AND METHODS

Animals

All mice used in this study were of the C57BL/6 strain.For immunocytochemistry and in situ hybridization, thelitters of six untimed pregnant mice were used as well aseight young adult mice, aged 2–4 months. Pregnant moth-ers were checked frequently around the expected parturi-tion time; the first 24 hours after birth was designated aspostnatal day (P)0. Both males and females were used atall time points. Adult animals were deeply anesthetizedwith Avertin (12.5 mg/ml i.p.) and perfused through theascending aorta with cold 1% paraformaldehyde dissolvedin 0.1 M phosphate-buffered saline (PBS; pH 7.4) for 1minute followed by 4% paraformaldehyde-PBS for 10 min-utes. For the developmental studies, postnatal animalswere killed at temporally staggered time points rangingfrom P0 to P30; a minimum of three animals was used foreach time point examined. Developing animals wereanesthetized with ketamine (40 mg/kg, i.p.) and perfusedtranscardially with normal saline followed by 4% para-formaldehyde-PBS for 10 minutes. The brains and spinalcords from all animals were removed and postfixed for 6hours in 4% paraformaldehyde-PBS. The treatment of allanimals was in strict accordance with institutional andNIH guidelines.

Abbreviations

AD anterior dorsal nucleusAV anterior ventral nucleusAM anterior medial nucleusCdP caudate-putamencl central lateral nucleuscm centromedian nucleusCP cortical platecs corticospinal tractdc dorsal columnsdLG dorsal lateral geniculate nucleusgp globus pallidusLD lateral dorsal nucleusLP lateral posterior nucleusLhb lateral habenular nucleusMD medial dorsal nucleusMhb medial habenular nucleusnp nucleus proprius

pc paracentral nucleusPo posterior nucleusRe Reuniens nucleusRT reticular nucleusS1 primary somatosensory cortexsb submedial nucleussg substantia gelatinosasm stria medullarisSN substantia nigraSP subplateV lateral ventricleVB ventrobasal nuclear complexVL ventral lateral nucleusvLG ventral lateral geniculate nucleusVM ventral medial nucleuswm white matter

216 J.Z. BARLOW AND G.W. HUNTLEY

Tissue processing and immunocytochemistry

Adult brains were sectioned by one of two differentmethods. Five brains and spinal cords were sectioned on avibratome at a setting of 50 mm and parcelled into threeadjacent series: one for immunocytochemistry; a secondfor in situ hybridization histochemistry; and a third whichwas Nissl-stained with cresyl violet. The remaining threeadult brains and spinal cords were cryoprotected in 30%sucrose and cut into 30-mm-thick sections on a freezingmicrotome. A rat monoclonal anti-Thy-1.2 antibody(30H12; Collaborative, Bedford, MA) was used to localizeThy-1 immunocytochemically in all tissue sections. Non-specific antibody binding was blocked by preincubation for1–2 hours in 4% goat serum and 0.05% bovine serumalbumin (BSA) in PBS. The effects of permeabilization onThy-1 immunolabeling was investigated in both vi-bratome and frozen sections by preincubating some sec-tions for 12–36 hours at 4°C in an identical diluent whichalso contained 0.3% Triton X-100. Sections were then in-cubated for 12 hours at 4°C in diluent containing 4% goatserum and 0.05% BSA in PBS to which primary antibodywas added (1:12,000). For each of the different sectioningand processing conditions, antibody binding was visual-ized in one of two ways: by using a species-appropriateimmunoperoxidase ABC kit (Vector Labs, Burlingame,CA) in the presence of 0.05% diaminobenzidine and0.003% H202, or immunofluorescently by incubation witha biotinylated anti-rat secondary antibody (Vector Labs)followed by streptavidin conjugated to Oregon Green (Mo-lecular Probes, Eugene, OR). All developmental tissue wascut on a vibratome and processed immunocytochemicallywithout incubation with Triton X-100. Antibody bindingwas visualized by immunoperoxidase labeling in the pres-ence of diaminobenzidine and H202 as described.

For cultured rat hippocampal neurons (see below), amouse monoclonal anti-Thy-1.1 antibody (OX-7; Serotec,Bicester, U.K.) was used to localize Thy-1. This differentprimary anti-Thy-1 antibody was used because there aretwo allelic forms of Thy-1 (Thy-1.1 and Thy-1.2), due to aconservative one amino acid substitution. The formpresent in rats is Thy-1.1, whereas most mice have Thy-1.2. Cells were first permeabilized with 0.25% TritonX-100 for 5 minutes followed by preincubation in a solu-tion of 10% normal goat serum in PBS to prevent nonspe-cific antibody binding. Cells were incubated at 4°C for 12hours in PBS diluent containing the anti-Thy-1.1 antibodyand either 1% BSA or 1% goat serum. Antibody bindingwas visualized immunofluorescently by subsequent incu-bation with a goat anti-rat biotinylated secondary anti-body (1:100 in PBS) followed by incubation with Texas-Red conjugated streptavidin (1:200 in PBS, both fromVector).

Two immunocytochemical control procedures were fol-lowed to verify antibody specificity (Fig. 1). First, tissuesections were processed immunocytochemically with theexception that the anti-Thy-1.2 antibody was replacedwith nonimmune serum (Fig. 1a). Second, anti-Thy-1.2antibody binding was localized immunocytochemically byusing perfused brain tissue taken from transgenic micelacking the Thy-1 gene (Fig. 1b). Such Thy-1 null mutantmice were generated by using standard embryonic stemcell techniques as part of a parallel set of forthcomingstudies; the absence of the Thy-1 coding region and mRNAwas confirmed by Southern and Northern blotting, respec-

tively (data not shown). In both sets of controls, no specificimmunolabeling was evident.

Hippocampal cell cultures

Hippocampal cell cultures were prepared from embry-onic day (E)18 Sprague-Dawley rats as described previ-ously (Withers and Banker, 1998). Briefly, cells were dis-sociated by treatment with 0.25% trypsin for 15 minutesat 37°C followed by trituration through a fire-polishedpasteur pipette. The cells were plated at a density of50,000 cells per 60-mm plastic petri dish on poly-L-lysine-coated coverslips in minimum essential medium (MEM;Gibco, Grand Island, NY) containing 10% fetal horse se-rum. When cells attached, coverslips were transferred todishes containing a monolayer of cortical astroglia andmaintained in MEM-containing supplements (Bottensteinand Sato, 1979), sodium pyruvate (1 nM), and ovalbumin(0.1%). Prior to processing for immunocytochemistry, cul-tures were fixed with 4% paraformaldehyde-PBS contain-ing 0.12 M sucrose for 30 minutes at 37°C.

Construction of RNA probes

RNA probes were synthesized from a DNA plasmid tem-plate, pKK62, which contains a 354-bp insert of the secondand third coding exons of the Thy-1.2 gene. Antisensestrand probes were transcribed from BglII-linearizedpKK62 in the presence of 35S-UTP with SP6 RNA poly-merase. Control sense strand probes were transcribedfrom BamHI-linearized pKK62 in the presence of 35S-UTPwith T7 RNA polymerase.

In situ hybridization histochemistry

In situ hybridization histochemistry was performed fol-lowing a modified protocol of Gall and Isackson (1989)

Fig. 1. a–c: Immunocytochemical controls showing specificity ofanti-Thy-1 antibody localization. Parasagittal sections of adult wild-type (a,c) or Thy-1 null mutant (b) mouse brains show a lack ofspecific immunolabeling when nonimmune serum replaced the pri-mary anti-Thy-1 primary antibody (a) or when the anti-Thy-1 anti-body was incubated with brain sections from null mutant mice (b). Incontrast, immunocytochemistry performed on wildtype mice withanti-Thy-1 antibody showed specific immunolabeling (c). d–e: Speci-ficity of Thy-1 probe hybridization. Inverted film autoradiographsshow the lack of specific probe hybridization in sections from wildtypeanimals hybridized with the Thy-1 sense-strand control probe (d),whereas a specific and intense signal (white) was seen when tissuewas hybridized with the Thy-1 antisense probe (e). Scale bars 5 1 mmfor a–c; 2 mm for d,e.

217THY-1 IN THE DEVELOPING SENSORY-MOTOR SYSTEM

Fig. 2. Cellular features of Thy-1 immunolabeling and probe hy-bridization. a, b: Thy-1 immunoreactivity was densely distributedthroughout the neuropil and labeled the plasmalemma of somata(arrows) and proximal dendrites (arrowheads) in layer V of sensory-motor cortex (a), or in the globus pallidus (b). Immunopositive fiberswere evident in white matter tracts such as the subcortical whitematter (arrow, c). The dotted line indicates the boundary betweenwhite matter (wm) and primary somatosensory cortex (S1; asterisk),V indicates the lateral ventricle. d–h: Comparison of immunolabelingquality by using different immunoprocessing conditions. d: Frozen,nonpermeabilized section through cerebral cortex. e, f: Vibratomesections through the ventral horn of spinal cord treated with (e) orwithout (f) Triton X-100. Arrowheads denote large motoneurons. g, h:Confocal images of immunofluorescently labeled fibers cut in cross-section through the corticospinal tract from vibratome sectionstreated with (g) or without (h) Triton X-100. Note the loss of fine

cellular details in the frozen or Triton-permeabilized sections. i, j: Pairof images of the same field of dissociated hippocampal cells grown invitro shown by differential interference contrast optics (i) or by fluo-rescence optics (j) following Thy-1 immunocytochemistry. The astro-cyte (arrowheads) is Thy-1- immunonegative in comparison with theimmunofluorescent neurons. k–l: Thy-1 probe hybridization. Bright-field emulsion autoradiograph (k) showing silver grain clusters indi-cating Thy-1 antisense probe hybridization overlying the larger, palenuclei of neurons (arrow), but not overlying the darker and smallernuclei of neuroglial cells (arrowhead). Section is from layer III of S1cortex. Darkfield photomicrograph (l) showing a few, scattered silvergrains indicating background probe hybridization in the subcorticalwhite matter (wm) in direct contrast to denser hybridization in theneuron-rich layers of the overlying S1 cortex or the underlyingcaudate-putamen (CdP). Scale bars 5 10 mm in a,g,h; 50 mm inb–d,e,f,i,j; 25 mm in k; 100 mm in l.


according to Huntley et al. (1992). Briefly, slide-mountedsections were washed in 0.1 M glycine in 0.1 M phosphatebuffer (pH 7.4), then in proteinase K (1 mg/ml; pH 8);0.25% acetic anhydride in 0.1 M triethanolamine (pH 8);

and two washes of 23 saline sodium citrate (SSC). Probeswere diluted to a concentration of 107cpm/mg in a hybrid-ization solution containing 50% deionized formamide, 503Denhardt’s, 10% dextran sulfate, 0.15 mg/ml yeast tRNA,

Fig. 3. Laminar patterns of Thy-1 immunolabeling and probe hy-bridization change during postnatal development in the somatosen-sory cortex. For each of the indicated ages, triplicate photomicro-graphs of adjacent coronal sections through primary somatosensorycortex (S1) show Thy-1 immunolabeling (left panel), Nissl-stainedcortical lamination (middle panel), and Thy-1 probe hybridization(right panel). Note that Thy-1 immunoreactivity is relatively moreintense in layer V at birth (a) and remains one of the most intenselyimmunoreactive layers into adulthood (f), whereas in contrast, layerIV is lightly labeled through postnatal day (P)6 (d), and remains one

of the most weakly labeled layers into adulthood (f). The intenselabeling at the outermost edges of some of the sections is an edgeartifact. Asterisks in f indicate the four major bands of more intenseimmunoreactivity evident in the adult. Probe hybridization in layer Vrises from birth through the first postnatal week (a–d) and thenbecomes more clustered in layers III and V by adulthood (e,f). Probehybridization intensity in layer IV is very low during the first post-natal week (a–d). Layer IV remains one of the least intensely hybrid-ized layers into adulthood (f). CP, cortical plate; SP, subplate; WM,white matter. Scale bars 5 100 mm.


Figure 4


0.33 mg/ml denatured salmon sperm DNA, and 40 mMdithiothreitol (DTT). Sections were hybridized for 20hours at 50°C in a humidified chamber. Following hybrid-ization, sections were washed in 43 SSC at 50°C, treatedwith ribonuclease A (20 mg/ml; pH 8), then washed in SSCsolutions of increasing stringency to a final wash of 0.13SSC at 50°C for 1 hour. Sections were then dried andexposed to autoradiographic film (b-max, Amersham, Ar-lington Heights, IL) for 2–6 days. Slides were lipid-extracted in 1:1 ethanol and chloroform for 20 minutes,rehydrated and air-dried prior to dipping in Kodak NTB-2 liquid emulsion diluted 1:1 with distilled water. Follow-ing exposure times of 1–2 weeks at 4°C, slides were devel-oped with Kodak D-19 developer at 17°C, fixed, counter-stained with cresyl violet, dehydrated, and coverslipped inDPX (distyrene and xylene). Control procedures consistedof processing sections in the manner described above fol-lowing hybridization with an 35S-UTP-labeled sense ribo-probe (Fig. 1d,e).

Data analysis and figure preparation

Immunoperoxidase-processed tissue was analyzed on aZeiss Axiophot photomicroscope. Immunofluorescentlyprocessed tissue was examined on a Zeiss 410 Laser Scan-ning Confocal microscope; high-resolution confocal imageswere obtained with a 1003 objective and an electroniczoom of 43. All figures were prepared from digital images.Brightfield images of immunoperoxidase material, bright-field and darkfield images of emulsion autoradiographs,and images of Nissl-stained sections were photographed(35-mm film). Film autoradiographs and 35-mm negativeswere scanned; these and the confocal images were im-ported into Adobe Photoshop (Adobe Systems; MountainView, CA). Only minimal adjustments in brightness andcontrast were made, and in some cases separate imageswere photomontaged into a composite assembly. Quark-Xpress (Quark, Inc., Denver, CO) was used to completefinal figure layout, lettering and other graphics.

RESULTS

Determining optimal immunocytochemicalconditions: General features

of immunolabeling

Thy-1 immunoreactivity was present in all structuresexamined in adult mice, which included sensory-motorcortex, thalamus, caudate-putamen, globus pallidus, andspinal cord. Thy-1 localization in adult tissue sections cuton a vibratome and processed without permeabilization byTriton X-100 was characterized by a dense distribution offine-grained reaction product in the neuropil, often ap-pearing punctate (Figs. 2a,b,f; 9a,b). Such neuropilarstaining outlined some unlabeled cell somata (Figs. 2a,b;9a), whereas Thy-1 immunoreactivity also prominentlylabeled the plasmalemma of cell somata and proximaldendrites in all structures examined (Figs. 2a,b,f; 9b).Intense immunoreactivity was also present within whitematter (myelinated) tracts such as the dorsal funiculus ofthe spinal cord (Fig. 2h) and that subjacent to the neocor-tex (Fig. 2c), where clearly labeled fibers were identified.These features of Thy-1 immunolocalization are consis-tent with previous studies using similar (e.g., Shults andKimber, 1993) or different (e.g., Morris and Barber, 1983)tissue fixation and processing methods.

In contrast, the fine cellular details of immunolabelingwere generally diminished or lost entirely in frozen sec-tions processed with or without Triton (Fig. 2d), as well asin those cut on a vibratome and incubated with TritonX-100 (Fig. 2e,g). Thus, results are based primarily onsections cut on a vibratome and processed without TritonX-100.

To verify the cellular identity of Thy-1-expressing cells,Thy-1 immunocytochemistry was performed on dissoci-ated hippocampal cell cultures containing a mixed popu-lation of neurons and glia in which the two cell types canbe readily distinguished morphologically (Withers andBanker, 1998). Under these conditions, Thy-1 immunore-activity was restricted to neurons, with no evidence ofimmunoreactive neuroglial cells (Fig. 2i,j).

General features of probe hybridization

The cellular localization of Thy-1 mRNA transcripts inadult brain and spinal cord was visualized autoradio-graphically, where silver-grain clusters representingprobe hybridization overlay larger, pale Nissl-stained nu-clei characteristic of neurons in comparison with thesmaller, darkly Nissl-stained nuclei typical of neuroglia(Figs. 2k; 4l; 9f). Further, there was no evidence for sig-nificant aggregations of silver grains in any of the whitematter tracts examined (Figs. 2l; 9e), or in the neuron-poor layer I of neocortex, suggesting that Thy-1 gene ex-pression was mostly, if not completely, restricted to neu-rons.

Regional and developmental patterns ofThy-1 immunolabeling and gene expression

Sensory-motor cortex

Adult. The regions of sensory-motor cortex analyzedincluded the lateral agranular field (primary motor cortex;Wise and Donoghue, 1986) and granular primary somato-sensory cortex (S1). The overall patterns of Thy-1 proteinand mRNA transcript localization in adults, and the de-velopmental sequence leading to the mature pattern, were

Fig. 4. Thy-1 immunolabeling and probe hybridization delineatetransiently a barrel-related pattern in layer IV of primary somatosen-sory cortex (S1). Tangential sections through layer IV at differentages showing immunolabeling (a–f) or probe hybridization (g–l). Be-tween postnatal day (P)3 and P7 (a–c), more intense immunoreactiv-ity delineates the barrel septa in comparison with light labeling ofbarrel centers (asterisks). By P11 through adulthood (d–f, Ad), immu-nolabeling (top panels) is light and homogeneous across barrel septal/center compartments, which can be seen in adjacent sections pro-cessed for cytochrome oxidase activity (lower panels). Arrows indicatesame blood vessels in matched images. g–k: Top panels are darkfieldimages of emulsion-dipped sections, bottom panels are brightfieldimages of same sections, which have been Nissl-counterstained toindicate cytoarchitectonically the barrels (asterisks denote barrel cen-ters). Probe hybridization shows a developmental pattern in whichdenser silver grain accumulations delineate the barrel septa in com-parison with more diffusely scattered grains within the barrel centers.By adulthood (k), the pattern looks mostly homogeneous. However, ahigher-power image (l) of the boxed region in K shows denser silvergrain accumulations overlying neurons of the barrel septa and walls(brackets), in comparison with lighter probe hybridization to neuronsof the barrel centers (arrowhead). Scale bars 5 100 mm in a–k; 25 mmin l. Scale bar in c also applies to a,b; scale bar in f also applies to d,e;scale bar in k also applies to g–j.


largely similar between these areas; thus, they are de-scribed together as sensory-motor cortex.

All layers of sensory-motor cortex were immunopositivefor Thy-1, although laminar variations in intensity of im-munoreactivity in the neuropil were evident (Fig. 3f). Fourbands of more intense immunoreactive neuropil were ev-ident, corresponding to layer I, the deeper half of layer III,layer V, and layer VIb (Fig. 3f, asterisks). In the remain-ing layers, neuropil labeling was light. Small, labeled so-mata were present in layers II, IV, and VI, whereas largerand more intensely labeled neurons were evident in layersIII and V (Fig. 2a). In the barrel field of S1, analysis offlattened tangential sections through layer IV showedthat Thy-1 immunoreactivity exhibited a moderate, essen-tially homogeneous neuropil labeling pattern throughout,and was not differentially concentrated in barrel centersor septa (Fig. 4f).

The laminar distribution of Thy-1 mRNA transcriptsrevealed autoradiographically was largely similar to thepattern of immunoreactivity (Fig. 3f). Large grain clusterswere evident overlying large somata in layers III and V,whereas less intense probe hybridization was present inthe other layers. The weakest probe hybridization wasobserved in layers II and IV (Fig. 3f). Low-power images oftangential sections through layer IV of the barrel fieldshowed an overall light signal, but unlike the homoge-neous immunolabeling pattern, there were slightlygreater accumulations of silver grains overlying the barrelsepta in comparison with more diffusely scattered grainsin the barrel centers (Fig. 4k). Higher-power images ofsuch sections showed that a larger number of neuronscomprising the barrel walls and septa underlay silvergrain clusters (brackets, Fig. 4l), whereas neurons in thebarrel centers were more lightly labeled, with occasionalones showing denser silver grain accumulations (arrow-head, Fig. 4l).

Developmental sequence

P0–P3. Sensory-motor cortex at P0 (birth) consists ofthe marginal zone, the cell-dense cortical plate, differen-tiated cortical layers V and VI, and the subplate (Fig. 3a).At P0, light Thy-1 immunoreactivity was present in themarginal zone and cortical plate, whereas a prominentband of moderately intense immunoreactivity was presentin layer V, becoming light and diffuse in layer VI and thesubplate (Fig. 3a). Neuropil labeling within these layerswas homogeneous; some weak labeling of the plasma-lemma of cell somata was evident in layer V. Thy-1 im-munoreactivity in layer IV, which differentiated by P2,was very light (Fig. 3b). Tangential sections through lay-ers IV/V of the barrel field showed, by P3, slightly moreintense Thy-1 immunolabeling in the barrel septa in com-parison with the barrel centers (Fig. 4a).

The intensity of Thy-1 mRNA probe hybridization wasvery low at P0. The marginal zone and the cortical plateexhibited background levels of hybridization, whereas inlayer V, a slightly higher concentration of silver grainswas evident at P0 (Fig. 3a). However, by P2, a dense anduniform band of silver grains became evident in layer V,whereas a diffuse scattering of very fine grains could beseen over layer VI and the subplate (Fig. 3b).

P4–P7. All cortical layers found in the adult werepresent by P6 (Fig. 3d). By P4, three bands of more in-tensely immunoreactive neuropil were evident, corre-sponding to layers III, V, and a deep band which spannedlayer VIb and the subplate (Fig. 3c). In layer IV, the septal

pattern of neuropil labeling became, during this period,more distinct than earlier ages (Fig. 4b,c). By P6, theintensity of immunoreactivity in the upper half of layer VIappeared slightly greater than at earlier stages, matchingmore equivalently that of layer V (Fig. 3d), and thusappeared subdivided into a more intense upper half, and aless intense lower half, but with an intense deep bandcorresponding to layer VIb still evident.

Probe hybridization (Fig. 3c,d) became increasinglymore dense in layers II/III. A more prominent hybridiza-tion signal became evident in layer IV, although it wasstill lower in comparison with the higher intensity signalsof adjacent layers III and V. In tangential sectionsthrough layer IV of the barrel field, slightly denser silvergrain accumulations were evident in the barrel septa incomparision with more diffusely scattered grains in thebarrel centers (Fig. 4g). Cells in layer VIb and the sub-plate were overlaid with a denser collection of small grains(Fig. 3d).

P10–21. By P10–12, the intensity of immunoreactiv-ity overall began to increase (Fig. 3e). Tangential sectionsthrough layer IV of the barrel field showed a mostly ho-mogeneous pattern, with the previously distinct labelingof the septa blurred by the increasing levels of immuno-reactivity of the barrel centers (Fig. 4d,e). In addition, theintensity of immunoreactivity increased in the lower por-tion of layer VI, resulting in a uniform distribution whichis characteristic of the adult pattern. By P14, the deeplayer III band of more intense neuropil immunoreactivitybecame evident and remained into adulthood. By P21, anadult pattern of Thy-1 immunoreactivity was presentthroughout.

The laminar pattern of probe hybridization changedslightly during this period, characterized by a gradual emer-gence into the adult pattern of discrete grain clusters inlayers III and V (Fig. 3e). Tangential sections through thebarrel field showed a pattern of probe hybridization in layerIV (Fig. 4h–j) which reached its greatest distinction betweenbarrel septa and centers by P14 (Fig. 4i). The prominentsignal in the barrel septa at P14 stood in contrast to thehomogeneous immunolabeling of the two compartments ev-ident at these stages (compare Fig. 4e and 4i).

Thalamus

Adult. Immunoreactivity was present throughout theneuropil of all nuclei of the dorsal thalamus, the reticularnucleus, and the zona incerta, and was moderate andmostly homogeneous in intensity (Fig. 5e,j,o,t). Moderateimmunolabeling was detected in the lateral habenularnucleus and stria medularis, whereas light immunolabel-ing was evident in the medial habenular nucleus.

A relatively intense hybridization signal, characterizedby large and conspicuous silver grain clusters, was ob-served throughout most dorsal thalamic nuclei and thereticular nucleus (Fig. 6p–r), with the exception of theintralaminar nuclei (Fig. 6q) and the ventral lateral genic-ulate nucleus (vLG; Fig. 6r), which displayed relativelyweak probe hybridization signals.


P0–P4. Thy-1 immunoreactivity was restricted to cer-tain nuclei and displayed differential intensities of immu-nolabeling across nuclei (Fig. 5). The neuropil of anteriordorsal nucleus (AD) and lateral dorsal nucleus (LD) werestrongly immunoreactive (Fig. 5a,b), with prominent so-matic plasmalemmal immunostaining. Moderate neuropil


Fig. 5. a–t: Patterns and intensity of Thy-1 immunolabeling in the thalamus during postnatal development. Each column of brightfield photomicrographsrepresents a rostral (top) to caudalmost (bottom) progression through the thalamus at the indicated age, allowing comparisons across nuclei. Each row of imagesacross ages represents the same approximate level through the thalamus, allowing comparisons of the same set of nuclei across ages. The arrowheads denotethe thalamic reticular nucleus. See text for details. P, postnatal day; Ad, adult. For other abbreviations, see list. Scale bar 5 100 mm.

Fig. 6. a–r: Patterns and intensity of Thy-1 mRNA probe hybridization change during postnatal development in the thalamus. For each set of images,the darkfield photomicrograph of the emulsion-dipped section (left) is matched to the adjacent thionin-stained section (right) which indicatescytoarchitectonically-delineated nuclei. The top row is through a representative anterior (Ant) level, middle row through middle levels (Mid); bottom rowthrough posterior levels (Post). P, postnatal day. For other abbreviations, see list. See text for details. Scale bars 5 300 mm.

Figure 6 (continued)

labeling was evident in medial dorsal nucleus (MD), lat-eral posterior nucleus (LP), posterior nucleus (Po), andventral lateral nucleus (VL; Fig. 5a,b,f,g). All other nucleiof the dorsal thalamus, including ventrobasal nuclearcomplex (VB), dorsal lateral geniculate nucleus (dLG), andvLG, were only very weakly immunoreactive at P0–P2(Fig. 5a,f,k,p). By P3–P4, light immunoreactivity was ev-ident in VB, whereas the dLG acquired more intense la-beling during this period (Fig. 5q). The reticular nucleusand zona incerta were lightly immunoreactive, with mod-erately intense immunoreactivity present in the lateralhabenula and strong labeling of the fibers composing thestria medullaris.

The pattern of probe hybridization in most casesmatched the pattern and intensity of immunolabeling(Fig. 6a–f). An intense signal was evident in AD (Fig.6a,d), with a weaker signal evident in VM, LP, and in-tralaminar nuclei (Fig. 6b,c,e,f). There was no detectablehybridization in anterior ventral nucleus (AV), anteriormedial nucleus (AM), and vLG at P0 (Fig. 6a,c), but weaksignal was present in these nuclei by P4 (Fig. 6d,f). Amoderate signal was present throughout much of the restof the dorsal thalamus at these stages, whereas a rostro-caudal gradient of probe hybridization was evident in thereticular nucleus, with a less intense signal present moreanteriorally (Fig. 6a,d), becoming stronger more posteri-orally (Fig. 6b,c,e,f). A relative mismatch between an in-tense hybridization signal but very light immunolabelingwas evident in VB and dLG through P2–P3 (Figs. 5p, 6c),which persisted in VB through P4 (Fig. 7).

P6–P10. The overall intensity of immunoreactivitygradually increased throughout the dorsal thalamus dur-ing this period (Fig. 5c). The VB complex acquired anadult-like intensity of immunoreactivity by P6 (Fig. 5m,r),whereas the intralaminar nuclei, AM, AV, and vLG alsoacquired more prominent labeling during this time (Fig.5c,h,m,r).

The overall intensity of probe hybridization increased inAV, AM, and VM during this period (Fig. 6g–l) to becomemostly adult-like. In particular, increases in the signalintensity in the intralaminar nuclei and VM were nowevident (Fig. 6j–l), contributing to an overall patternwhich appeared more homogeneous across the thalamusthan at earlier timepoints.

P14–P21. During this period, Thy-1 immunoreactivityin the thalamus became more homogeneous across nuclei,further blurring the previous distinctions between nuclei,and adding to the more uniform appearance of the immu-nostaining characteristic of the adult pattern, which wasachieved by P21 (Fig. 5d,i,n,s).

The overall pattern and intensity of probe hybridizationwas adult-like by these stages (Fig. 6m–o), although thesignal intensity in AD remained, at P21, higher than thatobserved by adulthood (Fig. 6m,p).

Caudate-putamen and globus pallidus

Adult. A mostly homogeneous distribution of light im-munoreactivity was evident throughout the neuropil ofthe caudate-putamen and globus pallidus (Fig. 8i), with nopatches of differential immunostaining evident in theformer which might correspond to the striosome and ma-trix compartments. Some cells exhibited light staining ofthe plasmalemma (arrows, Fig. 8i). The globus palliduswas slightly more intensely immunoreactive than the ad-jacent caudate-putamen (Fig. 8i), with fine-caliber pro-

cesses in the neuropil and distinct labeling of the plasma-lemma of cell somata evident (Fig. 2b).

A very light uniform scattering of silver grains wasobserved overlying the caudate-putamen, not substan-tially different from background levels of hybridization(Fig. 8j). In contrast, large clusters of silver grains wereobserved overlying the globus pallidus (Fig. 8j).


P0–P6. From P0 until P6, a hint of patchy immunore-activity was evident in the caudate-putamen (Fig. 8a,c). Astrong band of immunoreactivity was also evident at themost lateral margin of the nucleus (Fig. 8a,c). The medialcaudate-putamen exhibited moderate neuropilar immuno-staining. Within the globus pallidus, the more medialportion of the nucleus displayed stronger, but homoge-neous immunoreactivity in comparison with the ventro-lateral portion which was more moderately immunoreac-tive (Fig. 8a,c).

In situ hybridization histochemistry showed a high den-sity of silver grains overlying both the caudate-putamenand the globus pallidus (Fig. 8b,d). There was a slightlydecreasing gradient of intensity progressing from the dor-

Fig. 7. Higher-power photomicrographs of Thy-1 immunolabeling(a, brightfield) and probe hybridization in nearby emulsion-dippedsection (b, darkfield) through a posterior level of a postnatal day (P)4thalamus. The relative intensity of immunolabeling in the dorsal andventral lateral geniculate nuclei matches that of the hybridizationsignals, whereas in contrast, relatively weak immunolabeling of ven-trobasal nuclear complex (VB) is evident (a) in comparison with arelatively intense hybridization signal in this nucleus (b). Scale bar 5100 mm.


Fig. 8. Patterns of Thy-1 immunoreactivity and probe hybridiza-tion change over the course of development in the caudate-putamen(CdP) and globus pallidus (gp). a–d: Intensities of both immunoreac-tivity and probe hybridization were high in the CdP and the gp duringthe first postnatal week. e,f: Intensity of immunoreactivity and probehybridization decreases in the CdP during the second postnatal week;in the gp, intensity of immunoreactivity appears greater, whereasthat of probe hybridization appears mostly similar in comparison with

earlier ages. However, distinct silver-grain clusters appeared moreprominent in the gp during this period. g, j: By the third postnatalweek (g), immunolabeling in the gp reaches its greatest intensitybefore declining by adulthood (i). Labeled neurons in the CdP areindicated by the arrows in i. The intensity of probe hybridization inthe CdP by these stages was low and silver grains became diffuselyscattered (h,j); in the gp, distinct grain clusters became evident byadulthood (j). P, postnatal day; Ad, adult. Scale bar 5 150 mm.

sal to the ventral portion of the caudate-putamen. Grainclusters appeared to be more condensed and larger in theglobus pallidus than in the caudate-putamen (Fig. 8b,d).

P10–P18. By P10, the slightly patchy distribution ofThy-1 immunoreactivity in the caudate-putamen as wellas the sharply delineated lateral edge were no longerapparent, and the overall intensity of immunostaininghad also decreased, yielding a homogeneous distributionof light immunoreactivity (Fig. 8e,g). The globus palliduswas also labeled more uniformly at this age (Fig. 8e,g); byP18, the intensity of immunoreactivity reached its great-est level (Fig. 8g) before declining somewhat by adulthood.

An overall decrease in the density of silver grains wasevident by these stages in both the caudate-putamen andthe globus pallidus (Fig. 8f,h). Individual grain clustersremained more readily identifiable by their larger sizewithin the globus pallidus, in comparison with the morescattered grains in the caudate nucleus (Fig. 8f,h).

Spinal cord

Adult. Strong immunoreactivity was present through-out the grey matter neuropil at all levels (Fig. 9). In theventral horn, prominent labeling of the plasmalemma ofthe ventral motor neurons was evident (Fig. 9b). Immu-nopositive fiber bundles could be followed from the ventralhorn through the spinal white matter to the exit point ofthe ventral roots (Fig. 9b). The ventral, lateral, and dorsalfuniculi also displayed a fine meshwork of Thy-1-immunoreactive fibers. High-power confocal microscopeimages through the dorsal funiculus showed prominentlabeling of cross-sectioned fibers in the corticospinal tract(Fig. 9c) as well as the tracts of the dorsal columns (Fig. 9d).

In situ hybridization histochemistry yielded a fine scat-tering of small grains over the marginal zone and substan-tia gelatinosa, whereas large grain clusters were evidentoverlying neurons of the nucleus proprius, intermediategrey, and the ventral motor neurons (Fig. 9e,f). No probehybridization was detected in the spinal white matter.


P0–P3. Thy-1 immunoreactivity was present in thespinal grey matter at all cord levels at P0, but at any givenlevel, displayed an increasing gradient of intensity fromdorsal to ventral (Fig. 10a,b). The neurons of the dorsalhorn, and the neuropil of the marginal zone and substan-tia gelatinosa, appeared mostly unlabeled at both cervicaland lumbar levels (Fig. 10a). Moderately intense Thy-1immunoreactivity delineated the dorsal funiculus

Fig. 9. Thy-1 immunolabeling and probe hybridization in adultmouse spinal cord. a: Prominent labeling of the neuropil was evidentin the substantia gelatinosa (sg) and nucleus proprius (np). The holesrepresent unlabeled cell bodies, although in some cases, labeling ofthe somatic plasmalemma can be seen (arrowhead). b: Large mo-toneurons of the ventral horn displayed prominent labeling of theirsomatic plasmalemma (arrowheads) and proximal dendrites. Fasci-cles of labeled fibers (presumably the axons of ventral motoneurons)exit the ventral grey matter (arrow) and traverse the ventral funicu-lus. c,d: High-power, confocal images of Thy-1-immunolabeled corti-cospinal tract fibers (c) or fibers of the fasciculus gracilis (d). e: Largegrain clusters (arrowhead) overlying the motoneurons of the ventralhorn (cervical level shown). f: Brightfield emulsion autoradiograph ofsection through nucleus proprius showing clusters of silver grainsrepresenting Thy-1 probe hybridization (arrow) overlying pale nucleiof neurons, but not small and dark neuroglial cell nuclei. Scale bars 550 mm in a,b: 10 mm in c,d; 200 mm in e; 25 mm in f.


throughout the extent of the cord, where the labelingappeared as fine-caliber fibers. Such labeling presumablyrepresents the ascending somatosensory fibers of the fas-ciculi gracilis and cuneatus, which in rats are largelyestablished by birth (Asanuma et al., 1988; Stanfield,1992). In contrast, descending corticospinal fibers haveonly reached the first cervical segment by birth in rats(Jones et al., 1982) or the pyramidal decussation in mice(Barlow and Huntley, unpublished observations).

P4–P8. From P4 to P8, there was an overall increasein the intensity of immunolabeling of the neuropil of bothdorsal and ventral horns (Fig. 10c,d), particularly in themarginal zone and substantia gelatinosa. Labeling of fi-bers in the dorsal funiculus was also more prominent bythese stages.

P10–P14. During this period, the major changes incomparison with younger ages include a further overallincrease in the intensity of neuropil labeling in the spinalgrey, becoming adult-like by P12 (Fig. 10e,f ). By this age,some neurons in the dorsal and ventral horns exhibitedplasmalemmal immunolabeling, and their somata andproximal dendrites were delineated by the intense neuro-pil labeling (Fig. 10e,f).

DISCUSSION

This study details the cellular localization and regionaldistribution patterns of Thy-1 protein and gene expression

in sensory-motor and related regions of adult mouse brainand spinal cord and comprehensively examines the post-natal developmental time course through which suchadult patterns of expression emerge. The results reveal apreviously unrecognized spatially and temporally stag-gered sequence through which the adult pattern of wide-spread Thy-1 localization and expression is achieved.

Thy-1 expression by neuroglia has been observed invitro (Brown et al., 1984; Pruss, 1989), whereas, in con-trast, previous studies of Thy-1 immunolocalization intissue sections through rodent brain, spinal cord, andperipheral nerves suggest that Thy-1 expression is strictlyneuron-specific (Barclay and Hyden, 1978; Morris et al.,1983). Our results are consistent with these latter obser-vations, because: (1) no probe hybridization was detectedin white matter tracts; (2) silver grain accumulationswithin gray matter were clearly overlying Nissl-stainedsomata with the characteristics of neurons (Vaughan,1984); and (3) Thy-1 immunolocalization was restricted toneurons in dissociated hippocampal cell cultures contain-ing both cell types (e.g., Fig. 2j). Neuron-specific localiza-tion of Thy-1 in brain and spinal cord is thus similar tothat of other GPI-linked, immunoglobulin superfamilymembers such as LAMP (Horton and Levitt, 1988), neu-rotrimin (Struyk et al., 1995), and F3/F11 (Gennarini etal., 1989; Faivre-Sarrailh et al., 1992).

The use of vibratome sections without detergent perme-abilization yielded the most detailed labeling patterns in

Fig. 10. Thy-1 immunoreactivity in the developing spinal corddorsal (top row) and ventral (bottom row) horns. In the dorsal hornmarginal zone (mz) and substantia gelatinosa (sg), immunolabeling isweak at postnatal day (P)2 (a), increases in intensity by P4 (c), andbecomes prominent by P12 (e). In contrast, the neuropil of the ventral

horn at P2 is already relatively intensely labeled (b); by P4, immu-nolabeling of the neuropil has increased (d), and is adult-like by P12(f ), where somata and proximal dendrites (arrows) of large motoneu-rons are delineated by neuropilar labeling. Scale bar 5 50 mm.


comparison with the other methods tried, particularly inthe neuropil, where dense puncta, fine-caliber processes,and prominent plasmalemmal staining were evident. Inaddition, cross-sectioned axons were visualized in white-matter tracts such as the corticospinal and spinal dorsalcolumn tracts (e.g., Fig. 9c,d), indicating that myelinationdid not completely impede antibody binding at the cutsurface of the section. Although it is possible that alter-native tissue processing methods (Morris and Barber,1983) are more sensitive for revealing Thy-1 labeling ofaxons within heavily myelinated fiber tracts, the generalfeatures of localization observed in the present study areconsistent with those which would be expected for a cell-surface antigen, and are similar to previous studies show-ing that Thy-1 is distributed along the plasmalemmalsurfaces of neuronal somata, dendrites, and most axons(Stohl and Gonatas, 1977; Barclay and Hyden, 1978; Dottiet al., 1991). Thy-1 has also been isolated from synapto-somal preparations (Acton et al., 1978), suggesting local-ization to perisynaptic membranes, which may reflect inpart the very dense and fine neuropilar immunolabeling.

The distribution and developmental patterns of Thy-1expression form a unique profile in comparison with otherGPI-anchored neural CAMs of the immunoglobulin super-family. The distribution of F3/F11 and neurotrimin is alsoquite widespread, and therefore similar to Thy-1, butthese are expressed most highly during early development(Hosoya et al., 1995; Struyk et al., 1995), whereas North-ern blot analysis has shown that overall levels of Thy-1gene expression in brain is low at birth and rises steadilyto peak at adulthood (Morris, 1985). The high level ofThy-1 expression in adulthood is similar to that observedfor BIG-1, BIG-2, Kilon, and CEPU-1, but these othermolecules exhibit a highly restricted regional distribution(Yoshihara et al., 1994,1995; Hosoya et al., 1995; Struyket al., 1995; Spaltmann and Brummendorf, 1996; Funatsuet al., 1999). Similarly, LAMP and OBCAM, markers forneurons comprising the limbic and olfactory systems, re-spectively (Levitt, 1984; Horton and Levitt, 1988;Zhukareva and Levitt, 1995; Hachisuka et al., 1996; Pi-menta et al., 1996), play functional roles in the establish-ment of connectivity within these specific neural systems(Keller et al., 1989; Pimenta et al., 1995; Mann et al.,1998). The widespread localization of Thy-1 across tha-lamic relay nuclei related to somatosensory, visual, motor,and limbic systems in adults suggests that Thy-1 is prob-ably not, therefore, involved in providing system-specificguidance or maturation cues.

It has been suggested that in some systems the onset ofThy-1 expression occurs at a stage when migration iscomplete and dendritic growth has begun (Xue et al.,1991), but the results of this study suggest a more complexrelationship between neuronal maturation and the onsetof Thy-1 expression. The onset of Thy-1 expression insensory-motor cortex did not follow the “inside-out” pat-tern of neurogenesis and migration in which cells indeeper layers are born prior to ones in more superficiallayers (Caviness, 1982), and elaborate their dendritic pro-cesses earlier (Miller, 1988). Similarly, there was a mark-edly staggered onset of Thy-1 expression among thalamicnuclei of the anterior group, in which genesis, migration,and morphological development of the neurons comprisingindividual nuclei occur contemporaneously (Altman andBayer, 1988). The signals which govern such differentialThy-1 expression are unknown, but the as yet unidentified

ligand for Thy-1 (Tiveron et al., 1992; Dreyer et al., 1995)is a likely candidate. Speculatively, the differential onsetof Thy-1 expression across layers and nuclei may reflectstaggered rates of structural or functional maturationwhich proceed independently of the timing of neurogen-esis, migration, and dendritic growth. For example, theintense probe hybridization and immunoreactivity in cor-tical layer V in comparison with the protracted emergenceof Thy-1 expression in other layers could reflect the pre-cocious role for layer V neurons in establishing some of theearliest output pathways from the neocortex (Clasca et al.,1995).

The intensity of Thy-1 gene and protein expression in-creased gradually with maturation in all structures exam-ined with the exception of the caudate-putamen, whichdisplayed an opposite pattern in which mRNA expressionand intensity of immunoreactivity was greatest during thefirst postnatal week of life before declining during subse-quent development. It is unlikely that the gradual loss ofimmunolabeling represents a gradual masking of theepitope by astrocytic coverings of neuronal membranes,because, contemporaneously, increasingly robust immu-nolabeling was observed in all other grey matter struc-tures examined, where gradual astrocytic covering wouldbe expected as well. Although Thy-1 protein has a verylong surface half-life (Morris, 1985), it is likely, thoughspeculative, that the apparent down-regulation of proteinresults from the contemporaneous down-regulation ofgene expression which was also observed. Down-regulation of Thy-1 by Purkinje cells in developing mousecerebellum has been correlated with the genesis of synap-tic contacts furnished by ingrowing climbing fibers (Bolinand Rouse, 1986). Although the functional significance ofThy-1 down-regulation in the caudate-putamen is unclearat present, it has been suggested that Thy-1 may play arole in the initial establishment of the nigrostriatal pro-jection (Shults and Kimber, 1993), which commences pre-natally (Specht et al., 1981). Decreasing Thy-1 levelsshortly after birth may therefore reflect stabilization offinal nigrostriatal terminal fiber or synapse distributionwhich occurs during this early postnatal period.

A delay between Thy-1 gene and protein expression hasbeen described for Purkinje cells in mouse cerebellum(Xue and Morris, 1992), where it has been proposed Thy-1is excluded from growing axons and their distal ends untilgrowth and terminal arborization are complete (Xue et al.,1991). In the present study, a similar delay was observedfor the thalamic VB nucleus and dLG, where gene expres-sion was high at birth, but immunolabeling of these nucleiand in cortical layer IV, the target of their axon termina-tions, was not significant until ;P3. This may indicatethat, similarly, Thy-1 is also excluded from the terminalends of ingrowing thalamocortical axons until after theperiod in which such axons colonize layer IV and, in thecase of S1, formation of barrel cytoarchitecture is complete(Woolsey and Van Der Loos, 1970; Lund and Mustari,1977; Rice and Van der Loos, 1977; Crandall and Cavi-ness, 1984; Senft and Woolsey, 1991; Agmon et al., 1993;Kageyama and Robertson, 1993). Although extensive an-atomical remodeling of both thalamocortical terminalaxon arbors and the dendrites of layer IV neurons occursduring this period (Greenough and Chang, 1988; Senftand Woolsey, 1991; Agmon et al., 1993; Catalano et al.,1996), studies of Thy-1 knockout mice show that Thy-1does not play an essential role either in the normal devel-


opment of such patterned thalamocortical connections orin regulating their capacity for anatomical plasticity (Bar-low, Bozdagi, Kelley, and Huntley, unpublished observa-tions). On the other hand, functional plasticity of thalamo-cortical synaptic responses (the ability to generate long-term potentiation or depression) diminishes by P8–P10(Crair and Malenka, 1995; Feldman et al., 1998), coinci-dent with our observations of a significant increase in theintensity of Thy-1 immunolabeling in the neuropil of bar-rel centers. Because Thy-1 has been implicated in regu-lating long-term potentiation and properties of inhibitorypostsynaptic currents in the mature hippocampus(Nosten-Bertrand et al., 1996; Hollrigel et al., 1998), itmay play a role in regulating properties of synaptic func-tion in developing barrel cortex. In the immune system,Thy-1 modulates signal transduction in T-cells by regulat-ing the packing density and interactions between severalsignal transduction molecules which normally operatewithin the membrane as part of a complex (Hueber et al.,1997). Given that Thy-1 may constitute as much as2.5–7% of the total suface protein (Beech et al., 1983), aparticularly appealing hypothesis is that Thy-1 regulateschannel properties or other membrane components re-lated to synaptic signaling by physically modulating thedistribution of other cell-surface signaling proteins.

ACKNOWLEDGMENTS

This study was additionally supported by the EasternParalyzed Veterans Association and an Irma T. HirschlCareer Scientist Award. We thank Dr. Deanna L. Bensonfor help with hippocampal cell cultures, and are gratefulto Drs. Benson, Kevin A. Kelley and Victor Friedrich forhelpful comments and discussion.

LITERATURE CITEDActon RT, Addis J, Carl GF, McClain LD, Bridgers WF. 1978. Association

of Thy-1 differentiation alloantigen with synaptic complexes isolatedfrom mouse brain. Proc Natl Acad Sci USA 75:3283–3287.

Agmon A, Yang LT, O’Dowd DK, Jones EG. 1993. Organized growth ofthalamocortical axons from the deep tier of terminations into layer IVof developing mouse barrel cortex. J Neurosci 13:5365–5382.

Altman J, Bayer SA. 1988. Development of the rat thalamus: II. Time andsite of origin and settling pattern of neurons derived from the anteriorlobule of the thalamic neuroepithelium. J Comp Neurol 275:378–405.

Asanuma C, Ohkawa R, Stanfield BB, Cowan WM. 1988. Observations onthe development of certain ascending inputs to the thalamus in rats. I.Postnatal development. Brain Res Dev Brain Res 41:159–170.

Barclay AN, Hyden H. 1978. Localization of Thy-1 antigen in rat brain andspinal cord by immunofluorescence. J Neurochem 31:1375–1391.

Beech JN, Morris RJ, Raisman G. 1983. Density of Thy-1 on axonalmembrane of different rat nerves. J Neurochem 41:411–417.

Benson DL, Tanaka H. 1998. N-cadherin redistribution during synapto-genesis in hippocampal neurons. J Neurosci 18: 6892–6904.

Bolin LM, Rouse RV. 1986. Localization of Thy-1 expression during post-natal development of the mouse cerebellar cortex. J Neurocytol 15:29–36.

Bottenstein JE, Sato GH. 1979. Growth of a rat neuroblastoma cell line inserum-free supplemented medium. Proc Natl Acad Sci USA 76:514–517.

Brown RHJ, Schweitzer J, Dichter MA. 1984. Expression of the Thy-1antigen in long-term cultures of embryonic mouse spinal cord. BrainRes 296:87–91.

Catalano SM, Robertson RT, Killackey HP. 1996. Individual axon morphol-ogy and thalamocortical topography in developing rat somatosensorycortex. J Comp Neurol 367:36–53.

Caviness VS Jr. 1982. Neocortical histogenesis in normal and reeler mice:

a developmental study based upon [3H]thymidine autoradiography.Brain Res 256:293–302.

Clasca F, Angelucci A, Sur M. 1995. Layer-specific programs of develop-ment in neocortical projection neurons. Proc Natl Acad Sci USA 92:11145–11149.

Crair MC, Malenka RC. 1995. A critical period for long-term potentiationat thalamocortical synapses. Nature 375:325–328.

Crandall JE, Caviness VS. 1984. Thalamocortical connections in newbornmice. J Comp Neurol 228:542–556.

Dotti CG, Parton RG, Simons K. 1991. Polarized sorting of glypiatedproteins in hippocampal neurons. Nature 349:158–161.

Dreyer EB, Leifer D, Heng JE, McConnell JE, Gorla M, Levin LA, Bar-nestable CJ, Lipton SA. 1995. An astrocytic binding site for neuronalThy-1 and its effects on neurite outgrowth. Proc Natl Acad Sci USA92:11195–11199.

Faivre-Sarrailh C, Gennarini G, Goridis C, Rougon G. 1992. F3/F11 cellsurface molecule expression in the developing mouse cerebellum ispolarized at synaptic sites and within granule cells. J Neurosci 12:257–267.

Feldman DE, Nicoll RA, Malenka RC, Isaac JTR. 1998. Long-term depres-sion at thalamocortical synapses in developing rat somatosensory cor-tex. Neuron 21:347–357.

Funatsu N, Miyata S, Kumanogho H, shigeta M, Hamada K, Endo Y,Sokawa Y, Maekawa S. 1999. Characterization of a novel rat brainglycosylphosphatidylinositol-anchored protein (Kilon), a member of theIgLON cell adhesion molecule family. J Biol Chem 274:8224–8230.

Furley AJ, Morton SB, Manalo D, Karagogeos D, Dodd J, Jessell TM. 1990The axonal glycoprotein TAG-1 is an immunoglobulin superfamilymember with neurite outgrowth-promoting activity. Cell 61:157–170.

Gall CM, Isackson PJ. 1989. Limbic seizures increase neuronal productionof messenger RNA for nerve-growth factor. Science 245:758–761.

Gennarini G, Rougon G, Vitiello F, Corsi P, Di Benedetta C, Goridis C.1989. Identification and cDNA cloning of a new member of the L2/HNK-1 family of neural surface glycoproteins. J Neurosci Res 22:1–12.

Greenough WT, Chang FL. 1988. Dendritic pattern formation involvesboth oriented regression and oriented growth in the barrels of mousesomatosensory cortex. Brain Res 471:148–52.

Hachisuka A, Yamazaki T, Sawada J, Terao T. 1996. Characterization andtissue distribution of opioid-binding cell adhesion molecule (OBCAM)using monoclonal antibodies. Neurochem Int 28:373–379.

Hollrigel GS, Morris RJ, Soltesz I. 1998. Enhanced bursts of IPSCs indentate granule cells in mice with regionally inhibited long-term po-tentiation. Proc R Soc Lond B 265:63–69.

Horton HL, Levitt P. 1988. A unique membrane protein is expressed onearly developing limbic system axons and cortical targets. J Neurosci8:4653–4661.

Hortsch M. 1996. The L1 family of neural cell adhesion molecules: oldproteins performing new tricks. Neuron 17:587–593.

Hosoya H, Shimazaki K, Kobayashi S, Takahashi H, Shirasawa T, Tak-enawa T, Watanabe K. 1995. Developmental expression of the neuraladhesion molecule F3 in the rat brain. Neurosci Lett 186:83–86.

Hueber AO, Bernard AM, Battari CL, Marguet D, Massol P, Poa C, BrunN, Garcia S, Stewart C, Pierres M, He HT. 1997. Thymocytes in Thy-1-/- mice show augmented TCR signaling and impaired differentiation.Curr Biol 7:705–708.

Huntley GW, Benson DL, Jones EG, Isackson PJ. 1992. Developmentalexpression of brain derived neurotrophic factor mRNA by neurons offetal and adult monkey prefrontal cortex. Brain Res Dev Brain Res70:53–63.

Jones EG, Schreyer DJ, Wise SP. 1982. Growth and maturation of the ratcorticospinal tract. Prog Brain Res 57:361–379.

Kageyama GH, Robertson RT. 1993. Development of geniculocortical pro-jections to visual cortex in rat: evidence for early ingrowth and synap-togenesis. J Comp Neurol 335:123–148.

Keller F, Rimvall K, Barbe MF, Levitt P. 1989. A membrane glycoproteinassociated with the limbic system mediates the formation of the septo-hippocampal pathway in vitro. Neuron 3:551–561.

Lander A. 1989. Understanding the molecules of neural cell contacts:emerging patterns of structure and function. Trends Neurosci 12:189–195.

Levitt P. 1984. A monoclonal antibody to limbic system neurons. Science223:299–301.

Lund RD, Mustari MJ. 1977. Development of the geniculocortical pathwayin rats. J Comp Neurol 173:289–306.


Mahanthappa NK, Patterson PH. 1992a. Thy-1 involvement in neuriteoutgrowth: perturbation by antibodies, phospholipase C and mutation.Dev Biol 150:47–59.

Mahanthappa NK, Patterson PH. 1992b. Thy-1 multimerization is corre-lated with neurite outgrowth. Dev Biol 150:60–71.

Mann F, Zhukareva V, Pimenta A, Levitt P, Bolz J. 1998. Membrane-associated molecules guide limbic and nonlimbic thalamocortical pro-jections. J Neurosci 18:9409–9419.

Miller MW. 1988. Development of projection and local circuit neurons inneocortex. In: Peters A, Jones EG, editors. Cerebral cortex, vol 7.Plenum Press: New York. p 133–175.

Morris RJ. 1985. Thy-1 in developing nervous tissue. Dev Neurosci 7:133–160.

Morris RJ, Barber PC. 1983. Fixation of Thy-1 in nervous tissue forimmunohistochemistry. J Histochem Cytochem 31:263–274.

Morris RJ, Barber PC, Beech JN, Raisman G. 1983. The distribution ofThy-1 antigen in the P.N.S. of the adult adult. J Neurocytol 12:1017–1039.

Morris RJ, Beech JN, Barber PC, Raisman G. 1985. Late emergence ofThy-1 on climbing fibres demonstrates a gradient of maturation fromthe fissues to the folial convexities in developing rat cerebellum. J Neu-rocytol 14:453–467.

Nosten-Bertrand M, Errington ML, Murphy KPSJ, Tokugawa Y, BarboniE, Kozlova E, Michalovich D, Morris RGM, Silver J, Stewart CL, BlissTVP, Morris RJ 1996. Normal spatial learning despite regional inhibi-tion of LTP in mice lacking Thy-1. Nature 379:826–829.

Pimenta A, Zhukareva V, Barbe MF, Reinoso B, Grimley C, Henzel W,Fischer I, Levitt P. 1995. The limbic system-associated membraneprotein is an Ig superfamily member that mediates selective neuronalgrowth and targeting. Neuron 15:287–297.

Pimenta AF, Reinoso BS, Levitt P. 1996. Expression of the mRNAs encod-ing the limbic system-associated membrane protein (LAMP). 2. Fetalrat brain. J Comp Neurol 375:289–302.

Pruss RM. 1979. Thy-1 antigen on astrocytes in long-term cultures of ratcentral nervous system. Nature 280:688–690.

Reif AE, Allen JMV. 1964. The AKR thymic antigen and its distribution inleukemias and nervous tissues. J Exp Med 120:413–433.

Rice FL, Van der Loos H. 1977. Development of the barrels and barrel field inthe somatosensory cortex of the mouse. J Comp Neurol l 171:545–560.

Senft SL, Woolsey TA. 1991. Growth of thalamic afferents into mousebarrel cortex. Cereb Cortex 1:308–335.

Shults CW, Kimber TA. 1993. Thy-1 immunoreactivity distinguishespatches/striosomes from matrix in the early postnatal striatum of therat. Brain Res Dev Brain Res 75:136–140.

Spaltmann F, Brummendorf T. 1996. CEPU-1, a novel immunoglobulinsuperfamily molecule is expressed by developing Purkinje cells. J Neu-rosci 16:1770–1779.

Specht LA, Picket VM, Joh TH, Reis DJ. 1981. Light-microscopic immu-nocytochemical localization of tyrosine hydroxylase in prenatal ratbrain. I. Early ontogeny. J Comp Neurol 199:233–253.

Stanfield BB. 1992. The development of the corticospinal projection. ProgNeurobiol 38:169–202.

Stohl W, Gonatas NK. 1977. Distribution of the Thy-1 antigen in cellularand subcellular fractions of adult mouse brain. J Immunol 119:422–427.

Struyk AF, Canoll PD, Wolfgang MJ, Rosen CL, Deustachio P, Salzer JL.1995. Cloning of neurotrimmin defines a new subfamily of differen-tially expressed neural cell adhesion molecules. J Neurosci 15:2141–2156.

Tessier-Lavigne M, Goodman CS. 1996. The molecular biology of axonguidance. Science 274:1123–1133.

Tiveron MC, Barboni E, Pliego Rivero FB, Gormley AM, Seeley PJ, Gros-veld F, Morris RJ. 1992. Selective inhibition of neurite outgrowth onmature astrocytes by Thy-1 glycoprotein. Nature 355: 745–748.

Tuttle R, Schlaggar BL, Braisted JE, O’Leary DDM. 1995. Maturation-dependent upregulation of growth-promoting molecules in developingcortical plate controls thalamic and cortical neurite growth. J Neurosci15:3039–3052.

Vaughan DW. 1984. The structure of neuroglial cells. In: Jones EG, PetersA, editors. Cerebral cortex vol 2. New York: Plenum Press. p 285–329.

Walsh FS, Doherty P. 1997. Neural cell adhesion molecules of the immu-noglobulin superfamily: role in axon growth and guidance. Annu RevCell Dev Biol 13:425–456.

Wise SP, Donoghue JP. 1986. Motor cortex of rodents. In: Jones EG, PetersA, editors. Cerebral cortex, vol 5. New York: Plenum Press. p 243–270.

Withers GS, Banker G. 1998. Characterizing and studying neuronal cul-tures. In: Banker G, Goslin K, editors. Culturing nerve cells, vol 2.Cambridge, MA: MIT Press. p 113–151.

Woolsey TA,Van der Loos, H. 1970. The structural organization of layer IVin the somatosensory region (S1) of mouse cerebral cortex. Brain Res17:205–242.

Xue GP, Morris RJ. 1992. Expression of the neuronal surface glycoproteinThy-1 does not follow appearance of its mRNA in developing mousePurkinje cells. J Neurochem 58:430–440.

Xue GP, Calvert RA, Morris RJ. 1990. Expression of the neuronal cellsurface glycoprotein Thy–1 is under post-transcriptional control and isspatially regulated in the developing olfactory system. Development109:851–864.

Xue GP, Pliego Rivero B, Morris RJ. 1991. The surface glycoprotein Thy-1is excluded from growing axons during development: a study of theexpression of Thy-1 during axogenesis in hippocampus and hindbrain.Development 112:161–176.

Yoshihara Y, Kawasaki M, Tani M, Tamada A, Nagata S, Kagamiyama H,Mori K. 1994. Big-1: a new Tag-1/F3-related member of the immuno-globulin superfamily with neurite outgrowth promoting activity. Neu-ron 13:415–426.

Yoshihara Y, Kawasaki M, Tamada A, Nagata S, Kagamiyama H, Mori K.1995. Overlapping and differential expression of BIG-2, BIG-1, TAG-1and F3: four members of an axonal associated cell adhesion moleculesubgroup of the immunoglobulin superfamily. J Neurobiol 28:51–69.

Zhukareva V, Levitt P. 1995. The limbic system-associated membraneprotein (LAMP) selectively mediates interactions with specific centralneuronal populations. Development 121:1161–1172.


developmentally regulated expression of thy-1 in structures of the mouse sensory-motor system

Documents