Developmental Brain Research 140 (2003) 309–314www.elsevier.com/ locate/devbrainres

Short communication

P ostnatal development of the glutamate vesicular transporter VGLUT1in rat cerebral cortex

a b a a ,*Andrea Minelli , Robert H. Edwards , Tullio Manzoni , Fiorenzo Contia `Istituto di Fisiologia Umana, Universita di Ancona,Via Tronto 10/A, Torrette di Ancona, I-60020 Ancona, Italy

bDepartments of Neurology and Physiology, University of California at San Francisco, San Francisco, CA 94143-0435,USA

Accepted 25 November 2002

Abstract

The expression of the vesicular glutamate transporter VGLUT1 in the rat neocortex was studied during postnatal development usingimmunocytochemistry and Western blotting. At all ages, VGLUT immunoreactivity is localized to puncta that coexpress the presynapticmarker synaptophysin. VGLUT1 immunoreactivity is faint at birth, increases in the subplate during the first postnatal week, invades thesupragranular layers in the second week and reaches the adult pattern at P20–P30. Its spatial and temporal maturation patterns suggestthat VGLUT1 may be the vesicular transporter in developing corticocortical connections. 2002 Elsevier Science B.V. All rights reserved.

Theme: Development and regeneration

Topic: Cerebral cortex and limbic system

Keywords: Glutamate; Glutamate transporter; Vesicle; Neocortex; Callosal connection; Associative connection; Development

An essential step in synaptic transmission is neuro- most robustly expressed in the cerebral cortex, a regiontransmitter loading in synaptic vesicles by the action of where the role of Glu is prominent in both physiologicalspecific integral membrane proteins, named vesicular and pathological conditions. VGLUT1 mRNA is expressedtransporters. During the last few years, two distinct vesicu- in all cortical layers (whereas VGLUT2 and VGLUT3lar glutamate (Glu) transporters sharing similar pharmaco- mRNAs are virtually absent) and VGLUT1 immuno-logical characteristics have been identified, VGLUT1 (for- reactivity (ir), although more intense in layers I–III, is

1merly named brain-specific Na -dependent inorganic found throughout the depth of the neocortex (whilephosphate transporter, BNPI) [3,22,25] and VGLUT2 VGLUT2-ir is restricted to layers IV and VI and VGLUT3-

1(originally called differentiation associated Na -dependent ir is weak) [2,11,13,17]. In contrast to the large body ofinorganic phosphate transporter, DNPI) [1,27]. These data on VGLUT1 localization in the adult neocortex,transporters are localized to axon terminals forming nothing is known of the developmental expression ofasymmetric contacts, where they are primarily associated VGLUT1. We therefore used immunocytochemistry andto synaptic vesicles and are distributed in a complementary Western blotting to investigate qualitatively and quantita-fashion in the adult nervous system, with VGLUT1 being tively the postnatal development of VGLUT1 in the ratmainly expressed in telencephalon and VGLUT2 in dien- neocortex.cephalon and lower brainstem [2,11,17]. Recently, a third Sprague–Dawley albino rats (Charles River, Milan,vesicular Glu transporter, VGLUT3, has been identified Italy) were studied at the following time points: P0and detected in both asymmetric and symmetric axon (postnatal day 0, within 24 h of birth), P2, P5, P10, P15,terminals [12,13,24]. Of the VGLUTs, VGLUT1 is the P20, P30, and P60 (adult stage). Their care and handling

were approved by the Ethical Committee for AnimalResearch of the University of Ancona.*Corresponding author. Tel.:139-71-220-6056; fax:139-71-220-

For immunocytochemistry, rats (five per time point)6052.E-mail address: f.conti@popcsi.unian.it(F. Conti). were perfused transcardially with 4% paraformaldehyde in

0165-3806/02/$ – see front matter 2002 Elsevier Science B.V. All rights reserved.doi:10.1016/S0165-3806(02)00617-X

310 A. Minelli et al. / Developmental Brain Research 140 (2003) 309–314

0.1 M phosphate buffer (pH 7.4).Vibratome-cut coronal or period exhibits a superficial marginal zone, a denselyparasagittal sections (30–50-mm thick) were pretreated cellular cortical plate (CP), and a thick, less denselywith 1% H O , incubated in 10% non-immune newborn packed subplate region in which layers V and VI are2 2

calf serum (NBCS) with 0.2% Triton-X and then incubated clearly distinguishable [29].overnight at 48C with a rabbit polyclonal antibody For double-labeling studies, sections (from animals(1:1000) raised against a glutathioneS-transferase fusion sacrificed at P5 and P10) were incubated overnight in aprotein containing the last 68 amino acids (residues 493– mixture containing VGLUT1 (1:1000) and a mouse mono-560) of rat BNPI [22] (for details on antibody production clonal anti-synaptophysin (clone SVP-38; S-5768, Sigma;and characterization, see Ref [2].). The next day, sections 1:1000) antibodies, and the next day in a mixture ofwere processed and reacted as previously [21]. All ob- FITC-conjugated anti-rabbit IgG and TRITC-conjugatedservations were made in the first somatic sensory cortex anti-mouse IgG (both diluted 1:100) antisera. Double-(SI) or in the presumptive SI, which in the early postnatal labeled sections were examined using a Bio-Rad Mi-

Fig. 1. Representative low-power photomicrographs showing the distribution of VGLUT1 ir in rat SI during postnatal development.VGLUT1 at P0 (A), P2(B), P5 (C), P10 (D), P15 (E), P20 (F), P30 (G), and P60 (H). Roman numerals indicate cortical layers; CP, cortical plate. Bar: 200mm for (A)–(H).

A. Minelli et al. / Developmental Brain Research 140 (2003) 309–314 311

croradiance confocal laser scanning microscope equipped Nitrocellulose filter blots were incubated overnight withwith an argon (488 nm) and a helium-neon (543 nm) laser. VGLUT1 antibody (1:1000), then processed and reacted asRed and green immunofluorescence was imaged sequen- previously [21]. Labeled bands were visualized with atially through the relative channels using a360 oil BioRad Chemi Doc and densitometric analyses wereimmersion lens (numerical aperture 1.4) using a confocal performed using the Quantity One software from BioRad.aperture of 1 mm, and emissions were separated with At P0–P2, VGLUT1-ir was weak throughout the cortex,515/30 nm band-pass (for FITC) and 570 nm long-pass with layer VIb exhibiting the highest level of signal (Fig.(for TRITC) filters. Simultaneous visualization of the two 1A,B); here, small immunoreactive puncta were foundimmunoreactivities was achieved by merging paired green finely dispersed in the neuropil (Fig. 2E). Few large punctaand red images with a Bio-Rad LaserSharp processing were found scattered in layer I (Fig. 2A). The white mattersoftware (version 3.0). Control sections incubated with two underlying SI was virtually devoid of VGLUT1-ir (Fig.primary and one secondary antibody or with one primary 1A,B). At P5, VGLUT1-ir was increased in deep layers,and two secondary antisera revealed no appreciable cross- whereas the superficial layers and CP were moderatelyreactivity. stained (Fig. 1C): in layers V and VI, VGLUT1-positive

For immunoblotting, animal perfusion, tissue homogeni- puncta were dense and more heavily stained than at P0–P2zation, separation and resuspension of crude membrane and they were occasionally found in apposition withfraction were done as described previously [21]. Given the unlabeled cell bodies (Fig. 2C,F). At P10–P15, VGLUT1-small size of the cerebral cortex of early postnatal animals ir was homogeneously distributed through the cortex andand the consequent paucity of the protein extract we could was increased in layer V (Fig. 1D,E). Positive puncta wereharvest from it, for each time point studied we mixed the still mainly dispersed in the neuropil (Fig. 2B), but theycortical homogenates obtained from five rats. Aliquots were now more commonly observed around unstainedcontaining 30mg of protein extract were mixed with equal cellular profiles (Fig. 2G). At P20, VGLUT1-ir wasvolumes of 23 electrophoresis sample buffer plus 4 M increased throughout the cortex (notably in layers I, II–IIIurea (final concentration) and subjected to 7.5% SDS– and V; Fig. 1F) and its laminar distribution resembled thatPAGE with a 3% stacking gel under reducing conditions. observed in adult rats (Fig. 1H). At P30, the adult pattern

Fig. 2. Morphological details of maturation of VGLUT1 ir in the developing cerebral cortex. Layer I at P0 (A) and P15 (B). Layer V at P5 (C) and P30(D). Layer VI at P2 (E), P5 (F), and P15 (G). Bar: 10mm for (A)–(G).

312 A. Minelli et al. / Developmental Brain Research 140 (2003) 309–314

was fully expressed [2,11,17]: VGLUT1-ir was prominentin layers I and II–III, followed by layers V and VI, and lowin layer IV (Fig. 1G). In all cortical layers, positive punctawere found both scattered in the neuropil and in associa-tion with unlabeled cell bodies of different size and shape(Fig. 2D).

To investigate the nature of VGLUT11 puncta observedat early phases of cortical maturation, sections at P5 andP10 were double-stained for VGLUT1 and synaptophysin,a synaptic vesicle glycoprotein associated to axon termi-nals whose maturation profile follows that of synap-togenesis [10]. In single-labeled sections, synaptophysinimmunofluorescence was associated to puncta of differentsize located mostly in the marginal zone and subplate at P5and more evenly distributed through the cortical mantle atP10, as previously described [10]. In double-labeledsections at both P5 and P10, virtually all VGLUT11

profiles contained intense synaptophysin-ir (Fig. 3), thusindicating that the vast majority of VGLUT11 punctawere axon terminals.

Developmental changes of VGLUT1-ir were also ob-served in other telencephalic regions, where at all ages irwas associated exclusively to puncta. At P0–2,VGLUT1-irwas robust in the olfactory bulb, rhinal cortex and hip-pocampus (strata oriens and radiatum), moderate in theventricular zone (where it was absent after P5) and indentate gyrus, and weak in the striatum and thalamus.VGLUT1-ir increased gradually during the first and secondpostnatal week and reached the mature pattern betweenP10 and P20.

At all postnatal ages, we detected a single immuno-reactive band of approximately 60 kDa (Fig. 4A), con-sistent with the molecular weight reported in adult rat brainextracts [2]. The intensity of bands was faint at P0 and P2,

Fig. 3. VGLUT11 puncta coexpress synaptophysin-ir during corticalit increased progressively until P20, and peaked at P30,maturation. Single plane confocal images from the cortical subplate at P5exceeding adult levels (Fig. 4A). The relative density of show that all VGLUT11 profiles (green) express synaptophysin-ir (red).VGLUT1-ir bands (expressed as percentage of adult levels)Bar: 5 mm.

was 6.6% at P0, 13.5% at P2, 24% at P5, 47% at P10, 75%at P15, 92% at P20, and 114% at P30 (Fig. 4B).

Our results show that early after birth VGLUT1-ir is ficial and deep cortical layers already at birth [4,20]; (ii)associated to puncta which coexpress the presynaptic basalo-cortical fibers, though maturing with temporal andmarker synaptophysin and are mostly located in the spatial analogies with VGLUT1 [6], arise from neuronssubplate, a region where the highest density of established that are reportedly cholinergic or GABAergic [14]; (iii)synapses [5] and AMPA and NMDA receptor-mediated dopaminergic projections reach almost exclusively thesynaptic activity [15] have been detected in the neonatal prefrontal cortex [18]; (iv) fibers from the intralaminarcortex. Taken together, these finding suggest that in the thalamic nuclei invade the superficial cortical layers imme-developing neocortex VGLUT1-ir could reflect the protein diately after birth [26]; and (v) thalamo-cortical projectionsinserted in vesicle membranes localized in relatively from ‘specific’ relay nuclei are glutamatergic [19] butmature axon terminals, and that VGLUT1 may contribute develop before VGLUT1 and with different spatial featuresto mediate the synaptic activity occurring in the earliest [29], and all thalamic nuclei are virtually devoid ofphases of postnatal development. VGLUT1 mRNA during development [23]. A conspicuous

The developing neocortex is reached by various cor- fraction of corticopetal projections is represented by intra-ticopetal projections that follow specific patterns of matu- and interhemispheric corticocortical connections, whichration. Projections arising from subcortical structures are are considered glutamatergic [8]. In the somatic sensoryunlikely to use VGLUT1 since: (i) serotoninergic and cortex, callosal fibers and terminals are scanty and con-noradrenergic projections from the brainstem invade super- fined to the deep part of layer VI until the third day of life;

A. Minelli et al. / Developmental Brain Research 140 (2003) 309–314 313

R eferences

[1] Y. Aihara, H. Mashima, H. Onda H, S. Hisano, H. Kasuya, T. Hori,S. Yamada, H. Tomura, Y. Yamada, L. Inoue, I. Kojima, J. Takeda,

1Molecular cloning of a novel brain-type Na -dependent inorganicphosphate cotransporter, J. Neurochem. 74 (2000) 2622–2625.

[2] E.E. Bellocchio, H.L. Hu, A. Pohorille, J. Chan, V.M. Pickel, R.H.Edwards, The localization of the brain-specific inorganic phosphatetransporter suggests a specific presynaptic role in glutamatergictransmission, J. Neurosci. 18 (1998) 8648–8659.

[3] E.E. Bellocchio, R.J. Reimer, R.T. Fremeau, R.H. Edwards, Uptakeof glutamate into synaptic vesicles by an inorganic phosphatetransporter, Science 289 (2000) 957–960.

[4] C.A. Bennett-Clarke, N.L. Chiaia, R.S. Crissman, R.W. Rhoades,The source of the transient serotoninergic input to the developingvisual and somatosensory cortices in rat, Neuroscience 43 (1991)163–183.

[5] M.E. Blue, J.C. Parnavelas, The formation and maturation ofsynapses in the visual cortex of the rat. II. Quantitative analysis, J.Neurocytol. 12 (1983) 697–712.

[6] C.A. Calarco, R.T. Robertson, Development of basal forebrainprojections to visual cortex: DiI studies in rat, J. Comp. Neurol. 354(1995) 608–626.

[7] S. Clarke, G.M. Innocenti, Organization of immature intrahemis-pheric connections, J. Comp. Neurol. 251 (1986) 1–22.

[8] F. Conti, T. Manzoni, The neurotransmitters and postsynapticactions of callosally projecting neurons, Behav. Brain Res. 64(1994) 37–53.

[9] T.A. Coogan, A. Burkhalter, Sequential development of connectionsbetween striate and extrastriate visual cortical areas in the rat, J.Comp. Neurol. 278 (1988) 242–252.

[10] S.H. Devoto, C.J. Barnstable, Expression of the growth cone specificepitope CDA1 and the synaptic vesicle protein SVP38 in thedeveloping mammalian cerebral cortex, J. Comp. Neurol. 290(1989) 154–168.

Fig. 4. Expression of VGLUT1 in developing neocortex. (A) Representa-[11] R.T. Fremeau, M.D. Troyer, I. Pahner, G.O. Nygaard, C.H. Tran,

tive blots show the temporal intensity variations of immunoreactive bandsR.J. Reimer, E.E. Bellocchio, D. Fortin, J. Storm-Mathisen, R.H.

migrating at approximately 60 kDa. Reference molecular weights, in kDa,Edwards, The expression of vesicular glutamate transporters defines

are reported on the left. (B) Densitometric analysis is reported in thetwo classes of excitratory synapse, Neuron 31 (2001) 247–260.

graph showing the relative immunoblot densities given as percentage of[12] R.T. Fremeau Jr., J. Burman, T. Qureshi, C.H. Tran, J. Proctor, J.

adult levels considered as 100%.Johnson, H. Zhang, D. Sulzer, D.R. Copenhagen, J. Storm-Mathisen,R.J. Reimer, F.A. Chaudhry, R.H. Edwards, The identification ofvesicular glutamate transporter 3 suggests novel modes of signalingthey are denser in infragranular layers at P5 and start toby glutamate, Proc. Natl. Acad. Sci. USA 99 (2002) 14488–14493.invade supragranular layers at the beginning of the second

[13] C. Gras, E. Herzog, G.C. Bellenchi, V. Bernard, P. Ravassard, M.postnatal week [28], and in the visual cortex associative Pohl, B. Gasnier, B. Giros, S. El Mestikawy, A third vesicularprojections appear to follow the same pattern [7,9]. More- glutamate transporter expressed by cholinergic and serotoninergicover, the rat neocortex expresses detectable levels of neurons, J. Neurosci. 22 (2002) 5442–5451.

[14] I. Gritti, L. Mainville, M. Mancia, B.E. Jones, GABAergic and otherVGLUT1 mRNA already at birth [23], and recent electro-noncholinergic basal forebrain neurons, together with cholinergicphysiological studies revealed that neonatal subplate neu-neurons, project to the mesocortex and isocortex in the rat, J. Comp.

rons receive AMPA and NMDA receptor-mediated synap- Neurol. 383 (1997) 163–177.tic inputs arising from neocortical sources [16]. These [15] I.L. Hanganu, W. Kilb, H.J. Luhman, Spontaneous synaptic activityfindings indicate that the developmental profile of of subplate neurons in neonatal rat somatosensory cortex, Cereb.

Cortex 11 (2001) 400–410.VGLUT1 parallels that of corticocortical connections and[16] I.L. Hanganu, W. Kilb, H.J. Luhman, Functional synaptic projectionssuggest that VGLUT1 may act at developing corticocorti-

onto subplate neurons in neonatal rat somatosensory cortex, J.cal axons. Neurosci. 22 (2002) 7165–7176.

[17] E. Herzog, G.C. Bellenchi, C. Gras, V. Bernard, P. Ravassard, C.Bedet, B. Gasnier, B. Giros, S. El Mestikawy, The existence of asecond vesicular glutamate transporter specifies subpopulations ofA cknowledgementsglutamatergic neurons, J. Neurosci. 21 (2001) U1–U6.

[18] A. Kalsbeek, P. Voorn, R.M. Buijs, C.W. Pool, H.B.M. Uylings,This work was supported by a grant from CNR (Proget- Development of the dopaminergic innervation in the prefrontal

to Strategico Neuroscienze). We thank F. Quagliano for cortex of the rat, J. Comp. Neurol. 269 (1988) 58–72.technical help. [19] V.N. Kharazia, R.J. Weinberg, Glutamate in thalamic fibers ter-

314 A. Minelli et al. / Developmental Brain Research 140 (2003) 309–314

minating in layer IV of primary sensory cortex, J. Neurosci. 14 Molecular cloning and functional identification of mouse vesicular(1994) 6021–6032. glutamate transporter 3 and its expression in subsets of novel

[20] M. Latsari, I. Dori, J. Antonopoulos, M. Chiotelli, A. Dinopoulos, excitatory neurons, J. Biol. Chem. 2002 Oct 15 [epub ahead ofNoradrenergic innervation of the developing and mature visual and print].motor cortex of the rat brain: a light and electron microscopic [25] S. Takamori, J.S. Rhee, C. Rosenmund, J. Reinhard, Identificationimmunocytochemical analysis, J. Comp. Neurol. 445 (2002) 145– of a vesicular glutamate transporter that defines a glutamatergic158. phenotype in neurons, Nature 407 (2000) 189–194.

[21] M. Melone, L. Vitellaro-Zuccarello, A. Vallejo-Illaramendi, A. [26] C.G. Van Eden, Development of connections between the mediodor-Perez-Samartin, C. matute, A. Cozzi, D.E. Pellegrini-Giampietro, sal nucleus of the thalamus and the prefrontal cortex in the rat, J.J.D. Rothstein, The expression of glutamate transporter GLT-1 in Comp. Neurol. 244 (1986) 349–359.the rat cerebral cortex is down-regulated by the antipsychotic drug [27] H. Varoqui, M.K.-H. Schafer, H. Zhu, E. Weihe, J.D. Erickson,clozapine, Mol. Psychiatry 6 (2001) 380–386. Identification of the differentiation-associated Na1 /Pi transporter as

[22] B. Ni B, P.R. Rosteck, N.S. Nadi, S.M. Paul, Cloning and expression a novel vesicular glutamate transporter expressed in a distinct set of1of a cDNA encoding a brain-specific Na -dependent inorganic glutamatergic synapses, J. Neurosci. 22 (2002) 142–155.

phosphate cotransporter, Proc. Natl. Acad. Sci. USA 91 (1994) [28] S.P. Wise, E.G. Jones, The organization and postnatal development5607–5611. of the commissural projection of the rat somatic sensory cortex, J.

[23] B. Ni, X. Wu, G.-M. Yan, J. Wang, S.M. Paul, Regional expression Comp. Neurol. 168 (1976) 313–344.1and cellular localization of the Na -dependent inorganic phosphate [29] S.P. Wise, E.G. Jones, Developmental studies of thalamocortical and

cotransporter of the rat brain, J. Neurosci. 15 (1995) 5789–5799. commissural connections in the rat somatic sensory cortex, J. Comp.[24] M.K. Schafer, H. Varoqui, N. Defamie, E. Weihe, J.D. Erickson, Neurol. 178 (1978) 187–208.