A developmentally regulated nerve growth factor-induced gene, VGF, is expressed in geniculocortical afferents during synaptogenesis

Neuroscience Vol. 65, No. 4, pp. 997 1008, 1995

~ )

Pergamon

0306-4522(94)00538-9

Elsevier ScienceLid Copyright ~ 1995 1BRO Printed in Great Britain. All rights reserved 0306-4522/95 $9.50 + 0.00

A DEVELOPMENTALLY REGULATED NERVE GROWTH FACTOR-INDUCED GENE, VGF, IS EXPRESSED IN GENICULOCORTICAL AFFERENTS DURING SYNAPTOGENESIS A. L O M B A R D O , * S. A. RABACCHI,f~.]] F. C R E M I S I , ++* T. PIZZORUSSO,+:~ M. C. C E N N I , t R. POSSENTI,§ G. B A R S A C C H I * and L. M A F F E I t ++ *Lab. di Biologia Cellulare dello Sviluppo, Dip di Fisiologia e Biochimica, Universit~i di Pisa, Italy ?lstituto di Neurofisiologia del CNR, Pisa, Italy +Scuola Normale Superiore, Pisa, Italy §Lab. di Neurobiologia del CNR, Roma, Italy Ahstrae~The expression of the nerve growth factor-inducible gene VGF has been examined by in situ hybridization, Western blot and immunohistochemical studies in the developing and adult rat central nervous system, with particular emphasis on the visual system. Both the messenger RNA and the protein are particularly abundant in the developing dorsal lateral geniculate nucleus, appearing, respectively, at embryonal day 16 and 18. After its onset at El6, VGF messenger RNA expression increases progressively in the dorsal lateral geniculate nucleus and remains high during the first two post-natal weeks; afterwards, it gradually decreases and, at the offset of the plasticity period, it reaches very low levels maintained in adulthood. A similar time course has been observed for VGF protein in the dorsal lateral geniculate nucleus area, by semi-quantitative Western blots. In addition to the presence of the protein in the geniculate neurons, a strong, transient immunoreactivity has been found at the embryonic cortical subplate at El8, reflecting the presence of the antigen in axonal terminals originating from thalamic neurons. Interestingly, we found that the blockade of afferent electrical activity by intraocular injection of tetrodotoxin strongly reduces the level of VGF messenger RNA in the dorsal lateral geniculate nucleus. Although the function of the VGF protein is not known, it had been previously proposed that VGF could be a precursor for neuropeptide/s. The spatiotemporal expression of VGF, together with the observation of a regulation by electrical activity, suggest that this protein may be relevant in the process of synaptogenesis and/or synaptic stabilization in the developing geniculocortical connections.

Neuronal development comprises several steps including axonal elongation and arborization, synaptogenesis and synapse maturation. All these processes can be influenced by the exchange of information between the developing neuron and its environment. The nature of this cross-talk involves both biochemical factors and electrical activity. Several studies, many of which focused on the developing visual system, have shown that interfering with trophic factors, 7'13"21'22'25"2s neuropeptides, 32 adhesion molecules, 4° as well as neuronal electrical activity 5'3s dramatically affects the development and plasticity of neuronal projections, The molecular mechanisms underlying these processes are yet to be elucidated;

[iTo whom correspondence should be addressed. Abbreviations: DAB, 3,3' diaminobenzidine; dLGN, dorsal

lateral geniculate nucleus; MAP2, microtubule associated protein 2; Mg, medial geniculate nucleus; NGF, nerve growth factor; P0, postnatal day 0; S D S ~ A G E , sodium dodecyl sulphate-polyacrylamide gel electrophoresis; TCA, trichloroacetic acid; TTX, tetrodotoxin; vLGN, ventral lateral geniculate nucleus; VPL, ventral posterolateral nucleus; VPM, ventral posteromedial nucleus. 997

the modulation of expression of specific genes is likely to occur. One strategy that can be adopted to identify these molecules is the in vivo study of genes that are involved in neuronal differentiation in vitro. In the present study we analyse the developmental expression and regulation by electrical activity of a gene that has been previously identified in the neuronal differentiation pathway of the rat pheochromocytoma cell line PC12.~7 The exposure of these cells to nerve growth factor ( N G F ) activates a series of different genes leading to the acquisition of their neuronal phenotype. 24 The gene here described, designated VGF, 23 belongs to the class of the late responsive genes; it codes for a polypeptide of mol. wt 65,000 that shares similarities with the secretogranin/chromogranin family. 34 In vivo studies on the centraP 6 as well as peripheral ~° nervous system have shown that V G F distribution is restricted to neuronal and neuroendocrine cells. In P C I 2 cells, V G F polypeptide is located in synaptic large dense core vesicles; 34 cell depolarization regulates both its transcription and release. 34'3637 The V G F amino acid sequence shows a high content of basic residues which represent potential targets for a proteolytic

998

A. Lombardo et al.

cleavage. F u r t h e r m o r e , in the differentiation of PC 12 cells a n d cerebellar granule cells, the V G F protein is cleaved into two m a j o r p r o d u c t s o f mol. wt 10,000 a n d 18,000 t h a t are preferentially secreted u p o n depolarization. 36 All these properties strongly support the hypothesis that V G F protein is a precursor for biologically active peptides i m p o r t a n t in n e u r o n a l c o m m u n i c a t i o n . 10,34,37 T h r o u g h in situ hybridization, Western blot a n d i m m u n o h i s t o c h e m i c a l analysis we studied the distrib u t i o n of V G F m R N A a n d protein in the rat central nervous system with particular emphasis on the developing visual system. The s p a t i o t e m p o r a l expression o f V G F in n e u r o n s of the dorsal lateral geniculate nucleus ( d L G N ) a n d in their terminals at the cortex d u r i n g the period o f synaptogenesis as well as the regulation of V G F transcription by electrical activity have been investigated. A partial report of these results has been published in form o f a n abstract. 26 EXPERIMENTAL PROCEDURES

in situ hybridization Six timed pregnant Long Evans rats (Charles Rivers, I) were used to obtain embryos ranging from El4 to El9. Under chloral hydrate anaesthesia (4 ml/kg of a 0.64 M solution), embryos were removed and fixed overnight in 4% paraformaldehyde, dehydrated with ethanols and xylene, and embedded in paraffin. Embryos were sectioned at 7 ram, mounted on organosilanized slides and stored at 4°C until processed. Postnatal rats (ranging from postnatal day 0, P0, to adulthood) were anaesthetized with chloral hydrate. After being killed, the brains (except for the P0, which were treated as the embryos) were removed and frozen on dry ice. Cryostat sections (15 mm) were mounted on organosilanized slides, fixed with 4% paraformaldehyde, dehydrated in graded ethanols and stored at -80°C. The in situ hybridization experiments were essentially performed as described, 45 with minor modifications. A EcoRI-SalI 580bp fragment of the VGF eDNA clone23 was subcloned into pGEM3Z (Promega) to generate a riboprobe for in situ hybridization. Radioactive, antisense and sense cRNA probes were copied from the linearized plasmid, by using [35S]UTP and T7 or SP6 polymerase, respectively. Paraffin sections were deparaffinized before prehybridization treatments. Sections were then digested with proteinase K (3 mg/ml; embryos and P0) or permeabilized with chloroform and rehydrated (postnatal brains). All sections were then acetylated with acetic anhydride (0.3% in 0.1 M triethanolamine) and hybridized overnight with 2 x 105 c.p.m, of 35S-labelled riboprobe at 50°C. The sections were washed, dehydrated and dipped in Kodak emulsion NTB2 as described. 47 After drying, the slides were stored for one to two weeks at 4°C, developed in Kodak D 19 (3 min at 16°C) and fixed in Kodak Unifix. The specificity of the in situ hybridization results was confirmed by applying two criteria: the use of a sense strand riboprobe which showed no detectable signals, and the addition of an excess of the same unlabelled template in the hybridization buffer, that resulted in the complete elimination of the hybridization signal. Western blot Dissected samples of visual cortex and thalamus (area containing the dorsal geniculate nucleus; n = 3) were rapidly frozen in liquid nitrogen and stored at - 8 0 ° C until process-

ing. Samples were resuspended in 500 ml of distilled water and rapidly homogenized with Polytron at the maximum speed (Polytron PT1200 Kinematica AG Swiss) for 30s. Samples were then boiled for 10min and centrifuged at 4000 r.p.m, to remove cellular debris. The supernatants were centrifuged for 30 min at 35,000 r.p.m, in a 50,000 Ti rotor. The supernatants were then TCA precipitated. The pellets were resuspended in 200 ml of Tris-HC1 50 mM (pH 8) and analysed for protein concentration (Lowry). Same amounts of protein (100mg) were loaded in a gradient gel 7.5/20% acrylamide sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) and analysed for Western blot using an antibody that recognizes the carboxy-terminal region of the VGF protein (Trani et al., unpublished observations). The detection kit was purchased from Amersham Life Science (England; ECL Western blotting detection reagents RPN2106) and used as indicated. Immunohistochemistry

Thirty-two developing (prenatal and postnatal) and seven adult Long Evans rats were anaesthetized with chloral hydrate (4 ml/kg of a 0.64 M solution) and perfused transcardially with 4% paraformaldehyde or, in some cases, with polylysine paraformaldehyde. 3° Some brains were cut on a Vibratome (40 mm) and the sections reacted with the antibody solution. Other brains were cryoprotected with 25% sucros~phosphate buffer, cut on a cryostat (13 mm) and the sections attached to gelatin-coated slides. Immunohistochemistry was performed by treating the sections (or the slides) for 2-3 h with an anti-VGF (1:300 1:500) antibody. Three rabbit polyclonal antibodies were raised 35against different fusion proteins corresponding to three non-overlapping sequences of the gene. The preimmune serum from each rabbit was used as a control of specificity. As secondary antibody, a biotynilated antirabbit antibody (Vector, 1 : I00) was used followed by ABC kit (ABC, Vector, U.K.); the peroxidase reaction was visualized by using the glucose oxidase-nickel~tiaminobenzidine (DAB) method. 42 Alternatively, a fluorescein avidin complex (Vector, U.K.) was applied after the secondary antibody. In some experiments, another primary antibody was applied in parallel, on adjacent sections: anti-MAP2 monoclonal (Boehringer, 0.5mg/ml). A secondary biotynilated antibody against mouse IgG (Calbiochem) was used at 1:1000 dilution for microtubule associated protein 2 .(MAP2) visualization, followed by avidin fluorescein. The specificity of the VGF antibodies has been verified by Western blot experiments on cerebral cortex tissue (data not shown), according to Ref. 35. Horse radish peroxidase labelling

Horse radish peroxidase (Boehringer, type I) was diluted (30%) in saline plus 2% dimethylsulphoxide. Nine microlitres of this solution were injected in one eye of a P24 rat under ether anaesthesia. After 24 h the rat was anaesthetized with chloral hydrate and perfused as described in Ref. 31 and the brain was removed and cryoprotected With 30% sucrose~hosphate buffer. After sinking, the brain was cut with a CO2-freezing microtome at 40mm. Sections were then reacted using tetramethylbenzidine as a chromogen as described in Ref. 31. Animal manipulations Tetrodotoxin administration. In order to block afferent activity to the dorsal dLGN, we administered tetrodotoxin (TTX) (Sigma), a blocker of the sodium channel, by intraocular injections with a pulled micropipette under ether anaesthesia. TTX was dissolved in a 3.5 mM citrate buffer solution (pH 4.8). Three treatments were performed. (1) TTX injected monocularly: P4 rats (n = 10) were treated with a 250 nl injection of a 0.25 mg/ml solution of

Developmental expression of VGF in the rat visual system TTX every 6-8 h (for 24 h). Each injection site was reused as many times as possible. The protocol of TTX injection has been based on the electrophysiological analysis previously performed at similar ages by Galli-Resta et aL ~2 P23 and adult rats were treated with a 1 ml injection of a I mg/ml solution. One group of P23 rats (n = 5) and one group of adult rats (n = 3 ) were killed 24h after the injection. A third group of P23 rats (n = 3) was killed 48 h after. In P23 or adult rats the pupillary response to illumination was adopted to monitor the efficacy of TTX in silencing the activity of optic nerve fibre. In all cases (P4, P23 and adults) the same citrate buffer vehicle solution in which the TTX was dissolved was used for control injections in the contralateral eye. The same injection protocol used for the TTX-treated eye was adopted. (2) TTX injected bilaterally: three P23 rats were treated by injecting TTX in both eyes as described for TTX injected monocularly. (3) TTX injected in visual cortex: l - l . 5 m l of TTX (1 mg/ml) was injected in the visual cortex of P23 rats (n = 2), while control solution was administered to the contralateral side. Dark rearing. Two rats were kept in total darkness from birth until killing; to ensure that the environment was completely light-tight a piece of fast photographic film was left in the room for a day.

999

Monocular deprivation. Seven rats were monocularly deprived for one month by means of eyelid suture starting immediately after eye opening (P14); this period corresponds to the entire critical period of the rat visual cortex. 8 The suture was placed under a brief ether anaesthesia. RESULTS

Developmental expression of VGF gene and protein In situ hybridization. By in situ hybridization experiments, V G F gene expression becomes detectable in the rat central nervous system at e m b r y o n i c day 16 (El6). A t this stage, V G F m R N A is nearly absent in the brain, except for low levels o f labelling present in the cells of the d L G N (not shown). D u r i n g the following days of development, the n u m b e r o f brain structures expressing the V G F gene appears to increase (Fig. 1). A t E l 9 the V G F a u t o r a d i o g r a p h i c signal is higher in the d L G N , but expression is also present in the M G (medial geniculate; n o t shown), V P L and V P M (ventral posterolateral a n d ventral posteromedial; Fig. IA). Several brain structures

Fig. 1. Expl~ession of the VGF gene in the rat brain. Dark field photomicrographs of in situ hybridizations with VGF 35S-labelled riboprobe on coronal sections at embryonic day 19 (A), postnatal day 7 (B) and postnatal day 23 (C). (D) VGF hybridization signal in the cell layer V (arrow) of the cerebral cortex of an adult rat. agd, amigdaloid nuclei; hip, hippocampal formation; hyp, hypothalamus; lgn, lateral geniculate nucleus; .prs, presubiculum; rsg/rsa, retrosplenial granular and agranular cortices; vpm/vpl, ventral posteromedial and posterolateral thalamic nuclei. Scale bars = 125 mm (A-C); 200 mm (D).

1000

A. Lombardo et al.

Fig. 2. Developmental expression of VGF gene in the dLGN. Dark field photomicrographs showing the hybridization signal obtained with a VGF riboprobe on dLGN coronal sections during development. (A) Embryonic day 19 (B-E). Postnatal days 7, 15, 23, 45, respectively. (F) Adult. Scale bars = 200 mm.

such as the suprachiasmatic, supraoptic, paraventricular, ventromedial hypothalamic nuclei, amygdala, cingulate and entorhinal cortex, become positive by postnatal day 7 (P7), although the labelling still prevails in the thalamic areas (Fig. 1B). The increase in the number of m R N A positive structures continues until adulthood when labelling is noticeable in a great part of the thalamus, in the cerebral cortex (layer V), as well as in the hippocampus (Fig. 1C, D). A detailed study of the development expression of V G F m R N A was performed in the d L G N (Fig. 2). While at El4 no signal was detected, a faint labelling

was found two days later (El6, data not shown). In the following days a gradual increase in V G F m R N A expression was observed (Fig. 2A, El9) with a peak of expression between P5 and P10 (Fig. 2B). At later ages the labelling in the d L G N decreases gradually (Fig. 2C-E) to the low levels found in the adult (Fig. 2F). Western blot. A comparable time course of expression has been observed for the V G F protein, as assessed by a semi-quantitative analysis of the levels of V G F protein in the visual cortex and dLGNcontaining area of the thalamus (Fig. 3). The anti-

Developmental expression of VGF in the rat visual system bodies used recognize the precursor forms of V G F (mol. wt 90-85,000) and the processed endoproteolitic products (lower bands). 34 In the thalamus, highest levels of the V G F protein have been found between P5 and P15 (Fig. 3B); this result is in agreement with both the in situ hybridization and the immunohistochemistry data (see below). The Western blot signal found in the visual cortex increases gradually throughout development (Fig. 3A), consistent with the immunohistochemical results showing that the early labelling of thalamic fibres is later followed by the appearance of many pyramidal cells (see below). lmmunohistochemistry--cell bodies. A complementary immunohistochemical analysis was performed in the developing and adult brain with three polyclonal antibodies to the V G F proteins. 35 All three antibodies, which were raised against different portions of the V G F polypeptide, gave comparable results. The following stages of development have been analysed: El6, El8, P0, P4, P10, P15, P23 and adult. The overall pattern of V G F cellular labelling reflects its m R N A distribution. The time course of V G F immunostaining parallels that of V G F m R N A with a delay of approximately two days. In fact, while no signal is found at embryonic day 16 (El6) with the antibodies, V G F protein becomes first detectable at El8, that is two days after the beginning of V G F m R N A expression. At this age immunoreactivity is found exclusively in the d L G N neurons, in fibres scattered in the cortex and, as a strong punctate pattern, at the cortical subplate (see below, Fig. 4A,D,F). Later in development, starting from Pl, other parts of the thalamus start to become labelled (MG, VPM, VPL; Fig. 4C). At P10 immunoreactivity is also found in other structures including a few cells in the cingulate cortex and in the stratum oriens of the hippocampus (Fig. 5A). A much weaker but significantly wider expression of V G F protein is seen in the adult brain (Fig. 5B). Structures labelled

VC 0

7

14

I001

with V G F antibodies in the adult include: dLGN (Fig. 5E), ventral lateral geniculate nucleus (vLGN; Fig. 5B), hypothalamic nuclei (paraventricular, suprachiasmatic and supraoptic), thalamic nuclei (MG, VPM, VPL), amygdala, hippocampus, cingulate and entorhinal cortices, many pyramidal cells in the cerebral cortex, particularly layer V (Fig. 5C,D), as well as scattered fibres in the cerebral cortex (Fig. 5F). Immunohistochemistrv--geniculocortical fibres. Immunoreactivity in the cerebral cortex appears first at El8 as a strong, punctate stripe running parallel to the surface at the level of the subplate (Fig. 4A); in addition, at this age, labelled fibres are seen below the subplate (Fig. 4F). At P0, although just vaguely recognizable, the punctate pattern seen in the cortex becomes more diffused, appearing as a thicker stripe (not shown). At P4 VGF cortical immunoreactivity is more or less uniformly distributed in the cortex, both in the form of dots and fibres. The punctate appearance of the protein staining and the absence of detectable VGF m R N A in the embryonic cortex by in situ hybridization (Fig. 1A) suggest that the presence of V G F protein at the subplate reflects an axonal terminal location. Since at El8 the only labelled neurons are found in the dLGN, the signal observed in the cortex most likely represents the axons of those neurons. To confirm the cortical subplate distribution of the V G F protein, we carried out immunostaining assays by using anti-MAP2 antibodies. The MAP2 protein is known to be specifically expressed by differentiated neurons. 6 Since subplate neurons are the first differentiating neurons in the rat embryonal brain, MAP2 can be used at early stages of development to specifically label subplate neurons? ~ While at El6 we found MAP2 distribution in the cortex to be mainly confined to a stripe corresponding to the subplate neurons, at El8 we found (Fig. 4B) MAP2 antibody labelling both at the subplate and in the cortical plate area (as cortical plate neurons start to

LGN 0

ad

CG 7

14

ad

- 120 - 88 - 70 - 4 6 " Jl

"

- 38- 32 -

"

,..,. I

~ I

I

"~~"I

IIII I

I Ill III

I II

Fig. 3. Time course of expression of the VGF protein in the visual cortex and dorsal lateral geniculate nucleus area (Western blot). The 85,000 90,000 mol. wt bands represent the precursor forms of VGF, while the smaller bands are its endoproteolitic products. VC, visual cortex; LGN, region of the thalamus containing the lateral geniculate nucleus; CG, cerebellar granule cells (primary cell culture) used as a control sample. The numbers indicate the age in postnatal days and the adult (Ad.).

1002

A. Lombardo et al.

Fig. 4. Distribution of the VGF protein in the rat developing nervous system. Coronal sections through the brain. (A, B) Adjacent sections of cerebral cortex (El8) labelled with anti-VGF antibody (A) and anti-MAP2 antibody (B), showing that the area intensely labelled in A corresponds to the cortical subplate indicated by the arrow in B. (C) Section through the thalamus of a P4 rat pup showing high immunoreactivity of the dorsal lateral geniculate nucleus; similar staining has been observed on El8 thalamus. Some labelling is also detected in the ventrolateral thalamic nucleus. (D, E) Higher magnification view of, respectively, A and 13, showing that while VGF immunoreactivity is characterized by a punctate pattern resembling a terminal location, MAP2 shows a typical cellular distribution. (F) High magnification view of a fiber located below the cortical subplate, labelled with anti-VGF antibody. Scale bars=A-C, 100mm; D F, 20mm.

mature2°'33). On adjacent E 18 sections the localization of VGF immunoreactivity (Fig. 4A) corresponds to the subplate zone, as revealed by MAP2 immunostaining (Fig. 4B). However, while MAP2 displays a typical cellular distribution (Fig. 4E), VGF immunolabelling reveals a punctate pattern at the subplate level (Fig. 4D), consistent with a distribution at the axonal terminals afferent to the subplate neurons. In addition, the observation of scattered fibres labelled with VGF antibody located below the cortical subplate (Fig. 4F) provides further support for the axonal terminal distribution of the protein. Interestingly, the distribution of VGF immunoreactivity in the embryonic and early postnatal visual cortex strictly parallels the arrival of geniculocortical fibres at the cortical subplate (El8) and the subsequent invasion of the cortical plate in early postnatal life, where they encounter an intense period of synap-

togenesis with neurons of various cortical layers}° All these observations support the idea that the VGF protein is synthetized by thalamic cells, and then conveyed to the axonal terminals in the cortical subplate and later in the developing cortical layers, where it could exert its action.

Modulation o f VGF expression by electrical activity Electrical activity is known to play a key role in the process of synaptogenesis and synaptic stabilization in the developing visual system (reviewed in Ref. 38). Given the temporal correlation between the peak of VGF expression in the dLGN and synaptogenesis in the visual cortex, we decided to investigate whether retinal electrical activity could regulate VGF gene expression in the visual system. Two different approaches were chosen to modulate retilaal afferent activity. First, we sutured the eyelid of

Developmental expression of VGF in the rat visual system one eye or we dark-reared rats to subtract from the retinal activity the component due to patterned vision. Second, we completely silenced retinal ganglion cells action potentials by monocular or binocular injections of TTX, a potent voltagedependent Na + channel blocker. Manipulations were performed at P4, P23 (during the rat critical period

for plasticity of the geniculocortical pathway 8) and on adult animals. In situ hybridization and immunohistochemistry were performed to assess possible changes in V G F m R N A and protein expression in the dLGN. While eyelid suture (n = 7) or dark-rearing (n = 2) did not lead to detectable alterations in V G F m R N A

Fig. 5. Developmental expression of the VGF protein in the rat postnatal brain (coronal sections). (A, B) Section through the thalamus at PI0 (A) and P45 (B). While at P10 the immunoreactivity is particularly intense in the dorsal lateral geniculate nucleus (arrow), the labelling becomes more diffused in the adult brain (B). (C) Adult cerebral cortex (P45) showing labelled pyramidal cells. (D) High magnification view of C. (E) High magnification photomicrograph of adult dorsal lateral geniculate neurons. (F) High magnification of a fibre located in the cerebral cortex at Plg, apparently engulfing a cell. Scale bars A, B, 1 ram; C, 240ram; D, 30mm; E, 17mm and F, 14mm. N S C 65/4

L)

1003

1004

A. Lombardo et al.

Fig. 6. Modulation of VGF expression by electrical activity. Dark field photomicrographs of in situ hybridizations with a VGF riboprobe on coronal sections of a P23 rat brain. Binocular injection of TTX completely abolished the VGF signal in both dLGNs (arrows). Scale bar = 1 mm.

expression, TTX treatments induced a striking reduction in the m R N A level in the d L G N cells deprived of afferent retinal activity when performed at P23 or in the adult rat. Binocular injections of TTX (P23, n = 3) completely abolished V G F m R N A signal in both dLGNs (Fig. 6) without altering nuclei that are not the target of retinal afferents, such as the thalamic VPL/VPM nuclei (Fig. 6). When TTX was given monocularly, a peculiar pattern of V G F expression was observed in the geniculate nucleus. Figure 7(C,D) shows V G F m R N A labelling in the d L G N ipsitateral (Fig. 7C) and contralateral (Fig. 7D) to the eye which had been injected with TTX. Figure 7(A,B) shows the subfields of the right (Fig. 7B) and left (Fig. 7A) d L G N innervated by the right eye, as revealed by intraocular horse radish peroxidase injection. By comparing Fig. 7(C,D) with Fig. 7(A,B) it is apparent that a dramatic reduction of V G F m R N A takes place in the d L G N subfields receiving input from the TTX-treated eye. No change in V G F m R N A labelling is present in the LGN subfields connected with the eye injected with the control solution. The effect of TTX is totally recovered in 48 h, as shown by the reappearance of homogeneous labelling of both dLGNs 48 h after a monocular injection of TTX (Fig. 7E,F). This result is consistent with the recovery of the light-evoked pupillary reflex. The reduction of the amount of V G F m R N A induced by intraocular TTX is not secondary to a reduction in the electrical activity of the visual cortex. TTX application onto the monocular region of the right visual cortex has no effect on the expression of V G F m R N A in the ipsilateral d L G N (n = 2; data not shown). TTX administered in vivo to younger animals (P4) failed to show a down-regulation of V G F mRNA. However, TTX treatment at this age proved to be difficult due to both its shorter duration and to its

high toxicity. Evidence that electrical activity can regulate V G F gene expression also at early stages of development comes from a study of V G F expression in cultured embryonic d L G N neurons. Preliminary in vitro experiments, in which explants of embryonic d L G N were cultured in a serum-free medium in depolarizing conditions (40 mM KC1), showed up regulation of V G F expression (Lombardo et al., unpublished observations). Immunohistochemical analysis of the d L G N 24-48 h after TTX administration did not show a similar modulation of the V G F protein by electrical activity; however, different hypotheses can be advanced to explain this apparently contradicting result (see Discussion). DISCUSSION

In several mammals thalamic fibre terminals are known to establish temporary synapses with subplate neurons in the early developing cerebral cortex 15'29 before invading the cortical plate and forming synapses with their definitive targets in layer IV. This transient population of cells is crucial for the development of the thalamocortical system. When the pharmacological ablation of subplate neurons takes place during cat embryonal life the formation of thalamocortical connections is inhibited; ~4 its destruction in early postnatal life prevents the development of ocular dominance columns. 16 In the rat, geniculocortical afferents arrive at the cortical subplate at El8, where they begin to establish synapses with the subplate neuronsfl '2° Subsequently, the geniculate terminals start invading the cortical plate and forming synaptic contacts with the primary target cells in layer IV as well as with other neurons in developing layers VI and V and marginal zone. z°~27The process of synaptogenesis continues in the rat cerebral cortex during the first two or three postnatal weeks. 45

Developmental expression of VGF in the rat visual system Our results show that the anatomical distribution of the V G F m R N A and protein during development in the geniculocortical pathway appears temporally and spatially correlated with the process of synaptogenesis. The following observations favour this hypothesis. (1) A temporal correlation between synaptogenesis and V G F gene expression is clearly

1005

present in the dLGN: V G F m R N A is first detectable by in situ hybridization at El6, whereas the protein appears two days later (El8), as shown by immunohistochemistry. V G F gene expression in the d L G N increases gradually and remains high during the first postnatal weeks, a period when new synapses with the definitive target neurons (layer IV) are established.

F

Fig. 7. Modulation of VGF expression by electrical activity. (A, B) Light field microphotographs showing the rat dLGN subfields of the left (A) and right (B) dLGN innervated by the right eye as revealed by intraocular horse radish peroxidase injection. (C F) Expression of the VGF gene in the dLGNs after a monocular TTX injection on a P23 rat. (C, D) Twenty-four hours after a TTX injection into one eye there is a dramatic reduction of VGF expression in the dLGN subfields receiving input from the TTX treated eye. (E, F) Forty-eight hours after the injection the effect of TTX on VGF expression is completely recovered. Scale bars = 200 mm.

1006

A. Lombardo et al.

Afterwards, both VGF m R N A and protein gradually decrease and at the offset of the plasticity period (P45) s they are found at very low levels. (2) The spatial distribution of VGF protein in the rat embryonal cerebral cortex correlates well with the development of thalamocortical projections. The appearance of the VGF protein in the visual cortex coincides with the arrival of the geniculocortical afferents at the level of the subplate (El8). 2°,27 Similarly, our observations of a gradual spreading of immunoreactivity to the developing layers of the cortex from P0 onwards correspond closely with the period of invasion of the cortical plate by the geniculocortical thalamic fibres, 45 as previously assessed by both retrograde degeneration studies 27 and application of carbocyanine dies and wheat germ agglutinin-horse radish peroxidase. 2° The inference that VGF protein is present in geniculate fibres is based on several observations: (1) the punctate pattern observed in the embryonal cortical subplate, (2) the complete absence of cellular labelling at El8, except for thalamic nuclei, (3) the precise correspondence of VGF immunoreactivity in the visual cortex with the distribution of the geniculocortical afferents at different developmental stages. 2°'27 The present observations suggest that VGF protein is synthesized by dLGN neurons and then transported along their axons to their postsynaptic cortical neurons, where it exerts its function. In addition to the description of VGF gene expression during development, we also show that the blockade of afferent retinal activity induces a dramatic reduction of VGF m R N A in the dLGN neurons. The reduction in VGF expression following blockade of afferent activity is probably not due to unspecific damaging effects of the TTX injections, since this down-regulation is reversible, in agreement with the reversible properties of the drug. In addition, since neither the eye injections nor the vehicle solution in which the TTX was dissolved affected VGF m R N A levels, this suggests that the reduction in VGF here described is due to TTX. These data, together with the temporal correlation between innervation of dLGN by retinal afferents and beginning of VGF m R N A expression (both occurring at E163), seem to indicate that VGF gene expression is highly regulated by electrical activity. It has to be underlined that VGF mRNA is dramatically down-regulated in the dLGN only if retinal activity is completely abolished. Monocular deprivation or dark rearing appear to be ineffective. These results indicate that VGF gene expression is regulated by spontaneous, vision-independent, retinal activity. The molecular mechanisms mediating the effect of TTX on the regulation of VGF m R N A are not known. The signal inducing the down-regulation must pass through the retinal afferents, since no alteration was observed in other areas not receiving input from the TTX-treated eye. The first trigger is

likely to be the lack of electrical activity; this, in turn, is known to reduce neuronal metabolism. 44'4s However, the diminished metabolism does not determine a generalized reduction in gene expression since both up-regulation and down-regulation of different genes have been observed, l't9 We cannot rule out that other genes may mediate the down-regulation of VGF mRNA induced by silencing retinal afferents. However, supporting evidence that electrical activity could directly regulate VGF expression, comes from the presence of a calcium-regulatory sequence present in the promoter region of this gene. 35'37This regulation could reflect a role of VGF gene in activity-dependent plasticity phenomena. The result that VGF gene expression is regulated only by spontaneous visionindependent retinal activity is not contrary to this hypothesis. In fact, spontaneous electrical activity is important in defining topographical maps before eye opening, 11,39but also plays an important role after eye opening. 43 Several possibilities should be considered in interpreting the lack of modulation of the VGF protein by TTX. (1) The modulation of a gene is more easily observed at the level of transcription; a high stability of the protein would require much longer TTX treatments in order to make a change in VGF protein levels perceivable. When longer periods of treatment with repetitive injections were attempted the survival was very low, especially on young animals. (2) VGF release has been shown to be activity-dependent in PC12 cells; if this also occurs i n vivo, TTX injection, by preventing VGF release, could mask a reduction in VGF synthesis. (3) Besides the possible low turnover rate of the protein and its activity-dependent release, the low sensitivity of the immunohistochemical technique could hinder the detection of differences in antigen levels. As to the possible mechanisms of action of VGF in the developing visual system, we propose that VGF may act as an anterograde factor possibly released from the presynaptic terminals in an activitydependent way. Some observations support this hypothesis: (1) in vitro studies on PC12 cells show that the VGF protein is present in synaptic large dense core vesicles and released after depolarization; 35 (2) our immunohistochemical studies reveal the presence of VGF in the terminals of the dLGN in the visual cortex; (3) VGF shares some properties with neuropeptides: its induction by different factors and by cAMP, 1s,35,37its release in PC12 cells induced by several secretagogue agents, 34 the regulation of its transcription by electrical activity in vitro 37 and in vivo (present work) and the presence of many products of proteolytic cleavage. In particular, recent experiments suggest that the VGF polypeptide is cleaved in smaller peptides of mol. wt 18,000 and 10,000. 36 Likewise VGF, other peptides have been found to be present at higher levels in the developing brain compared with the adult (for references see Ref. 39). This has raised the hypothesis that

Developmental expression of VGF in the rat visual system peptides can be involved in the modelling of synaptic connectionsf CONCLUSION

I O07

for a role during the process of synaptogenesis in the developing visual system.

Acknowledgements--We are grateful to the technical staff of

In conclusion, the study of V G F distribution a n d regulation d u r i n g d e v e l o p m e n t gives hints on the possible functions o f this gene a n d candidates V G F

the Istituto Neurofisiologia CNR of Pisa for its excellent support. This project was partially supported by CNR (Progetto finalizzato Ingegneria Genetica), The EPA Cephalosporin Fund, and the HFSPO grant RG93/93.

REFERENCES

1. Benson D. L., Isackson P. J., Gall C. M. and Jones E. G. (1991) Differential effects of monocular deprivation on glutamic acid decarboxylase and type II calcium-calmodulin-dependent kinase gene expression in the adult monkey visual cortex. J. Neurosci. 11, 3147. 2. Blue M. E. and Parnavelas J. G. (1983) The formation and maturation of synapses in the visual cortex of the rat--I. Qualitative analysis. J. Neurocytol. 12, 599~16. 3. Bunt S. M., Lund R. D. and Land P. W. (1983) Prenatal development of the optic projection in albino and hooded rats. Devl Brain Res. 6, 149-168. 4. Chun J. J., Nakamura M. J. and Shatz C. J. (1987) Transient cells of the developing mammalian telencephalon are peptide-immunoreactive neurons. Nature 325, 617~20. 5. Constantine-Paton M., Cline H. T. and Debski E. (1990) Patterned activity, synaptic convergence and the NMDA receptor in developing visual pathways. A. Rev. Neurosci. 13, 129-154. 6. De Camilli P., Miller P. E., Navone F., Theurkauf W. E. and Vallee R. B. (1984) Distribution of microtubule-associated protein 2 in the nervous system of the rat studied by immunofluorescence. Neuroscience 11, 819-846. 7. Domenici L., Cellerino A. and Maffei L. (1993) Monocular deprivation effects in the rat visual cortex and lateral geniculate nucleus are prevented by nerve growth factor (NGF)--II. Lateral geniculate nucleus. Proc.LR. Soc. Lond. B 251, 25-31. 8. Fagiolini M., Pizzorusso T., Berardi N., Domenici L. and Maffei L. (1994) Functional postnatal development of the rat primary visual cortex and role of visual experience: dark rearing and monocular deprivation. Vision Res. 34, 709 720. 9. Ferrer I., Soriano E., Del Rio J. A., Alcantara S. and Auladell C. (1992) Cell death and removal in the cerebral cortex during development. Prog. Neurobiol. 39, 1~43. 10. Ferri G. L., Levi A. and Possenti R. (1992) A novel neuroendocrine gene product: selective VGF protein in endocrine and neuronal population. Molec. Brain Res. 13, 139-143. 11. Galli L. and Maffei L. (1988) Spontaneous impulse activity of rat retinal ganglion cells in prenatal life. Science 242, 90-91. 12. Galli-Resta L., Ensini M., Fusco E., Gravina A. and Margheritti B. (1993) Afferent spontaneous electrical activity promotes the survival of target cells in the developing retinotectal system of the rat. J. Neurosci. 13, 243 250. 13. Garofalo L., Ribeiro-da-Silva A. and Cuello A. C. (1992) Nerve growth factor-induced synaptogenesis and hypertrophy of cortical cholinergic terminals. Proc. natn. Acad. Sci. U.S.A. 89, 2639-2643. 14. Ghosh A., Antonini A., McConnel S. K. and Shatz C. J. (1990) Requirements for subplate neurons in the formation of thalamocortical connections. Nature 347, 179-181. 15. Ghosh A. and Shatz C. J. (1992) Pathfinding and target selection by developing geniculocortical axons. J. Neurosci. 12, 39-55. 16. Ghosh A. and Shatz C. J. (1992b) Involvement of subplate neurons in the formation of ocular dominance columns. Science 255, 1441-1443. 17. Greene L. A. and Tischler A. S. (1976) Establishment of a noradrenergic clonal line of rat adrenal pheocromocytoma cells which respond to nerve growth factor. Proc. natn. Acad. Sci. U.S.A. 73, 2424-2428. 18. Hawley R. J., Scheibe R. J. and Wagner J. A. (1990) NGF induces the expression of the VGF gene through a cAMP response element. J. Neurosci. 12, 2573 2581. 19. Hevner R. F. and Wong-Riley M. T. T. (1993) Mitochondrial and nuclear gene expression for cytochrome oxidase subunits are disproportionally regulated by functional activity in neurons. J. Neurosci. 13, 1805-1819. 20. Kageyama G. H. and Robertson R. T. (1993) Development of geniculocortical projections to visual cortex in rat: evidence for earlyingrowth and synaptogenesis. J. comp. Neurol. 335, 123-148. 21. Korsching S. (1993) The neurotrophic factor concept: a reexamination. J. Neurosci. 13, 2739-2748. 22. Levi-Montalcini R. (1987) Nerve Growth Factor 36 years later. Science 237, 1154-1162. 23. Levi A., Eldridge J. D. and Paterson B. M. (1985) Molecular cloning of a gene sequence regulated by nerve growth factor. Science 229, 393 395. 24. Levi A. and Alemfi S. (1991) The mechanism of action of nerve growth factor. A. Rev. Pharrnac. Toxic. 31,205-228. 25. Lohof A. M., Ip N. Y. and Poo M. (1993) Potentiation of developing neuromuscular synapses by neurotrophins NT-3 and BDNF. Nature 363, 350-353. 26. Lombardo A., Cremisi F., Rabacchi S. A., Pizzorusso T., Cenni M. C., Barsacchi G., Maffei L. and Possenti R. (1993) A developmentally regulated NGF-induced gene, VGF, is regulated by electrical activity in the rat visual system. Soc. Neurosci. Abstr. 456.7. 27. Lund R. D. and Mustari M. J. (1977) Development of the geniculocortical pathway in rats. J. comp. Neurol. 173, 289-306. 28. Maffei L., Berardi N., Domenici L., Parisi V. and Pizzorusso T. (1992) Nerve growth factor (NGF) prevents the shift in ocular dominance distribution of visual cortical neurons in monocularly deprived rats. J. Neurosci. 12, 4651~,662. 29. McConnel S. K., Ghosh A. and Shatz C. J. (1989) Subplate neurons pioneer the first axons pathway from the cerebral cortex. Science 245, 978-982.

1008

A. Lombardo et al.

30. McLean I. W. and Nakane P. K. (1974) Periodate-lysine-paraformaldehyde fixative a new fixative for immunoelectron microscopy. J. Histochem. Cytochem. 22, 1077-1083. 31. Mesulam M. M. (1978) Tetramethyl benzidine for horseradish peroxidase neurohistochemistry: a non-carcinogenic blue reaction-product with superior sensitivity for visualizing neural afferents and efferents. J. Histochem. Cytochem. 26, 106-117. 32. Mudge A. W. (1989) Neuropeptides find a role. Nature 339, 663~564. 33. O'Leary D. D. and Koester S. E. (1993) Development of projection neuron types, axon pathways, and patterned connections of the mammalian cortex. Neuron 10, 991 1006. 34. Possenti R., Eldridge J. D., Paterson B. M., Grasso A. and Levi A. (1989) A protein induced by NGF in PC 12 cells is stored in secretory vesicles and released through the regulated pathway. EMBO J. 8, 2217-2223. 35. Possenti R., Di Rocco G., Nasi S. and Levi A. (1992) Regulatory elements in the promoter region of VGF, a nerve growth factor-inducible gene. Proc. Natl. Acad. Sci. U.S.A. 89, 3815-3819. 36. Possenti R., Rinaldi A. M., Canu N., Levi A., Ferri G. L. and Trani E. (1994) Neuronal specific processing of VGF protein leads to the production and regulated secretion of peptides derived from its carboxi-terminal. Soc. Neurosci. Abstr. (In press). 37. Salton S. R. J., Fischberg D. J. and Dong K.-W. (1991) Structure of the gene encoding VGF, a nervous system-specific mRNA that is rapidly and selectively induced by Nerve Growth Factor in PC12 cells. Molec. cell. Biol. 11, 2335 2349. 38. Shatz C. J. (1990) Impulse activity and the patterning of connections during CNS development. Neuron 5, 745-756. 39. Shatz C. J. and Stryker M. P. (1988) Prenatal tetrodotoxin infusion blocks segregation of retinogeniculate afferents. Science 242, 87-89. 40. Schubert D. (1991) The possible role of adhesion in synaptic modification. Trends Neurosci. 14, 127 131. 41. Schwartz J. P. (1992) Neurotransmitters as neurotrophic factors: a new set of functions. Int. Rev. Neurobiol. 34, l 23. 42. Shu S., Ju G. and Fan L. (1988) The glucose oxidase-DAB-nickel method in peroxidase histochemistry of the nervous system. Neurosci. Lett. 85, 169 171. 43. Stryker M. P. and Harris W. (1986) Binocular impulse blockade prevents the formation of ocular dominance columns in cat visual cortex. J. Neurosci. 6, 2117-2133. 44. Thurlow G. A. and Cooper R. M. (1991) Activity-dependent changes in eye influence during monocular blockade: increases in the effects of visual stimulation on 2-DG uptake in the adult rat geniculostriate system. J. comp. Neurol. 306, 697 707. 45. Uylings H. B. M., Van Eden C. G., Parnavelas J. G. and Kalsbeek A. (1990) The prenatal and postnatal development of rat cerebral cortex. In The Cerebral Cortex of the Rat (ed. Kolb B. and Tees R. C.). MIT Press, Cambridge, MA. 46. Van den Pol A. N., Decavel C., Levi A. and Paterson B. (1989) Hypotalamic expression of a novel gene product, VGF: immunocytochemical analysis. J. Neurosci. 9, 4122~,137. 47. Wilkinson D. G. and Green J. (1990) In situ hybridization and three-dimensional reconstruction of serial sections. In Practical Approach Series: Postimplantation Mouse Embryos: A Practical Approach (ed. Rickwood D. and Crockcroft D. I.), pp. 155-171. IRL Press, Oxford. 48. Wong-Riley M. T. T. (1989) Cytochrome oxidase: an endogenous marker for neuronal activity. Trends Neurosci. 12, 94 101. (Accepted 19 October 1994)