Localization and seizure-regulation of integrin β1 mRNA in adult rat brain

Molecular Brain Research 55 Ž1998. 265–276

Research report

Localization and seizure-regulation of integrin b 1 mRNA in adult rat brain Jason K. Pinkstaff a , Gary Lynch a

b,c

, Christine M. Gall

a,b,)

Department of Anatomy and Neurobiology, UniÕersity of California, IrÕine, CA 92697-1275, USA b Department of Psychobiology, UniÕersity of California, IrÕine, CA 92697-1275, USA c Department of Psychiatry, UniÕersity of California, IrÕine, CA 92697-1275, USA Accepted 23 December 1997

Abstract Recent findings indicate that RGD-binding integrin receptors play a critical role in the maintenance of long-term potentiation but the identity and location of the integrin proteins involved are not known. The integrin b 1 is of particular interest in regard to synaptic plasticity because it is a component of many of the RGD-binding integrins and b 1-immunoreactivity has been localized within synaptic density fractions. The present study used in situ hybridization to evaluate the distribution of b 1 mRNA in adult rat brain and to determine if expression is altered by seizures. In untreated rats, b 1 mRNA is present at high levels in the ventricular epithelium and discrete neuronal groups including the magnocellular hypothalamic and efferent cranial nerve nuclei and the cerebellar Purkinje cells. Hybridization was less dense in the substantia nigra and hippocampal stratum pyramidale and low but present throughout the gray matter. Limbic seizures increased b 1 cRNA labeling of both neurons Že.g., hippocampal stratum pyramidale. and astroglial cells from 8 h through 48 h after seizure onset. These results indicate that in adult rat brain, b 1 mRNA is expressed by both neurons and glia; neuronal expression is highest in hypothalamic and peripherally projecting neurons capable of substantial morphological plasticity. Seizure effects demonstrate that b 1 is positively regulated by activity, and suggest that activity-dependent expression may play a role in synaptic plasticity in the adult brain. q 1998 Elsevier Science B.V. Keywords: Adhesion protein; Brain; Astrocyte; Hippocampus; Magnocellular nuclei; Long-term potentiation

1. Introduction Integrins are trans-membrane receptors that function in the recognition of extracellular matrix and cell-surface proteins. These receptors are heterodimers of non-covalently associated a and b subunits. To date, 22 subunits Ž14 a and 8 b . have been described w34x. Subunit composition varies across the integrin receptors and determines both ligand specificity and intracellular signalling activities. Integrins have been shown to be important in axonal guidance and proper synapse formation during the development of the central and peripheral nervous systems w55x but there is relatively little information on the expression and functional properties of these adhesion molecules in the adult brain. However, recent findings suggest that the integrins, and other types of adhesion proteins, are impor) Corresponding [email protected]

author.

Fax:

q 1-714-824-8549;

E-mail:

0169-328Xr98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S 0 1 6 9 - 3 2 8 X Ž 9 8 . 0 0 0 0 7 - 2

tant for synaptic plasticity w16x. In acute hippocampal slices, bath application of peptides that block the extracellular binding domain of several integrins Ži.e., the RGD-binding integrins., blocks the stabilization, but not the induction, of long-term potentiation ŽLTP. w65x. In similar preparations, antibodies to the immunoglobulin cell adhesion molecules L1 and neural cell adhesion molecule ŽNCAM. w36,56x, or enzymatic removal of the polysialic acid portions of NCAM w40x, prevent the induction but not the long-term stability of LTP. Moreover, spatial learning is reportedly impaired when NCAM levels are reduced by intraventricular antibody injection w14x or mutation w13,60x. Together these findings indicate that different classes of adhesion molecules play critical but differentiated roles in forms of synaptic plasticity underlying learning and memory. The identity and cellular localization of the integrin receptors involved in LTP are not known. In fact, the brain localization of just two subunits, a 8 and b 1, has been examined w15,25,44,47x. Within the hippocampal forma-

266

J.K. Pinkstaff et al.r Molecular Brain Research 55 (1998) 265–276

tion, a 8-immunoreactivity Žir. was localized to neuronal perikarya and dendrites, with particularly high levels in spines and dendritic post-synaptic densities in field CA3 w15x. Immunoreactivity for the b 1 subunit has been localized within hippocampus w25x, and a smaller Ž55 kDa. protein with b 1-ir is enriched in hippocampal post synaptic density fractions w4x, but the cellular localization of this subunit remains controversial. Grooms et al. Ž1993. reported that b 1-ir was localized within glial cells and fine puncta aligned with the apical dendrites of hippocampal pyramidal neurons, but also noted that staining was most prominently associated with the vasculature in adult rat brain. In contrast, in an analysis of the developing and adult cerebellum, Murase and Hayashi Ž1996. w44x found b 1 mRNA and immunoreactivity were exclusively localized to neurons, being prominently expressed by the Purkinje cells and neurons in the underlying vestibular nuclei. These data indicate that both a 8 and b 1 subunits are present in fields in which integrin involvement in LTP has been documented, but the cell types expressing b 1 mRNA have yet to be resolved. The present study used in situ hybridization techniques to evaluate the cellular and regional localization of b 1 integrin mRNA in adult rat brain. The b 1 subunit was considered important for further study because this subunit associates with many different a subunits in the formation of several RGD-binding receptors w34x that may play a role in hippocampal LTP. In addition, the involvement of integrin receptors in functional synaptic plasticity led us to question if b 1 gene expression is regulated by neuronal activity. To address this issue, the localization and levels of b 1 integrin mRNA were evaluated in rats killed at various time points after experimentally induced seizures. The results demonstrate that b 1 mRNA is normally expressed at high levels by discrete neuronal groups with notable capacities for morphological plasticity. Within the hippocampal formation, b 1 mRNA levels are very low in untreated rats but increase in both neurons and glia after seizures. Some of these results have been published in abstract form w50x.

2. Materials and methods 2.1. Animal treatments Adult, male Sprague–Dawley rats Ž250–300 g. were used Ž n s 38.. For induction of limbic seizures, rats were anesthetized with ketamine Ž50 mgrkg. and xylazine Ž10 mgrkg. and a unilateral electrolytic lesion was placed in the right dentate gyrus hilus using an insulated stainless steel electrode and anodal current Ž80 m A, 7 s.. The iron deposition from such lesions causes recurrent, bilateral limbic seizures that generally begin 2 h postlesion and continue for 8–10 h thereafter w12,48x. Only rats exhibiting at least two stage-four seizures Ži.e., rearing with clonus.

w54x were used in the present study. Anesthetic-control rats Ž n s 3. used for qualitative and quantitative in situ hybridization analysis, were similarly anesthetized but did not undergo surgery. Additional rats, used to evaluate the effects of electrode- and lesion-induced damage in the absence of seizures, included rats killed 24 h after Ži. a non-convulsant platinum–iridium wire hilar lesion Ž n s 4. w48x, Žii. a misplaced stainless steel wire hilar lesion that did not induce seizures Ž n s 1., and Žiii. a sham lesion Ži.e., electrode lowered but no current passed. Ž n s 1.. Rats were killed for Ž1. in situ hybridization using an overdose of sodium pentobarbital Ž100 mgrkg, i.p.. and transcardial perfusion with 0.9% saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4 ŽPB., and Ž2. RNA isolation by flourothane anesthesia followed by decapitation. For in situ hybridization analyses, hilus lesion ŽHL. rats were killed 4, 8, 18, 24, 48 and 96 h after the lesion; anesthetic control rats were killed 12 h after anesthesia. For Northern blot analysis, RNA was extracted from hilus lesion rats at 4, 12, 18, 24, 48 and 96 h post lesion, and from an untreated control rat, as described below Ž n s 1 each.. For comparison to the effects of hilus lesion-induced seizures, three additional rats received an i.p. injection of kainic acid Ž12 mgrkg., were observed for verification of seizure behavior and, at 3 h after kainic acid, received sodium pentobarbital Ž100 mgrkg i.p.. to terminate seizure activity. At 24 h after the initial injection, kainic acid treated rats and a paired control were killed and the brains fixed Žas above. for in situ hybridization analysis. Through all animal procedures, an effort was made to minimize the stress and discomfort of experimental animals. Protocols used have been approved by the appropriate institutional animal care and use committee which has been assigned Animal Welfare Assurance number 3416.01 Žon file with the National Institutes of Health.. 2.2. In situ hybridization Brains were post-fixed in 4% paraformaldehyde in PB Ž24 h, 48C., cryoprotected in 20% sucroser4% paraformaldehyde Ž24–48 h, 48C., and then sectioned on a freezing microtome Ž25 m m, coronal plane. and collected into 4% paraformaldehyderPB. For b 1 mRNA localization, the free-floating tissue sections were processed for in situ hybridization with the 35 S-labeled rat b 1 antisense cRNA, as previously described w21x, with the probe at a concentration of 1 = 10 7 cpmrml and hybridization at 608C for 36–48 h. Sections from experimental-seizure and control rats were processed together Žin cohorts of up to nine rats.. To assess labeling specificity, alternate sections from control and experimental seizure rats were similarly processed for hybridization with the 35 S-labeled sense sequence. As described in Section 3, no labeling was observed within brain tissue processed for hybridization with the sense riboprobe. After hybridization, the tissue was treated with

J.K. Pinkstaff et al.r Molecular Brain Research 55 (1998) 265–276

20 mgrml ribonuclease A Ž30 min, 458C. and washed through descending concentrations of saline sodium citrate buffer ŽSSC; 1 = SSC s 150 mM NaCl, 15 mM sodium citrate, pH 7.0. with the final wash in 0.1 = SSC at 608C. The sections were mounted on gelatin-coated slides, allowed to dry and exposed to Amersham b-Max autoradiographic film for 72 h. The tissue was then defatted in ethanol and chloroform, rehydrated through a descending series of alcohols, dipped in Kodak NTB2 emulsion Ž1:1 with H 2 O. and exposed for 3–5 weeks at 48C. Tissue autoradiograms were developed ŽKodak D-19., fixed, stained with Cresyl violet and coverslipped using Permount. Some tissue sections from rats killed at 18 h after a seizure-producing HL were processed for in situ hybridization with b 1 integrin 35 S-cRNA combined with the immunohistochemical localization of glial fibrillary acidic protein ŽGFAP. as described in Ref. w28x. Briefly, tissue sections were processed for isotopic in situ hybridization and rinsed through SSC Žas above., incubated overnight at 48C with polyclonal rabbit anti-GFAP ŽDAKO. at a dilution of 1:1000 in PB, and processed for localization of the antibody using the ABC system of Vector Laboratories and diaminobenzidine ŽDAB. as chromagen. After the DAB reaction, the tissue was rinsed through several changes of PB, mounted onto microscope slides and processed for emulsion autoradiography as described above. For better visualization of c-RNA labeled perikarya, some tissue sections from control rat brains were processed for colormetric in situ hybridization using a digoxigenin labeled cRNA probe and published techniques w10x. For this procedure, integrin b 1 cRNA was transcribed in the presence of digoxigenin-coupled UTP, and then hybridized and washed essentially as described above for isotopic in situ hybridization. The digoxigenin was localized by alkaline phosphatase immunohistochemistry w10,35x. Photomicrographs showing b 1 cRNArGFAP-ir double labeling were taken using a Sony Catseye digital camera and contrast-enhanced to better visualize autoradiographic grains. Fig. 3 shows digitized images of film autoradiograms captured using a Sony CCD video camera module and NIH Image software. All other photomicrographs are of emulsion autoradiograms and were taken using a Wild macroscope with Kodak Plus-X film. 2.3. Quantification of 35S-cRNA hybridization Quantification of 35 S-cRNA hybridization was performed by densitometric analysis of the film autoradiograms using an MCID imaging system ŽImaging Res., St. Catherines, Ont... Measures of labeling density were calibrated by interpolation from a standard curve derived from images of 14 C-labeled brain paste standards that were exposed to each sheet of film Žsee Ref. w23x for details on standards and quantification techniques.. Measurements

267

were taken from CA3b stratum pyramidale and the middle third of the dentate gyrus molecular layer Žexternal leaf. of the rostral hippocampal formation and from layer I and layers IIrIII of entorhinal cortex. At least five sections through each brain region were measured from each animal. All statistical analyses were conducted using the raw labeling density measures for individual rats from the HL-seizure and anesthetic-control groups Ži.e., experimental values were not ‘normalized’ or paired to control rat values.. The significance of the effect of treatment was evaluated using the one-way analysis of variance ŽANOVA. and, for individual comparisons, the Tukey–Kramer posthoc test. Differences had to reach the 95% confidence level Ž p F 0.05. to be considered statistically significant; unless otherwise stated, p values indicated are for the comparison to anesthetic-control values. 2.4. Northern blot analysis RNA was isolated from fresh brains following decapitation. The hippocampal formation, caudal cortex Žneocortex and entorhinalrpiriform cortex. and thalamus were dissected free and immediately frozen on dry ice; for each region, samples from right and left sides of the brain were pooled. Frozen tissue was homogenized in Tri–Zol reagent ŽGibco BRL. using the tight pestle of a Wheaton homogenizer. RNA was extracted through organic phase separation of phenolrchloroform, precipitated in isopropanol, and then quantified by UV spectrophotometry. Samples were denatured by adding 50% formamide, 2.2 M formaldehyde, 1 = gel buffer Ž20 mM HEPES pH 7.4, 1 mM ethylene diamine tetra-acetate wEDTAx., 10 m g ethidium bromide and heating to 55–608C for 15 min followed by the addition of 1 = sample buffer Ž10% glycerol, 0.2 mM EDTA pH 8.0, 0.05% bromophenol blue.. RNA samples Ž12 m g. were fractionated on a 1.2% agarose gel containing 2.3 M formaldehyde at 100 V for 3–4 h, and then transferred to a Zeta-probe ŽBIO-RAD. nylon membrane by capillary action in 10 = SSC. Following transfer to the nylon membrane, the RNA was cross-linked using ultraviolet radiation ŽStratalinker, Stratagene. and the membrane was incubated in hybridization buffer Ž5 = SSC, 50% de-ionized formamide, 50 mM Na 2 HPO4 , 5 = Denhardt’s, 0.1% SDS, 5% dextran sulfate and 200 m grml polyribonucleotide A. at 708C for 2–3 h. The 32 P-b 1 integrin cRNA probe Žsee below. was then added to the buffer and incubated overnight at 708C. The membrane was washed thoroughly with a final rinse of 0.1 = SSC, 0.1% sodium dodecyl sulfate at 708C and exposed to Amersham b-Max autoradiographic film with an intensifying screen ŽFisher. for 96 h at y808C. 2.5. b 1 integrin probes The plasmid containing rat b 1 integrin cDNA was the kind gift of Dr. Leonard Rome w37x and contains a 2.8-kb

268

J.K. Pinkstaff et al.r Molecular Brain Research 55 (1998) 265–276

cDNA inserted into the multiple cloning site of Bluescript KSŽy.. For in situ hybridization, the antisense RNA was transcribed from the Xho I digested plasmid with T3 RNA polymerase and the sense RNA was transcribed from the Xba I digested plasmid with T7 RNA polymerase, both in the presence of 35 S-UTP. For Northern hybridizations, a 300-bp fragment was excised from the plasmid by Rsa I and the antisense probe was transcribed using T3 RNA polymerase in the presence of 32 P-CTP.

3. Results 3.1. Normal distribution of hybridization The distribution of antisense riboprobe hybridization was evaluated from coronal tissue sections spanning the rostrocaudal distance from the mid-septum through lower brainstem. As shown in Fig. 1, the b 1 integrin cRNA sparsely labeled most regions of neuropil and densely labeled pia mater, the ventricular epithelium, the choroid plexus, and a few discrete neuronal groups. Neurons within the magnocellular hypothalamic nuclei, including the para and periventricular nuclei, the supra optic nuclei ŽFig. 1B., and nucleus circularis, were particularly densely labeled. Within the midbrain, there were well labeled cells in the red nucleus, the oculomotor nuclei ŽFig. 1C., and the mesencephalic reticular nucleus. Moderate densities of hybridization labeled the substantia nigra pars compacta and ventral tegmental area, the median eminence and the infundibular stem Žnot shown.. In some cases, low density labeling was evident in hippocampal stratum pyramidale as well. At more caudal levels, there was dense labeling of the Purkinje cells of cerebellum, cells in the brainstem efferent cranial nerve nuclei including the trochlear, motor trigeminal, facial motor, salivatory and hypoglossal nuclei, the dorsal motor nucleus of the vagus and nucleus ambiguus as well as in the mesencephalic trigeminal nucleus. To better visualize the b 1 mRNA containing cells in some of these fields, digoxigenin-labeled cRNA was hybridized to sections through control midbrain and lower brainstem and processed for the colormetric localization of hybridization using alkaline phosphatase histochemistry. With this colormetric marker the labeled cells within the cranial nerve nuclei appeared large and neuron-like. Fig. 2 shows alkaline phosphatase labeling in the facial and hypoglossal motor nuclei. At high magnification one can see the large neurons of the hypoglossal nucleus are b 1 integrin mRNA positive ŽFig. 2B.. Greater than background levels of labeling Žas determined by ‘sense’ probe hybridization. were associated with the neuropil through most gray matter regions of forebrain and brainstem. This was conspicuous within the hippocampal formation where cRNA labeling appeared to be somewhat laminated and generally at a higher density

Fig. 1. Photomicrographs showing the autoradiographic localization of in situ hybridization to b 1 integrin mRNA in adult rat brain. ŽA. Coronal section through rostral hippocampus showing generally low levels of hybridization with somewhat greater grain densities overlying the dentate gyrus moleculare layer Žml.; arrow indicates the hippocampal fissure Žh, hilus.. ŽB. Within hypothalamus, the paraventricular ŽPVN. and supraoptic ŽSO. nuclei are densely labeled. ŽC. In brainstem, moderate grain densities label neurons of the red nucleus ŽRMC. and the oculomotor nuclei ŽONC.. Arrowheads indicate dense labeling of the ventricular epithelium in ŽA. and a blood vessel in ŽB.. Calibration bar s 400 m m.

than in the surrounding white matter Že.g., corpus callosum.. As shown in Fig. 1A, relatively greater autoradiographic grain densities were associated with stratum lacunosumrmoleculare, the dentate gyrus molecular layer and the subgranular hilus. Within the molecular layer, loose clusters of autoradiographic grains were distributed over cells that appeared to be Nissl-stained astroglia Ži.e., round or oval nuclei, of intermediate size and staining density, lacking nucleoli..

J.K. Pinkstaff et al.r Molecular Brain Research 55 (1998) 265–276

Fig. 2. Integrin b 1 mRNA is expressed by efferent cranial nerve nuclei. Photomicrographs show the localization of integrin b 1 mRNA using colormetric Ždigoxigenin-cRNA. in situ hybridization. Panels ŽA. and ŽB. show labeled cells in the hypoglossal ŽXII. nucleus at low ŽA. and higher ŽB. magnification. Panel C shows labeling in the facial ŽVII. motor nucleus Calibration bar s 350 m m for ŽA. and ŽC.; 50 m m for ŽB..

269

Fig. 3. Digitized images of film autoradiograms showing the density and distribution of autoradiographic labeling in semi-adjacent sections from an anesthetic-control rat processed for in situ hybridization with 35 Slabeled b 1 antisense ŽA. and sense ŽB. riboprobes. ŽA. The antisense sequence diffusely labels the neuropil with somewhat higher density labeling of hippocampal stratum pyramidale Žsp. and dense labeling of the pia mater Žarrow in A.. ŽB. The sense sequence does not label the brain tissue but there is modest labeling associated with the pia mater Žarrow in B.. Calibration bar s1 mm.

Finally, in both control and experimental tissue processed with the sense riboprobe, there was no labeling within the brain tissue ŽFig. 3.. However, the pia mater was moderately well labeled with the sense transcript raising the possibility that the antisense labeling of this meningeal layer was non-specific. 3.2. Effect of seizure actiÕity Hilus lesion ŽHL.-induced seizures caused broadly distributed, bilateral increases in b 1 integrin mRNA content. Northern blot analysis detected low b 1 mRNA levels in control hippocampal formation, cortex and thalamus. As shown in Fig. 4, the cRNA probe detected a single mRNA species of about 3.8 kb. Within hippocampus, hybridization densities were increased at 8 h postlesion and continued to increase through 24 h postlesion. Hybridization was somewhat lower at 48 h and 4 days post-HL but was still greater than in the sample from the paired control rat. The effects of seizures were further evaluated using 35 S-cRNA in situ hybridization to tissue sections spanning the rostrocaudal distance from the hippocampal commis-

Fig. 4. RNA blot analysis showing relative b 1 integrin mRNA levels in areas of control and experimental seizure Ži.e., hilus lesion. rat brain. For each sample, 12 m g of total RNA was size-separated on an agarose gel and hybridized with a 32 P-labeled integrin b 1 cRNA probe. Samples from thalamus ŽThl., hippocampus ŽHip. and cortex ŽCtx. of a control Žcon. rat show low levels of hybridization in a single band at 3.8 kb. The b 1 mRNA was more abundant in samples from HL-seizure hippocampus at 8 h through 96 h postlesion, with peak levels occurring at 24 h. Levels of hybridization declined by 48 h, but remained greater than control densities through 96 h postlesion. Migration of 28S rRNA is indicated on the left side of blot.

270

J.K. Pinkstaff et al.r Molecular Brain Research 55 (1998) 265–276

sure to the caudal limit of hippocampus. Labeling was markedly and reliably increased within hippocampal formation, cortex, and amygdala of experimental-seizure as compared to control rats. At both 4 and 8 h after the HL, labeling was clearly increased within hippocampal stratum pyramidale ŽFig. 5C and Fig. 6B. and the dentate gyrus hilus ŽFig. 5C.. In two of four rats killed at 4 h post-HL, hybridization was also increased in layers II and III of the piriform and entorhinal cortices and in the posterior corti-

comedial amygdala and, in these fields, conformed to the neuronal cytoarchitectonics Žnot shown.. Increases were bilateral although the labeling within the dentate gyrus hilus and entorhinal cortex appeared greater ipsilateral to the lesion. In distinction from the earlier time point, by 8 h post-HL hybridization was also elevated along the hilar side of the dentate gyrus granule cell layer ŽFig. 5C, arrow., within the more rostral aspects of the subiculum, and diffusely within the tuberal region of hypothalamus.

Fig. 5. Seizures increase integrin b 1 mRNA in hippocampal neurons and glia. ŽA. Photomicrograph of a Nissl-stained section through the hippocampus showing the cytoarchitectonic field pictured in all panels. ŽB–F. Darkfield photomicrographs showing the distribution of b 1 integrin cRNA labeling in tissue from a control rat ŽB. and rats killed 8 ŽC., 12 ŽD., 18 ŽE., and 24 ŽF. h after a contralateral hilus lesion. In control hippocampal formation ŽB., modest grain densities are distributed over the dentate gyrus molecular layer Žml.. At 8 h postlesion, labeling is increased within CA3 and CA1 stratum pyramidale, the dentate gyrus molecular layer, and the subgranular hilus Žarrow in C.. At 12 h ŽD. there is a transition from the predominantly neuronal labeling of early postlesion time points, to what appears to be predominantly glial expression in the molecular layers. The latter is prominent at 18 ŽE. and 24 ŽF. h postlesion with peak levels occurring at 24 h. Abbreviations: h, hilus; sg, stratum granulosum. Arrows in ŽE. and ŽF. indicate the hippocampal fissure. Calibration bar s 350 m m.

J.K. Pinkstaff et al.r Molecular Brain Research 55 (1998) 265–276

271

Fig. 6. Low magnification darkfield photomicrographs showing the distribution of b 1 integrin cRNA labeling in tissue from a control rat and rats killed after HL- and kainic acid-induced seizures. ŽA. In control tissue there is low density, diffuse labeling in cortex and thalamus in addition to hippocampal Žhip. areas previously illustrated. Sections from experimental-seizure animals were taken from rats killed 4 h ŽB. and 18 h ŽC. after a contralateral HL and 24 h after kainic acid ŽD.. At 4 h post-HL, labeling is elevated in hippocampal CA3 stratum pyramidale Ž‘CA3’ in B. but not in neocortical fields. By 18 h post-HL ŽC., labeling is dramatically elevated in the molecular layers of hippocampus and in neocortex Žopen arrow.. Kainic acid treatment induced somewhat lesser increases in hybridization within hippocampus and neocortex but did elevate expression in thalamus Žarrow.. Calibration bar s 1 mm.

Moreover, labeling was further increased in olfactory cortex and within these regions no longer appeared to be laminar but spanned the largely neuron free layer I and in deeper layers containing both neurons and glia. In the deeper layers of piriform cortex, autoradiographic grains were clustered over both small Nissl-dense cells and some larger Nissl-pale cells giving the impression that hybridization was associated with both glial cells and neurons. From 12 to 24 h after the HL, there was an increase in labeling within the molecular layers of the hippocampal formation ŽFig. 5D–Fig. 6C. as well as across the depth of the neocortex ŽFig. 6C, open arrow.. Hybridization was further elevated in the entorhinal and piriform cortices and posterior amygdala and within these areas, as well as in the neocortex, extended into layer I. When examined at high magnification, much of the hybridization at 18 and 24 h post-HL appeared to be associated with astroglial cells. To specifically evaluate this possibility, tissue sections from rats killed at 18 h post-HL were processed for 35 S-cRNA labeling combined with the immunohistochemical localization of the astroglial protein GFAP. As shown in Fig. 7, many GFAP-ir cells were also b 1 mRNA-positive in the hippocampal molecular layers ŽFig. 7A–C. and in regions of white matter near the damage created by lesion placement ŽFig. 7D.. At this time point and in these fields most

clusters of autoradiographic grains appeared to be associated with GFAP-ir cells but there were additional GFAP-ir cells that were not autoradiographically labeled. Densitometric measures of b 1 35 S-cRNA labeling in HL-seizure rats are presented in Fig. 8, and are consistent with the above qualitative observations. As shown, hybridization was initially most clearly increased in neuronal cell layers Že.g., hippocampal CA3 stratum pyramidale and entorhinal cortex layer II by 8 h. but by 18–24 h postlesion was well elevated in these fields as well as within relatively neuron-poor molecular layers Že.g., dentate gyrus molecular layer and layer I of entorhinal cortex.. The seizure-producing HL had a significant effect on labeling densities in each of these fields Ž p - 0.0001, ANOVA. and labeling was significantly greater than in control rats for all fields measured from 18 through 48 h postlesion. Several observations indicate that, excepting the field immediately surrounding lesion placement, increases in b 1 mRNA content in HL rats are attributable to the effects of seizures as opposed to the damage of lesion placement. First, after HL-induced seizures increases in hybridization were bilateral and, within the contralateral hemisphere, did not correspond with the topography of deafferentation expected from the hilus lesion Ži.e., within the inner molecular layer of the dentate gyrus. w1x. Second, rats with the

272

J.K. Pinkstaff et al.r Molecular Brain Research 55 (1998) 265–276

increased in dorsal thalamus ŽFig. 6D, arrow.. These differences appear attributable to differences in the topography of seizure activity in these two paradigms as revealed by other activity dependent changes in gene expression. Most notably, in distinction from effects of HL-seizures w19,20,22x, kainic acid-seizures increase c-fos expression in thalamus w51x and have less influence on nerve growth factor mRNA content in the neocortex w19,22x. In tissue from HL-seizure rats, hybridization was more greatly and persistently increased in the region surrounding the lesion than on the contralateral side ŽFig. 9.. The greater effect surrounding the lesion was evident beginning at 8 h and continued through 4 days postlesion. Moreover, on the lesioned side only there was a second increase in the density of cRNA labeling within hippocampal CA3 stratum pyramidale by 48 h post-HL ŽFig. 8.. This second and larger increase in hybridization occurred as the density

Fig. 7. b 1 integrin mRNA is expressed by astroglial cells. Photomicrographs showing the autoradiographic localization of b 1 35 S-cRNA labeling Žblack grains. of GFAP-immunoreactive astroglial cells Žarrows. in tissue from a rat killed 18 h post-HL. Panels show cells in ŽA. contralateral stratum radiatum, ŽB. ipsilateral stratum oriens, ŽC. ipsilateral stratum radiatum and ŽD. ipsilateral corpus callosum near the damage created by the lesion wire. Cells shown in ŽB. and ŽC. are over 1 mm away from the lesion. Calibrations 20 m m.

electrode lowered but no current passed or with an electrolytic lesion of the hilus with platinum–iridium wire Žboth 24 h survival. had no behavioral seizure activity and no contralateral increases in hybridization. Ipsilateral to a platinum–iridium wire lesion, labeling was increased within both the hippocampal molecular layers and stratum pyramidale surrounding the field of damage. Finally, in one rat killed 24 h after HL surgery, the stainless steel wire lesion was placed slightly off the intended coordinates; this rat exhibited no seizure behavior and, by in situ hybridization, bilateral increases in integrin b 1 mRNA were not detected. To further verify that seizures, as opposed to other consequences of HL placement, increase b 1 integrin mRNA content, the effects of a convulsant systemic dose of kainic acid were evaluated. As shown in a representative rat in Fig. 6D, kainic acid induced an increase in b 1 cRNA hybridization although the pattern of labeling was somewhat different than seen after HL seizures. In the kainic acid treated rats, hybridization was elevated within both the pyramidal cell and molecular layers of the hippocampal formation and across the depth of neocortex. However, in contrast to HL rats, in kainic acid rats labeling was more modestly increased in neocortex and was

Fig. 8. Quantification of hybridization to b 1 integrin mRNA in tissue from control and HL-seizure rats. Densitometric measures of film autoradiograms were collected from the ipsilateral Žipsi.. and contralateral Žcontra.. CA3 pyramidal cell layer, the contralateral dentate gyrus molecular layer ŽDG-ML. and entorhinal cortical layers I and II ŽEC-I and EC-II, respectively; ns 3 per time point; means"S.E.M. shown.. The HL significantly increased labeling densities in all fields measured Ž p0.0001 for the main effect of treatment in all fields, One-way ANOVA; ) p- 0.01 and )) p- 0.001, Tukey–Kramer post-hoc test for comparison to control values..

J.K. Pinkstaff et al.r Molecular Brain Research 55 (1998) 265–276

273

Fig. 9. Damage increases b 1 integrin mRNA content. Photomicrographs showing b 1 cRNA labeling in the hippocampal formation contralateral ŽA. and ipsilateral ŽB. to a hilus lesion in a rat killed 48 h after surgery. On the contralateral side ŽA., labeling is diffusely distributed over most lamina at somewhat greater densities than seen in control tissue Žrefer to Fig. 8 for quantification.. Within the ipsilateral hippocampus, labeling is much more dense near the lesion site Žcentered caudal to the plane of section shown here.: discrete clusters of autoradiographic grains overlie cells distributed vertically across the several lamina traversed by the electrode tract. The open arrow in B indicates the frayed edge of the lesion visible within the external blade of the dentate molecular layer. Abbreviations: h, hilus; ml, dentate gyrus molecular layer. Calibration bar s 250 m m.

of labeling within the molecular layers had begun to decline and extended well outside the field of elevated hybridization immediately surrounding the lesion.

4. Discussion 4.1. Normal distribution of b 1 integrin expression The present results demonstrate that in untreated adult rat brain, integrin b 1 mRNA is expressed at low levels across most regions and at high levels by discrete neuronal groups. The low density labeling of the neuropil appears attributable to expression by astroglial cells for several reasons. This labeling was greater in areas of gray matter as opposed to white matter, did not correspond with neuronal cytoarchitectonics, and was most dense in regions predominantly populated by glial cells Že.g., dentate gyrus stratum moleculare.. In the latter region, clusters of autoradiographic grains labeled small cells of intermediate Nissl-staining density as is typical of astroglia in the in situ hybridization preparations w23,29x. In addition to this general neuropil labeling, neuronal groups were labeled with variable density including low density labeling of hippocampal stratum pyramidale and substantia nigra and very dense labeling of the magnocellular hypothalamic and efferent cranial nerve nuclei. As mentioned in Section 1, previous reports appeared contradictory in concluding that b 1-ir was most prominently localized to glia, in an analysis of hippocampus w25x, and to neurons, in an analysis of cerebellum w44x. Our in situ hybridization results are in agreement with findings in both of these reports and with the conclusion that integrin b 1 mRNA is expressed by both neurons and glial cells but with the higher levels of

neuronal expression being strikingly region- and cellspecific. The particular neuronal populations normally expressing b 1 mRNA at high levels are interesting because they include regions with ongoing structural synaptic plasticity andror a notable capacity for morphological plasticity and thereby suggest a possible association between b 1 expression and these processes. For example, within the magnocellular hypothalamic nuclei there are increases in direct neuronal membrane appositions Žsoma-somatic and dendro-dendritic., electrotonic coupling, and multiple synapses formed by individual axons in response to dehydration, lactation, stress and afferent stimulation w31,32,38x. These changes are rapid and reversible w9,39x and appear to reflect the normal ongoing morphological plasticity of these neurons as they modulate their functional output in response to changing stimulus conditions w30x. The presence of particularly high b 1 mRNA in the magnocellular hypothalamic neurons raises the possibility that within these cells there is a rapid turnover of b 1 protein as it is continually broken down and replaced with the loss and reformation of junctional contacts. It is noteworthy that these nuclei also exhibit high levels of the polysialylated form of NCAM which has been associated with regions and periods of morphological plasticity w63x. At more caudal levels of the neuraxis, b 1 mRNA was abundant in the efferent cranial nerve nuclei, the red nucleus and the Purkinje cells of the cerebellum. The neurons of the efferent cranial nerve nuclei are capable of axonal regeneration and, like spinal motor neurons, are likely to exhibit structural changes in response to dynamic trophic influences from the periphery w18x. Finally, within the red nucleus extensive axonal remodeling has been described following partial deafferentation w42x and for both red nucleus neurons w43x and Purkinje cells w24x changes in innervation

274

J.K. Pinkstaff et al.r Molecular Brain Research 55 (1998) 265–276

patterns have been observed in association with motor learning. The integrin receptors function as dimers containing one b and one a subunit. The b 1 subunit can dimerize with a number of a subunits w34x but we do not know as yet which dimers are present in different regions of brain. Indeed, among the a subunits only the brain distribution of a 8 has been evaluated. As described by Einheber et al. w15x, a 8-ir is densely distributed within the dentate gyrus hilus and hippocampal field CA3 with lesser staining evident in field CA1. In addition, a 8-ir is prominent in pyramidal layers of neocortex, the mitral, tufted and periglomerular cells of olfactory bulb, the substantia nigra and superior olive. Immunoreactivity was low to absent in the diencephalon and the cranial nerve nuclei of adult rat brain. In comparing this distribution to that of b 1 mRNA, we find overlap within the hilus of the dentate gyrus and hippocampus proper Ži.e., strata pyramidale and oriens., the red nucleus and the substantia nigrarventral tegmental area and notable mismatches in the cranial nerve nuclei and magnocellular hypothalamus Žwhich express b 1 but not a 8. and in the superior olive and pyramidal layers of neocortex Žwhich contain a 8-ir but not b 1 mRNA.. The a 8 b 1 dimer is known to act as a receptor for fibronectin, vitronectin, and tenasin C and to mediate adhesion and process outgrowth within in vitro assay systems w11,41,59,64x. The distribution data indicate that this receptor is likely to be present within a subset of b 1-expressing regions, including hippocampal field CA3, but that b 1 integrin forms different dimer pairs in other regions including the hypothalamus and brainstem cranial nerve nuclei. 4.2. Seizures stimulate b 1 expression in neurons and glia The effects of seizures on b 1 mRNA levels demonstrate that neuron and astroglial cells that express quite low levels of b 1 mRNA in untreated rats, can dramatically increase the level of expression in response to intense neuronal activity. The distribution and time course of seizure effects on b 1 mRNA, as well as the positive but dissociable effects of activity and damage on b 1 mRNA content, are reminiscent of the effects of seizures on basic fibroblast growth factor mRNA content in the same experimental paradigm w23x. In both cases, mRNA levels increased first in neurons, peaked with predominant astroglial expression at 24 h, and declined in contralateral fields of seizure activation more quickly than in ipsilateral fields of direct damage. The observations that bilateral increases in b 1 mRNA are also elicited by systemic kainic acid administration, and not by non-convulsant platinum wire hilar lesions, reinforces the conclusion that these effects are induced by seizures as opposed to the damage of hilus lesion placement. Although the present results do not speak to the threshold for activity-induction of b 1 mRNA expression, they do demonstrate that expression is activated by intense neuronal activity and raise the possi-

bility that b 1 expression may be regulated across a wide range of activity levels. 4.3. Potential inÕolÕement of b 1 integrin in hippocampal long-term potentiation Previous electrophysiological studies of acute hippocampal slices demonstrated that application of synthetic peptides that block the binding domain of the RGD integrins, prior to or some minutes after stimulation, irreversibly blocks the maintenance of LTP w4,65x. This suggests that either the adhesive or the signal transduction properties of the RGD integrins are critical for processes of LTP stabilization. What do the present results add to this story? Most importantly, these results indicate that b 1 is expressed by both neurons and glial cells in the hippocampal formation. This suggests several roles the integrins could play in association with LTP. Integrin receptors can mediate both cell–cell and cell–extracellular matrix adhesion w34,57x. The present evidence that b 1 is expressed by hippocampal pyramidal cells, suggests that b 1-containing receptors integral to the pyramidal cell dendritic membrane may stabilize that membrane andror local cellular morphology by binding to the extracellular matrix or to cell adhesion molecules expressed by adjacent neurons or glia w34x. Alternatively, pyramidal cell integrins could be activated and bound by ligands in an activity dependent fashion leading to a variety of transduction events that may contribute to LTP including increases in intracellular calcium, protein phosphorylation, and gene expression w61,66x. There is evidence that the integrins are activated by concomitants of seizures and subseizure neuronal activity. For example, intense neuronal activity stimulates the efflux of platelet activating factor ŽPAF. w7x which is well documented to activate integrin receptors in other systems w45x. Moreover, both seizure and LTP-inducing activity stimulates the synthesis w52x and efflux w27x of tissue plasminogen activating factor ŽtPA. and other serine proteases w26,46,58x that can cleave extracellular matrix proteins w6,53x and generate cleavage products that act as ligands for the integrin receptors w49x. Interestingly, like peptide inactivation of RGD-integrin ligand binding, disruption of PAF w3,8x and tPA w33x activities has been reported to disrupt the late or maintenance phase of LTP.

5. Conclusion As described here, seizure-induced increases in b 1 mRNA content were just barely evident at 4 h postlesion Žor, 2 h after seizure onset. and grew for at least 12 h thereafter. With this time course, activity-induced increases in integrin expression are not likely to play a role in the early phases of LTP. Indeed, within hippocampal field CA1 the induction and initial stabilization of LTP

J.K. Pinkstaff et al.r Molecular Brain Research 55 (1998) 265–276

occurs within 30 min of stimulation w2,62x and is not disrupted by protein synthesis inhibition w5x. However, late LTP Ži.e., that persisting over 3–4 h after stimulation. is reportedly vulnerable to protein and RNA synthesis inhibition w5,17x. Activity-induced changes in integrin gene expression could play a role in these longer latency effects.

Acknowledgements Supported by grants NS26748 from NINCDS and BNS 9024143 from NSF to C.M.G. and grant F49620-95-J-0301 from the AFOSR to G.L. The authors thank Dr. Julie Lauterborn for comments on the manuscript.

References w1x D.G. Amaral, M.P. Witter, Hippocampal formation, in: G. Paxinos ŽEd.., The Rat Nervous System, Academic Press, San Diego, 1995, pp. 443–486. w2x A. Arai, J. Larson, G. Lynch, Anoxia reveals a vulnerable period in the development of long-term potentiation, Brain Res. 511 Ž1990. 353–357. w3x A. Arai, G. Lynch, Antagonists of the platelet-activating factor receptor block long-term potentiation in hippocampal slices, Eur. J. Neurosci. 4 Ž1992. 411–419. w4x B. Bahr, G. Lynch, Purification of an Arg–Gly–Asp selective matrix receptor from brain synaptic plasma membranes, Biochem. J. 281 Ž1992. 137–142. w5x C.H. Bailey, D. Bartsch, E.R. Kandel, Toward a molecular definition of long-term memory storage, Proc. Natl. Acad. Sci. U.S.A. 93 Ž1996. 13445–13452. w6x C. Basbaum, Z. Werb, Focalized proteolysis: spatial and temporal regulation of extracellular matrix degradation at the cell surface, Curr. Opin. Cell Biol. 8 Ž1996. 731–738. w7x N.G. Bazan, G. Allan, Platelet-activating factor in the modulation of excitatory amino acid neurotransmitter release and of gene expression, J. Lipid Mediators Cell Signal. 14 Ž1996. 321–330. w8x N.G. Bazan, M.G. Packard, L. Teather, G. Allan, Bioactive lipids in excitatory neurotransmission and neuronal plasticity, Neurochem. Int. 30 Ž1997. 225–231. w9x G.H. Beagley, G.I. Hatton, Rapid morphological changes in supraoptic nucleus and posterior pituitary induced by a single hypertonic saline injection, Brain Res. Bull. 28 Ž1992. 613–618. w10x J.L. Bizon, J.C. Lauterborn, P.J. Isackson, C.M. Gall, Acidic fibroblast growth factor mRNA is expressed by basal forebrain and striatal cholinergic neurons, J. Comp. Neurol. 366 Ž1995. 379–389. w11x B. Bossy, E. Bossy-Wetzel, L.F. Reichardt, Characterization of the integrin a 8 subunit: a new integrin b 1-associated subunit, which is prominently expressed on axons and on cells in contact with basal laminae in chick embryos, EMBO J. 10 Ž1991. 2375–2385. w12x K.A. Campbell, B. Bank, N.W. Milgram, Epileptogenic effects of electrolytic lesions in the hippocampus: role of iron deposition, Exp. Neurol. 86 Ž1984. 506–514. w13x H. Cremer, R. Lange, A. Christoph, M. Plomann, G. Vopper, J. Roes, R. Brown, S. Baldwin, P. Kraemer, S. Scheff, Inactivation of the NCAM gene in mice results in size reduction of the olfactory bulb and deficits in spatial learning, Nature 367 Ž1994. 455–459. w14x E. Doyle, P. Nolan, R. Bell, C. Regan, ICV infusion of anti-neural cell adhesion molecules in a discrete post-training period impair consolidation of passive avoidance response in rats, J. Neurochem. 59 Ž1992. 1570–1573.

275

w15x S. Einheber, L. Scnapp, J. Salzer, Z. Cappiello, T. Milner, The regional and ultrastructural distribution of the a 8 integrin subunit in the developing and adult rat brain suggests a role in synaptic plasticity, J. Comp. Neurol. 370 Ž1996. 105–134. w16x R.D. Fields, K. Itoh, Neural cell adhesion molecules in activity-dependent development and synaptic plasticity, Trends Neurosci. 19 Ž1996. 473–480. w17x U. Frey, S. Frey, F. Schollmeier, M. Krug, Influence of actinomycin D, a RNA synthesis inhibitor, on long-term potentiation in rat hippocampal neurons in vivo and in vitro, J. Physiol. 490 Ž1996. 703–711. w18x H. Funokoshi, N. Belluardo, E. Arenas, Y. Yamamoto, A. Casabona, H. Persson, C.F. Ibanez, ´˜ Muscle-derived neurotrophin-4 as an activity-dependent trophic signal for adult motor neurons, Science 268 Ž1995. 1495–1499. w19x C. Gall, Seizure-induced changes in neurotrophin expression: implications for epilepsy, Exp. Neurol. 124 Ž1993. 150–166. w20x C. Gall, A. Arai, J. White, Localization of increased c-fos mRNA content in rat CNS following recurrent limbic seizures, Soc. Neurosci. Abstr. 14 Ž1988. 1161. w21x C. Gall, J.C. Lauterborn, K.M. Guthrie, In situ hybridization: a sensitive measure of activity dependent changes in neuronal gene expression, in: W. Stumpf, H. Solomon ŽEds.., In VitrorIn Vivo Autoradiography and Correlative Imaging, Academic Press, San Diego, 1995, pp. 379–399. w22x C. Gall, K. Murray, P.J. Isackson, Kainic acid-induced seizures stimulate increased expression of nerve growth factor in adult rat hippocampus, Mol. Brain Res. 9 Ž1991. 113–123. w23x C.M. Gall, R. Berschauer, P.J. Isackson, Seizures increase basic fibroblast growth factor mRNA in adult rat forebrain neurons and glia, Mol. Brain Res. 21 Ž1994. 190–205. w24x W.T. Greenough, B.J. Anderson, Cerebellar synaptic plasticity. Relation to learning versus neural activity, Ann. New York Acad. Sci. 627 Ž1991. 231–247. w25x S. Grooms, L. Terracio, L. Jones, Anatomical localization of b 1 integrin-like immunoreactivity in rat brain, Exp. Neurol. 122 Ž1993. 253–259. w26x T.P. Gschwend, S.R. Krueger, S.V. Kozlov, D.P. Wolfer, Neurotrypsin, a novel multidomain serine protease expressed in the nervous system, Mol. Cell. Neurosci. 9 Ž1997. 207–219. w27x A. Gualandris, T.E. Jones, S. Strickland, S.E. Tsirka, Membrane depolarization induces calcium-dependent secretion of tissue plasminogen activator, J. Neurosci. 16 Ž1996. 2220–2225. w28x K. Guthrie, T. Nguyen, C. Gall, Insulin-like growth factor-1 mRNA is increased in deafferented hippocampus–spatiotemporal correspondance of a trophic event with axon sprouting, J. Comp. Neurol. 352 Ž1995. 147–160. w29x K.M. Guthrie, A. Woods, T. Nguyen, C.M. Gall, CNTF mRNA expression in rat brain: normal distribution and increased astroglial expression of axonal and terminal degeneration, J. Comp. Neurol., in press, 1997. w30x G. Hatton ŽEd.., Emerging Concepts of Structure-Function Dynamics in the Adult Brain: The Hypothalamo–Neurohypophysial System, Prog. Neurobiol, Pergamon, Oxford, 1990, pp. 437–504. w31x G.I. Hatton, Plasticity in the hypothalamic magnocellular neurosecretory system, Fed. Proc. 45 Ž1986. 2328–2333. w32x G.I. Hatton, Q.Z. Yang, K.G. Smithson, Synaptic inputs and electrical coupling among magnocellular neuroendocrine cells, Brain Res. Bull. 20 Ž1988. 751–755. w33x Y.Y. Huang, M.E. Bach, H.P. Lipp, M. Zhuo, D.P. Wolfer, R.D. Hawkins, L. Schoonjans, E.R. Kandel, J.M. Godfraind, R. Mulligan, Mice lacking the gene encoding tissue-type plasminogen activator show a selective interference with late-phase long-term potentiation in both Schaffer collateral and mossy fiber pathways, Proc. Natl. Acad. Sci. U.S.A. 93 Ž1996. 8699–8704. w34x R. Hynes, Integrins, versatility, modulation, and signaling in cell adhesion, Cell 69 Ž1992. 11–25.

276

J.K. Pinkstaff et al.r Molecular Brain Research 55 (1998) 265–276

w35x J.C. Lauterborn, T. Tran, P. Isackson, C.M. Gall, NGF mRNA is expressed by GABAergic neurons in rat hippocampus, NeuroReport 5 Ž1993. 273–276. w36x A. Luthl, J. Laurent, A. Figurov, D. Muller, M. Schachner, Hip¨ pocampal long-term potentiation and neural cell adhesion molecules L1 and NCAM, Nature 372 Ž1994. 777–779. w37x S. Malek-Hedayat, L. Rome, Cloning and sequence of the cDNA encoding the rat oligodendrocyte integrin beta 1 subunit, Gene 158 Ž1995. 287–290. w38x S. Miyata, T. Itoh, O. Matsushima, T. Nakashima, T. Kiyohara, Not only osmotic stress but also repeated restraint stress causes structural plasticity in the supraoptic nucleus of the rat hypothalamus, Brain Res. Bull. 33 Ž1994. 669–675. w39x S. Miyata, T. Nakashima, T. Kiyohara, Structural dynamics of neural plasticity in the supraoptic nucleus of the rat hypothalamus during dehydration and rehydration, Brain Res. Bull. 34 Ž1994. 169–175. w40x D. Muller, C. Wang, G. Skibo, N. Toni, H. Cremer, V. Calaora, G. Rougon, J.Z. Kiss, PSA-NCAM is required for activity-induced synaptic plasticity, Neuron 17 Ž1996. 413–422. w41x U. Muller, B. Bossy, K. Venstrom, L.F. Reichardt, Integrin a 8 b 1 ¨ promotes attachment, cell spreading, and neurite outgrowth in fibronectin, Mol. Biol. Cell 6 Ž1995. 433–448. w42x F. Murakami, H. Katsumaru, K. Saito, N. Tsukahara, A quantitative study of synaptic reorganization in red nucleus neurons after lesions of the nucleus interpositus of the cat: an electron microscopic study involving intracellular injection of horseradish peroxidase, Brain Res. 242 Ž1982. 41–53. w43x F. Murakami, Y. Oda, N. Tsukahara, Synaptic plasticity in the red nucleus and learning, Behav. Brain Res. 28 Ž1982. 175–179. w44x S. Murase, M. Hayashi, Expression pattern of integrin b 1 subunit in Purkinje cells of rat and cerebellar mutant mice, J. Comp. Neurol. 375 Ž1996. 225–237. w45x R. Newton, M. Thiel, N. Hogg, Signalling mechanisms and the activation of leukocyte interins, J. Leukocyte Biol. 61 Ž1997. 422– 426. w46x A. Okabe, Y. Momota, S. Yoshida, A. Hirata, J. Ito, H. Nishino, S. Shiosaka, Kindling induces neuropsin mRNA in the mouse brain, Brain Res. 728 Ž1996. 116–120. w47x P. Pesheva, R. Juliano, M. Schachner, Expression and localization of the fibronectin receptor in the mouse nervous system, J. Neurosci. Res. 20 Ž1988. 420–430. w48x R. Pico, C. Gall, Hippocampal epileptogenesis produced by electrolytic iron deposition in the rat dentate gyrus, Epilepsy Res. 19 Ž1994. 27–36. w49x M. Pierschbacher, E. Ruoslahti, Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule, Nature 304 Ž1984. 30–33. w50x J. Pinkstaff, J. Detterich, G. Lynch, C. Gall, Localization and seizure-regulation of integrin b 1 mRNA in the adult rat hippocampus, Soc. Neurosci. Abstr. 22 Ž1996. 1982.

w51x T. Popovici, A. Represa, V. Crepel, ´ G. Barbin, M. Beaudoin, Y. Ben-Ari, Effects of kainic acid-induced seizures and ischemia on c-fos-like proteins in rat brain, Brain Res. 536 Ž1990. 183–194. w52x A. Quan, M. Gilbert, M. Colicos, E. Kandel, D. Kuhl, Tissue-plasminogen activator is induced as an immediate–early gene during seizure, kindling and long-term potentiation, Nature 361 Ž1993. 453–457. w53x J. Quigley, L. Gold, R. Schwimmer, L. Sullivan, Limited cleavage of cellular fibronectin by plasminogen activator purified from transformed cells, Proc. Natl. Acad. Sci. U.S.A. 84 Ž1997. 2776–2780. w54x R. Racine, Modulation of seizure activity by electrical stimulation: II. Motor seizure, Electoencephalogr. Clin. Neurophys. 32 Ž1972. 281–294. w55x L. Reichardt, K. Tomaselli, Extracellular matrix molecules and their receptors: functions in neural development, Annu. Rev. Neurosci. 14 Ž1991. 531–570. w56x L. Ronn, E. Bock, D. Linnemann, H. Jahnsen, NCAM-antibodies modulate induction of long-term potentiation in rat hippocampal CA1, Brain Res. 677 Ž1995. 145. w57x M. Ruppert, S. Aigner, M. Hubbe, H. Yagita, P. Altevogt, The L1 adhesion molecule is a cellular ligand for VLA-5, J. Cell Biol. 131 Ž1995. 1881–1891. w58x I.A. Scarisbrick, M.D. Towner, P.J. Isackson, Nervous system specific expression of a novel serine protease: regulation in the adult rat spinal cord by excitotoxic injury, J. Neurosci. 17 Ž1997. 8156–8168. w59x L.M. Schnapp, J.M. Bruess, D.M. Ramos, D. Sheppard, R. Pytela, Sequence and tissue distribution of the human integrin a 8 subunit: a b 1 associated subunit expressed in smooth muscle cells, J. Cell Sci. 108 Ž1995. 537–544. w60x A. Scholey, S. Rose, M. Zamani, E. Bock, M. Schachner, A role for the Neural Cell Adhesion Molecule ŽNCAM. in the late, consolidating phase of glycoprotein synthesis 6 hours following passive avoidance training of the young chick, Neuroscience 55 Ž1993. 499–509. w61x M. Schwartz, M. Schaller, M. Ginsberg, Integrins: emerging paradigms of signal transduction, Annu. Rev. Cell Dev. Biol. 11 Ž1995. 549–599. w62x U. Staubli, D. Chun, Factors regulating the reversibility of long term potentiation, J. Neurosci. 16 Ž1996. 853–860. w63x D.T. Theodosis, G. Rougon, D.A. Poulain, Retention of embryonic features by an adult neuronal system capable of plasticity: polysialylated neural cell adhesion molecule in the hypothalamo–neurohypophysial system, Proc. Nat. Acad. Sci. U.S.A. 88 Ž1991. 5494–5498. w64x B. Varnum-Finney, K. Venstrom, U. Muller, R. Kypta, C. Backus, ¨ M. Chiquet, L.F. Reichardt, The integrin receptor a 8 b 1 mediates interactions of embryonic chick motor and sensory neurons with tenascin-C, Neuron 14 Ž1995. 1213–1222. w65x P. Xiao, B. Bahr, U. Staubli, P. Vanderklish, G. Lynch, Evidence that matrix recognition contributes to stabilization but not induction of LTP, NeuroReport 2 Ž1991. 461–464. w66x K. Yamada, S. Miyamoto, Integrin transmembrane signaling and cytoskeletal control, Curr. Opin. Cell Biol. 7 Ž1995. 681–689.