The Journal of Neuroscience, May 1993, 13(5): 1965-1975 a-Bungarotoxin Binding to Hippocampal Interneurons: lmmunocytochemical Characterization and Effects on Growth Factor Expression

Robert Freedman,‘,* Cynthia Wetmore,’ lngrid Striimberg,’ Sherry Leonard,* and Lars Olson’ ‘Department of Histology and Neurobiology, Karolinska Institute, Stockholm, Sweden S-10401 and *Departments of Psychiatry and Pharmacology, Denver Veterans Administration Medical Center and University of Colorado Health Sciences Center, Denver, Colorado 80262

The nicotinic cholinergic antagonist a-bungarotoxin (a-BT) memory (Lippa et al., 1980), modulation of seizure threshold binds throughout the rat hippocampal formation. The binding (Green et al., 1989; Marks et al., 1989), and habituation of is displaceable by dtubocurarine. The most heavily labeled responseto sensory stimuli (Vinogradova, 1975). The impor- cells are GABA-containing interneurons in the dentate and tance of this pathway is emphasized by its apparent trophic in Ammon’s horn. These neurons have several different mor- regulation by several growth factors, including NGF and brain- phologies and contain several neuropeptides. a-BT-labeled derived neurotrophic factor (BDNF), even in adult life (Korsch- interneurons in the dentate are small cells between the gran- ing et al., 1985; Phillips et al., 1990). Cholinergic neurotrans- ular and molecular layers that often contain neuropeptide Y. mission involves a diverse set of postsynaptic receptors and a-BT-labeled interneurons in CA1 are medium-sized inter- mechanisms.The G-protein-coupled muscarinic receptors are neurons, occasionally found in stratum pyramidale, but more closelyassociated with cholinergicgeneration of the theta rhythm often found in stratum radiatum and stratum lacunosum mo- (Vanderwolf, 1975)and short-term memory (Lippa et al., 1980). leculare. These neurons often contain cholecystokinin. The Thesefunctions are alsoprimarily associatedwith the CA 1 area. largest a-BT-labeled interneurons are found in CA3, in both Muscarinic receptors are more concentrated in this area and stratum radiatum and stratum lucidum. These neurons are relatively lessdense in CA3 and the dentate (Joyce et al., 1989). multipolar and frequently are autofluorescent. They often Seizuregenesis and habituation of responseto sensorystimuli contain somatostatin or cholecystokinin. These large inter- have been linked to nicotinic mechanisms(Marks et al., 1989; neurons have been found to receive medial septal inner- Luntz-Leybman et al., 1992). These functions are closely as- vation and may also have projections that provide inhibitory sociated with the CA3 field (Schwartzkroin, 1986; Bickford- feedback directly to the medial septal nucleus. The cholin- Wimer et al., 1990). Nicotinic responsesare mediated by two ergic innervation of the hippocampus from the medial septal major classesof receptors, ganglionic type and neuromuscular nucleus is under the trophic regulation of NGF and brain- type. Pharmacological analysisof both seizure genesisand ha- derived neurotrophic factor, even in adult life. Expression of bituation in rat brain implicates mediation by a neuromuscular mRNA for both these factors is increased in CA3 and the type (Miner and Collins, 1989; Luntz-Leybman et al., 1992). dentate after intraventricular administration of a-BT, but not These data are supported by the prominent binding of the neu- after administration of the muscarinic antagonist atropine. romuscular-type antagonist oc-bungarotoxin(a-BT) in the CA3 a-BT-sensitive cholinergic receptors on inhibitory interneu- field of the hippocampus (Hunt and Schmidt, 1978; Segalet al., rons may be critical to medial septal regulation of the hip- 1978; Clarke et al., 1985; Harfstrand et al., 1988; Pauly et al., pocampal activity, including the habituation of response to 1989). The ganglionic antagonist called neuronal or K-BT also sensory input. binds in CA3, but the binding is fully displacedby a-BT (Schulz [Key words: bungarotoxin, hippocampus, interneurons, et al., 1991). In most other brain areas, there is no such dis- NGF, brain-derived neurotrophic factor, cholecystokinin, so- placement. The binding data are supported by in situ hybrid- matostatin, neuropeptide Y, nicotinic receptors] ization for several of the a-subunits of the nicotinic receptor. a-subunits 2, 3, and 4 form receptorsthat are sensitiveto neu- Medial septal innervation of the hippocampus has been asso- ronal BT and other ganglionic antagonists(Boulter et al., 1987). ciated with a number of functions of the hippocampus,including cDNA probes for these subunits do not hybridize strongly in generation of the theta rhythm (Vanderwolf, 1975), short-term the hippocampus(Wada et al., 1989). However, a probe for the (r7 subunit, which produces an a-BT-sensitive receptor, hy- Received Apr. 2, 1992; revised Oct. 30, 1992; accepted Nov. 13, 1992. bridizes strongly in the hippocampus(Johnson et al., 1991). This research was suooorted bv the Medical Research Council of Sweden. the The identity of hippocampal neuronsthat bind LU-BTwould Veterans Administration Medical Research Service, and National Institute of seemto be critical for further understanding of the role of nic- Mental Health Awards MH-442 I2 and MH-3823 1. We thank Monica Nyman, Eva Lindquist, Dorothy Dill, and Susanne Almerstriim for technical assistance. otinic neurotransmissionin hippocampal function. Although Laura Lee Lamothe prepared the manuscript. or-BT binding in the hippocampus has already been demon- Correspondence should be addressed to Robert Freedman, M.D., Department of Psychiatry C-268-71, University of Colorado Health Sciences Center, 4200 strated by a number ofinvestigators, there have beenfew studies Fast Ninth Avenue, Denver, CO 80262. that have extended these findings to the cellular level. In one Copyright 0 1993 Society for Neuroscience 0270-6474/93/131965-l 1$05.00/O study at the ultrastructural level, a-BT was found to bind near 1966 Freedman et al. . a-Bungarotoxin Binding to Hippwampal Interneurons

determine the possibleoverlap between the presenceof (Y-BT- binding proteins and these neuropeptides. Becauseof the importance of the NGF family to the integrity of the cholinergic medial septal pathway, in situ hybridization was usedto determine the effectsof a-BT on mRNA expression of several growth factors in hippocampalneurons. Lesion of the fomix increasesthe expressionof NGF in hippocampalneurons (Ayer-LeLievre et al., 1988). However, treatment with the mus- carinic cholinergic antagonist scopolamine does not increase levels of NGF in the hippocampus(Alberch et al., 1991). In the present study, the effectsof ol-BT and atropine were compared on expressionof the messagefor NGF, BDNF, and hippocam- pal-derived neurotrophic factor or neurotrophin-3 (NT3).

Materials and Methods

I a-BT binding. cu-BT binding was performed in four animals by the protocol of Fuchs (1989). Male rats (150 gm) were obtained from Alab

I (Stockholm, Sweden). Animals were anesthetized with pentobarbital and the brain was rapidly removed and frozen on a CO, freezing stage. i i Fourteen-micrometer cryostat sections, thaw mounted on slides coated i i with poly-L-lysine, were incubated for 1 hr at room temperature in a

I solution containing 5 nM 12sI-o(-BT (specific activity, 2000 Ci/mmol; Amersham) in 1.25% bovine serum albumin, 0.05 M Tris HCI, pH 7.4. This concentration of (Y-BT has been found to saturate a single class of receptors (Clarke et al.. 1985). Preincubation for 30 min in lO-4 M d-tubocurarine was used to detect nonspecific binding. After incubation with a-BT, the slides were washed in ice-cold buffer for 15 min, followed by three 15 min washes in ice-cold distilled water. The slides were then b dehydrated in 70% and 95% ethanol, dried, and dipped in autoradio- graphic emulsion diluted 1: 1 in water (NTB-2, Kodak, Rochester, NY). After a 10 d exposure at -2o”C, the slides were developed in Kodak D- 19 diluted 1: 1 in water, fixed. counterstained with cresvl violet, and coverslipped. Four animals received (u-BT in viva These animals were anesthetized with pentobarbital and placed in a stereotaxic instrument. 1251-a-BT, 0.01 mCi, was injected into the lateral ventricle, and the animal was allowed to survive for 30 min. This time was chosen because it is the interval at which maximal effects of a-BT on electrophysiological re- sponses in hippocampus are observed (Luntz-Leybman et al., 1992). The animals were then perfused for immunocytochemistry, followed by autoradiography, as described below. Immunocytochemistry. Six animals were anesthetized with pento- barbital and then perfused through the aorta with calcium-free Tyrode’s solution, followed by 4Oh parafonnaldehyde with 0.4% picric acid in 0.1 M phosphate buffer, pH 7.0. The brain was blocked and immersed in the fixative for several hours and then rinsed overnight in 10% sucrose in 0.1 M phosphate buffer. Cryostat sections (14 pm) were collected on chrom-alum-treated slides. rinsed in phosphate-buffered saline (PBS). Figure I. Binding of 12*I-a-BT in rat brain. a shows a pattern of binding and then incubated overnight at 4°C in a humidified chamber with to frozen rat brain sections. b shows binding with the same concentration primary antibody diluted in PBS containing 0.3% T&on. Sections were of 12+a-BT, blocked in the presence of lO-4 M d-tubocurarine. c is a rinsed twice in PBS and then exposed for 1 hr at room temperature to cresyl violet-stained section. In this and subsequent figures, the follow- a species-specific fluorescein-labeled secondary antibody. The slides were ing abbreviations are used: stratum oriens (o), stratum pyramidale @), then coverslipped with 90% glycerin in PBS, with 0.1% p-phenylene stratum radiatum (r), stratum lacunosum moleculare ([m), stratum lu- diamine as an antifading agent. cidum (I), dentate granule cells (g), and hilus (h). Scale bar, 1 mm. Primary antibodies and dilutions used were anti-GABA (guinea pig, 1:200; Sequala et al., 1984), anti-CCK (rabbit, 1: 1000; Abbott, Chicago, IL), and anti-NPY (rabbit, 1:400; Peninsula, Belmont, CA). A mouse monoclonal antibody for somatostatin ( 1:400) was a gift from Prof. T. synapses of both pyramidal and nonpyramidal neurons (Hunt Hiikfelt. Control slides were reacted only with secondary antibody. and Schmidt, 1978). Both types of neurons receive cholinergic Immunoreactive neurons in the hippocampus were photographed un- innervation (Frotscher and Leranth, 1985). The aim of the pres- der epi-illumination. The coverslips were then removed and the slides ent study was to identify the nonpyramidal neuronsthat bind were exposed to 1Z51-a-BT as described above, except for slides from wBT. Autoradiographic localization of ‘*V-CU-BTand immu- animals that had previously received W-ol-BT intraventricularly. After autoradiographic exposure and counterstaining with cresyl violet, the nocytochemical detection of GABA were usedto identify a pop- slides were examined for correspondence between neurons labeled with ulation of interneurons that have cu-BT-sensitivereceptors. In- ol-BT and those that had been previously found to be immunoreactive. terneurons of the hippocampus contain a variety of In situ hybridization. Animals were anesthetized briefly with halo- neuropeptides, including cholecystokinin (CCK, Sloviter and thane and placed in a stereotaxic apparatus. Intraventricular injection of a-BT (1 pg, two animals, or 0.3 pg, three animals), or atropine sulfate Nilaver, 1987) neuropeptideY (NPY, Deller and Leranth, 1990), (1 pg, two animals) was performed. Both drugs were obtained from and somatostatin(Kohler and Chan-Palay, 1982). Simultaneous Sigma Chemical (St. Louis, MO). Assuming a ventricular volume of (u-BT binding and immunohistochemistry were also used to 100 ~1, final concentrations were 300 nM to 1 PM for CX-BT and 1.4 NM The Journal of Neuroscience, May 1993, 13(5) 1967

Figure 2. Binding of 1251-a-BT to a frozen section of rat hippocampus (a). In the CA1 region (b), there are small labeled neuronal cell bodies in the distal portion of stratum oriens, near the al- veus, as well as some labeled cells in stratum lacunosum moleculare. There are also occasional small cells (arrow) labeled within stratum pyramidale. In the CA3 region (c, e), there are larger cells (arrows) labeled in stratum oriens (c)and in stratum lucidum (e). Labeling in the dentate gyrus (d) is heaviest in small cells (arrow) located at the edges of the granular layer. Scale bar, 50 pm.

for atropine sulfate. Animals were matched with controls that received in Whittemore et al., 1988), BDNF (Wetmore et al., 1991), and NT3 injections of the same volume of Ringer’s solution (10 pl), as well as (nucleotides 667-7 17 in Emfors et al., 1990). The probes were selected with unoperated animals. A 3 hr survival interval was chosen on the to hybridize with unique sequences in each of the different rat growth basis of reports that nonspecific effects of the injection procedure itself factor sequences. A probe for rat NGF receptor with similar GC content on growth factor mRNA expression are no longer present by that time was used to control for nonspecific hybridization. The sections were (Ballarin et al., 1991). The animals were then anesthetized with pen- hybridized 15 hr at 42°C in a solution containing 50% formamide, 4 x tobarbital, and the brains were removed and frozen. Cryostat sections saline-sodium citrate (SSC), 1 x Denhardt’s solution, 10% dextran sul- (14 pm), thaw mounted on poly-L-lysine-coated slides, were used for fate, 0.5 mg/ml sheared salmon sperm DNA, 1% sarcosyl, 0.02 M NaPQ hybridization. (pH 7.0), 50 mM dithiothreitol, and 10’ cpm/ml of probe. AAer hy- In situ hybridization was performed using synthetic 50-mer oligo- bridization, the slides were rinsed in five changes of 1 x SSC at 55”c, nucleotide probes, 3’-end labeled with “S-dATP (New England Nuclear, dried, and dipped in emulsion. After 6 weeks exposure at -2O”c, the Boston, MA). Probes were constructed for NGF (nucleotides 868-918 slides were developed and counterstained as described above. Silver 1968 Freedman et al. * a-Bungarotoxin Binding to Hippocampal Interneurons

Figure 3. Comparison between neurons with GABA-immunocytochemical reactivity as detected by fluorescence histochemistry (left panels) and I*%x-BT labeling. For all panels but b, the fluorescence histochemistrv was netformed on nerfused and fixed brain sections. followed bv 1251-~-BT binding. For b, the Y-~-BT was injected intraventricularly, and then-the brain was perfused, fixed, and sectioned. The section was prepared first for fluorescence histochemistry and then for autoradiography. a and b show large neurons (arrowheads) that contain GABA and also bind a-BT, located in the stratum oriens of CA3. Similar CA3 neurons are shown in stratum oriens and stratum lucidum in d. c shows a series of smaller neurons on the inner layer of the distal portion of the dorsal blade of the dentate gyrus. e shows several small neurons in CAl, in striatum radiatum, and in stratum lacunosum moleculare. Scale bar: 50 pm for a-d, 100 pm for e. grain densities were quantified in 20 x dark-field images by a computer- gyrus. a-BT binding was blocked by d-tubocurarine (Fig. lb). based video image analyzer (MAGISCAN, Joyce-Loebl Ltd., Tyne and The binding was most densein the dentate and more diffuse in Wear, UK). For each hippocampal field, three 500 pm* regions were Ammon’s horn (Fig. 2). In both areas, however, the densest identified. A threshold was established that identified grains and rejected the counterstained cell bodies. The same threshold was used for all binding wasto the cell bodiesand principal dendritesof neurons analyses. MAGISCAN then computed the fraction of area in the region outside the granule cell and pyramidal cell layers. In the dentate occupied by grains. these neurons were most often found at the boundariesof the granule cell layer, as well as scatteredthrough the remainder of Results the hilus (Fig. 2d). Frequently, labeled neuronswere found on (r-BT bound extensively to sectionsof rat hippocampus(Figs. the interior edgeof the dorsal blade of the dentate gyrus, near 1, 2). A similar pattern of binding was observed in both fixed its lateral end. In CA 1, small cY-BT-labeledneurons are present and unfixed tissue,although the amount of binding wasgreater at the margin of the stratum oriens and the alveus (Fig. 2b). in unfixed tissue. In viva binding of a-BT after intraventricular There were also scattered neurons in stratum pyramidale and administration showed a similar pattern to that observed in stratum lacunosummoleculare. vitro in CA1 and CA3, but generally failed to reach the dentate The largestneurons that were labeledwith (u-BT were found The Journal of Neuroscience, May 1993, f3(5) 1969

Figure 4. Simultaneous labeling of cells (arrowheads) in fixed brain sections for CCK by fluorescencehistochemistry, followed by 1251-a-BTbinding. a andb showlabeled cells in the hilusat low andhigher magnification. c andd showlabeled neurons in CA3, in stratumlucidum. e showslabeled cellsin CA2 in stratumoriens and in stratumpyramidale. Scale bar: 50pm for a, c, d, ande; 100pm for b.

in the CA3 region, in both stratum oriensand stratum lucidum, ticular neuropeptide. CCK, somatostatin,and NPY were stud- that is, outside the pyramidal layer or interior to it (Fig. 2c,e). ied. Overlap between CCK and ol-BT binding was apparent in The cell bodies and principal dendritic branchesof these cell large cells in the hilus (Fig. 4a,b), as well as in stratum lucidum bodieswere often quite heavily labeled. and stratum oriensofthe CA2 and CA3 region (Fig. 4c-e). Some of the oc-BT-labeledcells in the hilus in CA2 and CA3 were also Immunocytochemical identification somatostatin positive (Fig. Sa,c,d). Somatostatin-positive cells To determine the identity of the neurons labeled with or-BT, that bound a-BT were also prominent in stratum radiatum of identification of neurochemicalcontent by immunocytochem- CA1 (Fig. 5b). NPY-positive neuronsthat also bound ar-BTwere istry wascompared with a-BT binding in the samesection (Fig. most prominent in the boundary betweenthe dentate gyrus and 3). A subpopulation of the neurons reactive with an antibody the hilus (Fig. 6a). The large neuronsin stratum oriens of CA3 to GABA were labeled by a-BT. Although only about one-fifth that bind (u-BT are also notable for their prominent autoflu- of the GABA-immunoreactive neuronswere labeledwith (u-BT, orescence(Fig. 66). Thus, no neuropeptide uniquely definesthe very few neuronswere labeled with a-BT that did not contain (Y-BT population. It should further be noted that, as for GABA GABA immunoreactivity. Thus ol-BT-labeled neurons appear itself, ol-BT binding is found in only a subpopulationof the cells to be a subsetof inhibitory interneurons located throughout the that bind any particular peptide. hippocampus,including CA3 (Fig. 3a,b,d), CA1 (Fig. 3e) and the dentate (Fig. 3~). Eflects of a-BT and other cholinergic drugs on expressionof Hippocampal interneurons contain several neuropeptides. growth factors Immunocytochemistry and CX-BTbinding in the same section Intraventricular administration of ol-BT increasedexpression of were usedto determine if (Y-BT binding colocalized with a par- NGF in rat hippocampus(Fig. 7). With the hybridization con- 1970 Freedman et al. * a-Bungarotoxin Binding to Hippccampal lntemeurons

Figure 5. Simultaneous labeling of cells (arrowheads) in fixed brain sections for somatostatin by fluorescence histochemistry, followed by 1251-.(y- BT binding. a shows neurons in CA2, stratumoriens. b showsneurons in CAl, stratumoriens. c and d showneurons in CA3, stratumoriens. Scalebar, 50 pm. ditions used in this study, NGF expression in control brains (Figs. 8, 9). BDNF hybridization can bc observed in virtually was limited to hilar neurons and scattered cells in Ammon’s all neurons of the hippocampusin some control animals, with horn. The same pattern was observed in unoperated animals appropriate hybridizations conditions (Wetmore et al., 1991). and in those that received injections of Ringer’s solution (Fig. Hybridization in control animals in this study was limited to 7b). A similar distribution of expressionwas seenusing these scattered cells throughout Ammon’s horn (Fig. 8b) and to the hybridization conditions by Ballarin et al. (1991), in their con- hilus (Fig. 84. Similar patterns were observed in both unoper- trol specimens.After a-BT administration, NGF hybridization ated animals and those that received Ringer’s solution, as well increasedin both the dentate (Fig. 7a) and to a lesserextent in as in those reported by Ballarin et al. (1991). Reasonsfor the Ammon’s horn (Table 1). An increasednumber of hilar cells, low basalrate of BDNF expressionmay include the removal of many of them putative interneurons near the margin of the the brain while the animal wasunder barbiturate anesthesiaand granule cell layer, as well as the granule cells themselves,were the relatively short autoradiographic exposure used.After a-BT now labeled. In CA3, cells both within and external to the py- administration, there wasextensive BDNF hybridization in the ramidal layer were labeled, but most pyramidal neurons were hilus and the dentate gyrus (Fig. 8c) and in CA3 (Fig. 8a), not labeled. There was no increasein labeling in CA 1. encompassingboth pyramidal and nonpyramidal neurons.The BDNF expressionalso increasedafter a-BT administration BDNF labeling in the CA3 pyramidal layer wasmore extensive The Journal of Neuroscience, May 1993, 73(5) 1971

Figure 6. a shows labeling of a fixed section of the dentate gyms for NPY by fluorescence histochemistry, followed by 1251-~-BTbinding. b shows large neurons in CA3, stratum oriens, which contain autofluorescent fat granules. These cells are also labeled by 1251-o(-BT.Scale bar, 25 pm. than that observed for NGF. There was no increase in labeling CA3 that contain CCK and somatostatin, NPY-positive inter- in CA1 (Fig. 9). neurons in the dentate, and somatostatin-positive interneurons For both NGF and BDNF, the increase in expression was in CA1 comprise the most prominently labeled populations. bilateral, although increase in expression contralateral to the Only about one-fifth of immunocytochemically defined inter- injection site was significant only for NGF (Table 1). For both neurons bind a-BT in significant amounts. However, all major growth factors, the increase was greatest in CA3 and the dentate. types of hippocampal intemeurons appear to be represented NT3 expression was limited to CA1 and the subiculum; it did among the a-BT-labeled group. It is not unexpected that cu-BT not increase with a-BT administration. None of the growth binding is not colocalized with a particular neuropeptide, as the factors increased expression after atropine administration. neuropeptides themselves may have overlapping distributions among the intemeurons (Sloviter and Nilaver, 1987; Deller and Discussion Leranth, 1990). CX-BT binds throughout the hippocampus, but the most heavily The large intemeurons of CA3 labeled by c+BT may include labeled structures appear to be GABAergic intemeurons in both a group described by Alonso and Kohler (1982). This group of the dentate gyrus and Ammon’s horn. Large intemeurons in large multipolar neurons, similar in description to the neurons 1972 Freedman et al. l a-Bungarotoxin Binding to Hippocampal Interneurons

Figure 7. In situ hybridization using a probe for NGF mRNA in rat hippocampus, 3 hr after intraventricular administration of a-BT (1 pg in 10 11; a) or and equal volume of Ringer’s solution (b). The animal treated with c~-BT shows prominent labeling in the dentate (a). The animal treated with Ringer’s shows no labeling in the dentate (b), with scattered cells labeled in the hilus.

Table 1. Effect of wBT on in situ hybridization of mRNA for BDNF shown in Figure 6b, projects from stratum oriens and stratum and NGF lucidum of CA3 to the medial septal nucleus. Although the major projections of GABAergic neurons in CA3 are thought (u-BT (1 pg, i.c.v.) to be intrahippocampal, these neurons may also exert some Control Contra- inhibitory feedback control over choline@ input into the hip- (Ringer’s) Ipsilateral lateral pocampus.This group of neuronshas also beenshown to receive BDNF innervation from the medial septal nucleus, on the basis of Dentate 0.033 0.670 0.731 transfer of 3H-adenosine(Rose and Schubert, 1977). 0.037 0.635 0.672 The existence of a functional role for (u-BT receptors in the Hilus 0.061 0.130 0.012 CNS has been debated for some time, becausethe distribution 0.063 0.222 0.090 of a-BT binding does not correspond with the distribution of CA3 0.099 0.370 0.282 high-affinity nicotine and ACh binding sites(Hsrfstrand et al., 0.096 0.734 0.379 1988)and becausemany electrophysiologicalstudies found that CA1 0.058 0.102 0.082 the effectsof nicotine and ACh were not blocked by (Y-BT (Dug- 0.049 0.083 0.091 gan et al., 1976; Bursztajn and Gershon, 1977). The (~3and (~4 receptor subunits also do not form a-BT-sensitive receptorsin NGF oocytes (Boulter et al., 1987).Recently, however, severalstudies Dentate 0.047 0.321 0.334 0.042 0.372 0.302 t Hilus 0.114 0.080 0.120 Each value is the mean of three grain density determinations over a 500 pm2 area, 0.120 0.133 0.102 expressedas fraction of the area occupied by grains; values from two animals are CA3 0.037 0.140 0.030 shown. ANOVA showed a significant effect oftreatment on hybridization ofprobes 0.042 0.170 0.042 for both BDNF (F = 10.2; df 1,3; p c 0.05) and NGF mRNA (F = 32.8; df 1,3; p < 0.05) in the hippocampus ipsilateral to the ventricle used for administration. CA1 0.058 0.021 0.033 There were significant contralatemf effects onty for NGF (F = 70.6; df 1,3; p < 0.049 0.025 0.026 0.05). There were no significant effects for hippocampal region, which was used as a nested factor in the ANOVA. The Journal of Neuroscience, May 1993, U(5) 1973

Figure 8. In situ hybridization of BDNF mRNA in rat hinDocamous.3 hr after intraventricularadmini&atibn of a-BT (1 fig; a, c) or Ringer’ssolution (b, d). a-BT administrationcaused an increasedbinding in CA3 (a)and in the dentate gyrus and hilus (c). Ringer- treatedanimals showed only scattered cellslabeled in CA3 (b)and in the hilus (d), with little labelingin the dentate. Scalebar, 100pm. have shown blockade by (u-BT of effects of ACh and nicotine dentateand throughout Ammon’s horn. In the experimentswith on inhibitory interneurons in the cerebellum (de la Garza et al., ol-BT reported here, the increasein expressionwas principally 1987) as well as a role for a-BT in the physiology of the CA3 in CA3 and the dentate gyrus. This finding suggeststhat a-BT region in the hippocampus(Marks et al., 1989; Luntz-Leybman doesnot nonspecifically excite the hippocampus,but rather acts et al., 1992). The expressionof the a7 subunit in oocytes further primarily in CA3 and the dentate, where the cholinergic septal supports a functional role for the a-BT-binding protein (Cou- innervation is greatest. Furthermore, within the dentate, the turier et al., 1990). The effects of cY-BT-sensitivereceptors may increasein expressionof growth factors is greatestin the upper have been obscuredin other regions by their coexistence with blade, where cu-BT-binding interneurons are concentrated.CA3 other nicotinic receptors. In isolated gene expressionsystems, is the area where physiological effects sensitive to cu-BT have the function of the a-BT-sensitive a7 nicotinic receptor subunit been reported (Miner and Collins, 1989; Luntz-Leybman et al., can be obscuredby coexpressionof the ~y3receptor (Listerud et 1992). By contrast, electrophysiological effects of nicotine in al., 1991). Although in situ hybridization suggeststhat cu7is CA1 are not blocked by a-BT (Freund et al., 1990). It is possible expressedthroughout the hippocampus(Johnson et al., 199 l), that the coexistence of (~3 subunits in some CA1 cells may a3 is also expressedin someareas, particularly in CA1 (Wada obscurethe effectsof a-BT-sensitive receptors, as in the single- et al., 1989). cell expressionsystems described above. The increasein BDNF and NGF expressionafter a-BT ad- The increasedexpression of mRNA for NGF and BDNF was ministration is further evidence for a functional effect of (u-BT- not limited to interneurons that bound CX-BT,but rather was sensitivereceptors. BDNF and NGF expressionincrease in the prominent in the CA3 pyramidal cell layer and the dentate hippocampusafter a variety of treatments, including kindling granulescells as well, where CPBTbinding itself is lessintense. and treatment with excitants (Gall and Issackson,1989; Ballarin The finding of increasedgrowth factor expressionthroughout et al., 1991). In those experiments, expressionincreased in the the hippocampus after treatment with excitants such as kainic 1974 Freedman et al. l a-Bungarotoxin Binding to Hippocampal Interneurons

Figure 9. In situ hybridizationof BDNF mRNA 3 hr after ipsilateralintraventricular administration of (Y-BT(1 pg).Labeling is mostprominent in the CA3 and the areadentata (AD) layers,and clearlyless in CAI. The dorsalblade of the dentategyms is moreprominently labeled than the inferior blade. acid suggeststhat such increasemay occur whenever hippocam- expression(Alberch et al., 1991). cu-BT-sensitivereceptors may pal neuronsare stimulated. Blockade of the cholinergic input therefore have a unique role in regulating hippocampal physi- to selectedinterneurons in the hippocampus by (w-BT may be ology, not sharedby other types of cholinergic receptors. Fur- sufficient to excite the dentate granule cells and the CA3 pyra- thermore, nicotinic receptorson interneurons may be a common midal cells and thus to produce the increasepattern of expres- mechanismin the CNS. In addition to the innervation of the sion observed. Excitation is commonly observed when inter- hippocampal interneurons observed here, there are nicotinic neuronal function is antagonized (Schwartzkroin, 1986). synapsesreported on interneurons in the cerebellum and the Receptors on the pyramidal neurons themselvesmight also be retina (de la Garza et al., 1987; Hamassakibritto et al., 1991). responsible(Hunt and Schmidt, 1978). Although ol-BT deliv- ered intraventricularly primarily binds in the ipsilateral hip- References pocampus,effects on growth factor expressionwere observed Alberch J, Carman-KrzanM, FabrazzoM, WiseBC (1991) Chronic bilaterally. This finding is consistent with the commissuralcon- treatmentwith scopolamineand physostigmine changes nerve growth factor (NGF) receptordensity and NGF contentin rat brain. Brain nections between CA3 and the dentate, which would support Res5421233-240. bilateral excitation even if only one side were affected by a-BT. AlonsoA, Kohler C (1982) Evidencefor separateprojections of hip- The NPY-containing interneurons of the dentate are also known pocampalpyramidal and non-pyramidalneurons to differentparts of to have such connections (Deller and Leranth, 1990), as well as the septumin the rat brain.Neurosci Lett 3 1:209-214. the CA3 pyramidal neuronsthemselves. Furthermore, although Ayer-LeLievre C, OlsonL, EbendalT, SeigerA, PerssonH (1988) Expressionof the p-nervegrowth factor gene in hippocampalneurons. the data are consistent with the action of o-BT on hippocampal Science240: 1339-l 341. receptors, actions of a-BT elsewherein the brain after intra- BallarinM, EmforsP, LindeforsN, PerssonH (1991) Hippocampal ventricular administration are also possible. damageand kainic acid injectioninduce a rapid increasein mRNA It is noteworthy that a-BT increasedgrowth factor expression for BDNF and NGF in the rat brain.Exp Neurol 114:3543. Bickford-WimerPC, Nagamoto H, JohnsonR, Adler LE, EganM, Rose but atropine, a muscarinic antagonist, did not. Measurementof GM, FreedmanR (1990) Auditory sensorygating in hippocampal NGF protein in hippocampus and mRNA in cortex has also neurons.a modelsystem in the rat. BiaI Psychiatry27:183-192. shown no effect of muscarinic antagonism on growth factor BoulterJ, ConnollyJ, DenerisE, GoldmanD, HeinemannS, Patrick J The Journal of Neuroscience, May 1993, f3(5) 1975

(1987) Functional expression of two neuronal nicotinic acetylcholine Korsching S, Auburger G, Heumann R, Scott J, Thoenen H (1985) receptors from cDNA clones identifies a gene family. Proc Nat1 Acad Levels of nerve growth factor and its mRNA in the central nervous Sci USA 84~7763-7767. system of the rat correlate with cholinergic innervation. Eur Mol Biol Bursztajn S, Gershon MD (1977) Discrimination between nicotinic Org J 4:1389-1393. receptors in vertebrate ganglia and skeletal muscle by alpha-bungaro- Lippa AS, Pelhman RW, Beer B, Critchett DJ, Dean RL, Bartus RT toxin and cobra venoms. J Physiol (Lond) 269: 17-3 1. (1980) Brain cholinergic dysfunction and memory. in aged- rats. Neu- Clarke PBS, Schwartz RD, Paul SM, Pert C, Pert A (1985) Nicotine robiol Aging 1:13-19.- - binding in rat brain: autoradiographical comparison of [)H]-acetyl- Listerud M. Brussaard AB. Devav P. Colman DR. Role LW (1991) choline, [‘HI-nicotine and [12SI]-alpha-bungarotoxin. J Neurosci Functional contribution of neuronal AChR subunits revealed by an: 5:1307-1315. tisense oligonucleotides. Science 254: 15 18-l 52 1. Couturier S, Bertrand D, Matter JM, Hemandez MC, Bertrand S, Millar Luntz-Leybman V, Bickford P, Freedman R (1992) Cholinergic gating N, Valera S, Barkas T, Ballivet M (1990) A neuronal nicotinic of response to auditory stimuli in rat hippocampus.__ _ Brain Res 587: acetylcholine receptor subunit (alpha 7) is developmentally regulated 130-l-36. and forms a homo-oligomeric channel blocked by alpha-BTX. Neu- Marks MJ, Romm E, Campbell SM, Collins AC (1989) Variation of ron 5847-856. nicotinic bindina sites among inbred strains. Pharmacol Biochem de la Gat-za R, McGuire TJ, Freedman R, Hoffer BJ (1987) Selective Behav 33~679-689. antagonism of nicotine actions in the rat cerebellum with cu-bungar- Miner LL, Collins AC (1989) Strain comparison of nicotine-induced otoxin. Neuroscience 23:887-89 1. seizure sensitivity and nicotinic receptors. Pharmacol Biochem Behav Deller T, Leranth C (1990) Synaptic connections of neuropeptide Y 33:469-475. (NPY) immunoreactive neurons in the hilar area of the rat hippo- Pauly JR, Stitzel JA, Marks MJ, Collins AC (1989) An autoradio- campus. J Comp Neurol 300:433-447. graphic analysis of cholinergic receptors in mouse brain. Brain Res Duggan AW, Hall JG, Lee CY (1976) Alpha-bungarotoxin, cobra 221453459. neurotoxin and excitation of Renshaw cells by acetylcholine. Brain Phillips HS, Hains JM, Laramee GR, Rosenthal A, Winslow JW (1990) Res 107:166-170. Widespread expression of BDNF but not NT3 by target areas of basal Emfors P, Iblfiez CF, Ebendal T, Olson L, Persson H (1990) Molecular forebrain choline& neurons. Science 250:290-294. cloning and neurotrophic activities of a protein with structural sim- Rose G, Schubert P (1977) Release and transfer of [)H] adenosine ilarities to nerve growth factor: developmental and topographical ex- derivatives in the cholinergic septal system. Brain Res 121:353-357. pression in the brain. Proc Nat1 AcadSci USA 87:545&5458. Schulz DW, Loring RH, Aizenman E, Zigmond RE (1991) Autora- Freund RK. Juneschaffer DA. Collins AC (1990) Nicotine effects in diographic localization of putative nicotinic receptors in the rat brain mouse hippocampus are blocked by mecamylamine, but not other using 1251-neuronal bungarotoxin. J Neurosci 11:287-297. nicotinic antagonists. Brain Res 5 11: 187-l 9 1. Schwartzkroin PA (1986) Regulation of excitability in hippocampal Frotscher M, Leranth C (1985) Cholinergic innervation of the rat neurons. In: The hiDDOCamDUS.__ _. Vol 3 (Issacson RL. Pribram KH. hippocampus as revealed by choline acetyltransferase immunocyto- eds), pp 113-136. New York: Plenum. . chemistry. J Comp Neurol 239:237-246. Segal M, Dudai Y, Amsterdam A (1978) Distribution of cu-bungaro- Fuchs JL (1989) [1251]ol-Bungarotoxin binding marks primary sensory toxin-binding cholinergic nicotinic receptor in rat brain. Brain Res areas of developina rat neocortex. Brain Res 50 1:223-234. 148:105-l 19. Gall CM, Isackson Pj (1989) Limbic seizures increase neuronal pro- Sequala P, Geffard M, Buijs RM, LeMoal M (1984) Antibodies against duction of messenger RNA for nerve growth factor. Science 245:758- gamma amino butyric acid: specificity studies and immunocytochem- 761. ical results. Proc Nat1 Acad Sci USA 8 1:3888-3892. Green RC, Blume HW, Kupferschmid SB, Mesulam MM (1989) Al- Sloviter RS, Nilaver G (1987) Immunocytochemical localization of terations of hippocampal acetylcholinesterase in human temporal lobe GABA-, cholecystokinin-, vasoactive intestinal polypeptide-, and so- epilepsy. Ann Neurol 26:347-35 1. matostatin-like immunoreactivity in the area dentata and hippocam- Hamassakibritto A, Brzozowskaprechtl A, Karten HJ, Lindstrom JM, pus of the rat. J Comp Neural 256:42-60. Keyser KT (199 1) GABA-like immunoreactive cells containing nic- Vanderwolf CH (1975) Neocortical and hippocampal activation in otinic acetylcholine receptors in the chick retina. J Comp Neurol3 13: relation to behavior. J Comp Physiol Psycho1 88:300-323. 394-401. Vinogradova 0 (1975) Functional organization of the Iimbic system Hlrfstrand A, Adem A, Fuxe K, Agnati L, Andersson K, Nordberg A in the process of registration of information: facts and hypotheses. (1988) Distribution of nicotinic cholinergic receptors in the rat tel- In: The hippocampus: neurophysiology and behavior, VoI 2 (Issacson and diencephalon: a quantitative receptor autoradiographical study RL, Pribram KH, eds), pp l-70. New York: Plenum. using [3H]-acetylcholine, [LU-I~~I]bungarotoxin and [3H]nicotine. Acta Wada E, Wada K, Boulter J, Deneris E, Heinemann S, Patrick J, Swan- Physiol Stand 132: 1-14. son LW (1989) Distribution of aloha2. alnha3. aloha4. and beta2 Hunt SP, Schmidt J (1978) The electron microscopic autoradiographic neuronal nicotinic receptor subunit mRNAs in the central nervous localization of alpha-bungarotoxin binding sites within the central system: a hybridization histochemical study in the rat. J Comp Neurol nervous system of the rat: Brain Res 142: 152-l 59. 284:314-335. Johnson DS. Stroessner-Johnson HM. Boulter J. Amaral DG. Heine- Wetmore C, Cao Y, Pettersson RF, Olson L (199 1) Brain-derived mann S (199 1) Distribution of alpha7 neuronal nicotinic receptor neurotrophic factor: subcellular compartmentalization and intemeu- mRNA in the rat CNS. Sot Neurosci Abstr 17:249. ronal transfer as visualized with anti-peptide antibodies. Proc Nat1 Jovce_~,~~_ JN. Gibbs RB. Cotman CW. Marshall JF (1989). , Reaulation of Acad Sci USA 88:9843-9847. muscarmic receptors in hippocampus following cholinergic dener- Whittemore SR, Friedman PL, Larhammar D, Persson H, Gonzalez- vation and reinnervation by septal and striatal transplants. J Neurosci Carvajal M, Holets VR (1988) Rat o-nerve growth factor sequence 912776-2791. and site of synthesis in the adult hippocampus. J Neurosci Res 20: Kohler C, Chart-Palay V (1982) Somatostatin-like immunoreactive 403-410. neurons in the hippocampus: an immunocytochemical study in the rat. Neurosci Lett 34:259-264.