Brain Research 943 (2002) 132–141 www.elsevier.com / locate / bres Interactive report Selective lesion of cholinergic neurons in the medial septum by 192 IgG-saporin impairs learning in a delayed matching to position T- maze paradigm a, a b * David A. Johnson , Natalie J. Zambon , Robert B. Gibbs a Division of Pharmacology-Toxicology, Mylan School of Pharmacy, Graduate School of Pharmaceutical Sciences, Duquesne University, Pittsburgh, PA 15282, USA b Department of Pharmaceutical Sciences, University of Pittsburgh School of Pharmacy, Pittsburgh, PA 15261, USA Accepted 21 March 2002 Abstract This study examined whether selective destruction of cholinergic neurons in the medial septum impairs acquisition of a delayed matching-to-position (DMP) spatial memory task. Either the selective immunotoxin 192 IgG-saporin (SAP; 0.22 or 1.0 mg) or the non-selective excitatory neurotoxin ibotenate (IBO; 5 mg), was infused directly into the medial septum of rats. Both doses of SAP, but not IBO, significantly impaired acquisition of the DMP task and blunted the initial alternating behavior of the rats in the T-maze. Histochemical staining revealed that both doses of SAP produced a near complete depletion of choline acetyltransferase (ChAT)-positive neurons in the medial septum. Some loss of parvalbumin staining was observed following administration of the higher dose, but not the lower dose, of SAP. In contrast, IBO produced a nearly complete depletion of parvalbumin-positive staining throughout the medial septum. IBO also produced a loss of ChAT-positive neurons and considerable local damage in the medial septum around the area of injection; however, many ChAT-positive neurons in the medial septum distal to the injection remained. A significant correlation between the number of days to reach criterion and ChAT activity in the frontal cortex and hippocampus was observed. The results suggest that low doses of SAP can be used to selectively destroy cholinergic neurons in the medial septum, and that selective destruction of these neurons significantly impairs acquisition of the DMP task. We propose that acquisition of the DMP task is a sensitive behavioral assay for the selective loss of basal forebrain cholinergic projections. 2002 Elsevier Science B.V. All rights reserved. Theme: Neurotransmitters, modulators, transporters, and receptors Topic: Behavioural pharmacology Keywords: Spatial memory; Acetylcholine; Basal forebrain; Ibotenate; Cognitive impairment 1. Introduction standing of the function of basal forebrain cholinergic neurons has accelerated since the development of 192 Based on experiments utilizing muscarinic antagonists IgG-saporin (SAP), a selective immunotoxin that destroys and / or nonselective physical or chemical lesions, choliner- cholinergic neurons that express the low affinity neuro- gic projections emanating from basal forebrain structures trophin receptor (p75NTR) [9,46]. Initial studies reported including the medial septum (MS), nucleus basalis mag- SAP induced impairments in working memory in the nocellularis (NBM), and the vertical limb of the diagonal Morris water maze (MWM) that were consistent with band of Broca (VDB), have been identified as playing an earlier reports using non-selective chemical lesions [28]. important role in cognitive functions such as attention, However, these studies administered relatively large doses working memory and learning [2–4,14,20]. An under- of SAP into the cerebral ventricular system that sub- sequently were demonstrated to produce non-selective damage, including damage to Purkinje cells of the cere- *Corresponding author. Tel.: 11-412-396-5952; fax: 11-412-396- bellum [40,42]. Studies that injected smaller doses of SAP 4660. E-mail address: [email protected] (D.A. Johnson). directly into basal forebrain structures were demonstrated 0006-8993 / 02 / $ – see front matter 2002 Elsevier Science B.V. All rights reserved. PII: S0006-8993(02)02623-9
Transcript
Brain Research 943 (2002) 132–141www.elsevier.com/ locate /bres
Interactive report
Selective lesion of cholinergic neurons in the medial septum by 192IgG-saporin impairs learning in a delayed matching to position T-
maze paradigma , a b*David A. Johnson , Natalie J. Zambon , Robert B. Gibbs
aDivision of Pharmacology-Toxicology, Mylan School of Pharmacy, Graduate School of Pharmaceutical Sciences, Duquesne University, Pittsburgh,PA 15282, USA
bDepartment of Pharmaceutical Sciences, University of Pittsburgh School of Pharmacy, Pittsburgh, PA 15261, USA
Accepted 21 March 2002
Abstract
This study examined whether selective destruction of cholinergic neurons in the medial septum impairs acquisition of a delayedmatching-to-position (DMP) spatial memory task. Either the selective immunotoxin 192 IgG-saporin (SAP; 0.22 or 1.0 mg) or thenon-selective excitatory neurotoxin ibotenate (IBO; 5 mg), was infused directly into the medial septum of rats. Both doses of SAP, but notIBO, significantly impaired acquisition of the DMP task and blunted the initial alternating behavior of the rats in the T-maze.Histochemical staining revealed that both doses of SAP produced a near complete depletion of choline acetyltransferase (ChAT)-positiveneurons in the medial septum. Some loss of parvalbumin staining was observed following administration of the higher dose, but not thelower dose, of SAP. In contrast, IBO produced a nearly complete depletion of parvalbumin-positive staining throughout the medialseptum. IBO also produced a loss of ChAT-positive neurons and considerable local damage in the medial septum around the area ofinjection; however, many ChAT-positive neurons in the medial septum distal to the injection remained. A significant correlation betweenthe number of days to reach criterion and ChAT activity in the frontal cortex and hippocampus was observed. The results suggest that lowdoses of SAP can be used to selectively destroy cholinergic neurons in the medial septum, and that selective destruction of these neuronssignificantly impairs acquisition of the DMP task. We propose that acquisition of the DMP task is a sensitive behavioral assay for theselective loss of basal forebrain cholinergic projections. 2002 Elsevier Science B.V. All rights reserved.
Theme: Neurotransmitters, modulators, transporters, and receptors
1. Introduction standing of the function of basal forebrain cholinergicneurons has accelerated since the development of 192
Based on experiments utilizing muscarinic antagonists IgG-saporin (SAP), a selective immunotoxin that destroysand/or nonselective physical or chemical lesions, choliner- cholinergic neurons that express the low affinity neuro-gic projections emanating from basal forebrain structures trophin receptor (p75NTR) [9,46]. Initial studies reportedincluding the medial septum (MS), nucleus basalis mag- SAP induced impairments in working memory in thenocellularis (NBM), and the vertical limb of the diagonal Morris water maze (MWM) that were consistent withband of Broca (VDB), have been identified as playing an earlier reports using non-selective chemical lesions [28].important role in cognitive functions such as attention, However, these studies administered relatively large dosesworking memory and learning [2–4,14,20]. An under- of SAP into the cerebral ventricular system that sub-
sequently were demonstrated to produce non-selectivedamage, including damage to Purkinje cells of the cere-*Corresponding author. Tel.: 11-412-396-5952; fax: 11-412-396-bellum [40,42]. Studies that injected smaller doses of SAP4660.
E-mail address: [email protected] (D.A. Johnson). directly into basal forebrain structures were demonstrated
0006-8993/02/$ – see front matter 2002 Elsevier Science B.V. All rights reserved.PI I : S0006-8993( 02 )02623-9
D.A. Johnson et al. / Brain Research 943 (2002) 132 –141 133
to more selectively destroy cholinergic neurons in the basal Natick, MA; 5 mg in 1 ml, (n59), infused into the medialforebrain, but the resulting deficits in working memory septum over 5 min at a rate of 0.2 ml /min. Followingwere either modest [15], or nonexistent [5,12]. Since the infusion, the cannula was left in place for 5 min to allowMS has both cholinergic and GABAergic neurons that diffusion into the tissues before withdrawal of the cannula.project to the hippocampus, a more prominent role for The incision was closed and the animals were allowed toGABAergic septal projections in working memory has recover for 2 weeks.been hypothesized [1]. It is possible that other tests may bemore sensitive for detecting cognitive deficits associated 2.2. Behavioral testingwith the selective loss of cholinergic projections from theMS, particularly tests which assess learning as opposed to The delayed matching-to-position (DMP) task wasworking memory and retention. performed as recently described [22,24]. We elected to use
Infusions of low doses of the excitotoxin ibotenic acid a matching to position as opposed to a non-matching to(IBO) into the basal forebrain have been shown to position paradigm to offset the animals’ natural tendencypreferentially damage non-cholinergic neurons relative to to alternate [16,39]. Animals were food deprived to 85% ofcholinergic neurons as measured by cortical ChAT activity their normal body weight and then adapted to the T-mazeor by counting cortical acetylcholinesterase (AchE)-posi- by placing them in the maze with sweetened reward foodtive fibers [8,18]. The intent of the current study was to pellets (45 mg Noyes pellets; Lancaster, NH) once per daycompare the effects of septal injections of SAP with IBO for 5 days. Animals were then trained to run to the ends ofon acquisition of a spatial memory task, and to correlate the goal arms by using a series of eight forced ‘choices’effects on performance with markers for cholinergic and per day for 3 days, each rewarded with two Noyes pellets.GABAergic projections. Right and left arms were alternated in a random, balanced
(4R, 4L) fashion to avoid the introduction of a side bias.Animals then began DMP training. DMP training was
2. Materials and methods performed in trial pairs. Each animal received eight trialpairs /day. The first trial consisted of a forced ‘choice’ in
All chemicals were purchased from Sigma, Inc., St. which one goal arm was blocked, forcing the animal toLouis unless stated otherwise. enter the unblocked arm to receive two reward pellets. The
animal was then returned to the approach alley for the2.1. Intraseptal infusions second trial during which both goal arms were open.
Returning to the same arm visited on the previous trialAll experiments followed guidelines of the N.I.H. for resulted in a reward (four pellets). Entering the incorrect
the care and use of laboratory animals and Duquesne arm resulted in no reward and confinement to the arm forUniversity policies and with the approval of the University 10 seconds. Animals were run in squads of three or four.IACUC. Forty-two male Sprague–Dawley rats weighing After each trial pair, the animals were returned to theirbetween 275 and 300 g were obtained from Hilltop Lab cages for 5–10 min (variability based on whether thereAnimals (Scottdale, PA) and housed in a temperature and were intratrial delays), while training proceeded with thehumidity controlled facility with a 12:12 h light /dark cycle other animals. Animals continued to receive eight trialwith food and water readily available. On the day of pairs per day until they reached a criterion of 15 correctsurgery, rats were anesthetized with pentobarbital (50 mg/ choices out of 16 consecutive trials. Animals that reachedkg; IP), and placed in a stereotaxic frame (Stolting, Wood criterion were re-tested 7 days later to assess maintenanceDale, IL). An incision was made exposing the dorsal aspect of performance. For 2 days after reaching criterion,of the skull and a small hole was drilled through which a animals were tested using the standard paradigm. Animalsstainless steel cannula was placed into the medial septum, then received an additional 6 days of testing with increas-using coordinates from Bregma: AP 10.2 mm, L 0.0, DV ing delays (45, 60, 90 s) between the forced and open25.6 mm. Two doses of SAP were used. A relatively high choice (2 days per delay). Performance during acquisitiondose (1.0 mg) was selected from the literature to produce a and during post-criterion testing was analyzed using thesubstantial loss of cholinergic projections to the hippocam- non-parametric Freidman test for repeated measures. Dayspus and frontal cortex. A lower dose (0.22 mg) was also to criterion (DTC) were analyzed using non-parametricincluded to help control for non-selective damage to the Kruskal–Wallis ANOVA. Individual comparisons utilizedMS that might be produced by the higher dose. A single the Dunn’s test post-hoc. Comparisons in performance fordose of IBO (5.0 mg) was selected based on previous post-criterion testing utilized one-way ANOVA. GraphPadinvestigations that assessed the effects of IBO lesions to Prism version 3.02 was utilized for all statistical tests.the MS or NBM on spatial memory [30,44]. Animalsreceived either normal saline (sham control) 1 ml (n513); 2.3. ImmunostainingSAP (Chemicon, Temecula, CA; 0.22 (n57) or 1.0 mg(n513) in 1 ml prepared from the same lot); or IBO (RBI, Following behavioral testing, several rats from each
134 D.A. Johnson et al. / Brain Research 943 (2002) 132 –141
group were anesthetized with pentobarbital (100 mg/kg; protein [10]. Three 5-ml aliquots of each sample wereIP), and perfused with ice-cold saline followed by 50 ml of incubated for 30 min at 37 8C in a medium containing [3H]4% paraformaldehyde (PARA) in 50 mM sodium acetate acetyl-CoA (50 000–60 000 d.p.m. / tube, final concentra-(pH 6.5) and then with 50 cc of 4% PARA in 50 mM tion 0.25 mM acetyl CoA, Sigma, Inc., St. Louis, MO),Tris–HCl (pH 9.0). Brains were removed and placed into choline chloride (10.0 mM), physostigmine sulfate (0.24% PARA in 50 mM phosphate buffer at 4 8C for 3 h, and mM), NaCl (300 mM), sodium phosphate buffer (pH 7.4,then transferred to 15% sucrose in 50 mM phosphate 50 mM), and EDTA (10 mM). The reaction was termi-buffered saline (PBS) at 4 8C overnight. To assess the loss nated with 4 ml sodium phosphate buffer (10 mM)of markers for cholinergic and GABAergic projections in followed by the addition of 1.6 ml of acetonitrile con-the MS, thirty micron frozen sections through the MS taining 5 mg/ml tetraphenylboron. The amount of [3H](corresponding to plates 14–21 of [36]) were cut and acetylcholine produced was determined by adding 8 ml ofplaced into cold PBS and processed for the immuno- EconoFluor scintillation cocktail (Packard Instruments,cytochemical detection of either ChAT- or parvalbumin Meriden, CT) and counting total cpms in the organic phase(PARV)-like immunoreactivity (IR) as previously de- using an LKB beta-counter. Background was determinedscribed [25,26]. Briefly, adjacent sections were incubated using identical tubes to which no sample was added. Forwith an antibody against ChAT (goat anti-ChAT, each sample, the three reaction tubes containing sampleChemicon International, Inc., Temecula, CA, AB144P were averaged and the difference between total cpms anddiluted, 1:3500) or against PARV (mouse anti-parvalbumin, background cpms was used to estimate the total amount ofSigma, Inc., St. Louis, MO, F3171, diluted 1:12 000) for 3 acetylcholine (ACh) produced per sample. ChAT activitydays at 4 8C. All antibodies were diluted with PBS was then calculated for each sample as pmol AChcontaining 0.05% Triton X-100 and either 5% normal goat manufactured /h /mg protein. ChAT activity in the HIP, FC,serum (for ChAT staining) or 5% normal horse serum (for and MS/DB was correlated with DTC by performingPARV staining) (serums obtained from Vector Laboratories, Spearman rank order correlations. Effects on ChAT activi-Inc., Burlingame, CA). After incubating with the primary ty were analyzed using ANOVA.antibody, sections were rinsed with PBS and incubated for1 h at room temperature with an appropriate biotinylated 2.5. AChE assaysecondary antibody (either biotinylated rabbit-anti-goatIgG diluted 1:1000, or biotinylated horse-anti-mouse IgG The sonicates described above were also assayed fordiluted 1:100). The sections were then rinsed with PBS, acetylcholinesterase (AChE) activity using methodsincubated for 1 h in a solution containing avidin–HRP adapted from Ellman et al. [19]. Briefly, three 5-ml aliquotscomplex (Vectastain Elite Kit), and rinsed again. of each sample were placed into separate wells of aBiotinylated antibodies and Vectastain Elite kit were 96-well plate containing 175 ml of reaction solutionobtained from Vector Laboratories, Inc., Burlingame, CA. composed of sodium phosphate buffer (pH 7.4, 50 mM),The sections were then rinsed with Tris acetate solution tetraisopropyl pyrophosphoramide (IsoOMPA, 0.03 mM),(50 mM, pH 7.6), and transferred to a Tris acetate solution and 5,59-dithiobis-(2-nitrobenzoic acid) (0.33 mM). Platescontaining 3,39-diaminobenzidine (0.5 mg/ml), H O were preincubated for 30 min at 37 8C to inactivate non-2 2
(0.01%), and NiCl (0.032%) for 10–15 min. Sections specific esterases. The reaction was initiated by adding 202
were then rinsed well with PBS, mounted onto glass slides, ml of acetylthiocholine iodide to a final concentration ofdehydrated, coverslipped, and examined with a Leitz 0.5 mM. The plates were immediately mixed and scannedphotomicroscope. Sections through the cerebellum of these at 410 nm using an Emax plate reader (Molecular Devices,animals were mounted and stained with Cresyl violet to Inc., Menlo Park, CA). Plates were incubated in the darkassess viability of the cerebellar Purkinje cells. for an additional 30 min at 37 8C, and then scanned again
at 410 nm. AChE activity (mmol /mg/h) was calculated as2.4. ChAT assay the average net [OD 3reaction volume] / [extinction410
coefficient (13.6 nMol /ml)3sample concentration3sampleTwenty-six animals were processed for the quantifica- volume3incubation time]. The rate of ACh hydrolysis was
tion of brain ChAT activity. These animals were anes- expressed as mmol /h /mg protein. All samples from athetized with pentobarbital (100 mg/kg; IP), the brains given brain region were assayed at the same time. Effectsremoved, and tissues from the MS/DB, frontal cortex on AChE activity were analyzed using ANOVA.(FC), hippocampus (HIP), and cerebellum (CB), dissected,frozen, and processed. ChAT activity was measured as 2.6. Resultspreviously described [13,23]. Briefly, frozen tissues werethawed and dissociated by sonication in a medium con- Of the original 42 animals, four animals (one control,taining EDTA (10 mM) and Triton X-100 (0.5%) and one ibotenate, two SAP 1.0 mg) would not run the mazediluted to a concentration of 10 mg tissue /ml). An aliquot and were excluded from the study. Analysis of theof each sample was used for the determination of total remaining animals (12 control, seven SAP 0.22 mg, 11
D.A. Johnson et al. / Brain Research 943 (2002) 132 –141 135
SAP 1.0 mg and eight IBO 5.0 mg) indicated that SAPtreatment, but not IBO, significantly impaired acquisitionof the DMP task. Specifically, both SAP treatment groups,but not the IBO group, required significantly more days toreach criterion compared to sham operated controls (Fig.1). Analysis of the learning curves showed that controlanimals, but not other groups, performed significantlybelow chance on the first day of testing, consistent with theanimals’ innate tendency to alternate (Figs. 2 and 3). Incontrast, the SAP and IBO treated groups did not differsignificantly from chance on the first day of testing. Theperformance of all groups improved over time; however,the overall performance of the SAP treated animals wassignificantly lower than controls (P,0.05 for the 0.22 mggroup, P,0.001 for the 1.0 mg SAP group; rank sumdifferences of 34.0 and 82.5 respectively; Fig. 2). Incontrast, the performance of IBO-treated animals was notsignificantly different from controls (P.0.05; rank sumdifference 24.0). Moreover, the effects of SAP on DMPperformance were dose related with the greater impairmentobserved with the higher dose.
Fig. 2. Acquisition of the DMP Task. The performance of all groupsTo compare treatment effects at different stages of improved over time; however, the overall performance of the SAP-treated
acquisition, the data were blocked into three periods animals was significantly worse than controls (P,0.05 for the 0.22 SAPgroup, P,0.001 for the 1.0 mg SAP group; rank sum differences wererepresenting days 1–10, 11–20, and 21–30 of training34.0, and 82.5, respectively). The performance of IBO-treated animals did(Fig. 3). Analysis of the blocked data for days 11–20not differ significantly from controls (P.0.05; rank sum difference 24.0).revealed a significant decrease in performance of SAP*P,0.05 relative to controls.
treated animals (P,0.01 for 0.22 mg, P,0.001 for 1.0 mgSAP; rank sum differences of 24.5 and 40.0 respectively).In the 21–30-day block, only the 1.0 mg SAP group was Increasing the inter-trial delay to 60 s decreased per-significantly different from controls (P,0.001; rank sum formance in all treatment groups (18.2%, control; 32.2%,difference 38.5). 0.22 SAP; 11.0%, 1.0 SAP; and 27.5%, IBO). However,
Of the 38 animals that completed training, four (1.0 mg there were no significant differences in performanceSAP), did not reach criterion. Of the remaining 34 animals, between treatment groups with the delay (F51.22; P.
no significant treatment effects were detected when per- 0.05).formance was reassessed 1 week after reaching criterion. Immunostaining revealed a near complete loss of ChAT-
positive cells in the MS and the vertical limb of thediagonal band of Broca (VDB) of SAP-treated animalswith both the low and high doses (Fig. 4). With the highdose, loss of ChAT positive cells extended back to thehorizontal limb of the diagonal band of Broca (HDB), andeven to the rostral NBM. Some loss of PARV staining wasobserved in the MS of animals that received the 1 mg doseof SAP, but not in animals that received the 0.22 mg dose.No loss of ChAT-positive cells in other regions of thebrain was detected, including the adjacent striatum. Incontrast, IBO produced a substantial loss of PARV stainingin the MS (Fig. 4). IBO produced considerable localdamage around the infusion site and a substantial loss ofChAT-positive cells proximal to the injection; however,many ChAT-positive cells distal to the injection weredetected, indicating that a considerable number ofcholinergic projections were retained following injectionsof IBO. No loss of Purkinje cells in the cerebellum was
Fig. 1. Days to criterion. Bars indicate median days of training requireddetected following injections of SAP or IBO (not shown).to reach criterion for each group. Note that SAP treated animals required
Analysis of ChAT activity was consistent with thesignificantly more days to reach criterion than the controls. *P,0.05relative to controls. immunohistochemical results. Significant decreases in
136 D.A. Johnson et al. / Brain Research 943 (2002) 132 –141
Fig. 3. DMP performance blocked for time periods. Charts show performance on Day 1 of training, and blocked across Days 1–10, 11–20, and 21–30 oftraining. On Day 1 of training, the number of correct choices made by controls was significantly fewer than for SAP or IBO treated animals (P,0.05),suggesting a difference in the rate of spontaneous alternation; however, there were no significant group differences in overall performance across Days1–10 of training. In contrast, SAP-treated (but not IBO-treated) animals performed significantly worse than controls during days 11–20 of training(P,0.01 for 0.22 mg SAP, P,0.001 for 1.0 mg SAP; rank sum differences were 24.5 and 40.0 respectively). Animals that received the highest dose of SAPcontinued to perform significantly worse than controls during Days 21–30 of training (P,0.001, rank sum difference was 38.5). **P,0.01, ***P,0.001relative to controls.
ChAT activity were detected in the MS, HIP, and FC staining for ChAT protein, indicating a loss of cholinergicfollowing injections of SAP, but not IBO (P,0.001; Table neurons in that structure. The loss of ChAT-positive1). Significant decreases in AChE activity were likewise profiles in the MS and VDB was quite pronounced for bothdetected in the HIP and FC of SAP, but not IBO-treated doses of SAP, with very few ChAT positive profilesanimals (P,0.001; Table 2). detected throughout the MS or VDB. This is consistent
Correlations between DMP acquisition and ChAT activi- with the substantial decreases in ChAT and AChE activityty are shown in Fig. 5. Spearman rank order analysis detected in the HIP and FC, after injecting the higher doserevealed a significant negative correlation between acquisi- of SAP.tion (days to criterion) and ChAT activity in the HIP Recently, the selectivity of SAP for cholinergic neurons(P,0.01) and FC (P,0.005). in the MS has been questioned. High doses of SAP
injected into the lateral ventricles reportedly leads to lossof non-cholinergic p75NTR-expressing cells located else-
3. Discussion where in the brain, including Purkinje cells in the cere-bellum [27]. Many of these distal effects can be avoided by
3.1. Histology injecting small doses of SAP directly into the MS [37];however, non-selective effects on GABAergic projection
The results demonstrate that, consistent with previous neurons in the MS remain a concern since GABAergicreports [27], injections of SAP into the MS decrease innervation of the hippocampus plays an important role in
D.A. Johnson et al. / Brain Research 943 (2002) 132 –141 137
Fig. 4. Effects of saline, 192 IgG-saporin (Sap; 1.0 mg, 0.22 mg) and ibotenate (Ibo; 5 mg) on ChAT and PARV staining in the MS. Note the loss ofChAT-positive profiles in the SAP-treated animals. A slight decrease in PARV staining is apparent following administration of the highest dose, but not thelowest dose, of SAP. In contrast, note the loss of PARV staining and the presence of some remaining ChAT-positive profiles in the Ibo-treated animal.Arrows denote the borders of staining in the MS. Scale bar50.5 mm in all panels.
Table 1Effects of treatment on levels of ChAT activity in different brain regions
Tx/Brain region MS FC HIP CB
Control 3.0560.22 1.5460.09 1.6660.10 0.2660.03(n56) (n57) (n57) (n57)
ANOVA F[2,16]50.1, P50.94 F[2,17]552.5, P,0.00001 F[2,17]515.4, P,0.0002 F[2,17]51.3, P50.30† ‡Values are mean6S.E.M. (mmol /h /mg protein). *P,0.005, **P,0.0005 relative to controls; P,0.05, P,0.0005 relative to IBO-treated group.
138 D.A. Johnson et al. / Brain Research 943 (2002) 132 –141
Fig. 5. Correlations between DMP acquisition and ChAT activity. Spearman rank order analysis revealed a significant negative correlation betweenacquisition (days to criterion) and ChAT activity in the HIP (P,0.01) and FC (P,0.005).
regulating hippocampal function [29,35]. Parvalbumin is a the effectiveness of using SAP to selectively destroycalcium binding protein that serves as a marker for cholinergic neurons in the MS, and reinforce the necessityGABAergic projection neurons in the MS. In the present of performing preliminary dose range finding studies whenstudy, some loss of parvalbumin staining was observed in using SAP in order to ensure selectivity as well as efficacy.the MS of animals treated with the higher dose (1 mg) of As expected, injection of the neurotoxin IBO into theSAP, suggesting some impairment of GABAergic projec- MS resulted in the loss of both ChAT-positive neurons andtions. No apparent loss of PARV staining was observed in PARV staining, indicating a loss of both cholinergic andanimals treated with the 0.22 mg dose of SAP, suggesting GABAergic neurons from that structure. In addition, IBOthat non-selective effects of SAP on GABAergic projection appeared to produce more necrotic injury around theneurons in the MS are dose-dependent. Furthermore, no injection site than did SAP. The loss of PARV staining wasapparent effect of SAP injected directly into the MS on quite pronounced, with very few PARV-positive profilescerebellar Purkinje cells was observed. These results verify detected throughout the MS of IBO-treated animals. This
D.A. Johnson et al. / Brain Research 943 (2002) 132 –141 139
suggests that IBO produced a substantial loss of GABAer- demonstrated that IBO can induce memory deficits whengic projection neurons in the MS. In contrast, the loss of injected into the basal forebrain, using doses large enoughChAT-positive cells, as well as non-selective damage to to destroy large numbers of GABAergic and cholinergicthe MS, was most pronounced at the site of injection, and neurons [20,38,41]. Consistent with the current results,was less apparent distal from the injection. As a result, however, a recent study by Pang et al. [34], demonstratedmany ChAT-positive profiles were detected in the MS that when a dose of kainate that damages GABAergicdistal to the injection, suggesting that many cholinergic neurons but spares cholinergic neurons was utilized, thereprojections from the MS to the hippocampus and cortex in was no impairment in performance in Morris water mazeIBO-treated animals remained intact. The data are con- or radial arm maze tasks. The fact that the IBO lesions didsistent with the results of our ChAT assays that showed not significantly impair acquisition of the DMP tasksignificant reductions in ChAT activity in the MS, HIP, and despite the dramatic loss of PARV staining in the MSFC of SAP-treated, but not IBO-treated, animals. The indicates that loss of GABAergic neurons in the MSdifferential effect of IBO on cholinergic vs. GABAergic combined with a moderate loss of cholinergic neurons isneurons in this study is consistent with previous reports not sufficient to significantly impair DMP acquisition. Weshowing a limited loss of cholinergic relative to GABAer- conclude, therefore, that cholinergic neurons in the basalgic neurons following IBO injection into the basal fore- forebrain innervating the hippocampus and frontal cortexbrain [8,18]. play an important role in DMP acquisition, and that the
DMP task can be used to detect impairments in cognitive3.2. Behavior performance associated with deficits in basal forebrain
cholinergic function.The behavioral results show that SAP-treated animals, The specific cognitive processes that are affected by
but not IBO-treated animals, were significantly impaired in SAP lesions in the MS are still unknown. The DMP task isacquisition of the DMP task. Specifically, SAP-treated a spatial task, and rats can use a variety of strategies toanimals took longer to learn the task, and improved at a acquire the task [17]. Rotating the maze 1808 between theslower rate relative to both controls and IBO-treated first and second trials of a trial pair severely disruptsanimals. This was true for both high and low doses of SAP. performance on the task [21], suggesting that the animalsIn addition, the impairment produced by SAP was dose use extramaze spatial cues and spatial working memory torelated, with the higher dose of SAP producing the greater solve the task.impairment. Similar results have been reported by others Several studies, however, have reported limited effects[40,43,47]. For example, Walsh et al. [43] showed that of SAP lesions on measures of spatial working memory inSAP lesions to the MS produced a dose-related impairment rats. For example, Walsh et al. [43], reported a mildin acquisition of a delayed non-matching to sample radial working memory deficit using a delayed non-matching toarm maze task in which animals had to re-learn the sample radial arm maze task following SAP lesions to thelocations of baited and un-baited arms. The fact that these MS. Chappell et al. [12] reported no deficit in spatialeffects were dose-related may reflect more complete working memory in rats trained to perform a radial armdestruction of basal forebrain cholinergic projections at the maze and then retested after administering SAP lesions tohigher dose, particularly projections originating in parts of the MS. Dornan et al. [15] likewise reported no effect ofthe basal forebrain distal to the injection site (e.g., NBM). SAP lesions in the MS, nucleus basalis magnocellularisIn a separate study by Gibbs, intraseptal injection of 1.0 (NBM), or MS1NBM on spatial working memory in themg SAP was shown to produce a significant decrease in Morris water maze. Cahill and Baxter [11] report facili-ChAT activity in both the hippocampus and frontal cortex, tated acquisition of a discrimination task in a plus maze inwhereas intraseptal injection of 0.22 mg SAP decreased rats following SAP lesions to the MS/VDB, suggesting noChAT activity in the hippocampus, but not the frontal deficit in spatial working memory. We have observed nocortex [21]. In the present study, acquisition of the DMP significant difference between SAP or IBO-treated andtask was inversely correlated with ChAT activity in the control animals in performance decrements produced byfrontal cortex as well as the hippocampus. These data are increasing the intertrial delay once animals reach criterionconsistent with the possibility that the greater impairment on the DMP task. These findings suggest that any workingobserved with the higher dose is related to a greater loss of memory deficit produced by selective loss of septalbasal forebrain cholinergic neurons. Alternatively, the fact cholinergic neurons is mild at best, and is not likely tothat the higher dose of SAP produced some loss of PARV account for the deficit in DMP acquisition.staining in the MS raises the possibility that loss of Recent studies suggest that basal forebrain cholinergicGABAergic neurons could also have contributed to the projections may be more involved in attentional processesgreater impairment. This alternative is supported by a than in working memory processes per se. For example,recent study by Pang et al. [34]. Further studies are needed several studies [31,32,42] have reported significant deficitsto differentiate between these possibilities. on tests of sustained visual attention in rats following
The IBO treatment group did not show impairment in selective lesions to cholinergic cells in the NBM. Choliner-acquisition of the DMP task. A number of studies have gic neurons in the MS are less involved in sustained visual
140 D.A. Johnson et al. / Brain Research 943 (2002) 132 –141
attention [20], but may influence the acquisition of spatial bell, Beth E. Baker, Mark A. Liberatore, Payal Aggarwal,tasks by affecting other modes of attention. For example, Douglas Nelson.lesions of cholinergic inputs to the hippocampus reportedlydisrupt reductions in the processing of uninformativestimuli, as revealed by a failure to produce reductions in Referencesconditioned stimulus processing during a test of latentinhibition [6,7]. [1] M. Alreja, M. Wu, W. Liu, J.B. Atkins, C. Leranth, M. Shanabrough,
Muscarinic tone sustains impulse flow in the septohippocampalRats also have a natural tendency to alternate (i.e.,GABA but not cholinergic pathway: implication for learning andspontaneous alternation), which affects performance onmemory, J. Neurosci. 20 (2000) 8103–8110.
spatial alternation tasks [16,33]. As a result, it is often [2] D.M. Araujo, P.A. Lapchak, Y. Robitaille, S. Gauthier, R. Quirion,difficult when using a simple alternation paradigm to Differential alteration of various cholinergic markers in cortical anddetect changes in the rate of acquisition, or to distinguish sub-cortical regions of the human brain in Alzheimer’s disease, J.
Neurochem. 50 (1988) 1914–1923.between effects on rate of acquisition vs. overall per-[3] R.T. Bartus, R.L. Dean, B. Beer, A.S. Lippa, The cholinergicformance. For example, Wenk et al. [45] reported that SAP
hypothesis of geriatric memory dysfunction, Science 217 (1982)lesions to the NBM in rats had no effect on acquisition of a 408–417.T-maze reinforced alternation task. Inspection of the [4] R.T. Bartus, C. Flicker, R.L. Dean, M. Pontecorvo, J.C. Figueiredo,learning curves, however, reveal that the performance of S.K. Fisher, Selective memory loss following nucleus basalis
lesions: long term behavioral recovery despite persistent cholinergiccontrol animals was near criterion at the start of training,deficiencies, Pharmacol. Biochem. Behav. 23 (1985) 125–135.making any effect on learning difficult to detect. By
[5] M.G. Baxter, D.J. Bucci, L.K. Gorman, R.G. Wiley, M. Gallagher,employing a matching to position paradigm, the DMP Selective immunotoxic lesions of basal forebrain cholinergic cells:protocol counters the animals’ natural tendency to alter- effects on learning and memory in rats, Behav. Neurol. 109 (1995)nate, thereby increasing the initial difficulty of the task and 714–722.
[6] M.G. Baxter, D.J. Bucci, P.C. Holland, M. Gallagher, Impairmentsthe sensitivity of the paradigm for assessing acquisition.in conditioned stimulus processing and conditioned responding afterEvidence of the effect of the animals’ natural tendency tocombined selective removal of hippocampal and neocortical
alternate is apparent on the first training trial in which cholinergic input, Behav. Neurosci. 113 (1999) 486–495.control animals performed below chance. The fact that [7] M.G. Baxter, P.C. Holland, M. Gallagher, Disruption of decrementstreatment with either SAP or IBO abolished this tendency in conditioned stimulus processing by selective removal of hip-
pocampal cholinergic input, J. Neurosci. 17 (1997) 5230–5236.suggests that these lesions reduce the animals’ natural[8] I. Bednar, X. Zhang, R. Dastrani-Sedghi, A. Nordberg, Differentialtendency to alternate. Subsequent to day 1, all groups
changes of nicotinic receptors in the rat brain following ibotenicimproved with training and the performance of controls acid and 192-IgG saporin lesions of the nucleus basalis magnocel-quickly surpassed that of the SAP-treated animals. This lularis, Int. J. Dev. Neurosci. 16 (1998) 661–668.demonstrates that despite the better performance of [9] A.A. Book, R.G. Wiley, J.B. Schweitzer, Specificity of 192-IgG-
saporin for NGF receptor-positive cholinergic basal forebrain neu-lesioned animals vs. controls at the start of training,rons in the rat, Brain Res. 590 (1992) 350–355.animals with SAP lesions, but not IBO lesions, were
[10] M. Bradford, A rapid and sensitive method for the quantification ofsignificantly impaired with respect to the rate of improve- microgram quantities of protein utilizing the principal of protein dyement on the task. We conclude, therefore, that selective binding, Analyt. Biochem. 72 (1976) 248–253.destruction of cholinergic neurons in the MS impairs [11] J.F. Cahill, M.G. Baxter, Cholinergic and noncholinergic septal
neurons modulate strategy selection in spatial learning, Eur. J.acquisition of the DMP task.Neurosci. 14 (2001) 1856–1864.In conclusion, we propose that the DMP task provides a
[12] J. Chappell, R. McMahan, A. Chiba, M. Gallagher, A re-examina-sensitive behavioral assay for revealing selective deficits in tion of the role of basal forebrain cholinergic neurons in spatialthe functional status of basal forebrain cholinergic projec- working memory, Neuropharmacology 37 (1998) 481–487.tions. Further work is needed to identify the specific [13] J.T. Coyle, D.L. Price, M.R. DeLong, Alzheimer’s disease: a
disorder of cortical cholinergic innvervation, Science 219 (1983)subsets of basal forebrain cholinergic neurons that are most1184–1190.critical for this task, as well as the specific cognitive
[14] S.T. DeKosky, S.W. Scheff, W.R. Markesbery, Laminar organizationprocesses that are affected and which underlie the deficits of cholinergic circuits in human frontal cortex in Alzheimer’sin acquisition observed. This should provide a useful disease and aging, Neurology 35 (1985) 1425–1431.model for studying the role of basal forebrain cholinergic [15] W.A. Dornan, A.R. McCampbell, G.P. Tinkler, L.J. Hickman, A.W.
Bannon, M.W. Decker, K.L. Gunther, Comparison of site–specificprojections in cognitive processes and for assessing theinjections into the basal forebrain on water maze and radial armability of drugs to reduce cognitive deficits caused by basalperformance in the male rat after immunolesioning with 192-IgG-
forebrain cholinergic impairment associated with aging and saporin, Behav. Brain Res. 82 (1996) 93–101.disease. [16] R.J. Douglas, A.C. Raphelson, Spontaneous alternation and septal
lesions, J. Comp. Physiol. Psychol. 62 (1966) 320–322.[17] P.A. Dudchenko, How do animals actually solve the T maze?,
cholinergic system: interpreting the functional consequences ofGrants: NIH AG16261-01 (DAJ) and NSF IBN- excitotoxic lesions, Trends Neurosci. 14 (1991) 494–501.
9905676 (RBG). Technical Assistance: Ronald J. Camp- [19] G.L. Ellman, K.D. Courtney, V.A. Andres Jr., R.M. Featherstone, A
D.A. Johnson et al. / Brain Research 943 (2002) 132 –141 141
new and rapid colorimetric determination of acetylcholinesterase [34] K.C.H. Pang, R. Nocera, A.J. Secor, R.M. Yoder, GABAergicactivity, Biochem. Pharmacol. 7 (1961) 88–95. septohippocampal neurons are not necessary for spatial memory,
[20] B.J. Everitt, T.W. Robins, Central cholinergic systems and cognition, Hippocampus 11 (2001) 814–827.Ann. Rev. Psychol. 48 (1997) 649–684. [35] O. Paulsen, E. Moser, A model of hippocampal memory encoding
[21] R.B. Gibbs, Basal forebrain cholinergic neurons are necessary for and retrieval: GABAergic control of synaptic plasticity, Trendsestrogen to enhance acquisition of a delayed matching to position Neurosci. 21 (1998) 273–278.T-maze task, Horm. Behav. (2002) in press. [36] G. Paxinos, C. Watson, The Rat Brain in Stereotaxic Atlas, 2nd
[22] R.B. Gibbs, Estrogen replacement enhances acquisition of a spatial Edition, Academic Press, London, 1986.memory task and reduces deficits associated with hippocampal [37] D.P. Pizzo, J.J. Waite, L.J. Thal, J. Winkler, Intraparenchymalmuscarinic receptor inhibition, Horm. Behav. 36 (1999) 222–233. infusions of 192 IgG-saporin: development of a method for selective
[23] R.B. Gibbs, Effects of gonadal hormone replacement on measures of and discrete lesioning of cholinergic basal forebrain nuclei, J.basal forebrain cholinergic function, Neuroscience 101 (2000) 931– Neurosci. Methods 91 (1999) 9–19.938. [38] R. Schliebs, Basal forebrain cholinergic dysfunction—experimental
[24] R.B. Gibbs, Long-term treatment with estrogen and progesterone approaches and the diseased brain, Int. J. Dev. Neurosci. 16 (1998)enhances acquisition of a spatial memory task by ovariectomized 591–603.aged rats, Neurobiol. Aging 21 (2000) 107–116. [39] A.W. Still, Spontaneous alternation and exploration in rats, Nature
[25] R.B. Gibbs, C. Martynowski, Nerve growth factor induces Fos-like 210 (1966) 657–658.immunoreactivity within identified cholinergic neurons in the adult [40] J.J. Waite, A.D. Chen, M.L. Wardlow, R.G. Wiley, D.A. Lappi, L.J.rat basal forebrain, Brain Res. 753 (1997) 141–151. Thal, 192 Immunoglobulin G-saporin produces graded behavioral
[26] R.B. Gibbs, D.W. Pfaff, Effects of estrogen and fimbria / fornix and biochemical changes accompanying the loss of cholinergictransection on p75NGFR and ChAT expression in the medial neurons of the basal forebrain and cerebellar Purkinje cells, Neuro-septum and diagonal band of Broca, Exp. Neurol. 116 (1992) science 65 (1995) 463–476.23–39. [41] J.J. Waite, L.J. Thal, Lesions of the cholinergic nuclei in the rat
[27] S. Heckers, T. Ohtake, R.G. Wiley, D.A. Lappi, C. Geula, M.-M. basal forebrain: excitotoxins vs. an immunotoxin, Life Sci. 58Mesulam, Complete and selective cholinergic denervation of rat (1996) 1947–1953.neocortex and hippocampus but not amygdala by an immunotoxin [42] J.J. Waite, M.L. Wardlow, A.E. Power, Deficit in selective andagainst the p75NGF receptor, J. Neurosci. 14 (1994) 1271–1289. divided attention associated with cholinergic basal forebrain im-
[28] G. Leanza, O.G. Nilsson, R.G. Wiley, A. Bjorklund, Selective munotoxic lesion produced by 192-saporin; motoric / sensory deficitlesioning of the basal forebrain cholinergic system by intraventricu- associated with Purkinje cell immunotoxic lesion produced by OX7-lar 192 IgG-saporin: behavioural, biochemical, and stereological saporin, Neurobiol. Learn Mem. 71 (1999) 325–352.studies in the rat, Eur. J. Neurosci. 7 (1995) 329–343. [43] U. Walsh, C.D. Herzog, C. Gandhi, R.W. Stackman, R.G. Wiley,
[29] N.A. Manuel, C.H. Davies, Pharmacological modulation of Injection of 192 IgG-saporin into the medial septum producesGABA(A) receptor-mediated postsynaptic potentials in the CA1 cholinergic hypofunction and dose-dependent working memoryregion of the rat hippocampus, Br. J. Pharmacol. 125 (1998) deficits, Brain Res. 726 (1996) 69–79.1529–1542. [44] G.L. Wenk, C.A. Harrington, D.A. Tucker, N.E. Rance, L.C. Walker,
[30] M. McLaughlin, M. Inglis, B.M. Ross, K.C. Breen, J. McCulloch, Basal forebrain neurons and memory: a biochemical, histological,Modest cholinergic deafferentation fails to alter hippocampal G- and behavioral study to differential vulnerability to ibotenate andproteins, Neurochem. Int. 35 (1999) 59–64. quisqualate, Behav. Neurosci. 106 (1992) 909–923.
[31] J. McGaughy, M. Sarter, Effects of ovariectomy, 192 IgG-saporin- [45] G.L. Wenk, J.D. Stoehr, G. Quintana, S. Mobley, R.G. Wiley,induced cortical cholinergic deafferentation, and administration of Behavioral, biochemical, histological and electrophysiological ef-estradiol on sustained attention performance in rats, Behav. Neuro- fects of 192 IgG-saporin injections into the basal forebrain of rats, J.sci. 113 (1999) 1216–1232. Neurosci. 14 (1994) 5986–5995.
[32] J.L. Muir, B.J. Everitt, T.W. Robbins, AMPA-induced excitotoxic [46] R.G. Wiley, T.N. Oeltmann, D.A. Lappi, Immunolesioning: Selec-lesions of the basal forebrain: a significant role for the cortical tive destruction of neurons using immunotoxin to rat NGF receptor,cholinergic system in attentional function, J. Neurosci. 14 (1994) Brain Res. 562 (1991) 149–153.2313–2326. [47] C.C. Wrenn, D.A. Lappi, R.G. Wiley, Threshold relationship be-
[33] H. Ohta, X.H. Ni, K. Matsumoto, H. Watanabe, Working memory tween lesion extent of the cholinergic basal forebrain in the rat anddeficit in aged rats in delayed nonmatching to position task and working memory impairment in the radial maze, Brain Res. 847effect of physostigmine on performance of young and aged rats, Jpn. (1999) 284–298.J. Pharmacol. 56 (1991) 303–309.
