Regional distribution of nicotinic receptor subunit mRNAs in human brain: comparison between Alzheimer and normal brain

Ž .Molecular Brain Research 66 1999 94–103

Research report

Regional distribution of nicotinic receptor subunit mRNAs in humanbrain: comparison between Alzheimer and normal brain

Ewa Hellstrom-Lindahl a,), Malahat Mousavi a, Xiao Zhang a, Ritva Ravid c,¨Agneta Nordberg a,b

a Department of Clinical Neuroscience, Occupational Therapy and Elderly Care Research, DiÕision of Molecular Neuropharmacology, KarolinskaInstitute, Huddinge UniÕersity Hospital, S-14186 Huddinge, Sweden

b Geriatric Clinic, Huddinge UniÕersity Hospital, S-14186 Huddinge, Swedenc Netherlands Brain Bank, Meibergdreef 33, 1105 AZ Amsterdam ZO, Netherlands

Accepted 5 January 1999

Abstract

Ž .The regional expression of mRNA for the nicotinic acetylcholine receptor nAChR subunits a3, a4 and a7 was examined inŽ .postmortem brain tissues from controls and patients with Alzheimer’s disease AD by using quantitative RT-PCR. In parallel, the

numbers of nAChRs were measured by receptor binding. Relative quantification of the nAChR gene transcripts in control brains showedthat expression of a3 was highest in the parietal cortex, frontal cortex and hippocampus, and lower in the temporal cortex andcerebellum. The highest level of a4 mRNA was found in the temporal cortex and cerebellum, while a7 mRNA was equally distributed inall brain regions except for hippocampus where it was less abundant. In comparison with AD brains, no differences in the expression ofa3 and a4 in the temporal cortex, hippocampus and cerebellum were found. The level of a7 mRNA was significantly higher in the

w3 x w3 x w125 xhippocampus of AD brains compared to controls. The binding sites for H epibatidine and H nicotine in the temporal cortex and Iw3 xa-bungarotoxin in hippocampus were significantly decreased in AD patients compared to controls. Saturation analysis of H epibatidine

binding revealed two classes of binding sites, with a significant reduction of the higher affinity epibatidine binding sites in the temporalcortex of AD brain. The results show that there is a regional distribution of the expression of the different nAChRs subunits in humanbrain. The findings that the a3 and a4 mRNA levels were not changed in AD brains suggest that the loss of higher affinity epibatidinebinding sites observed in AD patients cannot be attributed to alternations at the transcriptional level of the a3 and a4 genes and thatcauses have to be searched for at the translational andror posttranslational level. The increased mRNA level of a7 previously found inlymphocytes, and now also in the hippocampus of AD patients, indicate that subunit specific changes in gene expression of nAChRs isassociated with AD. q 1999 Elsevier Science B.V. All rights reserved.

Keywords: Nicotinic acetylcholine receptor; Alzheimers’s disease; Human brain; mRNA; RT-PCR

1. Introduction

Ž .Neuronal nicotinic acetylcholine receptors nAChRsare ligand-gated ion channels comprised of a and b

Ž . Ž .subunits. Presently six a a2–a7 and three b b2–b4subunits have been identified and cloned from human

w xbrain 7,11,23,38 . Numerous combinations of a and b

) Corresponding author. Department of Clinical Neuroscience, Occupa-tional Therapy and Elderly Care Research, Division of Molecular Neu-ropharmacology, Huddinge University Hospital, B 84, S-14186 Hud-dinge, Sweden. Fax: q46-8-6899210; E-mail:[email protected]

subunits, and in the case of a7, the a subunits alone, cangenerate many subtypes of nAChRs with different func-tional and pharmacological properties. Precise subunitcomposition of most subtypes is still unclear and the roleof the many potential nAChR subtypes in brain remains tobe fully elucidated. Several distinct classes of nAChRs canbe distinguished in brain by radioligand binding studies.The most abundant nAChR subtype, which accounts formost of the high affinity nicotine binding sites, is made upof the a4 and b2 subunits and has high affinity for

Ž .epibatidine, y -nicotine and cytisine and low affinity fora-bungarotoxin. Epibatidine has been shown to be themost potent nicotine agonist yet reported and has alsohigh-affinity to the a3 subtype to which nicotine shows

0169-328Xr99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.Ž .PII: S0169-328X 99 00030-3

( )E. Hellstrom-Lindahl et al.rMolecular Brain Research 66 1999 94–103¨ 95

w xlower affinity 9,10,15,16 . The a7 containing nAChRsaccount for most of the high affinity a-bungarotoxin bind-ing.

The nAChRs are widely distributed in human brainw x Ž .4,11,37 . In Alzheimer’s disease AD , cholinergic deficitsare one of the most characteristic neurochemical features,and the cholinergic system in brain is the one that corre-

w xlates best with cognitive function 31 . A reduction ofnAChRs in the cerebral cortex and hippocampus of ADbrains has been reported both in vitro, with post mortem

w xbrain tissue 8,27,34,36,43,46 and in vivo with positronw xemission tomography 30,33 . It is not exactly known

which nAChR subtypes are affected in the aging processor in neurodegenerative diseases such as AD and Parkin-son’s disease. An increased understanding of the underly-ing mechanism leading to losses of nAChRs in AD andknowledge of which nAChR subtypes are most affectedmay lead to the development of more effective subtypespecific nicotinic agonists and possible treatment strate-gies.

Only a few studies have specifically compared theexpression of the different nAChR subunit mRNAs and thelocalization of nicotine binding sites in human brainw x2,14,37 . Moreover, no studies investigating the changesof nAChRs at the transcriptional level in AD subjects havebeen reported. In the present study the regional expressionof a3, a4 and a7 mRNA was examined in post mortemhuman brain tissue by quantitative reverse transcriptase-

Ž .polymerase chain reaction RT-PCR and compared withthe binding of three different nicotinic ligands, with partic-ular emphasis on comparisons with AD brains. Sincemarked losses of nAChRs consistently have been observedin cortical areas and hippocampus from AD brains, but not

w xin cerebellum 31 , the expression of nAChR mRNA wereinvestigated in these regions.

2. Materials and methods

2.1. Human brain postmortem tissue

Brain tissues were obtained from the Netherlands BrainBank. Samples were taken from a group of 11 clinicallydiagnosed and neuropathologically confirmed AD cases,and from 8 age-matched controls who had no history ofprimary neurological or psychiatric disorders. Table 1summarizes the postmortem data for the brains used in ourstudy. The clinical diagnosis of ‘probably AD’ was based

w xon the NINCDS-ADRDA criteria 24 . All cases wereneuropathologically confirmed using conventionalhistopathological stainings including Bodian, Hema-toxylinrEosin, Congo Red and ALZ-50. The neuropatho-logical diagnosis was based on a combination of the Cerad

w x w xCriteria 26 and the Braak staging 1 . The cortical tissuesused, medial temporal gyrus, medial frontal gyrus andsuperior parietal gyrus, always were pieces exactly fromthe same area.

Table 1Case history data of patients and control subjects

Controls AD

Ž .Brains n 8 11Ž .Sex MrF 3r5 7r4

Ž .Age at death years 74"4 78"4Range 54–84 54–98

Ž .Postmortem delay h 12"4 9"3Range 5–42 3–31Postmortem delay )24 h 1 1Tissue pH 6.73"0.19 6.64"0.29

Ž .Brain weight g 1284"48 1102"64

Cause of deathMyocardial infarct 2 –Cardiovascular insufficiency 1 2Pneumonia – 4Respiratory insufficiency 1 –

Ž .Bleeding artery carotis comm. 1 –Carcinoma 3 1Cachexia – 2Dehydration – 1Kidney insufficiency – 1

SeÕerity of diseaseAdvanced – 5Severe – 6

Ž .Duration of disease years – 6.6"1.3

2.2. RNA extraction and RT-PCR

Total RNA was isolated from the temporal, parietal andfrontal cortex, the hippocampus and cerebellum of controlbrains and from the temporal cortex, hippocampus andcerebellum of AD brains by a guanidinium thiocyanate-

w xphenol-chloroform extraction method 3 . Traces of con-taminating genomic DNA were removed in all RNA sam-ples by treatment with 3 units of RQ RNase-free DNaseŽ .Promega for 30 min at 378C. Reverse transcription wasperformed from 1 mg total RNA with reagents from the

Ž .Gene-Amp RNA PCR kit Perkin Elmer Cetus . The 20 mlreaction mixture consisted of 5 mM MgCl , 2 ml 10=2

PCR buffer II, 1 mM of all four deoxynucleotide triphos-Ž .phates dNTPs , 20 units RNase inhibitor, 2.5 mM random

hexamers, 50 units M-MLV reverse transcriptase and 1 mlsterile water. The reaction mixture was incubated in a

Ž .Mastercycler 5330 Eppendorf, Germany at 428C for 30min then 958C for 5 min and cooled. The PCR was carriedout in a reaction volume of 50 ml using a master mixcontaining 2 mM MgCl , 1= PCR buffer, 0.2 mM dNTPs,2

0.5 mM each of the forward and reverse primers, 0.5 mlŽ .Taq DNA polymerase Promega , and 5 ml of RT reaction

mixture. After an initial denaturation at 948C for 3 min, 30cycles were performed consisting of 948C for 30 s, 528CŽ . Ž .a7 primers or 608C a3, a4 and cyclophilin for 30 sand 728C for 30 s followed by a final extension at 728C for7 min. To increase the yield of the a4 product the Expand

Ž .High Fidelity PCR system Boehringer Mannheim was

( )E. Hellstrom-Lindahl et al.rMolecular Brain Research 66 1999 94–103¨96

used and the organic solvent DMSO was added to a finalconcentration of 2.5%. The oligonucleotide primers usedfor PCR were designed and synthesized as previously

w xdescribed 13 .Aliquots of the DNA samples were examined by elec-

trophoresis on 1.5% agarose gels containing 0.5 mgrmlethidium bromide. Control reactions were carried out withsamples without reverse transcription to cDNA to ensurethat the detected product was not the result of genomicDNA contamination. No products were detected from thesenegative controls or when PCR amplification was per-formed with sterile water instead of cDNA. The PCRproducts for a3, a4 and a7, subunits were purified using

Ž .QIAquick PCR purification kit Qiagen, Germany andsequenced using ABI PRISM Dye Termination Cycle Se-

Ž .quencing Ready Reaction Kit Perkin Elmer .

2.3. RelatiÕe quantification of a3, a4, and a7 subunitmRNAs

Relative quantification of nAChR subunit mRNAs wasw xperformed as earlier described 13,14 . Briefly, cDNA

corresponding to 20 ng of reverse transcribed RNA wassubjected to 22 cycles of amplification. PCR was per-

Žformed simultaneously from a single master mix but in.different tubes with the primers for a3, a4 or a7 and the

control cyclophilin primers. The amplified products werenon-radioactive labeled by incorporation of digoxigenin-

Ž . Ž .11-dUTP dig-dUTP Boehringer Mannheim directly intothe PCR product. After PCR, the dig-labeled PCR productswere separated on 1.5% agarose gel. The gels were denatu-rated, neutralized, the DNA transferred to a nylon filterŽ .Nytran 13, Schleicher and Schuell and crosslinked underUV light. Chemiluminescent detection was carried outwith a DIG Luminescent Detection Kit following the

Ž .manufacturer’s instructions Boehringer Mannheim . PCRproducts were detected by multiple film exposures rangingfrom 5–60 min. Scanned images were quantitated using

Žthe public domain NIH image program written by W.Rasband at the U.S. National Institutes of Health andavailable from the Internet by anonymous ftp from

.zippy.nimh.nih.gov . The a3, a4 and a7 signals werenormalized to their respective cyclophilin signal and ex-pressed as a ratio of subunit mRNA to cyclophilin mRNAwhen comparing levels of a specific mRNA subunit indifferent brain regions. The results from densitometricanalysis of mRNA expression in control and AD brains arereported, after normalization to the cyclophilin signal, as

Ž .percentage of the mean value of controls 100% obtainedfor the specific mRNA subunit and brain region analyzed.The results were evaluated following at least two indepen-dent experiments using the same RNA samples with simi-lar results obtained on both occasions. Groups used forstatistical comparison were always determined within thesame blot.

2.4. Membrane preparation

Tissues were homogenized in 20 volumes of 0.32 Msucrose solution and centrifuged at 1000=g for 10 min at48C. The pellet was discarded and the supernatant cen-trifuged at 17,000=g for 30 min at 48C. After one washthe new pellet was resuspended and homogenized in 20volumes of 50 mM Tris–HCl buffer, pH 7.4. The proteincontent was measured according to the method of Lowry

w xet al. 20 .

2.5. Binding assays

Ž .Membrane fractions 0.2 mg protein were incubated inŽ . Ž . w3 x50 mM Tris buffer pH 8.0 with 5 nM y - H nicotine

Ž .specific activity 64.4 Cirmmol, Du PontrNEN, USA inw3 xa total volume of 1 ml at 48C for 60 min. The H

epibatidine binding assay was performed by incubatingŽ . Ž .membrane fractions 0.2 mg protein with 0.02 nM " -

w3 x ŽH epibatidine specific activity 54.6 Cirmmol, Du. ŽPontrNEN in 50 mM Tris–HCl buffer pH 7.4; 120 mM

.NaCl, 5 mM KCl, 1 mM MgCl , 2.5 mM CaCl in a total2 2

volume of 1 ml at 258C for 3 h. Non-specific binding wasy4 Ž .determined in the presence of 10 M y -nicotine. After

incubation the samples were filtered through WhatmanGFrC glass filters presoaked with 0.3% polyetylenaminesolution. The filters were washed twice with 3 ml coldbuffer and the filters were transfered to vials containing 5ml scintillation fluid and counted in a liquid scintillationcounter. a-Bungarotoxin binding was performed by incu-

Ž . w125 xbating membranes 0.2 mg with 3.5 nM I a-Žbungarotoxin specific activity 128.8 Cir mmol, Du

. ŽPontrNEN in phosphate buffer pH 7.4; 10 mM.Na HPO , 10 mM NaH PO , 50 mM NaCl, 0.1% BSA2 4 2 4

in a total volume of 0.2 ml for 30 min at 378C. Theincubation was stopped by adding 1 ml cold buffer and thesamples were centrifuged for 17,000=g for 5 min at 48C.The pellet was washed twice with 1 ml cold buffer andcentrifuged as above. The bottoms of the tubes were cutoff and placed in scintillation vials and counted. Non-

w125 x Ž .specific binding of I a-bungarotoxin 71"2%, ns15was determined by preincubating the samples with 10y6

M a-bungarotoxin for 30 min at 378C before addition ofw125 xI a-bungarotoxin. Saturation binding analyses using

w3 x Ždifferent concentrations of H epibatidine 0.006 nM–3.nM were performed on membranes prepared from tempo-

ral cortices from control and AD brains. All assays wereperformed in duplicates or triplicates for each sample.

2.6. Data analyses

Saturation binding data were analyzed by nonlinearŽregression analyses GraphPad PRISM, version 2.0,

.GraphPad Software . The data was fitted to both one-site

( )E. Hellstrom-Lindahl et al.rMolecular Brain Research 66 1999 94–103¨ 97

Fig. 1. Ethidium bromide-stained agarose gel showing PCR productsamplified from temporal cortex of control brain with primers specific forthe a3, a4 and a7 nicotinic receptor subunits and cyclophilin. TotalRNA from temporal cortex was reverse transcribed to cDNA and thenamplified for 30 cycles by PCR. Negative control PCRs performed onRNA when the reverse transcriptase step was omitted are shown in lanes

Ž .2, 4, 6 and 8. A DNA size marker 100 bp DNA ladder is shown in laneŽ .9 M . mRNA for a3 was amplified as 271 bp, a4 as 226 bp, a7 as 257

Ž .bp and cyclophilin cp as 235 bp.

and two-site models. The one-site model was acceptedunless the two-site model gave a statistically better fit of

Ž .the data P-0.05 by F-test . Statistical analysis were

made using one-factor ANOVA followed by Fisher PLSDtest. All values are expressed as mean"S.E.M.

3. Results

3.1. Regional distribution of a3, a4 and a7 nAChRsubunit mRNAs in normal human brain

PCR products for the a3, a4 and a7 subunit mRNAscould be detected in all regions investigated, parietal,frontal and temporal cortex, hippocampus and cerebellum.A representative example, of the PCR products obtainedafter performing RT-PCR on RNA isolated from the tem-poral cortex, is shown in Fig. 1.

The RT-PCR procedure was also used for relative quan-tification of the levels of a3, a4 and a7 mRNAs incontrol brains by incorporation of digoxigenin-11-dUTPdirectly into the PCR product and detected by chemilu-miniscence. Cyclophilin was chosen as internal standardand each sample was normalized on the basis of its

Fig. 2. Expression of nAChR subunit mRNAs in the parietal cortex, frontal cortex, temporal cortex, hippocampus and cerebellum of control brains.Reverse transcribed total RNA was amplified 22 cycles for a3, a4, a7 and cyclophilin in the presence of 0.01 mM digoxigenin-11-dUTP. The dig-labelledPCR products were separated on 1.5% agarose gel, transferred to a nylon membrane and detected by chemiluminiscence. The amounts of specific productsin each sample were normalized for their cyclophilin content. The results are expressed as the ratio of the detected signal for each subunit mRNA to their

Ž .respective cyclophilin signal mean"S.E.M., ns4–7 . Statistical analysis was performed by one-factor ANOVA followed by Fisher PLSD test. P cx,Parietal cortex; F cx, frontal cortex; T cx, temporal cortex; Hip, hippocampus; Cb, cerebellum. ) P-0.05, )) P-0.01, ))) P-0.001, significantlydifferent from parietal cortex, frontal cortex and hippocampus. ))

¶P-0.01, significantly different from the other brain regions.

( )E. Hellstrom-Lindahl et al.rMolecular Brain Research 66 1999 94–103¨98

cyclophilin cDNA content to correct any errors due toRNA quantification or efficiency of cDNA synthesis. Therewas no significant difference in the level of cyclophilin

Ž .mRNA in the five brain regions studied Fig. 2 . Theamount of amplified products was linearly proportional tothe amount of starting template over a range of cDNAconcentrations when subjected to 22 cycles of amplifica-

Ž .tion data not shown .Relative quantification of mRNA for the different

nAChR subunits in parietal cortex, frontal cortex, temporalcortex, hippocampus and cerebellum showed that the levelof mRNA for a3, was evenly distributed in the parietalcortex, frontal cortex and hippocampus. In comparisonwith these three regions the level of a3 mRNA wassignificantly lower in the temporal cortex and cerebellumŽ .P-0.05, Fig. 2 . On the other hand, expression of a4

Ž .was significantly higher in the temporal cortex P-0.001Ž .and cerebellum P-0.01 when compared to parietal

cortex, frontal cortex and cerebellum. a7 mRNA was

almost equally distributed in all regions except for hip-pocampus where the detected level of a7 mRNA wassignificantly lower in comparison to all other brain regions

Ž .studied P-0.01, Fig. 2 .

3.2. Expression of a3, a4 and a7 nAChR subunit mRNAsin Alzheimer brains

Relative quantification of gene transcripts for the a3,a4 and a7 nAChR subunits in the cerebellum, hippocam-pus and temporal cortex showed that there was no differ-ence in the expression of a3 or a4 mRNA between AD

Ž .and control individuals Fig. 3 . A significantly higherlevel of mRNA for a7 was observed in hippocampus of

ŽAD brains in comparison to the control group P-0.01,.Fig. 3 . This significant difference in the a7 mRNA level

was obtained in three independent experiments. The levelsof the internal standard, cyclophilin, were similar in both

Ž .groups Fig. 3 .

Fig. 3. Expression of nAChR subunit mRNAs in cerebellum, hippocampus and temporal cortex of Alzheimer and control brains. Reverse transcribed totalRNA was amplified 22 cycles for a3, a4, a7 and cyclophilin in the presence of 0.01 mM digoxigenin-11-dUTP. The dig-labelled PCR products wereseparated on 1.5% agarose gel, transferred to a nylon membrane and detected by chemiluminiscence. The amounts of specific products in each sample

Ž .were normalized for their cyclophilin content. The results are expressed as the percentage of the mean value of controls 100% for each subunit mRNAŽ .and brain region mean"S.E.M., ns5–6 . Statistical analysis was performed by one-factor ANOVA. T cx, temporal cortex; Hip, hippocampus; Cb,

cerebellum. )) P-0.01, significantly different from control brain.

( )E. Hellstrom-Lindahl et al.rMolecular Brain Research 66 1999 94–103¨ 99

Table 2Changes in nicotinic receptors in brains of Alzheimer patients

Ligand Brain region Controls Alzheimer patients % ChangeŽ . Ž .fmolrmg fmolrmg

3 bw x Ž . Ž .H nicotine Temporal cortex 16.8"3.1 7 9.6"0.4 9 y43Ž . Ž .Hippocampus 14.5"2.8 6 10.3"1.1 10 y29Ž . Ž .Cerebellum 16.7"2.1 8 16.6"1.2 9

3 bw x Ž . Ž .H epibatidine Temporal cortex 9.6"1.9 7 5.4"0.3 8 y44Ž . Ž .Hippocampus 8.0"1.6 6 5.6"0.7 10 y30

aŽ . Ž .Cerebellum 14.2"1.0 8 15.0"1.1 10125w x Ž . Ž .I a-bungarotoxin Temporal cortex 15.5"1.0 7 17.0"1.5 9

bŽ . Ž .Hippocampus 18.2"1.8 4 13.7"0.9 10 y25a bŽ . Ž .Cerebellum 11.0"1.1 8 15.3"1.3 9 q39

w3 x w3 x w125 xMembranes were prepared from each region and assayed in triplicates using 5 nM of H nicotine, 0.02 nM H epibatidine and 3.5 nM of IŽ .a-bungarotoxin. The results are expressed as mean"S.E.M. n .

aSignificantly different from temporal cortex and hippocampus of control group, P-0.05.bSignificantly different from control group, P-0.05.

[3 ] [3 ]3.3. Comparison of specific binding of H nicotine, H[125 ]epibatidine and I a-bungarotoxin in brains of controls

and patients with Alzheimer’s disease

w3 xSingle concentrations of the radioligands H epibati-Ž . w3 x Ž . w125 xdine 0.02 nM , H nicotine 5 nM and I a-

Ž .bungarotoxin 3.5 nM were used to measure distributionof nAChRs in temporal cortex, hippocampus and cerebel-lum and to make comparisons of receptor densities withAD brains. In control brains, the highest specific binding

w3 x Žwith H epibatidine was detected in cerebellum P-. w125 x Ž .0.05 , I a-bungarotoxin in hippocampus P-0.05 ,

w3 xwhereas binding sites for H nicotine were fairly equallyŽ .distributed in these regions Table 2 . Compared to con-

w3 x w3 xtrols, H nicotine and H epibatidine binding was signif-Ž .icantly lower in temporal cortex of AD brains P-0.05 .

Also in the hippocampus of AD brains the number ofw3 x w3 xbinding sites for H nicotine and H epibatidine was

decreased by 30% but did not reach significance. Now3 x w3 xchanges in H nicotine and H epibatidine binding were

detected in the cerebellum of AD brains. The number ofw125 xbinding sites for I a-bungarotoxin was, when com-

Table 3w3 xBinding properties of H epibatidine in human temporal cortex of

Alzheimer and control brains

Ž . Ž .Parameter Controls ns5 AD ns4

Ž .K nM 0.015"0.001 0.013"0.001d1Ž .K nM 1.66"0.36 1.68"0.22d2

aŽ .B fmolrmg protein 8.66"0.41 6.46"0.65max1Ž .B fmolrmg protein 5.25"0.57 4.45"0.41max2

Ž .Hill’s coefficient n 0.76"0.05 0.73"0.01H

Results are means"S.E.M. of independent experiments, each performedin duplicates.Membranes were prepared from temporal cortex as described in Section2. K and B values were calculated from nonlinear regressiond max

analysis. The data were fit to one-site and two-site models. The two-siteŽ .model gave a statistical better fit of the data P -0.05 .

aSignificantly different from controls, P -0.05.

pared with controls, unchanged in the temporal cortex,significantly reduced by 25% in hippocampus and in-

Žcreased by 39% in cerebellum of AD brains P-0.05,.Table 2 .

[3 ]3.4. Analysis of H epibatidine binding parameters

w3 xBinding of H epibatidine to membranes preparedfrom the temporal cortex of control and AD brains wasmeasured and characterized over concentration ranges from

w3 x0.006 nM–3.0 nM. Saturation analysis of H epibatidinebinding in both control and AD brains gave Hill coeffi-

Žcients significantly less than 1 n s0.76 and 0.73, re-H

w3 xFig. 4. H epibatidine saturation curves in membrane preparations fromŽ .temporal cortex of Alzheimer and control brains. A Specific binding of

w3 x Ž .H epibatidine 0.006 nM–3 nM . Data were analysed by non-linearŽ .regression analysis giving best fit to a two-site model. B Scatchard plot

of specific binding data shown in A, with curve fit for the high affinitybinding sites indicated as dashed line for control and solid line for AD.The curves are representative of one of 4–5 individual assays and asummary of values obtained from analysis of all data is given in Table 3.

( )E. Hellstrom-Lindahl et al.rMolecular Brain Research 66 1999 94–103¨100

.spectively, P-0.001 and nonlinear regression analysis ofw3 xbinding data for H epibatidine were best fit to a two

binding sites model, each site comprising about 60 and40% of the total number of binding sites and with a100-fold difference in affinity constants. The estimated Kd

values for controls were approximately 15 pM and 1.7 nMŽ .Table 3 . There was no significant difference between theK values obtained from controls and AD cases. Thed

maximum number of high affinity binding sites, B formax1w3 x ŽH epibatidine was significantly reduced by 25% P-

.0.05 in AD brains compared to controls whereas only a15% decline in numbers of lower affinity sites was ob-

Ž . w3 xserved not significant . A representative example of Hepibatidine saturation curves in membrane preparationsfrom temporal cortex of AD and control brains is shown inFig. 4.

4. Discussion

The expression of a3, a4 and a7 mRNA was exam-ined in several cortical areas, hippocampus and cerebellumof human brain. A regional difference in the pattern ofexpression of nAChR subunit mRNAs was shown in thisstudy. In the temporal cortex and cerebellum the levels ofa4 gene expression were significantly higher compared toother brain regions studied, whereas the lowest expressionof a3 mRNA was detected in these two regions. TheRT-PCR method used here for quantification of mRNA isrelative and therefore does not allow us to compare theabsolute levels of expression of the different nAChR sub-units. Earlier studies of the human brain using in situhybridization have demonstrated strong signals for a3mRNA in thalamus, weak signals in cerebral cortex and

w xvery weak signals in hippocampus 37 . a3 mRNA hasbeen shown to be predominantly present in pyramidalneurons of layers III–VI of the motor cortex, whereastranscripts for the a4 subunit appeared throughout all

w xcortical layers 44 . Laminar distribution of nAChRs incortex using autoradiography have, on the other hand,

w3 x w3 xshown high binding of H nicotine and H epibatidinew xin layers III and V of motor cortex 41 .

In the present study mRNA for a7 was relativelyuniformly distributed in the regions studied except forhippocampus where the expression of a7 was much lessabundant. These findings are not compatible with the

w125 xbinding profile pattern showing the highest density of Ia-bungarotoxin binding sites in the hippocampus. Using insitu hybridization, a7 mRNA has been shown to be local-

w x w xized in human cortex 45 , hippocampus and thalamus 37 .In the hippocampus, one study showed high expression of

w xa7 in the dentate granular layer 37 , whereas anotherstudy reported intensive signals in the hilar neurons of the

w xhippocampus 45 . Recently, a study comparing the re-w125 xgional expression of a7 mRNA and I a-bungarotoxin

binding in human postmortem brain showed high to mod-

erate a7 hybridization signals in the CA1–CA3 cell layersw125 xof hippocampus, but there was very little I a-

w xbungarotoxin binding on these cells 2 . These studiesw125 xshow that a7 mRNA and I a-bungarotoxin binding

sites are not uniformly distributed throughout the entirehippocampus. It is important however to consider the factthat while techniques like autoradiography and in situhybridization permit analysis of binding sites and mRNAwith a fairly good resolution in small discrete brain regionsand cell layers, membrane preparations and RNA isolationfrom rather large tissue sample as performed in the presentstudy only gives an average estimation of the regionstudied. The results obtained for a3, a4 and a7 mRNA inthe present study are in good agreement with our recentfindings from a study comparing expression of a and b

nAChR subunit mRNAs in the frontal cortex and cerebel-w xlum from prenatal and aged human brain 14 . Those brain

w xregions also express a5, b2, b3 and b4 mRNA 14 . ThemRNA levels for a4, a5, a7, b2, and b4 were allsignificantly lower in aged brain compared with the prena-tal brain, which is consistent with the earlier reported lossof nAChRs in the cerebral cortex and hippocampus during

w xnormal aging 4,8,32 .w xIn agreement with earlier studies 22,31,43 significant

w3 x w3 xreductions of H nicotine and H epibatidine bindingsites were observed in the temporal cortex of AD brains in

w3 xcomparison with controls. A significant loss of H nico-tine binding sites has also been detected in the hippocam-

w xpus 22,34,36 and the same results, although not signifi-cant, was found in the present study with a 30% decrease

w3 x w3 xin H nicotine and H epibatidine binding in the hip-pocampus of AD brains. Analysis of saturation curves forw3 xH epibatidine showed a selective reduction of the highaffinity binding sites in the temporal cortex of AD brainswhich is in agreement with earlier findings showing achange in the proportion of high and low affinity nicotine

w xbinding sites in the temporal cortex of AD brains 29,43 .Since the a4b2 receptor subtype is considered to beresponsible for most of the high-affinity nicotine bindingsites the selective loss of high-affinity nicotinic bindingsites in AD brains suggest that the a4b2 receptors might

w xbe a vulnerable nAChR subtype in AD 43 . Epibatidineexhibits high affinity to both a4 and a3 nAChR subunitsw x9,16,35 , while nicotine reveals less affinity to a3 com-pared to the a4b2 nAChR subtype. In the present studynicotine and epibatidine both detected a similar reductionof number of binding sites in temporal cortex suggesting aselective loss of the major high-affinity a4b2 nAChRs inAD brains.

Saturation analysis of 3H-epibatidine binding in tempo-ral cortex revealed two classes of binding sites with Kd

values in the pM and nM range which is consistent withw xearlier published data for human brain 14,15,21 . The Kd

w3 xvalues for the high- and low affinity binding sites for Hepibatidine were unchanged in the temporal cortex of ADbrains compared to age-matched controls. Saturation bind-

( )E. Hellstrom-Lindahl et al.rMolecular Brain Research 66 1999 94–103¨ 101

w3 xing studies with H nicotine was not analyzed in thepresent study since earlier studies have shown unchanged

w3 xaffinity for H nicotine in several regions of AD brainsw x4,43 . The unchanged affinity constant suggests that areduction in the number of high-affinity binding sites isresponsible for the decreased binding observed in brains ofAlzheimer patients.

w3 xThe regional distribution of H epibatidine binding inw xhuman brain 21 corresponds closely to the distribution

w xreported for rat brain 15,25,35 except for the cerebellum.The cerebellum in human brain has shown proportionately

w3 xmany more H epibatidine binding sites compared tow xcerebral cortex 21 , as also shown in the present study,

w xwhereas the opposite was found for the rat brain 15,25 .Since the a7 containing nAChRs account for most of

the high affinity a-bungarotoxin binding we compared thew125 xI a-bungarotoxin binding in the temporal cortex, hip-pocampus and cerebellum of AD brains with controls.

w125 xEarlier studies using I a-bungarotoxin have shown thatw125 xlevels of I a-bungarotoxin binding sites are not re-

w xduced in frontal cortex of AD brains 5,18,42 , whereas aw125 xsignificant loss of about 40% of the I a-bungarotoxin

w x w xbinding 5 , or no reduction 18 has been reported for thetemporal cortex. In the hippocampus the binding sites forw125 xI a-bungarotoxin have been reported to be unchangedw x18,28 . Those findings are in contrast to the present results

w125 xwhich indicate a 25% loss of the I a-bungarotoxinbinding sites in the hippocampus of AD brains. Thisdiscrepancy may be due to methodological differences andselection of the areas of hippocampus used, or severity ofthe disease among the AD cases. We also detected a

w125 xsignificant increase in I a-bungarotoxin binding sitesin the cerebellum of AD brains which as far as we know,

w3 xhas not been reported before. An increased H a-bungarotoxin binding has also been reported in the thala-

w x w3 xmus of AD brains 28 . Interestingly, no change in Hnicotine binding has been reported in the cerebellum and

w xthalamus 31 . The exact functional role of a7 nAChRs inbrain remains unknown. This nAChR subtype has thegreatest Ca2q permeability of any known ligand-gatedchannel and may play a role in the regulation of intra-cellular calcium homeostasis and to modulate release of

w xneurotransmittors such as glutamate 12,23 . Moreover,recent studies indicate that a7 nAChRs may play a neuro-protective role against neurotoxicity of glutamate and b-

w xamyloids 6,17,40 .The present study indicates a general lack of correspon-

dence between levels of nAChR mRNAs and ligand bind-w xing which also have been observed by others 2,19,37,47 .

No significant differences in the expression of the a3 anda4 mRNA were found between AD subjects and controlindividuals in temporal cortex, hippocampus and cerebel-

w3 xlum. These results suggest that the reduced binding of Hw3 xnicotine and H epibatidine observed in the temporal

cortex, and to some lesser extent in the hippocampus ofAD brains, does not strictly correlate to changes of the a3

and a4 mRNA levels in these brain regions. Similarly, aw3 xdecreased H nicotine binding with no changes in the

density of a4-expressing neurons has been reported in thew xfrontal cortex of Parkinson’s disease 39 . The reduced

number of nAChRs may be due to several contributingfactors such as a decline in translation or modification ofthe receptor proteins, impaired receptor assembly, intra-cellular transport andror insertion into the plasma mem-brane. One should also consider that the b-subunits arevery important for receptor assembly but at this point it isunclear whether there are any changes in the expression ofthe b-subunits in AD brain. An alternative explanation isthat the receptor protein is transported to distinct nerveterminals, while the corresponding mRNA is expressed inthe cell somata of another region. Interestingly the a7mRNA was increased compared to controls in hippocam-

w125 xpus of AD brains whereas an decrease in the I a-bungarotoxin binding sites was found. We have also de-tected in lymphocytes from AD patients a significantincrease in a7 mRNA level compared to age-matchedcontrols, whereas the expression of a3, a4, b2 and b4

w xwas decreased 13 . The physiological significance of aselective increase of the a7 mRNA level in both hip-pocampus and lymphocytes in AD patients is still un-known since not much is known about the transcriptionalregulation of the human nAChRs and gene promotors ofcertain subunits.

In conclusion, this study shows that a regional distribu-tion of the expression of the different nAChRs subunitsexists in human brain. The a7 mRNA level was increasedin hippocampus of AD brains whereas the a3 and a4mRNA levels were not changed in any of the investigatedbrain regions, suggesting that the mechanism underlyingthe selective loss of the high-affinity nicotine binding sitesin AD patients cannot be attributed to alterations at thetranscriptional level of the a3 and a4 genes and isprobably more related to yet undefined defects at thetranslational andror posttranslational level.

Acknowledgements

This work was financially supported by grants from theŽ .Swedish Medical Research Council projects no 05817 ,

˚Stiftelsen for Gamla Tjanarinnor, Ake Wiberg’s founda-¨ ¨tion, Loo and Hans Osterman’s foundation, AlzheimerFoundation Sweden, Stohne’s foundation, KI foundationsand Swedish Match.

References

w x1 H. Braak, E. Braak, Evolution of the Neuropathology of AD, ActaNeurol. Scand. Suppl., 1996, 3–12.

w x2 C.R. Breese, C. Adams, J. Logel, C. Drebing, Y. Rollins, M.Barnhart, B. Sullivan, B.K. Demasters, R. Freedman, S. Leonard,

( )E. Hellstrom-Lindahl et al.rMolecular Brain Research 66 1999 94–103¨102

Comparison of the regional expression of nicotinic acetylcholinew125 xreceptor a7 mRNA and I -a-bungarotoxin binding in human

Ž .postmortem brain, J. Comp. Neurol. 387 1997 385–398.w x3 P. Chromczynski, N. Sacchi, Single-step method of RNA isolation

by acid guanidium thiocyanate-phenol-chloroform extraction, Anal.Ž .Biochem. 162 1987 156–159.

w x4 J. Court, F. Clementi, Distribution of nicotinic subtypes in humanŽ . Ž .brain, Alzheimer Dis. Assoc. Disord. 9 2 1995 6–14, Suppl.

w x5 P. Davies, S. Feisullin, Postmortem stability of a-bungarotoxinŽ .binding sites in mouse and human brain, Brain Res. 216 1981

449–454.w x6 D. Donnelly-Roberts, J. Xue, S. Arneric, J. Sullivan, In vitro

Ž . Ž .neuroprotective activator ChCA ABT-418, Brain Res. 719 199636–44.

w x7 K.J. Elliott, S.B. Ellis, K.J. Berckhan, A. Urrutia, L.E. Chavez-Noriega, E.C. Johnson, G. Velicelebi, M.M. Harpold, Comparativestructure of human neuronal a2– a7 and b2–b4 nicotinic acetyl-choline receptor subunits and functional expression of the a2, a3,

Ž .a4, a7, b2 and b4 subunits, J. Mol. Neurosci. 7 1996 217–228.w x w3 x8 D.D. Flynn, D.C. Mash, Characterization of L- H nicotine binding

in human cerebral cortex: comparison between Alzheimer’s diseaseŽ .and the normal, J. Neurochem. 47 1986 1948–1954.

w x9 W. Gerzanich, X. Peng, F. Wang, G. Wells, R. Anand, S. Fletcher,J. Lindstrom, Comparative pharmacology of epibatidine: a potentagonist for neuronal nicotinic acetylcholine receptors, Mol. Pharma-

Ž .col. 48 1995 774–782.w x10 M. Gopalakrishnan, L.M. Monteggia, D.J. Anderson, E.J. Molinari,

M. Piattoni-Kaplan, D. Donelly-Roberts, S.P. Arneric, J.P. Sullivan,Stable expression, pharmacologic properties and regulation of thehuman neuronal nicotinic acetylcholine a4b2 receptor, J. Pharma-

Ž .col. Exp. Ther. 276 1996 289–297.w x11 C. Gotti, D. Fornasari, F. Clementi, Human neuronal nicotinic

Ž .receptors, Progr. Neurobiol. 53 1997 199–237.w x12 R. Gray, A. Rajan, K. Radcliffe, M. Yakehiro, J. Dani, Hippocampal

synaptic transmission enhanced by low concentrations of nicotine,Ž .Nature 383 1996 713–716.

w x13 E. Hellstrom-Lindahl, X. Zhang, A. Nordberg, Expression of nico-¨tinic receptor subunit mRNAs in lymphocytes from normal and

Ž .patients with Alzheimer’s disease, Alzheimer’s Res. 3 1997 29–36.˚w x14 E. Hellstrom-Lindahl, O. Gorbounova, A. Seiger, M. Mousavi, A.¨

Nordberg, Regional distribution of nicotinic receptors during prena-tal development of human brain and spinal cord, Dev. Brain Res.

Ž .108 1998 147–160.w x15 R.A. Houghtling, M.I. Davila-Garcia, K.J. Kellar, Characterization

Ž . w3 xof " - H epibatidine binding to nicotinic cholinergic receptors inŽ .rat and human brain, Mol. Pharmacol. 48 1995 280–287.

w x16 K.J. Kellar, Epibatidine: its pharmacological actions and utility forŽ .studing neuronal nicotinic receptors, Neurotransmissions 11 1995

1–5.w x17 T. Kihara, S. Shimohama, H. Sawada, J. Kimura, T. Kume, H.

Kochiyama, T. Maeda, A. Akaike, Nicotinic receptor stimulationŽ .protects neurons against b amyloid toxicity, Ann. Neurol. 42 1997

159–163.w x18 W. Lang, H. Henke, Cholinergic receptor binding and autoradiog-

raphy in brains of non-neurological and senile dementia ofŽ .Alzheimer-type patients, Brain Res. 267 1983 271–280.

w x19 C. Liu, A. Nordberg, X. Zhang, Differential co-expression of nico-tinic acetylcholine receptor a4 and b2 subunit genes in various

Ž .regions of rat brain, NeuroReport 7 1996 1645–1649.w x20 O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, Protein

measurement with the Folin phenol reagent, J. Biol. Chem. 193Ž .1951 265–275.

w x21 A. Marutle, U. Warpman, N. Bogdanovic, A. Nordberg, Regionaldistribution of subtypes of nicotinic receptors in human brain and

Ž . w3 x Ž .effect of aging studied by q - H epibatidine, Brain Res. 801 1998143–149.

w x22 A. Marutle, U. Warpman, N. Bogdanovic, L. Lannfelt, A. Nordberg,

Deficits in neuronal nicotinic receptors in Alzheimer patients withthe swedish APP 670r671 mutation, J. Neurochem., 1999, in press.

w x23 D.S. McGehee, L.W. Role, Physiological diversity of nicotinicacetylcholine receptors expressed by vertebrate neurons, Annu. Rev.

Ž .Physiol. 57 1995 521–546.w x24 G. Mc Khann, D. Drachman, M. Folstein, R. Katzman, D. Prince,

E.M. Stadlan, Clinical diagnosis of Alzheimer’s disease: report ofNINCDS-ADRDA Work Group under the auspices of department ofhealth and human services task forces on Alzheimer’s disease,

Ž .Neurology 34 1984 939–944.w x25 H. Miao, C. Liu, K. Bishop, Z.-H. Gong, A. Nordberg, X. Zhang,

Nicotine exposure during a critical period of development leads topersistent changes in nicotinic acetylcholine receptors of adult rat

Ž .brain, J. Neurochem. 70 1998 752–762.w x26 S.S. Mirra, A. Heyman, D. Mc Keel, B.J. Crain, L.M. Brownlee,

F.S. Vogel, J.P. Hughes, G. van Belle, L. Berg, The consortium toŽ .establish a registry for Alzheimer’s disease CERAD : Part II.

Standardization of the neuropatholic assessment of Alzheimer’sŽ .disease, Neurology 41 1991 479–486.

w x 327 A. Nordberg, B. Winblad, Reduced number of H-nicotine and3H-acetylcholine binding sites in the frontal cortex of Alzheimer

Ž .brains, Neurosci. Lett. 72 1986 115–119.w x28 A. Nordberg, B. Winblad, Brain nicotinic and muscarinic receptors

in normal aging and dementia, in: A. Fisher, I. Hanin, C. LachmanŽ .Eds. , Alzheimer’s and Parkinson’s Diseases, Plenum, New York1986, pp. 95–108.

w x29 A. Nordberg, A. Adem, J. Hardy, B. Winblad, Change in nicotinicreceptor subtypes in temporal cortex of Alzheimers brains, Neurosci.

Ž .Lett. 86 1988 317–321.w x30 A. Nordberg, P. Hartvig, A. Lilja, M. Viitanen, K. Amberla, H.

Lundqvist, Y. Anderson, L. Ulin, B. Winblad, B. Langstrom, De-˚ ¨creased uptake and binding of 11C-nicotine in brain of Alzheimerpatients as visualized by positron emission tomography, J. Neural

Ž .Transm. 2 1990 215–224.w x31 A. Nordberg, Neuroreceptor changes in Alzheimer disease, Cere-

Ž .brovasc., Brain Metab. Rev. 4 1992 303–328.w x32 A. Nordberg, I. Alafuzoff, B. Winblad, Nicotinic and muscarinic

subtypes in the human brain: changes with aging and dementia, J.Ž .Neurosci. Res. 31 1992 103–111.

w x33 A. Nordberg, P. Hartvig, H. Lundqvist, A. Lilja, B. Langstrom,˚ ¨Ž .Ž .w11 xKinetic analysis of regional S y C nicotine in normal and

Alzheimer brains—in vivo assessment using positron emission to-Ž .mography, Alzheimer Dis. Assoc. Disord. 9 1995 21–27.

w x34 E.K. Perry, C.M. Morris, J.A. Court, A. Cheng, A.F. Fairbairn, I.G.Mc Keith, D. Irving, A. Brown, R.H. Perry, Alternation in nicotinebinding sites in Parkinsons’s disease, Lewy body dementia andAlzheimer’s disease: possible index of early neuropathology, Neuro-

Ž .science 64 1995 385–395.w x w3 x35 E.K. Perry, K.J. Kellar, H epibatidine labels nicotinic receptors in

rat brain: an autoradiographic study, J. Pharmacol. Exp. Ther. 275Ž .1995 1030–1034.

w x36 J.O. Rinne, T. Myllykyla, P. Lonnberg, P. Marjamaki, A post-¨ ¨ ¨mortem study of brain nicotinic receptors in Parkinson’s and

Ž .Alzheimer’s disease, Brain Res. 547 1991 167–170.w x37 F. Rubboli, J.A. Court, C. Sala, C. Morris, B. Chini, E. Perry, F.

Clementi, Distribution of nicotinic receptors in the human hip-Ž .pocampus and thalamus, Eur. J. Neurosci. 6 1994 1596–1604.

w x38 P.B. Sargent, The diversity of neuronal nicotinic acetylcholine recep-Ž .tors, Annu. Rev. Neurosci. 16 1993 403–443.

w x39 H. Schroder, R.A.I. de Vos, E.N.H. Jansen, Ch. Birtsch, A. Wevers,¨Ch. Lobron, Ch.S. Nowacki, R. Schroder, A. Maelicke, Gene ex-¨pression of the nicotinic acetylcholine receptor a4 subunit in thefrontal cortex in Parkinson’s disease patients, Neurosci. Lett. 187Ž .1995 173–176.

w x40 S. Shimohama, Nicotinic agonists prevent neuronal cell death byŽ .glutamate and amyloid beta protein, Neurobiol. Aging 17 1996

S40.

( )E. Hellstrom-Lindahl et al.rMolecular Brain Research 66 1999 94–103¨ 103

w x41 W. Sihver, P.G. Gillberg, A. Nordberg, Laminar distribution ofnicotinic receptor subtypes in human cerebral cortex as determined

w3 xŽ . w3 x w3 xby H y nicotine, H cytisine and H epibatidine in vitro auto-Ž .radiography, Neuroscience 85 1998 1121–1133.

w x42 K. Sugaya, E. Giacobini, E.V.A. Chiappinelli, Nicotinic acetyl-choline receptor subtypes in human frontal cortex: changes in

Ž .Alzheimer’s disease, J. Neurosci. Res. 27 1990 349–359.w x43 U. Warpman, A. Nordberg, Epibatidine and ABT-418 reveal selec-

tive losses of a4b2 nicotinic receptors in Alzheimer brains, Neu-Ž .roReport 6 1995 2419–2423.

w x44 A. Wevers, A. Jeske, Ch. Lobron, Ch. Birtsch, S. Heinemann, A.Maelicke, R. Schroder, H. Schroder, Cellular distribution of nico-¨ ¨tinic acetylcholine receptor subunit mRNAs in the human cerebral

cortex as revealed by non-isotopic in situ hybridization, Mol. BrainŽ .Res. 25 1994 122–128.

w x45 A. Wevers, J.P. Sullivan, T. Giordano, Ch. Birtsch, L.M. Monteggia,S. Nowacki, S. Arneric, H. Schroder, Cellular distribution of the¨mRNA for the a7 subunit of the nicotinic acetylcholine receptor in

Ž .the human cerebral cortex, Drug Dev. Res. 36 1995 103–110.w x46 P.J. Whitehouse, A.M. Martino, P.G. Antuono, P.R. Lowenstein,

J.T. Coyle, D.L. Price, K. Kellar, Nicotinic acetylcholine bindingŽ .sites in Alzheimer’s disease, Brain Res. 371 1986 146–151.

w x47 X. Zhang, Z.-H. Gong, E. Hellstrom-Lindahl, A. Nordberg, Regula-¨tion of a4b2 nicotinic acetylcholine receptors in M10 cells follow-

Ž .ing treatment with nicotinic agents, NeuroReport 6 1995 313–317.