Neuronal Nicotinic Receptor Deficits in Alzheimer Patients with the Swedish Amyloid Precursor Protein 670/671 Mutation


  • Amelia Marutle,

  • Ulrika Warpman,

  • Nenad Bogdanovic,

  • Lars Lannfelt,

  • Agneta Nordberg

  • The present address of Dr. U. Warpman is Departmnet of Pharmacology, Pharmacia & Upjohn, S-751 82 Uppsala, Swden.

  • Abbreviations used : Aβ, β-amyloid : AD, Alzheimer's disease ; APOE, apolipoprotein E ; APP, amyloid precursor protein ; nAChR, nicotinic acetylcholine receptor ; NFT, neurofibrillary tangels ; NP. neuronal plaque.

Address correspondence and reprint requests to Dr. A. Marutle at Department of Neuroscience and Family Medicine, Division of Molecular Neuropharmacology, Karolinska Institute, Huddinge University Hospital, B84, S-141 86 Huddinge, Sweden.


Abstract : The influence of β-amyloid on cholinergic neurotransmission was studied by measuring alterations in nicotinic acetylcholine receptors (nAChRs) in autopsy brain tissue from subjects carrying the Swedish amyloid precursor protein (APP) 670/671 mutation. Significant reductions in numbers of nAChRs were observed in various cortical regions of the Swedish 670/671 APP mutation family subjects (-73 to -87%) as well as in sporadic Alzheimer's disease (AD) cases (-37 to -57%) using the nicotinic agonists [3H]epibatidine and [3H]nicotine, which bind with high affinity to both α3 and α4 and to α4 nAChR subtypes, respectively. Saturation binding studies with [3epibatidine revealed two binding sites in the parietal cortex of AD subjects and controls. A significant decrease in Bmax (-82%) for the high-affinity site was observed in APP 670/671 subjects with no change in KD compared with controls (0.018 nM APP 670/671 ; 0.036 nM control). The highest load of neuronal plaques (NPs) was observed in the parietal cortex of APP 670/671 brains, whereas the number of [3H]nicotine binding sites was less impaired compared with other cortical brain regions. Except for a positive significant correlation between the number of [3H]nicotine binding sites and number of NPs in the parietal cortex, no strict correlation was observed between nAChR deficits and the presence of NPs and neurofibrillary tangles, suggesting that these different processes may be closely related but not strictly dependent on each other.

Alzheimer's disease (AD) is the most common form of dementia and is characterized by neuropathological features such as amyloid plaques, neurofibrillary tangles (NFTs), and disturbances in cholinergic transmitter activity. The etiology of the disease is complex, and mutations in four genes on chromosomes 21, 14, 1, and 12 have so far been identified in families with a history of AD (Dewji and Singer, 1996 ; Lannfelt, 1996 ; Stephenson, 1997). In addition, apolipoprotein E (APOE) has been found to constitute a susceptibility gene, such that the presence of an ε4 allele increases the risk for AD (Strittmatter et al., 1993). A double mutation at codon 670/671 on the amyloid precursor protein (APP) gene on chromosome 21 was described in a Swedish family (Mullan et al., 1992). This mutation is unique in the sense that it is the only AD mutation that has been shown to alter APP metabolism resulting in an overexpression of amyloid and, subsequently, plaque formation. Kidney cells transfected with the Swedish APP 670/671 mutation showed an increased β-amyloid (Aβ) production compared with cells expressing normal APP (Citron et al., 1992). It is therefore assumed that this mismetabolism is causing the disease (Hardy, 1996). AD has been traced through eight generations of this Swedish APP 670/671 mutation family, and the onset of disease occurs between 44 and 61 years of age (Axelman et al., 1994). The diagnosis of AD has been histopathologically verified in four family members who have died (Lannfelt et al., 1994). Longitudinal studies of brain metabolism by positron emission tomography in carriers and noncarriers of the APP 670/671 mutation have shown an impairment of glucose metabolism in cortical brain regions with increasing age, compared with noncarriers, with a pattern similar to that in sporadic AD cases (Nordberg et al., 1995b).

Multiple subtypes of nicotinic cholinergic receptors (nAChRs) exist in the brain (Sargent, 1993), and these receptors are involved in several functional processes, including cognitive function (Levin, 1992 ; Newhouse and Potter, 1994). Since profound losses of nAChRs have consistently been observed in brain tissue of sporadic AD cases (with unknown cause of the disease) (Nordberg et al., 1992), we have chosen to investigate whether APP 670/671 AD cases (with a known mutation in the APP gene) display similar neurochemical changes, i.e., losses of nAChRs, as sporadic AD cases. To characterize changes in subtypes of nAChRs, the radioligands (-)-[3H]nicotine and (±)-[3H]Epibatidine were used in the receptor binding assays. (±)-[3H]Epibatidine binds with high affinity to both α3 and α4β2 nAChR subunits (McKay et al., 1994 ; Gerzanich et al., 1995), whereas nicotine binds with higher affinity to the α4β2 nAChR subtype than the α3 nAChR subunit. In the present study, there was an attempt to correlate deficits in nAChRs to the neuropathological features, e.g., number of neuritic plaques (NPs) and NFTs, in cortical regions of the APP 670/671 brain to obtain a further understanding of the relationship between these various processes and the possible interactive mechanisms.


Human brain postmortem tissue

Brain tissues from 18 sporadic AD subjects (mean ± SEM age, 81.6 ± 1.6 years) and 19 age-matched controls (mean ± SEM, 76.7 ± 1.6 years) were obtained from the Huddinge University Hospital Brain Bank, Huddinge, Sweden (Table 1). Brain tissue from the four AD subjects carrying the Swedish APP 670/671 mutation (mean ± SEM age, 63.3 ± 3.71 years) and nine age-matched controls (mean ± SEM, 64.4 ± 1.6 years) were obtained at specific request from the Huddinge University Hospital Brain Bank Coordinator, because this unique autopsy tissue is not generally available for distribution from the Brain Bank. All AD patients had received the diagnosis “probably AD” according to NINCDS-ADRDA criteria (McKhann et al., 1984), and the diagnosis of definitive AD was confirmed by neuropathological examination using CERAD criteria (Mirra et al., 1991) adjusted for brain banking (Bogdanovic and Morris, 1995). The control brains had no clinical history of psychiatric or neurological disorders and no neuropathological changes indicative of dementia. Postmortem delay was 28.1 ± 4.0 h for the sporadic AD group, 19.6 ± 1.5 h for the control group, 21.7 ± 1.3 h for the Swedish AD family, and 22.0 ± 4.5 h for the age-matched controls. Individual data for the Swedish 670/671 APP-encoded mutation family are shown in Table 2. Patient 1 had been a heavy smoker until 51 years of age, when he was institutionalized. During the following years he only smoked occasionally until his death at 56 years of age. The other patients had no history of smoking.

Table 1. Case details of patients and control subjectsNA, no data available.
CaseAge (years)GenderPostmortem delay (h)APOECause of death
Sporadic AD     
184F93/4Carcinoma of pancreas
279F363/4Myocardial infarct
388F242/4Heart disease
487F493/4Heart disease
679F73/4Gastric bleeding, liver cirrhosis
786M103/3Heart disease
982M323/4Myocardial infarct
1083F603/4Heart disease
1180F124/4Heart disease
12100F413/3Heart disease
1382F243/3Myocardial infarct
1571M293/4Myocardial infarct
1675M243/3Myocardial infarct
1779M17NAHeart disease
Control subjects     
173M143/3Myocardial infarct
270F193/3Respiratory insufficiency
468M24NACarcinoma of prostate
583F183/3Myocardial infarct
669M223/3Myocardial infarct
778F33/3Heart disease
880M163/4Heart disease
971M263/3Myocardial infarct
1085F283/3Carcinoma of pancreas
1182M183/3Heart disease
1279F173/3Heart disease
1374M233/3Myocardial infarct
1465M113/3Myocardial infarct
1582F242/3Carcinoma of intestine
1668M273/3Heart disease
1778M28NAHeart disease
1891F24NACarcinoma of liver
1978M263/4Heart disease
Table 2. Individual data of AD patients carrying the Swedish 670/671 APP mutationNP and NFT data are mean ± SEM numbers of NPs or NFTs per square millimeter.
      Frontal cortexParietal cortexTemporal cortexOccipital cortex
PatientSmokingGenderAge (years)AD duration (years)APOENPsNFTsNPsNFTsNPsNFTsNPsNFTs
P1SmokerM5611ε4/4 44 ± 528 ± 364 ± 637 ± 443 ± 324 ± 339 ± 48 ± 3
P2NonsmokerF6212ε3/3 38 ± 635 ± 127 ± 229 ± 233 ± 522 ± 327 ± 220 ± 5
P3NonsmokerM665ε2/3 33 ± 410 ± 349 ± 525 ± 432 ± 416 ± 530 ± 516 ± 2
P4NonsmokerM6812ε2/3 51 ± 322 ± 458 ± 532 ± 246 ± 439 ± 227 ± 21 ± 0.5



For biochemical studies, left hemisphere frozen (-70°C) blocks were obtained from the frontal (Brodmann area 9), parietal (Brodmann area 39/40), temporal (Brodmann area 21), and occipital (Brodmann area 17/18) cortices, caudate nucleus, putamen, thalamus, and hippocampus. For histopathological investigation the right hemisphere of the brain was fixed in 4% formaldehyde for 4 weeks. Blocks of paraffin-embedded tissue were obtained from the same regions mentioned above. Eight-micrometer-thick sections were subjected to Nissl, hematoxylin, Bielschowsky, and Congo stainings.

Quantification of neuropathological features

Two-dimensional quantification of the NPs and NFTs was done on the sections stained according to the Bielschowsky method. NP and NFT profiles were counted throughout cortical layers in the 1-mm2 counting field. Two consecutive measurements in the superficial and deep layers of the cortex were performed on three randomly chosen areas where cytoarchitectonic characteristics of the region were well preserved. In the cortical regions, the base, middle, and top of the gyrus were chosen for quantification. Parts of the sections where structures were likely to be tangentially cut were avoided. Under high magnification (×100), a 1-mm2 stereological graticule with 100 counting points was superimposed perpendicular to the surface of the region. Only NP and NFT profiles that were completely inside the frame or crossing only two free but constantly defined edges of the frame were counted. The number of NPs and NFTs was corrected for the shrinkage factor and reported as a numerical density in the particular cortical area.

(-)-[3H]Nicotine and (±)-[3H]epibatidine binding

Membrane fractions (P2) (Gray and Whittaker, 1962) were prepared from frontal, temporal, occipital, and parietal cortices, thalamus, hippocampus, putamen, caudate nucleus, and cerebellum and stored at -80°C. Protein was measured by the method of Lowry et al. (1951), using albumin as a standard. The membrane suspensions (0.2 mg of protein/100 μ1 of homogenate) were thawed and incubated with 0.1 nM (±)-[3H]epibatidine (specific activity, 52 Ci/mmol ; Du Pont/NEN) in 50 mM Tris-HCl buffer, containing 120 mM NaCl, 5 mM KCl, 1 mM MgCl2, and 2.5 mM CaCl2 (pH 7.4), in a final volume of 1 ml for 3 h at 25°C. The same membrane suspension was incubated with 5 nM (-)-[3H]nicotine (specific activity, 64.4 Ci/mmol ; Du Pont/NEN) in 50 mM Tris buffer in a final volume of 1 ml for 40 min at 0°C. Nonspecific binding was determined in the presence of 0.1 mM (-)-nicotine. Incubation was terminated by vacuum filtration through Whatman GF/C filters, presoaked with 0.3% polyethylenimine solution for at least 5 h. Samples were washed four times with assay buffer. Radioactivity was counted in a Wallac scintillation counter. All assays were performed in triplicate.

Data analysis

Saturation binding data were analyzed by nonlinear curve-fitting analyses (Prism ; GraphPad Software, San Diego, CA, U.S.A.) to determine the total number of binding sites (Bmax) and affinity constant (KD). Statistical calculations were performed using Student's t test, and correlation coefficients were calculated by linear regression (ANOVA).

APOE genotyping

DNA from four 670/671 APP mutation AD subjects was extracted from brain tissue by standard methods. The APOE genotype was determined by using the INNO-LiPA APOE kit (Innogenetics N.V.).


(-)-[3H]Nicotine (specific activity, 64.4 Ci/mmol) and (±)-[3H]epibatidine (specific activity, 52 Ci/mmol) were obtained from Du Pont/NEN, Boston, MA, U.S.A. The INNO-LiPA APOE kit was purchased from Innogenetics N.V. (Ghent, Belgium). All other chemicals were of analytical grade.

Use of brain tissue obtained from the Huddinge University Hospital Brain Bank was permitted by the Karolinska Institute Ethical Committee (no. 92 : 268) and Ministry of Health (Socialstyrelsens tillstånd) (no. 5254-736/87).


Single concentrations of [3H]epibatidine (0.1 nM) and [3H]nicotine (5.0 nM), saturating ~60-70% of the maximal number of binding sites, were used to measure the distribution of nAChRs in brain of sporadic AD and APP 670/671 subjects and age-matched controls. In controls, the [3H]epibatidine binding was higher than that of [3H]nicotine in the cerebellum and thalamus, indicating somewhat different regional binding patterns for these two ligands in human brain Fig. 1). The numbers of nAChRs were significantly decreased in temporal cortex (-44%), frontal cortex (-50%), parietal cortex (-45%), occipital cortex (-57%), and hippocampus (-37%) of the sporadic AD cases compared with age-matched controls, when measured by [3H]nicotine (Fig. 1). Similar reductions in numbers of nAChRs, as obtained with [3H]nicotine, were also observed in [3H]epibatidine binding sites in temporal cortex (-40%), frontal cortex (-45%), parietal cortex (-52%), occipital cortex (-53%), and hippocampus (-38%) in sporadic AD brains. With [3H]epibatidine, an additional reduction in number of binding sites was detected in the thalamus (-38%) of sporadic AD cases (Fig. 1).

Figure 1.

Distribution of nicotinic receptors in temporal cortex (Tempctx), frontal cortex (Front ctx), parietal cortex (Par ctx), occipital cortex (Occ ctx), caudate nucleus (Caud nucl), putamen, hippocampus (Hippo), cerebellum (Cereb), and thalamus of sporadic AD (n = 7-14) and age-matched controls (n = 7-19) measured by (A) 5 nM [3H]nicotine and (B) 0.1 nM [3H]epibatidine as described in Materials and Methods. *p < 0.05 by Student's t test.

FIG. 1.

Individual data for the APP 670/671 subjects are shown in Table 2. Since the four AD subjects with the APP 670/671 mutation were ~15 years younger than the sporadic AD cases, a separate control group was collected. The control group showed a higher number of nicotinic receptors as measured with both [3H]epibatidine and [3H]nicotine, compared with the age-matched control group for the sporadic AD cases (Fig. 1) and Table 3).

Table 3. Changes in numbers of nAChRs in AD patients carrying the Swedish APP 670/671 mutationData are mean ± SEM values (no. of subjects).
  [3H]Nicotine (fmol/mg of protein) [3H]Epibatidine (fmol/mg of protein)
Brain regionControlSwedish APP 670/671 mutationControlSwedish APP 670/671 mutation
  1. ap <0.01

  2. bp <0.05

  3. c<0.001 by Student's t test.

Temporal cortex20.9 ± 2.7 (3)5.7 ± 0.8 (4)a21.1 ± 4.0 (5)4.8 ± 1.2 (4)b
Frontal cortex28.2 ± 3.5 (4)5.5 ± 1.8 (3)a27.2 ± 4.2 (5)3.5 ± 0.9 (3)a
Parietal cortex25.7 ± 2.5 (4)4.7 ± 1.1 (4)c32.8 ± 2.2 (5)6.2 ± 1.1 (4)c
Occipital cortex18.2 ± 2.1 (5)4.4 ± 0.8 (4)a25.6 ± 3.2 (5)6.3 ± 0.9 (3)a
Caudate nucleus23.0 ± 5.5 (5)10.8 ± 3.3 (4)45.4 ± 6.3 (5)28.7 ± 2.1 (3)
Putamen26.7 ± 7.0 (3)21.4 ± 2.9 (3)45.9 ± 3.9 (3)29.5 ± 3.2 (3)b
Thalamus15.2 ± 5.8 (4)15.6 ± 6.8 (3)48.1 ± 12.9 (5)22.3 ± 11.8 (3)
Cerebellum12.2 ± 1.3 (6)10.0 ± 1.9 (3)33.8 ± 4.6 (5)23.9 ± 3.5 (4)


[3H]Nicotine and [3H]epibatidine binding sites were significantly reduced in temporal (-73 and -77% for the respective ligand), frontal (-80 and -87%), parietal (-82 and -81%), and occipital (-76 and -75%) cortices of patients carrying the Swedish 670/671 APP mutation when compared with age-matched controls (Table 3). In addition, [3H]epibatidine binding sites were significantly reduced in the putamen (-36%) (Table 3). The deficits in nAChR binding sites in the cerebral cortices of 670/671 APP mutation cases were similar to those of sporadic AD cases, although the deficits appeared to be more pronounced in 670/671 APP mutation subjects when compared with age-matched controls. However, this could be attributed to the younger age and age-related loss of receptors in the cerebral cortices of these individuals.

Saturation analysis of [3H]epibatidine binding in the parietal cortices of APP 670/671 subjects revealed two classes of binding sites : a high-affinity site (KD1 = 0.018 ± 0.002 nM ; Bmax1 = 7.6 ± 1.2 fmol/mg of protein) and a low-affinity site with KD2 in the micromolar range, which was not quantified owing to a large variation in number of binding sites. The Bmax value and the estimated KD value for the high-affinity site in age-matched controls were KD1 = 0.036 ± 0.007 nM and Bmax1 = 42.9 ± 9.8 fmol/mg of protein, respectively. A significant reduction of the high-affinity [3H]epibatidine binding sites, Bmax1, was observed in the APP 670/671 subjects compared with age-matched controls (p <0.05) with no significant change in KD values. A representative example of [3H]epibatidine saturation binding curves in membrane preparations from the parietal cortex of APP 670/671 mutation subjects and age-matched control brains is shown in Fig. 2.

Figure 2.

A : Saturation binding of (±)-[3H]epibatidine (2 pM-10 nM) in P2 membrane fractions from parietal cortex of APP 670/671 subjects and age-matched controls. Inset : Saturation binding of (±)-[3H]epibatidine as shown in A but in the 2 pM-1 nM range. B : Scatchard plots of specific binding data shown in A. The curves are representative of one of three individual assays. B, bound ; B/F, bound/free.

FIG. 2.

To evaluate the association of nAChR deficits in AD APP 670/671 subjects with neuropathological features, the numbers of nAChRs, NPs, and NFTs were compared in four cortical brain areas. Among the cortical brain regions, the parietal cortex showed the highest load of NPs, whereas the reduction in number of [3H]nicotine binding sites appeared to be less compared with the occipital cortex and temporal cortex (Fig. 3). When studying each region separately, a significant increase in the number of [3H]nicotine binding sites was interestingly observed with an increased density of NPs (p < 0.05) in the parietal cortex (Fig. 4A). A tendency toward negative correlation was observed between the [3H]nicotine binding sites and the density of NPs in the temporal cortex of APP 670/671 subjects (Fig. 4A), but owing to the low number of subjects, the correlation of [3H]nicotine/NPs did not reach statistical significance. [3H]Epibatidine binding did not show any significant correlation to neuropathological features in any of the cortical brain regions studied (data not shown).

Figure 3.

Correlation of regional pathology (neuronal plaques) with number of [3H]nicotine binding sites in the Swedish APP 670/671 subjects. Data are mean ± SEM (bars) values of the four APP 670/671 cases.

Figure 4.

A : Relationship between NPs and [3H]nicotine binding in temporal and parietal cortex of four patients (P1-P4) carrying the APP 670/671 mutation. A significant positive correlation (p <0.05) was observed in the parietal cortex. r = regression coefficient, linear correlation. B : Relationship between NFTs and [3H]nicotine binding in temporal and parietal cortex of four patients (P1-P4) carrying the APP 670/671 mutation. No statistical significance was obtained.

FIG. 3.

FIG. 4.

A regional pattern similar to that observed for NPs was also noticed for the correlation of [3H]nicotine binding and NFT counts in each of the cortical areas studied in the APP 670/671 carriers. A tendency toward a negative correlation between the number of nAChRs and the density of NFTs was observed in the temporal cortex, and a tendency toward a positive correlation was observed in the parietal cortex (Fig. 4B). No correlation was observed between [3H]epibatidine binding sites and NFTs (data not shown).


The nAChRs are involved in memory processes in brain (Levin, 1992 ; Newhouse and Potter, 1994) and are markedly impaired in brains of sporadic AD patients (Nordberg et al., 1992). After the APP 670/671 mutation was discovered in one Swedish family (Mullan et al., 1992), five members of the family carrying the mutation have died with AD, four of which have come to autopsy. In the present investigation, the alteration in nAChRs was investigated for the first time in the Swedish APP 670/671 mutation family and compared with that in sporadic AD cases. Because several subtypes of nAChRs exist in the brain (Sargent, 1993 ; McGee and Role, 1995), two selective nicotinic agonists, (±)-[3H]epibatidine and (-)-[3H]nicotine, were used as receptor ligands to distinguish among different nAChR subtypes. Epibatidine exhibits high affinity to both α4 and α3 nAChR subunits (McKay et al., 1994 ; Gerzanich et al., 1995), whereas nicotine reveals highest affinity to the α4 nAChR subunit.

In agreement with previous studies (Nordberg, 1994), significant reductions in numbers of nAChRs were observed using [3H]nicotine in the cerebral cortices and hippocampus of sporadic AD brains. Epibatidine and nicotine both detected a selective loss of binding to the major high-affinity site, i.e., the α4β2 nAChR subtype, in the cerebral cortex of AD brains (Warpman and Nordberg, 1995), which was confirmed in this study. In addition, [3H]epibatidine, but not [3H]nicotine, revealed a significant reduction in number of nAChRs in the thalamus of sporadic AD cases compared with controls. The reduction of nAChRs observed with [3H]epibatidine binding in thalamus may reflect losses in the α3 nAChR subtype, because high levels of the α3 nAChR subunit have been found in the thalamus of human brain (Rubboli et al., 1994).

Significant reductions of [3H]nicotine and [3H]epibatidine binding sites were observed in the cerebral cortex of APP 670/671 AD subjects. Furthermore, [3H]epibatidine, but not [3H]nicotine, displayed an additional significant loss of binding in the putamen. A marked loss of nAChRs measured by [3H]nicotine and [3H]epibatidine has been found in normal aging (Nordberg et al., 1992 ; Marutle et al., 1998). Since APP 670/671 mutation cases were ~15 years younger than the sporadic AD cases, the reduction in number of nAChRs compared with age-matched controls appeared to be more pronounced compared with the deficits in nAChRs observed in sporadic AD cases.

Saturation analysis of [3H]epibatidine binding in the parietal cortex of APP 670/671 subjects revealed a significant reduction in number of high-affinity binding sites with no apparent change in the affinity constant, compared with age-matched controls.

The APP 670/671 mutation on chromosome 21 in the Swedish AD family is considered to be the cause of the disease, because it alters the APP metabolism leading to an accumulation of amyloid (Aβ) peptides and plaque formation (Citron et al., 1992 ; Mullan et al., 1992). Recent findings imply that an interaction between Aβ and the cholinergic nervous system may exist in the brain. Auld et al. (1998) postulated that Aβ might have a neuromodulatory function in the brain and thereby contribute to the vulnerability of selected cholinergic neuronal populations in AD. In addition, an activation of muscarinic cholinergic receptors has been reported to influence APP metabolism (Nitsch et al., 1992 ; Nitsch and Growdon, 1994). Furthermore, application of Aβ into the rat nucleus basalis has been shown to induce hypofunction of cholinergic neurons (Giovanneli et al., 1995), and Aβ inhibited in vitro release of endogenous acetylcholine from rat hippocampal and cortical slices at very low concentrations (picomolar to nanomolar) (Kar et al., 1996). Considering the neurotoxic potential of Aβ (Yankner et al., 1990 ; Mattson et al., 1993), the neurodegeneration of cholinergic neurons in AD (Coyle et al., 1983), and the possible involvement of nAChRs in mechanisms that protect against Aβ toxicity (Court et al., 1996 ; Salmon et al., 1996 ; Zamani et al., 1997), it is of great interest to investigate whether there is a relationship between the nAChRs and NPs inthe Swedish APP 670/671 family. A negative correlation between the density of senile plaques and the number of nAChRs has previously been observed in the frontal, temporal, and parietal cortices of normal and sporadic AD cases (Perry et al., 1989). In the present study, the parietal cortex also showed the highest abundancy of NPs while showing a smaller decrease in number of [3H]nicotine binding sites compared with other cortical brain regions.

Also interesting was the positive correlation seen between the number of [3H]nicotine binding sites and the density of NPs in the parietal cortex. A tendency toward a negative correlation was seen solely in the temporal cortex. Because the load of NPs in APP 670/671 cases is thought to be higher than in sporadic cases, it is quite plausible that a correlation may exist earlier in the course of the disease. As the [3H]epibatidine binding (high affinity to both α4β2 and α3 nAChRs) did not show similar tendencies in correlation to NPs as [3H]nicotine binding (high affinity to α4β2 nAChRs), the data suggest that the α4β2 nAChR subunit may be most closely related to the content of NPs. It is interesting that earlier studies have also shown a selective loss in α4β2 nAChRs in the temporal cortex of sporadic AD brain (Warpman and Nordberg, 1995). The lack of a strict negative correlation between nicotine binding sites and NPs might also be due to a more widespread disturbance of nicotinic synaptic pathology not limited to plaques.

It has earlier been reported that patients carrying APOE ε4 alleles have a higher risk for developing AD (Strittmatter et al., 1993 ; Corder et al., 1994) and a higher content of amyloid plaques in brain (Schmechel et al., 1993 ; Pirtillä et al., 1997). Some studies have shown an increased nAChR loss in AD patients with the APOE ε4 genotype (Poirier et al., 1995 ; Soininen et al., 1995), whereas we have not found any difference in the number of nAChRs in AD patients carrying the APOE ε4 allele compared with noncarriers (Svensson et al., 1997). Similarly, in the present study it was shown that the APP 670/671 AD subject who was ε4 homozygous (patient 1), although he had the highest load of NPs, also had the highest number of nAChRs. This patient had also been a heavy smoker, and although it is not certain whether the smoking continued up to his time of death, this could have contributed to the increased number of nAChRs (Benwell et al., 1988 ; Nybäck et al., 1989).

The presence of tangles is considered to be a more specific marker for damage of cell function than plaques and contributes to neuronal degeneration (Terry et al., 1991 ; DeKosky et al., 1992). However, in this study, we found no strict correlation between the numbers of nAChRs and tangles in cortical areas of APP 670/671 AD brains.

A significant positive correlation has been observed in vivo between cognitive function and the binding of [11C]nicotine in AD brains, indicating that there is a continuing loss of nAChRs with progression of the disease (Nordberg et al., 1995a,b). It is suggested that losses of nAChRs probably occur early in the disease, even before irreversible degeneration takes place (Nordberg et al., 1995a ; Perry et al., 1995). This fact and the possibility that distinct regions of the brain may have differing kinetics for the development of NFTs and NPs might explain why no strict correlation between nAChR deficits or neuropathological lesions was observed.

In conclusion, AD patients with the Swedish APP 670/671 mutation showed more pronounced reductions in numbers of nAChRs than sporadic AD cases compared with age-matched controls. No strict correlation between nAChRs and NPs or NFTs could be observed except for a significant positive correlation in the parietal cortex, suggesting that the correlation between different pathological components and neurotransmitters might be more complex and also that regional variances in brain exist, reflecting physiological differences in the structural dynamics of cortical organization.