Cortical α7 Nicotinic Acetylcholine Receptor and β-Amyloid Levels in Early Alzheimer Disease FREE

Cholinergic synaptic dysfunction contributes to cognitive impairment in Alzheimer disease (AD). These changes may be due, in part, to increased concentrations of β-amyloid (Aβ) peptides¹ and their interactions with nicotinic acetylcholine receptors (nAChRs), which are essential for normal cognitive function.^{2 - 3} β-amyloid binds to nAChRs, particularly the α7 subclass^{4 - 5} ; this may alter receptor function^{6 - 10} and also result in Aβ internalization, fibrillization, and deposition into plaques and cerebral vasculature.^{4 ,11 - 13} The status of α7 nAChRs in AD is controversial as there are reports of increases, decreases, or stability in AD.^{14 - 18} While non-α7 nAChR binding in frontal cortex declines early in AD,¹⁹ quantitative biochemical studies specific for α7 nAChRs in subjects with preclinical and early AD remain to be performed. The current study quantified α7 nAChR binding and total Aβ peptide concentration in superior frontal cortex (SFC) from subjects who participated in the Religious Orders Study.^{20 - 21} The status of these 2 biochemical measures was examined across subjects' groups defined by clinical diagnoses of no cognitive impairment (NCI), mild cognitive impairment (MCI), and early AD stage (mild to moderate AD [mAD]), or neuropathological diagnosis.

This study included 29 participants in the Religious Orders Study, a longitudinal clinical-pathological study of aging and AD in retired Catholic nuns, priests, and brothers.²⁰ Inclusion criteria and a description of the clinical evaluation have been published.^{20 - 21} At the last clinical evaluation (<12 months prior to death), subjects were classified as having NCI, MCI, or mAD (Table 1). Diagnosis of AD dementia was made using standard criteria.²² Mild cognitive impairment was defined as impairment on neuropsychological testing but without a diagnosis of dementia by the examining neurologist,²⁰ criteria similar to those describing patients who were not cognitively intact but nonetheless did not meet the criteria for dementia.^{23 - 26} A consensus conference of neurologists and neuropsychologists reviewed all the clinical and neuroimaging data, medical records, and interviews with family members and assigned a final diagnosis.

Neuropathological diagnosis of AD (possible, probable, or definite AD) or not AD (Table 1) was based on modified criteria by the Consortium to Establish a Registry for Alzheimer's Disease (CERAD),²⁷ which applied semiquantitative estimates of neuritic plaque density by a board-certified neuropathologist blinded to the clinical diagnosis.²⁸ Subjects were also assigned a National Institute on Aging (NIA)–Reagan neuropathological diagnosis²⁹ and a Braak score based on the presence of neurofibrillary tangles.³⁰ Subjects with pathology other than AD were excluded from the study.

Fresh frozen SFC (Brodmann area 9) gray matter was divided into aliquots for nAChR binding and Aβ peptide enzyme-linked immunosorbent assay (ELISA). For α7 nAChR binding, samples were homogenized in 10 volumes of 50mM Tris-hydrochloride (Tris-HCl) buffer (pH = 7.0), centrifuged twice at 40 000 × g for 10 minutes, resuspended in Tris buffer, and stored at −80°C. Samples were thawed and resuspended in an equal volume of Tris buffer containing 0.1% bovine serum albumin.^{31 - 32} Samples, 0.5 mg of protein each, were combined with 9.5nM [³H]methyllycaconitine (MLA) (159.1×10¹⁰ Bq/mmol; Tocris Cookson Ltd, Bristol, England) in Tris-HCl buffer containing 0.1% bovine serum albumin. Nonspecific binding was measured in the presence of 1mM nicotine. After a 2-hour incubation on ice, bound ligand was separated from free ligand using Whatman GF/B filters (Whatman, Florham Park, New Jersey), presoaked in 0.3% polyethyleneimine. Filters were rinsed with Tris-HCl buffer, placed in scintillation vials, and shaken in scintillation fluid for 1 hour before radioactivity was determined. Specific binding was calculated as the difference between total and nonspecific binding. Results were expressed as femtomoles per milligram of protein.

The Aβ assay was performed using a previously reported protocol.³³ Frozen SFC samples were homogenized (150 mg of tissue wet weight/mL of phosphate-buffered saline, pH = 7.4) and 30 mg of homogenized tissue were sonicated in 70% formic acid and centrifuged at 109 000 × g at 4°C for 1 hour, resulting in samples containing both soluble and formic acid–extracted insoluble Aβ peptides. The supernatant was neutralized with 1M Tris and 0.5M sodium phosphate, and the samples were assayed using a fluorescent-based ELISA (BioSource, Carlsbad, California) following the kit's instructions, with a capture antibody specific for the amino terminus of Aβ (amino acids 1-16) and detection antibodies specific for Aβ₄₀ and Aβ₄₂ peptides. Values were determined from standard curves using synthetic Aβ_1-40 and Aβ_1-42 peptides (BioSource) and expressed as picomoles per gram of wet brain tissue. “Total” Aβ values represent a sum of Aβ_1-40 and Aβ_1-42 peptide values.

β-amyloid levels were log-transformed because of data skewness. Comparisons of demographic characteristics, MLA binding, and Aβ levels between clinically or neuropathologically defined groups were performed using the Wilcoxon rank sum test, Kruskal-Wallis test, or the Fisher exact test, as appropriate. The association between demographic characteristics and MLA binding or Aβ levels was assessed by Spearman rank correlation or Wilcoxon rank sum test. Partial correlation was used for additional analyses adjusting for age. The correlation between MLA binding and Aβ levels was assessed by Spearman correlation. The level of statistical significance was set at .05 (2-sided).

The 3 clinical groups differed in Mini-Mental State Examination (MMSE) scores (P < .001), with the mAD group performing worse than both the NCI and MCI groups (Table 1). Clinical diagnostic groups also differed in CERAD diagnosis (P = .04), Braak staging (P = .04), and the NIA-Reagan diagnosis (P = .02), with the mAD subjects being more advanced neuropathologically compared with both the MCI and NCI groups (Table 1).

Subjects with a CERAD diagnosis of AD (CERAD score < 4; possible, probable, or definite AD; positive for cortical plaques) had lower MMSE scores (P = .06) and more advanced Braak stages and NIA-Reagan neuropathologic diagnoses (P = .004 and P < .001) (Table 1) than the not AD group (CERAD score = 4; no cortical plaques).

The MLA binding levels were slightly higher in mAD subjects; however, the difference was not statistically significant (Table 2). Clinical groups differed in total Aβ (combined Aβ₄₂ and Aβ₄₀; P = .02) (Figure) and Aβ₄₂ (P = .02), but not Aβ₄₀, concentrations, with mAD subjects having the highest levels.

Subjects with a CERAD diagnosis of AD had higher MLA binding (P = .08) (Figure) and higher concentrations of Aβ₄₂ (P = .02), Aβ₄₀ (P = .06), and total Aβ (P = .02) (Table 2) (Figure) in SFC when compared with those with a CERAD diagnosis of not AD.

There was no association of MLA binding with any of the demographic or clinical variables examined. Higher MLA binding levels correlated with greater likelihood of AD by the NIA-Reagan diagnosis (r = −0.47; P = .02) (Table 3) and weakly with Braak staging. There was an association of higher Aβ concentrations with more advanced age (r = 0.48-0.56; P = .01-.03; Table 3), but not with sex, education, the presence of APOE ε4 allele, or postmortem interval. Higher concentrations of total Aβ and Aβ₄₂, but not Aβ₄₀, correlated with lower MMSE scores (r = −0.62 for both; P = .003 and .004) (Table 3). In addition, higher total Aβ and Aβ₄₂ concentrations correlated with worse neuropathological scores (r = 0.62-0.70; P < .01), as did Aβ₄₀ concentrations, although to a lesser extent (Table 3). Adjusting for age, partial correlation showed similar results, although it yielded smaller correlation coefficients. There was no correlation between MLA binding and Aβ protein concentrations.

This study examined SFC α7 nAChR binding and Aβ peptide concentrations across the clinical and neuropathological categories of AD. Both markers were assayed in the same samples of cortical tissue from subjects who were clinically characterized within 12 months before death and neuropathologically evaluated post mortem. We did not detect significant changes in SFC α7 nAChR binding across clinical diagnostic groups. However, a trend toward elevated α7 nAChR binding levels was evident in subjects with CERAD diagnoses of AD (possible, probable, or definite) relative to not AD subjects (without neuritic Aβ plaques). Total Aβ (sum of Aβ₄₂ and Aβ₄₀) and Aβ₄₂ concentrations were elevated in the CERAD-AD group compared with the CERAD–not AD group, as well as in the clinical mAD compared with the MCI and NCI groups. The increase in α7 nAChRs in Aβ plaque–positive subjects, despite the lack of an association with Aβ concentrations, indicates that cellular expression of this receptor is influenced, either directly or indirectly, by the presence of senile plaques. This is in agreement with previous studies in patients with AD and animal models^{34 - 35} and is supported further by the current observation that the correlation between α7 nAChR levels and neuropathological staging was stronger using NIA-Reagan criteria compared with Braak staging, the latter relying only on neurofibrillary tangles for stage designation.³⁰

The apparent stability of α7 nAChRs across clinical categories of NCI, MCI, and mAD, and the lack of an association with MMSE scores, could be explained by the presence of plaques in all clinical groups. Cortical plaques were present in more than half of our NCI cases (Table 1), in agreement with previous reports of a substantial AD pathology in cognitively normal aged individuals.^{28 ,36 - 40} Although these studies would benefit from examining larger numbers of cognitively intact subjects free of any Aβ pathology, such individuals are rare, as Aβ plaques are a common feature in brains of elderly individuals.^{36 - 37 ,39 - 40} Furthermore, the pathological burden of Aβ includes not only insoluble fibrils in plaques, but also soluble Aβ oligomers.⁴¹ The impact of these distinct pools of Aβ on α7 nAChR binding in preclinical, early/moderate, and severe end-stage AD cases will be an important question to answer in future studies.

There are several possible explanations for the observed association between Aβ plaques and increased α7 nAChR binding. Plaques may serve as reservoirs of soluble Aβ species,¹ which can bind with high affinity to neuronal α7 receptors^{4 - 5} into a complex that is subsequently internalized.¹² This may result in a compensatory increase in expression of α7 nAChR on the cell surface. Additionally, excessive intracellular accumulation of Aβ₄₂ and subsequent neuronal lysis may contribute to plaque pathology,¹² potentially creating a cycle of neuronal degeneration and Aβ plaque deposition in AD. High concentrations of fibrillar Aβ in plaques, or soluble Aβ in the vicinity of these structures, may also influence the upregulation of α7 nAChR by reactive astrocytes. Astrocytes proliferate and display increased α7 nAChR density in the presence of Aβ plaques^{35 ,42 - 43} and upregulate nAChR messenger RNA expression and protein levels when exposed to Aβ in vitro.⁴⁴ Receptor binding assays cannot differentiate the relative contribution of different cell types to the overall regional expression of α7 nAChRs detected in tissue homogenates. Adding to this complexity are the postsynaptic and presynaptic sites of expression of α7 nAChRs, involving both local neuronal circuitry and afferent projections from distant neuronal cell populations. In this regard, a recent single-cell expression profiling study demonstrated upregulation of α7 nAChR messenger RNA in cortical-projecting basal forebrain cholinergic neurons in mAD subjects,⁴⁵ suggesting that changes in cortical α7 protein levels involve presynaptic elements on an important cholinergic afferent system. Collectively, these studies suggest that in SFC, α7 nAChR levels reported herein reflect changes both in cortical-projecting cholinergic basal forebrain neurons and regional cell-specific expression of these receptors in response to Aβ pathology.

In conclusion, the present findings demonstrate that cognitive decline in mAD is not associated with detectable changes in cortical α7 nAChR binding levels. In contrast, Aβ concentrations increased in mAD and correlated with cognitive impairment, in accord with reported associations of increased Aβ load with cognitive decline in AD.^{46 - 47} The observed trend for increased SFC α7 in subjects with plaques is in agreement with a previously reported positive correlation between α-bungarotoxin binding and Aβ plaque density³⁴ and warrants further investigation. These changes are in contrast with reports of reduced cortical α4 nAChR immunoreactivity with increased Aβ plaque densities and a loss of epibatidine binding with increased Aβ₄₂ concentrations.³⁴ Thus, α7 and non-α7 nAChRs may be differentially affected by Aβ pathology. In vivo positron emission tomography imaging techniques using radiolabeled probes for early detection of Aβ plaques^{48 - 49} and changes in select nAChRs⁵⁰ may act as early biomarkers for AD and will enable the timely implementation of appropriate therapies.

Correspondence: Steven T. DeKosky, MD, University of Virginia School of Medicine, PO Box 800793, Charlottesville, VA 22908 (dekosky@virginia.edu).

Accepted for Publication: November 5, 2008.

Author Contributions:Study concept and design: Ikonomovic, Wecker, Abrahamson, Ginsberg, Mufson, and DeKosky. Acquisition of data: Ikonomovic, Wecker, and DeKosky. Analysis and interpretation of data: Ikonomovic, Wecker, Abrahamson, Wuu, Counts, Ginsberg, Mufson, and DeKosky. Drafting of the manuscript: Ikonomovic, Wecker, Abrahamson, Wuu, and Mufson. Critical revision of the manuscript for important intellectual content: Ikonomovic, Wecker, Abrahamson, Wuu, Counts, Ginsberg, Mufson, and DeKosky. Statistical analysis: Wuu. Obtained funding: Ikonomovic, Ginsberg, Mufson, and DeKosky. Administrative, technical, and material support: Ikonomovic, Wecker, Ginsberg, and DeKosky. Study supervision: Ikonomovic.

Financial Disclosure: None reported.

Funding/Support: This work was supported by NIA grants AG14449 and AG10610.

Additional Contributions: Theresa Landers-Concatelli, MS, and William R. Paljug, MS, provided expert technical assistance. We are indebted to the support of the participants in the Religious Orders Study; for a list of participating groups, see http://www.rush.edu/rumc/page-R12394.html.