The study was performed with data obtained from a sample of individuals recruited at the Institute Universitaire de Geriatrie de Montreal, all of which underwent neurological examination, neuroimaging studies, laboratory tests, computed tomography (CT) and/or magnetic resonance imaging (MRI) scans of the brain, neuropsychological testing and longuitudinal studies for cognitive deficits [5] . All the recruited individuals underwent recurrent six months evaluations using a validated neuropsychological battery. The Protocole d’evaluation neuropsychologique optimal (PENO) battery was used as clinical measure of global cognition [14] ; this battery evaluated logical memory, comprehension, word production, discursive abilities, visual discrimination, and bell test through 20 psychological tests [5] (Table 1). Highly trained research psychologists made the evaluations but for the purpose of our study, we included one final evaluation before the patient’s death. A diagnosis of dementia was made during a consensus meeting using DSM-IV criteria [15] . The neuropsychological evaluation of the control group was taken from a standardized tables result, depending on the patient’s age and education. APOE

genotype data were available but the individuals that were selected for this study were not selected based on their APOE genotypes. The clinical profile of the individuals includes bioethic, deontological, clinico-patholog- ical, epidemiological, genetic, molecular, epidemiological, sociographic, toxic-environmental, biochemical, psychological, and neurological information. All patients and/or related family received full written and oral information regarding all study procedures and signed an informed consent. All experiments were done in accordance with the Helsinki Declaration of 1975 and the local ethics committee of the Institute de Geriatrie de Montreal, in accordance with current legislation on this field, approved the protocol as well as the details of the informed consent.

Brain tissue was obtained immediately after the death of non-traumatic origin of the patients. The study was conducted considering obtaining biopsies of 24 people: 11 individuals with cognitive decline-Alzheimer Diseases cases (mean age 76 ± 5.5 years old) and 13 non-demented controls (mean age 74.1 ± 6.9 years old) diagnosed clinically and histopathologically by two neuropathologists according to the NINCDS/RCDA criterion [16] [17] . Cortex tissue fragments ranging in size from 0.3 to 0.5 × 2 × 2 cm were obtained by an experimented technician within 24 hr after death and fixed by immersion in 4% paraformaldehyde for at least 36 hr. Eleven biopsy samples were acquired from the brain anatomical areas that correspond to the Brodmann’s areas described in Table 2. As part of the protocol biopsy samples from the eleven areas were stained to determine the expression of amyloid-β, phosphorylated tau protein and the presence of neurofibrillary tangles, neuritis dystrophy, and meganeuritic clusters. To quantitate the differences of the above mentioned markers only the results of brain anatomical areas 2, 4, 6 and 7 in both groups were considered.

Brain biopsies were paraffin embedded. Serial sections, 6-µm thick, were used for immunohistochemistry. Paraffin-embedded sections were dewaxed and rehydrated in decreasing concentrations of ethanol. When needed slides were pre-treated in 80% formic acid for 5 min to enhance amyloid-β immunoreactivity [11] [18] . Antigen retrieval was achieved by boiling the samples in 0.01 M sodium citrate pH 6 for 20 - 40 min. The quenching of the endogenous peroxidase was achieved by incubation with 0.3% hydrogen peroxide in methanol for 30 min at room temperature. After washing with phosphate-buffered saline (PBS) the slides were incubated with 1% albumin solution for 2 hr at room temperature, and extensively washed with PBS. The slides were incubated at 4˚C in a humid chamber overnight with primary antibody diluted in Tris-buffered saline containing 1% albumin. Mouse monoclonal anti-human hyper-phosphorylated PHF-tau (MABAD2) diluted 1:500 kindly provided by Dr. André Delacourte, mouse monoclonal anti-growth-associated protein 43 (MAB347) diluted 1:500, rabbit anti-neurofibrillary tangles (AB1518) antibody diluted 1:1000, and mouse monoclonal anti-synaptophysin (MAB332) diluted 1:1000, were from Millipore Co. (Billerica, MA). Rabbit anti-human tau (A002401) diluted 1:1000 was from Dako Co. (Carpinteria, CA).

Some biopsy slides were double stained to determine AD2 and amyloid-β co-expression. Briefly, sections were incubated overnight in a humidity chamber at 4˚C with the primary ant-AD2 antibody diluted 1:500 in PBS pH 7.4. After washing with PBS the slides were incubated for 1 hr at room temperature with a FITC-conjugated anti-isotype antibody (Jackson Immunoresearch, Canada). Samples were washed and incubated with a polyclonal anti-amyloid-β antibody (1:50, Boeringher-Manheim, Germany) for 2.5 hr at room temperature, washed with PBS and incubated for 1 hr at room temperature with a lissamine-rhodamine-conjugated anti-isotype antibody (Jackson Immunoresearch, Canada). Finally, the samples were washed and mounted with Vectashield (Vector laboratories, Burlingame).

Double-labeled slides were analyzed with a Zeiss microscope equipped with epi-ilumination and a dual laser confocal system (Zeiss LSM 410, Germany). Using a 40 × oil-immersion objective lens, consecutive images were captured in two channels (FITC A = 492; E = 520; lissamine-rhodamine A = 570; E = 590). The images were projected on the two-dimensional plane and were merged using a pseudocolor display (green for FITC and red for lissamine-rhodamine). In the merged image, a yellow color was interpreted as a co-localization between two fluorochromes.

Quantification of the number of plaques, neurofibrillary tangles and dystrophic neuritis: The density of pathological structures was evaluated using scans of all the tissue available on each sample, obtained with a 40X oil objective. Plaque subsets and subtypes were visually attributed according to their immunoreactivity. Since there were variations in the extent of tissue in the different sections, the subtype densities was normalized by dividing the number of pathological structures by the area of tissue. The gray matter area in each sample was determined by placing the slides under a CCD camera interfaced to an image analysis system (MCID Image Analysis System, Imaging Research Inc., ON. Canada); the density of plaques was automatically computed by image analysis software (IBAS, Kontron Electronik. Germany). After confocal analysis, total gray matter surface was reported as true density of pathological structures subsets/mm².

Quantification of meganeuritic clusters. 11 cortical brain regions (in both hemispheres) were scanned using a 10× objective (total magnification = 100×). Counting the number of clusters on the three richest fields and then calculating an average value determined the density of the clusters. Total numbers of synaptophysin positive meganeuritic clusters were reported by mm².

Data were expressed as the mean ± SD of each group. We compared the neuropsychological evaluation and the mean value of the different markers whose expression was analyzed, using the Mann Whitney test. A p ≤ 0.05 was considered statistically significant. Due to the effect of some variables such as age, scholarity, income, etc. some results were normalized using the z-test (from −1 to +1) where the higher the punctuation the greater the deterioration. Analyses were carried out using the SPSS 13.0 software (Addinsoft; SPSS Inc., Chicago, IL, USA).

All individuals asked to perform the test complied without difficulty. Neither educational status nor sensory disturbances such as mild visual or hearing disorders, limited the ability of the individuals to take the test. The time was spent to apply the test varied from 10 to 15 min. The results demonstrated significant differences in the cognitive function evaluated by the longitudinal studies battery for cognitive deficits between the control and the Alzheimer disease group. These results shown in Figure 1 corresponded to the normalized evaluation. The overall assessment of language ability that includes word comprehension, oral denomination, verbal fluency, discourse and reading abilities, and word repetition demonstrated significant differences between Alzheimer’s

individuals (mean 5.9 score vs. −9.8 in control individuals; p < 0.0001) (Figure 1(A)). The differences recorded and control in relation to memory (mean 0.01 vs. −0.62 score) (Figure 1(B)) indicated a decline in memory processes in the AD individuals but the difference was not statistical significant. The evaluation viso-constru- ctive evaluation (mean 11.18 vs. 1.28 score) (Figure 1(C)) showed significant differences between groups (p < 0.001). The most evident and highly significant (p < 0.0005) deteriorations were in spatial attention (mean 25.7 vs. −13.7 score) (Figure 1(D)), praxis (mean 20.3 vs. −26.1 score) (Figure 1(E)) and gnosis (mean 7.39 vs. −118.9 score) (Figure 1(F)) in the AD group compared to the control group. These results clearly reveal that with the exception of memory patients with AD exhibit severe cognitive decline compared to the control group.

The evaluation of the same parameters with the PENO protocol considering non-normalized data uniquely from Alzheimer’s disease individuals showed interesting differences. PENO evaluation was only adjusted to age and schoolarity. Although the overall normalized PENO score was significantly higher in Alzheimer’s disease individuals than in the control individuals (Figure 2(A)) the results with non-normalized data showed that the

less affected function was spatial attention whereas memory followed by gnosis were the most severely affected functions (Figure 2(B)). It was also interesting to observe the oscillatory pattern of cognitive decline, but as the figure shows some of the individuals did not conform to the pattern. Patient number 1 and 11 had Lewy bodies, patient 3 had an Alzheimer’s disease variant, and patient number 9 had severe vascular damage.

The quantification and comparison of histopathological hallmarks in the 2, 4, 6, and 7 anatomical areas of brain samples of the control and the Alzheimer’s disease group showed that the presence and quantitation of pathological structures were significantly higher in the Alzheimer disease group than in the control group. The only exception was the density of amyloid-β deposits in brain samples of individuals 60 years old or elder, independently of their status i.e., AD or control, although strangely, the density was lower in the AD samples (2.29 vs. 5.76, p = 0.777). Our most interesting result indicated a highly significant increase in Meganeuritic clusters density in Alzheimer’s disease (2.20 vs. 0.02, p < 0.001) but surprisingly there was an inverse and highly significant correlation between the neuropsychological evaluation results and the presence of meganeuritic clusters synaptophysin-reactive (Figure 3).

The neurofibrillary degeneration recognized by the AD2 antibody and the amyloid-β deposits demonstrated six singular lesions in the brain biopsy slides: 1) amyloid-β deposits [diffuse senile plaques, SPs], 2) classic neuritic plaque with an amyloid-β core [NP], 3) clusters of amyloid-β-free dystrophic neurites [DNCs], 4) dystrophic neurites [DNs], 5) neurofibrillary tangles [Ab + pTau + plaques; NFT], and 6) meganeuritic clusters immunoreactive to synaptophysin [MCs]. The images shown in Figure 4 are representative of the above-mentioned lesions found in the hippocampus of AD brain biopsies. A further Spearman correlation analysis to identify whether two variables can relate to a monotonic function showed a positive correlation between density of meganeuritic clusters and PENO, and density of Ab + pTau + plaques with PENO. The other correlations gave negative coefficients (Figure 5).