Neuropil deposition of beta-amyloid (A beta) peptides is believed to be a key event in the neurodegenerative process of Alzheimers disease (AD). An early and consistent clinical finding in AD is olfactory dysfunction with associated pathology. Interestingly, transgenic amyloid precursor protein (Tg2576) mice also show early amyloid pathology in olfactory regions. Moreover, a recent study indicates that axonal transport is compromised in the olfactory system of Tg2576 mice, as measured by manganese-enhanced magnetic resonance imaging (MEMRI). Here we tested whether the putative axonal transport deficit in the Tg2576 mouse model improves in response to a selective gamma-secretase inhibitor, N-[cis-4-[(4-chlorophenyl)-sulfonyl]-4-(2,5-difluorophenyl)cyclohexyl]-1,1,1-trifluoromethanesulfonamide (MRK-560). Tg2576 mice or wild-type (WT) littermates were treated daily with MRK-560 (30 mu mol/kg) or vehicle for 4 (acute) or 29 days (chronic). The subsequent MEMRI analysis revealed a distinct axonal transport dysfunction in the Tg2576 mice compared with its littermate controls. Interestingly, the impairment of axonal transport could be fully reversed by chronic administration of MRK-560, in line with the significantly lowered levels of both soluble and insoluble forms of A beta found in the brain and olfactory bulbs (OBs) following treatment. However, no improvement of axonal transport was observed after acute treatment with MRK-560, where soluble but not insoluble forms of A beta were reduced in the brain and OBs. The present results show that axonal transport is impaired in Tg2576 mice compared with WT controls, as measured by MEMRI. Chronic treatment in vivo with a gamma-secretase inhibitor, MRK-560, significantly reduces soluble and insoluble forms of A beta, and fully reverses the axonal transport dysfunction.

Translational bionnarkers in Alzheimer's disease based on non-invasive in vivo methods are highly warranted. H-1 magnetic resonance spectroscopy (MRS) is non-invasive and applicable in vivo in both humans and experimental animals. In vivo H-1 MRS and 3D MRI were performed on brains of double transgenic (tg) mice expressing a double mutant human P-amyloid precursor protein APP(K670N,M671L) and human mutated presenilin gene PS1M146L, and wild-type (wt) littermates at 2.5, 6.5 and 9 months of age using a 9.4T magnet. For quantification, LCModel (TM) was used, and the data were analyzed using multivariate data analysis (MVDA). MVDA evidenced a significant separation, which became more pronounced with age, between tg and wt mice at all time points. While myo-inositol and guanidoacetate were important for group separation in young mice, N-acetylaspartate, glutamate and macrolipids were important for separation of aged tg and wt mice. Volume segmentation revealed that brain and hippocampus were readily smaller in tg as compared to wt mice at the age of 2.5 months. Amyloid plaques were seen in 6.5 and 9 months, but not in 2.5 months old animals. In conclusion, differences in brain metabolites could be accurately depicted in tg and wt mice in vivo by combining MRS with MVDA. First differences in metabolite content were readily seen at 2.5 months, when volume defects in tg mice were present, but no amyloid plaques. (c) 2007 Elsevier Inc. All rights reserved.

Early diagnosis of Alzheimer's disease (AD) would be supported by the detection of new in vivo biomarkers. Magnetic resonance spectroscopy (MRS) provides means to study metabolites of the brain over time. A non–hypothesis driven analysis of the spectroscopy data may be appropriate to characterize the fingerprint of complex MRS data with large numbers of variables. In vivo MRS was performed using a 9.4/30 T horizontal Bruker magnet. 8–17 transgenic (tg) APP/PS1 mice and 11–13 wild type (wt) littermates per time point were investigated at the age of 2.5, 4.5, 6.5 and 9 months, respectively. A water suppressed PRESS pulse sequence was used to acquire the spectrum in a volume of 8 mm3 of hippocampus. Data acquisition was followed by processing using LCModel and multivariate data analysis using PLS–DA. Relative values of 10 variables (Glutamate (Glu), Glutamate/Glutamine (Glx), myo–Inositol (Ins), N–Acetylaspartate (NAA), N–Acetylaspartate/N–Acetylaspartylglutamate, (tNAA), Taurine (Tau), Glycerophosphocholine/Phosphocreatine (tCho), macromolecules/lipids (MM09), (MM09Lip09), (MM14) were derived with LCModel. Scatter plots of PLS–DA models showed a significant separation between tg and wt at the age of 4.5 (Q2=0.134), 6.5 (Q2=0.134) and 9 (Q2=0.146) months. 2.5 month old mice showed a tendency for separation without reaching statistical significance. Leave one out prediction with cut off value of 0.5 classified samples with the following sensitivity and specificity: 82% and 75% at 2.5, 71% and 77% at 4.5, 72% and 85% at 6.5, 55% and 83% at 9 months. The most important variables for separation were tCho and Glu or NAA and tNAA at 6.5 or 9 months, respectively. Regression analysis revealed a significant decrease of Glu, Glx, MM09, MM09LIP09 with age for tg and wt. No significant change with age was observed for NAA, Ins, Tau or tCho. tNAA significantly decreased with age in tg but not wt. Differences in composition and changes with age of brain metabolites could be accurately depicted in tg and wt mice in vivo by combining MRS with LC–Model and PLS–DA. Separation between tg and wt was possible earliest at 4.5 months of age. The sensitivity and specificity of the method averaged 80% and 70% respectively.