Metabolite Profile of Alzheimer's Disease in the Frontal Cortex as Analyzed by HRMAS 1H NMR

Abstract

Background: Investigation on neurochemical changes in the frontal cortex in individuals with Alzheimer's disease (AD) and different Apolipoprotein E (APOE) genotypes, using ex vivo solid-state high-resolution NMR analysis, may lead to a better understanding of the neurochemistry associated with AD as well as new AD-specific metabolite biomarkers that might potentially improve the clinical diagnosis of AD. Methods: Intact tissue samples of the frontal cortex were obtained from 11 patients and 11 age-matched non-demented controls. Metabolite profiles in all samples were analyzed ex vivo, using solid-state high-resolution magic angle spinning (HRMAS) 600 MHz ¹H nuclear magnetic resonance (NMR). A logistic regression analysis was used to rank metabolites based on their level of contribution in differentiating the AD patient tissues and the controls, and different AD-associated APOE genotypes (APOE ε4 vs. APOE ε3). Results: Tissue samples from the AD patients showed significantly lower NAA/Cr (p = 0.011), Ace/Cr (p = 0.027), GABA/Cr (p = 0.005), Asp/Cr (p < 0.0001), mI/Cr (p < 0.0001), and Tau/Cr (p = 0.021), and higher PCho/Cr (p < 0.0001), GPCho/Cr (p < 0.0001), and α&β-Glc/Cr (p < 0.0001) than the controls did. Specifically, a newly observed resonance at 3.71 ppm, referred to as α&β-Glc, was observed in 90.9% of the AD samples (10/11). Samples with APOE ε4 also exhibited higher PCho/Cr (p = 0.0002), GPCho/Cr (p = 0.0001), α&β-Glc/Cr (p < 0.0001), and lower Asp/Cr (p = 0.004) and GABA/Cr (p = 0.04) than the samples with APOE ε3 did. In the logistic regression analysis, PCho, GPCho, ASP, and α&β-Glc were found to be the most relevant metabolites for differentiating the AD patient tissues and the controls, and different APOE genotypes. Conclusion: HRMAS ¹H NMR with high spectral resolution and sensitivity offers a powerful tool to gain quantitative information on AD associated neurochemical changes. There are important neurochemical differences in the frontal cortex between the AD patient tissues and the controls, and between those with different APOE genotypes. The resonance (α&β-Glc) found at 3.71 ppm in the AD patient tissues may be further investigated for its potential in the diagnosis and monitoring of AD.

Keywords: Alzheimer’s disease; brain; metabolic change; neurochemistry; nuclear magnetic resonance.

FIGURE 1

Expanded region of the HRMAS NMR spectra of tissue samples from the frontal cortex of the typical AD and control brains. The “signature” resonance of each metabolite is labeled. The newly observed AD-specific resonance (3.71 ppm) named as α&β-Glc in this study is indicated by the solid red arrow, while at the corresponding location of 3.71 ppm in control samples, no resonance peak was observed, as indicated by the dashed red arrow.

FIGURE 2

Comparison of different metabolite ratios (relative to creatine) calculated from the samples between the AD and control groups (^∗indicating p < 0.05).

FIGURE 3

Comparison of ratios of metabolite/Cr between samples with no APOE allele (APOE ε3; n = 12) and with at least one APOE allele (APOE ε4; n = 10, ^∗indicating p < 0.05).

FIGURE 4

The AD and control groups were compared first without regard for the APOE genotypes (AD vs. Control) and then with only the APOE ε4 samples in the AD group and only the APOE ε3 samples in the control group (AD ε4 vs. Control ε3). Metabolic changes (AD_metabolites/Control_metabolites - 1) found between the AD and control groups are greater in several metabolites (GABA, Asp, Cho, PCho, GPCho, and α&β-Glc at 3.71 ppm) when considering the APOE genotype of the samples with less pronounced differences in NAA, Gln, mI, and Tau.

FIGURE 5

Using a logistic regression analysis, metabolites were ranked, such as PCho, GPCho, Asp, and α&β-Glc, were the top metabolites based on their contributions to both differentiate the AD subjects from the controls (A), and classifying the different APOE genotypes (B), with double asterisks and single asterisk indicating a p < 0.01 and p < 0.05, respectively.