Alterations of the volatile metabolome in mouse models of Alzheimer's disease

Abstract

In the present study, we tested whether the volatile metabolome was altered by mutations of the Alzheimer's disease (AD)-implicated amyloid precursor protein gene (APP) and comprehensively examined urinary volatiles that may potentially serve as candidate biomarkers of AD. Establishing additional biomarkers in screening populations for AD will provide enhanced diagnostic specificity and will be critical in evaluating disease-modifying therapies. Having strong evidence of gross changes in the volatile metabolome of one line of APP mice, we utilized three unique mouse lines which over-express human mutations of the APP gene and their respective non-transgenic litter-mates (NTg). Head-space gas chromatography/mass spectrometry (GC/MS) of urinary volatiles uncovered several aberrant chromatographic peak responses. We later employed linear discrimination analysis and found that the GC/MS peak responses provide accurate (>84%) genotype classification of urinary samples. These initial data in animal models show that mutant APP gene expression entails a uniquely identifiable urinary odor, which if uncovered in clinical AD populations, may serve as an additional biomarker for the disease.

Figure 1. Sensor mice distinguish APP mouse urine from NTg urine.

Time spent investigating by bl6 ‘sensor’ mice (n = 10) in an odor cross-habituation task. bl6 mice in home cages were acclimated to the task with successive presentations of a cotton stick laced with ddH20, followed by successive presentations of a cotton stick laced with juvenile (2-4mo old) NTg urine (a) or separately, juvenile NTg urine followed by juvenile APP (Tg2576) urine (b), or in a final separate test, aged (14-16mo old) NTg urine followed by aged APP urine (c). ***p < 0.0001, vs. preceding trial #4, ANOVA followed by Fisher’s PLSD. Each presentation consisted of a homogenized sample from an individual mouse. Odor cross-habituation (time spent investigating the first trial of the APP urine sample vs. the last trial of NTg urine sample) was similar (n.s., p > 0.05, ANOVA followed by Fisher’s PLSD) between juvenile and aged mice (d).

Figure 2. Distinct GC/MS peaks in APP mouse urine.

2-dimensional histograms of head-space gas chromatography and mass spectroscopy (GC/MS) results from NTg and APP male mice, across three strains. Y-axis bins = individual mouse results. Data are scaled similarly across all strains and genotypes. Data along the z-axis represent peak areas generated in MSClust (see Materials and Methods). Location of some notable compounds are indicated (also see Table 4).