Novel diagnostics of metabolic dysfunction detected in breath and plasma by selective isotope-assisted labeling

Abstract

Metabolomics is the study of a unique fingerprint of small molecules present in biological systems under healthy and disease conditions. One of the major challenges in metabolomics is validation of fingerprint molecules to identify specifically perturbed pathways in metabolic aberrations. This step is crucial to the understanding of budding metabolic pathologies and the ability to identify early indicators of common diseases such as obesity, type 2 diabetes mellitus, metabolic syndrome, polycystic ovary syndrome, and cancer. We present a novel approach to diagnosing aberrations in glucose utilization including metabolic pathway switching in a disease state. We used a well-defined prenatally exposed glucocorticoid mouse model that results in adult females with metabolic dysfunction. We applied the complementary technologies of nuclear magnetic resonance spectroscopy and cavity ring-down spectroscopy to analyze serial plasma samples and real-time breath measurements following selective (13)C-isotope-assisted labeling. These platforms allowed us to trace metabolic markers in whole animals and identify key metabolic pathway switching in prenatally glucocorticoid-treated animals. Total glucose flux is significantly proportionally increased through the major oxidative pathways of glycolysis and the pentose phosphate pathway in the prenatally glucocorticoid-treated animals relative to the control animals. This novel diagnostics approach is fast, noninvasive, and sensitive for determining specific pathway utilization, and provides a direct translational application in the health care field.

Figure 1. Proposed carbon fate following injection with [1-¹³C]-glucose

Schematics of different fates of labeled carbon in glucose metabolism following injection with [1-¹³C]-glucose in: (a) control mice (b) cortisol mice. Arrow width indicates the relative amount of pathway usage. Cortisol mice show significant increase in usage of the pentose phosphate cycle (PC) relative to Embden Meyerhof (EM) indicating an overall increase in glucose flux relative to control mice. Light gray boxes indicate CO₂ detection in breath by CRDS. Dark gray boxes indicate small metabolites detection in plasma via NMR.

Figure 2. SIAL labeling and tracing of ¹³C-stable isotope biomarkers in breath by CRDS

Traces are δ¹³C values from exhaled carbon dioxide shown for: (a) control mice, and (b) cortisol animals. The blue traces in both panels are the δ¹³C values following injection with [1-¹³C]-glucose (vehicle (n=3), cortisol (n=4)). The red traces in both panels are the δ¹³C values following injection with [6-¹³C]-glucose (vehicle (n=3), cortisol (n=3)). Animals prenatally treated with cortisol show elevated glucose flux. All traces are with respect to the injection time. The traces reflect the average δ¹³C values every fifteen minutes over a period of six hours for all animals relative to the injection time and are not corrected for matching maximum δ¹³C values. A time delay for peak δ¹³C value exists following treatment with [6-¹³C]-glucose in the cortisol mice. The control mice reach a peak δ¹³C value at 0.57 hours post injection whereas the prenatally cortisol-injected mice reach a peak δ¹³C value at 0.75 hours post injection. The unit of measurement for the δ¹³C values reported is %_o (parts per mil). Error bars represent the SEM. δ¹³C values normalized to the maximum δ¹³C value are reported in Table 1.

Figure 3. SIAL tracing of ¹H attached to ¹³C-lactate by NMR

Traces a-c of selected NMR regions after ¹³C-glucose enrichment in control and cortisol female mice. The plasma was obtained 2 h after injection with one of three different labeled glucose precursors: [1-¹³C]-glucose, [6-¹³C]-glucose, or ¹³C-U-glucose as indicated in the figure. The spectra shown contain only peaks from ¹H that are attached to ¹³C. These spectra were obtained by subtracting a ¹³C-filtered spectrum (containing only protons attached to ¹²C) from the corresponding unfiltered experiment (containing both ¹³C and ¹²C species). Panel (a) shows lactate and alanine peaks. Cortisol animals show a decrease in the amount of ¹³C-lactate derived from both [1-¹³C]-glucose and ¹³C-U-glucose. Panel (b) shows ¹H-[1-¹³C]-glucose assigned peaks indicating a decrease in plasma ¹³C-glucose in cortisol mice. Panel (c) depicts recycling of ¹³C-label to the rest of ribose protons (¹H-¹³C) in designated spectral region, * indicates peak arising from methanol. Assignments for ribose protons are shown in Supplementary Fig. 2A,B. Similar recycling is observed between the control and cortisol animals in each of labeled glucose.