Application of comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometry method to identify potential biomarkers of perinatal asphyxia in a non-human primate model

Abstract

Perinatal asphyxia is a leading cause of brain injury in infants, occurring in 2-4 per 1000 live births. The clinical response to asphyxia is variable and difficult to predict with current diagnostic tests. Reliable biomarkers are needed to help predict the timing and severity of asphyxia, as well as response to treatment. Two-dimensional gas chromatography-time-of-flight-mass spectrometry (GC×GC-TOFMS) was used herein, in conjunction with chemometric data analysis approaches for metabolomic analysis in order to identify significant metabolites affected by birth asphyxia. Blood was drawn before and after 15 or 18 min of cord occlusion in a Macaca nemestrina model of perinatal asphyxia. Postnatal samples were drawn at 5 min of age (n=20 subjects). Metabolomic profiles of asphyxiated animals were compared to four controls delivered at comparable gestational age. Fifty metabolites with the greatest change pre- to post-asphyxia were identified and quantified. The metabolic profile of post-asphyxia samples showed marked variability compared to the pre-asphyxia samples. Fifteen of the 50 metabolites showed significant elevation in response to asphyxia, ten of which remained significant upon comparison to the control animals. This metabolomic analysis confirmed lactate and creatinine as markers of asphyxia and discovered new metabolites including succinic acid and malate (intermediates in the Krebs cycle) and arachidonic acid (a brain fatty acid and inflammatory marker) as potential biomarkers. GC×GC-TOFMS coupled with chemometric data analysis are useful tools to identify acute biomarkers of brain injury. Further study is needed to correlate these metabolites with severity of disease, and response to treatment.

Figure 1. A GC x GC-TOFMS chromatogram of a primate pre-birth and post-birth in the asphyxia group

(a) Two-dimensional unfolded chromatogram showing increased signal of metabolite peaks in the post-birth sample compared to the pre-birth sample. The x-axis shows the unfolded retention times for the first and second columns. The y-axis shows signal intensity on m/z 73. (b) Contour plots of same pre- and post-birth samples showing increased signal post-birth in the asphyxia group. The x-axis shows column 1 retention time (s), and the y-axis shows column 2 retention time (s). The color bar demonstrates the signal intensity of each peak plotted on the chromatogram on m/z 73.

Figure 1. A GC x GC-TOFMS chromatogram of a primate pre-birth and post-birth in the asphyxia group

(a) Two-dimensional unfolded chromatogram showing increased signal of metabolite peaks in the post-birth sample compared to the pre-birth sample. The x-axis shows the unfolded retention times for the first and second columns. The y-axis shows signal intensity on m/z 73. (b) Contour plots of same pre- and post-birth samples showing increased signal post-birth in the asphyxia group. The x-axis shows column 1 retention time (s), and the y-axis shows column 2 retention time (s). The color bar demonstrates the signal intensity of each peak plotted on the chromatogram on m/z 73.

Figure 2. Principal Component Analysis (PCA) Plot and Fisher Ratio Plot on all m/z for primates A1 through A11

(a) The contour 2D plot of absolute values of PCA loadings. Peaks correspond with areas of the chromatograms with increased variance upon direct comparison. The x-axis shows the column 1 retention time (s), and the y-axis column 2 retention time (s). The color bar demonstrates the signal intensity of each peak plotted on the chromatogram. PCA plot demonstrates the peaks identified by PCA that have variance between samples without controlling for within group variation. (b) The contour 2D sum of Fisher ratio plots for m/z 40–600. The x-axis shows column 1 retention time (sec). The y-axis shows column 2 retention time (s). The color bar demonstrates the signal intensity of each peak plotted on the chromatogram. Peaks correspond with areas of the chromatograms with variance between sample groups (pre- versus post-birth) while controlling for within class variation (primate biological variance). The reduced number of peaks compared to (a) demonstrates a refined subset of data targeting specific areas of interest.

Figure 2. Principal Component Analysis (PCA) Plot and Fisher Ratio Plot on all m/z for primates A1 through A11

(a) The contour 2D plot of absolute values of PCA loadings. Peaks correspond with areas of the chromatograms with increased variance upon direct comparison. The x-axis shows the column 1 retention time (s), and the y-axis column 2 retention time (s). The color bar demonstrates the signal intensity of each peak plotted on the chromatogram. PCA plot demonstrates the peaks identified by PCA that have variance between samples without controlling for within group variation. (b) The contour 2D sum of Fisher ratio plots for m/z 40–600. The x-axis shows column 1 retention time (sec). The y-axis shows column 2 retention time (s). The color bar demonstrates the signal intensity of each peak plotted on the chromatogram. Peaks correspond with areas of the chromatograms with variance between sample groups (pre- versus post-birth) while controlling for within class variation (primate biological variance). The reduced number of peaks compared to (a) demonstrates a refined subset of data targeting specific areas of interest.

Figure 3. Heat map demonstrating change in concentration of metabolites identified by Fisher ratio comparing asphyxia group to control group

The 15 identified metabolites are listed on the y-axis. The individual primates are listed on the x-axis. Each square represents one metabolite from one primate sample. Samples are standardized to 1 to demonstrate degree of change. Samples with colors along the yellow/red spectrum have increased PARAFAC intensity demonstrating increased metabolite concentration compared to the mean. This heat map demonstrates that the pre-birth cord blood samples are similar to each other in both the control and asphyxia group. The asphyxia group shows marked elevation in metabolite concentration compared to control group demonstrating the affects of asphyxia on the primate metabolome. Bold font designates metabolites with significant changes in metabolite concentration as shown in Table 3 (p<0.05). * unable to perform unpaired Student’s t-test due to control n=1

Figure 4. Principal Component Analysis plot confirming variance of pre- and post-birth and asphyxia group to control

Principal Component 1 (PC1) loadings demonstrate variance between samples that were previously identified by Fisher ratio. The x-axis plots the 4 groups of metabolomes – pre- and post-birth of the control and asphyxia groups. The y-axis plots the PC1 loading score, which demonstrates the degree of variance each primate metabolome has when compared to the others. Pre-birth samples have similar PC1 loadings (unfilled shapes). All post-birth samples have increased PC1 loadings compared to pre-birth. There are increased PC1 loadings (i.e. increased variance) in post-birth asphyxia group compared to the control group. The square shape represents control C2 who was eliminated from data analysis. The outlier in the pre-birth asphyxia group represents A14, a primate that had a pre-birth profile more comparable to the post-birth asphyxia group.