Untargeted metabolomics analysis of differences in metabolite levels in congenital heart disease of varying severity

Abstract

Abstract Background Congenital heart disease (CHD) is characterized by various phenotypes, however, differences in metabolic profiles associated with CHD of various severity have not been elucidated. In this study, differences in metabolite concentrations among mild, moderate, and severe forms of CHD were explored, providing novel clues for our understanding of the mechanism of CHD. Methods Maternal amniotic fluid samples from fetuses with mild (n = 15), moderate (n = 7), and severe (n = 29) CHD lesions were analyzed by GC-TOF/MS. PCA, PLS-DA, and differential metabolite analysis among these three groups were conducted. Results PCA and PLS-DA models showed that metabolic profiles were comparable among CHD of different severity. Significant differences between mild and moderate CHD lesions were observed in the levels of gluconolactone, ornithine, threonine, sorbose, pentadecanoic acid, and the uric acid/xanthine ratio. Of these six differential metabolites, gluconolactone (r = 0.469, P = 0.028), sorbose (r = 0.577, P = 0.005) and the uric acid/xanthine ratio (r = 0.438, P = 0.041) were positively correlated with moderate CHD lesions, while ornithine (r=-0.531, P = 0.011), threonine (r=-0.546, P = 0.009), and pentadecanoic acid (r=-0.454, P = 0.034) were negatively associated. We found 9 differential metabolites between mild and severe CHD lesions, among which the alpha-ketoisovaleric acid/valine ratio (r=-0.383, P = 0.010), gluconolactone (r = 0.391, P = 0.009), and 4-hydroxycinnamic acid (r = 0.342, P = 0.023) were correlated with severe CHD lesions. Only sorbose showed significant differences between moderate and severe CHD lesions, and was negatively associated with severe CHD lesions (r=-0.341, P = 0.042). Conclusions Compared with mild CHD, specific differences were observed in metabolites or metabolite ratios in moderate and severe CHD lesions of CHD, several of which were significantly correlated with CHD severity. These results can help to understand the metabolic status of the affected fetus and provide new possibilities for exploring the pathological mechanism of CHD.