Exploratory Untargeted Metabolomics of Dried Blood Spot Samples from Newborns with Maple Syrup Urine Disease

Abstract

Currently, tandem mass spectrometry-based newborn screening (NBS), which examines targeted biomarkers, is the first approach used for the early detection of maple syrup urine disease (MSUD) in newborns, followed by confirmatory genetic mutation tests. However, these diagnostic approaches have limitations, demanding the development of additional tools for the diagnosis/screening of MUSD. Recently, untargeted metabolomics has been used to explore metabolic profiling and discover the potential biomarkers/pathways of inherited metabolic diseases. Thus, we aimed to discover a distinctive metabolic profile and biomarkers/pathways for MSUD newborns using untargeted metabolomics. Herein, untargeted metabolomics was used to analyze dried blood spot (DBS) samples from 22 MSUD and 22 healthy control newborns. Our data identified 210 altered endogenous metabolites in MSUD newborns and new potential MSUD biomarkers, particularly L-alloisoleucine, methionine, and lysoPI. In addition, the most impacted pathways in MSUD newborns were the ascorbate and aldarate pathways and pentose and glucuronate interconversions, suggesting that oxidative and detoxification events may occur in early life. Our approach leads to the identification of new potential biomarkers/pathways that could be used for the early diagnosis/screening of MSUD newborns but require further validation studies. Our untargeted metabolomics findings have undoubtedly added new insights to our understanding of the pathogenicity of MSUD, which helps us select the appropriate early treatments for better health outcomes.

Keywords: LysoPI; genetic testing; inborn errors of metabolism (IEMs); liquid chromatography high-resolution mass spectrometry; maple syrup urine disease (MSUD); methionine sulfoxide; newborn screening alloisoleucine; untargeted metabolomics.

Figure 1

Sample clustering and group separation based on a group of 20,568 features. (A) PLS-DA shows a clear separation between the two groups: MSUD newborn and healthy control. (B) OPLS-DA shows a clear separation between the two groups: MSUD newborns and healthy controls. The robustness of the created models was evaluated by the fitness of the model (R²Y: 0.963) and predictive ability (Q²: 0.353) values.

Figure 2

Volcano plot (moderated t-test, cut-off: p < 0.05, FC 1.5) between two groups: MSUD newborns and healthy controls. The heatmap revealed 1040 significantly dysregulated metabolites, where 303 (red) and 737 (blue) were up- and downregulated, respectively.

Figure 3

Hierarchical clustering (HAC) and heatmap analyses demonstrating (A) Upregulated metabolites. (B) Downregulated metabolites in MSUD newborns compared with healthy controls. The color scaled bar, green referred to upregulated metabolites, and red referred to downregulated metabolites.

Figure 4

Metabolomics profiling and biomarker evaluation between MSUD newborn and healthy control groups. (A): A receiver operating characteristics (ROC) curve was created by the OPLS-DA model, with area under the curve (AUC) values calculated from the combination of 5, 10, 15, 25, 50, and 100 metabolites (B): The frequency plot shows the top 15 identified metabolites. (C,D): Examples of metabolites methionine sulfoxide and L-alloisoleucine were upregulated in MSUD newborn patients with (AUC:0.81) and (AUC: 0.926), respectively. (E): lysoPI downregulated in MSUD newborns compared to healthy control (AUC: 0.86).

Figure 5

Pathway analysis of the significant metabolites dysregulated in MSUD newborns. In total, 210 metabolites were finally identified as human endogenous metabolites; 51 were upregulated, and 159 were downregulated. The color variation (yellow to red) shows the different significance levels of metabolites in the data.

Scheme 1

The workflow of data analyses performed in this study.