Serum Metabonomics Reveals Risk Factors in Different Periods of Cerebral Infarction in Humans

Guoyou Chen, Li Guo, Xinjie Zhao, Yachao Ren, Hongyang Chen, Jincheng Liu, Jiaqi Jiang, Peijia Liu, Xiaoying Liu, Bo Hu, Na Wang, Haisheng Peng, Guowang Xu, Haiquan TaoFrontiers in Molecular Biosciences 2022 · DOI

TL;DR

This study uses serum metabonomics to identify distinct metabolic profiles associated with different stages of cerebral infarction (CI) in humans. By analyzing over 100 metabolites, the researchers pinpoint key metabolic pathways and specific molecules that change over time, offering potential biomarkers for diagnosis and insights into CI's complex pathophysiology.

What Problem Does This Paper Address?

Cerebral infarction (CI), or stroke, is a leading cause of disability and mortality. While our understanding of CI's pathophysiology has grown, how metabolic changes evolve across different stages of the disease and whether these changes can predict disease progression remain largely unknown. Current diagnostic tools like CT and MRI primarily assess structural damage and timing, but a deeper understanding of the underlying biochemical shifts could lead to earlier diagnosis, better risk stratification, and more personalized treatment strategies. This paper aims to fill this gap by systematically investigating serum metabolite profiles at various time points after CI onset.

Methods

The researchers employed a non-targeted metabonomics approach using ultra-performance liquid chromatography-triple time of flight mass spectrometry (UPLC-Triple-TOF-MS). This technique allows for the simultaneous detection and quantification of a wide range of metabolites in biological samples.

Here's a breakdown of the key methods:

  • Sample Collection: Serum samples were collected from 143 CI patients categorized into four distinct time-based groups (within 3 days, 3-5 days, 5-7 days, and after 7 days of onset) and a group of 59 healthy controls. An additional group of CI patients with glycosuria was also included. A two-batch system (discovery and validation) was used to ensure robustness.
  • Metabolite Extraction and Analysis: Serum metabolites were extracted using acetonitrile and then analyzed using UPLC coupled with mass spectrometry in both positive and negative ion modes.
  • Data Processing and Statistical Analysis: Raw data were processed to extract and align metabolite peaks. Principal Component Analysis (PCA) and Orthogonal Partial Least-Squares Discriminant Analysis (OPLS-DA) were used to differentiate between the various groups. Statistical significance for metabolite differences was determined using the Wilcoxon Mann-Whitney test, with p < 0.05 considered significant.
  • Pathway Analysis: Identified metabolites were mapped onto metabolic pathways using the online MetaboAnalyst website to understand their biological relevance.

Datasets

The study utilized serum samples from:

  • 143 patients diagnosed with cerebral infarction (CI).
  • 59 healthy controls.

The CI patients were further divided into four time-based groups:

  • Group A: Within 3 days of cerebral infarction.
  • Group B: 3-5 days after cerebral infarction.
  • Group C: 5-7 days after cerebral infarction.
  • Group D: After 7 days of cerebral infarction.
  • Group T: Cerebral infarction patients with glycosuria.

The study employed a two-batch approach for sample collection:

  • First batch (discovery group): 100 CI patients and 41 healthy controls.
  • Second batch (validation group): 48 CI patients and 19 healthy controls.

No new benchmark datasets were introduced; the study focused on analyzing existing patient and control samples.

Key Insights and Ideas

  • Dynamic Metabolic Signatures: The most striking insight is that CI is not a static metabolic event. Instead, distinct sets of metabolites and metabolic pathways are significantly altered at different stages of the disease, suggesting a dynamic progression.
  • Phenylalanine Metabolism as a Key Player: Phenylalanine metabolism consistently emerged as a critical pathway involved in the development of CI, particularly within the first 3 to 7 days. This highlights the importance of amino acid metabolism in stroke pathophysiology.
  • Potential Biomarkers Identified: The study identified several metabolites that showed significant changes across different CI stages. Notably, phenylacetyl-L-glutamine was found to be elevated and proposed as a potential biomarker for CI.
  • Divergent Roles of Lipid Metabolism: Different classes of lipids, including medium-chain and long-chain acylcarnitines, phosphatidylcholines (PCs), and free fatty acids (FFAs), showed stage-specific alterations, suggesting complex roles in CI development and progression.
  • Glycosuria as a Distinct Metabolic State: CI patients with glycosuria (Group T) exhibited unique metabolic pathway differences compared to healthy controls, indicating that co-existing conditions like diabetes can significantly influence the metabolic landscape of stroke.

Key Results

  • Early Stage (within 3 days): Key metabolites included Phe-Phe, carnitine C18:1, palmitic acid, cis-8,11,14-eicosatrienoic acid, palmitoleic acid, 1-linoleoyl-rac-glycerol, MAG 18:1, MAG 20:3, phosphoric acid, 5α-dihydrotestosterone, Ca, K, and GGT.
  • Intermediate Stage (3-7 days): Significant changes were observed in GCDCA, glycocholate, PC 36:5, and LPC 18:2.
  • Late Stage (after 7 days): Metabolites like trans-vaccenic acid, linolenic acid, linoleic acid, all-cis-4,7,10,13,16-docosapentaenoic acid, arachidonic acid, DHA, various FFAs (18:1 to 22:6), PCs (34:1, 36:3, 38:4), ALP, and Crea showed alterations.
  • Vital Roles Identified: Phenylalanine metabolism, medium-chain acylcarnitines, long-chain acylcarnitines, choline, DHEA, LPC 18:0, LPC 18:1, FFA 18:0, FFA 22:4, TG, ALB, IDBIL, and DBIL were found to play crucial roles in the development of CI across different periods.
  • Validation: The second batch of samples confirmed significant changes in 13 metabolites, including 5a-dihydrotestosterone sulfate, choline, di-n-butylphthalate, several FFAs, L-phenylalanine, specific PCs, phenylacetyl-L-glutamine, and phosphoric acid.

Limitations and Future Work

  • Sample Size for Validation: The validation batch had a relatively modest sample size (n=18 for healthy controls), which the authors acknowledge might limit the power of the validation.
  • Underpowered for Associated Diseases: The study was underpowered to thoroughly evaluate the impact of co-existing conditions (other than blood sugar), dietary supplements, or specific diets on the metabolic profiles.
  • Need for Replication: The authors emphasize the need for replication in larger, independent cohorts to further validate the findings.
  • Complementary Diagnostic Tools: While current clinical diagnostics rely on imaging (CT/NMR), this study highlights the potential of metabonomics as a complementary tool for improving clinical prediction and understanding CI at different stages.
  • Personalized Treatment: Future work should focus on experimentally elucidating the pathophysiological mechanisms underlying these metabolic changes to inform personalized treatment approaches for CI.

Summary updated

Abstract

Studies of key metabolite variations and their biological mechanisms in cerebral infarction (CI) have increased our understanding of the pathophysiology of the disease. However, how metabolite variations in different periods of CI influence these biological processes and whether key metabolites from different periods may better predict disease progression are still unknown. We performed a systematic investigation using the metabonomics method. Various metabolites in different pathways were investigated by serum metabolic profiling of 143 patients diagnosed with CI and 59 healthy controls. Phe-Phe, carnitine C18:1, palmitic acid, cis-8,11,14-eicosatrienoic acid, palmitoleic acid, 1-linoleoyl-rac-glycerol, MAG 18:1, MAG 20:3, phosphoric acid, 5α-dihydrotestosterone, Ca, K, and GGT were the major components in the early period of CI. GCDCA, glycocholate, PC 36:5, LPC 18:2, and PA showed obvious changes in the intermediate time. In contrast, trans-vaccenic acid, linolenic acid, linoleic acid, all-cis-4,7,10,13,16-docosapentaenoic acid, arachidonic acid, DHA, FFA 18:1, FFA 18:2, FFA 18:3, FFA 20:4, FFA 22:6, PC 34:1, PC 36:3, PC 38:4, ALP, and Crea displayed changes in the later time. More importantly, we found that phenylalanine metabolism, medium-chain acylcarnitines, long-chain acylcarnitines, choline, DHEA, LPC 18:0, LPC 18:1, FFA 18:0, FFA 22:4, TG, ALB, IDBIL, and DBIL played vital roles in the development of different periods of CI. Increased phenylacetyl-L-glutamine was detected and may be a biomarker for CI. It was of great significance that we identified key metabolic pathways and risk metabolites in different periods of CI different from those previously reported. Specific data are detailed in the Conclusion section. In addition, we also explored metabolite differences of CI patients complicated with high blood glucose compared with healthy controls. Further work in this area may inform personalized treatment approaches in clinical practice for CI by experimentally elucidating the pathophysiological mechanisms.
BibTeX
@inproceedings{chen2022serum,
  author = {Guoyou Chen and Li Guo and Xinjie Zhao and Yachao Ren and Hongyang Chen and Jincheng Liu and Jiaqi Jiang and Peijia Liu and Xiaoying Liu and Bo Hu and Na Wang and Haisheng Peng and Guowang Xu and Haiquan Tao},
  title = {{Serum Metabonomics Reveals Risk Factors in Different Periods of Cerebral Infarction in Humans}},
  booktitle = {Frontiers in Molecular Biosciences},
  year = {2022},
  doi = {10.3389/fmolb.2021.784288},
}
Figure