Introduction

Ischemic stroke (IS), a neurological disorder caused by stenosis or occlusion of cerebral artery, is a leading cause of chronic disability worldwide and a serious threat to human life [1]. In addition to vascular occlusion caused by major atherosclerosis, IS can also be caused by acute cerebral infarction with definite etiology such as cardiogenic embolism, arteriolar occlusion, infectious disease, non-immune vascular disease, hypercoagulable state, and cryptogenic stroke with unclear etiology. Among them, the middle cerebral artery is a common occluded blood vessel, while occlusion of the vertebrobasilar artery, posterior cerebral artery, and anterior cerebral artery is less frequent [2, 3]. In the United States alone, over 750,000 stroke cases occur annually, making it the fifth leading cause of death and the top cause of disability [4]. Globally, IS ranks second in mortality and first in disability, with a concerning rise in younger individuals experiencing the disease [5]. Strokes can be hemorrhagic or ischemic, with over 85% being ischemic due to interrupted blood flow to the brain, leading to irreversible cell damage [6]. Inflammatory response [7], autophagy [8], and cell apoptosis are the primary mechanisms underlying brain tissue damage in IS [9]. Treatment for IS patients depends on the time of onset, neurological deficits, and neuroimaging results [10]. Despite significant advancements in thrombolysis and mechanical thrombectomy over the past decade, IS remains a major contributor to global healthcare burden [11]. Recombinant human tissue plasminogen activator (rt-PA) is an established intervention for acute IS, with clinical studies demonstrating increasing usage since 2006, highlighting the effectiveness of intravenous thrombolytic therapy [12]. However, the therapeutic window for rt-PA is narrow, and its effectiveness is limited by slow reperfusion. Additionally, rt-PA carries a significant risk of bleeding, which is roughly 10 times higher than in patients who do not receive rt-PA. According to clinical data statistics, the incidence of bleeding transformation after thrombolytic therapy in acute IS patients ranges from 10 to 43%, with symptomatic intracranial hemorrhage occurring at a rate of 1.7–10.3% [13, 14]. This highlights the urgent need for new diagnostic tools and therapies, particularly for early IS detection. Metabolomics, with its ability to detect rapid changes in body metabolites during disease onset, holds immense potential for early diagnosis and improved treatment strategies for IS. This review explores the current research landscape of metabolomics in IS and summarizes its promising applications in clinical settings.

Overview of metabolomics

What is metabolomics

Metabolomics, an emerging field within systems biology and an extension of omics technologies like genomics, transcriptomics, and proteomics, has gained widespread application in life sciences due to rapid advancements in analytical techniques [15]. First proposed by Professor Nicholson in 1999 [16], metabolomics focuses on studying the metabolic product profiles and dynamic changes of organisms, tissues, or cells under different physiological and pathological stimuli [21]. Conversely, non-targeted metabolomics involves a comprehensive analysis of all measurable analytes (including unidentified metabolites) within a sample under given conditions, providing a more comprehensive picture and avoiding potential biases in research direction [22]. This method is often used for exploratory studies, analyzing metabolites without prior knowledge of their identities. Non-targeted metabolomics holds immense potential for a holistic approach in biomedical research, enabling the discovery of novel biomarkers through comparison with metabolomics libraries, ultimately improving disease diagnosis and understanding of underlying pathological mechanisms [23]. In clinical practice, biological samples for metabolomics analysis are primarily obtained from blood, urine, feces, cerebrospinal fluid, saliva, and tissues. The core objective lies in analyzing the relationships between metabolites and physiological/pathological changes within the body from these samples, ultimately aiming to reveal the pathogenesis of diseases at a holistic level [24]. As a relatively new research method, metabolomics offers several advantages, including ease of sample acquisition, simplified protein detection, minimal requirements for large-scale database construction, convenient data processing, and high detection efficiency. These attributes have contributed to the growing interest in this field in recent years.

Research methods and data analysis of metabolomics

The rapid development of metabolomics is inextricably linked to advancements in its technology. Liquid chromatography (LC) and gas chromatography (GC) are the most prevalent methods for metabolite separation, while metabolite detection primarily relies on nuclear magnetic resonance (NMR) and mass spectrometry (MS) [25]. In metabolomics analysis, LC and GC are often coupled with MS, whereas NMR typically functions as a standalone tool. These techniques have become the mainstream platform for identifying and quantifying metabolites [26]. The combination of high-performance separation chromatography with highly specific and sensitive mass spectrometry enables rapid metabolite identification and accurate quantitative analysis. GC-MS excels at analyzing thermally stable and volatile metabolites with minimal matrix effects from complex samples [27]. LC-MS, on the other hand, boasts broader analytical capabilities. It can be combined with diverse chromatographic columns and other conditions for analysis, enabling the separation and identification of a wider range of metabolites within the sample without extensive pre-treatment. Notably, its high sensitivity makes it ideal for analyzing thermally unstable, non-volatile, and higher molecular weight substances [28]. NMR offers several advantages, including fast analysis speed, high reproducibility, suitability for high-throughput analysis, non-destructive nature, minimal bias, and simple sample preparation. It can simultaneously detect multiple organic compounds, making it widely used in metabolomics analysis [29].

The primary analytical methods in metabolomics encompass univariate and multivariate analysis. Univariate analysis, characterized by its simplicity, intuitiveness, and ease of understanding, is commonly employed in metabolomics research to swiftly examine differences between metabolite categories. Due to the challenges of metabolomics data meeting the assumptions of parametric testing, non-parametric methods like the Wilcoxon rank sum test, Kruskal-Wallis test, and t-test are frequently used. Additionally, calculating fold changes in metabolite concentration between groups and the area under the ROC curve (AUC) are common practices. Multivariate analysis encompasses various techniques such as principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), orthogonal projection to latent structures (OPLS), and cluster analysis (CA). PCA leverages the relationships between original variables, transforming them into a set of independent, comprehensive indicators (principal components) based on maximizing variation. Typically, 2–3 principal components are plotted to visually depict differences in metabolic patterns and clustering between groups. Load plots are then used to identify original variables contributing to group classification as potential biomarkers [150]. They identified 26 differential metabolites, including γ-aminobutyric acid (GABA), lysine, and glutamate, which are involved in various processes such as inhibiting oxidative stress (Fig. 5) and promoting angiogenesis. The reversal of these changes by folic acid treatment suggests its potential for neuroprotection through multiple mechanisms.

Ma et al. investigated the mechanism of action of Edaravone (EDA), a medication used clinically for IS, despite an unclear mechanism [151]. Using metabolomic analysis of urine and serum samples, they found that EDA treatment normalized levels of metabolites involved in valine, leucine, isoleucine biosynthesis, and taurine metabolism. The most significant change was observed in taurine metabolism, with EDA increasing the activity of the enzyme cysteine sulfite decarboxylase, which inhibits endothelial cell apoptosis. These findings suggest that EDA may exert its neuroprotective effects by regulating taurine metabolism and endothelial cell function (Fig. 6).

Zhou et al. investigated the effects of Danshen Chuanxiongqin injection (DSCXQ) on IS in a rat model [152]. They observed that DSCXQ treatment modulated the levels of various metabolites, including those involved in lipid metabolism (L-tryptophan, LysoPC), sphingolipid metabolism (dihydrosphingosine 1-phosphate), and oxidative stress(Fig. 5) (indole-3-methylacetate). These findings suggest that DSCXQ exerts its neuroprotective effects through regulation of multiple metabolic pathways.

Hou et al. studied the mechanism of Duzhi Wan (DZW), a TCM formula used for IS prevention and treatment [153]. They identified complement C3 (C3) and C5a complement factor receptor 1 (C5ar1) as core targets of DZW, while also pinpointing key metabolites involved in its neuroprotective effects, such as acetylcholine and inosine 5’-monophosphate. In vivo studies showed that DZW reduced levels of inflammatory markers after IS treatment, suggesting that it exerts its effects by inhibiting neuroinflammation (Fig. 7).

Ye et al. investigated the mechanism of Dengzhan Shengmai capsule (DZSM), a TCM formula used for brain dysfunction [154]. They found that DZSM treatment significantly increased the concentration of 2-ketoglutarate, a metabolite involved in the citric acid cycle and glutamate metabolism. 2-ketoglutarate can be converted to glutamate, which can act as a neurotransmitter or be converted into an inhibitor of the NF-κB signaling pathway, an inflammatory pathway. These findings suggest that DZSM may exert neuroprotective and anti-inflammatory effects through the regulation of the citric acid cycle and glutamate metabolism (Fig. 7).

Outlook

Compared to other strokes, IS is characterized by high mortality and disability rates, which seriously affect the quality of life of patients, especially those with acute IS [155]. While advancements in treatment methods and clinical management have reduced the incidence and disability rates, they remain high [156]. Therefore, new methods for early diagnosis and prognosis of IS are crucial.

Metabolomics has been applied to IS research for over a decade. Initial research focused on blood metabolomic changes during IS to identify potential biomarkers. However, these haven’t been widely used in clinical practice for diagnosis, progression assessment, or prognosis of IS. Cerebrospinal fluid (CSF) offers a better reflection of metabolic changes during IS due to its proximity to the brain environment [157]. Identifying CSF biomarkers with high sensitivity and specificity for IS would be revolutionary for guiding clinical decisions, improving survival rates, and reducing disability. It would allow for accurate diagnosis of ongoing or imminent cerebral ischemia or infarction, and predict outcomes. Current limitations in biomarker research include insufficient sample size, and inconsistent effects of factors like age and gender on metabolites. Future efforts should focus on expanding research cohorts, standardizing analysis methods, develo** a wider range of cost-effective biomarkers for clinical decision-making, and acknowledging the need for multi-biomarker approaches due to the heterogeneity of stroke [158].

While metabolomics has revealed various metabolic changes in IS across age, gender, and severity, its application in understanding the pathogenesis of IS remains limited, partly due to high costs [159]. Prior research focused on downstream damage mechanisms like excitotoxicity [160], early inflammatory damage [161], oxidative stress response [162], immune response [163], and various forms of cell death [164]. This review proposes a bolder approach – linking metabolic changes to the mechanisms of post-IS damage. By doing so, metabolomics can reveal the pathogenesis of IS and deepen our understanding of the disease.

Another recent application of metabolomics in IS research is exploring drug targets for TCM. This approach shows promise in analyzing the material basis of TCM efficacy in IS and develo** new TCM drugs for the disease. The rapid development of emerging metabolomic technologies like stable isotope tracing metabolomics and mass spectrometry imaging space metabolomics will create new opportunities for IS diagnosis, TCM drug development, and a more comprehensive understanding of disease mechanisms [165]. Stable isotope tracing metabolomics can elucidate the role of metabolites in metabolic pathways, while mass spectrometry imaging space metabolomics allows for quantitative localization analysis of metabolites in large samples, both of which can provide valuable insights [166].

Limitations

Metabolomics is a high-throughput technology used to study the metabolic status of organisms. Although it has important application value in early disease diagnosis, drug development, and other aspects, it also has limitations [167]. Firstly, it requires processing a large amount of data, which can be time-consuming and requires skilled personnel [168]. Furthermore, standardization across various aspects of metabolomics techniques, including sample collection, quality control, and data analysis, is crucial to enhance the data’s credibility and comparability. This will require addressing current limitations in these areas.

IS is a complex multifactorial disease. In this article, we mainly review the potential pathogenesis of metabolic factors in IS. However, factors such as oxidative stress, apoptosis, pyroptosis, and inflammatory damage are also important contributors to post-IS brain injury. Additionally, research on other omics technologies [169], such as transcriptomics (gene expression), proteomics (protein analysis), imaging omics, and single-cell sequencing are limited in studying IS progression, drug targets, and pharmacological mechanisms. Integrating metabolomics with these other approaches in future studies can provide a more comprehensive understanding of IS.

Conclusion

In conclusion, future studies on IS using metabolomics should consider integrating these new technologies with other omics approaches like transcriptomics and proteomics. This comprehensive and systematic analysis will enhance our understanding of the pathological mechanisms of IS and the efficacy mechanisms of TCM, ultimately promoting the development of precision medicine for IS.