Transcriptomics, single-cell sequencing and spatial sequencing-based studies of cerebral ischemia

Abstract

With high disability and mortality rate as well as highly complex pathogenesis, cerebral ischemia is highly morbid, prone to recurrence. To comprehensively understand the pathophysiological process of cerebral ischemia and to find new therapeutic strategies, a new approach to cerebral ischemia transcriptomics has emerged in recent years. By integrating data from multiple levels of transcriptomics, such as transcriptomics, single-cell transcriptomics, and spatial transcriptomics, this new approach can provide powerful help in revealing the molecular mechanisms of cerebral ischemia occurrence and development. Key findings highlight the critical roles of inflammation, blood–brain barrier dysfunction, and mitochondrial dysregulation in cerebral ischemia, offering potential biomarkers and therapeutic targets for early diagnosis and personalized treatment. A review of the research progress of cerebral ischemic injury mechanism under the analysis of the comprehensive transcriptomics research method was presented in this article, aiming to study the potential mechanism to provide new, innovative therapeutic strategies for this disease.

Introduction

In recent years, the increasing high mortality rate due to cerebral ischemia has become a major threat to the lives and health of people worldwide [1]. Ischemic stroke is a pathological state of cerebral ischemia, which is a complex pathophysiological process due to the interaction of various factors and mechanisms [2]. Although the mechanisms of cerebral ischemia have been studied to some extent, many underlying mechanisms of this remain unknown, causing an urgent need to unravel this mystery. It is estimated that one out of six people will suffer a stroke in their lifetime, making it the second most common cause of death worldwide and the first in China. Stroke causes serious health problems for individuals while imposing a huge socio-economic burden on society. Stroke is usually divided into two main types: ischemic stroke and hemorrhagic stroke. The former is the most common type of stroke, also known as ischemic stroke and cerebral ischemia, and accounts for about 85–90% of stroke cases. Cerebral ischemia is caused by a blockage of a blood vessel in the brain, resulting from a blood clot that has formed in the blood vessel or an embolus that has dislodged from another part of the body. The obstruction of the cerebral blood vessels leads to restricted blood supply to the brain region, resulting in a lack of oxygen and nutrients, which can damage brain cells. Despite advances in thrombolytic therapy, its narrow therapeutic time window and inability to address secondary mechanisms of injury limit clinical efficacy. Research has demonstrated that multiple pathways, including inflammation, excitotoxicity, oxidative stress, apoptosis, impaired energy metabolism, and cellular autophagy, are implicated in the pathological process of cerebral ischemia-induced brain injury [3]. These mechanisms are interconnected and communicate with each other, resulting in irreversible damage to brain tissue. The sequelae of cerebral ischemia are associated with high rates of disability, and the main clinical manifestations of ischemic stroke are hemiparesis and impaired consciousness. As the global population ages, the economic burden of cerebral ischemia is increasing. For people over 60, Stroke is the most common cause of permanent disability and the second leading cause of death and dementia. As the global population ages, the urgency of addressing stroke-related issues needs to be heightened. Fortunately, today's transcriptomics, single-cell sequencing and spatial sequencing technologies are able to bridge the gap between molecular mechanisms and clinical transformation by shedding new light on the pathological mechanisms and therapeutic targets of cerebral ischemia.

Transcriptomics is the study of the transcription of all genes within a cell and the laws of transcriptional regulation. In recent years, important advances have been made in high-throughput sequencing technology, data analysis methods, and application areas [4]. Transcriptome sequencing (RNA-Seq) uses high-throughput sequencing technology to reverse the transcription of the RNA in a tissue or cell into a cDNA library for sequencing [5]. This approach allows researchers to calculate the expression of different RNAs, discover new transcripts, and, with genomic reference sequences, determine the location and splicing of transcripts. With the development of sequencing technology, transcriptome sequencing has enabled more systematic studies of transcript types and expression, providing new insights into variable splicing events, new genes and transcripts, and fusion genes. The transcriptome in cerebral ischemia can reveal the key molecular changes involved, identify key regulatory genes in ischemia-related pathways, and provide a molecular basis for targeted drug development. In addition, single-cell sequencing technology, based on second-generation sequencing, has also been developed significantly. Single-cell messenger RNA (mRNA) sequencing was first accomplished in 2009 and that of single cells was later completed in 2011 [3]. Single-cell sequencing is currently one of the most advanced sequencing technology, which enables the study of cellular and microenvironmental heterogeneity at the individual cell level. This technology will enable more comprehensive studies of thousands of single cells in terms of transcription, paving the way for molecular mechanisms of single nucleotide variation, copy number variation and structural variation [6]. Beyond developmental biology, single-cell transcriptome sequencing (scRNA-seq) and related technologies have provided new ideas for the discovery of new molecular markers, rare subpopulations and evolutionary patterns [7]. The application of this technology in cerebral ischemia is also broad, as it can identify specific cell types in cerebral ischemia and provide a more precise molecular basis for disease typing by identifying ischemia-specific cell subpopulations and molecular markers in the clinic. However, single-cell sequencing technology cannot truly reflect the spatial information of cells, whereas spatial transcript-mics (ST) can precisely localise gene expression among cells in tissues, thus allowing for a more intuitive revelation of gene expression in different regions of tissues. Recent technological advances leveraging second-generation sequencing and imaging sequencing have demonstrated that spatial transcriptomics is capable of can systematically detecting most, if not all, levels of gene expression throughout tissue space [8]. With this technology, we can better understand the spatial distribution of specific molecules in cerebral ischemic injury and pinpoint the area of ischemic injury for subsequent precise treatment.

This paper reviews the recent applications of transcriptomics, single-cell transcriptomics, and spatial transcriptomics in the study of cerebral ischemia to provide theoretical support for clinical treatment, open up new possibilities for future therapeutic research directions.

Comprehensive transcriptome sequencing technology

Transcriptomics

Transcriptomics is the study of gene transcriptional activity and the laws of transcriptional regulation in cells at the holistic level. It reveals the regulatory networks and molecular pathways of gene expression after cerebral ischemia by measuring changes in the transcriptional levels of genes (Fig. 1). Velculescu et al. [9] were the first to introduce the concept of transcriptome for the study of gene expression in yeast cells. In a strict sense, the transcriptome refers to the process by which a cell produces messenger RNAs (mRNAs) coding for proteins in a particular environment; In a broader sense, it includes all kinds of RNAs expressed by a particular cell, tissue, or an entire organism under a particular physiological or pathological condition, including mRNAs as well as other non-coding RNAs that play a regulatory role in the process of gene expression and development (ncRNA), such as ribosomal RNA (rRNA), transfer RNA (tRNA), long-chain non-coding RNA (lncRNA), circular RNA (circRNA), and microRNA (miRNA) [10]. Transcriptome analysis provides helps in understanding the structure and function of genes, regulation of gene expression, and genomic plasticity, revealing key alterations in the process of biological changes that trigger human diseases, facilitating studies of underlying mechanisms and providing better molecular diagnostics and clinical treatments [11]. However, transcriptomics still has limitations in terms of experimental design, data analysis and technology application, for instance, it is difficult to obtain sufficient quantity or high quality of mRNA in some experiments. What’s more, the determination of sequencing depth and number of replicates depends on multiple factors, such as experimental purpose, sample complexity and statistical efficacy, which lacks clear standards and requires complex calculations and assessments, thus the results are greatly affected by different methods of differential expression analysis. Therefore, the accuracy and reliability of the analysis results are challenged by small sample sizes, low gene expression levels, complicated data integration, and inconsistent standards and methods. If above problems can be addressed, it will better facilitate the research of clinical diseases [12].

Fig. 1

Transcriptome sequencing technology flowchart

Single-cell sequencing technology

In recent years, advances in single-cell isolation, sequencing, cDNA library preparation and other analytical technologies have significantly propelled the development of single-sequencing technology (Fig. 2). The procedure of single-cell sequencing mainly includes single-cell isolation and capture, cell lysis, reverse transcription (conversion of its RNA to cDNA), cDNA amplification and library preparation [13]. Single-cell isolation and capture techniques are designed to obtain high-quality single cells with precise genetic and biochemical information that can help to reveal specific genetic and molecular mechanisms [14]. The most common single-cell isolation and capture techniques include limiting dilution, fluorescence-activated cell sorting (FACS), magnetically activated cell sorting, microfluidic systems, and laser microdissection [15]. After the conversion of the RNA into the first cDNA, amplification is performed using Polymerase Chain Reaction (PCR) or In Vitro Transcription (IVT). Among them, the former uses the Switching Mechanism At 5' end of the RNA Transcript (SMART) technique and the other is a linear amplification method. cDNA sequences of individual cells or individual nuclei can be sequenced using different sequencing techniques after obtaining them. For example, DNA Nanobal (DNB) technology first selects DNA fragments, blunts their ends and modifies at the 3' end to obtain externally extended ends of deoxyadenosine triphosphate (dATP). A Linker at the end of deoxythymidine triphosphate (dTTP) is used to connect the DNA fragments at each end. Next, the Linker was amplified for several cycles and was sequenced in single-stranded loops. On this basis, a specific strand of the PCR product is reverse-complemented with a specific molecule, which is then linked to the corresponding single strand by DNA ligase to obtain a single-stranded circular DNA library [16].

Fig. 2

Single-cell sequencing process

Single-cell sequencing technology provides a high-resolution perspective for understanding the pathophysiological process of cerebral ischemia by analyzing cell heterogeneity, revealing the molecular mechanism of disease, discovering new cell subsets, and dynamically monitoring the disease process. Concurrently, this technology assists in screening potential therapeutic targets and biomarkers, promoting the development of personalized treatment strategies, and establishing a novel approach for accurate diagnosis and treatment of cerebral ischemia. However, single-cell transcriptomics has several limitations related to experimental techniques, data processing, and multi-omics integration, which involve issues with single-cell capture efficiency and accuracy, transcription issues due to cell dissociation, insufficient throughput, high cost, and limited applicability to certain cell types. In addition, quality control that may mistakenly delete rare cells defective methods for quantifying technical noise affects the accuracy of the results; definitions of cell types are ambiguous, and the lack of standardized methods for cluster analysis affects subpopulation studies. Moreover, challenges remain in integrating transcriptome data with other information, and spatial resolution methods exhibit low sensitivity. These limitations hinder its further application in biological research [13].

Spatial transcription technology

Spatial transcription technologies commonly used today are classified as second-generation sequencing (NGS) and imaging sequencing [17] (Fig. 3). Second-generation sequencing technologies capture the transcriptome of an entire tissue by adding spatial barcodes before library preparation, allowing for the unbiased study of large areas of tissue without the need to select a particular region or set of gene targets [18]. Imaging sequencing technologies can be subdivided into in situ hybridisation (ISH) and In Situ Sequencing (ISS); In Situ Sequencing reads transcriptional sequences directly from tissues, while in situ hybridisation detects target sequences through the hybridisation of complementary fluorescent probes. Imaging-based sequencing technology generates a gene expression matrix after processing the images and then combines the gene expression data to depict the cells [19].

Fig. 3

Spatial sequencing technology process

Spatial transcriptome sequencing has overcome the spatial limitations of traditional technology in the study of cerebral ischemia diseases. By accurately locating the spatial distribution of gene expression, it reveals the molecular heterogeneity between ischemic areas and surrounding tissues, especially the key difference between ischemic core areas and penumbra. This technology facilitates the analysis of the dynamic interaction of cell types in different brain regions, including the coordination or imbalance of neurons, glial cells, and the vascular system. In addition, it clarifies the temporal and spatial evolution law of the local microenvironment, such as inflammation, oxidative stress, and angiogenesis. This provides a significant foundation for identifying disease-specific targets, optimizing neuroprotective strategies, and accurately determining intervention time windows. Consequently, this technology significantly promotes the study of the pathological mechanisms of cerebral ischemia and its clinical transformation. Although spatial transcriptomics has been greatly improved on the basis of the original technology, further research is still urgently needed in terms of resolution and number of detections. For example, although spatial transcriptomics based on in situ hybridisation and fluorescence microscopy can detect the spatial distribution of transcripts with high resolution and accuracy, the total number of RNA transcripts that can be detected is limited, which makes it difficult to obtain comprehensive transcriptome information; although next-generation sequencing-based methods can capture expressed RNAs from spots in space on a whole-transcriptome scale, each spot (with a radius of 10–100 µm) may contain multiple cells, making it difficult to obtain comprehensive information on the transcriptome. May contain multiple cells, which limits spatial resolution to the single-cell level and hinders the capture of whole-transcriptome data at single-cell resolution in space [20].

Common data analytics pipelines and challenges

Normalisation methods

Normalisation is a key step in the analysis of RNA-seq data, aiming to correct data bias and enable more accurate comparison of gene expression [21]. In RNA-seq data, gene expression level is measured by the number of transcripts in the mapped fragments, but it is affected by a variety of factors, such as the sequencing depth, gene length, and GC content, resulting in biased data, and therefore, the data need to be normalised. The purpose of normalisation is to eliminate or reduce the technical variations in RNA-seq data, including the bias caused by library size differences, gene length and GC content, so as to make comparisons of gene expression between different samples more reliable, accurately identify differentially expressed genes (DEGs) and provide reliable data support for subsequent studies. Common normalisation methods include Total Counts (TC), which adjusts the raw reads by the total library size of the sample; Upper Quartile (UQ) and Median (Med) scale the raw reads by calculating a normalisation factor based on the upper quartile and median of the sample's genetic reads, respectively; DESeq uses the Median Ratio method, Trimmed Mean of M values (TMM) using empirical Bayesian estimation, both of which perform better in terms of detection ability but suffer from lower specificity; Full Quantile (FQ) may increase the within-sample variation in some cases; Reads Per Kilobase per Million mapped reads (RPKM) and Fragments Per Kilobase per Million mapped fragments (FPKM) take gene length into account, but the correction process may introduce bias to low abundance genes [22].

Dimensionality reduction analysis

Dimensionality reduction analysis is the process of transforming the high-dimensional data into the low-dimensional data, aiming at retaining key information, improving data processing efficiency and enhancing the interpretability of data [23]. In spatial transcriptomics, it can fuse high-dimensional gene expression data with spatial information for processing, which may facilitate subsequent visual display and cluster analysis to identify biologically relevant spatial regions or cell types. Common methods contain traditional principal component analysis (PCA), non-negative matrix factorisation (NMF), etc., but they suffer from the lack of expressive power and the inability to integrate spatial information. The emerging spatial transcriptomics analysis with topic modeling to uncover spatial patterns (STAMP) approach, which combines thematic modelling with deep generative models, effectively overcoming some related shortcomings and showing good performance on multiple data sets [24].

Spatial resolution

Spatial resolution is the ability to distinguish and identify information at spatial locations during analysis and detection, and is of key importance in spatial transcriptomics technology [25]. For example, when studying gene expression in tissue sections, traditional RNA-seq analysis of homogenised biopsy tissues averages out the transcriptome information, resulting in the loss of information on the spatial location of gene expression. Spatial transcriptomics technology, however, has the spatial resolution to visualise and quantify the transcriptome while retaining the two-dimensional positional information of tissue sections by placing them on arrays of reverse transcription primers with unique positional barcodes. Therefore, it enables spatial transcriptomics technology to detect both highly expressed and lowly expressed genes. At the same time, it can accurately measure the transcript diffusion distance and distinguish the gene expression differences in different cell layers [26]. However, its detection accuracy is not high enough, so the technology still needs to be enhanced to resolve the spatial pattern of gene expression in cells.

Integrated multi-omics

The above advances in histological techniques have revolutionised our understanding of cerebral ischemia, while we recognise that each method has inherent limitations that constrain a comprehensive understanding of the pathology (Supplementary Table 1). Bulk transcriptomics, although powerful in profiling genome-wide expression changes, suffers from average cellular heterogeneity and poor detection of low-abundance transcripts. Moreover, single-cell RNA sequencing addresses cellular diversity but lacks spatial context and struggles with technical noise. Last but not least, spatial transcriptomics preserves organisation but faces resolution limitations and may include multiple cells per sequencing site. These above methodological gaps highlight the need for integrated multi-omics strategies.

Synergistic integration of complementary histological layers can effectively address these limitations. Proteomic validation of transcriptomic data confirms whether observed mRNA changes translate into functional protein changes, as demonstrated by mass spectrometry validation of TNF-α elevation in neuroinflammatory pathways [27]. Metabolomic correlation analyses linked gene expression patterns to biochemical transformations, such as linking upregulation of glycolytic transcripts to lactate accumulation in the penumbra. This multidimensional approach translates isolated observations into a coherent biological narrative. For cellular resolution studies, the integration of single-cell data with spatial techniques generates a more complete picture. Computational deconvolution methods can map scRNA-seq defined subpopulations to tissue coordinates, thereby pinpointing cell types such as Clec7a microglia to infarct boundary zones [28]. Simultaneous measurement of transcripts and proteins by CITE-seq validates cellular status, while imaging mass cytometry enables protein-level spatial validation of transcript findings. This integration addresses the spatial ambiguity of single-cell data and the cellular ambiguity of spatial approaches.

The true power of multi-omics lies in deciphering complex pathological cascades (Fig. 4). In neuroinflammation, the combination of scRNA-seq and phosphorylated proteomics revealed coordinated activation of the TREM-1/SYK axis at the transcriptional and post-translational levels [29]. Blood–brain barrier dysfunction studies also benefit from the correlation of endothelial Slc22a8 downregulation (scRNA-seq) with leaky areas (spatial transcriptomics and MRI) [30]. Mitochondrial dysfunction studies combine neuronal COX6C deletion (scRNA-seq) with oxidative phosphorylation defects (metabolomics) and subcellular damage patterns (imaging). These above examples illustrate how multi-omics transcends the capabilities of any single approach. In a word, this integrative paradigm has fundamentally advanced cerebral ischemia research by revealing novel properties that cannot be discovered by single-method studies. A comprehensive understanding of spatiotemporal dynamics, cellular interactions, and molecular networks provides unprecedented opportunities to develop targeted therapies. With the development of the technology and analytical methods, multi-omics approaches will increasingly lead to precision medicine strategies for cerebral ischemia treatment and rehabilitation.

Fig. 4

Multi-omics studies of cerebral ischemia

The pathophysiologic features and applications of cerebral ischemia

The pathophysiologic features of cerebral ischemia are characterized by an inflammatory response, blood–brain barrier impairment, mitochondrial dysfunction, and oxidative stress, among others. These features have been progressively unveiled through advancements in transcriptomic technologies. These technologies collectively provide a multi-scale view of cerebral ischemia, from regional vulnerability to cellular-level responses, accelerating the discovery of therapeutic targets and biomarkers (Supplementary Table 2).

Inflammation response

The inflammatory response is a common pathological condition following the onset of cerebral ischemia, despite its unclear underlying mechanisms, which is closely related to the mechanism of cerebral ischemic injury and its prognostic treatment. N-myc downstream regulatory gene 2 (NDRG2) is a cytoplasmic protein belonging to the family of N-myc downstream genes, which is widely expressed in a wide range of normal tissues. and the genes also have a variety of biological functions, such as cellular growth and differentiation, stress response and hormone response, etc. [31]. NDRG2 expression is upregulated during the formation of reactive astrocytes, and high levels of lactate also in turn increase intracellular NDRG2 expression levels. In addition, NDRG2 deficiency triggers dysregulation of inflammatory and immunomodulatory activities, leading to elevated levels of TNF-α, which results in further disruption of the local microenvironment. Xu et al. [32] learned by RNA-seq analysis that NDRG2 silencing causes the up-regulation of a large number of inflammatory and chemokine factors, as well as inflammation-related signalling pathways, which suggests that the lactate/NDRG2/TNFα signalling axis may provide extended mechanistic clues and valuable approaches to regulating neuroinflammation in the early stage of ischemia. TNF-α is a pro-inflammatory cytokine which plays a crucial role in the development of cerebral ischemia [33]. Li et al. [27] performed transcriptome sequencing in the cerebral cortex of ischemia–reperfusion (I/R) -injured mice, and found that the genes involved in inflammatory signalling pathways were upregulated. In addition, TNF-α levels in the cerebral cortex and serum were also significantly elevated after I/R, and TNF-α promoted platelet activation and aggregation through the RIP1/RIP3/AKT pathway, exacerbating brain injury, whereas the addition of the anti-TNF-α antibody attenuated ischemic injury, indicating a protective effect of the anti-TNF-α antibody against I/R injury in mice. Clec7a, a pattern recognition receptor on microglia, may mediate cell–cell interactions [34]. Shi et al. [35] first performed genome-wide transcriptional profiling and then discovered that compared with young mice, brain infiltration of peripheral immune cells in aged mice had significantly reduced interactions with microglia in neighbouring cells after cerebral ischemia. In addition, Clec7a protein expression, immune cell infiltration, and young and old microglia had cell–cell interactions after cerebral ischemia. These findings have provided future elucidation of microglia-dependent cellular and molecular pathways that may contribute to the discovery of new therapeutic targets to promote functional recovery in elderly cerebral ischemia patients. Receptor-interacting protein kinase 2 (RIPK2) is a serine/threonine kinase that propagates inflammatory signals by interacting with pattern recognition receptors (PRRs) and subsequently activating the TAK1, NF-κB and MAPK pathways [36]. It is now well-acknowledged that RIPK2 plays a key role in inflammatory processes, activating multiple pro-inflammatory and cell death pathways. Through RNA sequencing analysis, there are important differences between microglia from RIPK2^−/− mice and RIPK2^+/+ mice. RIPK2 has been shown to take part in the acute and late stages of cerebral ischemia injury and are engaged pathologically in the progression of cerebral ischemia injury by promoting neuroinflammation and increasing infarct size. This identifies RIPK2 as a novel therapeutic target for the treatment of cerebral ischemia and other neuroinflammatory diseases. Anxa2 is a multifunctional calcium (Ca²⁺) and phospholipid-binding protein which involved in the inflammatory response [37]. Tian et al. [38] found that multiple immune-inflammation-related pathways, including the NF-κB signalling pathway, can be activated in the hippocampus, after brain I/R by brain tissue transcriptomics. In addition, Anxa2 showed a significant increase and exhibited specific upregulation in microglia. It is found that Anxa2 regulates microglia activation and inflammation by mediating the downstream NF-κB signalling pathway, which subsequently regulates neurons through brain I/R. This finding suggests that Anxa2 deficiency may have an impact on attenuating microglia-induced inflammatory responses and protecting neurons from death, implying that intervening in the targeted expression of Anxa2 in microglia may be a potential therapeutic strategy to alleviate the progression of cerebral ischemia pathology. Perivascular Macrophages (PVM) are innate immune cells closely associated with the brain microvasculature and are now a major contributor to neurovascular dysfunction [39]. Drieu et al. [40] found, through transcriptomics data analysis, that alcohol-exposed mice exhibit an exacerbation of the initiation of the inflammatory response of the neural vasculature after cerebral ischemia, resulting in larger ischemic lesions. Perivascular macrophages (PVMs) are the aggravation of the effect of alcohol Key players, and their specific depletion prevented alcohol-induced exacerbation of ischemic lesions, which provides a new pathway to block the exacerbation of alcohol-induced neurovascular inflammatory responses triggered after cerebral ischemia. Cavin-1 (CAV1), associated with endocytosis, extracellular matrix organisation, cholesterol distribution, cell migration, is an oncogenic membrane protein [41]. Cav-1 can determine the integrity of the endothelial barrier after cerebral ischemia. Zhang et al. [42] learned by transcriptome analysis that Rxrg was the most significantly changed molecule after the Cav-1 knockdown. The decrease in RXR-g was more pronounced after Cav-1 ablation. This may indicate a specific role for endothelial Cav-1/RXR-g signalling in diseases involving the thromboinflammatory circuit. Triggering receptor expressed on myeloid cells (TREM)-1, an amplifier of the innate immune response, is a critical regulator of inflammation [43, 44]. Xu et al. [29] altered expression of multiple immune genes in the peri-infarct region after cerebral ischemia was identified by RNA-seq, among which TREM-1 and NLRP3 inflammatory vesicle-related genes were significantly elevated. Microglial TREM-1 activates the CARD9/NF-κB signaling pathway and NLRP3 inflammasome by interacting with SYK, which subsequently leads to the production of inflammatory cytokines and chemokines and focal death. Blockade of TREM-1 inhibits activation of the TREM-1/SYK pathway and the associated inflammatory response, and it also improves cerebral ischemia prognosis. Through these studies, we found that cerebral ischemia is closely related to inflammatory responses using transcriptome analysis.

The impairment of blood–brain barrier

The blood–brain barrier (BBB) is a unique barrier structure consisting of endothelial cells (ECs), tight junction cells (TJs), pericytes, astrocyte telangiectasia, and the basement membrane [45]. When blood–brain barrier function is impaired, it may lead to neuroinflammation and the release of inflammatory cytokines, which in turn contributes to the impairment of the blood–brain barrier. This phenomenon of the impairment of the blood–brain barrier is an important pathophysiological feature of acute cerebral ischemia.

Factor VII-activated protease (FSAP), also known by its genetic name hyaluronan-binding protein 2 (HABP2), is a plasma serine protease with multiple functions [46]. Tian et al. [47] found by transcriptomics that FSAP may regulate endothelial function through the Wnt signaling pathway. Studies have demonstrated that FSAP inhibition could exert pro-angiogenic and neuroprotective effects by activating the Wnt5a signaling pathway and improve neurological function in experimental cerebral ischemia models, suggesting that FSAP may be a potential target for the treatment of AIS–LVO. Similarly, sodium/proton exchanger heterodimer 1 (NHE1), a membrane glycoprotein encoded by the NHE1/SLC9A1 gene, is ubiquitously expressed and plays a key role in the regulation of intracellular pH and volume homeostasis [48, 49]. Song et al. [50] found 340 genes significantly differentiated after cerebral ischemia by RNA-seq analysis, with the Wnt/β-catenin signalling pathway significantly up-regulated. They found that deletion of NHE1 in astrocytes can play a protective role in cerebral ischemia by activating the Wnt/β-catenin signaling pathway, reducing endothelial cell transcytosis, decreasing BBB injury, and increasing vascular repair and cerebral blood flow. This finding reveals the critical role of astrocytes in regulating the cerebrovascular Wnt/β-catenin signaling pathway and provides a new potential target for the treatment of cerebral ischemia. In addition, non-muscle myosin II (NMM II) is a motor protein essential for the maintenance of intracellular homeostasis and is involved in a variety of cellular functions [51]. Gong et al. [52] found that NMMHC IIA is involved in the regulation of BBB-related genomic changes in cerebral I/R injuries by transcriptomics and that Hippo and PPARγ/NF-κBp65 signalling regulation promotes brain endothelial barrier integrity by preventing TJs degradation. Furthermore, the mechanism through which NMMHC IIA controls signalling pathways is complex and involves multiple transcription factors, including YAP, PPARγ and NF-κB. Thus, NMMHC IIA may play a crucial regulatory role in I/R-induced brain endothelial barrier damage by modulating multiple signalling pathways. Slc22a8, also known as organic anion transporter protein (OAT3), is primarily found in the kidney, but it is also detected in almost all barrier epithelia within the body, along with endothelial and other cells [53]. Liu et al. [30] uncovered by transcriptomics that Slc22a8 overexpression had a significant protective effect on blood–brain barrier integrity after I/R, which improves TJ protein loss and reduces blood–brain barrier leakage. In addition, Slc22a8 overexpression increased the transcriptional activity of β-catenin and the level of the key TJ protein Claudin-5. These findings suggest that downregulation of Slc22a8 in endothelial cells may be a key mechanism of blood–brain barrier disruption after cerebral ischemia. Therefore, targeting Slc22a8 may provide a novel therapeutic approach to enhance blood–brain barrier integrity and improve cerebral ischemia prognosis. Osteopontin (OPN), encoded by the secreted phosphorylated protein 1 (Spp1) gene, is a multifaceted stromal cell glycoprotein secreted by a variety of immune cells, such as macrophages and T cells. As an immunomodulatory cytokine, it is highly expressed by bone-marrow-derived macrophages (BMMF) and regulates a variety of cellular immune responses including migration, communication, and immune responses [54]. Spitzer et al. [55] revealed that bioinformatic dissection of the neurovascular unit (NVU) transcriptome has identified osteobridging proteins as a potential therapeutic target for the dysregulation of multiple NVU cell types in acute cerebral ischemia. Through transcriptome analysis of NVU cells, the researchers identified hundreds of genes with altered expression across cell types after cerebral ischemia, including genes involved in pathways, such as angiogenesis, extracellular matrix interactions, and neuroinflammatory responses. Moreover, OPN was found to be dysregulated in multiple NVU cell types and to play a negative role in cerebral ischemia, leading to blood–brain barrier (BBB) destruction. Antibody treatment that neutralises OPN improves NVU cell function, restores BBB integrity, and improves survival and neurological outcomes. These findings highlight that OPN, a novel target for acute cerebral ischemia therapy, may augment existing reperfusion therapies and reduce the adverse effects of BBB destruction. Transcriptome sequencing revealed the important relationship between cerebral ischemia and BBB injury.

Mitochondrial dysfunction

Mitochondria are not only the energy factories of the cell but also are intimately involved in other biological processes, such as calcium homeostasis, reactive oxygen species (ROS) production, hormone biosynthesis and cell differentiation [56,57,58]. Mitochondria have played an important role in many diseases. Studies have shown that mitochondrial transplantation has certain benefits on different phenotypes, such as improving behavioural assessment, reducing infarct size, decreasing ROS levels and inhibiting apoptosis, etc. Mitochondrial intervention attenuates I/R injury, improves neurological prognosis, and reduces the size of cerebral infarcts [59,60,61,62]. Xie et al. [63] showed that by transcriptome analysis, transplanted mitochondria may affect metabolic pathways, especially those related to lipid metabolism. Mitochondrial transplantation is expected to attenuate cerebral ischemia–reperfusion injury. The proposed mechanism involves selective isolation of mitochondrial components and regulation of cellular metabolism, reduction of ROS and oxygen-dependent apoptosis, Intracellular mitochondrial transfer may be linked to intercellular junctions, potentially providing new therapeutic options for cerebral ischemia. By RNA-seq analysis, Wu et al. [64] demonstrated the high expression of RNA-binding proteins (RBPs) and mitochondrial mRNAs in extracellular vesicles (HypEVs) released from neurons under hypoxic conditions. Crucially it was determined that fusion proteins serve as the pivotal regulators in this context. Relevant extracellular vesicle proteins for mRNA metabolism were identified. The study revealed the protective role of HypEVs in cerebral ischemia, highlighting that fusion proteins and the mitochondrial mRNAs they carry are released through HypEVs and are important for neuroprotection. These findings suggest that HypEVs could be a potential therapy for ischemic injury.

Oxidative stress

Transcriptome sequencing results showed that in the acute phase of cerebral ischemia, vascular blockade leads to calcium overload and energy depletion of neuronal cells, which, after oxygen and nutrient replenishment, induces the accumulation of ROS, causing brain injury. In addition, the high level of ROS can lead to neuronal apoptosis [65]. γ-Glutamylcysteine (γ-GC) has been regarded as a precursor of glutathione (GSH) substance, which has the effect of alleviating oxidative stress and neuronal apoptosis [66]. LI et al. [67] found that γ-GC may play a neuroprotective role by attenuating apoptosis, ameliorating oxidative damage and alleviating endoplasmic reticulum stress through transcriptomic analysis. It was found that γ-GC inhibited the activation of endoplasmic reticulum stress-related proteins PERK and IRE1α, reduced the expression of downstream pro-apoptotic proteins, and blocked the apoptotic signalling pathways including PERK–eIF2α-CHOP and IRE1α-TRAF2–JNK. γ-GC alleviated the oxidative stress and inhibited neural apoptosis induced by endoplasmic reticulum stress through the enhancement of GSH and the removal of ROS. Neuronal apoptosis, providing a potential new pathway for neuroprotection in vitro and in vivo and new possibilities for cerebral ischemia treatment. Nuclear factor E2-related factor-2 (Nrf2) is an important regulator of the antioxidant cell defense system. Heme oxygenase-1 (HO-1) possesses anti-oxidative stress properties and is now thought to protect tissues by repairing redox balance and reducing inflammation [68]. Studies have shown that stimulation of the Nrf2/HO-1 pathway is effective in attenuating oxidative stress injury. Induction of Nrf2 expression may serve as an effective target to ameliorate oxidative stress injury and mitochondrial protection [69,70,71]. The above results suggest that targeting Nrf2 may ameliorate oxidative stress injury in cerebral ischemia and offer a new avenue for the treatment of cerebral ischemia.

MicroRNA expression

miRNA is a single-stranded non-coding molecule consisting of 18–25 nucleotides that controls gene expression by promoting the degradation or inhibiting the translation of mRNA [72]. Kong et al. [73] revealed that cerebral ischemia significantly affected the expression profile of miRNAs in circulating NK cells using microRNA sequencing, with miRNA-451a and miRNA-122-5p expression levels significantly increasing. These miRNAs may regulate NK cell function after cerebral ischemia and affect the defence mechanism of the immune system. This finding uncovers a new mechanism for the effects of cerebral ischemia on the immune system and provides clues for understanding the increased risk of infection after cerebral ischemia. Loppi et al. [74] combined de novo transcriptomics data with miRNA sequencing to identify neuron-specific miRNome, miR-127 were identified in the brain, which is highly expressed in neurons and has been associated with increased cell death in senescent, lipopolysaccharide injected ischemic mice, thus shedding new light on the role of miRNAs in the brain under such conditions. Previous studies have shown that miR-127 is associated with apoptosis and proteasome activity. Notably, a reduced level of miR-127 has been observed in elderly cerebral ischemia patients, suggesting a potential influence on the prognosis of cerebral ischemia in the elderly.

Gene expression pattern

Gene expression pattern is a common pathophysiological response after cerebral ischemia, and the mechanisms involved are wide-ranging. Current studies have not been able to identify this crucial yet underlying development. Filippenkov et al. [75] analysed the gene expression change pattern in the contralateral hemisphere of the rat brain after a localised cerebral ischemia by RNA-seq. They identified four patterns of gene expression, including genes specifically expressed in the hemisphere on the side of injury, genes specifically expressed in the contralateral hemisphere, bilateral isotropic changes, and bilaterally inverted changes. The study found that cerebral ischemia not only affects the injured region, but also induces gene responses in the contralateral hemisphere. A total of 164 genes with significant changes in mRNA levels existed, and 96 genes were up-regulated and 68 genes were down-regulated, and the up-regulated genes were mainly related to inflammation, immune response and stress response, such as orexin receptor, IL6, IL7 and other signalling pathways. The down-regulated genes were mainly associated with neurotransmission and metabolic processes. Although this study lacked dynamic timepoint analysis to track the changes in gene expression over time, and further functional experiments are still needed to verify the specific roles of these genes after cerebral ischemia, it still revealed the potential role of the contralateral hemisphere after cerebral ischemia, and the changes in its gene expression provide a new perspective to understand the repair and neuroprotection of the brain after cerebral ischemia.

Single-cell regulatory mechanisms

Single-cell transcriptomics techniques reveal a complex molecular regulatory network during cerebral ischemia. A study by Cho et al. [76] analysed the circulating immune cell landscape in patients with mild cerebral ischemia. An increase in circulating NK cell populations in the early stages of mild ischemia averts the occurrence of post-ischemic infections by generating more cytokines. The IFN-p38 MAPK–nuclear factor kappa B (NFκB) pathway is the main signaling cascade for activating NK cells. Recent single-cell studies [3, 13] show this pathway exhibits cell-type specific activation patterns, with NK cells demonstrating particularly robust NFκB response compared to other lymphocytes. In general, NFκB activation induces pro-inflammatory and pro-apoptotic responses. In cerebral ischemia, NFκB signaling is involved in acute responses and plays a role in blood–brain barrier breakdown, inflammation, and neuronal cell death. Spatial transcriptomic analyses [8, 26] reveal these effects are most pronounced in peri-infarct regions, where blood–brain barrier disruption is most severe. NFκB activity is also associated with cerebral ischemia severity, as cerebral ischemia size decreases with inhibition of NFκB in animal models28 and cerebral ischemia patients. The study showed that the NF-κB pathway of NK cells would be a more specific target for the treatment of mild cerebral ischemia, revealing the role of immune cells in the inflammatory response after cerebral ischemia. This aligns with emerging precision medicine approaches [11] that target specific cell populations rather than systemic immune suppression. A study by Kirdajova et al. [77] found that a transient astrocyte-like NG2 glial subpopulation expressing the NG2 antigen only emerged after permanent cerebral ischemia, whereas NG2 glial cells play a key role in maintaining brain health as they can proliferate and produce new oligodendrocytes during development and adulthood. Single-cell RNA-seq studies [7, 20] have identified this as part of a broader phenomenon of glial cell plasticity following ischemic injury. Moreover, the study utilized single-cell sequencing to classify the cells based on gene expression patterns. After using single-cell sequencing techniques to cluster the cells to study neurogenic potential after cerebral ischemia, it was found that NG2 glial cell subsets expressing reactive astrocyte markers (GFAP) are present only momentarily after focal cerebral ischemia [78]. NG2 glial cells may be involved in the repair process after brain injury. This transient population appears to represent a critical transitional state in the glial response, as demonstrated by trajectory analysis in recent studies [13, 74]. In addition, Jin et al. [79] used single-cell sequencing to discover the effects of senescence on cerebral ischemia recovery. Senescence was found to impair the ability of microglia (MG)/macrophages (MΦ) to promote angiogenesis and oligodendrogenesis through paracrine cells. Notably, the permanent depletion of MG/MΦ was shown to impede angiogenesis and oligodendrogenesis in both young and old mice, thereby hindering long-term cerebral ischemia recovery. Furthermore, MG and invasive MΦ contribute to the brain repair process after cerebral ischemia as these two myeloid cell populations express various paracrine factors (VEGFA, IGF1, Spp1, etc.). Romay et al. [80] also found that a significant decrease in Notch3 signalling in the cerebral vasculature that significantly associated with ageing. Inactivation of Notch3 was found to alter the regulation of calcium and contractile function. Notch3 is critical for maintaining the differentiated state of smooth muscle cells, and the loss of Notch3 leads to a progressive loss of contractile function, which adversely affected vascular reactivity, Defects and loss of contractile force lead to structural changes in the microvascular system, some of which are unique to the brain (microaneurysms and vascular beading), and a combination of impaired vasocontractile force and altered spatial structure around blood vessels is manifested as reduced lymphatic flow leading to significant accumulation of glycosaminoglycans in the brain parenchyma. This phenomenon ultimately manifests as the neuronal transcriptional alterations observed in neurodegenerative disorders, characterized concurrently by a reduction in chaperone expression (for instance, HSP70, evidenced by decreased levels of Hspa1a and Hspa1b transcripts) and an upregulation of ubiquitination pathways along with heightened oxidative stress (as indicated by changes in NDUFA4 and COX6C gene expression).

LD Li et al. [81] showed that endoplasmic reticulum (PER)-related genes, especially ADCY5, CAMK2A, PLCB1, NTRK2, and DLG4, may be the key genes triggering apoptosis during cerebral ischemia/reperfusion injury in a hypoxic environment by single-cell sequencing. ADCY5, a member of the adenylate cyclase family, is widely present in the brain, especially in the striatum and olfactory nodes, where it is selectively expressed at high levels and can promote the conversion of adenine-5'-triphosphate to 3', 5'-cyclic adenosine-5'-triphosphate (cAMP). CAMK2 is a calcium-reactive serine/threonine kinase composed of four distinct subunits (CAMK2A, 2B, 2D, and 2G), of which type 2A is associated with cognitive function, neurodevelopment, and may provide neuroprotection to neurons in CI/reperfusion through NF-KB signaling. PLCB1, located primarily in brain regions associated with various psychiatric disorders, is a key biomarker of cognitive improvement, modulates the abundance of PLCB1 expression in the cerebral cortex, and helps to improve associated cognitive dysfunction. The NTRK2 gene encoding tropomyosin receptor kinase B (TrkB), which plays an important role in neuronal development and survival, may reduce the potential therapeutic benefits of ischemia/reperfusion injury by increasing TRKB expression. DLG4 is responsible for the synthesis of postsynaptic dense protein 95 (PSD95), a member of the membrane-associated guanylate kinase family. As a major scaffold protein in excitatory postsynaptic densification, PSD95 contributes to synaptic plasticity and can be used to modulate neurotransmitter release of synaptic vesicles by methylating DLG4 and activating neuronal and neurovascular coupling, thereby stimulating interneuronal neuroprotection and reducing brain damage. In addition, a study by Wang et al. [82] revealed a novel phenotype of B cells after cerebral ischemia, Known as MLB, which is defined as the co-expression of B cell and macrophage markers at the gene and protein levels. Moreover, in cerebral ischemic disease, these macrophage-like B cells display elevated expression and enhanced phagocytosis and chemotaxis. Single-cell sequencing also revealed the mechanism through which transcription factors CEB or Pax mediate on inducing transdifferentiation of this cell. The forced expression of CEBP was found to drive the transdifferentiation of B cells into macrophages across all developmental stages, from pre-B cells to mature B cells. MLB showed inhibition of Pax5, which is consistent with other studies showing that epigenetic silencing of Pax5 induces B lymphoma cells to develop into macrophages, but not into other cell type, which may play a role in immune regulation after cerebral ischemia. Benakis et al. [83] suggested that microglia response to early activation in response to cerebral ischemia is regulated by different T cell subsets. Using single-cell sequencing technology with single-cell trajectory inference analysis to reveal that T cells influence the transition of microglia from normal homeostatic microglia root clusters to reactive cerebral ischemia microglia terminal clusters, T cells regulate gene sets associated with chemotactic mediated mechanisms (Ccl2, Ccl7, and Cxcl10) in microglia, and also mediate downregulation of markers associated with reactive microglia, such as the expression of Trem2. These markers have previously been described as regulated by the T cytokine IL-10. T cells overexpressing IL-10 down-regulated the complement system of microglia. They prevented excessive elimination of synapses and consequent neuronal dysfunction, suggesting that the manipulation of lymphocytes and the cytokines they secrete may be an effective therapeutic strategy to modulate inflammation and improve prognosis after cerebral ischemia. Furthermore, according to Ashley McDonough et al. [84] The use of single-cell sequencing has revealed that specific gene expression profiles in microglia shift microglia to an immunomodulatory or protective phenotype in response to cerebral ischemia. It can be targeted by administering toll-like receptor agonists, interferon beta cytokines, and drugs that induce preconditioning through cross-tolerance mechanisms. This sheds lights on targeting specific molecular signalling pathways in microglia to develop novel and effective drugs for cerebral ischemia. In addition, Liu et al. [85] used single-cell sequencing technology to discover a new synergistic mechanism of 11-keto-beta-boswellic acid and Z-boswellic acid (KBA) in cerebral ischemia, which synergistically regulate inflammatory responses in microglia as well as cell metabolism and iron death in astrocytes. The study identified Spp1 may be the target of the synergistic action, allowing the precise development of drugs targeting Spp1 for the treatment of brain injury. Therefore, Spp1 can be used to treat cerebral ischemia as a potential therapeutic approach.

The use of single-cell sequencing technology provides ideas for finding new markers. These studies suggest that the study of the mechanisms of cerebral ischemia can help to better understand the pathogenesis of cerebral ischemia and provide new ideas for the treatment of cerebral ischemia.

Cellular spatial regulatory mechanism

The spatial context provides information for cell characterization and functional analysis. Spatial transcriptomics is the spatially resolved and high-dimensional assessment of gene transcription [86]. This technology measures gene expression in situ,determining the cell type structure of a tissue, searching for cell–cell interactions, and monitoring molecular interactions between different components of a tissue [87]. Meanwhile, advancements in sensitivity, multiplexing and throughput are driving rapid progress in this field. Cutting-edge developments have made it possible to characterize cells in their tissue environments, further improving our understanding of tissue structure and its molecular basis in health and disease [86].

Limited knowledge exists about cell type- and location-specific specific molecular changes that occur in the center versus the periphery of ischemic injury. Even less is understood about cell type- and location-specific molecular changes. Therefore, spatially resolved single-cell transcriptome analysis of cerebral ischemia is urgently needed. The application of spatial transcriptomics enables the identification of ischemic regions with specific cellular and molecular features, which suggests that spatial transcriptomics is a promising method for the accurate assessment of brain damage following cerebral ischemia. Spatial transcriptomics, in the context of cerebral ischemia, allows for transcriptomic profiling of tissues, maintaining the spatial distribution of heterogeneous cellular and molecular features around the ischemic core while maintaining the spatial localization of sequenced molecules. This technique helps to study cellular communication after ischemic injury by orthogonal integration of single-cell data and high-dimensional spatial data from normal and diseased tissues [88]. Spatial transcriptomics experiments using a model of cerebral ischemia revealed that microglia clusters in the ischemic core and penumbra regions are distinct. Spatial transcriptomics is a landmark for diseases with focal manifestations (acute ischemia affecting specific parts of the brain or chronic cerebral small vessel disease). Spatial transcriptomics revealed ischemia-induced transcriptomic differences among specific glial cell subpopulations in the gray and white matter in the context of cerebral ischemia or chronic cerebral small vessel disease [89]. For the spatial transcriptomic analysis of ischemic hemispheres in mice, four coronal sections of the ischemic hemispheres of photothrombotic mice and the ipsilateral hemispheres of sham-operated mice were collected. These sections were observed under a light microscope, transcriptome profiles were generated, clustered and visualized, resulting in a total of 19,777 transcriptome profiles. It is possible to both visualize the transcriptional landscape within the tissue and identify gene expression profiles associated with specific histological entities. In this study, spatial transcriptomics revealed a remodeling of the overall spatial and transcriptional patterns in the ischemic hemisphere after cerebral ischemia. Clustering of spots based on different gene expression patterns identified ischemic regions, including the infarct core area (ICA), the proximal area of the peri-infarct area and the distal area of the peri-infarct area, and the distal area of the peri-infarct area. This approach enables the definition of the peri-infarct region in a clear and unambiguous manner without interference from other factors. Second, the progressive effects of the proximal and distal peri-infarct regions on cortical reorganization were assessed based on different transcriptional profiles, especially on primary and secondary motor cortices. Spatial transcriptomics revealed the spatial characteristics of these cells, exploring previously uncharacterized cellular interactions and their relationship to specific localization under cerebral ischemia [88].

Spatial transcriptomics greatly enhances biodiscovery capabilities, bridging tissue biology and genomics to further, improve our understanding of human tissues in the research, diagnostic and therapeutic environments [86, 87]. Spatial genomics technology can improve our understanding of tissue structure and its molecular basis in health and disease. However, the technology has its shortcomings, and current issues include those related to resolution and the number of transcripts that can be assessed at a given time. Although some platforms can provide transcript resolution at the single-cell level, they still have other limitations such as the number of gene probes and/or the number of transcripts that can be detected at a given time [89]. In short, this technique plays an extremely important role in studying the functional characteristics of different cell subtypes. As spatial transcriptomics continues to evolve, this will greatly contribute to our understanding of the responses within discrete brain regions following specific injuries.

Clinical value

The application of comprehensive transcriptomics in the treatment of cerebral ischemia provides an important tool for revealing the disease mechanism and developing new therapies. It can identify potential therapeutic targets, help to screen biomarkers for early diagnosis and prognosis evaluation, and provide a basis for personalized treatment strategies. Generally speaking, comprehensive transcriptomics promotes the development of precision medicine in cerebral ischemia treatment and has broad clinical application prospects (Supplementary Table 3). Its clinical significance is embodied in the following aspects.

Prognosis

Drieu et al. [40] found that alcohol-exposed mice exhibited an exacerbation of the initiating response to neurovascular inflammation after cerebral ischemia, suggesting that alcohol-exposed mice exhibited larger ischemic lesions and that cerebral perivascular macrophages were key players in the aggravating effects of alcohol as their specific depletion prevented alcohol-induced aggravation of ischemic lesions. This finding provides a new pathway to blocking alcohol-induced neurovascular inflammatory response. It also suggests a positive correlation between alcohol intake and exacerbation of cerebral ischemia. The study by Baumann et al. [90] predicted delayed cerebral ischemia following rupture of an intracranial aneurysm by analysing gene expression profiles of blood cells. Neuregulin 1 associated with vascular reactivity, inflammation, and angiogenesis, is a possible candidate predictor of delayed cerebral ischemia, suggesting that gene expression analysis can be a useful tool for diagnosing and predicting cerebral ischemia. After acute cerebral ischemia occurs, dynamics change of immune system affect clinical progression quite rapidly and significantly. The gene expression of peripheral immune cells is related to the disease state of the brain after cerebral ischemia. Therefore, it is considered a potential biomarker to help diagnose mild cerebral ischemia. The study by Cho et al. [76] analysed circulating immune cell landscapes in patients with mild cerebral ischemia, providing a new perspective on understanding the role of immune cells in the pathogenesis of cerebral ischemia.

Pharmacology

The study by Andrijevic et al. [91] provides insights into cellular recovery after prolonged warm ischemia throughout the body. Li et al. [92] found that Ginkgo biloba extracted from Shu Xuning Injection (SXNI) could inhibit granulocyte adhesion and infiltration by down-regulating the expression of key genes in the pathway. G-CSF, E-selectin, and MAC-1, and that high expression of inflammatory mediators and adhesion molecules associated with the pathway could also be significantly down-regulated by SXNI, which ameliorated the brain damage in mice with subacute cerebral ischemia and promoted the recovery of neuromotor function. This may contribute to the development of new pharmacological strategies, especially in protecting neurons and promoting functional recovery.

Neuroprotection and repair

Carpenter et al. [93] provided new insights into the diagnosis and treatment of early preneoplastic lesions by analysing the transcriptomic profiles of the donor pancreas, and these findings may shed light on neuroprotective strategies in cerebral ischemia. Xie et al. [63] suggested that the therapeutic effects of mitochondria may be related to alterations of metabolism, especially lipid metabolism, through transcriptomic data, providing clues for future studies on the mechanisms of cerebral ischemia. Wu et al. [64] found that both mitochondria-associated proteins and RNAs are transported to small extracellular vesicles (sEVs) of neurons, which contain sarcomere fusions (FUS), and that FUS contributes to the maintenance of mitochondrial morphology and function under conditions of ischemic injury. FUS-mediated hype mainly reduces mitochondria-associated apoptosis through the maintenance of neuronal synapse integrity and the reduction of mitochondria-associated apoptosis, which has specific neuroprotective effects on cerebral ischemia. In addition, mitochondrial intervention can attenuate I/R injury, improve neurological prognosis, and reduce the size of cerebral infarction [59,60,61,62].

Cell therapy

Liu et al. [94] revealed key transcriptomic changes promoting neuroplasticity after cerebral ischemia by single-cell mapping and found that infiltrating monocyte-derived macrophages (MDM) exhibited a gradual shift in fate trajectory toward activated MDM. Intercellular crosstalk among MDM,microglia orchestrated anti-inflammatory, repair-promoting microglial cell phenotypes and facilitated post-cerebral ischemia neurogenesis, suggesting that cell therapy may contribute to recovery after cerebral ischemia. Olig2 is an oligodendrocyte marker. In addition, Xu et al. [95] found that transplantation of Olig2-OL progenitor cells (OPCs) could promote the recovery of learning and memory deficits by promoting neuronal survival and increasing myelin regeneration in cerebral ischemia rats by studying single OPCs. Thus, Olig2-OPC transplantation may be an ideal cellular resource for cerebral ischemia cell therapy.

Improvement of the blood–brain barrier

Spitzer et al. [55] found significant up-regulation of osteobridging protein (OPN) encoded by the Spp1 gene in the early acute phase of cerebral ischemia, as well as a reduction in its level during long-term recovery of cerebral ischemia patients, and that treatment with anti-OPN antibodies accelerated the restoration of blood–brain barrier function in cerebral ischemia, leading to a reduction in cerebral oedema and a lower risk of haemorrhagic transformation. Therefore, anti-OPN antibody therapy may augment currently approved reperfusion therapy for acute cerebral ischemia by minimising the deleterious effects of ischemia-induced blood–brain barrier disruption. In the future, OPN could be used as a new therapeutic target for acute cerebral ischemia, and the blood–brain barrier (BBB) could be improved by anti-OPN antibodies for the treatment of cerebral ischemia.

With the trend toward multi-omics integration, there exists a huge potential for multi-omics in clinical diagnosis and treatment. Single-cell sequencing and spatial transcriptomics technologies have the potential to revolutionize the diagnosis and treatment of cerebral ischemia by providing high-resolution insights into cellular and molecular dynamics, which can be integrated into clinical workflows to identify patient-specific therapeutic targets and to monitor treatment response in real time. In addition, the development of novel therapeutic agents, such as small molecule drugs targeting specific pathways identified through transcriptomics, can be accelerated by utilizing these advanced technologies. Integrating cutting-edge technologies into clinical practice is expected to improve patient prognosis through personalized and precision medicine. Therefore, future research should focus on exploring the molecular mechanisms underlying the heterogeneity of cell populations during cerebral ischemia; conducting longitudinal studies using single-cell and spatial transcriptomics to track the progression and recovery of ischemic injury over time, leading to a deeper understanding of the temporal dynamics of the disease; and accelerating the conduct of clinical trials to validate the efficacy of identified biomarkers and molecular targets. The application of artificial intelligence and machine learning to analyze large-scale transcriptomics data can reveal new patterns and predictive models of cerebral ischemia, paving the way for more effective diagnostic and therapeutic strategies. Future efforts should prioritize bridging the gap between high-resolution molecular discoveries and clinical implementation, ensuring that advancements in multi-omics directly inform diagnostic tools, biomarker validation, and targeted therapies.

Conclusions and future perspectives

Cerebral ischemia is one of the leading causes of death and disability worldwide with high morbidity and mortality rate, and it is also a complex pathophysiological process involving multiple cell types and molecular pathways. Transcriptomics technology has revealed the regulatory networks and molecular pathways of gene expression after cerebral ischemia by analysing changes in the transcriptional levels of genes. The development of single-cell sequencing technology enables us to study the heterogeneity of cells and microenvironments at the single-cell level, providing new ideas for the discovery of new molecular markers, rare subpopulations, and evolutionary patterns. Spatial transcriptomics technology, on the other hand, can localise gene expression in different cells within a tissue at the original spatial location and discover differences in gene expression in different parts of the tissue. Conducting multi-omics studies can provide us with a deeper understanding of the pathological process of cerebral ischemia and provide a scientific basis for the development of new therapeutic approaches. Given the inherent limitations of transcriptomics, future research endeavors can leverage multi-omics data integration strategies to gain a more comprehensive understanding of ischemic conditions. Specifically, by integrating genomics data, we can identify ischemia-related gene variants through a comparative analysis of the genomes of affected patients and healthy controls while simultaneously elucidating gene interactions by cross-referencing with transcriptomics data sets. Furthermore, the integration of proteomics data offers a valuable opportunity to compare protein expression profiles in ischemic tissues with corresponding transcriptomics data, thereby facilitating the discovery of post-transcriptional regulatory mechanisms and novel signaling pathways. This approach not only enhances our understanding of the molecular underpinnings of ischemia but also provides insights into potential therapeutic targets. In the realm of metabolomics, the integration of small molecule metabolite data with transcriptome data enables the construction of a gene-metabolite network. This network serves as a powerful tool for analyzing the mechanisms underlying cellular energy metabolism disorders and identifying potential intervention targets. By unraveling the intricate interplay between genes and metabolites, we can gain a deeper understanding of the pathological processes involved in ischemia and develop more effective therapeutic strategies. In terms of the utilization of emerging computational tools, machine learning algorithms can be employed to construct diagnostic or prognostic models of cerebral ischemia with the help of SVM, RF and other algorithms with the transcriptome and clinical information as inputs, and screen for key gene markers. Network analysis tools can be used to construct gene regulatory and PPI networks based on transcriptome differential genes, and analyze the topology to lock the key regulatory factors and the bioinformatics databases can be integrated to combined differential genes with KEGG and other metabolites. At the same time, we integrate bioinformatics databases and compare the transcriptome differential genes with KEGG and other databases to clarify the biological processes and signaling pathways involved in the genes. In the mining and application of potential new therapeutic targets, we will first validate the potential targets found in the transcriptome through gene knockout and overexpression in cell or animal models, and evaluate their effects on cerebral ischemic injury. Then, we will design small molecules or biologics based on the validated targets using CADD, and evaluate the efficacy of the drugs with transcriptome data. Then we will sequence the transcriptome of patients with cerebral ischemia, and formulate a targeted therapeutic plan based on the gene expression profiles of the individuals, improving the precision of treatment. Finally, the transcriptome sequencing of patients with cerebral ischemia will be used to formulate targeted therapeutic programs based on individual gene expression profiles and improve the accuracy of treatment, providing a more comprehensive view of the pathophysiological mechanisms of cerebral ischemia and new therapeutic methods for clinical practice. In conclusion, the integration of transcriptomics, single-cell sequencing and spatial transcriptomics into the study of cerebral ischemia has led to a deeper understanding of the pathological process of cerebral ischemia, opened up new avenues for understanding the complexity of the disease and the development of innovative therapeutic approaches, and provided a scientific basis for the development of new treatments.