1. Identification of 183 DEGs and 19 DEMs.

Intracranial hemorrhage (ICH) is a prevalent cerebrovascular disease and a critical medical emergency (Zhang et al., 2024). Approximately 75%–80% of subarachnoid hemorrhage (SAH) cases result from ruptured intracranial aneurysms, predominantly affecting individuals aged 40 to 60. Cerebral vasospasm (CVS), one of the most severe complications, occurs in up to 70% of cases, typically manifesting within days to 2 weeks after SAH (Treggiari-Venzi et al., 2001). CVS-induced cerebral ischemia, delayed ischemic neurological deficits (DIND), cerebral infarction, and fatalities significantly contribute to the high mortality and disability rates associated with this disease (Zarrin et al., 2024). Therefore, promptly identifying, preventing, and managing CVS in individuals with cerebral hemorrhage is crucial to minimize brain tissue damage and enhance patient outcomes.

CVS generally occurs in the brainstem or major arteries, characterized by localized, segmental, or diffuse spastic manifestations due to ultrastructural and functional changes in endothelial cells and smooth muscle cells. The resulting vasoconstriction and impaired vasodilation lead to delayed stenosis and narrowing of the vascular lumen, reducing cerebral perfusion pressure and blood flow and ultimately causing cerebral infarction (Harrod et al., 2005; Macdonald, 2014). The precise etiologies of CVS are complex and multifactorial. Factors such as red blood cells, metabolic byproducts, and bloody cerebrospinal fluid within blood clots can act as mechanical stimulants on intracranial vessels, enhancing vasoconstrictive function and leading to CVS. Oxyhemoglobin is critical in triggering CVS by generating oxygen free radicals through lipid peroxidation, resulting in serious tissue cell membrane impairment (Dietrich and Dacey, 2000; Ishiguro et al., 2006; Yang et al., 2019). This oxidative process increases endothelin (ET) production and disrupts vascular relaxation by interacting with nitric oxide (Zhang et al., 2001). Additionally, the inflammatory response initiated by the rupture of an intracranial aneurysm results in blood accumulation in the subarachnoid space and brain ventricles. This accumulation leads to endothelial cell apoptosis and decreased endothelial cell function, ultimately contributing to CVS development. Research by Aihara and Kasuya has shown that expression levels of IL-6, IL-1α, and Intercellular Adhesion Molecule-1 (ICAM-1) following intracerebral hemorrhage correlate with the onset of CVS (Aihara et al., 2001).

Brain vascular endothelial and smooth muscle cells play crucial roles in the development and progression of CVS. Following cerebral hemorrhage, arterial remodeling, inflammation, altered extracellular matrix, and cell damage result in prolonged vasoconstriction and impaired vasodilation, leading to vascular spasm. The lysis of red blood cells and arachidonic acid and their metabolic byproducts can damage endothelial cells, release bioactive substances such as endothelin and serotonin, inhibit prostacyclin synthase activity, and promote platelet adhesion and aggregation. This process increases thromboxane A2 release, causing an imbalance between vasoconstriction and vasodilation, promoting cerebrovascular spasm (Cahill et al., 2006; Viderman et al., 2023). Activation of Ca2+ channels on smooth muscle cell membranes further contributes to vasoconstriction. Oxyhemoglobin induces endothelin-1 secretion by smooth muscle cells via the NF-κB and Rho kinase/PKC pathways, leading to vasoconstriction (Peeyush Kumar et al., 2019).

Additionally, oxyhemoglobin can increase and activate voltage-gated calcium channels in cerebral arterial smooth muscle cells while inhibiting voltage-gated potassium channels, leading to a significant influx of Ca2+ (Ishiguro et al., 2006). The phosphorylation of calcium-dependent myosin light chain kinase causes contraction of smooth muscle cells, representing the classical theory of vasoconstriction. Various factors such as inflammation, endothelial cell apoptosis, oxidative stress response, NO signaling pathway, and cell membrane dysfunction contribute to the development of CVS, complicating the identification of relevant biomarkers for accurate prediction and diagnosis, ultimately impacting patient outcomes. Thus, extensive identification of underlying biomolecules and key pathways related to CVS pathophysiological mechanisms is urgently needed.

Microarray and high-throughput sequencing, combined with integrated bioinformatics analysis, have significantly advanced the identification of novel genes and pathways involved in disease pathology. Studies have examined changes in gene expression and miRNA levels in fluid systems such as blood and cerebrospinal fluid as potential biomarkers for CVS following intracerebral hemorrhage. Antoine Baumann et al. demonstrated that NRG1 and HTRA1 impact the proliferation and migration of vascular smooth muscle cells, contributing to CVS development (Baumann et al., 2013). Stanley S. discovered that miR-27a-3p, miR-516a-5p, miR-566, and miR-1197 are significantly upregulated in the vasospasm group through cerebrospinal fluid (CSF) miRNA sequencing, suggesting these miRNAs could serve as CSF biomarkers predisposing individuals to CVS (Stylli et al., 2017). Additionally, elevated expression levels of miR-15a (Kikkawa et al., 2017), miR-24 (Li et al., 2018), and miR-195-5p (Tsai et al., 2021) in SAH patients are closely associated with CVS onset. miR-199a, miR-497, and miR-365 target TGF-β and MAPK signaling, influencing inflammation, matrix degradation, vascular smooth muscle cell apoptosis, and vessel wall rupture (Supriya et al., 2022). Despite studies on genes with DEGs and DEMs, significant individual variability among patients has hindered clinical validation, resulting in discrepancies in using these mRNAs and miRNAs for CVS diagnosis. Furthermore, few studies have conducted integrative analyses of miRNA-mRNA regulatory networks to elucidate their biological functions, necessitating further investigation.

In the present study, mRNA expression profiles (GSE37924) and miRNA expression datasets (GSE165698) from the GEO database were selected to identify DEGs between CVS and non-CVS groups. Potential pathological mechanisms of CVS were explored through functional enrichment analysis, including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG), and a protein-protein interaction (PPI) network was constructed. Furthermore, a miRNA-mRNA regulatory network was developed to investigate the roles of key miRNAs and their target genes in cultured cells and brain tissues affected by SAH. This approach facilitates a comprehensive understanding of molecular mechanisms and aids in discovering potential diagnostic biomarkers or therapeutic targets for CVS.

Cerebrovascular spasm, a critical complication, can result in delayed cerebral ischemia, delayed ischemic neurological deficits, and even cerebral infarction and fatality. Early diagnosis and effective treatment of CVS significantly reduce mortality in hemorrhagic patients, potentially saving lives. Previous studies have examined expression profiles involved in vasospasm induced by ICH to identify DEGs and DEMs. However, the results varied significantly due to differences in sequencing platforms and sample variations. Thus, a more reliable strategy to identify and screen biomarkers and targets for early detection of CVS is needed. To address this, we selected DEMs associated with CVS from the GEO database and constructed a miRNA-mRNA network. Additionally, core miRNAs (LET-7A-5P, MIR-24-3P, MIR-29A-3P, and MIR-132-3P) and their target mRNAs (CDK6 and SLC2A1) were screened from the network and validated in patients and in vitro. These findings suggest that these miRNAs and their target mRNAs may be involved in the regulatory network of CVS. Identifying and analyzing these core molecules can unveil the potential pathogenesis of CVS and improve the accuracy of diagnosis and treatment.

Various factors contribute to CVS, with cerebrovascular smooth muscle dysfunction being a primary cause. These factors include red blood cells and their breakdown products, impaired endothelial cell function, endothelin-1 (ET-1), nitric oxide (NO), and inflammatory mediators. The diverse inducing factors and mechanisms of CVS ultimately lead to brain tissue death. Histological staining revealed higher HE intensity in hematoma areas than in NAH, where neurons showed swelling. TUNEL and caspase-3 staining in hematoma areas indicated significant neuronal injury, apoptosis, and degeneration during intracerebral hemorrhage stages. Advances in proteomic technologies have identified potential biomarkers in brain injury and CVS, including S100β, GFAP, neurofilament proteins, APP, and tau (Tobieson et al., 2021; Batista et al., 2023). Our results align with prior research indicating elevated levels of APP, S100β, and NFL proteins in regions affected by hematomas.

Through mRNA dataset analysis, we identified 183 differentially expressed genes (114 upregulated and 69 downregulated) between the vasospasm and non-vasospasm groups (GSE37924). Subsequently, we performed enrichment analyses of GO categories and KEGG pathways to comprehensively elucidate the functions and underlying mechanisms of these genes.

The GO and KEGG pathway analyses revealed that the differentially expressed genes in the vasospasm group were enriched in the pronucleus and ATPase complex categories within the cellular component (CC) category. Additionally, they were enriched in the folic acid-containing compound biosynthetic or metabolic process and pteridine-containing compound metabolic process categories within the BP category, compared to the non-vasospasm groups. Pathway enrichment analysis suggested that the mechanisms underlying cerebral vascular vasospasm might involve regulating various pathways, including the one-carbon pool by folate, folate biosynthesis, RNA degradation, oxidative phosphorylation, and PI3K-AKT-mTOR signaling pathways. The PI3K-AKT-mTOR and oxidative phosphorylation pathways were particularly significant. The PI3K-AKT signaling pathway, a classical intracellular signaling pathway, regulates various cellular functions, such as metabolism, proliferation, cell survival, growth, and angiogenesis. This pathway is also implicated in CVS (Jin et al., 2022). Administration of Simvastatin has been demonstrated to effectively modulate the expression of PI3K-AKT-eNOS in endothelial cells, reducing the incidence of CVS following brain hemorrhage (Sugawara et al., 2008). Evidence suggests that thiazolidinediones or the calcimimetic R-568 can inhibit the expression of PI3K/Akt in vascular smooth muscle cells, suppressing vasospasm occurrence (Sinagra et al., 2013; Gulec et al., 2021). The oxidative phosphorylation signaling pathway is also involved in CVS. SAH-induced vasospasm and cerebral edema can lead to cerebral ischemia, which stimulates the production of substantial reactive oxygen species (ROS), such as superoxide ions (O₂-) and hydrogen peroxide (H₂O₂). Excessive ROS production can damage neurovascular units, cause neurological injury, induce endothelin production, and impair vasodilation, leading to worsened vascular spasms (Osuka et al., 2010; Xu et al., 2022; Zhang et al., 2023). Melatonin, an antioxidant, can scavenge free radicals stimulate the expression of antioxidant enzymes (such as superoxide dismutase, glutathione peroxidase, catalase, glutathione reductase, glutathione-S-transferase, heme oxygenase, and nitric oxide synthase), promote glutathione synthesis, reduce electron leakage in the mitochondrial electron transport chain, and suppress CVS, alleviating secondary brain injury (Xu et al., 2022). Another antioxidant, edaravone, has been reported to effectively improve neurofunctional deficits, protect vascular endothelium, and alleviate DCVS in patients with SAH (Munakata et al., 2009). Excessive RNA breakdown may increase intracellular ROS production, causing damage to the neurovascular unit. These research findings are consistent with our RNA-seq analysis results.

In this research, we constructed a WGCNA network to identify modules associated with CVS. Functional annotation and pathway analysis revealed that genes within these modules are involved in various biological processes, including neuroactive ligand-receptor interaction, pantothenate and Coenzyme A (CoA) biosynthesis, glycolysis, starch and sucrose metabolism, and epithelial-mesenchymal transition, among others. Within vasospasm, neurotransmitters and hormones such as vasoconstrictors like endothelin and catecholamines induce vascular smooth muscle contraction by binding to their respective receptors, thereby precipitating vasospasm (Lin et al., 2022). CoA, in addition to its role in cellular metabolism, including fatty acid synthesis and energy metabolism, appears to regulate certain neurotransmission pathways and cell signaling, influencing the contractile state of vascular smooth muscle (Indolfi et al., 2000; Kuzuya et al., 2004). ATP, generated within the glycolytic pathway, acts as a neurotransmitter that regulates the contraction and relaxation dynamics of vascular smooth muscle, thereby influencing vascular tone and hemodynamics (Leach et al., 2002). Additionally, the glycolytic pathway plays a role in other vasospasm-related factors, including lactate production. Excessive lactate levels can induce vessel acidosis, leading to vasoconstriction and spasm. Moreover, the glycolytic pathway affects intracellular redox balance, which regulates the production of vasodilators such as nitric oxide (NO), influencing the occurrence of vasospasm (Asgari et al., 2013; Xu et al., 2023). Elevated blood glucose levels may impair endothelial function and increase intracellular calcium ions in vascular smooth muscle cells, which promotes vasospasm (Matano et al., 2019).

Our study rigorously screened and identified 15 key genes implicated in CVS and vascular smooth muscle contraction. Four of these genes, Acetylcholinesterase (AChE), diacylglycerol kinase zeta (DGKZ), NLRP3 inflammasome, and Phospholipase D2 (PLD2), have been linked to smooth muscle contraction. AChE and DGKZ directly modulate smooth muscle function (Jewulski et al., 2021), while NLRP3 inflammasome contributes to endothelial dysfunction and smooth muscle cell pyroptosis (Bai et al., 2020; Pang et al., 2022), and PLD2 is involved in angiotensin generation (Parmentier et al., 2001). Among the remaining 11 genes associated with CVS, GPD1L and GPD2, both involved in glucose metabolism, stand out, suggesting a potential link between glucose metabolism disorders and CVS. However, genes such as AGL, COL1A1, HLA-DMB, PPP1R3C, SGCD, STAT4, TAGLN, TNNT1, and WAS have been less frequently reported in relation to vasospasm and warrant further investigation. Co-expression and correlation analysis revealed strong interactions among the 15 genes, highlighting their significant roles in the occurrence and progression of CVS.

PPI networks were constructed to comprehensively explore the potential interactions and functions of significant DEGs associated with CVS. The top hub genes identified were ARID1B, EZH2, HDAC2, GART, IPO5 (RANBP5), RPL23, AKAP9, DHFR, RB1, and BRD9 based on the degree method. Most key genes were closely related to vascular smooth muscle cell or vascular endothelial cell injury. ARID1B (AT-rich interactive domain-containing protein 1B) is a component of the SWI/SNF ATP-dependent chromatin remodeling complex. The SWI/SNF complex is essential for vasculogenesis and the proliferation and differentiation of vascular smooth muscle (Julian et al., 2015). Fei Fang et al. indicated that the silencing of Brg1/Brm (components of the SWI/SNF complex) can mitigate endothelial dysfunction under inflammatory conditions, thereby alleviating phenotypes in an animal model of atherosclerosis (Fang et al., 2013). EZH2, a crucial Polycomb Repressive Complex 2 (PRC2) component, catalyzes histone H3 trimethylation at lysine 27 (H3K27me3), leading to gene repression. Elevated EZH2 levels in pulmonary arterial hypertension (PAH) patients can be targeted with GSK126, demonstrating decreased PASMC proliferation and improved mitochondrial function in vitro (Habbout et al., 2021; Dave et al., 2023). EZH2 also influences VSMC proliferation, migration, and vascular remodeling in conditions like atherosclerosis and abdominal aortic aneurysms (Li et al., 2022; Dave et al., 2023; Liu et al., 2023). HDAC2 (Histone Deacetylase 2) is involved in vascular smooth muscle cell (VSMC) phenotypic switching, proliferation, and migration, contributing to cardiovascular diseases (Ryu et al., 2019; Chakraborty et al., 2022). It also maintains endothelial cell barrier function, promotes angiogenesis, and regulates inflammatory responses (Hori et al., 2020). BH4 (Tetrahydrobiopterin), a product of DHFR (Dihydrofolate reductase), regulates NO signaling in endothelial cells (Thoms et al., 2016). RB1 (Retinoblastoma protein 1) is a tumor suppressor gene that inhibits the progression of the cell cycle, limits the proliferation of vascular endothelial cells, and regulates the differentiation of vascular smooth muscle cells. Studies indicated that Rb1 deficiency impairs arterial vasodilation by reducing DHFR activity and NO production in aortic endothelial cells (Cao et al., 2019). These findings strongly suggest a potential link between these key genes and the occurrence of CVS.

miRNA, a stable single-stranded non-coding RNA, regulates gene expression and has been used in diagnosing, monitoring, and treating diseases like cerebral hemorrhage-induced vasospasm (Kikkawa et al., 2017; Wang et al., 2021a; Wang et al., 2021b; Zhang Y. et al., 2022; Huang et al., 2023). In this study, we identified CVS-associated DEMs and their target genes using miRTarBase, and constructed a miRNA-mRNA network based on the negative correlation between miRNA and mRNA expression levels. Through validation with clinical specimens and in vitro cell experiments, we identified several miRNAs (LET-7A-5P, MIR-24-3P, MIR-29A-3P, and MIR-132-3P) and genes (CDK6 and SLC2A1) that may play significant roles in the occurrence and development of CVS. MIR-132-3P, for instance, has been shown to suppress pulmonary artery smooth muscle cell proliferation and migration after pulmonary embolism (Zhang Y. et al., 2022). Its target gene CDK6, a core component regulating the cell cycle, is considered a mutation target in malignant tumors (Meng et al., 2004).

Additionally, CDK6 has been implicated in endothelial and smooth muscle cell proliferation. CDKs were upregulated in the pulmonary vasculature of individuals with pulmonary arterial hypertension (PAH) and in animal models. Remarkably, dinaciclib and palbociclib effectively block cell cycle progression in primary pulmonary smooth muscle and endothelial cells (Weiss et al., 2019). MiR-132-3P regulates the transformation between contractile and synthetic vascular smooth muscle cells (VSMCs) (Zhang J. et al., 2022). As a target gene of miR-132-3P, SLC2A1 (GLUT1) is crucial for maintaining cellular energy supply and is closely associated with cancer prognosis (Liu et al., 2022). In various brain diseases, including malignant brain tumors, cerebral blood vessels show dysregulated expression of GLUT1 and certain junctional proteins. GLUT1 is expressed in the vascular endothelial cells of the BBB, and knocking down GLUT1, specifically in endothelial cells, increases BBB permeability, leading to vasogenic brain edema (Zheng et al., 2010). Studies on Alzheimer’s disease have shown that mice with BBB-specific GLUT1 deficiency exhibit reduced cerebral microvascular degeneration, decreased cerebral blood flow, and accelerated APP deposition (Winkler et al., 2015). Our experimental findings and these studies suggest that abnormal GLUT1 expression in cerebral arterial smooth muscle cells is implicated in the pathogenesis of cerebrovascular diseases. MiR-24-3P and MiR-29A-3P are downregulated in VSMCs and regulate abnormal proliferation induced by hypoxia and atherosclerosis (You et al., 2020; Lee and Kang, 2023; Ji et al., 2024). MiR-24-3p inhibits VSMC proliferation and promotes phenotypic transformation in atherosclerosis. Administration of miR-29a-3p also alleviates pulmonary vascular remodeling, reducing pulmonary artery pressure (Luo et al., 2015).

Therefore, combining these previous studies with our validations, we infer that these four miRNAs and two genes could serve as recovery biomarkers for patients with CVS. However, certain limitations should be noted. The sample size was relatively small, and further validation of the selected miRNA and mRNA candidates is needed to confirm their universality and reliability across larger clinical samples. Additionally, the mechanisms by which miRNA-target gene regulatory networks influence the function and contraction of vascular endothelial and smooth muscle cells require further investigation and validation.

In summary, our study identified 183 DEGs and 19 DEMs associated with CVS following early diagnosis of intracranial hemorrhage. Through miRNA-mRNA network analysis, we elucidated the intricate and interdependent regulatory relationships between these critical miRNAs and their corresponding target genes, contributing to CVS development. Subsequent experimental verification confirmed the upregulation of two genes (CDK6 and SLC2A1) and the downregulation of four miRNAs (LET-7A-5P, MIR-24-3P, MIR-29A-3P, and MIR-132-3P) among the total of 6 DEGs and 8 DEMs. These results suggest a potential correlation between these genes and miRNAs, indicating their possible interplay in the development of CVS. The identified biomarkers are promising tools for the diagnosis of CVS, providing valuable insights for future research into its underlying mechanisms.