背景与目的：脑中风（cerebral stroke），又称脑卒中，是一种急性脑血管疾病，主要分为缺血性脑中风和出血性脑中风两大类^[1]。其中，脑出血（Intracerebral Hemorrhage, ICH）是出血性脑中风的主要类型之一，占所有急性中风的20%-30%, 但其急性期病死率高达30%-40%，造成的残疾调整生命年（disability-adjusted life years, DALYs）损失远大于缺血性中风^[2–4]，且预后较差，发病 1 年后的死亡率大于 50%，幸存者中多数遗留不同程度的神经功能障碍^[5–7]，是中国成人致死致残率最高的疾病之一，。近年来，中国ICH的发病年龄呈现年轻化趋势，30-40岁人群的发病率显著上升，可能与高血压控制不足、生活方式改变及工作压力增加有关，尤其是在农村地区，ICH发病率显著高于城市^[8]。有研究指出，约50%-70%的ICH病例与高血压相关^[9]，但是单纯控制血压只能降低30%-40%脑出血风险，提示存在其它重要的危险因素。白细胞（包括中性粒细胞、单核细胞/巨噬细胞、淋巴细胞等）在血管疾病的发生、发展和修复中具有双重作用，既可促进炎症损伤，也参与保护性调节，但在高血压脑出血血管病变中的作用尚未可知。单细胞转录组测序技术（single-cell RNA sequencing, scRNA-seq）通过解析单个细胞的基因表达谱，揭示细胞异质性和功能状态，为研究复杂生物系统提供了前所未有的精度。本研究通过单细胞测序技术对人与小鼠外周血白细胞进行检测，筛选高血压脑出血血管病变过程可能发挥重要作用的关键细胞群体及差异表达基因，旨在探究高血压脑出血血管病变潜在的发病机制及可能靶点。

方法：将 8 月龄的野生型 C57BL/6 雄性小鼠（28-34g）通过皮下埋入血管紧张素 II（1000 ng/kg/min）的微量渗透泵，并同时于饮水中给予给与 L-NAME（100 mg/kg/天）的方式构建自发性的高血压脑出血模型^[10]。在模型构建后每日三次观察小鼠的行为学特征来评价脑出血的发病情况。在此过程中分别收取对照组、高血压组、脑出血组的外周血样本，此外，收取人外周血对照组、高血压组、脑出血组外周血样本，以上样本在裂解红细胞后对全部白细胞进行单细胞转录组测序，并进行生物信息学分析，注释细胞群并找出病程中发生变化的细胞群体。通过聚类分析细胞的动态变化筛选与脑出血相关的细胞。为进一步探究这群细胞的功能，我们进行了对照组-疾病组间的差异表达基因（Differential expressed genes, DEGs）分析，并针对这些差异基因开展了GO和KEGG 富集，通过网络药理学利用网络药理学的方法寻找能够逆转目的细胞差异基因表达的小分子化合物,此外还进行scenic分析寻找潜在的转录因子，并通过cellchat探究T细胞各个亚群与中性粒细胞间的关系。

结果：对我们20个人外周血白细胞样本及11个小鼠外周血白细胞样本进行单细胞转录组测序，质量过滤后分别捕获了123,202及128,212个白细胞，人外周血白细胞被注释为中性粒细胞、嗜碱性粒细胞、髓细胞、T 细胞、B 细胞、自然杀伤细胞、血小板66大类型；小鼠外周血白细胞被注释为中性粒细胞、嗜碱性粒细胞、T 细胞、B 细胞、自然杀伤细胞、红细胞、单核细胞、血小板86种细胞类型。有部分细胞在疾病进程中存在数量比例上的差异，与正常对照组相比，人与小鼠外周血T细胞在ICH过程中细胞百分比均发生了不同比例的降低。由于T细胞存在较大的异质性，我们对其进行了进一步的分群处理以便分析其功能，在人与小鼠中的T细胞群中我们分别获得了我们获得了8个及9个异质性亚群。其中人与小鼠CD4+Tn、CD4+Treg亚群在脑出血的过程中细胞比例一致降低，CD8+TEMRA占比一致升高。在功能分析过程中我们发现CD4+Tn及Treg细胞在高血压脑出血血管病变过程中可能通过调节细胞杀伤、促进淋巴细胞失能、T细胞失能、抗NK细胞介导的细胞毒性等方式发挥保护作用，这群细胞在脑出血过程中占比降低了，而CD8+TEMRA细胞恰好在脑出血的过程中升高且可能发挥细胞毒性作用的功能，在接下来的细胞通讯分析中我们进一步发现，CD4+Tn细胞能够通过MHC I类分子等信号干预涉及CD8+TEMRA细胞在内的细胞毒作用调节免疫稳态，二者相互关联共同促使了脑出血的发生。随后，此外，我们还使用网络药理学的方式在cmap数据库中查找药物转录组与疾病状态中CD4+Tn细胞的差异基因特征相反的药物，有望能够改善疾病表型，并通过转录因子分析锁定了CD4+Tn细胞群中的关键转录因子BACH2、E2F6及下游靶基因，为后续的疾病干预筛选提供了关键位点，并通过网络药理学分析在cmap数据库中查找药物转录组与疾病状态中CD4+Tn细胞的差异基因特征相反的药物，筛选出的小分子化合物Naringin、Terazosin经动物实验验证具有脑出血预防作用。

结论：在脑出血过程中，人与小鼠外周血中具有保护功能的CD4+Tn细胞和CD4+Treg细胞比例减少，而具有细胞毒性作用的CD8+TEMRA比例增多，二者的动态变化可能协同促进高血压脑出血的发生。小分子化合物柚皮苷Naringin和Terazosin可通过调控CD4+Tn发挥对高血压脑出血的保护作用。在脑出血过程中，人与小鼠外周血白细胞中行使保护作用的CD4+Tn、CD4+Treg减少的同时参与细胞毒性作用的CD8+Temra细胞增多，二者之间可能存在此消彼长的关联共同促进了高血压脑出血的发生。

Background and Objectives: Cerebral stroke, a major acute cerebrovascular disease, is broadly classified into ischemic and hemorrhagic subtypes. Intracerebral hemorrhage (ICH), the predominant form of hemorrhagic stroke, accounts for 20%-30% of all acute strokes, yet exhibits disproportionately high morbidity and mortality. The acute-phase fatality rate reaches 30%-40%, with a 1-year mortality exceeding 50%. Survivors often suffer severe neurological deficits, making ICH a leading cause of disability-adjusted life years (DALYs) in Chinese adults. Notably, ICH incidence is rising among younger populations (30-40 years), linked to poorly controlled hypertension, lifestyle shifts, and occupational stress, particularly in rural areas. Although hypertension underlies 50%-70% of ICH cases, blood pressure management alone reduces risk by only 30%-40%, implicating additional pathogenic factors. Leukocytes (e.g., neutrophils, monocytes/macrophages, lymphocytes) play dual roles in vascular injury and repair, but their dynamics in hypertensive ICH remain unclear. Single-cell RNA sequencing (scRNA-seq) enables high-resolution dissection of cellular heterogeneity and functional states. Here, we applied scRNA-seq to human and murine peripheral blood leukocytes to identify critical cell populations and differentially expressed genes (DEGs) driving hypertensive ICH vasculopathy, aiming to elucidate novel mechanisms and therapeutic targets.

 Cerebral stroke is an acute cerebrovascular disease primarily classified into two major types: ischemic stroke and hemorrhagic stroke [1]. Among these, intracerebral hemorrhage (ICH) is one of the main subtypes of hemorrhagic stroke, accounting for 20%-30% of all acute strokes. However, it has an acute-phase mortality rate of 30-40% and contributes to a significantly higher burden of disability-adjusted life years (DALYs) compared to ischemic stroke. Moreover, ICH exhibits a poor prognosis, with a 1-year mortality rate exceeding 50%, and most survivors suffer from varying degrees of neurological dysfunction. In China, ICH is one of the leading causes of death and disability in adults. Notably, recent epidemiological trends indicate a younger age of onset, with a significant increase in incidence among individuals aged 30–40 years, likely attributable to poor hypertension control, lifestyle changes, and increased work stress. This trend is particularly pronounced in rural areas, where the ICH incidence is significantly higher than in urban regions . Studies suggest that 50%–70% of ICH cases are associated with hypertension. However,  blood pressure control alone reduces the risk of ICH by only 30%–40%, indicating the involvement of other critical risk factors. Leukocytes—including neutrophils, monocytes/macrophages, and lymphocytes—play a dual role in vascular diseases, contributing to both inflammatory injury and protective regulation. Yet, their specific role in hypertension-induced cerebrovascular pathology in ICH remains unclear. Single-cell RNA sequencing (scRNA-seq) technology, by analyzing the gene expression profiles of individual cells, reveals cellular heterogeneity and functional states, offering unprecedented precision for studying complex biological systems. In this study, scRNA-seq was employed to examine peripheral blood leukocytes in humans and mice, aiming to identify key cell populations and differentially expressed genes involved in the progression of hypertensive intracerebral hemorrhage. The goal is to explore the potential mechanisms and possible therapeutic targets underlying hypertensive intracerebral hemorrhage.

Methods:Spontaneous hypertensive ICH was induced in 8-month-old male C57BL/6 mice (28-34g) via angiotensin II (1000 ng/kg/min, osmotic pump) and L-NAME (100 mg/kg/day, drinking water). Behavioral monitoring was performed thrice daily to assess ICH onset. Peripheral blood samples were collected from control, hypertensive, and ICH groups in both humans and mice. After erythrocyte lysis, total leukocytes underwent scRNA-seq and bioinformatics analysis for cell annotation and dynamic population tracking. Cluster analysis identified ICH-associated subsets. DEG analysis, GO/KEGG enrichment, network pharmacology (CMAP database screening), SCENIC (transcription factor inference), and CellChat (intercellular communication) were employed to characterize CD4+Tn cell functions and therapeutic candidates.

 Eight-month-old wild-type C57BL/6 male mice (28-34 g) were used to construct a spontaneous hypertensive intracerebral hemorrhage (ICH) model. This was achieved by subcutaneously implanting an osmotic pump delivering angiotensin II (1000 ng/kg/min) and simultaneously administering L-NAME (100 mg/kg/day) in their drinking water. After model establishment, the behavioral characteristics of the mice were observed three times daily to assess the onset of ICH. During this process, peripheral blood samples were collected from the control group, the hypertension group, and the ICH group. After lysing red blood cells, single-cell RNA sequencing (scRNA-seq) was performed on all leukocytes, followed by bioinformatics analysis to annotate cell populations and identify specific cell subsets that changed during the disease progression. Clustering analysis was conducted to track dynamic changes in cells and identify those associated with ICH. To further investigate the functions of these cell populations, differential expressed gene (DEG) analysis was performed between the control and disease groups. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were conducted on these DEGs. Network pharmacology was employed to identify small-molecule compounds capable of reversing the differential gene expression in target cells. Additionally, SCENIC analysis was performed to identify potential transcription factors, and CellChat was used to explore the interactions between T cell subsets and neutrophils.

Results: scRNA-seq of 20 human and 11 murine samples yielded 123,202 and 128,212 high-quality leukocytes, respectively. Human cells were annotated as neutrophils, basophils, myeloid cells, T cells, B cells, and NK cells; murine cells included monocytes. T cell proportions declined consistently during ICH. Subclustering revealed 8 (human) and 9 (murine) T cell subsets, with conserved reductions in CD4+Tn and CD4+Tregs and expansion of CD8+TEMRA. Functional analyses indicated that CD4+Tn/Tregs exert vasculoprotection via cytotoxicity suppression, lymphocyte anergy, and NK cell inhibition, while their loss coincided with CD8+TEMRA-mediated cytotoxicity. CellChat implicated MHC-I signaling in CD4+Tn-mediated immune homeostasis. Key transcription factors (BACH2, E2F6) and downstream targets were identified. Network pharmacology prioritized naringin and terazosin as CD4+Tn-modulating compounds, validated in vivo for ICH prevention.

Single-cell transcriptome sequencing was performed on 20 human peripheral blood leukocyte samples and 11 murine peripheral blood samples. After quality filtering, a total of 123,202 and 128,212 leukocytes were captured, respectively. Human peripheral blood leukocytes were annotated into six major cell types, including neutrophils, basophils, myeloid cells, T cells, B cells, natural killer (NK) cells, and platelets. Murine peripheral blood leukocytes were annotated into eight cell types, including neutrophils, basophils, T cells, B cells, NK cells, erythrocytes, monocytes, and platelets. Significant differences in the proportional abundance of certain cell types were observed during disease progression. Compared to the normal control group, the percentage of T cells in both human and murine peripheral blood decreased to varying degrees during intracerebral hemorrhage (ICH). Given the high heterogeneity of T cells, we performed further subclustering to analyze their functional characteristics. In human and murine T cell populations, we identified 8 and 9 heterogeneous subpopulations, respectively. Among these, the proportions of CD4+ naïve T cells (CD4+Tn) and regulatory T cells (CD4+Treg) were consistently reduced during ICH in both humans and mice, while the proportion of CD8+ terminally differentiated effector memory T cells (CD8+TEMRA) was consistently increased. Functional studies revealed that CD4+Tn and Treg cells may exert protective effects during ICH by modulating cell cytotoxicity, promoting lymphocyte dysfunction, inducing T cell exhaustion, and counteracting NK cell-mediated cytotoxicity. However, the proportions of these cells were significantly reduced during ICH. In contrast, the proportion of CD8+TEMRA cells was significantly increased during ICH and may contribute to cytotoxic effects. Further cell-cell communication analysis demonstrated that CD4+Tn cells could regulate immune homeostasis by modulating the cytotoxic activity of CD8+TEMRA cells through MHC class I molecules and other signaling pathways. The interplay between these two cell types collectively contributed to the pathogenesis of ICH. Additionally, using a network pharmacology approach, we screened the CMAP database for drugs with transcriptomic profiles opposite to the differential gene signatures of CD4+Tn cells in the disease state, which may potentially ameliorate the disease phenotype. Through transcription factor analysis, we identified key transcription factors, BACH2 and E2F6, and their downstream target genes within the CD4+Tn cell population, providing critical targets for subsequent disease intervention.

Conclusion: Hypertensive ICH involves a pathogenic imbalance between depleted protective CD4+Tn/Tregs and expanded cytotoxic CD8+TEMRA. The small-molecule compounds naringin and terazosin confer protection by restoring CD4+Tn homeostasis.