血管重塑是多种重大心血管与呼吸疾病中的核心病理改变。根据重塑血管管径大小，可分为主动脉重塑与外周小血管重塑，不同类型的血管重塑过程对于疾病的发生发展具有重要意义。然而目前对于不同疾病中血管重塑过程的细胞分子机制研究程度并不一致，因此全面系统地解析不同类型的血管重塑机制十分必要。

为了探究不同类型的血管重塑的细胞分子机制，我们选择了腹主动脉瘤作为经典的主动脉重塑引起的疾病，同时选择了以肺小血管重塑为主要病理改变的肺动脉高压作为主要的研究对象，利用单细胞转录组测序对主动脉重塑与肺小血管重塑的细胞组分与异质性进行解析。

首先，我们探究了腹主动脉瘤发生发展过程中的细胞异质性。我们构建腹主动脉瘤的小鼠模型，获取腹主动脉瘤单细胞转录组学数据。通过生物信息学分析，我们描绘了腹主动脉瘤的细胞异质性，同时创新性地发现了纤维细胞参与腹主动脉瘤细胞外基质调控的潜在作用。

另一方面，我们利用多种跨物种的肺动脉高压动物模型探究了肺血管重塑中的关键细胞类型与共性机制。通过不同模型中的细胞类型优先级的分析，我们发现巨噬细胞与内皮细胞是肺动脉高压中贡献和改变最为显著的细胞类型。对巨噬细胞亚型分析发现，多种肺动脉高压模型中Dhcr24^high肺泡巨噬细胞亚型占比一致上调，且在肺动脉高压发生过程中存在促炎极化的可能。对内皮细胞亚型的分析揭示了与肺动脉高压内皮功能障碍对应的内皮细胞亚型，包括凋亡的内皮细胞、过度增殖的内皮细胞、Nox2⁺促炎表型的内皮细胞以及具有保护作用的Ednrb⁺内皮细胞。通过进一步的细胞互作分析，我们发现了在肺动脉高压中巨噬细胞与内皮细胞亚型之间显著的细胞通讯信号改变，提出了CXCL12、ANGPTL4以及SEMA3作为新的治疗靶点的可能。

为了更系统大量地计算组织中细胞类型占比，我们进一步结合人类细胞景观数据库建立能够解析计算组织中细胞组分占比情况的新算法tranSig。我们应用贝叶斯模型对跨组织的单细胞数据表达进行融合，从而生成更为准确可靠的细胞类型特征表达矩阵。利用经典贝叶斯去除单细胞数据与批量转录组数据之间的批次效应，结合去卷积化算法CIBERSORTx对细胞类型占比进行估计。通过模拟数据以及真实数据应用，即外周血转录数据，肺泡灌洗液转录数据以及动脉瘤转录数据，对tranSig的准确性进行验证，通过与现有方式的基准测试比较，我们发现tranSig算法对于细胞类型占比的估计有更高的准确性和可靠性。

综上，我们建立了以主动脉重塑为特征的腹主动脉瘤与以肺血管重塑为特征的肺动脉高压疾病的单细胞图谱，从细胞亚型的分辨率描绘血管重塑中的细胞分子机制，发现了纤维细胞减少腹主动脉瘤的发生与死亡，同时揭示了肺泡巨噬细胞亚型在肺动脉高压中潜在的极化机制以及Ednrb⁺内皮细胞的潜在保护作用。研发了能够准确计算血管重塑中细胞组分占比的去卷积化算法，为今后的血管重塑研究提供了充分的数据支持与准确的计算工具。

Vascular remodeling is the key pathological changes in multiple cardiovascular diseases and respiratory diseases. There are two types of vascular remodeling (i.e., aortic and peripheral microvascular remodeling) based on the diameters of vessels whereas remodeling occurred. However, cellular and molecular mechanism of different types of vascular remodeling were poorly understood, thus systemically deciphering the mechanisms was needed.

To investigate the heterogeneities of remodeling mechanism, abdominal aortic aneurysm (AAA) as a classical outward aortic remodeling and pulmonary hypertension as a typical inward microvascular remodeling were involved. Single cell RNA sequencing (scRNA-seq) was performed to demonstrate the cellular heterogeneities.

ScRNA-seq was performed on an angiotensin II-induced mouse model of AAA. Macrophages, B cells, T cells, fibroblasts, smooth muscle cells and endothelial cells were identified through bioinformatic analyses. Intriguingly, we defined CD45⁺COL1⁺ fibrocytes in AAA and extended our findings to the single cell dataset of ATAA patients, of which the existence was further validated by immunostaining in mouse and human AAA tissues. More importantly, the fibrocytes were proposed to attenuate AAA formation through modulating extracellular matrix (ECM) production and thus enhancing aortic stability.

ScRNA-seq was performed on lung tissues from mice exposed to chronic hypoxia or Sugen5416 combined with hypoxia, rats exposed to monocrotaline and control animals. Cell populations perturbed in rats and mice were similar to those found in human disease, with macrophages and endothelial cells being the most affected. A novel DHCR24^high macrophage population harboring both tissue remodeling and pro-inflammatory features were consistently increased across PH models. Phenotypic modulation of DHCR24^high macrophages corresponded with PH progression. Several functionally diverse endothelial subtypes were found, including novel ETB⁺ and NOX2⁺ subpopulations, reflecting enhanced apoptosis, dysregulated angiogenesis and proliferation, and reactive oxygen species mediated stress. These macrophage and endothelial subtypes expressed numerous PH drug target genes, and exhibited several potential intercellular interactions involving the ANGPTL4, CXCL12, and SEMA3 signaling axes.

In addition, computational cell type deconvolution on bulk transcriptomics data can be utilized to systemically demonstrate the cell proportions of remodeling tissues. A novel Bayesian framework, tranSig, was developed to improve signature matrix inference from scRNA-seq by leveraging shared cell type-specific expression patterns across different tissues and studies. Our simulations show that tranSig is robust to the number of signature genes and tissues specified in the model. Applications of tranSig to bulk RNA sequencing data from peripheral blood, bronchoalveolar lavage, and aorta demonstrate its accuracy and power to characterize biological heterogeneity across groups. In summary, tranSig offers an accurate and robust approach to defining gene expression signatures of different cell types, facilitating improved in silico cell type deconvolutions.

In conclusion, we leveraged scRNA-seq data of AAA uncovering a novel cell type fibrocyte involved in AAA pathogenesis. Notably, a comprehensive single-cell atlas of mainstream rodent PH models was established, highlighting several novel macrophage and endothelial subtypes and signaling motifs potentially contributing to human disease. Moreover, an add-on tool for cell type signature matrix was developed to accurately infer the cell proportions in remodeling tissues and enable researchers to comprehensively understand cellular alterations in diseases.