Meta-Analysis of Smooth Muscle Lineage Transcriptomes in Atherosclerosis and Their Relationships to In Vitro Models

Vascular smooth muscle cells (VSMC) exhibit phenotypic plasticity in atherosclerotic plaques, and among other approaches, has been modeled in vitro by cholesterol loading.

Identification of the sources of variation across single cell datasets in atherosclerosis will be an important step towards understanding VSMC fate transitions in vivo. Also, we conclude that cholesterol-loading in vitro is insufficient to model the VSMC cell state transitions observed in vivo, which underscores the need to develop better cell models. Mouse models, however, appear to reproduce a number of the features of VSMCs in human plaques.

Meta-analysis of scRNA-seq data from VSMC lineage traced cells across five experiments of murine atherosclerosis was performed. In vivo expression profiles were compared to three in vitro datasets of VSMCs loaded with cholesterol and three datasets of polarized macrophages.

We identified 24 cell clusters in the meta-analysis of single cells from mouse atherosclerotic lesions with notable heterogeneity across studies, especially for macrophage populations. Trajectory analysis of VSMC lineage positive cells revealed several possible paths of state transitions with one traversing from contractile VSMC to macrophages by way of a proliferative cell cluster. Transcriptome comparisons between in vivo and in vitro states underscored that data from three in vitro cholesterol-treated VSMC experiments did not mirror cell state transitions observed in vivo. However, all in vitro macrophage profiles analyzed (M1, M2, and oxLDL) were more similar to in vivo profiles of macrophages than in vitro VSMCs were to in vivo profiles of VSMCs. oxLDL loaded macrophages showed the most similarity to in vivo states. In contrast to the in vitro data, comparison between mouse and human in vivo data showed many similarities. Conclusions: Identification of the sources of variation across single cell datasets in atherosclerosis will be an important step towards understanding VSMC fate transitions in vivo. Also, we conclude that cholesterol-loading in vitro is insufficient to model the VSMC cell state transitions observed in vivo, which underscores the need to develop better cell models. Mouse models, however, appear to reproduce a number of the features of VSMCs in human plaques.