Revisiting vascular remodeling in the single cell transcriptome era

Perivascular adipose tissue (PVAT) is a local fat depot surrounding most blood vessels. “Outside-to-inside” signaling from PVAT to the underlying vessel wall actively contributes to vascular homeostasis, as well as to pathogenesis of cardiovascular diseases including hypertension, arterial stiffness, atherosclerosis, neointima formation and aneurysm.^1–3 Crosstalk between PVAT and blood vessels happens partly through paracrine effects of adipokines, such as adiponectin, leptin, IGFBP2 (insulin like growth factor binding protein 2), as well as inflammatory cytokines, which may act on the vascular wall to induce smooth muscle cell (SMC) and endothelial cell proliferation, migration and differentiation. In addition, PVAT produces vessel contracting and relaxing factors to regulate vascular tone and affect blood pressure.^{4, 5} Mouse thoracic PVAT shares most of the defined characteristics of brown adipose tissue, including cellular morphology and expression of thermogenic genes.^6–9 Its distinctive function in thermogenesis, in response to cold environment or β3-adrenergic receptor agonists, is critical to maintaining blood and body temperature.^{1, 6} Additionally, as seen in other adipose depots, PVAT undergoes a “whitening” process under obese or thermoneutral conditions, while circulating hormones from other tissues such as irisin, FGF21(fibroblast growth factor 21) or ANP (atrial natriuretic peptide) etc. might convert “whitening” PVAT back to “brown-like” PVAT.^10–12 Crosstalk between PVAT and blood vessels also happens through trans-differentiation of cells in the vessel wall. We reported that PVAT may share the same precursors with SMCs in blood vessels.⁶ Consistent with our findings, Long et al reported that beige adipocytes expressed a smooth muscle-like gene signature while ectopic expression PRDM16 (PR/SET domain 16) converts bona fide vascular SMCs into brown-like adipocytes in vitro.¹³ However, it remains unknown which groups of cells in the blood vessels will differentiate into adipocytes or other cell types and which are the specific stimuli driving the process in each case. A recent study suggests that PVAT harbors multipotent mesenchymal stem cells (MSCs) capable of differentiating into multiple cell types, including cells from the adipogenic, osteogenic and chondrogenic lineages.¹⁴ Whether the MSCs in PVAT play any role in vascular development and remodeling remained to be addressed, largely due to technical limitations imposed by their small numbers.

The assessment of the single-cell transcriptome through single-cell RNA sequencing (scRNA-seq) provides us the opportunity to reveal and directly visualize the heterogeneity of gene expression at the single cell resolution in certain organs and tissues.^{15, 16} For example, Xiong et al uncovered disruptions of the vascular signaling network during nonalcoholic steatohepatitis using single cell secretome gene analysis;¹⁷ Cochain and Winkels et al revealed the transcriptional landscape of aortic immune cells in atherosclerosis.^{18, 19} Furthermore, regarding the power of this emerging technology is highlighted by the finding one sub-type of disease associated macrophages, TREM2 (triggering receptor expressed on myeloid cells 2) high expressing macrophages in the atherosclerotic plaque, fatty liver, obese adipose tissues and central nervous system with neurodegenerative disease.^19–21 Therefore, monitoring and targeting this particular macrophage sub-type may help better understand the cellular heterogeneity responsible for the development of diseases and aid in development of new treatments.

Pan and colleagues demonstrated last year that there are resident PVAT stromal cells (PVASCs) in mouse thoracic PVAT.²² Through scRNA-seq analysis of primary cultured PVASCs (with >80% cells Sca1⁺/CD90⁺/CD146⁺/CD29⁺), they detected enormous transcriptional heterogeneity in PVASCs. There are 10 to 12 distinct sub-populations, including adipogenic, endothelial, epithelial, neural and cardiomyocyte lineages, in PVASCs both from young and old mice. Gene ontology (GO) analysis revealed that pathways related to vasculature development were enriched in PVASCs, which suggests the PVASCs may play a critical role in vascular remodeling. In addition, they identified and characterized the human PVASCs (hPVASCs) with a similar strategy. Using the mouse carotid artery ligation model, transplanted hPVASCs were found to be involved in vascular remodeling and neointima formation. In this ATVB issue, Gu and colleagues used multiple methods providing unique direct evidences that mesenchymal stem cells (CD29⁺/Sca1⁺/PLIN1⁻/PECAM1⁻) reside in the PVAT, hence called perivascular adipose tissue-derived mesenchymal stem cells (PV-ADSCs), and that they contribute to vascular remodeling through migration and SMC differentiation. This work, for the first time, demonstrated the existence of PV-ADSCs by performing the scRNA-seq assays in enzymatically digested, no-cultured primary PV-ADSCs, which preserves the in vivo characteristics of these cells. After applying stringent quality control filters, two distinct sub-populations of PV-ADSCs were identified in the mouse thoracic PVAT with specific signature gene expression patterns. The cluster 2 PV-ADSCs, similar to adventitial Sca1⁺ cells,²³ displayed enriched myogenesis and TGF-β signaling pathways, which indicates that it favors differentiation to the SMC lineage and is involved in vascular remodeling in the disease condition. In addition, scRNA-seq combined with pseudotime analysis were performed in cultured PV-ADSCs at low passages, which revealed their SMC differentiation trajectory. Using the vein graft model, the authors further indicate that, in vivo, PV-ADSCs transplanted into the adventitia side of the vein graft, significantly promoted neointima formation through migration and differentiation to SMCs. Through metabolic profiling during the differentiation of PV-ADSCs to SMC, this study uncovered that SMC differentiation induced by TGFβ1 and miR-378a-3p is associated with reprogramming of lipid metabolism.

There are a few limitations in this study. The relatively small cell numbers used for the initial scRNA-seq profiling from no-passaged PV-ADSCs may not be enough to fully reveal the actual heterogeneity of PV-ADSCs phenotypes in vivo. Additionally, the evidence for the contribution of PV-ADSCs to vascular remodeling via migration and SMC differentiation in vivo are to a certain degree preliminary. It has been suggested that in vitro expansion of the progenitor cells will at least cause partial loss of its signature characteristics. Therefore, PVAT transplantation and lineage tracing may help strengthen the conclusion.

In summary (Figure), the current study uses the power of the rapidly emerging and evolving scRNA-seq technology combined with metabolomics to uncover the unexpected heterogeneity of PV-ADSCs and propose the novel concept that the PV-ADSCs contribute to vascular remodeling through regulation of their metabolic reprogramming. These results support a model where PVAT is an intrinsic layer of the vascular wall and highlight the need to further explore the adipose-vascular wall crosstalk and the role of PV-ADSCs in vascular remodeling in various metabolic diseases, such as atherosclerosis and diabetes mellitus. Furthermore, scRNA-seq could lead to development of innovative strategies for vascular disease diagnosis, including those with atypical clinical manifestations, strong gender or age association, and rare syndromes and for treatment by uncovering and targeting specific cell sub-populations or metabolic pathways.

Cartoon showing the crosstalk between PVAT and aorta.

1) SM22α⁺ precursors from PVAT and blood vessels contribute to PVAT adipogenesis; 2) PVAT-derived contracting or relaxing factors regulate blood vessel tone and homeostasis; 3) PVAT-derived inflammatory adipokines/cytokines promote macrophage infiltration; 4) PVAT-derived anti-inflammatory adipokines/cytokines maintain blood vessel homeostasis; 5) ADSCs in PVAT contribute to blood vessel remodeling; 6) PVAT “Browning”/”Whitening” characteristics affect PVAT-aorta crosstalk to promote homeostasis or disease, respectively.