Single-cell technologies to decipher cardiovascular diseases

Abstract

Cardiovascular disease remains the leading cause of death worldwide. A deeper understanding of the multicellular composition and molecular processes may help to identify novel therapeutic strategies. Single-cell technologies such as single-cell or single-nuclei RNA sequencing provide expression profiles of individual cells and allow for dissection of heterogeneity in tissue during health and disease. This review will summarize (i) how these novel technologies have become critical for delineating mechanistic drivers of cardiovascular disease, particularly, in humans and (ii) how they might serve as diagnostic tools for risk stratification or individualized therapy. The review will further discuss technical pitfalls and provide an overview of publicly available human and mouse data sets that can be used as a resource for research.

Keywords: Atherosclerosis; Diagnostic; Heart failure; Hypertrophy; Single-cell sequencing; Single-nuclei sequencing.

Graphical Abstract

Single-cell analysis of tissues studied in cardiovascular disease. Opportunities for in-depth analysis of tissue heterogeneity, transcriptional responses, and cell communication are enabled with scRNA or snRNA sequencing.

Figure 1

Summary of sequencing approaches. (A) Following isolation of cells or nuclei, labelling and/sorting strategies are often employed via flow cytometry, magnetic sorting, or antibodies to enrich a cell-type or deplete unwanted material. Next, droplet sequencing approaches generate gel emulsions to capture single cells in high-throughput manner, whereas cells sorted into plates (e.g. SMART-seq3) are lysed in their component wells. SPLiT-seq uses multiple rounds of barcoding where an iterative process, wherein samples are split and labelled/barcoded, then pooled, and the process is repeated. (B) Spatial transcriptomics applies fixed tissue to a slide with spatially barcoded wells, allowing for relative anatomical location of RNA which is captured following RNA liberation.

Figure 2

Quality control in scRNA-seq experiments. (A) Overview of typical complications arising by cell and tissue type in cell preparation and impacts on quality of data. (B) Representation of how cell isolation may disrupt cell membranes, leach RNA, and how ambient RNA contamination arises from disrupted cells. (C) Depiction of how cytoplasmic RNA loss can modulate the ratio of cytoplasmic/mitochondrial RNA and impact genes per-cell ratios. IM, intact membrane; PM, perforated membrane. (D) Example of droplet sequencing with ambient RNA in fluidic channels and gel emulsions, following cell lysis and barcoding, ambient RNA will be associated to a particular cell. Examples of situations with high ambient RNA, e.g. after nuclei isolation of hearts, wherein cardiomyocytes will cluster in one region, but with cardiomyocyte-specific markers appearing in other non-cardiomyocyte clusters. Ambient RNA removal minimizes such artefacts (right panel). (E) Example of degrees of clustering resolution is shown for incompletely resolved, improved cluster resolution, and over resolved clusters. Example of well-defined clusters of myeloid cells, which can be assigned to classical (red: CD14^high/FCGR3A^low), intermediate (green: CD14^mid/FCGR3A^mid), non-classical monocytes (gold: CD14^low/FCGR3A^high), and dendritic cells (three clusters in dark blue, light blue, pink). Cell-type-specific markers are shown in right panels.

Figure 3

Cellular responses to cardiac injury. Representation of changes in cardiac cells after injury and stress identified by sc/snRNA-seq. (A) Cardiomyocytes express lncRNAs involved in re-expression of foetal genes, whereas fibroblast (B) showed a heterogenic response to injury or aging. (C) Endothelial cells show metabolic adaptations that coincide with transient enhanced mesenchymal gene signatures following myocardial infarction. (D) Immune cell alterations after myocardial infarction. Bone marrow vs. tissue-resident macrophages following cardiac insult, along with neutrophil and T-cell responses.

Figure 4

Atherogenic and atheroprotective responses. Representation of endothelial cells exposed to laminar (pink) vs. disturbed flow (purple) in a physiological setting (left portion of figure). Such responses were assessed in endothelial cells using scATAC-seq and scRNA-seq wherein accessible chromatin and RNA expression found signatures associated with vasoconstriction, EndMT, and EndICLT in disturbed flow endothelial cells. Trem2⁺ macrophages were detected in lesions, showing high expression of Il1b and specialized lipid metabolism. Naive T-cell-induced Tregs were critical for the maintenance of an atheroprotective environment. These iTregs diminished pro-inflammatory M1 monocyte signatures and sensitized M2 macrophages to anti-inflammatory responses. B cells and plasma cells demonstrated elevated clonality of antibodies targeting atherosclerotic lesions. Autoantibodies against ALDH4A1 were shown to be atheroprotective. EndMT, endothelial to mesenchymal transition; EndICLT, endothelial to immune cell-like transition; iTreg, induced (from Naive T cell) T regulatory cell; NRP1, Neuropilin; ALDH4A1, aldehyde dehydrogenase 4 Family Member A1.