The Power of Single-Cell Analysis for the Study of Liver Pathobiology

Abstract

Single cell transcriptomics has emerged as a powerful lens through which to study the molecular diversity of complex tissues such as the liver, during health and disease, both in animal models and in humans. The earliest gene expression methods measured bulk tissue RNA, but the results were often confusing because they derived from the combined transcriptomes of many different cell types in unknown proportions. To better delineate cell-type-specific expression, investigators developed cell isolation, purification, and sorting protocols, yet still, the RNA derived from ensembles of cells obscured recognition of cellular heterogeneity. Profiling transcriptomes at the single-cell level has opened the door to analyses that were not possible in the past. In this review, we discuss the evolution of single cell transcriptomics and how it has been applied for the study of liver physiology and pathobiology to date.

Figure 1 –

a) Flow diagram of different methods by which cells can be harvested for use in various single cell platforms. Fresh liver tissue can be digested and cells sorted using flow cytometry gated against specific antibodies when analyzing known cell subsets, which can then be processed for 1-cell-1-well, droplet, microwell, or combinatorial barcoding methods. Fresh tissue can also be separated into larger groups, such as parenchymal and non-parenchymal cells, using density gradient centrifugation. Freshly procured hepatocytes often do not survive the processing steps to become barcoded, so freezing blocks of liver tissue for nuclei extraction is currently the only method to perform single cell analysis on hepatocytes. Frozen blocks can then be sectioned, allowing tissue to remain intact for spatial methods to be deployed. b) TSNE plot used to visualize how each cell (each dot in the plot) clusters (cell types or states) based on differential gene expression in 2-D space. c) Heatmap of marker genes (y-axis) for each single cell (x-axis) grouped as clusters. Green is high expression and purple is low expression. The marker genes shown are those that best discriminate each cluster from all the other clusters and are interpreted as signatures of cell types and cell states. d) Violin plots with specific marker genes (y-axis) can be used to identify cell types in a cluster (x-axis).

Figure 2 –. Novel insights provided by single cell transcriptomics regarding hepatocytes and hepatic NPCs in nomal liver and during various injuries and pathological conditions.

a) Normal liver biology reveals heterogenous spatially zonated hepatocytes as well as liver endothelial cells. Hepatocytes in the periportal zones that receive more oxygen complete high ATP-demanding tasks such as cholesterol synthesis and urea cycle and use the Ras signaling pathway. Pericentral hepatocytes in a chronic hypoxic environment are used for tasks such as bile synthesis and drug metabolism and use the Wnt pathway. b) Liver fibrosis activates heterogeneous subsets of myofibroblasts (MFB) with only one subpopulation that expresses alpha-smooth muscle actin (α-SMA). S100a6 is proposed as a more universal marker for MFB activation. One subpopulation of MFBs that is termed “scar-associated” is PDGFRa⁺ and is influenced by HSCs and scar-associated macrophages (SAMac) which have a unique gene signature and are pro-fibrogenic. HSCs are also spatially zonated to be either portal vein-associated (PaHSC) or central vein-associated (CaHSC). CaHSCs are the pathogenic collagen producing cells and exclusively express LPAR1, which binds the lipid signaling molecule lysophosphatidic acid. c) In liver endothelial cells (LEC), NASH increased expression of genes involved in lipid metabolism, antigen presentation, and chemokine release, but down regulated those involved in vascular development and homeostasis. NASH expands the monocyte-derived macrophage (MDM) population whereas the Kupffer cell (KC) population is decreased. TREM2⁺ macrophages correlated with increased liver fibrosis and a higher NAFLD score. These MDMs also had downregulation of S100a8 and S100a9. NASH conditions led to increased expression of genes involved in extracellular matrix (ECM) structure and remodeling, but restricted expression of NGfr, which promotes HSC apoptosis. d) Expansion of macrophages were also found in visceral adipose tissue of mice fed a high fat diet. These macrophages were also heterogenous with one subpopulation being Trem2⁺.