A single-cell transcriptomic atlas characterizes liver non-parenchymal cells in healthy and diseased mice

Zheng Wang, Jingyang Qian, Xiaoyan Lu, Ping Zhang, Rongfang Guo, He Lou, Shuying Zhang, Jihong Yang, Xiaohui FanbioRxiv 2021 · DOI

TL;DR

This study presents a comprehensive single-cell transcriptomic atlas of liver non-parenchymal cells (NPCs) across healthy and five distinct liver disease models in mice. This resource reveals intricate cellular heterogeneity and disease-specific reprogramming, offering a deeper understanding of liver disease mechanisms and potential therapeutic targets.

What Problem Does This Paper Address?

The liver is a complex organ with diverse cell types, particularly its non-parenchymal cells (NPCs), which are crucial for its structure and function. However, our understanding of NPC heterogeneity, especially in the context of various liver diseases, has been limited. This lack of detailed knowledge hinders the development of precise therapies for liver conditions. The authors aim to address this by creating a high-resolution map of liver NPCs in both healthy and diseased states, identifying how these cells and their interactions change across different liver pathologies.

Methods

The researchers employed single-cell RNA sequencing (scRNA-seq) on a large scale, profiling over 171,000 NPCs from healthy mice and five different mouse models of liver disease: alcoholic liver disease (ALD), nonalcoholic steatohepatitis (NASH), drug-induced liver injury (DILI), cholestatic liver injury (BDL), and ischemia-reperfusion (IR) injury. They utilized standard bioinformatics pipelines for data processing, including quality filtering, normalization, and dimensionality reduction (t-SNE). Key analytical approaches included:

  • Clustering and Cell Typing: Unsupervised clustering to identify distinct cell populations and subtypes, followed by annotation using known marker genes and the SingleR package.
  • Transcriptional Regulatory Network Analysis: Using the SCENIC package to infer transcription factor (TF)-target gene networks and analyze TF activity across different cell types and disease models.
  • Metabolic Analysis: Quantifying metabolic pathway activity using a previously established scoring method based on KEGG pathways.
  • Cell-Cell Interaction (CCI) Analysis: Employing CellPhoneDB to predict ligand-receptor interactions and map communication networks between different cell types.
  • Pseudotime Analysis: Using Monocle2 and RNA velocity to infer developmental trajectories and dynamic changes within specific cell populations, particularly neutrophils and Kupffer cells (KCs).

Datasets

The study utilized scRNA-seq data generated from:

  • Healthy control mice.
  • Five mouse models of liver disease:
  • Alcoholic liver disease (ALD)
    • Nonalcoholic steatohepatitis (NASH) induced by a high-fat, methionine-choline deficient diet (HF-MCD)
    • Drug-induced liver injury (DILI) induced by acetaminophen (APAP)
    • Cholestatic liver injury induced by bile duct ligation (BDL)
    • Liver ischemia-reperfusion (IR) injury

The raw single-cell RNA sequencing data has been deposited into the Gene Expression Omnibus (GEO) database (GEO: GSE166178).

Key Insights and Ideas

  • Unprecedented NPC Heterogeneity: The study reveals a remarkable level of diversity within liver NPCs, identifying numerous subtypes for endothelial cells, hepatic stellate cells (HSCs), neutrophils, T cells, and mononuclear phagocytes (MPs). This fine-grained resolution is crucial for understanding cell-specific roles in disease.
  • Disease-Specific Cell Signatures: Different liver diseases are characterized by distinct alterations in NPC populations and their transcriptional profiles, suggesting that each disease has a unique cellular fingerprint.
  • Protective Neutrophil Subtype: A specific neutrophil subtype, characterized by high Chil3 expression, was identified and found to be significantly increased in drug-induced and cholestatic liver injury models, suggesting a protective role in these conditions.
  • Metabolic Reprogramming in Fibrosis: Fibrosis development is strongly associated with metabolic shifts, particularly increased glycolysis and retinol metabolism in activated HSCs and MPs.
  • Differential Cell-Cell Communication in Fibrosis: The study highlights how cell-cell interactions differ in driving fibrosis. Cholangiocytes and immune cells play a more prominent role in cholestatic liver fibrosis, while HSCs are more critical for NASH fibrosis.

Key Results

  • A total of 171,814 liver NPCs were profiled across control and five disease models.
  • 12 major NPC types and numerous subtypes were identified and characterized.
  • Transcriptional regulatory network analysis revealed disease-specific reprogramming of TFs.
  • Metabolic analysis showed increased glycolysis and retinol metabolism in activated HSCs and MPs in fibrotic conditions.
  • Cell-cell interaction analysis identified unique ligand-receptor pairs specific to different disease models, with distinct contributions of cholangiocytes and HSCs to fibrosis.
  • The study developed an interactive website ("Murine Liver NPCs Atlas") to provide universal access to the data.

Limitations and Future Work

  • Mouse Model Limitations: While the study uses established mouse models, the authors acknowledge that some models, like the MCD diet for NASH, may not fully recapitulate all aspects of human disease (e.g., obesity, insulin resistance).
  • Single-Cell Resolution vs. Spatiotemporal Heterogeneity: The study provides a single-cell landscape but notes that future work should consider the spatiotemporal heterogeneity of intrahepatic cells, including hepatocytes, to gain a more complete picture.
  • Human Data Comparison: While a comparison with human bulk RNA-seq data was performed, the authors suggest that further analysis with human single-cell data would be valuable.
  • Further Exploration of Subtypes: The authors suggest that the detailed characterization of cell subtypes and their specific roles in disease progression offers significant potential for developing targeted therapies. The interactive atlas is intended to facilitate this future research.

Summary updated

Abstract

ABSTRACT The heterogeneity of liver non-parenchymal cells (NPCs) is essential for liver structure and function. However, the current understanding of liver NPCs, especially in different liver diseases, remains incompletely elucidated. Here, a single-cell transcriptome atlas of 171,814 NPCs from healthy and 5 typical liver disease mouse models, including alcoholic liver disease, nonalcoholic steatohepatitis (NASH), drug-induced liver injury, cholestatic, and ischemia-reperfusion liver injury is constructed. The inter- and intra-group heterogeneity of 12 types (and numerous subtypes) of NPCs involving endothelial cells, hepatic stellate cells (HSCs), neutrophils, T cells, and mononuclear phagocytes (MPs) are summarized. A protective subtype of neutrophils characterized by Chil3 high is validated and found significantly increasing only in drug-induced and cholestatic liver injury models. Transcriptional regulatory network analysis reveals disease-specific transcriptional reprogramming. Metabolic activity analysis indicates that fibrosis is accompanied by increases in glycolysis and retinol metabolism in activated HSCs and MPs. Moreover, we found that cell-cell interactions between cholangiocytes and immune cells contribute more to cholestatic liver fibrosis compared with NASH, while HSCs are more important for NASH fibrosis. Our atlas, together with an interactive website provides a systematic view of highly heterogeneous NPCs and a valuable resource to better understand pathological mechanisms underlying liver diseases.
BibTeX
@misc{wang2021single,
  author = {Zheng Wang and Jingyang Qian and Xiaoyan Lu and Ping Zhang and Rongfang Guo and He Lou and Shuying Zhang and Jihong Yang and Xiaohui Fan},
  title = {{A single-cell transcriptomic atlas characterizes liver non-parenchymal cells in healthy and diseased mice}},
  year = {2021},
  doi = {10.1101/2021.07.06.451396},
  howpublished = {bioRxiv},
}
Figure