Two types of regenerative cell populations appear in acute liver injury

Abstract

The liver has a robust regenerative capacity. However, the mechanisms underlying this process remain unclear. Numerous studies on liver regeneration have been previously conducted using partial hepatectomy models, which may not fully represent acute liver injury with inflammation and necrosis. This is commonly observed in the majority of clinical cases. In this study, we conducted a single-cell RNA sequencing (RNA-seq) analysis of liver regeneration in acetaminophen-treated mice using publicly available data. We found that two distinct populations of regenerative cells simultaneously appeared within the same regenerative process. The two populations significantly differed in terms of cell morphology, differentiation, localization, proliferation rate, and signal response. Moreover, one of the populations was induced by contact with necrotic tissue and demonstrated a higher proliferative capacity with a dedifferentiated feature. These findings provide new insights into liver regeneration and therapeutic strategies for liver failure.

Keywords: acetaminophen; acute liver injury; damage-induced liver injury; dedifferentiation; hepatocytes; liver regeneration.

Graphical abstract

Figure 1

The specific cell population has a high proliferative capacity in hepatic regeneration (A) Uniform manifold approximation and projection, with 10,114 hepatocytes integrated. Colors represent clusters and time after APAP administration. (B) The percentage of cells in their cycle phase at each time point. (C) The percentage of Mki67-positive cells in each cluster. (D) A heatmap showing the average expression levels of metabolism-related genes in the 48 h clusters.

Figure 2

Two proliferating clusters emerge during the regeneration process (A) Violin plot of the feature genes in cluster 7. (B) Heatmap of gene expression related to the KEGG signal pathway that regulates stem cell pluripotency. The genes significantly upregulated in cluster 7 are presented in red. (C) A heatmap showing the average signal score for each cluster. (D) Violin plot of genes commonly upregulated in the signal pathways showing high scores in cluster 7. (E) Ridge plots for Alb and Trf for each cluster.

Figure 3

The cell population with high proliferation and dedifferentiation is cells facing necrosis (A) Experimental design. Mice were sacrificed at time points indicated by red arrows. (B) H&E staining at each time point. Arrows indicate cells with prominent nucleoli of interest. Inserts show representative cells indicated by arrows. (C) IHC for mixed antibodies for TP53INP1 and c-Jun. Upper: wide-field images. Lower: magnified images. (D) Positive rate of TP53INP1 or c-Jun in INTR cells. (E) Immunofluorescence for ANXA2 (TxRed) and DAPI (blue). (F) IHC for ZEB1. In all graphs, n = 5–6 mice/group. Data are shown as mean ± SD. Scale bar: 50 μm. C, central vein; P, portal vein.

Figure 4

INTR and non-INTR cells have different rates of cell-cycle progression (A) Immunofluorescence for Ki67. Arrows indicate representative positive cells; insert shows a magnified cell. Bar graph shows Ki67 positivity. TxRed, Ki67; Blue, DAPI. (B) Immunofluorescence for Cyclin D1. Red arrowheads, representative positive INTR cells; yellow arrowheads, representative positive non-INTR cells. (C) IHC for Cyclin B1. Red arrowheads indicate representative positive cells. (D) Percentage of binuclear cells at each time point. (E) Nuclear size and DAPI intensity distribution at each time point, shown as a 2D density plot using kernel density estimation. (F) Immunofluorescence for BrdU. Bar graph shows BrdU positivity. TxRed, BrdU; Blue, DAPI. (G) IHC for thymidylate synthase. Red arrowheads indicate representative positive cells from 0 to 24 h; insert shows a magnified cell. In all graphs, n = 5–6 mice/group. Data are shown as mean ± SD. Scale bar: 50 μm. C, central vein; P, portal vein.

Figure 5

Inhibition of EGFR, c-Jun, and mTOR significantly affects INTR cell proliferation (A) Immunofluorescence for Ki67 in the EGFR inhibition experiment. TxRed, Ki67; Blue, DAPI. Bars show Ki67 positivity. (B) Relative mRNA expression of Jun, normalized to Gapdh. (C) Immunofluorescence for Ki67 in the JNK inhibition experiment. TxRed, Ki67; Blue, DAPI. Bars show Ki67 positivity. (D) Relative mRNA expression of Egfr, normalized to Gapdh. (E) Immunofluorescence for Ki67 in the mTOR inhibition experiment. TxRed, Ki67; Blue, DAPI. Bars show Ki67 positivity. (F) Immunofluorescence for c-Jun in the mTOR inhibition experiment. Arrowheads indicate representative positive cells among non-INTR cells. TxRed, c-Jun; Blue, DAPI. Bars show c-Jun positivity in non-INTR cells. (G) Ridge plot comparing DAPI intensity of non-INTR cell nuclei in control and mTORi groups. In all graphs, n = 5–6 mice/group. Data are shown as mean ± SD. Scale bar: 50 μm. C, central vein; P, portal vein.

Figure 6

Contact with necrotic tissue is necessary for the conversion to INTR cells (A) H&E staining of mice treated with EGFRi, JNKi, or mTORi. Arrowheads indicate representative INTR cells. n = 5–6 mice/group. (B) IHC for c-Jun and TP53INP1 around necrotic tissue- or gel-injected areas. Arrowheads represent positive cells. n = 4 mice/group. (C) IHC for c-Jun and TP53INP1 in partially hepatectomized mice. Arrowheads indicate representative positive cells; insert shows a magnified cell. n = 2 mice. Scale bar: 50 μm. C, central vein; P, portal vein.

Figure 7

INTR cells also appear in human liver injury (A) H&E and IHC for c-Jun and TP53INP1 in patients with HBV, AIH, and DILI. Arrowheads indicate positive cells. Dotted lines outline necrotic areas. Scale bar: 50 μm.