Mid-zone hepatocytes exhibit the most pronounced proliferative decline in early acute liver injury.

Wildtype C57BL/6J mice were intraperitoneally injected with 300 mg/kg APAP for 0, 3, 6, 12 and 24 h. (A) Immunohistochemical staining of Ki67 to detect proliferating hepatocytes in liver sections at 0, 3, 6, 12 and 24 h post-APAP. Representative images of pericentral (PC), middle (Mid) and periportal (PP) are shown. Red arrows indicate Ki67+ hepatocytes. (B) Quantitative analysis of Ki67+ hepatocytes across liver zones (PC, Mid, PP) at 0, 3, 6, 12, and 24 h post-APAP. (C) Quantitative analysis of Ki67+ hepatocytes across 0, 3, 6, 12, and 24 h post-APAP in each liver zone (PC, Mid, PP). (D) Changes of Ki67+ hepatocytes across liver zones (PC, Mid, PP) at 0, 3, 6, 12 and 24 h post-APAP. Value change=number of Ki67+ hepatocytes at x h post-APAP - the number of Ki67+ hepatocytes at 0 h post-APAP. (a) denotes significance between PC and Mid regions, (b) denotes significance between PC and PP regions, and (c) denotes significance between Mid and PP regions. Data points represent mean hepatocyte counts from six high-power fields per zonal layer (n=3 mice/group) in B and C. Data are represented as means ± SD; One-way ANOVA (B,C,D).*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant.

Mid-zone hepatocytes show distinct gene expression profiles.

(A) Heatmap of the top 10 differentially expressed genes in periportal (PP), midzonal (Mid) and pericentral (PC) hepatocyte zones at 0, 3, 6 and 24 h post-APAP. Functional categories at the bottom were assigned using Gene Ontology (GO) terms (biological process, cellular component, molecular function); for genes annotated with multiple GO terms, the most representative or significantly enriched term was selected. (B) Immunofluorescence staining for glutamine synthetase (GS, a pericentral marker) with DAPI (blue) in liver sections at 0, 3, 6, and 24 h after APAP administration. Quantification of the zonal distribution of GS-positive cells at the indicated time points. (C) Immunohistochemical detection of p62 (Sqstm1 protein, a midzone marker during early injury) in liver sections at 0, 3, 6, and 24 h after APAP administration. Red arrows indicate representative p62-positive hepatocytes. Quantification of the zonal distribution of p62-positive cells at the indicated time points. (D) Immunohistochemical staining of Cyp2f2 (a periportal marker) in liver sections at 0, 3, 6, and 24 h after APAP administration. Quantification of the zonal distribution of Cyp2f2-positive cells at the indicated time points. Data are presented as mean ± SD. One-way ANOVA (B-D). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant.

Zonal metabolism of APAP leads to stress response in Mid-zone hepatocytes.

(A) Schematic figure illustrating the metabolic process of liver injury caused by APAP. (B) Dot plot displays the scaled gene expression levels of the Cyp family in hepatocyte zones (zones PP, Mid, and PC) at 0, 3, 6, 24 h post-APAP. (C) Immunohistochemical staining of Cyp1a2 to detect its dynamic changes in hepatocyte zones (from CV to PV) at 0, 3, 6, 24 h post-APAP. Injured area is outline by black dashed lines. (D) Quantification of zonal distribution of Cyp1a2-positive cells in liver sections at 0, 3, 6, and 24 h post-APAP, respectively. (E) Schematic figure illustrating expression changes of Cyp1a2 before and after APAP. (F) Immunofluorescence staining of Chop protein (green) to observe the dynamic changes in its expression across hepatocyte zones (from CV to PV) at 6 h post-APAP with doses of 100, 300, or 500 mg/kg. Cyp1a2 protein (red) staining highlights the area around the CV. Cell nuclei are stained with DAPI (blue). (G) Quantification of zonal distribution of Chop-positive cells in liver sections from mice treated with varying doses of APAP– 100, 300, or 500 mg/kg. (H) Schematic explaining why mid-lobular hepatocytes display a distinct gene expression program early after APAP injury. Cyp enzymes are expressed in a decreasing gradient from the central vein (CV) to the portal vein (PV). APAP is therefore metabolized primarily by pericentral Cyp-expressing hepatocytes; their necrosis shifts the main site of residual Cyp activity to the mid-lobular zone. High local APAP metabolism then drives a specific stress-response program, accounting for the unique early transcriptional signature of the mid-zone. Data are represented as means ± SD; One-way ANOVA (D,G). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant.

The Atf4-Ddit3 axis emerges as pivotal in Mid-zone during early acute injury.

(A) Heatmap displays the area under the curve (AUC) scores of transcription factor (TF) motifs, estimated per dot by Single-Cell Regulatory Network Inference and Clustering (SCENIC), highlighting gene regulatory networks and differentially activated TF motifs in hepatocyte zones (PP, Mid, PC) at 6 h post-APAP. Columns represent TF motifs, rows represent dots, and color intensity indicates AUC scores. (B) Heatmap shows inferred transcription factors (TFs) activity across different zonal regions at 6 h post-APAP. Activity was quantified as regulon enrichment scores using AUCell. The color scale represents regulon activity scores. (C) Violin plots shows regulon activity of Atf4 and Ddit3 in each zonation at 6 h post-APAP. Activity was quantified as regulon enrichment scores using AUCell. (D) GO pathway analysis reveals the top 15 enriched pathways for genes co-regulated by Atf4 and Ddit3 in mid hepatocyte zones at 6 h post-APAP. (E) Heatmap showing the target genes expression of Atf4 (left) and Ddit3 (right) in hepatocyte zones (zones PP, Mid, and PC) at 6h post-APAP. (F) Immunohistochemistry staining of p-eIF2A in liver sections from mice at 0 h and 3 h post-APAP. p-eIF2A-positive hepatocytes are indicated by red arrows. The number of p-eIF2A-positive cells per field of view (FOV) is quantified. n=3 mice/group. (G) Heatmap displays stress response and ER stress-related genes from differentially expressed genes (DEGs) identified in Mid hepatocyte zones at 0 and 6 h post-APAP. Data are represented as means ± SD; Unpaired Student’s t-test (F). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant.

The Atf4-Ddit3 axis protects hepatocytes from liver injury.

(A) Immunohistochemical staining of Atf4 in liver sections at 0 and 6 h post-APAP. Red arrows indicate Atf4-positive hepatocytes. Zonal distribution of Atf4-positive cells in liver sections at 6 h post-APAP is quantified. The statistic is the percentage of Atf4-positive hepatocytes in each layer over the total number of Atf4-positive hepatocytes. n=3 mice. (B) Schematic figure illustrating the overexpression of Atf4 via AAV in hepatocytes of wildtype C57BL/6J mice. Overexpression of EGFP is used as a control. (C) H&E staining showing liver morphology from mice overexpressing AAV-TBG-EGFP or AAV-TBG-Atf4 at 6 h post-APAP. The injured area is outlined by black dashed lines. The percentage of injured area is quantified. n=3 mice/group. Necrotic areas outlined by loss of cellular architecture on H&E. (D) Serum levels of ALT is measured in mice overexpressing AAV-TBG-EGFP or AAV-TBG-Atf4 at 6, 24, 48, and 72 h post-APAP. n=6 mice/group. (E) Immunohistochemistry staining of Atf4 in liver sections from mice overexpressing AAV-TBG-EGFP or AAV-TBG- Atf4 at 24 h post-APAP. Atf4-positive hepatocytes are indicated by red arrows. The number of Atf4-positive cells per field of view (FOV) is quantified. n=9 mice/group. (F) Immunofluorescence staining of Ki67 (red) was performed to assess proliferating hepatocytes in mice overexpressing AAV-TBG-EGFP or AAV-TBG-Atf4 at 72 h post-APAP. Cell nuclei were stained with DAPI (blue). The number of Ki67-positive cells per FOV is quantified. n=6 mice/group. (G) Immunohistochemistry staining of p-eIF2A in liver sections from mice overexpressing AAV-TBG-EGFP or AAV-TBG- Atf4 at 24 h post-APAP. p-eIF2A-positive hepatocytes are indicated by red arrows. The number of p-eIF2A-positive cells per FOV is quantified. n=6 mice/group. Data are represented as means ± SD; Unpaired Student’s t-test (C-G). One-way ANOVA (A). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant.

The Atf4-Ddit3 axis protects the liver by pausing proliferation via Btg2.

(A) Heatmap of Ddit3-binding signals in chromatin domains. (B) Overlap of Ddit3-binding genes by Cut&Run with differentially expressed genes (DEGs) in mid zone was shown. These genes were referred as Ddit3-binding DEGs. (C) Integrative Genomics Viewer (IGV) plot shows the Cut&Run peaks of Ddit3 and input on Btg2-site, Ifrd1-site, Fos-site and Egr1-site. (D) Immunohistochemical staining of Btg2 in liver sections at 0 and 6 h post-APAP. Red arrows indicate Btg2-positive hepatocytes. (E) Schematic figure illustrating overexpression of Ddit3 via AAV in hepatocytes of wildtype C57BL/6J mice. Overexpression of EGFP is used as a control. (F) Immunohistochemical staining of Btg2 in liver sections from mice overexpressing either AAV-TBG-EGFP or AAV-TBG-Ddit3 at 12 h post-APAP. Btg2-positive hepatocytes are indicated by red arrows. Btg2-positive hepatocytes in liver sections is quantified. n=4 mice/group. (G) Schematic figure illustrating overexpression of Btg2 via AAV in hepatocytes of wildtype C57BL/6J mice. Overexpression of EGFP is used as a control. (H) H&E staining shows morphology of livers from AAV-TBG-EGFP or AAV-TBG-Btg2 overexpressing mice at 6 h post-APAP. Injured area is outlined by black dashed lines. The percentage of injury area is quantified. n=3 mice/group. Necrotic areas outlined by loss of cellular architecture on H&E. (I) Serum levels of ALT is measured at 6h post-APAP in AAV-TBG-EGFP or AAV-TBG-Btg2 overexpressing mice. n=3 mice/group. (J) Immunohistochemical staining of Ki67 in liver sections from mice overexpressing either AAV-TBG-EGFP or AAV-TBG-Btg2 at 72 h post-APAP. Ki67-positive hepatocytes are indicated by red arrows. Ki67-positive cells in liver sections is quantified. n=4 mice/group. (K) Schematic figure illustrating knockdown of Btg2 via AAV in hepatocytes of wildtype C57BL/6J mice. IP479-EFS-sgEGFP is used as a control. (L) Serum levels of ALT is measured at 6h post-APAP in IP479-EFS-sgEGFP or IP479-EFS-sgBtg2 knockdown mice. n=3 mice/group. (M) H&E staining shows morphology of livers from IP479-EFS-sgEGFP or IP479-EFS-sgBtg2 knockdown mice at 6 h post-APAP. Injured area is outlined by black dashed lines. The percentage of injury area is quantified. n=3 mice/group. Necrotic areas outlined by loss of cellular architecture on H&E. (N) Immunohistochemical staining of Ki67 in liver sections from IP479-EFS-sgEGFP or IP479-EFS-sgBtg2 knockdown mice at 72 h post-APAP. Ki67-positive cells in liver sections is quantified. n=3 mice/group. Data are represented as means ± SD; Unpaired Student’s t-test (F, H-N). **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant.

The Atf4-Ddit3 axis mediates integrated stress protection in mid-zone hepatocytes at the expense of proliferative capacity during early AILI.

Schematic model depicting the adaptive response of mid-zone hepatocytes to acute liver injury, wherein activation of the Atf4-Ddit3 pathway upregulates Btg2 to induce transient cell cycle arrest, balancing survival and regeneration.

(qPCR primers)

The liver displays transcriptome-wide zonation in AILI.

(A) Schematic figure illustrating the experimental strategy for spatial transcriptome analysis. n=1 mouse/time point. (B) Representative H&E staining images illustrating liver morphology in mice following intraperitoneal injection of 300 mg/kg APAP at various time points (0, 3, 6 and 24 h). These liver sections were subsequently used for spatial transcriptome analysis. Injured area is outline by red line. Necrotic areas outlined by loss of cellular architecture on H&E. (C) Spatial map visualizing the spatiotemporal dynamics of hepatocyte zones (zones PP, Mid, and PC) at 0, 3, 6, 24 h post-APAP, respectively. (D and E) UMAP visualizing the spatiotemporal dynamics of hepatocyte zones (zones PP, Mid, and PC) at 0, 3, 6, 24 h post-APAP, based on zonal distribution and time points, respectively. (F) Spatiotemporally resolved heatmaps of representative PC marker Glul, PP marker Cyp2f2 in zoom-in area. (G) Volcano plot illustrating the DEGs in Mid-zone at 6 h compared to 0 h post-APAP. Gray dots denote genes that are not statistically significant. Red and blue dots represent genes that are upregulated and downregulated, respectively, in the sample tissue, at least a 0.5-fold difference from the matched control, with a false discovery rate (FDR) threshold of 0.05. (H) Scatter plot shows the dynamic changes of Sqstm1 mean expression log₂(TPM) in hepatocyte zones (zones PP, Mid, and PC) at 0 and 6 h post-APAP. TPM: Transcripts per million. (I) Percentage of Ki67-positive spots analyzed at 0, 3, 6 and 24 h post-APAP. (J) Average module score for S-phase (up) and G2/M-phase (down) genes across hepatocyte zones (zones PP, Mid, and PC) at 0, 3, 6, 24 h post-APAP. (K) Expression levels of the S-phase gene Nasp (up) and the G2/M phase gene Cks1b (down) in each hepatocyte zone (zones PP, Mid, and PC) at 0, 3, 6, 24 h post-APAP.

The zonal expression of the Cyp family dictates the zonal hepatocyte response to APAP.

(A) Scatter plot showing the dynamic changes of Cyp2e1 and Cyp1a2 mean expression log₂ (TPM), in hepatocyte zones (PP, Mid, PC) at 0, 3, 6, and 24 h post-APAP. TPM: Transcripts per million. (B) Representative H&E staining images illustrating liver morphology in mice following intraperitoneal injection of various doses of APAP (100, 300, 500mg/kg) at 6h post APAP. Injured area is outline by black dashed lines. Necrotic areas outlined by loss of cellular architecture on H&E. (C) The percentage of injured area is quantified. n=3 mice/group. Data are represented as means ± SD; One-way ANOVA (C). *p < 0.05; ***p < 0.001.

The Atf4-Ddit3 axis emerges as pivotal in mid-lobular hepatocytes during the initial stages of acute injury.

(A) Heatmap displays the area under the curve (AUC) scores of transcription factor (TF) motifs, estimated per dot by Single-Cell Regulatory Network Inference and Clustering (SCENIC), highlighting gene regulatory networks and differentially activated TF motifs in hepatocyte zones (PP, Mid, PC) at 3 h post-APAP. Columns represent TF motifs, rows represent dots, and color intensity indicates AUC scores. (B) The heatmap shows inferred transcription factors (TFs) activity across different zonal regions at 3 h post-APAP. Activity was quantified as regulon enrichment scores using AUCell. The color scale represents regulon activity scores. (C) Violin plots shows regulon activity of Atf4 (left) and Ddit3 (right) in each zonation at 3 h post-APAP. Activity was quantified as regulon enrichment scores using AUCell. (D) Gene regulatory network displays Atf4 and its target genes, as well as Ddit3 and its target genes at 3 h post APAP, respectively. Edge color indicates the regulatory strength of transcription factors (TFs) on target genes, as measured by AUCell scores. (E) GO pathway analysis reveals the top 15 enriched pathways for genes co-regulated by Atf4 and Ddit3 in mid hepatocyte zones at 3 h post-APAP. (F) GO pathway analysis reveals the top 15 enriched pathways for up-regulated genes in Mid hepatocyte zones at 3 h compared to the rest hepatocyte zones post-APAP. (G) Heatmap displays stress response and ER stress-related genes from differentially expressed genes (DEGs) identified in Mid hepatocyte zones at 0 and 3 h post-APAP.

Validation of zonal hepatocyte responses and transcriptional changes post-APAP using the GSE223561 scRNA-seq dataset.

(A) UMAP visualization of hepatocytes colored by cell type annotations. (B) UMAP visualization of hepatocytes colored by time post-APAP treatment. (C) UMAP visualization of hepatocytes colored based on the expression of centrally zonated genes (Glul, Cyp2e1, Cyp2a5), periportal genes (Alb, Cyp2f2, Sds), and mid-zonal genes (Hmox1, Atf3, Ubc). (D) GO pathway analysis showing the top 10 enriched pathways for upregulated genes in mid-zonal hepatocytes at 3 h post-APAP. (E) GO pathway analysis showing the top 10 enriched pathways for upregulated genes in mid-zonal hepatocytes at 6 h post-APAP. (F) Heatmap illustrating the area under the curve (AUC) scores of transcription factor (TF) motifs (Top 50), as estimated per cell using Single-Cell Regulatory Network Inference and Clustering (SCENIC). Differentially activated motifs in each zonal region are shown for 3 h post-APAP. (G) Heatmap illustrating the AUC scores of TF motifs (Top 50) estimated per cell using SCENIC, highlighting differentially activated motifs in each zonal region at 6 h post-APAP.

Transduction efficiency and confounding effects of AAV-TBG overexpression.

(A) Western blot analysis of Cyp2e1 protein in whole liver lysates from mice transduced with 1.2 × 10¹¹viral genome copies of AAV-TBG-EGFP, AAV-TBG-Atf4, or AAV-TBG-Btg2. Cyp2e1/Gapdh ratios are shown. n = 4 mice/group. (B) Transduction efficiency of AAV-TBG-EGFP, AAV-TBG-Atf4, and AAV-TBG-Btg2 in mouse liver was measured by immunostaining for the transgene proteins. n = 4–8 mice/group. (C) Correlation between transduction efficiency of AAV-TBG-Atf4 or AAV-TBG-Btg2 and serum ALT levels in the respective AAV-injected mice. Pearson correlation coefficient r and p values are indicated. (D) Btg2 mRNA expression in liver tissues from IP479-EFS-sgEGFP and IP479-EFS-sgBtg2 mice. n=9 mice/group. Data are represented as means ± SD; Pearson correlation (C); Unpaired two tailed Student’s t-test (B, D). *p < 0.05; **p < 0.01; ****p < 0.0001.

The ISR-Btg2 axis emerges in CCL4-induced acute liver injury.

(A) H&E staining shows morphology of livers from mice at 18 h post oil and CCL4 injection. Injured area is outlined by black dashed lines. The percentage of injury area is quantified. n=4 mice/group. Necrotic areas outlined by loss of cellular architecture on H&E. (B) Serum levels of ALT is measured at 18h post-CCL4 mice. n=4 mice/group. (C) Immunohistochemistry staining of p-eIF2A in liver sections from mice at 18 h post oil and CCL4 injection. p-eIF2A-positive hepatocytes are indicated by red arrows. n=4 mice/group. (D) Immunohistochemistry staining of Atf4 in liver sections from mice at 18 h post oil and CCL4 injection. Atf4 -positive hepatocytes are indicated by red arrows. n=4 mice/group. (E) Immunofluorescence staining of Chop protein (green) in liver sections from mice at 18 h post oil and CCL4 injection. Cyp1a2 protein (red) staining highlights the area around the CV. Cell nuclei are stained with DAPI (blue). n=4 mice/group. (F) Immunohistochemistry staining of Btg2 in liver sections from mice at 18 h post oil and CCL4 injection. Btg2 -positive hepatocytes are indicated by red arrows. n=4 mice/group. (G) Immunofluorescence staining of Ki67 (red) was performed to assess proliferating hepatocytes in mice at 18 h post oil and CCL4 injection. Cyp1a2 protein (red) staining highlights the area around the CV. Cell nuclei were stained with DAPI (blue). n=4 mice/group. Data are represented as means ± SD; Unpaired Student’s t-test (A, B). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant.

The ISR-Btg2 axis emerges after partial hepatectomy.

(A) H&E staining shows morphology of livers from mice at 24 h post sham and partial hepatectomy. Injured area is outlined by black dashed lines. The percentage of injury area is quantified. n=5 mice/group. Necrotic areas outlined by loss of cellular architecture on H&E. (B) Serum levels of ALT is measured at 24 h post sham and partial hepatectomy. n=5 mice/group. (C) Immunohistochemistry staining of p-eIF2A in liver sections from mice at 24 h post sham and partial hepatectomy. p-eIF2A-positive hepatocytes are indicated by red arrows. n=5 mice/group. (D) Immunohistochemistry staining of Atf4 in liver sections from mice at 24 h post sham and partial hepatectomy. Atf4 -positive hepatocytes are indicated by red arrows. n=5 mice/group. (E) Immunofluorescence staining of Chop protein (green) in liver sections from mice at 24 h post sham and partial hepatectomy. Cyp1a2 protein (red) staining highlights the area around the CV. Cell nuclei are stained with DAPI (blue). n=5 mice/group. (F) Immunohistochemistry staining of Btg2 in liver sections from mice at 24 h post sham and partial hepatectomy. Btg2 -positive hepatocytes are indicated by red arrows. n=5 mice/group. (G) Immunofluorescence staining of Ki67 (red) was performed to assess proliferating hepatocytes in mice at 24 h post sham and partial hepatectomy. Cyp1a2 protein (red) staining highlights the area around the CV. Cell nuclei were stained with DAPI (blue). n=5 mice/group. Data are represented as means ± SD; Unpaired Student’s t-test (A, B). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant.