Fig 2: Graphical summary illustrating the role of SALL4 and BMI1 in reparative HC-to-CC conversion in DDC-induced cholestatic liver injury and AY-driven malignant HC-to-CCA transformation. OE, overexpression.

Fig 3: SALL4 regulates regenerative HC-to-CC conversion in the DDC-fed cholestasis model. A, Schematic experimental design. Sall4fl/fl mice were injected with either AAV8-TBG-GFP (Sall4 WThep-GFP) or AAV8-TBG-Cre (Sall4 KOhepΔGFP) 2 weeks prior administration of a DDC-containing diet for 4 weeks. Liver tissues were analyzed at the end of the treatment period. B, LW/BW ratio comparison and serum biochemistry analysis comparing alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), and total bilirubin levels between Sall4 WThep-GFP and Sall4 KOhepΔGFP groups (n = 4–6 mice per group). C, Representative images of CD45 IHC and Sirius Red staining in livers sections from both groups. D and E, Quantification of (D) CD45+ inflammatory cells per 1,000 μm2 and (E) collagen deposition (%) by Sirius Red staining. F, Representative liver sections stained with hematoxylin and eosin (H&E), SOX9 (CC lineage marker), and Ki-67 (proliferation marker). G and H, Quantification of (G) SOX9+ cells (%) and (H) Ki-67+ cells in the periportal area (%). I, IF staining for SOX9, HNF4A, and GFP. Yellow arrowheads indicate intermediate cells expressing both CC (SOX9) and HC (HNF4A) markers. J, Quantification of SOX9+HNF4A+ double-positive intermediate cells among total HNF4A+ cells (%). K, IF staining for CK19 and GFP to visualize HC-derived CC. L, Quantification of CK19+ CCs per microscopic field. M, Quantification of CK19+GFP+ double-positive cells among total CK19+ cells (%). Data are presented as the mean ± SD; Student t test (*, P < 0.05; **, P < 0.01; ****, P < 0.0001; ns, not significant). Scale bars, (C–F) 100 μm; (K–M) 50 μm.

Fig 4: Sall4 overexpression disrupts YAP1-driven HC-to-CCA transformation while enhancing the clonal expansion of fatty LPC-like cells. A, Schematic representation of the experimental design. FVB mice underwent HDTVI with transposon vectors expressing HA-tagged myrAkt and Yap1S127A (AY) combined with GFP expression (AY-GFP) or V5-tagged Sall4 (AY-Sall4) plasmids and were analyzed at indicated time points (2.5 and 5 weeks after HDTVI). B, qRT-PCR showing Sall4 expression in AY-Sall4 compared with AY-GFP. C, Representative liver images from AY-GFP or AY-Sall4 mice at 2.5 (top) and 5 weeks (bottom) after HDTVI. D, LW/BW ratio analysis comparing AY-GFP and AY-Sall4 at indicated time points. E and O, IHC analysis for indicated markers (HA-tag, panCK, HNF4A, Ki-67, and V5-tag) at indicated time points. F–N, Quantification of indicated parameters [tumor area, panCK+ area, lipid droplet area, Ki-67+ cells, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), bilirubin, CD45, TUNEL, and α-SMA] comparing AY-GFP and AY-Sall4. Data are presented as the mean ± SD; Student t test (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001). Scale bars, (E and O) 1 mm (overview), 200 μm (high magnification); (K) 200 μm. ns, not significant.

Fig 5: Bmi1 modulation phenocopied Sall4 modulation in AY-driven HC-to-CCA tumorigenesis. A, qRT-PCR analysis showing Bmi1 expression in AY-Sall4 compared with AY-GFP. B, Schematic experimental design. C and H, Representative macroscopic images of livers harvested from indicated groups. D, LW/BW ratio comparison (n = 3–5 per group). E and I, IHC analysis for indicated markers (HA-tag, HNF4A, Ki-67, panCK, and V5-tag) in experimental groups. F and J, Quantification of Ki-67+ proliferating cells (%) and lipid droplets area per field (%) comparing Bmi1 KOCCAΔ and AY-Bmi1 tumors. G, qRT-PCR analysis showing Bmi1 expression in AY-Bmi1 compared with AY-GFP. Data are presented as the mean ± SD; Student t test (*, P < 0.05; ***, P < 0.001; ****, P < 0.0001). Scale bars, (E and I) 2 mm (overview), 200 μm (magnified views).