PAR3-mediated coordination of hepatocyte proliferation, maturation, and architecture in liver development and regeneration

Li Yang; Yu-Heng Zhou; Junya Hayase; Lei Wang; Weiguang Zhang; Erfei Bi; Cheng-Ran Xu

doi:10.1038/s41467-026-73856-w

. 2026 Jun 12;17:7611. doi: 10.1038/s41467-026-73856-w

PAR3-mediated coordination of hepatocyte proliferation, maturation, and architecture in liver development and regeneration

Li Yang ^1,^✉, Yu-Heng Zhou ², Junya Hayase ³, Lei Wang ⁴, Weiguang Zhang ⁴, Erfei Bi ^3,^✉, Cheng-Ran Xu ^2,^5,^✉

PMCID: PMC13424312 PMID: 42285921

Abstract

The liver’s ability to coordinate cell proliferation, polarization, and maturation with architectural assembly is fundamental to its metabolic functions and regenerative capacity. However, the mechanisms that integrate these processes remain incompletely understood. Using murine and human models, we show that hepatic polarity and bile canaliculi (BC) formation initiate at sites of cell division during the hepatoblast stage, and that BC elongation is temporally associated with oriented cell division during hepatocyte differentiation. This process involves PAR3- and YAP-dependent modulation of ERK/MAPK signaling. PAR3 forms a protein complex that localizes a subset of IQGAP1 to tight junctions, thereby restricting ERK/MAPK activation at the basolateral membrane. This regulation is required for efficient BC elongation and hepatocyte maturation. Together, these findings reveal a stage-dependent mechanism linking liver architecture to functional maturation and provide insights into liver development, regeneration, and potential therapeutic strategies for metabolic liver disease and liver failure.

Subject terms: Cell proliferation, Differentiation, Apicobasal polarity, Regeneration, Data processing

Here they show that PAR3 promotes bile canaliculus elongation and hepatocyte maturation by restricting ERK/MAPK signaling, linking liver architecture formation with functional maturation during development and regeneration.

Introduction

Coordinating cell proliferation, architectural assembly, and functional maturation is a central challenge in developmental biology, regenerative medicine, and metabolism research. These processes are particularly critical in the liver, where they underpin the establishment of metabolic homeostasis, bile acid secretion, and detoxification. Understanding the molecular mechanisms that integrate cellular organization with functional maturation is essential for elucidating normal organogenesis and advancing therapeutic strategies for liver repair and regeneration.

The liver, a central metabolic organ, relies on hepatocytes—the primary parenchymal cells—to perform its diverse functions, including bile acid secretion, nutrient metabolism, and detoxification. Hepatocytes emerge from endoderm-derived bipotential hepatoblasts around embryonic day (E) 14.5 in mice and approximately 9 weeks post coitum (W) in humans^1,2. These hepatoblasts undergo extensive developmental transitions from the embryonic to postnatal stages, culminating in mature hepatocytes with specialized polarity and architecture^1–3. Advances in transcriptomics, proteomics, genetics, and epigenetics have expanded our understanding of hepatocyte differentiation and function, highlighting the interplay between structural and metabolic maturation^1–8.

A hallmark of hepatocyte maturation is the formation of bile canaliculi (BC)—specialized apical structures that serve as conduits for bile acid excretion and metabolic waste removal. Hepatocytes are arranged in cord-like structures, with apical surfaces facing each other to form BC networks, while their basal surfaces engage closely with sinusoidal endothelial cells to facilitate oxygen and nutrient exchange^9,10. The establishment of hepatocyte polarity and BC formation is fundamental for liver metabolic functions. However, the molecular mechanisms driving these processes during development remain poorly understood, particularly their integration with metabolic maturation.

Previous studies using hepatic cell lines such as WIF-B and HepG2 identified several regulators of hepatocyte polarity and BC formation, including Par1b, myosin II, myosin Vb, Rab11a, Rho-associated protein kinase (ROCK), and the Vps34p-PI3K pathway^11–16. However, Pard3 knockdown yields divergent outcomes in culture—abolishing lumen formation in HepG2 cells¹⁷ but inducing multi-lumen phenotypes in MDCK cells due to aberrant spindle orientation¹⁸. In primary hepatocytes, bile acids activate Epac-Rap1-MEK-LKB1-AMPK signaling and the cAMP-PKA-LKB1 pathway to promote ABCB11 trafficking and BC network formation, independent of cell division^19–21. Recent evidence further indicates that Rab35 facilitates BC morphogenesis by promoting the formation of apical transversal structures²², while in vivo studies implicated Rab35, CDC42, and CD147 as key regulators of hepatocyte polarity and BC architecture^22–24. Despite these advances, the molecular and cellular mechanisms governing hepatocyte polarization and BC formation during liver development remain incompletely understood.

Our previous work using the rat hepatocyte line Can 10 demonstrated that the sequential localization of the polarity protein PAR3, tight-junction (TJ) protein ZO1, and apical cargo protein MDR1 to the division site during the terminal stage of cytokinesis is critical for linking cell division to the establishment of hepatocyte polarity and primordial BC formation, and oriented cell division further contributes to BC elongation^25,26.

Here, we present a comprehensive analysis of hepatocyte polarity and BC development in both murine and human models. We show that hepatocyte polarization and primordial BC formation initiate during the hepatoblast stage, with BC elongation accompanying hepatocyte differentiation. These processes are linked to cytokinesis and oriented cell division, reflecting the coordinated integration of cell proliferation, architectural assembly, and functional maturation. Mechanistically, we identify a polarity–signaling axis in which PAR3 interacts with IQGAP1—a scaffold that bridges ERK to its upstream activators RAF and MEK^27,28, and also regulates adherens junctions, TJs, actin organization, and Cdc42/Rac signaling^27–29. PAR3 forms a protein complex that localizes a subset of IQGAP1 to TJs, thereby limiting its availability at the basolateral membrane and restraining ERK/MAPK activation. Loss of PAR3 causes IQGAP1 relocalization at the basal membrane and elevated RAF and p-ERK levels, disrupting the cytoskeleton, impairing oriented cell division, and ultimately hindering BC elongation. Strikingly, hepatocyte-specific deletion of Pard3 or the Hippo pathway effector Yap—a potential PAR3 interactor at the TJ^30–32—produces comparable defects in BC elongation, hepatocyte proliferation, and maturation, all mediated through ERK/MAPK hyperactivation. Moreover, the PAR3–ERK/MAPK pathway regulates BC architecture during liver regeneration, concomitant with altered proliferation. Notably, ERK/MAPK suppression is also critical for spindle orientation during airway tube elongation³³, underscoring a conserved epithelial mechanism. Together, these findings establish PAR3 and ERK/MAPK as regulators that coordinate hepatocyte proliferation, polarity, and BC morphogenesis to ensure proper liver development, regeneration, and metabolic function.

Results

Development of hepatic cell polarity and BC architecture

To investigate the establishment of hepatic polarity and BC formation, we analyzed key polarity proteins and apical cargo molecules critical for BC development. The polarity protein PAR3 and the TJ protein ZO1 are known to co-localize at the TJs in hepatocytes²⁵, while the apical cargo protein MDR1/ABCB1 facilitates the transport of hydrophobic substrates into BCs (Fig. 1a)³⁴. Using immunofluorescence, we assessed these markers across developmental stages in mouse liver sections from embryonic day E11.5 to postnatal day (P) 60 (Fig. 1b).

Fig. 1 — a Schematic illustrating the stages of hepatic cell polarization and BC formation, showing the localization of PAR3, ZO1, and MDR1 on primordial and elongated BCs. b Experimental workflow for the analysis of hepatic cell polarity and BC architecture in mouse and human livers. Mouse and human schematics were created using BioRender. Xu, X. (2026) https://BioRender.com/luccdzh. c HNF4A, ZO1, and MDR1 staining in embryonic day E11.5 to P60 mouse livers. Scale bars, 50 μm. d Quantification of ZO1⁺MDR1⁺HNF4A⁺ polarized cells among total HNF4A⁺ cells and of cells forming primordial or elongated BCs in (c). e 3D reconstruction of ZO1 signals showing BC network organization in mouse livers from E11.5 to P60. Scale bars, 50 μm. f HNF4A, ZO1, and MDR1 staining in W5 to W14 human embryonic livers. Scale bars, 50 μm. g Quantification of ZO1⁺MDR1⁺HNF4A⁺ polarized cells and cells forming primordial or elongated BCs in (f). h, i, AURKB, ZO1, and MDR1 staining in mouse (h) and human (i) embryonic livers. Yellow arrowheads, AURKB signals localized at BCs. At least 3 biological replicates for each time point. Scale bars, 10 μm. Data (d, g) represent mean ± s.e.m.; N, the number of mice (d) or human embryos (g). Source data (d, g) are provided as a Source Data file.

At E11.5, approximately 45% of HNF4A⁺ hepatoblasts exhibited membrane co-localization of ZO1 and MDR1, indicating the initiation of hepatic polarity and BC formation. This co-localization increased to 65% by E13.5 and 72% by E14.5 (Fig. 1c-e and Supplementary Fig. 1a). These findings suggest that hepatic cell polarity and the formation of primordial BCs—initial BC structures formed between polarizing hepatoblasts—begin during the early hepatoblast stage. By E15.5, over 85% of hepatocytes displayed clear polarization, characterized by elongated or tubular BC structures extending from primordial BCs to connect rows of hepatocytes (Fig. 1c-e and Supplementary Fig. 1a)²⁵. This elongation closely coincided with the hepatoblast-to-hepatocyte transition, a critical phase of structural and functional maturation.

As development progressed, nearly all hepatocytes attained polarity, culminating in the formation of a complex BC network within cord-organized hepatocytes (Fig. 1c-e and Supplementary Fig. 1a,b). Notably, polarized hepatoblasts at E13.5 and elongated BCs at E15.5 and E17.5 appeared to be broadly distributed throughout the liver, without an obvious spatial bias relative to blood vessels or the hilum-periphery axis (Supplementary Fig. 1c-e).

To determine whether these processes are conserved in humans, we extended our analysis to human liver sections (Fig. 1b). At W5 of gestation, approximately 34% of HNF4A⁺ hepatoblasts exhibited polarization, as indicated by co-localization of ZO1 and MDR1 (Fig. 1f,g and Supplementary Fig. 1f). By W8, polarization increased to 70%, accompanied by the formation of elongated BC structures. By W14, nearly 92% of hepatocytes were polarized, with BC elongation observed in over 92% of these cells (Fig. 1f,g and Supplementary Fig. 1f).

These results suggest a conserved developmental trajectory of hepatic polarity and BC biogenesis between mouse and human models. In both species, BC elongation is closely linked to the hepatoblast-to-hepatocyte transition, underscoring the importance of coordinated polarization and structural assembly in establishing liver architecture and metabolic function.

Linking hepatic polarity and BC formation to cell division

Our previous in vitro studies using the rat hepatocyte line Can 10 demonstrated that hepatocyte polarization and the formation of primordial BCs are spatially coupled to cytokinesis²⁵. Live imaging of Can 10 cells co-expressing mScarlet–Radixin (an apical marker) and GFP–Mzt1 (a centrosomal/pericentriolar material protein) further showed that elongating BCs arose through oriented cell division (Supplementary Fig. 2a,b). To assess whether this relationship extends beyond immortalized cell lines, we also performed live imaging analyses of primary E14.5 DLK⁺ hepatoblasts, which represent a physiologically relevant and proliferative population in vivo. Real-time tracking of these primary cells revealed that approximately 71.1% (27 of 38 nascent BCs) and 73.9% (34 of 46 elongated BCs) are associated with cell division (Supplementary Fig. 2c,d). These observations support a mechanistic association between hepatic polarity, BC formation, and cell division²⁵.

To test this relationship in vivo, we used Aurora B (AURKB) as a marker of the midbody stage to visualize the location and orientation of dividing hepatocytes³⁵. AURKB signals were detected at sites of primordial BCs (co-localized ZO1⁺ or MDR1⁺) from E11.5 to E13.5 in mouse livers and from W5 to W8 in human fetal livers (Fig. 1h,i). Notably, AURKB was also present at the sites of elongated BCs in E15.5-E17.5 mouse and W11-W14 human liver sections (Fig. 1h,i). Co-staining with α-tubulin (TUBA) confirmed that AURKB signals correspond to midbody microtubule arrays, reinforcing the association between cell division and BC formation (Supplementary Fig. 2e).

Together, these data indicate that cell division is associated with de novo or primordial BC formation and BC elongation during the early stages of liver development in both mouse and human models. This highlights the contribution of coordinated cell division to proper liver architecture.

Disruption of Pard3 impairs BC elongation

PAR3 is a key polarity regulator essential for maintaining epithelial architecture and function³⁶. To investigate its role in hepatic polarity and BC development, we generated a conditional Pard3 knockout (KO) mouse using Foxa3-Cre (Fig. 2a), which is active in the definitive endoderm prior to hepatoblast specification³⁷. Reverse transcription-quantitative PCR (RT-qPCR) and immunofluorescence confirmed efficient Pard3 deletion in both hepatoblasts and hepatocytes (Supplementary Fig. 3a,b).

Fig. 2 — a Schematic of the *Pard3* conditional KO mouse model. b 3D reconstruction of ZO1 signals in E15.5 and E17.5 WT and *Pard3*-KO livers. Scale bars, 50 μm. c Quantification of 3D BC length in (b). d HNF4A, ZO1, and MDR1 staining in E15.5 and E17.5 WT and *Pard3*-KO livers. Scale bars, 50 μm. e Quantification of BC fragment length in (d). f Transmission electron micrographs of E15.5 WT and *Pard*3-KO BC ultrastructure. Red arrowheads, lumens; yellow arrowheads, TJs; blue arrowheads, microvilli. 3 biological replicates for each genotype. Scale bars, 500 nm. g PCA of E11.5 to E19.5 hepatoblasts and hepatocytes. Arrow, the developmental trajectory. h PCA of E17.5 WT and *Pard3*-KO hepatocytes. Cell numbers are indicated in parentheses. i Pseudotime distribution of hepatocytes in (h). j Volcano plot showing DEGs in E17.5 WT and *Pard3*-KO hepatocytes. Numbers indicate significantly upregulated genes. k Selected GO terms for DEGs in (j). l Expression of hepatocyte maturation-associated genes in E17.5 WT and *Pard3*-KO hepatocytes. Violin plots display median lines, and minimum and maximum values. m HNF4A, Ki67, and MDR1 staining in E17.5 WT and *Pard3*-KO livers. Scale bars, 10 μm. n Quantification of Ki67⁺ hepatocytes in (m). o Heatmap of differentially accessible chromatin regions in E17.5 WT and *Pard3*-KO hepatocytes. Numbers in parentheses indicate the number of regions. p Expression of genes associated with regions in (o). q Enriched transcription factor–binding motifs for regions in (o). Data (c, e, n) represent mean ± s.e.m.; N, the number of mice. Boxplots (i, p) display median lines, 25th–75th percentiles (box), whiskers extending 1.5 times the interquartile range from the edges, and minimum and maximum values at whisker ends. A two-tailed unpaired Student’s t-test (c, e, n), a two-tailed Wilcoxon rank-sum test (i), and a two-tailed Wilcoxon signed-rank test (l, p). n.s., not significant. Cells of each genotype (h, i, l) were derived from 2 mice (refer to Supplementary Data 1). Source data (c, e, h, i, l, n) are provided as a Source Data file.

To determine whether PAR3 is required for early polarity establishment, we performed ZO1 and MDR1 immunostaining. At E13.5, the proportions of polarized hepatoblasts and primordial BCs were comparable between Pard3-KO and wild-type (WT) livers (Supplementary Fig. 3c-e), indicating that PAR3 is dispensable for the initiation of hepatoblast polarity and primordial BC formation.

By E15.5, however, pronounced defects in BC elongation became evident in Pard3-KO livers. ZO1 immunostaining revealed a marked reduction in BC length compared with WT controls (Fig. 2b-e). Transmission electron microscopy confirmed intact TJs in Pard3-KO hepatocytes but showed BC lumina lacking microvilli—structures critical for bile secretion and metabolic function (Fig. 2f). As development progressed, these defects persisted: by E17.5, WT livers displayed elongated BCs, whereas Pard3-KO BCs remained short (Fig. 2b-e).

Together, these results demonstrate that PAR3 is dispensable for initial hepatoblast polarization but essential for BC elongation and microvilli formation. Loss of PAR3 disrupts these processes, impairing hepatic architecture and prompting further investigation into the underlying molecular mechanisms.

Disruption of Pard3 delays hepatocyte development

To assess the impact of Pard3 deletion on hepatoblast/hepatocyte development, we labeled hepatoblasts and hepatocytes from E11.5 to E19.5 both WT and Pard3-KO livers with DLK antibody^1,38, isolated the DLK⁺ cells using fluorescence-activated cell sorting (FACS), and performed high-precision single-cell RNA sequencing (scRNA-seq) with the Smart-seq3 method (Supplementary Fig. 3f,g and Supplementary Data 1)³⁹. To ensure consistent data quality across developmental stages, we reconstructed the hepatoblast/hepatocyte developmental trajectory from E11.5 to E19.5. Although similar trajectories have been reported in previous studies^1,2, this approach enabled detailed comparative analyses between WT and Pard3-KO livers. Using Smart-seq3 scRNA-seq, we sorted hepatoblasts and hepatocytes from E11.5 to E19.5 WT mice^1,3 and detected an average of 8000 genes per cell. Principal component analysis (PCA) revealed that cells from each time point clustered distinctly and aligned linearly along the developmental trajectory, representing a coherent progression of hepatocyte maturation (Fig. 2g and Supplementary Data 1). This comprehensive trajectory provided a robust framework for evaluating how Pard3-KO affects hepatocyte development and functional maturation.

When mapped onto the developmental trajectory, Pard3-KO and WT hepatoblast/hepatocytes at E11.5 to E15.5 exhibited similar developmental states with minimal differences in gene expression profiles, as confirmed by differential gene expression (DGE) analysis (Supplementary Fig. 3h-p). Consistently, in vitro colony formation assays indicated that Pard3 deletion did not affect the differentiation potential of E13.5 hepatoblasts (Supplementary Fig. 3q,r). By E17.5 and E19.5, however, Pard3-KO hepatocytes displayed a significant developmental delay, as evidenced by their altered positioning on the trajectory (Fig. 2h,i and Supplementary Fig. 3s,t). DGE analysis identified 640 and 233 upregulated genes in Pard3-KO hepatocytes compared to WT, while 139 and 222 genes were more highly expressed in WT hepatocytes at E17.5 and E19.5, respectively (Fig. 2j and Supplementary Fig. 3u). Gene ontology (GO) analysis revealed that Pard3-KO hepatocytes exhibited increased expression of genes associated with cell proliferation and MAPK signaling. In contrast, WT hepatocytes showed enrichment for genes critical to core hepatocyte functions, including glucose and lipid metabolism, as well as drug detoxification (Fig. 2k,l and Supplementary Fig. 3v). Further analysis using Ki67 staining corroborated these findings, showing elevated proliferation in E17.5 Pard3-KO hepatocytes compared to WT (Fig. 2m,n).

Additionally, ATAC-seq analysis revealed widespread changes in chromatin accessibility in E17.5 Pard3-KO hepatocytes. A total of 1,987 regions with increased accessibility were identified in Pard3-KO cells, significantly higher than the 461 regions observed in WT hepatocytes (Fig. 2o and Supplementary Data 2). Genes associated with these more accessible regions in Pard3-KO hepatocytes showed elevated expression levels compared to WT cells (Fig. 2p). Motif enrichment analysis revealed that these regions were enriched with transcription factors (TFs) linked to the MAPK pathway and cell proliferation, including FRA1/FOSL2, JUN, and AP1 (Fig. 2q)^40–45. These results indicate a shift in transcriptional regulation that promotes cell proliferation while disrupting pathways required for normal hepatic function.

To ensure liver-specific Pard3 deletion, we also generated Pard3-KO mice using Alfp-Cre, which is active exclusively in hepatoblasts from mid-gestation onward⁴⁶. The knockout efficiency achieved with Alfp-Cre was lower than that obtained with Foxa3-Cre (77.7% vs. 98.9%, respectively; Supplementary Fig. 3a,4a). Nevertheless, Alfp-Cre-mediated deletion still caused reduced BC elongation (Supplementary Fig. 4b-e) and delayed hepatocyte maturation, as confirmed by single-cell transcriptome analysis (Supplementary Fig. 4f,g). DGE analysis identified 167 downregulated genes enriched for hepatocyte functions and 116 upregulated genes associated with proliferation in Alfp-Cre-mediated Pard3 knockout livers (Supplementary Fig. 4h,i). By comparison, Foxa3-Cre-mediated deletion affected a larger number of genes overall (779 vs. 283) (Supplementary Fig. 4j-l), consistent with the higher recombination efficiency achieved in this model. Notably, Pard3 deletion did not disrupt apical–basal polarity in intrahepatic bile duct epithelial cells at E17.5. However, quantitative analysis revealed altered ductal morphogenesis, with an increased proportion of asymmetric ductal structures and a concomitant reduction in symmetric ones (Supplementary Fig. 4m-p). Importantly, at this developmental stage, BC remain fragmented and have not yet formed a continuous network with bile ducts^47,48. Therefore, the BC elongation defects observed upon Pard3 knockout cannot be attributed to secondary effects arising from bile duct morphogenesis. Together, analyses of both Foxa3-Cre and Alfp-Cre models demonstrate that PAR3 is required for efficient BC elongation and hepatocyte maturation. Consistent with the lower recombination efficiency, the phenotypes observed in Alfp-Cre mice were milder than those in Foxa3-Cre mice.

The consequences of Pard3 disruption extended beyond embryonic development, severely impairing postnatal survival. Most Pard3-KO mice failed to survive beyond two weeks after birth (Supplementary Fig. 5a). At P0, these mice exhibited jaundiced, shorter BCs, an absence of microvilli in the BC lumen, and elevated hepatocyte proliferation compared to WT littermates (Supplementary Fig. 5b-h). By P14, Pard3-KO mice displayed reduced body weight, yellowish livers, and an increased liver-to-body weight ratio (Supplementary Fig. 5i,j). Abnormal BC morphology and elevated hepatocyte proliferation persisted in these mice (Supplementary Fig. 5k-o), underscoring the long-lasting developmental defects caused by PAR3 loss. Single-cell transcriptome profiling analysis also reveals that Pard3 deletion delayed hepatocyte maturation at both P0 and P14 (Supplementary Fig. 5p-x).

Together, these findings highlight the essential role of PAR3 in maintaining hepatocyte structural integrity, functional maturation, and controlled proliferation. By balancing proliferation with differentiation, PAR3 ensures proper liver development, BC architecture, and postnatal survival.

YAP-PAR3 interaction in BC elongation and hepatocyte development

To further investigate PAR3’s role in BC elongation and hepatocyte development, we conducted co-immunoprecipitation followed by mass spectrometry (Co-IP/MS) on E17.5 WT and Pard3-KO hepatocytes using a PAR3 antibody (Fig. 3a and Supplementary Fig. 6a). This analysis identified 764 potential PAR3-interacting proteins (Supplementary Data 3), enriched in processes such as actin cytoskeleton organization, cell junction assembly, cell adhesion regulation, and polarity development (Fig. 3b). Key interactors included ZO1, CDC42, Radixin (RDX), and Ezrin (EZR) (Fig. 3c, Supplementary Fig. 6b,c and Supplementary Data 3)⁴⁹. Notably, the signaling molecule YAP was also identified as a PAR3 interactor (Fig. 3c and Supplementary Data 3). Immunofluorescence analysis confirmed partial co-localization of YAP and PAR3 near BC regions (Fig. 3d), supporting earlier reports of functional interactions between YAP and PAR3 in a cell line⁵⁰.

Fig. 3 — a Schematic of PAR3 Co-IP assays in E17.5 WT and *Pard3*-KO hepatocytes. b Selected GO terms associated with PAR3 Co-IP–identified proteins. Numbers in parentheses indicate protein counts. c Western blotting analysis of PAR3 Co-IP eluates from E17.5 WT and *Pard3*-KO hepatocytes. At least 2 biological replicates. d YAP and PAR3 staining in E13.5 and E17.5 WT and *Yap*-KO livers. Scale bars, 10 μm. e 3D reconstruction of ZO1 signals in E15.5 and E17.5 WT and *Yap*-KO livers. Scale bars, 50 μm. f Quantification of 3D BC length in (e). g HNF4A, ZO1, and MDR1 staining in E15.5 and E17.5 WT and *Yap*-KO livers. Scale bars, 50 μm. h Quantification of BC fragment length in (g). i PCA of E17.5 WT and *Yap*-KO hepatocytes. Arrow, the developmental trajectory. Cell numbers are shown in parentheses. j Pseudotime distribution of hepatocytes in (i). Boxplots display median lines, 25th–75th percentiles (box), whiskers extending 1.5 times the interquartile range from the edges, and minimum and maximum values at whisker ends. k Volcano plot showing DEGs in E17.5 WT and *Yap*-KO hepatocytes. Numbers indicate significantly upregulated genes. l Selected GO terms for DEGs in (k). m HNF4A, Ki67, and MDR1 staining in E17.5 WT and *Yap*-KO livers. Scale bars, 10 μm. n Quantification of Ki67⁺ hepatocytes in (m). o Heatmap of differentially accessible chromatin regions in E17.5 WT and *Yap*-KO hepatocytes. Numbers in parentheses indicate the number of regions. p Enriched transcription factor–binding motifs for regions in (o). q UMAP of E17.5 WT, *Pard3*-KO, and *Yap*-KO hepatocytes. r Venn diagram showing highly differentially expressed genes in E17.5 WT, *Pard3*-KO, and *Yap*-KO hepatocytes. s Selected GO terms for overlapped DEGs in (r). Data (f, h, n) represent mean ± s.e.m.; N, the number of mice. A two-tailed unpaired Student’s t-test (f, h, n), and a two-tailed Wilcoxon rank-sum test (j). Cells of each genotype (i, j) were derived from 2 mice (refer to Supplementary Data 1). Source data (c, f, h, i, j, n) are provided as a Source Data file.

To investigate YAP’s role in BC formation and hepatocyte development, we generated a conditional Yap-KO mouse model using Foxa3-Cre (Supplementary Fig. 6d). Similar to Pard3-KO, Yap-KO did not affect hepatoblast polarity or primordial BC formation at E13.5 (Supplementary Fig. 6e-g). However, Yap-KO livers exhibited significantly shorter BCs at E15.5 and E17.5 compared to WT controls (Fig. 3e-h). scRNA-seq of DLK⁺ hepatocytes revealed delayed hepatocyte maturation in Yap-KO mice at E17.5, despite normal developmental states at E15.5 (Fig. 3i,j and Supplementary Fig. 6h-j). DGE and GO analysis showed increased expression of 717 genes related to cell proliferation and MAPK signaling in Yap-KO hepatocytes, whereas WT hepatocytes exhibited higher expression of 103 genes associated with glucose and lipid metabolism, drug detoxification, and the urea cycle (Fig. 3k,l and Supplementary Fig. 6k). Consistent with findings in Pard3-KO livers, Yap-KO livers also exhibited elevated hepatocyte proliferation, as evidenced by increased Ki67 staining (Fig. 3m,n). ATAC-seq analysis identified 1,898 regions with increased chromatin accessibility in Yap-KO hepatocytes, significantly more than the 269 regions observed in WT cells (Fig. 3o). Genes associated with these regions showed increased expression in Yap-KO hepatocytes (Supplementary Fig. 6l), and motif enrichment analysis revealed TFs linked to the MAPK pathway and cell proliferation, including FRA1/FOSL2, JUN, and AP1 (Fig. 3p).

Comparative scRNA-seq analysis of E17.5 WT, Pard3-KO, and Yap-KO hepatocytes showed that Pard3-KO and Yap-KO cells clustered together, distinct from WT cells on the UMAP plot (Fig. 3q). A Venn diagram of upregulated genes revealed that 49.8% of the upregulated genes were shared between Pard3-KO and Yap-KO hepatocytes (Fig. 3r,s). Furthermore, the double KO (Pard3/Yap-KO) model exhibited phenotypes similar to those of individual KOs, including shorter BCs, delayed hepatocyte maturation, and increased proliferation at E17.5 (Supplementary Fig. 7a-j).

Co-IP assays using a ZO1 antibody confirmed interactions among ZO1, PAR3, and YAP (Supplementary Fig. 7k). However, despite their interaction, the absence of either PAR3 or YAP did not significantly alter the localization of the other near BC regions (Supplementary Fig. 7l-o).

Collectively, these findings indicate that PAR3 and YAP may function within a shared regulatory pathway, with their co-localization at TJs playing a pivotal role in BC elongation and hepatocyte development.

PAR3 regulates BC elongation and hepatocyte development by suppressing ERK activation

scRNA-seq and ATAC-seq analyses revealed activation of the MAPK pathway in E17.5 Pard3-KO hepatocytes (Fig. 2k,q), accompanied by elevated phosphorylated ERK1/2 (p-ERK1/2) levels (Fig. 4a,b). During normal liver development, p-ERK1/2 levels progressively decreased from E13.5 to E17.5 (Supplementary Fig. 8a,b), indicating a physiological downregulation of ERK/MAPK signaling during hepatocyte maturation.

To assess the role of ERK/MAPK signaling in hepatocyte development and BC elongation, we cultured E15.5 mouse liver lobes ex vivo (Fig. 4c and Supplementary Fig. 8c). After two days, Pard3-KO liver lobes displayed significantly shorter BCs than WT (Fig. 4d,e). Activation of PKC by TPPB⁵¹ phenocopied this defect, whereas MEK inhibition with PD0325901 or U0126 restored BC length in Pard3-KO lobes to WT levels and further enhanced BC elongation in WT (Fig. 4d,e and Supplementary Fig. 8d,e). These findings indicate that PAR3 promotes BC elongation by restraining ERK/MAPK activity.

To evaluate how ERK/MAPK activity influences hepatocyte maturation, we performed scRNA-seq on DLK⁺ hepatocytes isolated from treated liver lobes (Fig. 4c and Supplementary Fig. 8f). PCA revealed delayed developmental progression in TPPB-treated WT and Pard3-KO hepatocytes, whereas PD0325901 and U0126 treatments accelerated maturation (Fig. 4f,g). Complementary ATAC-seq analysis identified two major clusters of differentially accessible chromatin regions (Fig. 4h). Cluster 1, comprising regions with increased accessibility in TPPB-treated WT or Pard3-KO hepatocytes, were enriched for MAPK-responsive transcription factors (FRA1/FOSL2, JUN, and AP1). In contrast, cluster 2, enriched in PD0325901-treated hepatocytes, contained motifs for key regulators of hepatocyte differentiation and function, including HNF4A, FOXA2, HNF6, CEBPB, and FOXA1 (Fig. 4i). These data demonstrate that ERK/MAPK inhibition promotes hepatocyte maturation at both transcriptional and chromatin-accessibility levels.

To further define the role of the PAR3–ERK/MAPK axis in BC elongation, we used Can 10 hepatocytes. Pard3 knockdown (Pard3-KD) by siRNA increased p-ERK1/2 levels and markedly shortened BCs compared with controls (Supplementary Fig. 8g-j). Treatment with U0126 restored BC length in Pard3-KD cells and further extended BCs in control cells (Supplementary Fig. 8i,j), confirming that PAR3 regulates BC elongation by suppressing ERK/MAPK signaling—consistent with our in vivo observations.

We next tested whether ERK/MAPK activation contributes to the Yap KO phenotype. p-ERK1/2 levels were elevated in E17.5 Yap-KO hepatocytes (Supplementary Fig. 9a,b), and PD0325901 treatment of E15.5 Yap-KO liver lobes restored BC length to WT levels (Supplementary Fig. 9c-f). scRNA-seq analysis further showed that TPPB treatment delayed hepatocyte maturation, whereas PD0325901 treatment accelerated it (Supplementary Fig. 9g,h).

Together, these findings demonstrate that both PAR3 and YAP regulate BC elongation and hepatocyte maturation by restraining ERK/MAPK activity. This shared regulatory mechanism integrates polarity, proliferation, and differentiation programs essential for coordinated liver morphogenesis and metabolic maturation.

PAR3 drives BC elongation via IQGAP1-ERK signaling and actin coordination

PAR3 Co-IP/MS analysis identified IQGAP1 as a PAR3-interacting protein (Fig. 3c and Supplementary Data 3). IQGAP1 functions as a scaffold linking activated growth factor receptors at the basal membrane to RAF1, thereby promoting ERK/MAPK activation through MEK phosphorylation^52–54. During embryonic liver development (E13.5-E17.5), IQGAP1 expression gradually declined (Supplementary Fig. 10a,b). In E13.5 hepatoblasts, IQGAP1 was primarily localized near the cell membrane, with partial enrichment at ZO1-marked BC regions (Fig. 5a). This pattern persisted in E17.5 WT hepatocytes but shifted in Pard3-KO livers, where IQGAP1 was reduced at BCs and accumulated at the β-catenin–marked basal membrane (Fig. 5a-c). Consistently, RAF1 levels were increased at the basal membrane in Pard3-KO hepatocytes (Fig. 5d-f).

Fig. 5 — a IQGAP1, ZO1, and β-catenin staining in E13.5 and E17.5 WT and *Pard3*-KO livers. Scale bars, 50 μm. b Quantification of IQGAP1 intensity near apical and basal membranes in E17.5 WT and *Pard3*-KO hepatocytes in (a). c 3D reconstruction of IQGAP1, ZO1, and β-catenin signals in E17.5 WT and *Pard3*-KO livers. Scale bars, 10 μm. d RAF1, ZO1, and β-catenin staining in E13.5 and E17.5 WT and *Pard3*-KO livers. Scale bars, 50 μm. e Quantification of RAF1 intensity near the basal membrane in E17.5 WT and *Pard3*-KO hepatocytes in (d). f 3D reconstruction of RAF1, ZO1, and β-catenin signals in E17.5 WT and *Pard3*-KO livers. Scale bars, 10 μm. g Volcano plot showing differentially enriched proteins from IQGAP1 Co-IP assays in E17.5 WT and *Pard3*-KO hepatocytes. Numbers indicate significantly enriched proteins. h Selected GO terms associated with proteins in (g). Western blotting analysis (i) and quantification (j) of p-ERK1/2 in control, *Pard3*-KD, *Iqgap1*-KD, and *Pard3*/*Iqgap1*-double KD Can 10 cells. k HNF4A, ZO1, and MDR1 staining in 2-day cultured E15.5 liver lobes. Scale bars, 50 μm. l Quantification of BC fragment length in (k). m Selected GO terms associated with ERK1/2 Co-IP–identified proteins. Numbers in parentheses indicate protein counts. n AURKB and MDR1 staining in E17.5 WT and *Pard3*-KO livers. Yellow arrowheads, AURKB signals. Scale bars, 10 μm. o Quantification of AURKB localization at BCs in (n). p, r AURKB and MDR1 staining in 2-day cultured liver lobes. Yellow arrowheads, AURKB signals. Scale bars, 10 μm. q, s Quantification of AURKB localization at BCs in (p) or (r). Data (b, e, j, l, o, q, s) represent mean ± s.e.m.; N, the number of mice (b, e, l, o, q, s) or independent batches of cell culture (j). A two-tailed unpaired (b, e, o) or paired (j) Student’s t-test, and a one-way ANOVA with Tukey’s HSD (l) or Dunnett’s (q, s) test. Source data (b, e, i, j, l, o, q, s) are provided as a Source Data file.

IQGAP1 Co-IP/MS analyses of E17.5 WT and Pard3-KO hepatocytes revealed distinct interaction profiles, reflecting divergent regulatory networks (Fig. 5g and Supplementary Fig. 10c). DGE and GO analysis showed that 75 proteins enriched in WT hepatocytes were associated with epithelial polarity, development, and cell–cell adhesion, whereas 160 proteins enriched in Pard3-KO hepatocytes were linked to MAPK signaling and growth factor responses (Fig. 5g,h). Together, these findings support a model in which PAR3 forms a protein complex that localizes a subset of IQGAP1 to TJs, thereby limiting its availability at the basolateral membrane and restraining ERK/MAPK pathway activation.

We further examined IQGAP1’s role in the PAR3–ERK/MAPK axis using the Can 10 cells. In WT cells, IQGAP1 was predominantly localized at the BC region, whereas Pard3-KD caused its translocation from the apical to the basal membrane (Supplementary Fig. 10d,e), consistent with the in vivo findings (Fig. 5a-c). Notably, simultaneous KD of Pard3 and Iqgap1 markedly reduced p-ERK1/2 levels, restoring them to control levels compared with Pard3-KD alone (Fig. 5i,j). These results indicate that PAR3 regulates ERK/MAPK signaling by controlling IQGAP1 localization, thereby preventing its association with basal components that activate the pathway.

Despite the reduction in p-ERK1/2 levels, BC length in Pard3/Iqgap1 double KD cells remained similar to that in Pard3-KD cells, suggesting that IQGAP1 also contributes to BC elongation through ERK-independent mechanisms. This aligns with IQGAP1’s function as a multifunctional scaffold that interacts with small GTPases (Cdc42, Rac1, Rab27a), actin filaments, and components of adherens and TJs^27–29. These partners likely coordinate actin organization and junctional stability—processes essential for BC formation and elongation.

To further validate this mechanism, we introduced a GFP–PAR3 siRNA-resistant construct into Pard3-KD Can 10 cells via lentiviral transduction. Western blotting confirmed an approximately 2.8-fold increase in PAR3 expression (Supplementary Fig. 10f,g). Immunofluorescence staining of IQGAP1, ZO1, and F-actin, together with p-ERK1/2 analysis, showed that PAR3 overexpression fully rescued the Pard3-KD phenotype, restoring elongated BCs, re-establishing apical IQGAP1 localization, and reducing p-ERK1/2 levels (Supplementary Fig. 10f-j). These findings support that PAR3 suppresses ERK/MAPK activation by localizing a subset of IQGAP1 to TJs.

Immunofluorescence analysis of E17.5 Pard3-KO hepatocytes revealed a fragmented or punctate distribution of F-actin, although it remained primarily near TJs (Supplementary Fig. 10k). To evaluate the role of the actin cytoskeleton in BC development, E15.5 liver lobes were treated with the actomyosin inhibitor Y27632 or the F-actin polymerization inhibitor Cytochalasin D⁵⁵. Both treatments markedly shortened BCs after two days of culture (Fig. 5k,l and Supplementary Fig. 10l). Notably, co-treatment with the MEK inhibitor PD0325901 failed to restore BC length (Fig. 5k,l), and neither Y27632 nor Cytochalasin D significantly altered p-ERK1/2 levels (Supplementary Fig. 10m,n). These results indicate that MEK inhibition cannot rescue BC elongation without an intact actomyosin system.

Further evidence for ERK involvement in cytoskeletal regulation came from Co-IP/MS analysis, which identified 927 potential ERK1/2-interacting proteins (Supplementary Fig. 10o and Supplementary Data 3). GO analysis revealed that many of these proteins are linked to actin cytoskeleton organization (Fig. 5m). Consistently, TPPB-treated hepatocytes displayed fragmented or punctate F-actin structures (Supplementary Fig. 10p), suggesting that ERK modulates the cytoskeleton during BC development.

In addition to cytoskeletal regulation, oriented cell division also contributes to BC elongation. Immunofluorescence analysis showed reduced AURKB localization at BCs in E17.5 Pard3-KO or Yap-KO livers compared to WT (Fig. 5n,o and Supplementary Fig. 10q,r), indicating impaired division orientation²⁵. Ex vivo treatments with TPPB, Y27632, or Cytochalasin D further diminished AURKB localization signals, whereas PD0325901 enhanced them (Fig. 5p-s). Together, these findings define a functional link among PAR3, ERK/MAPK signaling, the actin cytoskeleton, and oriented cell division—coordinately regulating BC elongation and hepatocyte maturation.

ERK/MAPK regulates human BC elongation and hepatocyte development

To assess whether the ERK/MAPK pathway similarly governs human hepatocyte development and BC elongation, we cultured W7–8 human embryonic liver lobes for 2 days (Fig. 6a,b). ZO1 immunofluorescence showed that TPPB treatment shortened developing BCs, whereas PD0325901 markedly increased BC length (Fig. 6c,d).

Fig. 6 — a Schematic of ex vivo liver lobe culture from W7–8 human embryos. b Representative images of 2-day cultured W7–8 liver lobes. Scale bars, 1 mm. c HNF4A, ZO1, and MDR1 staining in 2-day cultured liver lobes. Scale bars, 50 μm. d Quantification of BC fragment length in (c). Data represent mean ± s.e.m.; N, the number of human embryos. A one-way ANOVA with Tukey’s HSD test. e PCA of W6–W19 human hepatoblasts and hepatocytes². Arrow, the developmental trajectory. f PCA of 2-day cultured cells. Cell numbers are shown in parentheses. g Pseudotime distribution of cells along the trajectory in (f). Boxplots display median lines, 25th–75th percentiles (box), whiskers extending 1.5 times the interquartile range from the edges, and minimum and maximum values at whisker ends. A two-tailed Wilcoxon rank-sum test. h Expression of hepatocyte maturation-associated genes in 2-day cultured cells. Violin plots display median lines, and minimum and maximum values. A two-tailed Wilcoxon signed-rank test. Cells of each treatment (f–h) were derived from 3 human embryos (refer to Supplementary Data 4). Source data (d, f–h) are provided as a Source Data file.

To further investigate the impact of ERK/MAPK modulation on human hepatocyte maturation, we isolated EpCAM⁺ hepatocytes using FACS and performed scRNA-seq analysis (Supplementary Fig. 10s and Supplementary Data 4). When projected onto the PCA plot of the human hepatoblast-to-hepatocyte developmental trajectory², TPPB-treated cells exhibited delayed maturation, whereas PD0325901-treated cells showed accelerated development (Fig. 6e-h).

Together, these findings demonstrate that ERK/MAPK signaling regulates both BC elongation and hepatocyte maturation in human liver development, revealing a conserved mechanism that coordinates structural and functional maturation across species.

PAR3 regulates BC architecture during liver regeneration

To investigate the role of PAR3 in maintaining hepatocyte polarity and BC architecture, we generated a conditional Pard3-KO mouse model by administering AAV8-Tbg-Cre to adult Pard3-f/f mice for four weeks (Fig. 7a). RT-qPCR confirmed efficient Pard3 deletion in hepatocytes (Fig. 7b). In quiescent adult livers, both the liver-to-body weight ratio and BC morphology were comparable between WT and Pard3-KO mice, indicating that PAR3 is dispensable for maintaining BC structure and hepatocyte polarity under homeostatic conditions (Fig. 7c-g).

Fig. 7 — a Schematic of the *Pard3* conditional KO strategy using AAV8-*Tbg-Cre* delivery in adult mice and the 70% hepatectomy procedure performed in WT and *Pard3*-KO animals. b RT–qPCR confirming *Pard3* deletion in hepatocytes. c Gross liver morphology in WT and *Pard3*-KO mice at baseline (day 0) and indicated time points after hepatectomy. Scale bars, 1 cm. d Quantification of liver-to-body-weight ratio in WT and *Pard3*-KO mice at baseline and post-hepatectomy time points. e HNF4A, ZO1, and MDR1 staining in WT and *Pard3*-KO livers at baseline and indicated post-hepatectomy time points. Scale bars, 50 μm. f Quantification of BC fragment length in (e). g, 3D reconstruction of ZO1 signals in WT and *Pard3*-KO livers at baseline and indicated post-hepatectomy time points. Scale bars, 50 μm. h 3D UMAP of WT and *Pard3*-KO hepatocytes at baseline, 2 days, and 7 days after hepatectomy. Cell numbers are shown in parentheses. Cells of each genotype or treatment were derived from at least 2 mice (refer to Supplementary Data 1). i Volcano plots showing DEGs in WT and *Pard3*-KO hepatocytes at baseline, 2 days, and 7 days post-hepatectomy. Numbers indicate significantly upregulated genes. j Selected GO terms for genes in (i). k Serum ALT and AST levels in WT and *Pard3*-KO mice at baseline, 2 days, and 7 days post-hepatectomy. l HNF4A, Ki67, and MDR1 staining in WT and *Pard3*-KO livers at baseline, 2 days, and 7 days post-hepatectomy. Scale bars, 10 μm. m Quantification of Ki67⁺ hepatocytes in (l). n, o Western blotting analysis (n) and quantification (o) of p-ERK1/2 levels in WT and *Pard3*-KO hepatocytes 7 days post-hepatectomy. p Model depicting the PAR3–ERK/MAPK signaling axis in regulating BC elongation and hepatocyte maturation. GFR, growth factor receptor. Created in BioRender. Xu, X. (2026) https://BioRender.com/luccdzh. Data (b, d, f, k, m, o) represent mean ± s.e.m.; N, the number of mice. A two-tailed paired (b) or unpaired (d, f, k, m, o) Student’s t-test. n.s., not significant. Source data (b, d, f, h, k, m, n, o) are provided as a Source Data file.

Given the dynamic nature of liver regeneration, we hypothesized that PAR3 might play a more critical role during this process, which involves active hepatocyte proliferation. To test this, we performed 70% hepatectomy on WT and Pard3-KO mice (Fig. 7a). At two days post-surgery, liver-to-body weight ratios and BC structures were comparable between the two groups (Fig. 7c-g). By day 7, however, Pard3-KO mice exhibited a ~ 25% higher liver-to-body weight ratio than WT, a difference that persisted through week 6 (Fig. 7c,d). Although BC networks remained present, their average length was significantly reduced in Pard3-KO livers (Fig. 7e-g). Despite this shortening, Pard3-KO mice did not develop jaundice at day 7, likely due to residual BC function supporting bile transport and metabolic activity (Fig. 7c,e-j).

To elucidate the transcriptional dynamics underlying these phenotypic differences, we performed scRNA-seq of hepatocytes during liver regeneration. At day 2 post-hepatectomy, WT hepatocytes formed distinct clusters from untreated cells on a three-dimensional (3D) UMAP plot, reflecting an active regenerative state, and by day 7 had largely reverted to a baseline profile (Fig. 7h), consistent with previous reports⁵⁶. In contrast, untreated WT and Pard3-KO hepatocytes clustered together, whereas at days 2 and 7 post-hepatectomy, they formed distinct groups (Fig. 7h). DGE analysis revealed minimal differences under basal conditions, but marked alterations during regeneration (Fig. 7i). GO analysis showed that WT hepatocytes upregulated pathways associated with lipid and bile acid metabolism and detoxification, whereas Pard3-KO hepatocytes upregulated genes linked to the inflammatory response, as well as steroid and cholesterol metabolism—processes known to regulate the cell cycle (Fig. 7j)^57–59. By day 7, Pard3-KO hepatocytes also showed increased expression of genes related to cytokine-mediated signaling and cell proliferation, consistent with increased hepatocyte turnover (Fig. 7j). In addition, serum levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were significantly higher in Pard3-KO mice compared to WT, indicating compromised hepatocyte function following Pard3 deletion (Fig. 7k).

Immunofluorescence further confirmed enhanced hepatocyte proliferation and elevated p-ERK1/2 levels in Pard3-KO livers at day 7 post-hepatectomy (Fig. 7l-o). Notably, IQGAP1 protein levels were barely detectable in both genotypes at this stage (Supplementary Fig. 10t). Together, these results demonstrate that the PAR3–ERK/MAPK signaling axis orchestrates hepatocyte proliferation and BC remodeling during liver regeneration. However, hepatocytes appear to engage distinct intermediary mechanisms to couple PAR3 and ERK/MAPK signaling during regeneration compared with embryonic development—a possibility warranting further investigation.

Discussion

This study delineates the mechanisms that coordinate cell proliferation, maturation, and structural organization during liver development and regeneration. By integrating scRNA-seq, immunofluorescence, biochemical assays, and genetic manipulation in mouse and human liver models, we delineate the temporal progression of hepatic polarization and BC biogenesis. Hepatic polarity is established immediately after hepatoblast specialization, with primordial BCs forming at cytokinetic sites between adjacent hepatoblasts. As differentiation proceeds, these BCs elongate and expand through oriented cell division, coupling morphogenesis with functional maturation. Together, these findings highlight the dynamic integration of cell division, polarity establishment, and hepatocyte differentiation in shaping liver architecture.

Our data identify PAR3 as a key regulator of BC elongation and hepatocyte maturation. PAR3 assembles a TJ–associated protein complex that includes ZO1, YAP, CDC42, IQGAP1, RDX, and EZR—factors collectively implicated in epithelial polarity, actin organization, and growth control. Rather than acting as a binary determinant, this PAR3-centered network contributes to the spatial modulation of ERK/MAPK signaling, thereby supporting actin cytoskeletal organization, appropriate spindle orientation, and efficient BC elongation during hepatocyte differentiation. In the absence of PAR3, IQGAP1 exhibits altered subcellular distribution toward basolateral domains, accompanied by elevated ERK/MAPK activity. Increased pathway activity correlates with disrupted actin organization, shortened BCs, enhanced proliferation, and delayed hepatocyte maturation, as reflected by activation of FOS-, JUN-, and AP1-associated transcriptional programs (Fig. 7p). These findings position PAR3 as a coordinator of structural and functional maturation during liver development.

Importantly, our results indicate that PAR3 is dispensable for the initial establishment of apical–basal polarity and primordial BC formation but is required for subsequent BC elongation. This distinction underscores the modular organization of hepatic polarization mechanisms and highlights a specific role for PAR3 in refining canalicular architecture during the hepatoblast-to-hepatocyte transition. Consistent with this view, PAR3 regulates the extent and rate of BC elongation rather than serving as the sole determinant of canalicular initiation.

BC formation is likely driven by cell division–dependent mechanisms during the early stages of liver development, when hepatoblasts are more proliferative, whereas BC network formation during later stages is more likely driven by cell division–independent mechanisms, as hepatocytes become less proliferative. At E17.5, approximately half of dividing hepatocytes display mitotic spindles positioned near elongated BCs and aligned parallel to the canalicular axis, a configuration likely to facilitate BC elongation²⁵. In contrast, divisions oriented obliquely or perpendicular to BCs may lead to asymmetric inheritance of canalicular membranes by daughter cells, thereby promoting network complexity and branching^15,19,20,25. These observations suggest that hepatocytes deploy distinct modes of cell division to balance canalicular elongation and network elaboration during development.

In contrast to liver development, where BC architecture forms de novo, liver regeneration involves reactivation of hepatocyte proliferation within a pre-existing canalicular network⁵⁵. Our data indicate that PAR3–ERK/MAPK signaling performs analogous functional roles in coordinating BC remodeling in both contexts, although the upstream cues, regulatory mechanisms, and cellular environments are likely to differ. By constraining ERK/MAPK activity, PAR3 supports preservation of BC architecture and facilitates hepatocyte functional recovery following injury, emphasizing its importance in maintaining liver structure and function during regeneration.

These findings carry implications for therapeutic strategies aimed at restoring liver function. Proper BC architecture and hepatocyte maturation are essential for hepatic physiology, and modulation of PAR3–ERK/MAPK signaling may represent a strategy to enhance both processes in diseased or injured livers. Moreover, these insights offer a framework for improving hepatocyte-based regenerative approaches. Current differentiation protocols from pluripotent stem cells frequently yield immature hepatocyte-like cells with limited polarity and canalicular organization; targeted modulation of PAR3–ERK/MAPK signaling could promote functional maturation and organized BC formation, advancing applications in disease modeling, drug testing, and cell therapy.

While IQGAP1 emerges as an important mediator linking PAR3 to ERK/MAPK signaling, further work will be required to define the precise molecular interactions within this network. In particular, many PAR3-associated factors identified by proteomic analyses remain functionally uncharacterized. Future work combining in vitro approaches—such as real-time imaging, gene editing, and biochemical assays in hepatocyte models—with in vivo validation will be essential for dissecting the dynamics and physiological relevance of these pathways during liver development and regeneration.

In summary, this study establishes a framework connecting cell division, polarity regulation, and BC biogenesis during liver development. PAR3-dependent spatial modulation of ERK/MAPK signaling, together with oriented cell division, contributes to BC elongation and hepatocyte maturation. These insights refine our understanding of hepatic morphogenesis and provide a conceptual basis for strategies aimed at generating mature, polarized hepatocyte-like cells for regenerative medicine and liver disease therapy.

Methods

Mice

All animal experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) of Peking University Health Science Center. Mice were housed under specific pathogen-free (SPF) conditions, maintained at a temperature of 23 ± 2 °C, with a 12-h day/night cycle. The mouse strain used in this study including B6C3F1 (the F1 hybrid of female C57BL/6 and male C3H), Foxa3-Cre⁶⁰, Alfp-Cre⁴⁶, Pard3-f (Nanjing Biomedical Research Institute of Nanjing University, T002765), and Yap-f ⁶¹ mice. Female mice aged 8 to 12 weeks were mated with adult male mice, and the morning on which the vaginal plug was detected was designated E0.5.

Human samples

All protocols involving human samples were approved by the Peking University Institutional Review Board (PU-IRB) (certificate number: IRB00001052-18083). Written informed consent was obtained from all donors prior to sample collection. Human embryos were obtained from elective pregnancy terminations at Haidian Maternal & Child Health Hospital in Beijing.

Cell lines

Can 10 cells are derived from Fao, a subclone of the male rat hepatoma cell line H4IIEC3. Can 10 cells were obtained from Dr. Doris Cassio²⁶.

Immunofluorescence staining

Tissues and cultured explants were fixed in 4% paraformaldehyde overnight at 4 °C. For paraffin sections, samples were dehydrated using an ethanol series, cleared with xylene, embedded in paraffin, and sectioned into 7-10 µm thick sections. After rehydration, antigen retrieval was performed by autoclaving sections in citrate buffer (10 mM citrate, 0.05% Tween-20, pH 6.0) for 10 min. For frozen sections, samples were incubated in 30% sucrose at 4 °C overnight, embedded in the optimum cutting temperature (OCT) compound, and sectioned at 10-50 µm.

Primary antibodies included: anti-HNF4A (Santa Cruz Biotechnology, sc-6556, 1:50; LifeSpan BioSciences, LS-C758303, 1:50; Cell Signaling Technology, 3113, 1:200), anti-ZO1 (Invitrogen, 61-7300, 1:200; Santa Cruz Biotechnology, sc-8147, 1:50), anti-MDR1 (Santa Cruz Biotechnology, sc-55510, 1:100; Santa Cruz Biotechnology, sc-8313, 1:50), anti-PAR3 (Abcam, ab64646, 1:100), anti-SOX9 (Sigma-Aldrich, ab5535, 1:200), anti-AURKB (Santa Cruz Biotechnology, sc-25426, 1:200), anti-TUBA (Sigma-Aldrich, T6199, 1:100), anti-Ki67 (Thermo Fisher Scientific, 14-5698-80, 1:200), anti-YAP (Cell Signaling Technology, 12395, 1:100), anti-F-actin (Phalloidin CruzFluor™ 594 Conjugate) (Santa Cruz Biotechnology, sc-363795), anti-IQGAP1 (Proteintech, 22167-1-AP, 1:100), anti-RAF1 (Proteintech, 26863-1-AP, 1:100), anti-β-catenin (Cell Signaling Technology, 37447, 1:100), anti-RDX (Proteintech, 13790-1-AP, 1:100), and anti-EZR (Proteintech, 26056-1-AP, 1:100). The following secondary antibodies were applied: Alexa Fluor 488 donkey anti-goat (Thermo Fisher Scientific, A-11055, 1:1000), Alexa Fluor 594 donkey anti-rabbit (Thermo Fisher Scientific, A-21207, 1:1000), Alexa Fluor 647 donkey anti-mouse (Thermo Fisher Scientific, A-31571, 1:1000), PE goat anti-rat (BioLegend, 405406, 1:1000), and Alexa Fluor 647 donkey anti-rabbit (Jackson ImmunoResearch, 711-607-003, 1:1000). Nuclei were counterstained with DAPI (Sigma, D9564, 0.5 µg/mL).

Confocal imaging was performed using a Leica TCS SP8 microscope. Z-stack images were acquired and projected onto the x–y plane using the maximum-intensity method for quantification. To quantify hepatoblast or hepatocyte polarity and BC structures during embryonic development, sections were stained with HNF4A, ZO1, and MDR1 antibodies. HNF4A was used to identify hepatoblasts and hepatocytes, while ZO1 and MDR1 marked cell polarity and BCs. HNF4A⁺ cells showing co-localized ZO1 and MDR1 signals on the plasma membrane were defined as polarized; those without such co-localization were classified as non-polarized. Structures showing ZO1 and MDR1 co-localization between two adjacent HNF4A⁺ cells were defined as primordial BCs, whereas those spanning more than two cells were defined as elongated BCs. To quantify hepatoblasts forming polarity during division and hepatocytes elongating BCs through oriented division, sections were stained for AURKB, ZO1, or MDR1. AURKB was used to define spindle position and orientation during late cytokinesis. AURKB⁺ hepatoblasts with signals overlapping ZO1 or MDR1 between adjacent cells were defined as forming polarity through cell division, whereas AURKB⁺ hepatocytes with AURKB signals located at or near ZO1- or MDR1-marked BCs were defined as elongating BCs through oriented division. BC length was determined by measuring the linear ZO1 signal using ZEN imaging software. To assess hepatoblast or hepatocyte proliferation, sections were stained with Ki67 and HNF4A antibodies, and Ki67⁺HNF4A⁺ cells were defined as proliferative. For quantification of IQGAP1 and RAF1 signal intensity, sections were stained with IQGAP1, RAF1, ZO1, and β-catenin antibodies, with ZO1 and β-catenin marking the apical and basal membranes, respectively. The intensity of IQGAP1 signals co-localized with ZO1 or β-catenin and RAF1 signals co-localized with β-catenin was quantified using LAS X software (Source data).

For three-dimensional reconstruction of BC structures⁴⁸, 50 µm frozen sections were stained with ZO1 antibody. For reconstruction of IQGAP1 and RAF1 localization, 20 µm frozen sections were stained with IQGAP1, RAF1, ZO1, and β-catenin antibodies. Images were acquired along the z-axis at 0.5 µm intervals, and 3D reconstructions were generated using LAS X imaging software. 3D filament-based measurements were performed at embryonic day (E) 15.5 and E17.5 using the Imaris Filament module (Imaris v10.2; Bitplane). Briefly, reconstructed 3D image stacks were subjected to automated filament tracing, canaliculi were traced end-to-end, and bile canalicular length was extracted for each individual segment.

For Can 10 cell staining, cells were cultured on 18-mm collagen-coated coverslips in Ham’s F-12K (Kaighn’s) medium (Thermo Fisher Scientific, 21127022) supplemented with 1% fetal bovine serum. For PAR3 staining, cells were fixed in 100% methanol for 2 min at 4 °C, followed by 1.8% formaldehyde (Santa Cruz Biotechnology, sc-203049A) for 10 min, washed twice with PBS (135 mM NaCl, 1.3 mM KCl, 3.2 mM Na₂HPO₄, 0.5 mM KH₂PO₄, pH 7.4), and blocked for 30 min with 3% BSA in PBS. For other antibodies, cells were fixed with 1.8% formaldehyde for 10 min, washed twice with PBS, and permeabilized in PBS containing 0.1% Triton X-100 and 3% BSA. Primary antibodies used in Can 10 cells included anti-PAR3 (Sigma-Aldrich, 07-330, 1:1000), anti-aPKC (Santa Cruz, sc-17781, 1:250), anti-IQGAP1 (Santa Cruz, sc-10792, 1:1000), anti-ZO1 (Santa Cruz, sc-33725, 1:400), and Alexa Fluor 647 phalloidin (Thermo Fisher Scientific, A-22287, 1:200). Confocal images were acquired using a Yokogawa CSU-X1 spinning-disk scanner mounted on a Nikon Ti2-U microscope equipped with a CFI Apo TIRF 100×/1.49 oil objective, Hamamatsu ORCA-Quest qCMOS camera (C15550-20UP), and a Stradus VersaLase four-wavelength laser system (Vortran Laser Technology). Super-resolution images were captured on a VisiTech VT-iSIM system (Olympus IX71 inverted microscope with UPLSAPO 60×/NA 1.2 water objective and Hamamatsu ORCA-Quest qCMOS camera) and deconvolved using the Microvolution plugin in ImageJ.

Transmission electron microscopy

Liver tissues were rinsed with PBS, cut into ~1 mm³ cubes, and fixed in a fixation buffer containing 2.5% glutaraldehyde and 2% freshly prepared paraformaldehyde in 0.1 M PBS. The samples were then dehydrated through a graded ethanol series, embedded in Epon resin, and sectioned into ultrathin sections. Images were performed on a FEI Tecnai G2 Spirit transmission electron microscope (Thermo Fisher Scientific).

Cell preparation and FACS sorting

Embryonic mouse liver tissues, and cultured explants were digested with 0.25% Trypsin-EDTA at 37 °C, with digestion terminated by adding FBS. Cells were labeled with anti-mouse DLK-FITC (MBL International, D187-4, 1:100) or anti-human EpCAM-APC (BioLegend, 324208, 1:100) antibodies. P14 and adult mice were anesthetized, and their livers were perfused through the inferior vena cava with perfusion buffer (PBS with 1 mM EGTA) followed by digestion buffer [RPMI 1640 with 1 mg/mL collagenase IV (Sigma-Aldrich, C5138)]. All samples were filtered through a 70 µm cell strainer and sorted using a BD FACSAria Fusion cell sorter (BD Bioscience). Dead cells were excluded using DAPI staining.

RT-qPCR

Total RNA was extracted using the RNAprep Pure Micro Kit (Tiangen, DP420). Reverse transcription was performed using the HiScript II Q RT SuperMix (Vazyme, R223-1). RT-qPCR was conducted with 2× M5 HiPer SYBR Premix EsTaq (with Tli RNase H) (Mei5 Biotechnology, MF787-01) on a LightCycler® 480 Instrument II (Roche). The specific primers (Sangon Biotech Co., Ltd.) used for the analysis are as follows: Pard3-F: 5’-CCAGTTCCAGCGAGACAACG-3’; Pard3-R: 5’-CGGTATCAGCATGTCCAACC-3’; Yap-F: 5’-TTCGGCAGGCAATACGGAAT-3’; Yap-R: 5’-TGCTCCAGTGTAGGCAACTG-3’; Gapdh-F: 5’-ATGGTGAAGGTCGGTGTGAAC-3’; Gapdh-R: 5’-GCCTTGACTGTGCCGTTGAAT-3’.

Western blotting

Targeted cells were lysed in RIPA buffer [150 mM NaCl, 50 mM Tris-HCl (pH 7.4), 1% Triton X-100, 1% sodium deoxycholate and 0.1% SDS] supplemented with a protease inhibitor cocktail (Selleck Chemicals, B14002) and a phosphatase inhibitor cocktail (Beyotime, P1081). Western blotting was performed according to the standard protocol. The antibodies used included anti-PAR3 (Abcam, ab64646, 1:500), anti-IQGAP1 (Proteintech, 22167-1-AP, 1:500), anti-p-ERK1/2 (Cell Signaling Technology, 4370, 1:500), anti-ERK1/2 (Proteintech, 11257-1-AP, 1:500), anti-Histone 3 (H3) (Abcam, ab1791, 1:1000), and goat anti-rabbit IgG-HRP (Bio-Rad, 170-6515, 1:1000). Band intensities were quantified using ImageJ software.

For Can 10 cells, cells were washed with ice-cold PBS containing Ca²⁺ and Mg²⁺, and lysed in 2× protein loading buffer (Bio-Rad, 1610747) with 1× protease inhibitor cocktail (Sigma-Aldrich, P8340). Western blotting followed standard protocols, using primary antibodies anti-PAR3 (Sigma-Aldrich, 07-330, 1:1000), anti-IQGAP1 (Santa Cruz Biotechnology, sc-10792, 1:1000), anti-p-ERK1/2 (Cell Signaling Technology, 4370, 1:250), anti-ERK1/2 (Cell Signaling Technology, 4695, 1:250), and anti-β-actin (GeneTex, GTX637675, 1:2,000). Secondary antibodies included Rabbit Optimized HRP Reagent, Pico (Thermo Fisher Scientific, 1862191) and Mouse Optimized HRP Reagent, Pico (Thermo Fisher Scientific, 1862116). The uncropped scans were supplied in the Source Data file or as a Supplementary Fig. in the Supplementary Information.

Co-immunoprecipitation (Co-IP) and mass spectrometry (MS)

Targeted cells were lysed in a lysis buffer [150 mM NaCl, 25 mM Tris-HCl (pH 7.4), 1% CA-630 (Sigma-Aldrich, 56741), 1 mM EDTA and 5% glycerol] supplemented with a protease inhibitor cocktail. The lysates were incubated overnight at 4 °C with antibodies including normal rabbit IgG (Sigma-Aldrich, 12-370, 1:50), anti-PAR3 (Abcam, ab64646, 1:50), anti-ZO1 (Invitrogen, 61-7300, 1:50), anti-IQGAP1 (Proteintech, 22167-1-AP, 1:50), anti-ERK1/2 (Proteintech, 11257-1-AP, 1:50), along with pre-cleared Protein A Dynabeads (Invitrogen, 10002D). The input and Co-IP elution samples were analyzed by Western blotting. The antibodies used included anti-PAR3 (Abcam, ab64646, 1:500), anti-ZO1 (Santa Cruz Biotechnology, sc-8147, 1:200), anti-YAP (Cell Signaling Technology, 12395, 1:500), anti-IQGAP1 (Proteintech, 22167-1-AP, 1:500), anti-CDC42 (Proteintech, 67212-1-Ig, 1:500), anti-RDX (Proteintech, 13790-1-AP, 1:500), anti-EZR (Proteintech, 26056-1-AP, 1:500), goat anti-rabbit IgG-HRP (Bio-Rad, 170-6515, 1:1,000), goat anti-mouse IgG-HRP (Invitrogen, 31430, 1:1,000), and donkey anti-goat IgG-HRP (Santa Cruz Biotechnology, sc-2020, 1:1,000). For MS analysis, Co-IP elution samples were resolved using 10% SDS-PAGE, stained with Coomassie Brilliant Blue R-250, and the targeted gel bands were excised for MS analysis using an Orbitrap Exploris 480 Mass Spectrometer (Thermo Fisher Scientific).

Liver lobe culture

Liver lobe culture was conducted following the previously described protocol¹, with some modifications. Liver tissues were cut into ~1 mm³ cubes. The explants were cultured and treated with DMSO (Sigma-Aldrich, D2650), 0.5 µM TPPB (Calbiochem, 508158), 0.5 µM PD0325901 (Selleck, S1036), 10 µM U0126 (Sigma-Aldrich, U120), 1 µM Y27632 (Selleck, S1049), or 0.5 µM Cytochalasin D (Selleck, S8184).

siRNA-mediated knockdown in Can 10 cells

Pre-designed double-stranded small interfering RNAs (siRNAs) targeting rat Pard3 and Iqgap1 with the following 25-nucleotide synthetic RNA sequences (sense strand) were used: Pard3 siRNA, 5’-AGAGUUGGAUGACAGAGAACGCAGG-3’; Iqgap1 siRNA, 5’-GUCAUGAAGUUGUUUGAUAGAGCUA-3’. Negative Control DsiRNA (IDT, 51-01-14-04) served as a control. Can 10 cells were seeded at 0.9 × 10⁴/cm², transfected with 4.8 pmol of siRNA using LipofectamineTM RNAi MAX Transfection Reagent (Thermo Fisher Scientific, 13778150), and cultured for 72 h before use in further experiments.

GFP-PAR3 overexpression in Can 10 cells

For GFP–PAR3 overexpression, the full-length Pard3 coding sequence was cloned, and nine nucleotides within the siRNA targeting region were mutated without altering the amino acid sequence (5′-TGAATTAGACGATAGGGAGCGGAGA−3’; mutated bases bold). The mutated Pard3 sequence was inserted into the lentiviral vector pLJM1-GFP to generate GFP–PAR3. Lentivirus was produced by co-transfecting GFP–PAR3 with the packaging plasmids pMLDg/pRRE, pRSV-Rev, and pMD2.G into HEK293T cells using Lipofectamine 3000 (Thermo Fisher Scientific, L3000015). Twelve hours after transfection, the medium was replaced with DMEM supplemented with 10% FBS, and the lentivirus-containing supernatant was collected at 48 h post-transfection. Viral supernatants were filtered and concentrated using Lenti-X Concentrator (Takara Bio, 631231). Can 10 cells were transduced with the GFP–PAR3 siRNA-resistant lentivirus and, after 24 h, were transfected with either control DsiRNA or Pard3 siRNA, followed by 72 h of culture before analysis.

AAV virus production and purification

AAV8-Tbg-Cre AAV virus production and purification were performed as previously reported⁶². In brief, the pX602-Cre-sgRNA, pRC2/8 and PHelper plasmids were transfected into HEK-293T cells with the help of PEI (Polysciences, 23966). Approximately 60 h after transfection, cells containing virus were collected and lysed. Virus was purified through a density gradient centrifugation procedure. Approximately 1 × 10¹² viral particles were injected into mice through the retro-orbital sinus.

70% hepatectomy

The mice were anesthetized using trichloroethanol with a dosage of 1 g/kg body weight. A midline incision was made through the skin and abdominal wall to expose the liver. The median and left lateral lobes of the liver were ligated using non-absorbable sutures and excised to remove approximately 70% of the total liver mass. The abdominal wall and skin incisions were closed, and warm saline were administered to replace fluid losses. The mice were then placed in a warm environment to recover, with access to food and water.

Serum ALT and AST measurement

Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were measured using an Alanine Aminotransferase Assay Kit (Nanjing Jiancheng Technology Co., Ltd., C009-2-1) and an Aspartate Aminotransferase Assay Kit (Nanjing Jiancheng Technology Co., Ltd., C010-2-1), respectively, according to the manufacturer’s instructions.

Hepatoblast differentiation in vitro

In vitro hepatoblast differentiation was performed as previously described^2,63. Hepatoblasts were isolated by FACS from E13.5 WT and Pard3-KO embryos. For hepatocyte differentiation, single hepatoblasts were cultured on gelatin-coated dishes for 6 days; for cholangiocyte differentiation, single hepatoblasts were cultured in Matrigel for 10 days. Both systems used DMEM/F12 medium supplemented with 10% FBS, ITS-X, antibiotics, EGF, HGF, and dexamethasone.

Single-cell RNA-seq

Single-cell RNA-seq was performed following the Smart-seq3³⁹ protocol. Individual cells were sorted into a 384-well plate filled with cell lysis buffer using FACS. After mRNA reverse transcription and cDNA amplification, the amplified cDNA was used to construct sequencing libraries with TruePrep DNA Library Prep Kits (Vazyme, TD501). DNA purification and size selection were carried out using VAHTS DNA Clean Beads (Vazyme, N411-03).

ATAC-seq

ATAC-seq was following a modified version of the previously described protocol⁶⁴. A total of 1000 cells were collected in 50 µL of lysis buffer [10 mM NaCl, 10 mM Tris-HCl (pH 7.4), 3 mM MgCl₂ and 0.5% CA-630] and incubated at 4 °C for 10 minutes. The samples were then centrifuged at 500 × g for 5 min, leaving 6 µL of buffer containing nuclear pellets. DNA fragmentation and library preparation were performed using a TruePrep DNA Library Prep Kit (Vazyme, TD502). The DNA was purified using VAHTS DNA Clean Beads.

Smart-seq3 scRNA-seq data alignment and preprocessing

For Smart-seq3 scRNA-seq, index, primer, and adapter sequences from paired-end reads were trimmed using Trimmomatic (v0.39)⁶⁵. The trimmed reads were aligned to the Mus musculus reference genome (GRCm38/mm10) or the Homo sapiens reference genome (GRCh38/hg38) using Hisat2 (v2.1.0) with default settings⁶⁶, depending on the species of origin. Gene assignments were performed using featureCounts (v2.0.1)⁶⁷ with default parameters. Transcripts per 0.1 million (TP0.1 M) were calculated and log-transformed. For mouse data, to avoid transcriptome perturbations caused by highly expressed genes, the extremely highly expressed genes Apoa1, Apoa2, Alb, Ahsg, Mt1, Fgb, Afp, and Ttr were excluded when calculating total transcripts in mouse. Only cells with the detection of more than 4,000 genes (TP0.1 M > 0) were retained for further analyses.

Dimension reduction and pseudotime analysis

We used the R package Seurat (v4.4.0)⁶⁸ to analyze scRNA-seq data. First, we selected the top 2,000 highly variable genes (HVGs) using the FindVariableFeatures function. Only protein-coding genes were retained, while cell cycle-related genes and potential contamination genes (from index switching or ambient RNA contamination) were excluded. The data was then scaled with the ScaleData function, and PCA was performed using the RunPCA function on the selected HVGs. The PCA embedding matrix was used for uniform manifold approximation and projection (UMAP) analysis using the RunUMAP function. For Fig. 7h, UMAP analysis was performed with the parameter ‘n.components = 3’.

For the mouse data, we integrated each knockout genotype and its corresponding control group with the wild-type reference data using the mnnCorrect function from the R package batchelor (v1.18.1)⁶⁹. With the integrated data, PCA was conducted as before, and the first two principal components (PCs) values were utilized for principal curve analysis using the principal_curve function from the R package princurve (v2.1.6)^70,71. The principal curve index values were used as the pseudotime of hepatocytes.

For the human data, we used our published 10x Genomics data as a reference and normalized the data using the NormalizeData function in Seurat (v4.4.0), with sizefactor = 100,000, keeping other parameters as default. HVGs selection and PCA calculation were conducted using the same approach as previously described. The human cultured scRNA-seq data was embedded into the first two PCs of the reference dataset to evaluate hepatocyte maturation. We performed linear regression on the reference cells to identify the hepatocyte maturation trajectory and projected all cells onto this trajectory to calculate pseudotime.

Differential gene expression analysis

Differential gene expression (DGE) analysis was performed using the FindMarkers function in the Seurat package (v4.4.0) in R, with the parameter test.use set to ‘wilcox’. Genes with an adjusted p-value of less than 10⁻⁴ and log₂(fold-change) > 0.5 were considered significantly differentially expressed.

Gene ontology enrichment analysis

Gene ontology (GO) enrichment analysis was performed using the R package clusterProfiler (v4.10.1)⁷² to identify significantly enriched GO terms. The parameters used were minGSSize = 3, maxGSSize = 1000, with all other settings at default.

ATAC-seq data alignment and preprocessing

ATAC-seq libraries were sequenced as 150-bp paired-end reads on an Illumina NovaSeq X Plus platform. Quality assessment was performed using FastQC with the parameters ‘--extract --o out_dir -t 4 fastqc.file’. Index, primer, and adapter sequences were trimmed from the paired-end reads using Trimmomatic v0.38 (http://www.usadellab.org/cms/?page=trimmomatic) with the parameters ‘java jar trimmoma 0.38.jar PE -phred33 fasq1.file fastq2.file ILLUMINACLIP:fasta.file:2:30:10 LEADING:3 TAILING:3 SLIDINGWINDOW:4:15 MINLEN:36’. The trimmed paired-end reads or raw single-end reads, were then aligned to the Mus musculus genome (mm10) using Bowtie2 (v2.2.5) with default settings. Reads with a mapping quality score of ≥30 were retained for further analysis.

PCR duplicates were removed using samtools (v1.3.1)⁷³ with the command ‘samtools sort in_bam|samtools rmdup -s - out_bam’. For downstream analysis, reads from biological replicates were pooled. TP10M values were calculated and stored in bedGraph files using Homer (http://homer.ucsd.edu/homer/) with the command ‘makeUCSCfile out_dir -o out.bdg -name sample_name -avg -fsize 1e20’.

Peaks in the ATAC-seq data were identified using MACS (v1.4.2)⁷⁴ with default settings. Peaks overlapping with ENCODE⁷⁵ blacklisted genomic regions were excluded from analysis. Tag densities of ATAC-seq peaks were computed for promoter or enhancer regions using the ‘ScoreMatrixBin’ function of the R package genomation (v1.34.0) (https://www.bioconductor.org/packages/release/bioc/html/genomation.html). To ensure consistency, ATAC-seq tag densities across samples were normalized using DESeq2 size factors.

Differential analysis and motif analysis of ATAC-seq regions

For in vivo ATAC-seq data, differential analysis was performed based on the normalized ATAC-seq tag densities using the R package limma (v3.58.1)⁷⁶. An adjusted p-value cutoff of 0.005 and a log₂(fold-change) threshold of log₂(3) were applied to identify significantly differential regions. For genes associated with the variable regions, we matched the nearest transcription start site (TSS) to the variable regions, retaining only those genes with TP0.1 M > 1. For ex vivo ATAC-seq data, all regions were grouped via the shared nearest neighbor (SNN) clustering algorithm, which was implemented via the FindNeighbors and FindClusters functions in the R package Seurat (v4.4.0), with the parameters ‘resolution = 4’.

Motif enrichment analysis was completed using Homer’s findMotifsGenome.pl script with the ‘-size given’ option and default settings to identify enriched DNA motifs in the ATAC-seq regions.

Mass spectrometry (MS) data alignment, processing, and differential analysis

MS data acquired on an Orbitrap Exploris 480 were processed using Proteome Discoverer 2.2. Database searches were conducted against the Mus musculus Uniprot database using the following parameters: trypsin digestion ( ≤ 2 missed cleavages), fixed cysteine carbamidomethylation, variable methionine oxidation and N-terminal acetylation, precursor ion tolerance of 10 ppm, fragment ion tolerance of 0.02 Da, peptide- and protein-level FDR ≤ 1%. Proteins passing high-confidence thresholds were used for differential analysis. Differential protein enrichment was determined from Sequest HT-identified peptides using the limma R package (v3.58.1). Significance thresholds were adjusted p < 0.05 and log₂(fold change) > 1 (for Fig. 5g: adjusted p < 0.01 and log₂(fold-change) > 0.25). Protein accessions were converted to gene symbols using Uniprot⁷⁷ for downstream GO enrichment analysis.

Statistics and reproducibility

For pseudotime analyses, p-values were calculated using a two-tailed Wilcoxon rank-sum test. For gene expression analyses associated with variable regions, p-values were calculated using a two-tailed Wilcoxon signed-rank test. For RT-qPCR measurements, BC length, cell counts, and signal intensity quantifications, data distribution and variance homogeneity were formally assessed for each comparison using the Shapiro–Wilk test and Levene’s test, respectively. When data satisfied assumptions of normality and homoscedasticity, two-tailed unpaired Student’s t-tests (for two-group comparisons) or one-way ANOVA with appropriate post hoc tests (Tukey’s HSD or Dunnett’s) were used. When assumptions of normality or equal variance were not met, appropriate alternative statistical methods were employed, including Welch’s ANOVA with Games–Howell post hoc correction, Friedman’s test with Conover’s post hoc test, or robust ANOVA, as specified in the corresponding figure legends. All statistical tests were two-sided. Adjusted P values were computed using the Benjamini–Hochberg (BH) procedure to control the false discovery rate (FDR) wherever multiple comparisons were conducted.

Biological replicates were defined as individual embryos, mouse livers, or independently prepared ex vivo culture wells. At least 2 biological replicates were performed with consistent results, ensuring reproducibility. For BC length analysis, all identifiable BC structures within 1–5 images per replicate (typically tens to hundreds of BCs) were measured, and the mean value was used as the representative value. A similar approach was applied to other quantifications, including polarized or proliferating cell proportions, signal intensity, and bile duct structural analysis, where tens to hundreds of cells or structures per replicate were analyzed and averaged prior to statistical testing.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information^{(4MB, pdf)}

Peer Review file^{(18.5MB, pdf)}

41467_2026_73856_MOESM3_ESM.pdf^{(89.6KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data 1^{(3.5MB, xlsx)}

Supplementary Data 2^{(13.7MB, xlsx)}

Supplementary Data 3^{(405.6KB, xlsx)}

Supplementary Data 4^{(28.7KB, xlsx)}

Reporting Summary^{(108.7KB, pdf)}

Source data

Source Data^{(1.3MB, xlsx)}

Acknowledgements

We thank the members of the Xu laboratory for their valuable advice and discussion; the Peking-Tsinghua Center for Life Science Computing Platform; Dr. Yan-Chun Wang from Haidian Maternal & Child Health Hospital in Beijing for assistance with human embryonic liver tissue collection; Dr. Ying-Chun Hu from the Cryo-EM Platform at Peking University for assistance with transmission electron microscopy sample preparation and imaging; and Dr. Dong Liu from the National Center for Protein Sciences at Peking University for assistance with mass spectrometry. We thank BioRender (BioRender.com) for providing elements used to create figures in this article.

Author contributions

C.-R.X. and E.B. conceived the project; L.Y., E.B., and C.-R.X. designed the research; L.Y., L.W., and J.H. performed the research; L.Y., Y.-H.Z., J.H., E.B., W.Z., and C.-R.X. analyzed the data; L.Y., E.B., and C.-R.X. wrote the manuscript.

Peer review

Peer review information

Nature Communications thanks Naoki Tanimizu, Marino Zerial, and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This research was supported by the National Key Research and Development Program of China (grants 2023YFA1800600 and 2024YFA1802900 to C.-R.X.), the National Natural Science Foundation of China (grants 32125014 and 32030034 to C.-R.X., and 32470871 to L.Y.), and the National Institute of Diabetes and Digestive and Kidney Diseases of the USA (grant R01DK128861 to E.B.), the funding from the Peking-Tsinghua Center for Life Sciences to C.-R.X., the Noncommunicable Chronic Diseases-National Science and Technology Major Project (grant 2024ZD0531400 to L.Y.), and the State Key Laboratory of Reproductive Regulation & Breeding of Grassland Livestock (grant 2025KYPT0081 to L.Y.).

Data availability

The mouse RNA-seq and ATAC-seq data generated in this study have been deposited in the Gene Expression Omnibus (GEO) under accession code GSE281035. The human RNA-seq data have been deposited in the Genome Sequence Archive (GSA) under accession code HRA006394. Access to the human RNA-seq data is controlled in accordance with Chinese regulations on human genetic resources and is granted through GSA upon reasonable request. The mass spectrometry proteomics data have been deposited in the ProteomeXchange Consortium via the iProX partner repository with the dataset identifier IPX0017102000 and ProteomeXchange accession PXD078219. All relevant raw data from figures generated in this study have been provided in the Source Data file. Source data are provided with this paper.

Code availability

All original analysis scripts and code used in this study have been deposited at Zenodo under the 10.5281/zenodo.17314838.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Li Yang, Email: liyang@imu.edu.cn.

Erfei Bi, Email: ebi@pennmedicine.upenn.edu.

Cheng-Ran Xu, Email: cxu@pku.edu.cn.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-73856-w.

References

1.Yang, L. et al. A single-cell transcriptomic analysis reveals precise pathways and regulatory mechanisms underlying hepatoblast differentiation. Hepatology66, 1387–1401 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Wang, X. et al. Comparative analysis of cell lineage differentiation during hepatogenesis in humans and mice at the single-cell transcriptome level. Cell Res.30, 1109–1126 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Yang, L. et al. Determination of key events in mouse hepatocyte maturation at the single-cell level. Dev. Cell58, 1996–2010 e1996 (2023). [DOI] [PubMed] [Google Scholar]
4.Yang, Y. et al. An optimized proteomics approach reveals novel alternative proteins in mouse liver development. Mol. Cell Proteom.22, 100480 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Rowe, C. et al. Proteome-wide analyses of human hepatocytes during differentiation and dedifferentiation. Hepatology58, 799–809 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Yang, L. et al. The default and directed pathways of hepatoblast differentiation involve distinct epigenomic mechanisms. Dev. Cell58, 1688–1700 e1686 (2023). [DOI] [PubMed] [Google Scholar]
7.Macchi, F. & Sadler, K. C. Unraveling the epigenetic basis of liver development, regeneration and disease. Trends Genet36, 587–597 (2020). [DOI] [PubMed] [Google Scholar]
8.Grindheim, J. M., Nicetto, D., Donahue, G. & Zaret, K. S. Polycomb repressive complex 2 proteins EZH1 and EZH2 regulate timing of postnatal hepatocyte maturation and fibrosis by repressing genes with euchromatic promoters in mice. Gastroenterology156, 1834–1848 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Si-Tayeb, K., Lemaigre, F. P. & Duncan, S. A. Organogenesis and development of the liver. Dev. Cell18, 175–189 (2010). [DOI] [PubMed] [Google Scholar]
10.Wang, L. & Boyer, J. L. The maintenance and generation of membrane polarity in hepatocytes. Hepatology39, 892–899 (2004). [DOI] [PubMed] [Google Scholar]
11.Wakabayashi, Y., Dutt, P., Lippincott-Schwartz, J. & Arias, I. M. Rab11a and myosin Vb are required for bile canalicular formation in WIF-B9 cells. Proc. Natl. Acad. Sci. USA102, 15087–15092 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Tuma, P. L., Nyasae, L. K., Backer, J. M. & Hubbard, A. L. Vps34p differentially regulates endocytosis from the apical and basolateral domains in polarized hepatic cells. J. Cell Biol.154, 1197–1208 (2001). [DOI] [PMC free article] [PubMed] [Google Scholar]
13.van der Wouden, J. M., van, I. S. C. & Hoekstra, D. Oncostatin M regulates membrane traffic and stimulates bile canalicular membrane biogenesis in HepG2 cells. EMBO J.21, 6409–6418 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Herrema, H. et al. Rho kinase, myosin-II, and p42/44 MAPK control extracellular matrix-mediated apical bile canalicular lumen morphogenesis in HepG2 cells. Mol. Biol. Cell17, 3291–3303 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Slim, C. L. et al. Par1b induces asymmetric inheritance of plasma membrane domains via LGN-dependent mitotic spindle orientation in proliferating hepatocytes. PLoS Biol.11, e1001739 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Lazaro-Dieguez, F. et al. Par1b links lumen polarity with LGN-NuMA positioning for distinct epithelial cell division phenotypes. J. Cell Biol.203, 251–264 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Tocan, V. et al. Hepatocyte polarity establishment and apical lumen formation are organized by Par3, Cdc42, and aPKC in conjunction with Lgl. J. Biol. Chem.297, 101354 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Hao, Y. et al. Par3 controls epithelial spindle orientation by aPKC-mediated phosphorylation of apical Pins. Curr. Biol.20, 1809–1818 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Fu, D., Wakabayashi, Y., Lippincott-Schwartz, J. & Arias, I. M. Bile acid stimulates hepatocyte polarization through a cAMP-Epac-MEK-LKB1-AMPK pathway. Proc. Natl. Acad. Sci. USA108, 1403–1408 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Fu, D., Wakabayashi, Y., Ido, Y., Lippincott-Schwartz, J. & Arias, I. M. Regulation of bile canalicular network formation and maintenance by AMP-activated protein kinase and LKB1. J. Cell Sci.123, 3294–3302 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Homolya, L. et al. LKB1/AMPK and PKA control ABCB11 trafficking and polarization in hepatocytes. PLoS One9, e91921 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Belicova, L. et al. Anisotropic expansion of hepatocyte lumina enforced by apical bulkheads. J. Cell Biol.220 (2021). 10.1083/jcb.202103003 [DOI] [PMC free article] [PubMed]
23.van Hengel, J. et al. Continuous cell injury promotes hepatic tumorigenesis in cdc42-deficient mouse liver. Gastroenterology134, 781–792 (2008). [DOI] [PubMed] [Google Scholar]
24.Lu, M. et al. Basolateral CD147 induces hepatocyte polarity loss by E-cadherin ubiquitination and degradation in hepatocellular carcinoma progress. Hepatology68, 317–332 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Wang, T., Yanger, K., Stanger, B. Z., Cassio, D. & Bi, E. Cytokinesis defines a spatial landmark for hepatocyte polarization and apical lumen formation. J. Cell Sci.127, 2483–2492 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Peng, X. et al. How to induce non-polarized cells of hepatic origin to express typical hepatocyte polarity: generation of new highly polarized cell models with developed and functional bile canaliculi. Cell Tissue Res.323, 233–243 (2006). [DOI] [PubMed] [Google Scholar]
27.Hedman, A. C., Smith, J. M. & Sacks, D. B. The biology of IQGAP proteins: beyond the cytoskeleton. EMBO Rep.16, 427–446 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Watanabe, T., Wang, S. & Kaibuchi, K. IQGAPs as key regulators of actin-cytoskeleton dynamics. Cell Struct. Funct.40, 69–77 (2015). [DOI] [PubMed] [Google Scholar]
29.Tanos, B. E., Yeaman, C. & Rodriguez-Boulan, E. An emerging role for IQGAP1 in tight junction control. Small GTPases9, 375–383 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Cong, W. et al. ASPP2 regulates epithelial cell polarity through the PAR complex. Curr. Biol.20, 1408–1414 (2010). [DOI] [PubMed] [Google Scholar]
31.Royer, C. et al. ASPP2 links the apical lateral polarity complex to the regulation of YAP activity in epithelial cells. PLoS One9, e111384 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Buckley, C. E. & St Johnston, D. Apical-basal polarity and the control of epithelial form and function. Nat. Rev. Mol. Cell Biol.23, 559–577 (2022). [DOI] [PubMed] [Google Scholar]
33.Tang, N., Marshall, W. F., McMahon, M., Metzger, R. J. & Martin, G. R. Control of mitotic spindle angle by the RAS-regulated ERK1/2 pathway determines lung tube shape. Science333, 342–345 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Kodan, A. et al. ABCB1/MDR1/P-gp employs an ATP-dependent twist-and-squeeze mechanism to export hydrophobic drugs. FEBS Lett.595, 707–716 (2021). [DOI] [PubMed] [Google Scholar]
35.Carmena, M., Earnshaw, W. C. & Glover, D. M. The Dawn of Aurora kinase research: from fly genetics to the clinic. Front Cell Dev. Biol.3, 73 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Thompson, B. J. Par-3 family proteins in cell polarity & adhesion. FEBS J.289, 596–613 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Lee, C. S., Sund, N. J., Behr, R., Herrera, P. L. & Kaestner, K. H. Foxa2 is required for the differentiation of pancreatic alpha-cells. Dev. Biol.278, 484–495 (2005). [DOI] [PubMed] [Google Scholar]
38.Tanimizu, N., Nishikawa, M., Saito, H., Tsujimura, T. & Miyajima, A. Isolation of hepatoblasts based on the expression of Dlk/Pref-1. J. Cell Sci.116, 1775–1786 (2003). [DOI] [PubMed] [Google Scholar]
39.Hagemann-Jensen, M., Ziegenhain, C. & Sandberg, R. Scalable single-cell RNA sequencing from full transcripts with Smart-seq3xpress. Nat. Biotechnol. (2022). 10.1038/s41587-022-01311-4 [DOI] [PMC free article] [PubMed]
40.Roskoski, R. Jr ERK1/2 MAP kinases: structure, function, and regulation. Pharm. Res.66, 105–143 (2012). [DOI] [PubMed] [Google Scholar]
41.Schneeweis, C. et al. AP1/Fra1 confers resistance to MAPK cascade inhibition in pancreatic cancer. Cell Mol. Life Sci.80, 12 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Zhang, S. et al. Chromatin accessibility uncovers KRAS-driven FOSL2 promoting pancreatic ductal adenocarcinoma progression through up-regulation of CCL28. Br. J. Cancer129, 426–443 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Zhang, W. & Liu, H. T. MAPK signal pathways in the regulation of cell proliferation in mammalian cells. Cell Res.12, 9–18 (2002). [DOI] [PubMed] [Google Scholar]
44.Pulverer, B. J., Kyriakis, J. M., Avruch, J., Nikolakaki, E. & Woodgett, J. R. Phosphorylation of c-jun mediated by MAP kinases. Nature353, 670–674 (1991). [DOI] [PubMed] [Google Scholar]
45.Whitmarsh, A. J. & Davis, R. J. Transcription factor AP-1 regulation by mitogen-activated protein kinase signal transduction pathways. J. Mol. Med.74, 589–607 (1996). [DOI] [PubMed] [Google Scholar]
46.Kellendonk, C., Opherk, C., Anlag, K., Schutz, G. & Tronche, F. Hepatocyte-specific expression of Cre recombinase. Genesis26, 151–153 (2000). [DOI] [PubMed] [Google Scholar]
47.Tanimizu, N. et al. Intrahepatic bile ducts are developed through formation of homogeneous continuous luminal network and its dynamic rearrangement in mice. Hepatology64, 175–188 (2016). [DOI] [PubMed] [Google Scholar]
48.Bebelman, M. P. et al. Hepatocyte apical bulkheads provide a mechanical means to oppose bile pressure. J Cell Biol222 (2023). 10.1083/jcb.202208002 [DOI] [PMC free article] [PubMed]
49.Hoeflich, K. P. & Ikura, M. Radixin: cytoskeletal adopter and signaling protein. Int J. Biochem Cell Biol.36, 2131–2136 (2004). [DOI] [PubMed] [Google Scholar]
50.Zhang, P. et al. Dual function of partitioning-defective 3 in the regulation of YAP phosphorylation and activation. Cell Discov.2, 16021 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Kozikowski, A. P. et al. New amide-bearing benzolactam-based protein kinase C modulators induce enhanced secretion of the amyloid precursor protein metabolite sAPPalpha. J. Med. Chem.46, 364–373 (2003). [DOI] [PubMed] [Google Scholar]
52.Ren, J. G., Li, Z. & Sacks, D. B. IQGAP1 modulates activation of B-Raf. Proc. Natl. Acad. Sci. USA104, 10465–10469 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
53.McNulty, D. E., Li, Z., White, C. D., Sacks, D. B. & Annan, R. S. MAPK scaffold IQGAP1 binds the EGF receptor and modulates its activation. J. Biol. Chem.286, 15010–15021 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Ren, J. G., Li, Z. & Sacks, D. B. IQGAP1 integrates Ca2+/calmodulin and B-Raf signaling. J. Biol. Chem.283, 22972–22982 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Meyer, K. et al. Bile canaliculi remodeling activates YAP via the actin cytoskeleton during liver regeneration. Mol. Syst. Biol.16, e8985 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Chembazhi, U. V., Bangru, S., Hernaez, M. & Kalsotra, A. Cellular plasticity balances the metabolic and proliferation dynamics of a regenerating liver. Genome Res.31, 576–591 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Fernandez, C., Lobo Md Mdel, V., Gomez-Coronado, D. & Lasuncion, M. A. Cholesterol is essential for mitosis progression and its deficiency induces polyploid cell formation. Exp. Cell Res.300, 109–120 (2004). [DOI] [PubMed] [Google Scholar]
58.Peng, J. et al. Enhanced liver regeneration after partial hepatectomy in sterol regulatory element-binding protein (SREBP)-1c-null mice is associated with increased hepatocellular cholesterol availability. Cell Physiol. Biochem.47, 784–799 (2018). [DOI] [PubMed] [Google Scholar]
59.Liu, Q. et al. Liver regeneration after injury: mechanisms, cellular interactions and therapeutic innovations. Clin. Transl. Med14, e1812 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Lee, C. S., Friedman, J. R., Fulmer, J. T. & Kaestner, K. H. The initiation of liver development is dependent on Foxa transcription factors. Nature435, 944–947 (2005). [DOI] [PubMed] [Google Scholar]
61.Zhang, N. et al. The Merlin/NF2 tumor suppressor functions through the YAP oncoprotein to regulate tissue homeostasis in mammals. Dev. Cell19, 27–38 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Wang, X., Xu, B. L. & Chen, X. W. Acute gene inactivation in the adult mouse liver using the CRISPR-Cas9 technology. STAR Protoc.2, 100611 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Lee, D. H. et al. LATS-YAP/TAZ controls lineage specification by regulating TGFbeta signaling and Hnf4alpha expression during liver development. Nat. Commun.7, 11961 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Buenrostro, J. D., Giresi, P. G., Zaba, L. C., Chang, H. Y. & Greenleaf, W. J. Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nat. Methods10, 1213–1218 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics30, 2114–2120 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Kim, D., Langmead, B. & Salzberg, S. L. HISAT: a fast spliced aligner with low memory requirements. Nat. Methods12, 357–360 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics30, 923–930 (2014). [DOI] [PubMed] [Google Scholar]
68.Hao, Y. et al. Integrated analysis of multimodal single-cell data. Cell184, 3573–3587 e3529 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Haghverdi, L., Lun, A. T. L., Morgan, M. D. & Marioni, J. C. Batch effects in single-cell RNA-sequencing data are corrected by matching mutual nearest neighbors. Nat. Biotechnol.36, 421–427 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Trevor Hastie, W. S. Principal curves. J. Am. Stat. Assoc.84, 502–516 (1989). [Google Scholar]
71.Cannoodt, R. Princurve 2.0: fit a principal curve in arbitrary dimension. (2018).
72.Yu, G., Wang, L. G., Han, Y. & He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS16, 284–287 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Danecek, P. et al. Twelve years of SAMtools and BCFtools. Gigascience10 (2021). 10.1093/gigascience/giab008 [DOI] [PMC free article] [PubMed]
74.Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol.9, R137 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
75.Consortium, E. P. An integrated encyclopedia of DNA elements in the human genome. Nature489, 57–74 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
76.Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res.43, e47 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
77.UniProt, C. UniProt: the Universal Protein Knowledgebase in 2023. Nucleic Acids Res.51, D523–D531 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(4MB, pdf)}

Peer Review file^{(18.5MB, pdf)}

41467_2026_73856_MOESM3_ESM.pdf^{(89.6KB, pdf)}

Description of Additional Supplementary Files

Supplementary Data 1^{(3.5MB, xlsx)}

Supplementary Data 2^{(13.7MB, xlsx)}

Supplementary Data 3^{(405.6KB, xlsx)}

Supplementary Data 4^{(28.7KB, xlsx)}

Reporting Summary^{(108.7KB, pdf)}

Source Data^{(1.3MB, xlsx)}

Data Availability Statement

All original analysis scripts and code used in this study have been deposited at Zenodo under the 10.5281/zenodo.17314838.

[CR1] 1.Yang, L. et al. A single-cell transcriptomic analysis reveals precise pathways and regulatory mechanisms underlying hepatoblast differentiation. Hepatology66, 1387–1401 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] 2.Wang, X. et al. Comparative analysis of cell lineage differentiation during hepatogenesis in humans and mice at the single-cell transcriptome level. Cell Res.30, 1109–1126 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Yang, L. et al. Determination of key events in mouse hepatocyte maturation at the single-cell level. Dev. Cell58, 1996–2010 e1996 (2023). [DOI] [PubMed] [Google Scholar]

[CR4] 4.Yang, Y. et al. An optimized proteomics approach reveals novel alternative proteins in mouse liver development. Mol. Cell Proteom.22, 100480 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR5] 5.Rowe, C. et al. Proteome-wide analyses of human hepatocytes during differentiation and dedifferentiation. Hepatology58, 799–809 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Yang, L. et al. The default and directed pathways of hepatoblast differentiation involve distinct epigenomic mechanisms. Dev. Cell58, 1688–1700 e1686 (2023). [DOI] [PubMed] [Google Scholar]

[CR7] 7.Macchi, F. & Sadler, K. C. Unraveling the epigenetic basis of liver development, regeneration and disease. Trends Genet36, 587–597 (2020). [DOI] [PubMed] [Google Scholar]

[CR8] 8.Grindheim, J. M., Nicetto, D., Donahue, G. & Zaret, K. S. Polycomb repressive complex 2 proteins EZH1 and EZH2 regulate timing of postnatal hepatocyte maturation and fibrosis by repressing genes with euchromatic promoters in mice. Gastroenterology156, 1834–1848 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Si-Tayeb, K., Lemaigre, F. P. & Duncan, S. A. Organogenesis and development of the liver. Dev. Cell18, 175–189 (2010). [DOI] [PubMed] [Google Scholar]

[CR10] 10.Wang, L. & Boyer, J. L. The maintenance and generation of membrane polarity in hepatocytes. Hepatology39, 892–899 (2004). [DOI] [PubMed] [Google Scholar]

[CR11] 11.Wakabayashi, Y., Dutt, P., Lippincott-Schwartz, J. & Arias, I. M. Rab11a and myosin Vb are required for bile canalicular formation in WIF-B9 cells. Proc. Natl. Acad. Sci. USA102, 15087–15092 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Tuma, P. L., Nyasae, L. K., Backer, J. M. & Hubbard, A. L. Vps34p differentially regulates endocytosis from the apical and basolateral domains in polarized hepatic cells. J. Cell Biol.154, 1197–1208 (2001). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR13] 13.van der Wouden, J. M., van, I. S. C. & Hoekstra, D. Oncostatin M regulates membrane traffic and stimulates bile canalicular membrane biogenesis in HepG2 cells. EMBO J.21, 6409–6418 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Herrema, H. et al. Rho kinase, myosin-II, and p42/44 MAPK control extracellular matrix-mediated apical bile canalicular lumen morphogenesis in HepG2 cells. Mol. Biol. Cell17, 3291–3303 (2006). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Slim, C. L. et al. Par1b induces asymmetric inheritance of plasma membrane domains via LGN-dependent mitotic spindle orientation in proliferating hepatocytes. PLoS Biol.11, e1001739 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.Lazaro-Dieguez, F. et al. Par1b links lumen polarity with LGN-NuMA positioning for distinct epithelial cell division phenotypes. J. Cell Biol.203, 251–264 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Tocan, V. et al. Hepatocyte polarity establishment and apical lumen formation are organized by Par3, Cdc42, and aPKC in conjunction with Lgl. J. Biol. Chem.297, 101354 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Hao, Y. et al. Par3 controls epithelial spindle orientation by aPKC-mediated phosphorylation of apical Pins. Curr. Biol.20, 1809–1818 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Fu, D., Wakabayashi, Y., Lippincott-Schwartz, J. & Arias, I. M. Bile acid stimulates hepatocyte polarization through a cAMP-Epac-MEK-LKB1-AMPK pathway. Proc. Natl. Acad. Sci. USA108, 1403–1408 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Fu, D., Wakabayashi, Y., Ido, Y., Lippincott-Schwartz, J. & Arias, I. M. Regulation of bile canalicular network formation and maintenance by AMP-activated protein kinase and LKB1. J. Cell Sci.123, 3294–3302 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Homolya, L. et al. LKB1/AMPK and PKA control ABCB11 trafficking and polarization in hepatocytes. PLoS One9, e91921 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Belicova, L. et al. Anisotropic expansion of hepatocyte lumina enforced by apical bulkheads. J. Cell Biol.220 (2021). 10.1083/jcb.202103003 [DOI] [PMC free article] [PubMed]

[CR23] 23.van Hengel, J. et al. Continuous cell injury promotes hepatic tumorigenesis in cdc42-deficient mouse liver. Gastroenterology134, 781–792 (2008). [DOI] [PubMed] [Google Scholar]

[CR24] 24.Lu, M. et al. Basolateral CD147 induces hepatocyte polarity loss by E-cadherin ubiquitination and degradation in hepatocellular carcinoma progress. Hepatology68, 317–332 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.Wang, T., Yanger, K., Stanger, B. Z., Cassio, D. & Bi, E. Cytokinesis defines a spatial landmark for hepatocyte polarization and apical lumen formation. J. Cell Sci.127, 2483–2492 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] 26.Peng, X. et al. How to induce non-polarized cells of hepatic origin to express typical hepatocyte polarity: generation of new highly polarized cell models with developed and functional bile canaliculi. Cell Tissue Res.323, 233–243 (2006). [DOI] [PubMed] [Google Scholar]

[CR27] 27.Hedman, A. C., Smith, J. M. & Sacks, D. B. The biology of IQGAP proteins: beyond the cytoskeleton. EMBO Rep.16, 427–446 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Watanabe, T., Wang, S. & Kaibuchi, K. IQGAPs as key regulators of actin-cytoskeleton dynamics. Cell Struct. Funct.40, 69–77 (2015). [DOI] [PubMed] [Google Scholar]

[CR29] 29.Tanos, B. E., Yeaman, C. & Rodriguez-Boulan, E. An emerging role for IQGAP1 in tight junction control. Small GTPases9, 375–383 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Cong, W. et al. ASPP2 regulates epithelial cell polarity through the PAR complex. Curr. Biol.20, 1408–1414 (2010). [DOI] [PubMed] [Google Scholar]

[CR31] 31.Royer, C. et al. ASPP2 links the apical lateral polarity complex to the regulation of YAP activity in epithelial cells. PLoS One9, e111384 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Buckley, C. E. & St Johnston, D. Apical-basal polarity and the control of epithelial form and function. Nat. Rev. Mol. Cell Biol.23, 559–577 (2022). [DOI] [PubMed] [Google Scholar]

[CR33] 33.Tang, N., Marshall, W. F., McMahon, M., Metzger, R. J. & Martin, G. R. Control of mitotic spindle angle by the RAS-regulated ERK1/2 pathway determines lung tube shape. Science333, 342–345 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Kodan, A. et al. ABCB1/MDR1/P-gp employs an ATP-dependent twist-and-squeeze mechanism to export hydrophobic drugs. FEBS Lett.595, 707–716 (2021). [DOI] [PubMed] [Google Scholar]

[CR35] 35.Carmena, M., Earnshaw, W. C. & Glover, D. M. The Dawn of Aurora kinase research: from fly genetics to the clinic. Front Cell Dev. Biol.3, 73 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Thompson, B. J. Par-3 family proteins in cell polarity & adhesion. FEBS J.289, 596–613 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Lee, C. S., Sund, N. J., Behr, R., Herrera, P. L. & Kaestner, K. H. Foxa2 is required for the differentiation of pancreatic alpha-cells. Dev. Biol.278, 484–495 (2005). [DOI] [PubMed] [Google Scholar]

[CR38] 38.Tanimizu, N., Nishikawa, M., Saito, H., Tsujimura, T. & Miyajima, A. Isolation of hepatoblasts based on the expression of Dlk/Pref-1. J. Cell Sci.116, 1775–1786 (2003). [DOI] [PubMed] [Google Scholar]

[CR39] 39.Hagemann-Jensen, M., Ziegenhain, C. & Sandberg, R. Scalable single-cell RNA sequencing from full transcripts with Smart-seq3xpress. Nat. Biotechnol. (2022). 10.1038/s41587-022-01311-4 [DOI] [PMC free article] [PubMed]

[CR40] 40.Roskoski, R. Jr ERK1/2 MAP kinases: structure, function, and regulation. Pharm. Res.66, 105–143 (2012). [DOI] [PubMed] [Google Scholar]

[CR41] 41.Schneeweis, C. et al. AP1/Fra1 confers resistance to MAPK cascade inhibition in pancreatic cancer. Cell Mol. Life Sci.80, 12 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR42] 42.Zhang, S. et al. Chromatin accessibility uncovers KRAS-driven FOSL2 promoting pancreatic ductal adenocarcinoma progression through up-regulation of CCL28. Br. J. Cancer129, 426–443 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Zhang, W. & Liu, H. T. MAPK signal pathways in the regulation of cell proliferation in mammalian cells. Cell Res.12, 9–18 (2002). [DOI] [PubMed] [Google Scholar]

[CR44] 44.Pulverer, B. J., Kyriakis, J. M., Avruch, J., Nikolakaki, E. & Woodgett, J. R. Phosphorylation of c-jun mediated by MAP kinases. Nature353, 670–674 (1991). [DOI] [PubMed] [Google Scholar]

[CR45] 45.Whitmarsh, A. J. & Davis, R. J. Transcription factor AP-1 regulation by mitogen-activated protein kinase signal transduction pathways. J. Mol. Med.74, 589–607 (1996). [DOI] [PubMed] [Google Scholar]

[CR46] 46.Kellendonk, C., Opherk, C., Anlag, K., Schutz, G. & Tronche, F. Hepatocyte-specific expression of Cre recombinase. Genesis26, 151–153 (2000). [DOI] [PubMed] [Google Scholar]

[CR47] 47.Tanimizu, N. et al. Intrahepatic bile ducts are developed through formation of homogeneous continuous luminal network and its dynamic rearrangement in mice. Hepatology64, 175–188 (2016). [DOI] [PubMed] [Google Scholar]

[CR48] 48.Bebelman, M. P. et al. Hepatocyte apical bulkheads provide a mechanical means to oppose bile pressure. J Cell Biol222 (2023). 10.1083/jcb.202208002 [DOI] [PMC free article] [PubMed]

[CR49] 49.Hoeflich, K. P. & Ikura, M. Radixin: cytoskeletal adopter and signaling protein. Int J. Biochem Cell Biol.36, 2131–2136 (2004). [DOI] [PubMed] [Google Scholar]

[CR50] 50.Zhang, P. et al. Dual function of partitioning-defective 3 in the regulation of YAP phosphorylation and activation. Cell Discov.2, 16021 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] 51.Kozikowski, A. P. et al. New amide-bearing benzolactam-based protein kinase C modulators induce enhanced secretion of the amyloid precursor protein metabolite sAPPalpha. J. Med. Chem.46, 364–373 (2003). [DOI] [PubMed] [Google Scholar]

[CR52] 52.Ren, J. G., Li, Z. & Sacks, D. B. IQGAP1 modulates activation of B-Raf. Proc. Natl. Acad. Sci. USA104, 10465–10469 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] 53.McNulty, D. E., Li, Z., White, C. D., Sacks, D. B. & Annan, R. S. MAPK scaffold IQGAP1 binds the EGF receptor and modulates its activation. J. Biol. Chem.286, 15010–15021 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Ren, J. G., Li, Z. & Sacks, D. B. IQGAP1 integrates Ca2+/calmodulin and B-Raf signaling. J. Biol. Chem.283, 22972–22982 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Meyer, K. et al. Bile canaliculi remodeling activates YAP via the actin cytoskeleton during liver regeneration. Mol. Syst. Biol.16, e8985 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Chembazhi, U. V., Bangru, S., Hernaez, M. & Kalsotra, A. Cellular plasticity balances the metabolic and proliferation dynamics of a regenerating liver. Genome Res.31, 576–591 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR57] 57.Fernandez, C., Lobo Md Mdel, V., Gomez-Coronado, D. & Lasuncion, M. A. Cholesterol is essential for mitosis progression and its deficiency induces polyploid cell formation. Exp. Cell Res.300, 109–120 (2004). [DOI] [PubMed] [Google Scholar]

[CR58] 58.Peng, J. et al. Enhanced liver regeneration after partial hepatectomy in sterol regulatory element-binding protein (SREBP)-1c-null mice is associated with increased hepatocellular cholesterol availability. Cell Physiol. Biochem.47, 784–799 (2018). [DOI] [PubMed] [Google Scholar]

[CR59] 59.Liu, Q. et al. Liver regeneration after injury: mechanisms, cellular interactions and therapeutic innovations. Clin. Transl. Med14, e1812 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR60] 60.Lee, C. S., Friedman, J. R., Fulmer, J. T. & Kaestner, K. H. The initiation of liver development is dependent on Foxa transcription factors. Nature435, 944–947 (2005). [DOI] [PubMed] [Google Scholar]

[CR61] 61.Zhang, N. et al. The Merlin/NF2 tumor suppressor functions through the YAP oncoprotein to regulate tissue homeostasis in mammals. Dev. Cell19, 27–38 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR62] 62.Wang, X., Xu, B. L. & Chen, X. W. Acute gene inactivation in the adult mouse liver using the CRISPR-Cas9 technology. STAR Protoc.2, 100611 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR63] 63.Lee, D. H. et al. LATS-YAP/TAZ controls lineage specification by regulating TGFbeta signaling and Hnf4alpha expression during liver development. Nat. Commun.7, 11961 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR64] 64.Buenrostro, J. D., Giresi, P. G., Zaba, L. C., Chang, H. Y. & Greenleaf, W. J. Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nat. Methods10, 1213–1218 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR65] 65.Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics30, 2114–2120 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR66] 66.Kim, D., Langmead, B. & Salzberg, S. L. HISAT: a fast spliced aligner with low memory requirements. Nat. Methods12, 357–360 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR67] 67.Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics30, 923–930 (2014). [DOI] [PubMed] [Google Scholar]

[CR68] 68.Hao, Y. et al. Integrated analysis of multimodal single-cell data. Cell184, 3573–3587 e3529 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR69] 69.Haghverdi, L., Lun, A. T. L., Morgan, M. D. & Marioni, J. C. Batch effects in single-cell RNA-sequencing data are corrected by matching mutual nearest neighbors. Nat. Biotechnol.36, 421–427 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR70] 70.Trevor Hastie, W. S. Principal curves. J. Am. Stat. Assoc.84, 502–516 (1989). [Google Scholar]

[CR71] 71.Cannoodt, R. Princurve 2.0: fit a principal curve in arbitrary dimension. (2018).

[CR72] 72.Yu, G., Wang, L. G., Han, Y. & He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS16, 284–287 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR73] 73.Danecek, P. et al. Twelve years of SAMtools and BCFtools. Gigascience10 (2021). 10.1093/gigascience/giab008 [DOI] [PMC free article] [PubMed]

[CR74] 74.Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol.9, R137 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR75] 75.Consortium, E. P. An integrated encyclopedia of DNA elements in the human genome. Nature489, 57–74 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR76] 76.Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res.43, e47 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR77] 77.UniProt, C. UniProt: the Universal Protein Knowledgebase in 2023. Nucleic Acids Res.51, D523–D531 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

PAR3-mediated coordination of hepatocyte proliferation, maturation, and architecture in liver development and regeneration

Li Yang

Yu-Heng Zhou

Junya Hayase

Lei Wang

Weiguang Zhang

Erfei Bi

Cheng-Ran Xu

Abstract

Introduction

Results

Development of hepatic cell polarity and BC architecture

Fig. 1. Hepatic cell polarity and BC architecture during mouse and human liver development.

Linking hepatic polarity and BC formation to cell division

Disruption of Pard3 impairs BC elongation

Fig. 2. Pard3 deletion disrupts BC elongation and delays hepatocyte maturation.

Disruption of Pard3 delays hepatocyte development

YAP-PAR3 interaction in BC elongation and hepatocyte development

Fig. 3. Yap deletion disrupts BC elongation and delays hepatocyte maturation.

PAR3 regulates BC elongation and hepatocyte development by suppressing ERK activation

Fig. 4. PAR3 suppresses ERK/MAPK signaling to regulate BC elongation and hepatocyte maturation.

PAR3 drives BC elongation via IQGAP1-ERK signaling and actin coordination

Fig. 5. Involvement of the actin cytoskeleton and oriented cell division in bile BC elongation.

ERK/MAPK regulates human BC elongation and hepatocyte development

Fig. 6. Modulation of the ERK/MAPK pathway regulates human BC elongation and hepatocyte development in ex vivo cultures.

PAR3 regulates BC architecture during liver regeneration

Fig. 7. PAR3 regulates BC architecture during liver regeneration.

Discussion

Methods

Mice

Human samples

Cell lines

Immunofluorescence staining

Transmission electron microscopy

Cell preparation and FACS sorting

RT-qPCR

Western blotting

Co-immunoprecipitation (Co-IP) and mass spectrometry (MS)

Liver lobe culture

siRNA-mediated knockdown in Can 10 cells

GFP-PAR3 overexpression in Can 10 cells

AAV virus production and purification

70% hepatectomy

Serum ALT and AST measurement

Hepatoblast differentiation in vitro

Single-cell RNA-seq

ATAC-seq

Smart-seq3 scRNA-seq data alignment and preprocessing

Dimension reduction and pseudotime analysis

Differential gene expression analysis

Gene ontology enrichment analysis

ATAC-seq data alignment and preprocessing

Differential analysis and motif analysis of ATAC-seq regions

Mass spectrometry (MS) data alignment, processing, and differential analysis

Statistics and reproducibility

Reporting summary

Supplementary information

Source data

Acknowledgements

Author contributions

Peer review

Peer review information

Funding

Data availability

Code availability

Competing interests

Footnotes

Contributor Information

Supplementary information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases