Skip to main content
NCBI home page
Nature Portfolio logoLink to Nature Portfolio
. 2026 Aug 20;45(41):4362–4376. doi: 10.1038/s41388-026-03960-0

PTEN-AKT2 regulates mixed lineage liver cancer development and sensitizes cancer cells to TGFβ treatment

Qi Tang 1, Ielyzaveta Slarve 1, Jingyu Chen 1, Ni Zeng 1, Yiwei Gu 1, Lina He 1, Shunan Hu 1, Diala Alhousari 1, Zifei Xu 1, Brittney Hua 1, Guo Zhang 1, Phillip Nguyen 1, Mario Alba 2, Ajay Bharadwaj 3, Jihyeon Lee 3, Simion Kreimer 3, Jian Xu 4, Baoan Ji 5, Shelly C Lu 6, Jennifer Van Eyk 3, Shefali Chopra 7, Gary Kanel 7, Liyun Yuan 8, Bangyan L Stiles 1,7,✉
PMCID: PMC13645676  PMID: 42624900

Abstract

Primary liver cancers, including hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCA), arise from the neoplastic transformation of hepatocytes and cholangiocytes, respectively. Loss or downregulation of PTEN, a tumor suppressor negatively regulating the PI3K/AKT pathway, is frequently observed in CCA and HCC. Notably, PTEN mutations are observed at nearly twice the frequency in combined CCA-HCC tumors than either HCC or CCA alone. Using lineage-specific liver-targeted PTEN-deficient mouse models, we demonstrate that PTEN loss drives cellular dedifferentiation and tumorigenesis, a process that is critically dependent on AKT2. Mechanistically, PTEN deficiency induces activation of NOTCH and upregulation of transcriptional factor SOX9, which plays a central role in tumor cell transformation. In parallel, PTEN loss increases SMAD4 expression and sensitizes the tumor cells to TGFβ signaling, with TGFβ treatment repressing SOX9 expression in tumor cells lacking PTEN. Together, our study defined a critical role for PTEN-AKT2 signaling in maintaining liver epithelial lineage fidelity and revealed how its disruption promotes the conversion of mature hepatocytes or cholangiocytes into liver cancer stem-like cells (LCSCs). Furthermore, we identify a PTEN-dependent crosstalk between NOTCH and TGFβ pathways that governs liver tumor development. Together, this work provides mechanistic insight into lineage plasticity in liver cancer with implications for pathway-directed therapy.

graphic file with name 41388_2026_3960_Figa_HTML.webp

Open in a new tab

Subject terms: Biliary tract cancer, Growth factor signalling, Cancer stem cells

Introduction

Hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (iCCA) are the most common primary malignancies in the liver. Hyperproliferative cholangiocytes and their association with liver cancer stem–like cell (LCSC) markers confer an aggressive phenotype in iCCA and combined HCC–CCA (cHCC-CCA). Consistent with high rates of therapeutic resistance and relapse [1–3], the median relative survival for mixed cHCC-CCA and iCCA is approximately 8 months and 7 months, respectively (SEER*Stat database [4]). This mortality rate is projected to increase in the US, in contrast to the declining trends observed in most other major cancers [5]. Recent whole-genome and epigenetic analyses revealed substantial molecular heterogeneity within iCCA, with frequent alterations in isocitrate dehydrogenase1 (IDH1), BRCA1-associated protein1, and FGFR2 gene fusions associated with poor clinical outcomes [2, 6]. Overexpression of ERBB family receptor tyrosine kinases, including EGFR and c-Met, similarly predicts adverse prognosis. Notably, aberrant activation of FGFR2, EGFR and c-Met drives tumor progression through downstream engagement of phosphoinositide 3-kinase (PI3K)/AKT and ERK [7, 8]. Moreover, iCCAs harboring IDH1 mutations also exhibit enhanced PI3K/AKT activity [9], indicating convergence at this oncogenic pathway across distinct genetic subtypes.

PTEN (phosphatase and tensin homolog deleted on chromosome 10), a central tumor suppressor and negative regulator of the PI3K/AKT pathway [10], is frequently downregulated in liver cancers with iCCA indications, with loss of expression reported in up to 70% of cases [11]. In murine models with liver-specific Pten deletion, we previously reported the development of mixed lineage tumors, underscoring the critical role of PTEN in cHCC-CCA development [12–14]. Here, we demonstrated that intact PTEN is necessary to maintain both cholangiocytes and hepatocytes in their differentiated states, whereas PTEN loss in either lineage is sufficient to drive cHCC-CCA development. Mechanistically, we identified AKT2 as the key downstream effector that promotes mixed lineage tumor formation, as it converges on SOX9 to sustain LCSC transformation, with opposing regulatory inputs from the NOTCH and TGFβ signaling pathways. Specifically, NOTCH signaling activates SOX9, whereas TGFβ suppresses SOX9 expression, and PTEN status determines the cellular responsiveness to TGFβ inhibition. Collectively, these findings establish AKT2 as a critical oncogenic driver in cHCC-CCA and reveal a PTEN-dependent regulatory axis in which TGFβ-mediated control of SOX9 governs LCSC transformation and tumor aggressiveness. Importantly, our data indicate that PTEN loss sensitizes tumor cells to TGFβ inhibition, highlighting potential therapeutic vulnerabilities in mixed-lineage liver cancers.

Materials and methods

Patient samples

Human samples were obtained from University of Southern California (USC) tissue bank and experiments performed under protocols approved by the Institutional Review Board (IRB). Tissue bank samples were collected with informed conscent under USC IRB. Formalin–fixed, paraffin-embedded tissue sections were obtained from the USC Tissue Bank under protocols HS-20-00168 (LY) and HS-16-00258 (BLS).

Animals

The LiPten (PtenLoxP/LoxP; Alb-Cre+/-) and ChoPten (PtenloxP/loxP; R26RYFP; SOX9-CreERT+Tamoxifen) mice were previously reported [13, 15, 16]. HepPten mice were generated via injection of AAV8-TBG-Cre to PtenloxP/loxP; R26RYFP at 1 month of age. LiPtenA2 (PtenLoxP/LoxP; Akt2LoxP/LoxP; Alb-Cre+/-) mice were developed by crossing Akt2loxP/loxP (LiA2) mice [17] with the LiPten mice. Control mice were Cre-/- mice from the respective cohorts. Male mice on C57BL/6 J background were housed under standard conditions with ad libitum access to food and water and a 12-h light/dark cycle. All animal procedures were approved by the Institutional Animal Care and Use Committee at the University of Southern California and conducted in accordance with institutional and national guidelines.

Database analysis

Median survival in patients with carcinoma of liver and intrahepatic bile duct, HCC or CCA was estimated using the Kaplan–Meier method in SEER*Stat with 95% confidence intervals obtained from the summary statistics [4] (SEER 1975–2022; November 2024 Submission, accessed Nov 11, 2025). Genomic mutation data for liver and biliary tract cancers were obtained from cBioPortal [18] (accessed June 21, 2025). TCGA and GTEx datasets were analyzed using GEPIA2 (http://gepia2.cancer-pku.cn/) to evaluate patient survival and differential gene expression between tumor and normal tissues in the Liver Hepatocellular Carcinoma and Intrahepatic Cholangiocarcinoma cohorts [19]. Expression profiling data analyzed in this study were previously reported to GEO (GSE70501) [20]. ChIP-seq and ATAC-seq data for HepG2 cells were downloaded from ENCODE [21] and processed using standard IGV pipelines.

Cell culture

Huh7 (JCRB0403) [22], PLC/PRF/5 (ATCC, #CRL-8024) [23] and mouse immortalized cholangiocytes [24] were cultured in DMEM (Corning Inc., #10-013-CV) supplemented with 10% FBS (Neuromics, #FBS007). Mouse immortalized liver cell lines (WT, Pten-/-, Pten-/-; A2-/-) were established from WT, LiPten, LiPtenA2 mouse livers and cultured as previously reported [25, 26]. Cells were treated with DAPT (N-[N-(3,5-difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester) (Cayman Chemical, #13197) or recombinant human TGFβ1 protein (Abcam, #ab50036) as indicated. SOX9 knockdown and NICD overexpression were achieved using siRNA- and plasmid-based approaches as described in the Supplementary Methods.

Sphere and colony formation

Log-phase growing cells were seeded in ultra-low attachment plates (Corning, #3473) in DMEM/F12 medium (Corning, #10-090-CV) supplemented with 0.5xB27, 10ug/ml EGF, 5 ng/ml FGF and allowed to form spheres for up to 5 days. Seeding density and plate format were adjusted according to experimental requirements: 300,000 cells were seeded in 60 mm dishes, 5000–25,000 cells per well were seeded in 24-well plates. For serial passaging, primary spheres were dissociated into single cells to generate secondary and tertiary spheres using the same protocol. Spheres ≥100 μm in diameter were counted at the indicated time points.

For colony formation, log-phase cells (1250 per well) were suspended in 0.35% agar and overlaid onto a base agar layer (2 mL per well). Cultures were maintained at 37°C with 5% CO₂ for 2–3 weeks, with fresh medium added every 3–4 days.

Immunoblotting

Total protein was isolated from cultured cells or liver tissue homogenates using RIPA lysis buffer as previously described [27]. Samples were resolved by SDS-PAGE and transferred onto polyvinylidene fluoride membranes. Membranes were incubated with the primary antibodies (see Supplementary Methods). The HRP-conjugated secondary antibodies and ECL substrate were applied to visualize the protein bands.

Immunofluorescence staining

Paraffin-embedded tissue sections were deparaffinized, rehydrated, and subjected to antigen retrieval as described [28]. Primary antibodies (see Supplementary Methods) were incubated overnight at 4°C. Alexa Fluor-conjugated secondary antibodies (Thermo Fisher, 1:200) were applied for 1 h at room temperature. Nuclei were counterstained with DAPI and imaged by Zeiss Axio Imager M1 fluorescence microscopy.

RNA analysis

Total RNA was extracted from cultured cells and mouse liver tissue using TRIzol reagent (Invitrogen, #15596018). Reverse transcription was performed with the M-MLV reverse transcriptase system (Promega, PR-M1705). Quantitative PCR (qPCR) was then conducted using the APExBIO qPCR kit (#K1070). Gene-specific primers are listed in Supplementary Methods.

Statistical analysis

Two-tailed Student’s t-test (2 samples) or ANOVA with Tukey posthoc test (>2 samples) were conducted for statistical analysis. * indicates p value < 0.05, ** indicates p value < 0.01, *** indicates p value < 0.001. Data represented as Mean ± SEM.

Results

Mutation of PTEN is associated with mixed tumor phenotype with iCCA indications and poor tumor staging

The PTEN/PI3K signaling pathway has been considered a permissive cell growth signal that allows tumor development in the liver. We reported previously that loss of the tumor suppressor PTEN (LiPten mice) leads to the development of dual-lineage liver cancer after chronic steatosis and inflammation [14, 20, 29]. To understand the significance of PTEN-regulated signals in liver cancer, we performed mutational analysis of patient data (from 24 dataset total: 10 CCA, 13HCC, 1 combined) available in cBioPortal [18]. Besides previously reported mutations of p53 (32.4%), CTNNB1 (14.5%), ARID1A (13.3%) and TERT (13.2%), mutations of genes within the PTEN-regulated pathway were also observed (PIK3CA,4%; TSC2, 3.2%; and PTEN, 2.4%). Deep deletion, truncation and nonsense or missense mutations accounted for most of the mutations for PTEN (Fig. 1A). We further explored the mutations of PTEN in samples with iCCA indications. Among the 1312 samples, 4% carried mutations of PTEN and 4.9% with PIK3CA mutations, ranking among the topmost frequently mutated genes. The mutation rate of PTEN was dramatically higher (7%) in samples that were clinically defined as both HCC and iCCA (190 HCC+iCCA; 49 iCCA+HCC), consistent with the development of dual lineage tumors observed in the LiPten livers [13, 29].

Fig. 1. Mutation of PTEN is associated with mixed tumor phenotype with iCCA indications and poor tumor staging.

Fig. 1

Open in a new tab

A Mutation counts for PTEN in liver and biliary tract cancer samples accessed through cBioportal. Samples are from 30 studies including hepatobiliary cancer (84.2%), hepatocellular carcinoma (HCC, 6.2%), cholangiocarcinoma (CCA, 1.5%) and cancers of the biliary tract (0.4%) for a total of 5519 patients and 5,648 samples. B Correlation analysis of PTEN mutation with tumor staging and histology scores.

Using the American Joint Committee on Cancer TNM staging system, we observed that PTEN mutation was associated with higher tumor staging, particularly with tumors that presented distant metastasis (stage IV) and those with lymph node involvement (Stage IIIC) (Fig. 1B). PTEN mutation was also associated with the lowest grade tumors (stage I), whereas those with middle stages had a lower frequency of PTEN mutations. Thus, PTEN mutation may allow cells to gain transformation advantage that permits tumor initiation as well as additional metastasis potential. Consistently, more microsatellite instability (MSI) was observed in samples with PTEN mutation vs. those with wild-type PTEN. Histology grading of the samples indicates that PTEN mutation was associated with more undifferentiated or poorly differentiated tumors (Fig. 1B). Together, these patient data analyses suggest that PTEN may play a crucial role in maintaining hepatocytes and cholangiocytes in differentiated states to prevent tumor development.

PTEN loss in the liver parenchyma leads to heterogeneous tumors

We showed previously that the LiPten mice develop dual lineage tumors when deletion of Pten is targeted to albumin-expressing cells [14, 15]. Histologic examination by hematoxylin and eosin (H&E) staining revealed tumor regions characterized by the presence of both trabecular hepatocellular and glandular biliary-type malignant architectures (Fig. 2B, C). Cytokeratin (CK) and Arginase1 (Arg1), two highly specific markers for cholangiocytes and hepatocytes respectively are used to further define CCA and HCC tumor regions (Fig. 2B, C). Staining for proliferating cell nuclear antigen (PCNA) demonstrated robust nuclear positivity within CK+ and Arg1+ tumor regions with minimum PCNA positivity in control livers (Fig. 2B, C). The HCC tumor regions are also positive for HepPar-1 (Supplementary Fig. 1A). To address whether the tumors originates from embryonic hepatoblasts, we developed two lineage tracing models that target Pten deletion to either mature hepatocytes or cholangiocytes using AAV8-TBG-Cre delivered to 1-month-old PtenloxP/loxP; R26RYFP mice (HepPten) or tamoxifen-induced Cre expression in 1-month old PtenloxP/loxP; R26RYFP; SOX9-CreERT mice (ChoPten) [13] respectively (Fig. 2 and Supplementary Fig. 1). All LiPten and HepPten mice developed tumors at 11-13 months of age with tumor incidence in the ChoPten mice being lower (63.6%) (Fig. 2A). Such reduced tumor incidence observed in ChoPten mice could result from less severe steatotic liver injury, since PTEN is still expressed in hepatocytes which prevents excessive lipid accumulation [13]. Similar to the LiPten mice that develop HCC and iCCA, immunostaining confirmed the development of both HCC and iCCA in HepPten and ChoPten mice though with different penetrations (Fig. 2B, C and Supplementary Fig. 1). Quantification of PCNA staining in CK+ and Arg1+ areas shows robustly increased cell proliferation of tumors in all three models. In addition, both CK+ iCCA and HNF4a+ HCC tumors are positive for the lineage marker yellow fluorescent protein (YFP), confirming they are derived from PTEN-targeted cells (Supplementary Fig. 1). These findings collectively suggest that PTEN loss promotes the mixed lineage tumors regardless of the cell types that are targeted for Pten deletion. Together, these genetic interrogations suggest that PTEN and the PTEN-regulated signals play critical roles in maintaining cellular plasticity of the liver cells.

Fig. 2. PTEN loss in liver parenchymal cells leads to heterogeneous tumors.

Fig. 2

Open in a new tab

A Schematics for generation of Pten deletion mouse models. LiPten is generated by crossing PtenloxP/loxP mice with Alb-Cre+ mice where albumin expression starts on embryonic day E12.5 and cells targeted include hepatocytes and hepatoblasts. HepPten is generated by injection of AAV8-TGB-Cre virus into the PtenloxP/loxP; YFPlox-Stop-lox mice at 1 month of age, which targets only mature hepatocytes. ChoPten mice are generated by tamoxifen injection at 1 month of age in PtenloxP/loxP; YFPlox-Stop-lox Sox9CreERT mice, which targets mature cholangiocytes. Representative H&E staining of liver tissue from 12-month-old respective mouse models showing HCC and CCA morphology (Top panels). Bottom panels show immunofluorescent co-staining of the proliferation marker PCNA (magenta) with CK (green) that marks the iCCA regions B or Arginase-1 (green) that marks the HCC regions C, with DAPI (blue) counterstaining nuclei. The percentage of PCNA⁺ cells among CK⁺ B or Arginase-1⁺ C cells was quantified and shown to the right of the images. Scale bar 50 μm. * p < 0.05, ** p < 0.01.

Loss of AKT2 rescues oncogenesis in livers lacking PTEN

We next explored the involvement of PI3K/AKT pathways in this PTEN regulated dual lineage tumor development using mouse models carrying deletions of Pten and Akt2, the most abundant Akt isoforms expressed in the liver [30, 31]. We showed previously that germline loss of AKT2 inhibits tumor development in the LiPten livers and that this result is due to the metabolic effect of AKT2 loss on liver steatosis and associated injury [14, 20, 32]. More recently, our analysis showed that AKT2 but not AKT1 regulates the tumor-associated osteopontin signal in hepatocytes [26]. To address whether AKT2 may directly regulate the PTEN-controlled oncogenesis, we deleted the liver Akt2 in LiPten mice (LiPtenA2, Akt2loxP/loxP; PtenloxP/loxP; Alb-Cre+) (Fig. 3Aand Supplementary Fig. 2A) and showed that loss of AKT2 in the liver arrested the Pten deletion-induced tumorigenesis (Fig. 3B, C). At 12 months of age, only benign fluid-filled cysts were observed in the LiPtenA2 livers, and they do not progress to malignancy up to 16 months of age (Fig. 3B, C). Quantitative analysis shows significantly reduced PCNA staining in LiPtenA2 vs. LiPten livers (Fig. 3D). Immunostaining for cytokeratin19 and glutamine synthase (GS) supports gross and morphology analysis showing dilated ducts but no tumors in the LiPtenA2 livers. Sirius red staining further demonstrated diminished fibrosis (Fig. 3C). KEGG and WIKI pathway enrichment analysis of proteomics analysis further confirms that cancer-related pathways are downregulated in the LiPtenA2 vs. LiPten livers (Supplementary Fig. 2B).

Fig. 3. Loss of AKT2 rescues oncogenesis in livers lacking PTEN.

Fig. 3

Open in a new tab

A Immunoblotting of liver tissue lysates verified PTEN and AKT status in the respective mouse models. Blots are cropped for clarity. B Tumor spectrum of WT, LiPten, and LiPtenA2 mice from 3-16 months. Each square represents one mouse. Filled squares indicate tumor-bearing mice while open squares are mice without tumors. C Representative images of the 12-month-old mouse livers from the respective genotypes. From top to bottom, gross morphology (arrows, tumor nodules; asterisks, cysts), H&E staining, and immunohistochemistry for CK19 (brown), Glutamine Synthesis (GS, brown), and Sirius red (red). Scale bar 50 μm. D Immunofluorescent co-staining of the proliferation marker PCNA (magenta) with DAPI (blue) counterstaining nuclei. The percentage of PCNA⁺ cells among DAPI⁺ cells was quantified and shown below the images. Scale bar 50 μm. E Sphere formation using immortalized liver cells established from WT, LiPten (Pten-/-) and LiPtenA2 (Pten-/-; AKT2-/-) mice. 300,000 cells per 60 mm dishes seeded. Representative images of day-3 spheres (top) and quantification (bottom) are shown below the images. n = 3. Scale bar 100 μm. Experiment repeated 3 times. ** p < 0.01, *** p < 0.001.

To further explore this pro-tumorigenic role of AKT2, we performed sphere-forming assay with immortalized hepatocytes established from WT, LiPten and LiPtenA2 livers to assess their self-renewal and stemness properties in anchorage-independent, non-adherent 3D conditions. While WT hepatocytes rarely form spheres (>100μM), Pten-/- hepatocytes readily formed spheres. Pten-/-; Akt2-/- hepatocytes formed significantly fewer spheres compared to Pten-/- hepatocytes. (Fig. 3E). Together, these data suggest that AKT2 and AKT2-regulated functions serve as critical regulators in malignant transformation in response to PTEN loss.

Consistent with a role of PTEN/PI3K/AKT signal in CCA development, the expression of all 4 PIK3 genes was higher in iCCA vs. their respective non-tumor controls with PIK3CA and PIK3CB being significantly upregulated in iCCA (Supplementary Fig. 3A). A similar trend was observed for expressions of the AKT isoforms and their targets (Supplementary Fig. 3A, B) with the expression of AKT2 being significantly upregulated in iCCA. Expression of AKT3 is also significantly higher though AKT3 expression is low in both normal and tumor liver tissues [30, 31]. Furthermore, expression of PTEN negatively correlated with overall survival of iCCA cohort (Supplementary Fig. 3C). Interestingly, these markers do not differ between tumor vs. non-turmors in HCC. Together with the phenotype observed with LiPten and the LiPtenA2 mice (Figs. 2, 3), these analyses suggest that the PTEN-regulated PI3K/AKT2 signaling may play a more critical role in iCCA and cHCC-CCA development than in HCC.

NOTCH-SOX9 signal is induced in iCCA-HCC developed with PTEN loss

Exploring how PTEN/AKT2-regulated signals promote iCCA and cHCC-CCA, we analyze the gene expression data from LiPten and LiPten livers lacking AKT2 (A2LiPten) that we previously generated (GSE70501). In the LiPten livers, we showed previously that Notch signals in the LiPten livers is robustly induced compared with control livers [33]. Here, we show that AKT2 loss reverse the upregulation of NOTCH signal observed in LiPten livers (Fig. 4A). NOTCH signal is critical for ductal development during embryogenesis and abnormal activation of NOTCH signal has been shown to induce the transformation of hepatocytes to malignant cholangiocytes and promote CCA development [34, 35]. Similar to our observation for PIK3 and AKT isoforms (Supplementary Fig. 3), expression of NOTCH signaling molecules is also upregulated in patient samples of iCCA and not HCC (Supplementary Fig. 4A). Additionally, the expression of NOTCH signaling molecules, including NOTCH1 and JAGGED1, correlated strongly with that of molecules within the PI3K/AKT signaling pathway (Supplementary Fig. 4B).

Fig. 4. NOTCH signal is induced in iCCA-HCC developed with PTEN loss.

Fig. 4

Open in a new tab

A Analysis of Core genes within Notch Signaling Pathway in 15-month-old LiPten, A2LiPten vs. control mice (GSE70681). Left, PCA plot showing gene expression in A2LiPten livers resembles that of Controls and not LiPten livers. Middle, Top differentially regulated pathways induced by PTEN loss and their reversal in A2LiPten livers. Right, heatmap for core Notch genes in the mouse livers. B Top, immunofluorescent staining of NICD (Notch intracellular cleaved domain, green) and Jag1, a Notch ligand (magenta) in the LiPten, HepPten and ChoPten livers. Tissues are co-stained with DAPI for nuclei (Blue). Bottom, qPCR analysis showing relative expression of Hes1 and Notch1 in the LiPten (n = 6, left), HepPten (n = 3, middle) and ChoPten (n = 3, right) livers. C Top, immunofluorescent staining of NICD (green) and Jag1 (magenta) showing minimum signal in LiPtenA2 samples. Tissues are co-stained with DAPI for nuclei (blue). Bottom, immunoblotting and qPCR analysis showing reduced Notch target Hes1 and Notch1 expression in LiPtenA2 vs. LiPten livers. Blots are cropped for clarity. n = 5. Scale bar 50 μm. * p < 0.05, ** p < 0.01, *** p < 0.001.

We performed immunostaining of JAG1 and NICD, the cleaved and activated NOTCH intracellular domain, in the livers of the three tumor models (LiPten, HepPten and ChoPten) (Fig. 4B). Our data showed that JAG1 and NICD were robustly detected in tumors developed in these livers. Upregulated mRNA expression of Notch 1 and its target gene Hes1 supported this finding (Fig. 4B). In the LiPtenA2 livers, NICD is not detected and JAG1 was present only in the cystic ducts (Fig. 4C). Furthermore, loss of AKT2 significantly inhibited the expression of Hes1 and Notch1 induced by deletion of Pten (Fig. 4C). HES1 protein was also lower in LiPtenA2 livers than that observed in LiPten livers which is higher than controls. Thus, NOTCH activation may play a role in the iCCA and iCCA-HCC tumors driven by PTEN loss and activation of PI3K/AKT2.

Previous work has suggested that NOTCH cascade drives CCA development [34, 35] and that SOX9 may be involved in this effect [34, 36]. Our co-expression analysis showed a strong correlation of SOX9 with NOTCH1 and JAG1 in TCGA liver cancer samples (Supplementary Fig. 5A). In tumors developed in our murine models– LiPten, ChoPten, and HepPten - SOX9 staining was detected, particularly in iCCA regions (Fig. 5A, B). Upregulation of pAKT and SOX9 was also observed when PTEN proteins were lost (HepPten and LiPten) or diminished (in ChoPten livers due to intact PTEN in untargeted hepatocytes) (Fig. 5B). Aligning with the lack of tumor development, SOX9 expression was significantly lower in the LiPtenA2 livers (Fig. 5C). Histological analysis showed that the macroscopic cysts developed in the LiPtenA2 livers were composed of dilated but well-organized ductal structures (Fig. 5D). Expression of SOX9 is localized only in cholangiocytes within these organized ductal structures in the LiPtenA2 livers vs. the randomly scattered pattern observed in the LiPten livers (Fig. 5D).

Fig. 5. Induction of SOX9 by PTEN loss and its rescue by AKT2 loss.

Fig. 5

Open in a new tab

A Representative images of the 12-month mouse livers from LiPten (left), HepPten (middle) and ChoPten (right) mice stained with SOX9 (magenta) and CK (green). Tissues are co-stained with DAPI for nuclei (blue). B Immunoblotting analysis of the 12-month mouse livers showing elevated SOX9 in LiPten, HepPten and ChoPten mouse livers. C Left, immunoblotting showing induction of SOX9 in LiPten and its inhibition in livers from LiPtenA2 mice. Blots are cropped for clarity. p-AKT (S473 and T308) shows AKT phosphorylation status that correlates with its activity as indicated by p-Pras40 (T246). Right, qPCR analysis of the 12-month mouse liver showing induction of Sox9 and two other LCSC markers in the LiPten livers and its reduction in LiPtenA2 livers. n = 5. * p < 0.05, ** p < 0.01, *** p < 0.001. D Representative H&E and immunofluorescent images showing induction and disorganized distribution of SOX9 (magenta) in LiPten livers and reduced expression and its limited expression in the organized ductal structure in LiPtenA2 livers. CK (green) is stained to show cholangiocytes and ductal structures. DAPI (blue) is stained to indicate nuclei. E Immunofluorescent staining of SOX9 (magenta) protein expression in patient tumor samples vs. adjacent non-tumor tissues with DAPI (blue) counterstaining nuclei. The percentage of SOX9⁺ among DAPI⁺ was quantified. Dotted lines connect samples from the same patient. NT non-tumor, T tumor. Scale bar 50 μm.

In patient samples, immunostaining shows that SOX9 was highly expressed in tumor samples vs. adjacent tissues (Fig. 5E). In the TCGA data repository, SOX9 expression was more significantly induced in iCCA vs. HCC (Supplementary Fig. 5B). Expression of SOX9 also negatively correlates with disease-free patient survival in iCCA (HR = 1.6) but not HCC (HR = 0.9) (Supplementary Fig. 5C). Together, these data suggest that induction of SOX9 likely plays a role in the tumor transformation process regulated by PTEN/AKT2.

NOTCH-SOX9 signal partially regulates tumor cell transformation controlled by PTEN-AKT2

Accumulating evidence suggests that SOX9 serves as a liver progenitor cell marker and that SOX9+ cells are capable of regenerating both hepatocytes and cholangiocytes after injury [37–39]. Supporting the role of SOX9 in sustaining LCSC state, expression of LCSC markers EpCAM [38] and CD133 (ref. [12]) strongly correlated with that of SOX9 in TCGA samples (Supplementary Fig. 6A). Consistent with the regulation of SOX9 by PTEN-AKT2 signal, expressions of Epcam and Prom1 are robustly induced in the LiPten livers while AKT2 loss attenuated their expression significantly (Fig. 5C). To test the potential of SOX9 in LCSC transformation, we knocked down Sox9 and showed inhibiting Sox9 expression resulted in reduced sphere and colony formation ability using Huh7 (ref.22) and PLC/PRF/5 (ref.23) cells, which both exhibit LCSC characteristics (Fig. 6Aand Supplementary Fig. 6B). Thus, these data indicated that the presence of SOX9 indeed sustains a LCSC phenotype. In the Pten-deleted hepatocytes [25] and cholangiocytes [24], we also observed fewer spheres and colonies when Sox9 was knocked down (Fig. 6B, C). Inhibiting SOX9 expression also reduced the expression of LCSC markers (Fig. 6D) such as PROM1 and EpCAM [12, 38]. Together, these data suggest that PTEN loss permits the LCSC transformation, with induction of SOX9 playing a role in this process.

Fig. 6. SOX9 regulates the transformation of LCSC.

Fig. 6

Open in a new tab

A Serial sphere formation was performed in Huh7 stem-like cells with the expression of siSox9 or control siRNA. Spheres are broken up and passaged every 5 days and each passage counts as a series. Left, representative image of the spheres. Right, quantification of the spheres per 25,000 cells plated in a 24-well plate. B Colony formation assay performed using WT and Pten-/- immortalized cells with or without siSox9 transfection. Left, immunoblotting showing reduced levels of SOX9, blots are cropped for clarity; Middle, representative image of the colonies; Right, quantification of colonies. n = 3. Experiment was performed twice. C Sphere formation assay performed using mouse immortalized cholangiocytes lacking PTEN with or without siSOX9 introduction (n = 3). Left, representative image. Right, quantification of the spheres. D qPCR analysis showing down-regulation of stem cell markers with shSOX9 transfected in HepG2 cells. n = 3. Scale bar 100 μm. * p < 0.05, ** p < 0.01.

During ductal plate development, NOTCH signal is necessary to program the cholangiocyte fate partially by inducing the upregulation of SOX9. Gain-of-function experiments showed that ectopic expression of NICD or exposure to JAG1 ligand-coated extracellular matrix both induced the expression of SOX9 (Fig. 7A). Thus, NOTCH signal indeed positively regulates the expression of SOX9 in liver cancer cells. We next tested if inhibition of NOTCH can attenuate tumor cell transformation due to PTEN-AKT2 alteration using sphere formation in Pten-/- mouse liver tumor cells treated with or without DAPT (Fig. 7B, C), the γ-secretase inhibitor that blocks NOTCH Signaling [40]. As expected, DAPT treatment time-dependently attenuated the expression of both HES-1 and SOX9 in WT cells. The Pten-/- cells displayed a more moderate response to DAPT with faster recovery. Functional analysis shows that Pten-/- cells formed larger and more spheres than the WT cells (Fig. 7C). Consistent with their HES-1 and SOX9 signal response, while significantly reducing sphere formation in WT cells, DAPT treatment is less effective at doing so in the Pten-/- cells.

Fig. 7. NOTCH-SOX9 signal partially regulates tumor cell transformation controlled by PTEN-AKT2.

Fig. 7

Open in a new tab

A Notch intracellular domain (NICD) was transfected to Huh7 cells. Expression of SOX9 and HES1 was analyzed with Vinculin as a loading control. Blots are cropped for clarity. Right, Huh7 cells were exposed to Jag1 ligand- or IgG control-coated extracellular matrix and expression of SOX9 was quantified using qPCR. n = 6. B Immunoblot analysis of SOX9 and HES1 expression in WT and Pten⁻/⁻ immortalized liver cells treated with 20 μM DAPT for 1, 2, 3, 6, and 24 h. DAPT-induced downregulation of SOX9 and HES1 in WT is partially regulated in Pten⁻/⁻ cells. C Sphere formation of WT and Pten-/- immortalized cells treated with or without 20 μM DAPT (300,000 cells seeded per 60 mm dishes, day 2). Top, representative images of the spheres. Bottom, quantification of the sphere numbers. n = 3. Experiment repeated 3 times. D Sphere formation of WT, Pten-/- and Pten-/-; Akt2-/- immortalized cells with or without NICD transfection (10,000 cells seeded per 24-well plates, day 2). Top, representative images of the spheres; Bottom, quantification of the sphere numbers. n = 3. The experiment was repeated 3 times. Scale bar 100 μm. * p < 0.05, ** p < 0.01, *** p < 0.001.

We next explored if introducing NICD could rescue the inability of sphere formation due to AKT2 loss (Fig. 7Dand Supplementary Fig. 7). Introduction of NICD indeed induced sphere formations in all three cell lines. Compared with the Pten-/- cells, Pten-/-; Akt2-/- cells formed significantly smaller and less spheres. Even though NICD introduction is able to rescue some of the sphere forming ability of the Pten-/-; Akt2-/- cells, it was not able to restore the full sphere formaing ability to the level of the Pten-/- cells. Together, these data shows that while inhibiting NOTCH activation blocks PTEN loss induced LCSC transformation, activating NOTCH itself is not sufficient to drive LCSC transformation when AKT2 signal is absent. Thus, NOTCH signal may only be partially responsible for the PTEN-AKT2-regulated LCSC transformation.

PTEN loss sensitizes tumor cells to TGFβ regulated LCSC repression

To determine the signals regulated by PTEN/AKT2 that intersect with the NOTCH–SOX9 axis during tumor transformation, we next examined the role of TGFβ signaling which was among the top signal pathways regulated by both PTEN and AKT2 (Fig. 4A). We showed previously that combining loss of PTEN and SMAD4, the co-SMAD for TGFβ signal, accelerates tumor development and shifts tumor identity toward iCCA [25], implicating TGFβ signaling as a modifier of lineage fate downstream of PTEN loss. Previous studies have shown that TGFβ induces Sox9 expression during ductal development [41, 42]. However, in bipotent progenitor models including hepatoblast-like cells, TGF-β was reported to promote hepatocyte differentiation while suppressing biliary markers, indicating context-dependent effects [41–43].

In the Pten-/- cells, we observed significant induction of SMAD4, concurrent with the upregulation of SOX9 (Fig. 8A). The levels of both were reduced in the Pten-/-; Akt2-/- cells. All three cell lines are responsive to TGFβ treatment as indicated by the upregulation of p-SMAD3 protein and induction of Ccn2 and Serpine1 mRNA expression (Fig. 8A).

Fig. 8. PTEN loss sensitizes liver tumor cells to TGFβ treatment.

Fig. 8

Open in a new tab

A Top, immunoblotting showing differently regulated SOX9 and SMAD levels in WT, Pten-/- and Pten-/-; Akt2-/- immortalized cells with or without TGFβ treatment. Bottom, qPCR analysis showing induction of TGFβ target gene Ccn2 and Serpine1 after TGFβ treatment. B Promoter analysis of SOX9 showing binding site for SMAD3, SMAD4 and RBPj SMAD3 and RBPj bind to the same site, where H3K27Ac marks a transcriptional activation site. SMAD4 binds to a different site. An H3K9Me site is adjacent to the SMAD4 binding site. C Immunoblotting analysis of spheres formed in WT, Pten-/- and Pten-/-; Akt2-/- immortalized cells treated with TGFβ, DAPT or both. TGFβ inhibits SOX9 in Pten-/- cells whereas DAPT inhibits SOX9 in WT cells. Left, representative blots. Blots are cropped for clarity. Right, Quantification of relative protein abundance. D Sphere formation in WT, Pten-/- and Pten-/-; Akt2-/- immortalized cells treated with TGFβ, DAPT or both (5000 cells seeded per 24-well plates). Left, representative images. Right, quantification of sphere numbers. Scale bar 100 μm. n = 6. Experiment repeated 3 times. E Proposed schematics for how PTEN-AKT2 signal regulates the response of LCSC to TGFβ through the regulation of SOX9. * p < 0.05, ** p < 0.01, *** p < 0.001.

Here, TGFβ regulation of SOX9 expression appears to depend on PTEN status. Specifically, TGFβ induced SOX9 expression in WT cells but suppressed it in Pten⁻/⁻ cells concurrent with the reduction of SMAD4 (Fig. 8Aand Supplementary Fig. 8A), indicating that PTEN loss fundamentally alters the cellular response to TGFβ. Consistent with this SOX9 response, TGFβ treatment robustly and significantly reduced sphere formation in Pten⁻/⁻ cells, whereas WT cells were largely unresponsive (Supplementary Fig. 8B). Deletion of Akt2 reduced SMAD4 expression and reversed the SOX9 and sphere formation response in Pten-/-; Akt2-/- cells (Fig. 8Aand Supplementary Fig. 8B). These data indicate that PTEN-AKT2 signal may alter TGFβ signal response to regulate the expression of SOX9 and tumor initiation capacity of the tumor cells.

To investigate the molecular basis by which TGFβ regulates SOX9 expression, we analyzed the SOX9 promoter using ENCODE datasets [21] from HepG2 cells. This analysis revealed binding of both SMAD3 and SMAD4 within 1 kb of the SOX9 transcriptional start site (Fig. 8B), supporting direct regulation by TGFβ signaling. Additionally, RBPj, the transcriptional effector of NOTCH signaling, co-localizes with SMAD3 at sites across TGFβ-responsive genes including Sox9, supporting crosstalk of the two signaling pathways (Fig. 8Band Supplementary Fig. 8C). Thus, the elevated SMAD4 in Pten⁻/⁻ cells may enhance NOTCH responsiveness by stabilizing transcriptional complexes and promoting chromatin accessibility at NOTCH target loci [44, 45], thereby amplifying lineage-specific gene expression programs. Supporting this notion, treatment of WT cells with TGFβ prevented DAPT from inhibiting the expression of SOX9 (Fig. 8B). On the other hand, TGFβ treatment appears to induce SOX9 with or without DAPT in WT cells, indicating this effect of TGFβ is independent of NOTCH (Fig. 8C). In fact, TGFβ more strongly induced SOX9 when NOTCH signal is blocked, suggest potential competition of the two signals. When PTEN is lost, TGFβ treatment has the opposite effect on SOX9 expression than what was observed in WT cells. TGFβ robustly inhibited the expression of SOX9 with or without DAPT (Fig. 8Cand Supplementary Fig. D). AKT2 loss, which reduced SMAD4 expression and attenuated SOX9 expression (Fig. 8A), rendered the cells less responsive to TGFβ though it did not restore its responsive to DAPT (Fig. 8C). Together, these data suggest that TGFβ may act as redundant signal for NOTCH regulation of SOX9 when PTEN is intact; hence, inhibition of NOTCH permits TGFβ to induce SOX9. When PTEN is lost and SMAD4 is induced, TGFβ signal becomes more dominant. At the same time, the chronic upregulation of SMAD4 turns TGFβ signal to inhibitory on the expression of SOX9.

Together, these molecular analyses suggest that the Pten-/- tumor cells may be more vulnerable to the tumor-inhibitory effects of TGFβ, due to its upregulation of SMAD4. To address this, we subjected the different cell lines to sphere formation with or without TGFβ or DAPT or both. Consistent with the inability of DAPT to reduce SOX9 levels in Pten⁻/⁻ cells, this treatment only moderately reduced sphere formation. WT cells, on the other hand, are sensitive to DAPT. TGFβ, while ineffective at reducing sphere formation in WT cells, robustly and significantly reduced sphere number and size in Pten⁻/⁻ cells (Fig. 8D). Cells lacking both PTEN and AKT2 form very few spheres and are less responsive to either treatment, consistent with reduced overall signal response to TGFβ and NOTCH pathway interrogation (Fig. 8C).

Together, these findings support a model in which cells with intact PTEN maintain progenitor or cholangiocytic identity with Notch whereas TGFβ signal acts as backup signal. In contrast, PTEN loss—via AKT2—potentiates NOTCH–SOX9 signaling to drive LCSC transformation, while simultaneously sensitizing cells to TGFβ-mediated repression, likely through altered SMAD4-dependent chromatin regulation (Fig. 8E). When AKT2 is absent, response to NOTCH and TGFβ are both blunted, likely due to the effect of AKT2 on sustaining NOTCH and SMAD4 signal input. Here, TGFβ is not able to either induce or inhibit the expression of SOX9, suggesting that TGFβ effect on SOX9-be it positive or negative may be dependent on some basal level activity of NOTCH or at least other signals enforced by the presence of AKT2.

Discussion

In this study, we report that PTEN, through AKT2, plays a crucial role in regulating the development of mixed lineage tumor phenotypes. We showed that (1) loss of PTEN in various parenchymal cells – albumin-positive, SOX9-positive, or mature hepatocytes- results in iCCA-HCC cancer development regardless of cell origin; (2) deletion of Akt2 blocks the PTEN-regulated mixed-lineage cancer development; (3) SOX9, a NOTCH-induced transcriptional factor, regulates LCSC transformation downstream of PTEN/AKT2; and (4) In the absence of PTEN, cells become sensitized to TGFβ regulated inhibition of LCSC transformation. Together, these data established the role of the PTEN/AKT2 signal in mixed lineage liver cancer development and delineated TGFβ regulated pathways that contribute to this development (Fig. 8E).

Primary liver cancer arises from either hepatocytic or biliary lineage cells, giving rise to HCC or iCCA, respectively. Although traditionally considered distinct entities, accumulating molecular and histological evidence supports a continuum of lineage plasticity across primary liver tumors, exemplified by the aggressive combined cHCC-CCA, which displays features of both lineages within the same lesion [46, 47]. Using a lineage tracing approach, we established that the PTEN signal plays critical roles in the cHCC-CCA development. Without PTEN, cHCC-CCA develops irrespective of the cell type that deletion of Pten is targeted to, indicating that PTEN loss overrides lineage-specific constraints. Importantly, this process requires AKT2, identifying it as a critical oncogenic effector downstream of PTEN. AKT2 is the most abundantly expressed AKT isoform in the liver [30, 31]. We have previously reported that germline AKT2 loss inhibits tumor development in the LiPten mice [14]. We showed that AKT2-regulated lipid biosynthesis produces a tumor microenvironment to promote the oncogenic events induced by PTEN loss [29, 32]. However, AKTs were found to be dispensable for chemical-induced liver carcinogenesis though cell context differences appear to determine how each AKT isoforms function [31]. We established here that AKT2 indeed plays a role in hepatocyte-cholangiocyte cell fate regulated by PTEN. Whether this function depends on it being a metabolic kinase remains to be determined. Given that activation of AKT is an unequivocal consequence observed with PTEN loss of function [48], our current work highlights a context-specific requirement for AKT2 in liver tumorigenesis and suggests that AKT isoform dependency may underlie divergent outcomes observed in other models of liver cancer. Future work is needed to directly compare the effects of each AKT isoform to decipher their unique cell-specific functions.

Molecularly, cHCC-CCA may originate from LCSCs capable of dual hepatocytic and cholangiocytic differentiation [46], or through transdifferentiation of mature hepatocytes into cholangiocyte-like cells, a process often triggered by chronic injury or oncogenic signaling [49]. These lineage conversions are orchestrated by key developmental pathways— including aberrant activation of developmental signaling networks including NOTCH and WNT/β-CATENIN— which enable the emergence of tumors containing both hepatocytic and cholangiocytic components [49–51]. Here, we report that loss of PTEN induces the upregulation of NOTCH signal, concomitant with the induction of WNT/β-CATENIN that we characterized previously [20], reinforcing a stemness-promoting transcriptional program. In addition to NOTCH and WNT signals, we identified TGFβ as a key context-dependent regulator of lineage fate decisions downstream of PTEN loss. While TGFβ can induce the expression of SOX9 during biliary differentiation [52], we show that sustained upregulation of SMAD4 in PTEN-deficient cells permits TGFβ to repress SOX9 and inhibit SOX9-driven LCSC transformation. These results suggest that TGFβ and NOTCH converge onto SOX9 to control LCSC and that PTEN loss fundamentally alters cellular responsiveness and sensitizes tumor cells to TGFβ mediated repression rather than NOTCH inhibition. Mechanistically, chronic upregulation of SMAD4 in the Pten-/- cells may promote recruitment of repressive SMAD4 complexes to the SOX9 promoter, facilitating a shift from an open to a closed chromatin configuration [44, 45] and converting TGFβ signaling from a pro–stemness to an anti–stemness program (Fig. 8E). Collectively, these findings define PTEN–AKT2 signaling as a critical determinant of lineage plasticity in liver cancer and reveal a previously unrecognized mechanism by which developmental signaling pathways are rewired to drive mixed-lineage tumor development.

In summary, our study demonstrates that PTEN–PI3K–AKT2 signaling is a key regulator of hepatocyte–cholangiocyte fate determination and a central driver of cHCC–CCA development. We identify AKT2 as the critical AKT isoform mediating this mixed-lineage tumor phenotype, providing a mechanistic rationale for targeting AKT signaling in iCCA and mixed lineage liver cancer. Furthermore, we uncover a PTEN-dependent crosstalk between TGFβ and NOTCH signaling that converges on SOX9 to regulate tumor lineage plasticity and show that PTEN loss sensitizes tumor cells to TGFβ-mediated repression. Together, these findings define a molecular framework linking tumor suppressor loss, developmental signaling, and lineage plasticity, with important implications for therapeutic strategies in aggressive liver cancers.

Supplementary information

Supplemental Methods (48.6KB, pdf)
supplemental Figures (35.9MB, pdf)

Acknowledgements

Dr. Stiles acknowledges grant funding support from NIH NCI CA154986 and NIH NIDDK R01DK131492. Dr. Stiles also acknowledge funding support from USC Mann School of Pharmacy for this project. Drs. Stiles and Yuan acknowledge funding support from USC Norris Cancer Center.

Author contributions

QT: study design, protocol/method development, model development, conducting research and data collection, extracting and analyzing data, interpreting results, manuscript writing and revision, updating reference lists. LS, JC&NZ: study designing, protocol/method development, model development, conducting the search. YG, LH, SH, DA, ZX, BH, GZ, PN, MA, AB, JL, &SK: conducting experiments and data collection. JX, BJ, SCL, JVE &LY, providing resources/samples and manuscript revision. SC&GK, pathology analysis and manuscript revision. BLS: study designing, protocol/method development, model development, data analysis, interpreting results, manuscript writing and revision, funding and project resources.

Funding

Open access funding provided by SCELC, Statewide California Electronic Library Consortium.

Data availability

The data generated in this study are available within the article and its supplementary data files. Data sourced from external databases are available at the public repositories indicated in the results section. The mouse liver proteomic dataset analyzed in this study will be uploaded to protein data commons upon acceptance of this manuscript.

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.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41388-026-03960-0.

References

  • 1.Valle JW, Lamarca A, Goyal L, Barriuso J, Zhu AX. REVIEW | New horizons for precision medicine in biliary tract cancers. Cancer Discov. 2017;7:943–62. 10.1158/2159-8290.CD-17-0245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Lowery MA, Ptashkin R, Jordan E, Berger MF, Zehir A, Capanu M, et al. Comprehensive molecular profiling of intra- and extrahepatic cholangiocarcinomas: potential targets for intervention. Clin Cancer Res. 2018;24:4154–61. 10.1158/1078-0432.CCR-18-0078. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Kokuryo T, Yokoyama Y, Nagino M. Recent advances in cancer stem cell research for cholangiocarcinoma. J Hepato-Biliary-Pancreat Sci. 2012;19:606–13. 10.1007/s00534-012-0542-6. [DOI] [PubMed] [Google Scholar]
  • 4.Surveillance, Epidemiology, and End Results (SEER) Program. SEER*Stat Database: Incidence - SEER Research Data, 8 Registries, Nov 2024 Sub (1975-2022) [Research database] [Internet]. National Cancer Institute, DCCPS, Surveillance Research Program; 2025 [cited 2025 Nov 11]. Available from: https://seer.cancer.gov/seerstat/
  • 5.Rahib L, Wehner MR, Matrisian LM, Nead KT. Estimated Projection of US Cancer Incidence and Death to 2040. JAMA Netw Open. 2021;4:e214708. 10.1001/jamanetworkopen.2021.4708. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Mody K, Jain P, El-Refai SM, Azad NS, Zabransky DJ, Baretti M, et al. Clinical, Genomic, and Transcriptomic Data Profiling of Biliary Tract Cancer Reveals Subtype-Specific Immune Signatures. JCO Precis Oncol. 2022;:e2100510. 10.1200/PO.21.00510 [DOI] [PMC free article] [PubMed]
  • 7.Nakazawa K, Dobashi Y, Suzuki S, Fujii H, Takeda Y, Ooi A. Amplification and overexpression of c-erbB-2, epidermal growth factor receptor, and c-met in biliary tract cancers. J Pathol. 2005;206:356–65. 10.1002/path.1779. [DOI] [PubMed] [Google Scholar]
  • 8.Clapéron A, Mergey M, Ho-Bouldoires THN, Vignjevic D, Wendum D, Chrétien Y, et al. EGF/EGFR axis contributes to the progression of cholangiocarcinoma through the induction of an epithelial-mesenchymal transition. J Hepatol. 2014;61:325–32. 10.1016/j.jhep.2014.03.033. [DOI] [PubMed] [Google Scholar]
  • 9.Rimini M, Loi E, Fabregat-Franco C, Burgio V, Lonardi S, Niger M, et al. Next-generation sequencing analysis of cholangiocarcinoma identifies distinct IDH1-mutated clusters. Eur J Cancer. 2022;175:299–310. 10.1016/j.ejca.2022.08.026. [DOI] [PubMed] [Google Scholar]
  • 10.Tu T, Chen J, Chen L, Stiles BL. Dual-Specific Protein and Lipid Phosphatase PTEN and Its Biological Functions. Cold Spring Harb Perspect Med. 2020;10:a036301. 10.1101/cshperspect.a036301. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Yothaisong S, Dokduang H, Techasen A, Namwat N, Yongvanit P, Bhudhisawasdi V, et al. Increased activation of PI3K/AKT signaling pathway is associated with cholangiocarcinoma metastasis and PI3K/mTOR inhibition presents a possible therapeutic strategy. Tumour Biol J Int Soc Oncodevelopmental Biol Med. 2013;34:3637–48. 10.1007/s13277-013-0945-2. [DOI] [PubMed] [Google Scholar]
  • 12.Rountree CB, Ding W, He L, Stiles B. Expansion of CD133-Expressing Liver Cancer Stem Cells in Liver-Specific Phosphatase and Tensin Homolog Deleted on Chromosome 10-Deleted Mice. Stem Cells. 2009;27:290–9. 10.1634/stemcells.2008-0332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Chen J, Debebe A, Zeng N, Kopp J, He L, Sander M, et al. Transformation of SOX9+ cells by Pten deletion synergizes with steatotic liver injury to drive development of hepatocellular and cholangiocarcinoma. Sci Rep. 2021;11:11823. 10.1038/s41598-021-90958-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Galicia VA, He L, Dang H, Kanel G, Vendryes C, French BA, et al. Expansion of hepatic tumor progenitor cells in Pten-null mice requires liver injury and is reversed by loss of AKT2. Gastroenterology. 2010;139:2170–82. 10.1053/j.gastro.2010.09.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Stiles B, Wang Y, Stahl A, Bassilian S, Lee WP, Kim YJ, et al. Liver-specific deletion of negative regulator Pten results in fatty liver and insulin hypersensitivity. Proc Natl Acad Sci. 2004;101:2082–7. 10.1073/pnas.0308617100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Kopp JL, Dubois CL, Schaffer AE, Hao E, Shih HP, Seymour PA, et al. Sox9+ ductal cells are multipotent progenitors throughout development but do not produce new endocrine cells in the normal or injured adult pancreas. Development. 2011;138:653–65. 10.1242/dev.056499. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Cho H, Mu J, Kim JK, Thorvaldsen JL, Chu Q, Crenshaw EB, et al. Insulin Resistance and a Diabetes Mellitus-Like Syndrome in Mice Lacking the Protein Kinase Akt2 (PKBβ). Science. 2001;292:1728–31. 10.1126/science.292.5522.1728. [DOI] [PubMed] [Google Scholar]
  • 18.Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2012;2:401–4. 10.1158/2159-8290.CD-12-0095. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Tang Z, Kang B, Li C, Chen T, Zhang Z. GEPIA2: an enhanced web server for large-scale expression profiling and interactive analysis. Nucleic Acids Res. 2019;47:W556–60. 10.1093/nar/gkz430. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Debebe A, Medina V, Chen CY, Mahajan IM, Jia C, Fu D, et al. Wnt/β-catenin activation and macrophage induction during liver cancer development following steatosis. Oncogene. 2017;36:6020–9. 10.1038/onc.2017.207. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Kagda MS, Lam B, Litton C, Small C, Sloan CA, Spragins E, et al. Data navigation on the ENCODE portal. Nat Commun. 2025;16:9592. 10.1038/s41467-025-64343-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Nakabayashi H, Taketa K, Miyano K, Yamane T, Sato J. Growth of human hepatoma cells lines with differentiated functions in chemically defined medium. Cancer Res. 1982;42:3858–63. [PubMed] [Google Scholar]
  • 23.MacNab GM, Alexander JJ, Lecatsas G, Bey EM, Urbanowicz JM. Hepatitis B surface antigen produced by a human hepatoma cell line. Br J Cancer. 1976;34:509–15. 10.1038/bjc.1976.205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Yang Y, Wang J, Wan J, Cheng Q, Cheng Z, Zhou X, et al. PTEN deficiency induces an extrahepatic cholangitis-cholangiocarcinoma continuum via aurora kinase A in mice. J Hepatol. 2024;81:120–34. 10.1016/j.jhep.2024.02.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Xu X, Kobayashi S, Qiao W, Li C, Xiao C, Radaeva S, et al. Induction of intrahepatic cholangiocellular carcinoma by liver-specific disruption of Smad4 and Pten in mice. J Clin Invest. 2006;116:1843–52. 10.1172/JCI27282. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Slarve I, Wang Y, Ding Y, Niu X, Tang Q, Jia C, et al. Isoform specific regulation of osteopontin by AKT2 in hepatocytes and livers. Cell Signal. 2025;132:111799. 10.1016/j.cellsig.2025.111799. [DOI] [PubMed] [Google Scholar]
  • 27.Zeng N, Li Y, He L, Xu X, Galicia V, Deng C, et al. Adaptive basal phosphorylation of eIF2α is responsible for resistance to cellular stress-induced cell death in Pten-null hepatocytes. Mol Cancer Res MCR. 2011;9:1708–17. 10.1158/1541-7786.MCR-11-0299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Peng Z, Aggarwal R, Zeng N, He L, Stiles EX, Debebe A, et al. AKT1 Regulates Endoplasmic Reticulum Stress and Mediates the Adaptive Response of Pancreatic β Cells. Mol Cell Biol. 2020;40:e00031-20. 10.1128/MCB.00031-20. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Tu T, Hong H, Alhousari D, He L, Alba M, Gu Y, et al. Proinflammatory macrophages release CXCL5 to regulate T cell function and limit effects of αPD-1 in steatosis-driven liver cancer. JHEP Rep Innov Hepatol. 2025;7:101385. 10.1016/j.jhepr.2025.101385. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Reyes-Gordillo K, Shah R, Arellanes-Robledo J, Cheng Y, Ibrahim J, Tuma PL. Akt1 and Akt2 Isoforms Play Distinct Roles in Regulating the Development of Inflammation and Fibrosis Associated with Alcoholic Liver Disease. Cells. 2019;8:1337. 10.3390/cells8111337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Wang Q, Yu WN, Chen X, Peng Xding, Jeon SM, Birnbaum MJ, et al. Spontaneous Hepatocellular Carcinoma after the Combined Deletion of Akt Isoforms. Cancer Cell. 2016;29:523–35. 10.1016/j.ccell.2016.02.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.He L, Hou X, Kanel G, Zeng N, Galicia V, Wang Y, et al. The Critical Role of AKT2 in Hepatic Steatosis Induced by PTEN Loss. Am J Pathol. 2010;176:2302–8. 10.2353/ajpath.2010.090931. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Jia C, Medina V, Liu C, He L, Qian D, Tu T, et al. Crosstalk of LKB1-regulated and PTEN-regulated signals in liver morphogenesis and tumor development in mice. Hepatol Commun. 2017;1:153–67. 10.1002/hep4.1027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Hu S, Molina L, Tao J, Liu S, Hassan M, Singh S, et al. NOTCH-YAP1/TEAD-DNMT1 Axis Drives Hepatocyte Reprogramming Into Intrahepatic Cholangiocarcinoma. Gastroenterology. 2022;163:449–65. 10.1053/j.gastro.2022.05.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Cigliano A, Wang J, Chen X, Calvisi DF. Role of the Notch signaling in cholangiocarcinoma. Expert Opin Ther Targets. 2017;21:471–83. 10.1080/14728222.2017.1310842. [DOI] [PubMed] [Google Scholar]
  • 36.Zhao J, Patel J, Kaur S, Sim SL, Wong HY, Styke C, et al. Sox9 and Rbpj differentially regulate endothelial to mesenchymal transition and wound scarring in murine endovascular progenitors. Nat Commun. 2021;12:2564. 10.1038/s41467-021-22717-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Font-Burgada J, Shalapour S, Ramaswamy S, Hsueh B, Rossell D, Umemura A, et al. Hybrid Periportal Hepatocytes Regenerate the Injured Liver without Giving Rise to Cancer. Cell. 2015;162:766–79. 10.1016/j.cell.2015.07.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Lin Y, Zhang F, Zhang L, Chen L, Zheng S. Characteristics of SOX9-positive progenitor-like cells during cholestatic liver regeneration in biliary atresia. Stem Cell Res Ther. 2022;13:114. 10.1186/s13287-022-02795-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Shang T, Jiang T, Cui X, Pan Y, Feng X, Dong L, et al. Diverse functions of SOX9 in liver development and homeostasis and hepatobiliary diseases. Genes Dis. 2023;11:100996. 10.1016/j.gendis.2023.03.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Dong Z, Huo J, Liang A, Chen J, Chen G, Liu D. Gamma-Secretase Inhibitor (DAPT), a potential therapeutic target drug, caused neurotoxicity in planarian regeneration by inhibiting Notch signaling pathway. Sci Total Environ. 2021;781:146735. 10.1016/j.scitotenv.2021.146735. [DOI] [PubMed] [Google Scholar]
  • 41.Antoniou A, Raynaud P, Cordi S, Zong Y, Tronche F, Stanger BZ, et al. Intrahepatic bile ducts develop according to a new mode of tubulogenesis regulated by the transcription factor SOX9. Gastroenterology. 2009;136:2325–33. 10.1053/j.gastro.2009.02.051. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Schaub JR, Huppert KA, Kurial SNT, Hsu BY, Cast AE, Donnelly B, et al. De novo formation of the biliary system by TGFβ-mediated hepatocyte transdifferentiation. Nature. 2018;557:247–51. 10.1038/s41586-018-0075-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Boulter L, Lu WY, Forbes SJ. Differentiation of progenitors in the liver: a matter of local choice. J Clin Invest. 2013;123:1867–73. 10.1172/JCI66026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Suriyamurthy S, Baker D, ten Dijke P, Iyengar PV. Epigenetic Reprogramming of TGF-β Signaling in Breast Cancer. Cancers. 2019;11:726. 10.3390/cancers11050726. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Hill CS. Transcriptional Control by the SMADs. Cold Spring Harb Perspect Biol. 2016;8:a022079. 10.1101/cshperspect.a022079. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Brunt E, Aishima S, Clavien PA, Fowler K, Goodman Z, Gores G, et al. cHCC-CCA: Consensus terminology for primary liver carcinomas with both hepatocytic and cholangiocytic differentation. Hepatol Baltim Md. 2018;68:113–26. 10.1002/hep.29789. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Park HS, Bae JS, Jang KY, Lee JH, Yu HC, Jung JH, et al. Clinicopathologic Study on Combined Hepatocellular Carcinoma and Cholangiocarcinoma: with Emphasis on the Intermediate Cell Morphology. J Korean Med Sci. 2011;26:1023–30. 10.3346/jkms.2011.26.8.1023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Chen CY, Chen J, He L, Stiles BL. PTEN: Tumor Suppressor and Metabolic Regulator. Front Endocrinol. 2018;9:9. 10.3389/fendo.2018.00338. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Sekiya S, Suzuki A. Hepatocytes, rather than cholangiocytes, can be the major source of primitive ductules in the chronically injured mouse liver. Am J Pathol. 2014;184:1468–78. 10.1016/j.ajpath.2014.01.005. [DOI] [PubMed] [Google Scholar]
  • 50.Holczbauer Á, Wangensteen KJ, Shin S. Cellular origins of regenerating liver and hepatocellular carcinoma. JHEP Rep. 2022;4. 10.1016/j.jhepr.2021.100416. [DOI] [PMC free article] [PubMed]
  • 51.Oikawa T. Cancer Stem cells and their cellular origins in primary liver and biliary tract cancers. Hepatol Baltim Md. 2016;64:645–51. 10.1002/hep.28485. [DOI] [PubMed] [Google Scholar]
  • 52.Coricor G, Serra R. TGF-β regulates phosphorylation and stabilization of Sox9 protein in chondrocytes through p38 and Smad dependent mechanisms. Sci Rep. 2016;6:38616. 10.1038/srep38616. [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

Supplemental Methods (48.6KB, pdf)
supplemental Figures (35.9MB, pdf)

Data Availability Statement

The data generated in this study are available within the article and its supplementary data files. Data sourced from external databases are available at the public repositories indicated in the results section. The mouse liver proteomic dataset analyzed in this study will be uploaded to protein data commons upon acceptance of this manuscript.

Articles from Oncogene are provided here courtesy of Nature Publishing Group

RESOURCES