Oxytocin Receptor Regulates the Hippo/YAP Axis to Drive Hepatocarcinogenesis

The FDA-approved oxytocin receptor antagonist atosiban can ameliorate Hippo signaling dysfunction in liver cancer to suppress tumor growth, providing an effective and rapidly translatable therapy for hepatocellular carcinoma.

Abstract

Dysregulation of Hippo signaling, especially the downstream effector YAP, is a critical driver of hepatocellular carcinoma (HCC). Therefore, identifying therapeutic targets to block Hippo signaling could help improve survival outcomes for patients with HCC. In this study, we conducted an unbiased siRNA screen on G protein–coupled receptors targeted by drugs approved in the United States and strongly associated with the Hippo/YAP pathway and identified the oxytocin receptor (OXTR) as an important activator of the Hippo/YAP axis in HCC. The OXTR was correlated with the Hippo gene signature and poor survival outcomes in HCC, and OXTR activation promoted HCC progression through the Hippo/YAP axis. The OXTR antagonist atosiban blocked the growth of HCC in xenograft, patient-derived explant, organoid, and MST1/2 double-knockout mouse models. Molecular studies revealed that activation of the OXTR facilitated YAP dephosphorylation, nuclear accumulation, and transcriptional activation in HCC. OXTR interacted with Gα_q/11 at several important sites (R137, I141, and I227) and induced YAP activation through the Gα_q/11/Rho-associated protein kinase/LATS axis. Chromatin immunoprecipitation assays showed that YAP bound to the enhancer region of the OXTR and facilitated its transcription, creating a positive feedback loop. Together, this study uncovered the interplay between Hippo signaling and the OXTR pathway in hepatocarcinogenesis and established OXTR inhibition with atosiban as a promising strategy for treating HCC.

Significance:

The FDA-approved oxytocin receptor antagonist atosiban can ameliorate Hippo signaling dysfunction in liver cancer to suppress tumor growth, providing an effective and rapidly translatable therapy for hepatocellular carcinoma.

Graphical Abstract

Introduction

Liver cancer is one of the most commonly diagnosed malignancies worldwide, and hepatocellular carcinoma (HCC) is the major subtype (1, 2). Although HCC can be treated with surgery, its overall prognosis is still poor, making it an urgent public health challenge (3). Recent genome-wide association studies and genetic engineering models have demonstrated that the Hippo/YAP axis is the driver of HCC formation; moreover, targeting the Hippo/YAP axis is thought to be a promising strategy for HCC treatment (4). However, pharmaceutical inhibitors that target the interaction of the YAP–TEAD complex, such as verteporfin and super-TDU (5–7), have failed several clinical trials, halting their clinical translation (8). Given the importance of YAP function in HCC, identifying novel targets in the Hippo signaling pathway for HCC therapy is essential.

The core of the Hippo pathway is a phosphorylation kinase cascade composed of several proteins, including MST1/2 and LATS1/2, which inhibits the activity of YAP/TAZ by promoting phosphorylation, cytosolic retention, and proteasome-dependent degradation (9–11). Numerous studies have revealed abnormalities in the Hippo pathway in human malignancies, including HCC (12). For example, abnormal activation of YAP and TAZ was detected in HCC cells, and this abnormal activation was found to be associated with poor cellular differentiation and reduced survival durations among patients with HCC (13). Furthermore, the roles played by the components of the Hippo pathway in regulating liver size and the oncogenic progression of HCC have undergone rigorous investigation (14, 15). For example, MST1/2 loss in albumin-Cre (Alb-Cre) mice results in HCC formation (16). Collectively, these data suggest that aberrations in the Hippo pathway might serve as the primary catalyst for HCC, thereby rendering the targeting of Hippo signaling a potentially effective approach for clinical HCC treatment.

G protein–coupled receptors (GPCR) play crucial roles in various physiologic processes, including the immune response, hormone regulation, and neurotransmission, making them versatile and essential proteins in the body (17). Given their diverse functions, GPCRs have become a central focus in pharmaceutical research and drug development. Notably, approximately 35% of currently marketed drugs are designed to directly target GPCRs, highlighting their importance as valuable therapeutic targets (18). The biological link between GPCRs and the Hippo pathway was established in 2012 by Professor Kunliang Guan (19). However, identifying the functional GPCRs involved in HCC progression is essential because the GPCR family comprises approximately 900 members and the expression profiles vary among different types of malignancies (20, 21). The oxytocin receptor (OXTR) belongs to the GPCR family and is highly expressed in the mammary gland and uterus (22, 23). The activation of the OXTR by oxytocin plays important roles in uterine contraction and milk ejection (24). In addition, OXTR signaling in the central nervous system also participates in the regulation of social, cognitive, and emotional behaviors (25). Recent studies have also revealed the involvement of the OXTR in breast cancer formation (26) and drug resistance in pancreatic cancer (27). Furthermore, high levels of the OXTR can promote invasion in oral squamous cell carcinoma via ERK5 signaling (28). However, the function of the OXTR in Hippo signaling and hepatocarcinogenesis is still unclear.

In the present study, the OXTR was identified as both an upstream signal and downstream target of the Hippo/YAP axis in HCC progression. OXTR facilitated YAP activation through the Gα_q/11–Rho-associated protein kinase (ROCK)–LATS axis in HCC. Our study revealed a novel positive feedback loop between OXTR signaling and the Hippo/YAP axis, suggesting that blocking the OXTR with atosiban may be a plausible strategy for treating HCC.

Materials and Methods

Cell culture

SNU449 and HEK293 cells were acquired from the ATCC. SNU761 cells were acquired from the Meisen Chinese Tissue Culture Collections (MeisenCTCC). SNU761 and SNU449 cells were cultured in RPMI-1640 medium (42401, Life Technologies) supplemented with 2 mmol/L L-glutamine (25030, Life Technologies). HEK293 cells were cultured in DMEM supplemented with 4.5 g/L glucose and 4 mmol/L L-glutamine (DMEM, D6429, Sigma-Aldrich). All the cell lines were cultured with 10% FBS (10270–106, Gibco) supplemented with 1% penicillin/streptomycin (C0222, Beyotime) at 37°C with 5% CO_2. All cells were routinely tested for Mycoplasma using Mycoplasma PCR Detection Kit (C0301S, Beyotime Biotechnology) and were confirmed to be Mycoplasma-free as of February 21, 2025. All the cell lines were characterized via cell line authentication. Cell line authentication via short tandem repeats was performed via the Promega PowerPlex 21 system.

GPCR-based siRNA screening

The 28 siGPCRs were synthesized and purchased from GenePharma. The sequences of the siRNAs are listed in Supplementary Table S1. SNU449 cancer cells were transfected with different siGPCRs. After transfection for 48 h, RNA was extracted, and the RNA was reverse-transcribed into cDNA. The levels of the classic Hippo/YAP downstream gene CTGF were measured to screen for significant regulation of Hippo/YAP signaling. The mRNA level of siControl CTGF was normalized to one.

DNA constructs

The Flag-OXTR and LATS1 plasmids used in the experiments were obtained from HANBIO (https://www.hanbio.net). The Myc-YAP and TEAD-reporter plasmids were used in our previous studies (29). Lenti-dCas9-KRAB-blast and lentiGuide-Puro were gifts from Dr. Guo (30). The lentiviral shOXTR vectors were created by ligating hybridized oligos into the pLVX lentiviral vector with T4 DNA ligase (New England Biolabs). Supplementary Table S2 displays the sense strand of the short hairpin RNA nucleotide sequence targeting the OXTR. The Flag-OXTRs I227A, I141A, and R137A were cloned with mutagenic primers containing the desired mutation using QuikChange II Site-Directed Mutagenesis Kit (Agilent Technologies).

CRISPR interference assay

Dr. Guo assisted with the CRISPR interference (CRISPRi) assay (30). To achieve stable expression of dCas9-KRAB in SNU449 cells, a lentiviral packaging system was used. This system included the Lenti-dCas9-KRAB-blast plasmid (Addgene, plasmid #89567), along with the pMD2.G and psPAX2 packaging plasmids. Lentiviral particles were produced in HEK293 cells using Lipofectamine 2000 (Thermo Fisher Scientific) following the manufacturer’s protocol. SNU449 cells were then infected with the lentiviral supernatant for 24 hours, followed by selection with 10 μg/mL blasticidin (ST018, Beyotime Biotechnology) for 10 to 14 days. Single-guide RNA (sgRNA) sequences targeting OXTR sites 1 to 4 were designed using CRISPOR (http://crispor.tefor.net) and subsequently cloned and inserted into the lentiGuide-Puro plasmid (Addgene, plasmid #52963). The specific sgRNA sequences are provided in Supplementary Table S2. Lentiviral particles for each sgRNA were generated in HEK293 cells, and dCas9-KRAB–expressing SNU449 cells were transduced and selected with puromycin for 72 hours. To assess the effects of CRISPRi, chromatin immunoprecipitation (ChIP)‒qPCR was performed to measure the relative enrichment of YAP-binding sites.

Transfection and reagents

The cells were transfected with different plasmids via Lipofectamine 2000 (1662298, Invitrogen) and with siRNAs via RNAi Max (Invitrogen, 13778150) according to the manufacturer’s instructions. The sequences of the OXTR and YAP are shown in Supplementary Table S2. Atosiban (MedChemExpress, cat. #HY-17572), carbetocin (MedChemExpress, cat. #HY-17573), verteporfin (MedChemExpress, cat. #HY-B0146), GSK429286A (MedChemExpress, cat. #HY-11000), Y27632 (MedChemExpress, cat. #HY-10071), C3 (Cytoskeleton, cat. #CT03), and VT104 (MedChemExpress, cat. #HY-134956) are shown in Supplementary Table S3.

For lentivirus OXTR silencing, the shOXTR construct was inserted into the pLKO.1 insertion vector and cotransfected with the pMD2.G envelope plasmid and psPAX2 packaging plasmid in HEK293 cells. Cotransfection aimed to express a specific gene. After 48 hours, the lentivirus mixture containing OXTR short hairpin RNA expression was harvested for further experimental procedures. The gene was expressed in SNU449 cells by coculturing them with 300 µL of lentiviral suspension in 3 mL of antibiotic-free medium.

Mice

All animal study protocols were approved by the Shandong University Animal Care and Use Committee (KYLL-2023LW121). Throughout the study, the mice were housed in a specific pathogen-free environment with controlled temperature and lighting conditions (12 hours light/12 hours dark cycle) and free access to food and water. For the tumorigenesis assay, 6-week-old female BALB/c nude mice were purchased from the SPF (Beijing) Biotechnology Co., Ltd., BALB/c nude mice were injected with 2  ×  10⁶ SNU449 cells in 100 μL of PBS subcutaneously (xenograft tumor model) or into the mouse hepatic lobe (31). The tumor width and length were measured at the indicated time points. The tumor volume was calculated via the formula volume = length × (width²)/2.

For mouse strains, Mst1^fl/fl and Mst2^fl/fl mice purchased from Cyagen Biosciences, Inc., were crossed with Alb-Cre mice purchased from Cyagen Biosciences, Inc., for liver-specific Mst1/2 double-knockout. Genotyping was performed via PCR using genomic DNA prepared from the mouse tail. The primer sequences used for genotyping are shown in the Supplementary Materials. Atosiban (1 mg/kg) was prepared separately in normal saline (32, 33).

Organoid culture

Human HCC tissues were obtained from the Second Hospital of Shandong University (KYLL2024453). Each patient provided written informed consent. This study was approved by the Institute’s Research Ethics Committee of the Second Hospital of Shandong University and conducted in accordance with the ethical guidelines of the World Medical Association Declaration of Helsinki. Informed consent was provided by all patients prior to this study. The related reagents were purchased from BioGenous Biotechnology Co. Ltd. The detailed steps were as follows: human HCC tissue pieces were collected in primary tissue storage solution in conical tubes, and the tissue samples were stored at 4°C until isolation. The tissues were then rinsed with DPBS twice. The tissue was minced into small fragments of 1 to 3 mm³ in a cell culture dish using surgical scissors. The tissue fragments were digested with 10 mL of tumor tissue digestion solution in a 15-mL conical tube at 37°C, with various incubation times ranging from 15 to 45 minutes. The digestion was terminated by adding 2% FBS, and the solution was filtered through a 100-μm cell strainer and centrifuged at 250 × g for 3 minutes at 4°C. In the case of a visible red pellet, the supernatant was aspirated, and the pellet was resuspended in 2 mL of red blood cell lysis solution to lyse the erythrocytes at room temperature for 1 minute and then centrifuged at 250 × g for 3 minutes at 4°C. The supernatant was aspirated, and the pellet was resuspended in the extracellular matrix (ECM). The ECM-containing organoids at the bottom of 24-well cell culture plates were plated in ∼30 μL droplets around the center of each well. The culture plate was placed in a humidified incubator at 37°C and 5% (vol/vol) CO₂ for 15 to 25 minutes to allow the ECM to solidify, and 500 μL of complete organoid medium was carefully added to each well. The culture plate was placed at 37°C in a 5% (vol/vol) CO₂ incubator. The medium was changed every 3 to 4 days by carefully aspirating the medium from the wells and replacing it with fresh, prewarmed complete organoid medium. The organoids were closely monitored. Ideally, patient-derived HCC organoids should be passaged for the first time 15 days after initial plating.

RNA isolation and RT-PCR

The manufacturer’s instructions were followed to extract cellular RNA using an RNA extraction kit (Tiangen, DP451) as previously described (34). The primer sequences designed for the qPCR assay are shown in Supplementary Table S4. The data were analyzed via the 2^−ΔΔC_t method, with 36B4 serving as a standard gene for normalization.

Cell Counting Kit-8 assay

As previously described (34), the assay for cell viability was performed using Cell Counting Kit-8 (CCK-8). A total of 4  ×  10³ cells were plated into 96-well plates. The proliferation levels were assessed at designated time intervals. By using a multifunctional enzyme-linked analyzer (BioTek), the optical density value at 450 nm was quantified.

Transwell assay

As previously described (34), the migration assay was conducted in uncoated Transwell chambers with 8-μm pores (Corning). First, the upper chamber was inoculated with cells suspended in 200 μL of serum-free medium, whereas the lower chamber was inoculated with 500 μL of medium containing 20% FBS. Following a 24-hour incubation period, 4% paraformaldehyde and hematoxylin were used to fix and stain the cells that crossed the lower surface of the membrane. The cells from three randomly selected fields were subsequently counted in triplicate for the experiment.

Colony formation and wound-healing assays

SNU449 and SNU761 cells were transfected with either siOXTR or siControl, whereas SNU761 cells were transfected with either the Flag vector or the Flag-OXTR plasmid. The methods used for the colony formation and wound-healing assays were described in our previous studies (35). The percentage of wound recovery was expressed as [1−(width of the wound at a given time/width of the wound at t = 0)] × 100%.

Western blotting and coimmunoprecipitation assay

For Western blotting, the cells were harvested and lysed with Western and IP lysis buffer (P0013J, Beyotime) supplemented with protease and phosphatase inhibitors (Thermo Fisher Scientific) as previously described (34). Secondary antibodies (Beyotime, A0216, 1:6,000 or Beyotime, A0208, RRID: AB_2892644, 1:6,000) were applied. The signals were visualized using Western blotting with an enhanced chemiluminescence substrate.

To conduct coimmunoprecipitation assays, 500 μg of protein lysate was precleared with 20 μL of Protein A+G agarose (Beyotime, P2028) and rabbit IgG (Beyotime, A7016, RRID: AB_2905533, 1:50) for 2 hours at 4°C. Immunoprecipitation was subsequently carried out for 4 hours at 4°C with the specified antibody. The negative controls included either rabbit IgG (Beyotime, A7016, RRID: AB_2905533, 1:50) or mouse IgG (Beyotime, A7028, RRID: AB_2909433, 1:50). Western blot analysis was performed using specific antibodies. The antibodies used are listed in Supplementary Table S5.

In vitro kinase activity assay

The Phos-tag acrylamide reagent (Wako, 93521), purchased from Wako Chemicals, was utilized to prepare gels containing phos-tag and MnCl₂ (Wako, 13446-34-9), following the guidelines provided by the manufacturer. For the primary antibody on these prepared phos-tag gels, anti-YAP antibody (Cell Signaling Technology, cat. #14074; RRID: AB_2650491, 1:500) was used. This antibody allows the YAP protein to be resolved into distinct, visible bands, which are naturally separated on the basis of their phosphorylation state, thus facilitating the observation of changes in phosphorylation.

Dual-luciferase reporter assay

Luciferase activity was measured via the Dual-Luciferase Reporter Assay System (E1910, Promega). The cells were transfected with the designated siRNA or plasmid along with the required chemicals, the TEAD luciferase reporter vector, or the Renilla plasmid. Following a 24-hour incubation period, the cells were lysed, and the luciferase activity was subsequently assayed. A GloMax-Multi Jr (Promega-GloMax) instrument was used to detect luciferase activity.

Immunofluorescence assay

SNU449 cells were fixed with 4% paraformaldehyde for 15 minutes, permeabilized with 0.25% Triton X-100 for 10 minutes, and blocked with 3% BSA for 30 minutes at room temperature. A rabbit anti-OXTR antibody (Abcam, ab217212, RRID: AB_2904223, 1:400) was then applied. Subsequently, Alexa Fluor 647–conjugated (Invitrogen) anti-mouse or Alexa Fluor 488–conjugated anti-rabbit secondary antibodies (Invitrogen) were used. The nuclei were counterstained with DAPI (Beyotime). Negative controls were prepared using samples incubated only with secondary antibodies, excluding the primary antibodies. The images were acquired using a Nikon A+ laser scanning confocal system.

Evaluation of IHC staining

The scores were determined by combining the proportion of positively stained tumor cells and the intensity of staining. The tumor cell proportions were recorded as follows: 0 (no positive tumor cells), 1 (≤20% positive tumor cells), 2 (21%–50% positive tumor cells), 3 (51%–70% positive tumor cells), and 4 (>70% positive tumor cells). The staining intensity was recorded as follows: 0 (no staining), 1 (weak staining = light yellow), 2 (moderate staining = yellow brown), and 3 (strong staining = brown). The staining index (SI) was calculated as the product of the staining intensity score and the proportion of positive tumor cells. Using this assessment method, we evaluated OXTR expression by determining the SI, with possible scores of 0, 1, 2, 3, 4, 6, 9, or 12. Tumors with a SI score of ≥6 were classified as having positive or high OXTR expression, whereas those with a SI score of <6 were classified as having negative or low OXTR expression.

Patient-derived explant assay

The excised tissues, intended for research and approved by the ethical committee of Qilu Hospital, Shandong University, were processed in accordance with the ex vivo culture protocol. Each patient provided written informed consent. Typically, liver cancer tissues were cultured on gelatine sponges with cell culture medium supplemented with 10% FBS. The tissues were then subjected to either vehicle or atosiban treatment for 48 hours. The tissues were subsequently fixed in 10% formaldehyde at 4°C overnight. To verify the quality of the samples, hematoxylin and eosin staining was performed. Subsequently, IHC analysis was conducted to examine specific markers.

RNA sequencing and data analysis

The RNA sequencing (RNA-seq) data were stored in the Gene Expression Omnibus repository (GEO, GSE249404 for siControl and siOXTR in SNU449; GSE266229 for vehicle and atosiban in human HCC organoids). Genes exhibiting significant differential expression [P < 0.05 and fold change (FC) >1.5] were subjected to ingenuity pathway analysis for further investigation. For gene set enrichment analysis (GSEA), the CORDENONSI_YAP_CONSERVED_SIGNATURE gene set was retrieved from the GSEA Molecular Signatures Database and implemented via GSEA 4.2.3 software. Additionally, Metascape (https://Metascape.org) was used for enrichment analysis, leveraging hallmark gene sets and Kyoto Encyclopedia of Genes and Genomes pathways to delve into pathways associated with differentially expressed genes. Furthermore, OmicStudio tools (https://www.omicstudio.cn/tool) were used to generate a volcano plot of the differentially expressed genes, with thresholds of P < 0.05 and a FC >1.5.

Analysis of publicly available data

Gene expression data for 373 human HCC tumor samples were retrieved from The Cancer Genome Atlas (TCGA) database. OXTR expression levels in the TCGA dataset were arranged in ascending order, and the cutoff value was determined by the median for the low and high groups. The relationship between OXTR expression and survival in patients with HCC was analyzed using the Xiantao online platform (https://www.xiantao.love/). The results were generated via GSEA online software with CORDENONSI_YAP_CONSERVED_SIGNATURE, which comes from the Molecular Signature Database (https://www.gsea-msigdb.org/gsea/msigdb/index.jsp). YAP-based ChIP sequencing (ChIP-seq) data were retrieved from the NCBI (GSE66081, GSE61852, GSE107013, and GSE131687).

Statistical analysis

The statistical analyses used in this study included Student t test, the χ² independence test, and ANOVA. The data are presented as the means ± SDs, with statistical significance defined as *, P < 0.05; **, P < 0.01; and ***, P < 0.001.

Ethics approval and consent to participate

The Ethical Board of Shandong University reviewed and approved this use of clinical samples (KYLL2024453) and animal models (KYLL-2023LW121). Each patient provided written informed consent.

Data availability

RNA-seq data for cell lines can be found in the GEO database (GSE249404). The RNA sequence data for human HCC organoids can be found in the GEO database (GSE266229). YAP-based ChIP-seq data were retrieved from the NCBI (GSE66081; GSE61852, GSE107013, and GSE131687). The Western blot data are provided in the supplementary materials. All other raw data are available upon request from the corresponding author.

Results

Identification of the OXTR as a novel mediator of the Hippo/YAP pathway in HCC

Approximately 35% of currently marketed drugs are designed to directly target GPCRs, highlighting their importance as valuable therapeutic targets (18). The GPCR family is a large group of cell-surface receptors that comprises approximately 900 members. We first analyzed the whole-genome expression profiles of HCC samples from the TCGA database. The correlation of the expression of each GPCR with the Hippo gene signature was analyzed. GSEA revealed 75 GPCRs with the highest normalized enrichment scores (NES, >1.5). Moreover, 134 GPCRs were identified as targets for drugs approved in the United States or European Union (36). First, a cross-comparison of the 134 GPCRs targeted by drugs approved in the United States and the 75 GPCRs strongly associated with the Hippo/YAP pathway identified by GSEA (NES >1.5, P < 0.05) revealed 28 common genes (Fig. 1A). Second, we further validated these 28 genes via qRT-PCR and siRNA screening. The expression of CTGF, the main Hippo target gene, was used as a readout (Fig. 1B), and OXTR depletion resulted in the lowest mRNA level of CTGF (Fig. 1C).

Figure 1.

Identification of the OXTR as a novel mediator of the Hippo/YAP pathway in HCC. A, Comparison of 75 GPCRs strongly associated with the Hippo/YAP pathway and 134 GPCRs targeted by drugs approved in the United States according to GSEA (NES >1.5; P < 0.05). B, The flowchart shows the siRNA screening procedure involved in modulating Hippo/YAP signaling for the 28 GPCRs shown in A. C, The relative expression level of CTGF in cells that were transfected with siGPCRs in the screening library. The qRT-PCR results of the siControl group were normalized to one. ns, nonsignificant. D, Relative RNA levels of OXTRs in HCC tumor samples (n = 371) vs. normal samples (n = 50) from the TCGA database (https://www.genome.gov/). FC = 2.16. E, IHC was used to detect OXTR expression in 90 liver cancer tissues and 60 normal liver tissues, and statistical analysis of OXTR expression was performed. Scale bar, 100 μm. F, Kaplan–Meier analysis showing overall survival depending on OXTR expression levels in HCC tumor samples from the TCGA database. G, GSEA revealed a significant positive correlation between the OXTR and YAP target gene signature in HCC samples from the TCGA database, with a threshold criterion of P < 0.05. H, Heatmap of the relationships between the OXTR and differentially expressed Hippo/YAP pathway–related genes in the TCGA database with threshold criteria of P < 0.05 and a FC >1.5. I, Volcano map of RNA-seq data from SNU449 cell lines treated with siControl or siOXTR. |log₂ FC| > 1 and P < 0.05 were set as screening criteria. NoDiff, no difference; Sig_Down, significant downregulation; Sig_Up, significant upregulation. J, Kyoto Encyclopedia of Genes and Genomes analysis of downregulated genes in RNA-seq data from SNU449 cell lines treated with siControl or siOXTR with threshold criteria of P < 0.001 and a FC > 1.5. K, GSEA showing enrichment of the CORDENONSI_YAP_CONSERVED_SIGNATURE in RNA-seq data from SNU449 cell lines treated with siControl or siOXTR. L and M, IHC was used to detect OXTR and YAP expression in 90 liver cancer tissues (L). Statistical analysis of OXTR and YAP expression in L (M). Scale bar, 100 μm. N, Correlations between the OXTR expression level in HCC tumor samples and the clinicopathologic characteristics of the corresponding patients. Data analysis revealed that OXTR expression was correlated with sex (P = 0.005), tumor invasion (P = 0.031), and lymph node metastasis (P = 0.027). All data are presented as the means ± SDs. ns, nonsignificant; *, P < 0.05; **, P < 0.01; ***, P < 0.001 for comparisons. B, Created in BioRender. Yang, H. (2025) https://BioRender.com/exsenbv.

We further investigated the expression pattern of the OXTR in HCC. Intriguingly, an analysis of the TCGA database revealed increased OXTR mRNA levels in HCC tissues compared with normal liver tissues (Fig. 1D). This finding was further validated through IHC analysis of HCC and normal liver samples, as shown in Fig. 1E. A subsequent survival analysis using the TCGA database revealed a correlation between high OXTR expression and poor survival outcomes among patients with HCC (Fig. 1F), which was further validated by analyzing survival curves through IHC analysis of the HCC samples shown in Fig. 1E (Supplementary Fig. S1A). Furthermore, upon analyzing the correlations between OXTR expression and the expression of the main downstream genes of the Hippo signaling pathway in the TCGA database, GSEA revealed a notable positive association between OXTR expression and the YAP target gene signature in HCC tissue (NES = 1.54; P < 0.05; Fig. 1G). Moreover, we further analyzed the correlation between the OXTR and YAP in a series of human cancers. Interestingly, a notable positive association was found between OXTR expression and the YAP target gene signature in pancreatic adenocarcinoma (PAAD; NES = 1.84; P < 0.001), kidney renal clear cell carcinoma (KIRC; NES = 2.17; P < 0.001), colon adenocarcinoma (COAD; NES = 1.71; P < 0.05), and lung adenocarcinoma/lung squamous cell carcinoma (LUAD/LUSC; NES = 1.75; P < 0.05) in the CORDENONSI_YAP_CONSERVED_SIGNATURE gene set (Supplementary Fig. S1B–S1E). Therefore, we believe that the effect of the OXTR on the Hippo/YAP axis could be a general regulatory phenomenon in a series of human malignancies.

In addition, a heatmap depicting the relationships between OXTR expression and differentially expressed Hippo/YAP pathway–related genes in the TCGA data, with a threshold of P < 0.05 and a FC >1.5, revealed a significant positive correlation between the OXTR and 27 genes, including CCN1 (CYR61) and CCN2 (CTGF). This finding effectively corroborated our initial screening results (Fig. 1H). Furthermore, this observation was validated through RNA-seq analysis of SNU449 cells (GSE249404). The volcano plot clearly demonstrated that OXTR depletion led to significant downregulation of the expression of key Hippo pathway target genes, including CYR61 and CTGF, as shown in Fig. 1I. In addition to the Hippo pathway, our Kyoto Encyclopedia of Genes and Genomes pathway analysis revealed that OXTR depletion also influenced several other oncogenic pathways, such as the metabolic pathway, Wnt pathway, and PI3K-AKT signaling pathway (Fig. 1J). GSEA revealed enrichment of the YAP target gene signature in the RNA-seq data from SNU449 cell lines treated with siControl or siOXTR (NES = −1.31; P < 0.05; Fig. 1K). Furthermore, we conducted IHC analysis to examine OXTR expression in HCC samples and its association with clinical characteristics. The IHC data revealed a significant positive correlation between YAP and OXTR expression (Fig. 1L and M). Clinical data analysis revealed that OXTR expression was positively correlated with sex, tumor invasion, and lymph node metastasis (Fig. 1N). Considering these findings, we hypothesize that the OXTR could serve as a significant positive regulator of the Hippo signaling pathway in human HCC.

OXTR depletion slows HCC progression in vivo and in vitro

We further explored the impact of the OXTR on the phenotypic characteristics of HCC cell lines by decreasing the expression levels of the OXTR. The depletion of the OXTR was validated through immunoblot analysis (Fig. 2A; Supplementary Fig. S2A) and significantly impaired the proliferation of HCC cancer cells, as indicated by the results of the CCK-8 assay (Fig. 2B; Supplementary Fig. S2B). Furthermore, the colony formation assay demonstrated a decrease in colony-forming ability when the OXTR level was reduced (Fig. 2C and D; Supplementary Fig. S2C and S2D). Additionally, the Transwell migration assay revealed a significant reduction in the migratory capacity of the HCC cell lines following OXTR depletion (Fig. 2E and F; Supplementary Fig. S2E and S2F). In addition, this conclusion was reinforced by the results of the wound-healing assay (Fig. 2G and H; Supplementary Fig. S2G and S2H). Moreover, to further evaluate the role of the OXTR, an in vivo orthotopic xenograft model of HCC was used. To establish this model, nude mice were randomly divided into two groups and subjected to hepatic lobe injection with either shControl or shOXTR to establish an orthotopic subcutaneous xenograft tumorigenesis model. These findings indicated that OXTR silencing hampered the tumorigenic capacity of HCC cells (Fig. 2I and J). Furthermore, subsequent IHC examination of the orthotopic xenograft tumors revealed decreased expression levels of Ki67 (Fig. 2K and L). In addition, this conclusion was reinforced by the results of a subcutaneous xenograft tumorigenesis model in which shControl or shOXTR was subcutaneously injected (Fig. 2M–S). Overall, OXTR depletion significantly inhibits the progression of HCC.

Figure 2.

OXTR depletion slows HCC progression in vivo and in vitro. A, Immunoblot analysis showing the expression level of the OXTR in SNU449 cell lines transfected with siControl or two independent OXTR siRNAs for 48 hours. β-Actin was used as the internal control. B, A CCK-8 assay was used to detect the viability of SNU449 cells transfected with siControl or siOXTR for 24 hours, and the results were tested at the indicated time points. C and D, Colony formation of SNU449 cell lines transfected with siControl or two independent OXTR siRNAs for 24 hours. D, Quantitative analysis of the colony formation assay results. E‒H, Transwell and wound-healing assays were used to detect the migration ability of SNU449 cells treated as indicated. F and H, The results of the quantitative analysis. I and J, Imaging (I) and number (J) of orthotopic xenograft HCC tumors harvested from four mice per group. K and L, IHC was used to detect Ki67 expression in orthotopic xenograft HCC tumors (I), and a statistical analysis of Ki67 expression (L) was performed. Scale bar, 40 μm. M‒O, Imaging (M), tumor weights (N), and tumor volumes (O) in BALB/c nude mice subcutaneously inoculated with stably transfected shControl or shOXTR SNU449 cells. P, Lysates of tumor tissues from M were subjected to immunoblotting analyses with the indicated antibodies. Q–S, IHC was used to detect OXTR, YAP, p-YAP, and Ki67 expression in M (Q), and statistical analysis of YAP (R) and Ki67 expression (S) was performed. Scale bar, 150 μm. Three independent experiments were conducted to obtain the results shown in A‒H. All data are presented as the means ± SDs. *, P < 0.05; **, P < 0.01; ***, P < 0.001 for comparisons (Student t test).

The OXTR antagonist atosiban inhibits HCC progression in vivo and in vitro

We further investigated the effect of atosiban, an OXTR antagonist, on HCC progression. OXTR protein levels were decreased in cell lines treated with atosiban (Fig. 3A; Supplementary Fig. S3A) and significantly impaired the proliferation of HCC cancer cells, as indicated by the CCK-8 assay (Fig. 3B; Supplementary Fig. S3B). Furthermore, the colony formation assay revealed that atosiban reduced the capacity for colony formation (Fig. 3C and D; Supplementary Fig. S3C and S3D). Transwell assays demonstrated that atosiban diminished the migratory potential of the HCC cell lines (Fig. 3E and F; Supplementary Fig. S3E and S3F). Moreover, this observation was further strengthened by the findings of the wound-healing assay (Fig. 3G and H; Supplementary Fig. S3G and S3H). Additionally, the efficacy of atosiban was further evaluated in vivo using a subcutaneous xenograft tumorigenesis model (Fig. 3I–L). Furthermore, subsequent IHC analysis of the xenograft tumors revealed decreased nuclear localization of YAP and expression of Ki67 following atosiban treatment (Fig. 3M–O). In summary, our findings, along with those of others, indicate that atosiban significantly inhibits the progression of HCC.

Figure 3.

The OXTR antagonist atosiban inhibits HCC progression in vivo and in vitro. A, Immunoblot analysis showing the expression level of the OXTR in SNU449 cell lines treated with DMSO or atosiban for 24 hours. β-Actin was used as the internal control. B, A CCK-8 assay was used to detect the viability of SNU449 cells treated with DMSO or atosiban for 24 hours at the indicated time points. C and D, Colony formation of SNU449 cell lines treated with DMSO or atosiban for 24 hours. D, Quantitative analysis of the colony formation assay results. E‒H, Transwell and wound-healing assays were used to detect the migration ability of SNU449 cells treated with DMSO or atosiban for 24 hours. F and H The results of the quantitative analysis. I‒L, Flowchart (I), imaging (J), tumor weights (K), and tumor volume (L) in BALB/c nude mice subcutaneously inoculated with stably transfected shControl or shOXTR SNU449 cells. Error bars, means ± SDs. M‒O, IHC was used to detect hematoxylin and eosin (H&E), YAP, p-YAP, and Ki67 expression in J (M), and statistical analysis of YAP (N) and Ki67 expression (O) was performed. Error bars, means ± SDs. Scale bar, 150 μm. Three independent experiments were conducted to obtain the results shown in A‒H. All data are presented as the means ± SDs. **, P < 0.01; ***, P < 0.001 for comparisons (Student t test). I, Image of mice created in BioRender. Yang, H. (2025) https://BioRender.com/twioz0d.

Atosiban inhibits HCC progression in Mst1^flox/flox; Mst2^flox/flox mice and patient-derived explant and organoid models

A series of studies highlighted the pivotal role of the Mst1, Mst2, and YAP genes in regulating growth and tumorigenesis in the liver (37). Previous studies revealed that Mst1^flox/flox; Mst2^flox/flox; Alb-Cre mice could spontaneously produce tumors at 3 months of age (<5 mm; ref. 38). To thoroughly study the potential inhibitory effects of OXTR gene therapy on HCC progression, we generated Mst1^flox/flox; Mst2^flox/flox; Alb-Cre mice. PCR analysis of the Cre, Mst1 and Mst2 alleles from the indicated genotypes is shown in Fig. 4A. Compared with Mst1^wt/wt; Mst2^wt/wt; Alb-Cre mice, Mst1^flox/flox; Mst2^flox/flox; Alb-Cre mice presented progressive hepatomegaly with a 1.5-fold increase in liver mass relative to total body mass at 2 months of age (Fig. 4B and C). Consistent with previously reported observations, we also detected tumors in Mst1^flox/flox; Mst2^flox/flox; Alb-Cre mice at 3 months of age. Saline or atosiban (1 mg/kg)/day was then intraperitoneally administered to these mice (Fig. 4D). The results revealed that atosiban inhibited HCC progression in the Mst1^flox/flox; Mst2^flox/flox; Alb-Cre mouse model (Fig. 4E and F). We further carried out additional studies on YAP expression/localization and downstream pathway expression. Compared with Mst1^wt/wt; Mst2^wt/wt; Alb-Cre mice, Mst1^flox/flox; Mst2^flox/flox; Alb-Cre mice presented low levels of p-LATS1 and p-YAP, whereas the mRNA levels of the YAP target genes CTGF and CYR61 were significantly increased (Fig. 4G and H). In addition, atosiban increased the levels of p-LATS1 and p-YAP and inhibited the mRNA levels of the YAP target genes CTGF and CYR61 in the Mst1^flox/flox; Mst2^flox/flox; Alb-Cre mouse model (Fig. 4G and H). As phosphorylation of YAP leads to its nuclear export and retention in the cytoplasm, we further investigated the subcellular localization of YAP. Compared with Mst1^wt/wt; Mst2^wt/wt; Alb-Cre mice, Mst1^flox/flox; Mst2^flox/flox; Alb-Cre mice presented high levels of nuclear localization of YAP. Moreover, the results of the nuclear–cytosol separation assay indicated that atosiban could reduce the nuclear localization of YAP in the Mst1^flox/flox; Mst2^flox/flox; Alb-Cre mouse model (Fig. 4I). Furthermore, subsequent IHC evaluation of the tumors revealed diminished expression of Ki67 following treatment with atosiban in Mst1^flox/flox; Mst2^flox/flox; Alb-Cre mice (Fig. 4J and K). In the patient-derived xenograft assay, we found that atosiban could reduce the proportion of nuclear YAP and increase the phosphorylation of YAP at the Ser127 site in the OXTR-high and -intermediate groups but had little effect on YAP in the OXTR-low group (Fig. 4L‒N; Supplementary Fig. S4A‒S4D).

Figure 4.

Atosiban inhibits HCC progression in Mst1^flox/flox; Mst2^flox/flox mice and patient-derived explant and organoid models. A, PCR analysis of Cre, Mst1, and Mst2 from the indicated genotypes. B, Gross images of livers from Mst1^wt/wt; Mst2^wt/wt; Alb-Cre and Mst1^flox/flox; Mst2 ^flox/flox; Alb-Cre mice at 2 months of age. C, Liver weight‒to‒body weight ratios in Mst1^wt/wt; Mst2^wt/wt; Alb-Cre and Mst1^flox/flox; Mst2 ^flox/flox; Alb-Cre mice at 2 months of age. Statistical significance (unpaired t test) is indicated. D, Flowchart of atosiban or saline solution treatment experiments in Mst1^flox/flox; Mst2 ^flox/flox; Alb-Cre mice. E and F, Hepatoma formation following the intraperitoneal administration of normal saline or 1 mg/kg atosiban (E). F, The relative tumor number is indicated. G, Immunoblot analysis showing the expression levels of Mst1, Mst2, p-YAP, YAP, p-LATS1, and LATS1 in the indicated mice subjected to the indicated treatments. β-Actin was used as the internal control. H, CTGF and CYR61 mRNA levels were determined via qRT-PCR in the indicated mice subjected to the indicated treatments. I, Nucleoplasm separation by immunoblotting was used to detect YAP in the indicated mice subjected to the indicated treatments. J and K, IHC was used to detect Ki67 expression, as shown in E (J), and to perform a statistical analysis of Ki67 expression (K). Error bars, means ± SDs. Scale bar, 40 μm. L‒N, In the patient-derived explant assay, atosiban treatment inhibited the proliferation potential of HCC tumors. HCC tumor xenografts were treated with vehicle or 10 μmol/L atosiban. The samples were fixed and stained for OXTR, YAP, p-YAP, and Ki67. The YAP- and Ki67-positive cells were counted for analysis (M and N). Scale bar, 150 μm. O and P, Single organoid expanding on days 3, 6, 9, 12, and 15 after treatment with vehicle or 10 μmol/L atosiban. Scale bar, 40 μm (O). Statistical analysis of organoid viability (P). ns, nonsignificant. Q, Heatmap of the relationship of relative mRNA expression (log₂-FC) of 45 Hippo/YAP target genes in RNA-seq data (atosiban vs. vehicle, GSEA266229) in three different HCC organoids treated with vehicle or 10 μmol/L atosiban. The upregulation and downregulation of mRNA levels are indicated by red and green, respectively, with threshold criteria of P < 0.05 and a FC > 1.5. R‒T, GSEA showing enrichment of the YAP target gene signature in RNA-seq data from three different HCC organoids treated with vehicle or 10 μmol/L atosiban (GSEA266229). All data are presented as the means ± SDs. *, P < 0.05; **, P < 0.01; ***, P < 0.001 for comparisons (Student t test). D, Image of mice created in BioRender. Yang, H. (2025) https://BioRender.com/twioz0d.

In recent years, interest in organoids, which are unique models that closely mimic native tumor tissue and are capable of in vitro culture, has increased. Various studies have focused on creating organoid models specifically for liver cancer (39). In this study, we effectively refined the methodology and implemented a culture protocol, leading to the development of larger-sized HCC organoids with consistent passaging and stable growth conditions. Furthermore, in vitro human hepatoma organoid models confirmed that atosiban inhibits the growth of human hepatoma organoids (Fig. 4O and P). Moreover, this conclusion was confirmed by RNA-seq analysis of three different human hepatoma organoid models treated with vehicle or atosiban. A heatmap of the relationships between atosiban and 45 differentially expressed Hippo/YAP pathway–related genes identified via RNA-seq analysis with a threshold criterion of P < 0.05 also revealed that atosiban was significantly negatively correlated with these genes (Fig. 4Q). GSEA consistently revealed enrichment of the YAP target gene signature in the RNA-seq data of the vehicle-treated group but not the atosiban-treated group (Fig. 4R–T). We also tested the effect of atosiban in HCC organoids with high, intermediate, and low OXTR expression. The data revealed that atosiban inhibited organoid growth in the high-OXTR and intermediate-OXTR groups but had little effect in the low-OXTR group (Supplementary Fig. S4E–S4K). Overall, in Mst1^flox/flox; Mst2^flox/flox mice and patient-derived explant and organoid models, we clarified that atosiban inhibits HCC progression.

OXTR overexpression and pharmacologic OXTR activation promote HCC progression in vivo and in vitro

We further examined the impact of the OXTR on the HCC phenotype. The overexpression of the OXTR and administration of carbetocin, a pharmacologic activator of the OXTR, increased the OXTR protein level in SNU761 cells, which express low basal levels of the OXTR (Supplementary Fig. S5A and S5B), and significantly promoted the proliferation of HCC cancer cells, as indicated by the CCK-8 assay (Supplementary Fig. S5C and S5D). Furthermore, the colony formation assay revealed that OXTR overexpression and carbetocin increased the colony formation capacity (Supplementary Fig. S5E–S5H), whereas the Transwell assay revealed that OXTR overexpression and carbetocin increased the migration capacity of SNU761 cells (Supplementary Fig. S5I–S5L). In addition, this conclusion was reinforced by the results of the wound-healing assay (Supplementary Fig. S5M–S5P). The function of the OXTR was further assessed through an in vivo experiment in which a xenograft model of HCC was used to evaluate the role of the OXTR. A subcutaneous xenograft tumorigenesis model was established by randomly assigning nude mice to two groups, with one group receiving SNU761 cells stably transduced with the indicated plasmid by a lentivirus through subcutaneous injection. These findings indicated that the OXTR augmented the tumorigenic ability of HCC cells (Supplementary Fig. S5Q–S5S).

The OXTR activates the Hippo/YAP axis by decreasing YAP phosphorylation

To validate its biological connection to Hippo signaling, we extended our investigation by depleting the OXTR in HCC cell lines. The depletion of the OXTR in HCC cell lines promoted YAP phosphorylation, as indicated by immunoblotting (Fig. 5A; Supplementary Fig. S6A). Similarly, blocking the OXTR with atosiban had a similar effect on HCC cells (Fig. 5B; Supplementary Fig. S6B). Moreover, the overexpression and pharmacologic activation of the OXTR through carbetocin reduced YAP phosphorylation (Fig. 5C and D). Additionally, we examined the mRNA levels of YAP target genes (CTGF and CYR61) in HCC cell lines and revealed that depletion of the OXTR suppressed the expression of CTGF and CYR61 (Fig. 5E; Supplementary Fig. S6C). Blockage of the OXTR with atosiban had a similar effect on CTGF and CYR61 mRNA expression (Fig. 5F; Supplementary Fig. S6D). Moreover, overexpression and pharmacologic activation of the OXTR by carbetocin increased YAP target gene expression (Fig. 5G and H). Consistently, the luciferase reporter experiment revealed a decrease in TEAD response element activity due to OXTR depletion or inhibition by atosiban in HCC cell lines, whereas an increase in TEAD response element activity was observed upon OXTR overexpression and pharmacologic activation by carbetocin (Fig. 5I–L; Supplementary Fig. S6E and S6F). As phosphorylation of YAP leads to its nuclear export and retention in the cytoplasm, we further investigated the subcellular localization of YAP. The results of the nuclear–cytosol separation assay indicated that OXTR silencing or inhibition significantly decreased the nuclear localization of YAP (Fig. 5M and N). However, the overexpression and pharmacologic activation of the OXTR by carbetocin induced the nuclear localization of YAP (Fig. 5O and P), which was further confirmed by an immunostaining assay (Fig. 5Q–T).

Figure 5.

The OXTR activates the Hippo/YAP axis by decreasing YAP phosphorylation. A‒D, Immunoblot analysis showing the expression levels of the OXTR, YAP phospho-tag, and total YAP in SNU449 and SNU761 cells treated with the indicated methods for 48 hours. β-Actin was used as the internal control. E‒H, CTGF and CYR61 mRNA levels were determined by qRT-PCR in SNU449 and SNU761 cells treated with the indicated method for 48 hours. I‒L, TEAD response element transcriptional activity was detected in SNU449 and SNU761 cells treated with the indicated method for 48 hours. M‒P, Nucleoplasm separation by immunoblotting confirmed that YAP was located in the cytoplasm and nucleus after 48 hours of incubation with the indicated methods. Q‒T, Immunofluorescence staining assay showing the localization patterns of YAP. OXTR overexpression and carbetocin promote the translocation of YAP from the cytoplasm to the nucleus. Endogenous YAP (green) and nuclei (blue) were stained with specific antibodies and DAPI, respectively. Scale bar, 10 μm. Quantification of the subcellular localization of YAP in at least 100 randomly selected cells. Three independent experiments were conducted to obtain the results. All data are presented as the means ± SDs. **, P < 0.01; ***, P < 0.001 for comparisons (Student t test).

Several studies have emphasized the role of TAZ in HCC progression (40, 41). We further depleted the OXTR or inhibited OXTR function using atosiban in SNU449 and observed that OXTR silencing or inhibition did not change the total TAZ protein level or TAZ phosphorylation at the Ser89 site (Supplementary Fig. S6G and S6H). Moreover, overexpression and pharmacologic activation of the OXTR by carbetocin did not affect the total TAZ protein level or TAZ phosphorylation at the Ser89 site (Supplementary Fig. S6I and S6J). In addition, OXTR inhibition or activation did not change TAZ mRNA levels in SNU449 and SNU761 cells (Supplementary Fig. S6K–S6N). On the basis of these findings, we can conclude that the OXTR functions in Hippo signaling mainly through YAP but not TAZ.

The OXTR facilitates HCC progression using activation of the Hippo/YAP axis

To delve deeper into the interplay between the Hippo pathway and OXTR function in the context of HCC, we performed a series of rescue experiments according to our previous experiments. The effectiveness of OXTR silencing and YAP overexpression was confirmed through Western blot analysis (Supplementary Fig. S7A). Furthermore, the results of the CCK-8 assay indicated that OXTR depletion resulted in the inhibition of HCC cell proliferation, a phenotype that was partially rescued upon further YAP overexpression (Supplementary Fig. S7B). Furthermore, the findings from the colony formation assay revealed that the ability of HCC cancer cells to form colonies was diminished following OXTR depletion. However, this suppressive influence on colony formation was partially alleviated by subsequent overexpression of YAP (Supplementary Fig. S7C and S7D). Similarly, the results obtained from the Transwell and wound-healing assays suggested that the migratory capacity of HCC cells was notably hampered after OXTR deprivation. Nevertheless, this inhibitory effect on cell migration was partially reversed by subsequent overexpression of YAP (Supplementary Fig. S7E–S7H). Furthermore, an orthotopic subcutaneous xenograft tumorigenesis model was established by randomly allocating nude mice into three groups, with each group treated through hepatic lobe injection of shControl, shOXTR, or shOXTR+YAP SNU449 cells. The results demonstrated that OXTR silencing suppressed the tumorigenic potential of HCC cells, a phenotype that was partially mitigated by subsequent YAP overexpression (Supplementary Fig. S7I and S7J). Additionally, subsequent IHC analysis of the orthotopic xenograft tumors revealed diminished expression of Ki67 (Supplementary Fig. S7K and S7L). This conclusion was reinforced by the results of a subcutaneous xenograft tumorigenesis model with shControl, shOXTR, or shOXTR+YAP through subcutaneous injection (Supplementary Fig. S7M–S7O). Overall, OXTR depletion significantly slows the progression of HCC.

The OXTR activates YAP through the Gα_q/11–ROCK–LATS axis in HCC

The OXTR, a 7-transmembrane GPCR capable of binding to either Gα_i or Gα_q proteins, activates a set of signaling cascades, such as the MAPK, PKC, PLC, or CaMK pathways (23). To determine the specific Gα protein that mediates YAP regulation following OXTR overexpression and pharmacologic activation with carbetocin, we silenced Gα_q/11 or Gα_i in HCC cells. Notably, depletion of Gα_q/11 significantly impeded the dephosphorylation of YAP triggered by the OXTR, whereas depletion of Gα_i had a negligible effect on SNU761 cells (Fig. 6A and B). This conclusion was further confirmed by an immunostaining assay (Fig. 6C–F). Because LATS1/2 are the major kinases responsible for GPCR proteins in Hippo signaling (42, 43), we investigated whether LATS1 is involved in the regulation of YAP by the OXTR. First, we measured the kinase activity of LATS1 immunoprecipitated from SNU761 cells. The overexpression and pharmacologic activation of OXTR via carbetocin inhibited LATS1 kinase activity, which coincided with YAP dephosphorylation (Fig. 6G and H). In addition, overexpression of the wild-type form of LATS1 blocked the dephosphorylation effect of YAP caused by OXTR activation, not the kinase-dead form of LATS1 (Fig. 6I and J). Several studies have highlighted the importance of the Rho/ROCK pathway as a crucial downstream signaling pathway responsive to GPCR activation and as an upstream regulator of the Hippo pathway (44–46). Therefore, we delved deeper into whether Rho GTPase plays a role in the YAP activation induced by the OXTR. Notably, overexpression of Rho-L63, the active mutant of Rho, led to significant dephosphorylation of YAP, regardless of OXTR overexpression and activation. Conversely, botulinum toxin C3, an inhibitor of Rho GTPase, effectively blocked the dephosphorylation effect of YAP caused by OXTR activation (Fig. 6K and L). To further investigate whether ROCK is involved in modulating the Hippo pathway via the OXTR, we evaluated the effects of two ROCK inhibitors, GSK429286A and Y27632 (46). Our immunoblotting analysis revealed that ROCK inhibition significantly impeded the dephosphorylation of YAP, which was triggered by the OXTR (Fig. 6M and N). In summary, our results showed that the OXTR activates YAP through the Gα_q/11–ROCK–LATS axis in HCC.

Figure 6.

The OXTR activates YAP through the Gα_q/11–ROCK–LATS axis in HCC. A and B, OXTR activation decreases YAP phosphorylation through Gα_q/11. SNU761 cells were transiently transfected with siControl, siGα_q/11, or siGα_i. After 24 hours, Flag-OXTR or carbetocin was added. The mixture was incubated for another 24 hours. The levels of YAP, phosphorylated YAP, Gα_q/11, Gα_i, and Flag were determined by immunoblotting. C‒F, OXTR activation induces YAP nuclear localization through Gα_q/11. Endogenous YAP (green) and nuclei (blue) were stained with specific antibodies and DAPI, respectively. Scale bar, 10 μm. Quantification of the subcellular localization of YAP in at least 100 randomly selected cells. D and F are presented as the means ± SDs. ***, P < 0.001 for comparisons (Student t test). ns, nonsignificant. G and H, OXTR activation inhibits LATS1 activity. SNU761 cells were treated via the indicated method. LATS1 was immunoprecipitated. Phosphorylation of YAP by LATS1 was determined with a phospho-YAP antibody. I and J, Ectopic expression of LATS1 blocks YAP dephosphorylation induced by the OXTR. SNU761 cells were transiently transfected with control, LATS1 wild-type (WT), or kinase-dead mutant (K/R) strains. Twenty-four hours later, the cells were treated with the OXTR plasmid or carbetocin for another 24 hours. The phosphorylation and protein levels of YAP were determined by immunoblotting. K and L, Rho GTPase is involved in the dephosphorylation of YAP induced by the OXTR. SNU761 cells were transiently transfected with control, Myc-Rho-L63, or C3. Twenty-four hours later, the cells were treated with the OXTR plasmid or carbetocin for another 24 hours. Total and phosphorylated YAP protein levels were determined by immunoblotting. M and N, ROCK is required for OXTR-induced YAP activation. SNU761 cells were pretreated with the OXTR plasmid or carbetocin for another 24 hours, followed by treatment with GSK429286 (1 mmol/L) or Y27632 (1 mmol/L) for 4 hours. The total and phosphorylated YAP protein levels were determined by immunoblotting. Three independent experiments were conducted to obtain the results.

The OXTR interacts with Gα_q/11 via several important sites (R137, I141, and I227) and facilitates YAP activity via the Gα_q/11–ROCK–LATS axis

Recent advances in structural biology have substantially improved our understanding of GPCR activation mechanisms and interactions with G proteins (47). We further analyzed the overall structure of the OXTR bound to oxytocin (PDB code: 7RYC) via software (https://pymol.org/2/). The results revealed that the OXTR may interact with Gα_q/11 at the I227, I141, and R137 sites (Fig. 7A and B). We further constructed I227A, I141A, and R137A as well as I227A, I141A, and R137A mutation plasmids. Additionally, immunoprecipitation experiments demonstrated that the interaction of the OXTR with Gα_q/11 in HCC cells depends on their I227, I141, R137, and I227 sites (Fig. 7C). Furthermore, the OXTR decreased YAP phosphorylation in HCC cells via its I227, I141, R137, and I227 sites (Fig. 7D). This conclusion was further confirmed by an immunostaining assay and a nuclear‒cytoplasmic separation assay (Fig. 7E‒G). In addition, we also tested the mRNA levels of YAP target genes (CTGF and CYR61) in SNU761 cells and found that overexpression of the OXTR promoted CTGF and CYR61 expression. However, this positive effect on cell migration was diminished or eliminated by overexpressing the I227A, I141A, and R137A or the I227A, I141A, and R137A mutation OXTR plasmids (Fig. 7H). Consistently, the TEAD response element transcriptional activity detected in SNU761 validated this result (Fig. 7I). Similarly, the results of the colony formation, Transwell and wound-healing experiments suggested that the colony formation capacity and migration ability of HCC cells were significantly impeded upon OXTR mutation (Fig. 7J–O).

Figure 7.

The OXTR interacts with Gα_q/11 via several important sites (R137, I141, and I227) and promotes YAP activity via the Gα_q/11–ROCK–LATS axis. A and B, Overall structure of the OXTR bound to oxytocin (PDB code: 7RYC). Oxytocin, the OXTR, Gα, Gβ, and Gγ are yellow, brown, cyan, green, and purple, respectively. The critical interactions between the OXTR and Gα are magnified (right), and the residues referred to in the text are labeled. Software: All the structure figures were analyzed and rendered via PyMOL (https://pymol.org/2/). Four mutant OXTR plasmids were constructed according to the potential binding sites. C, Immunoprecipitation assay showing the interaction between mutant OXTRs and Gα_q/11. D, Immunoblot analysis showing the expression levels of total YAP and p-YAP¹²⁷ in SNU761 cells treated with the indicated plasmids for 48 hours. β-Actin was used as the internal control. E and F, Immunofluorescence staining assay showing the localization patterns of YAP. Endogenous YAP (green) and nuclei (blue) were stained with specific antibodies and DAPI, respectively; scale bar, 10 μm. Quantification of the subcellular localization of YAP in at least 100 randomly selected cells. G, Nucleoplasm separation experiments by immunoblotting confirmed that I227, I141, and R137 of the OXTR were important for promoting YAP nuclear accumulation in MGC SNU761 cells. H, CTGF and CYR61 mRNA levels were determined by qRT-qPCR in SNU761 cells treated with the indicated method for 48 hours. I, TEAD response element transcriptional activity was detected in SNU761 cells treated with the indicated method for 48 hours. J and K, Colony formation (left) of SNU761 cells treated with the indicated method for 24 hours. L‒O, Transwell and wound-healing assays were used to detect the migration ability of SNU761 cells treated with the indicated method for 24 hours. All data are presented as the means ± SDs. ns, nonsignificant; *, P < 0.05; **, P < 0.01; ***, P < 0.001 for comparisons (Student t test).

YAP transcriptionally regulates OXTR expression, forming a feedback loop between YAP and the OXTR

Given the pivotal role of YAP in governing tumor progression, we performed a series of studies to investigate its global genomic binding patterns. We first analyzed the whole-genome expression profile of HCC samples from the TCGA database, and the analysis revealed that the OXTR was positively correlated with CTGF (R = 0.344) and CYR61 (R = 0.476; Fig. 8A). We further analyzed multiple YAP-based ChIP-seq datasets (GSE66081, GSE61852, GSE107013, and GSE131687) and observed four potential binding peaks around the OXTR gene, three of which shared binding peaks with enriched TEAD and H3K27ac (sites 1, 3, and 4; Fig. 8B). This might reflect the potential active binding sites at sites 1, 3, and 4. To determine the functional binding sites, we utilized a novel technology called the dCas9-KRAB strategy with the help of Dr. Guo (30). Such an experiment could introduce a transcriptional repressor to a certain domain of DNA and identify the critical binding site for OXTR expression. Our data revealed that each YAP-binding site was effectively blocked by dCas9-KRAB (Fig. 8C). However, OXTR mRNA levels were significantly decreased in the CRISPRi cell lines at sites 3 and 4 but not at sites 1 and 2 (Fig. 8D). Therefore, sites 3 and 4 of the OXTR enhancer region are the functional binding sites for the YAP protein.

Figure 8.

YAP transcriptionally regulates OXTR expression, which forms a positive feedback loop between Hippo/YAP and the OXTR. A, OXTR expression was significantly correlated with that of CCN1 (CYR61) and CCN2 (CTGF) in HCC samples from the TCGA database. B, OXTR genome schematic and database analysis of the binding region of YAP or TEAD to the OXTR. Sites 1‒4 CRISPRi sgRNAs were designed to target the four OXTR sites. C and D, YAP ChIP‒qPCR and qRT-qPCR with or without CRISPRi in YAP-overexpressing SNU449 cells. E‒G, qRT-qPCR indicated that siRNA, VP, or VT104 treatment via the indicated method in SNU449 cells decreased the CTGF and CYR61 mRNA levels. H‒J, Immunoblot analysis showing the expression level of the OXTR in SNU449 cells treated with siYAP, VP, or VT104 via the indicated methods. Treatment inhibited OXTR protein expression and mRNA expression. K‒M, RT-qPCR showing the mRNA expression levels of the OXTR with siYAP, VP, or VT104 treatment via the indicated method in SNU449 cells. N and O, Immunofluorescence imaging of the OXTR (red) and DAPI (blue) in SNU449 cells subjected to the indicated treatments. Scale bar, 10 μm. All data are presented as the means ± SDs. ns, nonsignificant; **, P < 0.01; ***, P < 0.001 for comparisons (Student t test).

Previous studies have shown that verteporfin and VT104 impede the functionality of YAP/TAZ/TEAD (41, 48). Notably, the mRNA levels of CTGF and CYR61 decreased in SNU449 cells treated with siYAP, VP, or VT104 (Fig. 8E–G). The expression levels of OXTR mRNA and protein in SNU449 cells significantly decreased following siYAP, VP, or VT104 treatment (Fig. 8H–M). This finding was further confirmed by immunostaining (Fig. 8N and O).

Discussion

In the present study, we discovered that the OXTR plays an important role in the regulation of Hippo signaling activity, which is important for HCC formation and progression. A cross-comparison of 134 GPCRs targeted by drugs approved in the United States and 75 GPCRs strongly associated with the Hippo/YAP pathway identified by GSEA and unbiased siRNA GPCR screening revealed the OXTR as a critical GPCR for Hippo signaling in HCC. Further experiments demonstrated that the activation of the OXTR facilitates the nuclear translocation of YAP and its target gene expression via the modulation of YAP phosphorylation. A mechanistic study revealed that the OXTR promotes YAP function via the Gα_q/11–ROCK–LATS axis, whereas blockade of the OXTR via the obstetric drug atosiban inhibited YAP function and HCC progression via in vitro, organoid, and patient-derived explant assays and genetic models. Surprisingly, the OXTR is a direct target gene of the Hippo/YAP axis, indicating that the interaction between the OXTR and YAP coordinates with Hippo pathway activity during HCC progression and that targeting the OXTR inhibitor atosiban could be a novel strategy for treating HCC.

The biological link between GPCRs and the Hippo pathway was established in 2012 by Professor Kunliang Guan (19). The GPCR family is a large group of cell-surface receptors that comprises approximately 900 members (21). GPCRs receive extracellular signals, such as hormones, peptides, and neurotransmitters, whereas the activation of GPCRs causes conformational changes and acts as a guanine nucleotide exchange factor to facilitate the exchange of GDP with GTP (49). The further effect of GPCRs depends on the subtype of Gα subunits, including Gα_s, Gα_i/o, Gα_q/11, and Gα_12/13 (50, 51). The OXTR, a 7-transmembrane GPCR capable of binding to either Gα_i or Gα_q proteins, activates a set of signaling cascades, such as the MAPK, PKC, PLC, or CaMK pathways (23). Although the overall regulatory effect of GPCRs on the Hippo pathway has been revealed, the specific functions of most GPCR members in Hippo signaling and HCC formation are still unclear. In another way, GPCRs can either suppress or enhance YAP function, and identifying the regulatory direction of the Hippo pathway for each member could be difficult. To address these questions, we carried out GPCR siRNA screening for crucial hits in Hippo regulation and HCC formation. Our data revealed that the OXTR is not only an important upstream activator of YAP and HCC via the Gα_q/11–ROCK–LATS axis but also a downstream effector of YAP activation. Moreover, we further analyzed the correlation between the OXTR and YAP in a series of human cancers. Interestingly, a notable positive association was found between OXTR expression and the YAP target gene signature in PAAD, KIRC, COAD, and LUAD. Therefore, we believe that the effect of the OXTR on the Hippo/YAP axis could be a general regulatory phenomenon in a series of human malignancies.

Moreover, we revealed that the OXTR facilitates HCC progression using activation of the Hippo/YAP axis. Furthermore, as shown in Fig. 1J, we also observed that the OXTR could influence metabolic pathways. Cross-talk between the Hippo pathway and metabolic pathways is also possible. Specifically, the downstream effectors of the Hippo pathway, YAP and TAZ, have been shown to interact with metabolic pathways such as glycolysis, glutaminolysis, and lipid metabolism (52, 53). For example, YAP/TAZ activation can promote glycolysis by regulating the expression of glycolytic enzymes, thereby supporting the metabolic demands of rapidly proliferating cells (54). In addition, Hippo signaling may influence metabolic reprogramming through its cross-talk with other signaling pathways, such as the mTOR and AMPK pathways, which are central to metabolic regulation. For example, YAP/TAZ activation has been linked to increased mTOR activity, enhancing anabolic processes such as protein and lipid synthesis (55). Conversely, inactivation of the Hippo pathway can lead to metabolic quiescence, which aligns with its role in maintaining tissue homeostasis. Further investigations into the metabolic implications of the OXTR/Hippo axis are warranted.

The positive feedback loop is a fundamental mechanism that increases signal transduction and is widely observed in human malignancies. One classic example is hypoxia-inducible factor (HIF) signaling in carcinogenesis, in which the growth of cancer cells results in a hypoxic microenvironment, whereas a lack of oxygen promotes HIF signaling activation, which subsequently increases HIF target gene expression and cancer progression (56). Precision therapy targeting HIF signaling has achieved great success in recent years, and several clinical trials on HIF inhibitors are ongoing. Therefore, the reciprocal promotion mechanisms between these two oncogenic pathways could be important in carcinogenesis and ideal targets for cancer treatment. In our study, we revealed that the unexpected feedback loop between the OXTR and the Hippo/YAP axis might have important implications for cancer biology. Notably, blocking the regulatory loop using atosiban markedly inhibited YAP function and HCC progression both in vitro and in vivo, which might indicate that targeting the YAP–OXTR interaction could be a promising strategy for HCC therapy. The regulation of Hippo signaling has been confirmed in only a limited number of GPCR members, therefore, whether reciprocal regulation between the OXTR and the Hippo pathway is common in other GPCR proteins or whether such regulation between GPCR and Hippo signaling is a common phenomenon among other types of membrane receptors warrants further investigation.

On the basis of the critical function of Hippo signaling in HCC, targeting the Hippo pathway is a promising strategy for treating HCC. However, effective inhibitors that target the Hippo/YAP axis are still unavailable in the clinic. Several preclinical drugs, such as verteporfin and Super-TDU (57, 58), which have been proposed to block the Hippo/YAP axis, have failed several preclinical studies on Hippo-driven cancer, but the reasons are not completely clear. For example, YAP inhibitors can penetrate the cytosol to block protein interactions (59). In addition, the stability of inhibitors or potential cellular toxicity in vivo might also be the rate-limiting factor for clinical applications. Therefore, we shifted our strategy to inhibiting the upstream GPCR of the Hippo/YAP axis in HCC. Historically, GPCR targeting has been proven to have a high success rate in pharmaceutics. With advances in GPCR pharmacology, GPCR target drugs constitute one thirds of the FDA-approved drug list and half of the drugs sold on the market (36, 60). In principle, the modulation of GPCRs with allosteric sites could effectively change the structure and function of the receptor and achieve potent therapeutic advantages, including enhanced selectivity, without worrying about the penetration rate across the cell membrane. Compared with intracellular inhibitors, targeting GPCRs is more selective, as they can preferentially activate the desired signaling pathway while avoiding or minimizing the side effects on other pathways. One example is the OXTR inhibitor atosiban, which has been successfully utilized in obstetrics to block OXTR signaling but has few side effects. As blocking the YAP/TEAD interaction could be difficult with pharmaceutical strategies, further investigations of potential upstream GPCR targets through library screening could open new avenues for HCC therapy.

Supplementary Material

Supplementary Figure 1

Identification of the OXTR as a novel mediator of the Hippo/YAP pathway

Supplementary Figure 2

OXTR depletion slows hepatocellular carcinoma progression in HCC cells

Supplementary Figure 3

The OXTR antagonist atosiban inhibits hepatocellular carcinoma progression

Supplementary Figure 4

Atosiban inhibits hepatocellular carcinoma progression in patient-derived explant and organoid models

Supplementary Figure 5

OXTR overexpression and pharmacological activation of OXTR promote hepatocellular carcinoma progression in vivo and in vitro

Supplementary Figure 6

The OXTR activates the Hippo/YAP axis by decreasing YAP phosphorylation

Supplementary Figure 7

The OXTR facilitates hepatocellular carcinoma progression via activation of the Hippo/YAP axis

Supplementary Table 1

Sequences of siGPCRs screen

Supplementary Table 2

Sequences

Supplementary Table 3

Chemicals

Supplementary Table 4

Primer sequences for RT-qPCR

Supplementary Table 5

Antibodies

Authors’ Disclosures

No disclosures were reported.

Authors’ Contributions

H. Yang: Data curation, funding acquisition, writing–original draft, writing–review and editing. J. Cui: Data curation. P. Su: Data curation. X. Cui: Data curation. H. Guo: Data curation. P. Yang: Data curation. S. Zhang: Data curation. C. Zhang: Data curation. M. Fu: Data curation. Z. Li: Resources, data curation, methodology. Y. Ding: Project administration, writing–review and editing. T. Zhuang: Funding acquisition, validation, project administration, writing–review and editing. J. Zhu: Funding acquisition, writing–original draft, project administration, writing–review and editing. X. Tan: Validation, writing–original draft.