NEXN targeting MYOCD by facilitating EMT-related β-catenin nuclear translocation modulates the metastasis of hepatocellular carcinoma.

Introduction
Liver cancer is one of the most prevalent malignant tumors, ranking third in global cancer-related deaths. Based on current research findings, the five-year survival rate for individuals diagnosed with liver cancer ranges from 5% to 30%. Furthermore, the global incidence of liver cancer is projected to rise from approximately 870,000 new cases in 2022 to an estimated 1.52 million by the year 2050.1,2 Hepatocellular carcinoma (HCC), the most common type of liver cancer, accounts for approximately 80% of primary liver cancers. Currently, surgical resection is the preferred treatment for stage IIa liver cancer (China Liver Cancer Staging, CNLC), while systemic therapy combined with radiofrequency ablation is commonly employed for patients with stages IIb and III liver cancer. Recent advancements in systemic therapies, such as sorafenib, regorafenib, and the combination of bevacizumab and atezolizumab, have offered modest survival benefits for patients with advanced HCC.3,4 However, the dysregulation of various genetic and epigenetic signaling pathways triggers the transformation of normal hepatocytes into malignant phenotypes, resulting in rapid disease progression. Liver cancer is characterized by high malignancy, significant tissue heterogeneity, and rapid progression. Therefore, ongoing research into the mechanisms behind HCC progression and metastasis is crucial for identifying effective biomarkers for diagnosing metastasis and extending the overall survival (OS) of patients with liver cancer.
NEXN is an F-actin-binding protein involved in cell adhesion and migration. Its deficiency can lead to the disorganization of the Z-disc structure in cardiomyocytes, resulting in dilated cardiomyopathy.5,6 Studies have shown that, as a direct substrate of CLK4, NEXN phosphorylation can alleviate cardiac hypertrophy induced by CLK4 knockout in mouse cardiomyocytes.7 This highlights NEXN as a potential target for preventing and treating heart failure. Furthermore, the long non-coding RNA NEXN-AS1 reportedly interacts with the 5′ flanking region of NEXN to upregulate its expression, thereby inhibiting the development of atherosclerosis.8 In cancer research, high NEXN expression was reportedly associated with shorter patient survival in osteosarcoma.9 Zhang et al. found that ceRNA has-miR-590-3p promoted melanoma progression by inhibiting NEXN expression.10 However, studies on the prognostic value and functional mechanisms of NEXN in liver cancer remain scarce.
The Wnt signaling pathway plays an important role in regulating the proliferation, differentiation, apoptosis, migration, invasion, epithelial-mesenchymal transition (EMT), and stemness in various malignant tumors, including colorectal, liver, non-small cell lung, and triple-negative breast cancers, where the abnormal expression of the WNT pathway has been observed.11 When Wnt signaling is active, Wnt binds to the Frizzled receptor and activates a Disheveled protein, thereby enhancing GSK3β phosphorylation. This causes unphosphorylated β-catenin to accumulate in the cytoplasm and translocate to the nucleus. In the nucleus, elevated β-catenin binds to TCF/Lef, driving transcriptional changes and activating downstream target genes such as matrix metalloproteinases-7 (MMP-7), E-cadherin, vimentin, cyclin-D1, FN, Slug, c-Myc, and Axin.11,12 Additionally, β-catenin mediates cell-cell adhesion by binding to the intracellular domain of E-cadherin at the cell membrane, where it interacts with α-catenin and actin to form the E-cadherin-actin complex.13,14 This complex anchors β-catenin to the membrane and facilitates intercellular adhesion, a process that is essential for maintaining epithelial cell polarity and integrity. This study aimed to elucidate the functional mechanisms of NEXN in HCC progression.

Results

Nexilin F-actin binding protein is mildly expressed in hepatocellular carcinoma and correlates with poor prognosis
The role of NEXN in liver cancer progression was first explored by examining its expression in 85 paired liver cancer and adjacent non-tumor tissues. NEXN exhibits heterogeneity in its cellular distribution. In this study, as illustrated in Figure 1A, we observed that NEXN was highly expressed in hepatocellular carcinoma, predominantly localized within the cytoplasm of these carcinoma cells (indicated by arrows in the figure). Conversely, in the adjacent non-cancerous tissues, NEXN was present in smaller quantities, dispersed within both the cytoplasm and interstitial tissues, without any discernible pattern of concentrated distribution. Paired analysis of cytoplasmic IHC staining (Figure 1B) revealed that NEXN expression was significantly lower in the tumor tissues of most patients with HCC compared to adjacent non-tumor tissues. Further validation at the mRNA level using TCGA database data confirmed that NEXN expression was significantly higher in normal liver tissues (n = 160) than in tumor tissues (n = 371, p < 0.001), as depicted in Figure 1C (Refer to Data S1 for raw data specifics).
Next, Kaplan−Meier survival analysis and log rank tests were conducted to assess the association between NEXN expression and OS, DFS, and progression-free interval (PFI) in patients with HCC. As shown in Figures 1D–1G, patients with low NEXN expression had significantly shorter OS, DFS, and PFI compared to those with higher expression (p < 0.05). Univariate Cox analysis revealed that NEXN ectopic expression, tumor size (≥5 cm), T stage (T1/T2), TNM stage (TNM1/TNM2), and gamma-glutamyl transpeptidase (GGT) levels influenced OS in patients with HCC (Table S2), while DFS was influenced by NEXN expression, tumor size (≥5 cm), alanine transaminase, and GGT levels (Table S3). Multivariate Cox regression analysis highlighted NEXN as the only independent prognostic factor significantly impacting OS and DFS (hazard ratio = 0.423, 0.380; 95% confidence interval: 0.223–0.802, 0.207–0.698, respectively) in patients with HCC. Nevertheless, the underlying mechanisms through which NEXN affects HCC progression and prognosis require further investigation.

Nexilin F-actin binding protein regulates the proliferation, migration, and invasion of hepatocellular carcinoma cells
To evaluate the role of NEXN in HCC cells, stable NEXN-FLAG overexpression models were created in Huh7-NC/Huh7-NEXN and BEL-7402-NC/BEL-7402-NEXN cell lines (Figures S1A and S1B), the successful establishment of which was confirmed by WB. As depicted in Figures 2A and 2B, compared to the control group, Huh7-NEXN and BEL-7402-NEXN cells exhibited reduced proliferation rates, significantly delayed growth curves, and substantially extended PDT (38.74 h vs. 59.53 h and 40.69 h vs. 60.87 h, respectively). Additionally, NEXN overexpression inhibited colony formation, evidenced by a reduction in the number of colonies (Figure 2C).
In this study, NEXN overexpression significantly suppressed the wound healing rate of HCC cells at 24 and 42 h after scratch (Figures 2D and 2F), suggesting compromised lateral migration. Transwell assays further revealed mitigated vertical migration and invasion (Figures 2E and 2G) in Huh7-NEXN and BEL-7402-NEXN cells compared to their respective control groups. Additionally, we effectively reduced NEXN expression in the Huh7-NEXN hepatoma cell line and confirmed the protein-level knockdown. The NEXN-Sh3 sequence demonstrated the highest knockdown efficiency (see Figure S1D) and was utilized in subsequent recovery experiments. Rescuing NEXN expression in stable NEXN-overexpressing HCC cells showed that knockdown of NEXN reversed the inhibitory effects of NEXN overexpression on lateral migration, vertical migration, and invasion, manifesting as increased migration and invasion abilities after NEXN knockdown (Figures 2H and 2I). Collectively, these findings indicate that NEXN plays a regulatory role in HCC cell proliferation, migration, and invasion at the cellular level.

Nexilin F-actin binding protein overexpression suppresses the tumorigenicity of hepatocellular carcinoma cells in vivo
To further verify the regulatory role of NEXN in HCC progression, a humanized nude mouse xenograft model was established (Figure 3A). As shown in Figures 3B and 3D, Huh7-NC cells, with normal HCC cell characteristics, rapidly formed tumors when transplanted subcutaneously into nude mice. In contrast, the tumorigenic capacity of Huh7-NEXN cells overexpressing NEXN was significantly inhibited, as evidenced by a marked reduction in tumor growth rate, smaller tumor volume, and decreased tumor weight (Figures 3C and 3D). Immunohistochemical staining further confirmed higher NEXN protein expression in the NEXN-overexpressing xenografts, predominantly in the cytoplasm. Correspondingly, the intensity of KI67 nuclear protein staining was significantly lower in the NEXN-overexpressing tumors compared to the control group (Figure 3F). These results indicate that NEXN suppresses the tumorigenicity of HCC cells in vivo.

Nexilin F-actin binding protein overexpression modulates hepatocellular carcinoma epithelial-mesenchymal transition by inhibiting nuclear β-catenin translocation
WB was conducted to assess the effect of NEXN overexpression on EMT-related proteins in HCC. The results (Figures 4A and 4B) indicated that N-cadherin, β-catenin, Vimentin, and MMP2 expression levels were reduced in Huh7-NEXN cells. Similarly, in BEL-7402-NEXN cells, a decreased expression of mesenchymal proteins N-cadherin, β-catenin, and MMP2 was observed, accompanied by an increased expression of epithelial protein E-cadherin (Figures 4C and 4D). Additionally, stable knockdown of NEXN in NEXN-overexpressing cell lines was conducted to evaluate its role in regulating the HCC EMT phenotype. As depicted in Figure 4E, NEXN knockdown in Huh7-NEXN cells restored N-cadherin, β-catenin, vimentin, and MMP2 expression levels (Figure 4F).
KEGG enrichment analysis was performed using TCGA data to explore how NEXN regulates HCC migration and invasion. The KEGG enrichment analysis identified a significant correlation between NEXN expression and the Wnt signaling pathway (Figure 4G). β-catenin, a key protein in the Wnt pathway, is essential for regulating EMT nuclear transcription. Additionally, its nuclear accumulation is crucial for Wnt/β-catenin signaling activation. We found that in Huh7 cells, β-catenin was predominantly localized in the cytoplasm and partially in the nucleus, whereas in BEL-7402 cells, β-catenin was mainly in the nucleus with some presence in the cytoplasm. However, NEXN overexpression significantly reduced nuclear β-catenin levels in both cell lines (Figures 4H–4K). Immunofluorescence further revealed that NEXN overexpression substantially inhibited the nuclear accumulation of β-catenin, with β-catenin being predominantly localized to the cell membrane and cytoplasm (Figures 4L and 4M). Collectively, these results suggest that NEXN modulates β-catenin nuclear accumulation in HCC, which may serve as a key mechanism in promoting liver cancer progression.

Nexilin F-actin binding protein binds to MYOCD to modulate the epithelial-mesenchymal transition process in hepatocellular carcinoma by regulating β-catenin nuclear translocation
Proteomic analysis was conducted to investigate the differentially expressed genes regulated by NEXN in HCC. As shown in Figure 5A, NEXN overexpression upregulated MYOCD, confirmed in both in vivo animal models (Figure S2A) and in vitro cell experiments (Figure 5B). We also found a strong positive correlation between MYOCD and NEXN expression in HCC (Figure 5C), as identified in the TCGA database (R = 0.645, p < 0.001). Next, the expression of MYOCD was consistently suppressed in NEXN-overexpressing hepatocellular carcinoma (HCC) cells. The MYOCD-Sh1 sequence was effectively silenced and utilized in subsequent experimental analyses (Figure S2B). As depicted in Figure 5D, NEXN overexpression significantly upregulated MYOCD protein levels, while MYOCD knockdown did not affect NEXN expression, suggesting that MYOCD functions downstream of NEXN. Moreover, knocking down MYOCD restored the colony formation, migration, and invasion capabilities of HCC cells (Figures 5H and 5I). Meanwhile, WB results demonstrated that MYOCD silencing reversed the suppression of mesenchymal proteins such as N-cadherin, β-catenin, vimentin, and MMP-2 (Figure 5F). Collectively, these findings indicate that MYOCD acts as a downstream effector of NEXN in regulating the EMT process of HCC.
To investigate the interaction between NEXN and MYOCD, Co-IP was performed by transiently co-transfecting NEXN-FLAG and MYOCD-HA plasmids in 293T cells in vitro. The results confirmed the presence of MYOCD protein with an HA tag in the FLAG pull-down complex, indicating an interaction between NEXN and MYOCD (Figure 5J). Immunofluorescence experiments further revealed partial co-localization of NEXN and MYOCD in the cytoplasm of Huh7-NEXN xenograft cells (Figure 5K).
Additionally, in NEXN-overexpressing, MYOCD-knockdown HCC cells, nuclear-cytoplasmic fractionation revealed that silencing MYOCD restored β-catenin nuclear translocation (Figures 6A and 6B). Immunofluorescence results also demonstrated that MYOCD knockdown induced the relocalization of β-catenin from the cell membrane to the nucleus (Figures 6C and 6D). In summary, NEXN binds to MYOCD to regulate β-catenin nuclear translocation, thereby influencing the EMT process in HCC (Figure 5L).

Nexilin F-actin binding protein regulates hepatocellular carcinoma prognosis in conjunction with MYOCD
Our previous study found that NEXN activates the Wnt signaling pathway by modulating β-catenin nucleation through its interaction with MYOCD. Subsequently, we examined the impact of MYOCD on the prognosis of hepatocellular carcinoma. Kaplan−Meier survival analysis using TCGA data indicated that patients with high MYOCD expression had longer OS, disease-specific survival (DSS), and PFI compared to patients with low MYOCD expression (Figures 6E–6G). Further co-expression analysis of NEXN and MYOCD revealed that patients with high expressions of both genes had significantly longer OS, DSS, and PFI compared to other groups (Figures 6H–6J), with statistical significance. These results suggest that elevated NEXN and MYOCD expression extends HCC prognosis.

Discussion
NEXN encodes a filamentous actin-binding protein. Gene ontology analysis suggests that this gene is involved in binding to the actin cytoskeleton and exhibits calmodulin-dependent protein kinase activity, potentially contributing to cell adhesion, cell motility, cell migration, and interactions. Numerous studies have shown that actin-binding proteins play a role in tumor growth, invasion, and metastasis, making them valuable as potential cancer biomarkers and therapeutic targets.15 For instance, fascin actin-bundling protein 1 (FSCN1) is overexpressed in over 80% of bladder and head and neck cancers, 27% of breast cancers, and 33% of gastric cancers, influencing the distant metastasis and prognosis of malignant tumors by regulating key pathways such as EMT, PI3K/AKT, WNT/β-catenin, and MAPK.15 NEXN was expressed at lower levels in papillary thyroid microcarcinoma than in normal thyroid tissue16 but was highly expressed in osteosarcoma,9 where it was associated with poor prognosis. This finding demonstrates its expression heterogeneity across cancer types.
In this study, through both clinical samples and analysis of the TGCA database, NEXN was confirmed to be downregulated in HCC, and its low expression was significantly associated with shorter OS, DFS, and PFI. NEXN was also identified as an independent prognostic factor for both OS and DFS in patients with HCC. Furthermore, NEXN overexpression significantly inhibited HCC cell proliferation, colony formation, migration, and invasion in vitro, along with suppressing tumor growth in vivo. Collectively, these results suggest that NEXN may play a key role in inhibiting HCC progression and metastasis by functioning as a tumor suppressor gene.
In tumor cells, the Wnt/β-catenin pathway induces EMT by upregulating mesenchymal markers such as Slug, Zeb1, and Twist, inhibiting E-cadherin-mediated adhesion, and promoting the expression of MMPs, including MMP-2, MMP-3, and MMP-7, to remodel the extracellular matrix.17 In this study, NEXN overexpression inhibited N-cadherin, vimentin, and MMP2 expression in Huh7 cells and reduced N-cadherin and MMP2 levels in BEL-7402 cells while upregulating E-cadherin expression on the cellular level. Furthermore, β-catenin protein levels were decreased. Based on these results, we speculate that NEXN may regulate the EMT process in HCC by modulating β-catenin expression and its translocation to the nucleus.
Abnormal activation of the Wnt/β-catenin signaling pathway and the nuclear accumulation of β-catenin are closely associated with the progression of gastrointestinal cancers. Among patients with HCC, about 30% exhibit mutations in the CTNNB1 gene, 8% in the Axin1/2 genes, and 3% in the APC gene.18 In HCC, β-catenin accumulated in the nucleus binds to TCF-4 to form a β-catenin/TCF-4 complex.19 This complex attaches to the TCF-4 binding site on the c-Myc gene promoter and induces abnormal c-Myc expression, inhibiting apoptosis and promoting rapid cell proliferation. Additionally, the β-catenin/TCF-4 complex stimulates VEGF transcription by interacting with a TCF-4 binding site on the VEGF gene promoter, further promoting angiogenesis.20 In this study, we observed that in HCC cells, β-catenin was predominantly localized in the nucleus and cytoplasm, while NEXN overexpression significantly inhibited its nuclear accumulation, shifting β-catenin to the cell membrane. Notably, in normal hepatocytes, β-catenin was mildly expressed and was primarily located on the cell membrane at low levels. Conversely, in cirrhosis and HCC, β-catenin expression on the membrane decreased while its cytoplasmic and nuclear levels increased. These findings suggest that NEXN, as a tumor suppressor, induces the cytoplasmic accumulation of non-phosphorylated β-catenin, its nuclear translocation, and the subsequent activation of downstream genes such as N-cadherin, MMP2, and MMP9 by activating the classical Wnt/β-catenin pathway, ultimately regulating the initiation and development of HCC.
In addition to its transcriptional activation role in the cytoplasm, β-catenin interacts with E-cadherin at the cell membrane to contribute directly to adherens junctions.21,22 The E-cadherin/catenin complex is involved in the initiation and development of HCC. Normally, E-cadherin connects to the actin cytoskeleton through the β-catenin/α-catenin complex, stabilizing epithelial cell adhesion. However, when β-catenin translocates from the membrane to the nucleus, E-cadherin-mediated adhesion weakens while the transition from E-cadherin to N-cadherin increases. In this study, we found that as an actin-binding protein, NEXN participated in cell adhesion, motility, and migration by interacting with the actin cytoskeleton. Additionally, NEXN overexpression led to increased E-cadherin levels and decreased N-cadherin expression in hepatocytes. These results suggest that NEXN may influence the catenin-actin interaction and regulate the expression of adhesion proteins, thereby contributing to the EMT process in liver cancer.
Proteomics and related technologies were utilized to reveal that NEXN regulation in HCC induced the ectopic expression of MYOCD. MYOCD is a cofactor of the transcription factor SRF, essential for the development and differentiation of cardiac and smooth muscle cells. MYOCD promotes muscle-specific gene expression in an SRF-dependent manner, playing important roles in cardiac contraction, smooth muscle differentiation, atherosclerosis, and cardiovascular diseases.23 Additionally, it has previously been demonstrated that MYOCD functions as a tumor suppressor by inhibiting the proliferation of cancer cells through its role as a transcriptional cofactor for SRF.24 For example, Liao et al. discovered that MYOCD induced apoptosis in breast cancer MCF-7 cells by upregulating Maspin expression.25 In smooth muscle sarcoma,26 MYOCD inhibits cell proliferation by activating p21 expression through SRF binding to the CArG sequence. However, the role of MYOCD in liver cancer has not been reported. To confirm its regulatory function on HCC prognosis, data from the TCGA database were analyzed, revealing that low MYOCD expression was associated with significantly shorter OS, DSS, and PFI compared to high expression. Additionally, the co-expression of NEXN and MYOCD led to markedly improved OS, DSS, and PFI in patients with HCC, suggesting that NEXN regulates the prognosis of HCC in conjunction with MYOCD. These findings offer new candidate molecules for predictive models of liver cancer prognosis and recurrence risk and provide a theoretical basis for identifying effective diagnostic markers for HCC metastasis.
Based on the synergistic effect of NEXN and MYOCD on liver cancer prognosis, further investigation into the functional role of MYOCD in HCC was carried out at the cellular level. The results revealed that NEXN overexpression upregulated MYOCD protein levels in HCC cells, with consistent results observed in in vivo animal models. Silencing MYOCD in NEXN-overexpressing HCC cells restored cell cloning and phenotypic characteristics, while no changes were observed in NEXN expression. This indicates that MYOCD may act as a downstream target of NEXN in regulating the initiation and development of liver cancer. Meanwhile, we found that NEXN interacted with MYOCD, and they exhibited partial co-localization in the cytoplasm. Notably, previous research reported that MYOCD, a transcriptional cofactor, is mainly expressed in the nucleus and partially in the cytoplasm in normal cardiac and smooth muscle tissues. Conversely, Tong et al.27 reported that in non-small cell lung cancer, MYOCD was predominantly localized in the cytoplasm and interacted with SMAD3 to regulate tumor metastasis. This suggests a shift in MYOCD distribution in tumor cells, which aligns with our findings. Additionally, we found, for the first time, that NEXN induced changes in β-catenin nuclear translocation. In contrast, when MYOCD expression was downregulated in NEXN-overexpressing hepatocytes, β-catenin shifted from the cell membrane to the cytoplasm and nucleus. This suggests that NEXN may regulate the EMT process in HCC through the NEXN-MYOCD-β-catenin signaling axis.
In summary, this study is the first to confirm that NEXN, in conjunction with MYOCD, modulates the adverse prognosis of HCC. Functionally, NEXN modulates the proliferation, migration, and invasion of HCC cells by binding to MYOCD and regulating the WNT/β-catenin signaling pathway. The Wnt signaling pathway is recognized for its role in regulating the proliferation, migration, invasion, and epithelial-mesenchymal transition (EMT) processes of hepatocellular carcinoma (HCC) cells, primarily through the canonical β-catenin-dependent pathway. Notably, missense mutations in the AXIN1 gene have been identified in 25% of HCC cases, leading to a loss or reduction in AXIN protein expression,28 which subsequently results in the abnormal nuclear accumulation of β-catenin. Molecular agents targeting Wnt ligands/receptors, such as OMP-54F28, β-catenin/TCF complexes, such as CWP232291, and Tankyrase, such as XAV939, have demonstrated potential in preclinical studies. Furthermore, drugs targeting mutations in APC or β-catenin, such as ETC-159, have progressed to clinical trials for gastrointestinal tumors.29,30 This study underscores the role of NEXN as a critical regulator of EMT in HCC, focusing on the NEXN-MYOCD-β-catenin signaling axis. It enriches the understanding of the activation mechanisms of the Wnt pathway in HCC and provides a preclinical foundation for the development of inhibitors targeting Wnt pathway-mediated metastasis in HCC.

Limitations of the study
Limitations of the study need further research: identifying suitable HCC cell lines and creating a nude mouse model to understand NEXN’s role in HCC recurrence and metastasis, and updating follow-up data to analyze NEXN expression in metastatic liver cancer. Additionally, the NEXN-MYOCD interaction in the Wnt signaling pathway requires more exploration.

Resource availability

Lead contact
For additional information and requests for resources and reagents, please contact the lead researcher, Dr. Wu Qiong, at wuqiong@zjcc.org.cn, who will facilitate these inquiries.

Materials availability
New, unique materials were not generated in the course of this study.

Data and code availability

•This article does not report the original code.

•Additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments
This research was supported by grants from the Key Project of the 10.13039/501100004731Natural Science Foundation of Zhejiang Province (Grant no. LZ20H290001), the 10.13039/501100004731Zhejiang Provincial Natural Science Foundation of China (Grant no. LQ23H160008), and the 10.13039/501100017594Medical and Health Science and Technology Project of Zhejiang Province (Grant no. 2025KY711).

Author contributions
Q.H.Y., Q.W., and J.H. contributed to the study conception and design; Q.W., X.L., Y.X., and H.W. performed the experiments; Q.W., X.L., X.F.X., and B.Z. analyzed and interpreted the data and statistical analysis; Q.W. wrote the article; X.L. and Q.H.Y. edited the article. All the authors approved the final version and agreed to publication.

Declaration of interests
The authors stated that they had no conflicts of interest.

STAR★Methods

Key resources table

Experimental model and study participant details

Cell lines and cell culture
The human HCC cell line BEL-7402 was cultured in Roswell Park Memorial Institute 1640 medium containing 10% fetal bovine serum (FBS, Gibco, NY, USA). Alternatively, the human HCC cell line Huh7 and the human embryonic kidney cell line 293T were cultured in Dulbecco’s Modified Eagle Medium containing 10% FBS (Gibco, NY, USA). All cell lines were purchased from the Shanghai Institute of Biochemistry and Cell Biology and were authenticated using short tandem repeat profiling. Mycoplasma contamination was tested in all cell lines using polymerase chain reaction.

In vivo animal experiments
Huh7 cells expressing either NC or NEXN (∼2×106 cells/mouse) were subcutaneously injected into the right armpits of 6–8-week-old male BALB/c nude mice (Experimental Animal Center, Hangzhou Medical College, Hangzhou, China). There are 8 mice in each group (n = 8). Tumor volume was measured every 3 days using calipers and calculated with the formula: Volume (mm3) = (L × W2)/2, where L represents the tumor length and W represents the tumor width. All animal experiments were approved by the Animal Ethics Committee of Zhejiang Laboratory Animal Center.

Ethics approval and consent to participate
The Ethics Committee of Shanghai Outdo Biotech Company gave its approval to this project involving human participants (SHYJS-CP-1607007). All animal experiments were approved by the Animal Ethics Committee of Zhejiang Laboratory Animal Center (ZJCLA-IACUC-20020142).

Method details

Clinical samples and immunohistochemistry
Precisely 90 human liver cancer tissue samples (HLivH180Su07) and paired adjacent non-cancerous tissues were obtained from Shanghai Outdo Biotech Company. Informed consent was obtained from all patients, and the study was approved by the Ethics Committee (the ethical approval number: SHYJS-CP-1607007). NEXN protein expression in the liver cancer tissue array was assessed by immunohistochemistry (IHC) (following previously described STAR methods).31 The samples were subsequently incubated with specific rabbit polyclonal antibodies against NEXN (1:1000, NBP1-92179, NOVUS) overnight at 4°C. Two experienced pathologists independently scored the samples based on the percentage of positive staining and staining intensity. Staining intensity was graded on a scale of 0 (negative), 1 (+), 2 (2+), and 3 (3+). Similarly, the percentage of positive staining was categorized as 0 (negative), 1 (1−25%), 2 (26−50%), 3 (51−75%), and 4 (76−100%). The total score was then calculated as the product of the staining intensity and percentage scores. Scores <8 were classified as low NEXN expression, while scores ≥8 were considered high NEXN expression. Additionally, Kaplan−Meier analysis and log rank tests were conducted to evaluate the OS and disease-free survival (DFS) of patients with HCC. Five sample pairs were excluded from the paired analysis of NEXN expression in HCC and adjacent tissues due to tissue loss or detachment. The clinical characteristics data of patients are detailed in Table S1.

Transfection and viral infection
NEXN shRNA and MYOCD shRNA target sequences were respectively cloned into the lentiviral vector pLent-U6-shRNA-CMV-copGFP-P2A-Blasticidin at the BamHI/MIUI sites (Vigene Bioscience, MD, USA). The shRNA sequences were as follows:
Human NEXN shRNA-1:5′-GCAAGCTGAAGAGGAAGCAAGTTCAAGAGACTTGCTTCCTCTTCAGCTTGCTTTTTT-3′;
Human NEXN shRNA-2:5′-GAGATGATTCACTACTTATAATTCAAGAGATTATAAGTAGTGAATCATCTCTTTTTT-3′;
Human NEXN shRNA-3:5′-GAGCAATTGACCTTGAAATTATTCAAGAGATAATTTVAAGGTCAATTGCTCTTTTTT-3′;
Human MYOCD shRNA-1:5′- GCTGTGAAAGAGGCCAATTCAAGAGATTATGGCCTCTTTCAGCTTTTTT-3′
Human MYOCD shRNA-2:5′- CAAAGGTGAAGAAGCTTAAATTTCAAGAGAATTTAAGCTTCTTCACCTTTGTTTTTT-3′
Human MYOCD shRNA-3:5′- ATCTGAAGGTCTCTGAATTAATTCAAGAGATTAATTCAGAGACCTTCAGATTTTTTT-3′
Human NEXN cDNA was cloned into the pLenti-EF1a-CMV-P2A-Puro expression vector. For lentiviral infection, the packaging plasmids pMD2.G and psPAX2 were co-transfected with the target plasmid into HEK-293T cells. After incubating for 48 h, the viral supernatant was collected, filtered, and concentrated. The virus was then used to infect Huh7 and BEL-7402 cells. Stable NEXN-overexpressing cell lines were selected using puromycin (#P816466, MACKLIN, Shanghai, China), while blasticidin (#ST018, Beyotime, Shanghai, China) was used to select NEXN and MYOCD stable knockdown cell lines.

Cell proliferation assay
Cell proliferation was monitored for 5 consecutive days using a live cell counting assay. Huh7-NEXN and BEL-7402-NEXN cells in the log phase were seeded into 24-well plates (∼1×104 cells/well, n = 3). Cell numbers were counted at 24, 48, 72, 96, and 120 h. Additionally, a growth curve was plotted using the average daily values, and the population doubling time (PDT) was calculated using the formula: PDT = t × lg2/(lgNt - lgN0), where N0 represents the initial cell count, Nt denotes the final cell count, and t is the time between N0 and Nt.

Colony formation assay
Liver cancer cells in the log phase were seeded into 6-well plates at 1,600 cells/well. The culture was terminated when 50 or more visible colonies had formed. Subsequently, cells were fixed with 4% methanol for 15 min, stained with 0.1% crystal violet for 30 min, washed with deionized water, air-dried, and photographed to count the number of colonies.

Scratch assay
Paired Huh7-NEXN and BEL-7402-NEXN cells (∼100×104 cells/well) were seeded into 6-well plates, respectively. When the cells reached 90% confluence, a sterile pipette tip was used to vertically scratch the cell layer, followed by the addition of 2% FBS. Photos were acquired at 0, 24, and 42 h after scratching. The wound areas at 0, 24, and 42 h were quantified using ImageJ software, and the wound healing rate was calculated as (Wound area at 0 h – Wound area at 24 h or 42 h)/Wound area at 0 h × 100 %.

Transwell migration and invasion assays
Approximately 8×104 liver cancer cells were resuspended in a medium containing 1% FBS and seeded into the upper chamber of a Transwell insert with an 8 μm pore size that was made of polycarbonate membrane (#3422, Corning, NY, USA). Meanwhile, 800 μL of medium with 20% FBS was added to the lower chamber. After 24 h of co-culturing, the cells were fixed with methanol for 15 min, stained with 0.1% crystal violet for 30 min, and randomly imaged at 100× magnification in eight fields of view. Cell counts were subsequently quantified using ImageJ. Alternatively, for the invasion assay, the upper chamber was pre-coated with 50 μL of Matrigel diluted 1:3 (#356234, BD, NJ, USA) and incubated at 37°C for 2 h before seeding the liver cancer cells. The subsequent steps followed the same procedure as the migration assay.

Co-immunoprecipitation and western blotting assays
Co-immunoprecipitation (Co-IP) and western blotting (WB) assays were conducted following previously described procedures.31 Briefly, pre-treated cells were lysed for 30 min at 4°C using NP40 buffer or IP lysis buffer (1 mM EDTA, 1% Triton X-100, 10% glycerol, 50 mM Tris pH 7.4, 150 mM NaCl). The protein lysate was then centrifuged at 12,000 rpm for 10 min at 4°C. The supernatant obtained from centrifugation was subsequently incubated with 5 μL of anti-FLAG M2 magnetic beads (#M8823, Sigma, Kawasaki, Japan) overnight at 4°C. The following day, the magnetic bead-protein complexes were washed seven times with IP lysis buffer and boiled in 1× sodium dodecyl sulfate loading buffer. Proteins were then separated via sodium dodecyl sulfate polyacrylamide gel electrophoresis, blocked, and incubated overnight at 4°C with the appropriate primary antibodies. Next, membranes were washed three times with 1× TBST, followed by incubation with secondary antibodies at room temperature for 1 h. Protein detection was visualized using ECL substrate (#BMU102-CN, Abbkine, Wuhan, China) and imaged using the Bio-Rad Molecular Imager ChemiDoc XRS+ System (CA, USA) to determine the grayscale of protein expression. A full list of antibodies used in this study is listed in the supplemental information.

Immunofluorescence assay
Cells were seeded onto glass-bottom confocal dishes (#801002, NEST, Wuxi, China). After complete adherence, the cells were fixed with 4% paraformaldehyde at room temperature for 15 min and permeabilized with 0.5% Triton X-100 for 15 min. Blocking was performed using 5% BSA at room temperature for 1 h. The cells were then incubated overnight at 4°C with several primary antibodies, namely FLAG (1:800, #8146, CST), MYOCD (1:500, #SAB4200539, Sigma), and β-catenin (#1:800, 8480, CST). The cells were then respectively incubated with TRITC-conjugated rabbit secondary antibodies and FITC-conjugated mouse secondary antibodies (Jackson ImmunoResearch, PA, USA) for 1 h at room temperature. After washing thrice with 0.5% phosphate-buffered saline with tween 20, the cells were stained with 4′,6-diamidino-2-phenylindole (#C1002, Beyotime, Shanghai, China) and mounted with an anti-fading agent. Images were captured using a Lecia TCS SP8 confocal microscope (Lecia TCS SP8, Hesse, Germany).

Nuclear-cytoplasmic fractionation
We utilized the Nuclear-Cytosol Extraction Kit (#P1200, ApplyGen Technology Co., Ltd., Beijing, China. Precisely 1×106 cells were resuspended in 100 μL of Cytosol Extraction Buffer A, shaken vigorously, and incubated on ice for 15 min, during which the solution was shaken for 15 s every 5 min. After centrifugation at 12,000 g for 5 min at 4°C, the cytoplasmic proteins were collected from the supernatant. The nuclear pellet was resuspended twice with 100 μL and 5 μL of Cytosol Extraction Buffer A and B, respectively, shaken for 10 s, incubated on ice for 1 min, and centrifuged. After repeating this process twice, the supernatant was discarded. Next, 100 μL of Nuclear Extraction Buffer was added, and the solution was shaken vigorously for 15 s, then incubated on ice for 30 min, with 15-s shaking intervals every 10 min. Finally, the sample was centrifuged to isolate the nuclear proteins.

Quantification and statistical analysis
Statistical analyses were performed using GraphPad Prism 9.0 and SPSS 26.0 software, with paired/unpaired t-tests or one-way ANOVA for comparisons. All data are presented and plotted as Mean ± SEM ± SD as indicated in figure and results. p < 0.05 was considered statistically significant. Differences were considered statistically significant when p < 0.05 (two-tailed, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001). Each experiment was conducted thrice.