Exosome-derived hsa_circ_0007132 promotes lenvatinib resistance by inhibiting the ubiquitin-mediated degradation of NONO.

1
Introduction
Hepatocellular carcinoma (HCC) ranks as the sixth most prevalent malignant tumor globally and represents the third most frequent cause of cancer-associated deaths worldwide. [1]. Due to its insidious onset, most HCC cases are identified during advanced stages. For these patients, lenvatinib is one of the preferred first-line therapeutic agents [2]; however, the emergence of resistance significantly undermines its efficacy [3]. Therefore, there is an urgent need to delve deeper into the mechanisms of lenvatinib resistance (LR) to identify potential therapeutic targets that could facilitate the reversal of resistance.
Circular RNAs (circRNAs) represent a category of non-coding RNAs formed through back splicing of precursor mRNAs, resulting in closed circular RNA molecules. Growing evidence indicates that circRNAs are directly involved in the onset and progression of numerous tumors, playing significant regulatory roles [4]. A key mechanism through which circRNAs exert their biological functions is via interaction with RNA-binding proteins (RBPs) [5]. For instance, circASAP1 enhances HNRNPC protein expression by binding to it and obstructing its ubiquitination, leading to increased association of HNRNPC with GPX4 mRNA, thereby stabilizing its mRNA and inhibiting tumor progression of clear cell renal cell carcinoma [6]. In HCC, circPTPN12, generated through the mediation of ESRP1, functions as a molecular scaffold that promotes the interaction between OTUD6B and PDLIM2, thereby facilitating deubiquitination. The elevated levels of PDLIM2 subsequently bind to P65, augmenting its ubiquitination and degradation. This cascade ultimately leads to the inhibition of the NF-κB signaling pathway, effectively inhibiting HCC proliferation [7]. According to research conducted by Dong et al., circUPF2 was identified as being markedly upregulated in exosomes secreted by sorafenib-resistant HCC cell lines, where it increases resistance to sorafenib by upregulating SLC7A11 expression and suppressing ferroptosis. Further mechanistic investigations revealed that circUPF2 acts as a molecular scaffold that promotes the binding of IGF2BP2 to SLC7A11 mRNA, stabilizing the mRNA and ultimately leading to its upregulation [8]. Despite a plethora of studies confirming the crucial roles of circRNAs in cancer development, the relationship between exosomal circRNAs and LR remains unexplored and warrants further investigation.
NONO is a key component of the Drosophila Behavior/Human Splicing (DBHS) protein family and serves as a versatile protein involved in the regulation of gene expression, cellular proliferation, and invasion across various cancers [9]. In HCC, the abnormal expression of NONO also plays a crucial regulatory role in its progression. For instance, in the hypoxic microenvironment of HCC, NONO binds to and stabilizes the mRNA of HIF-1 and HIF-2. The upregulated HIF-1 and HIF-2 subsequently activate the transcription of hypoxia-inducible genes, thereby promoting HCC progression [10]. Another study has demonstrated that NONO can function as an RBP interacting with DIO3OS, a highly conserved lncRNA. This interaction markedly diminishes NONO's ability to mediate the nuclear export of ZEB1 mRNA, resulting in the suppression of ZEB1 expression and consequently limiting tumor progression. [11]. However, there is currently a lack of research reporting on the involvement of NONO in the process of HCC LR.
In this research, we identified a novel exosomal circRNA, hsa_circ_0007132, that is highly associated with HCC LR. Further in vitro and in vivo experiments confirmed that hsa_circ_0007132 is highly expressed in both LR tissues and cells, playing a pivotal role in regulating LR by promoting EMT. Mechanistically, hsa_circ_0007132 interacts with NONO to obstruct its ubiquitination and degradation. The resulting upregulation of NONO expression contributes to the malignant progression of HCC. In conclusion, our research elucidates the critical role of the hsa_circ_0007132/NONO axis in HCC LR, offering a potential therapeutic target for reversing resistance to lenvatinib.

2
Results
2.1
Screening and identification of hsa_circ_0007132
In order to identify exosomal circRNAs that play a significant role in the process of LR, we collected peripheral blood samples from three HCC patients who exhibited progressive disease (PD) and three who achieved complete response (CR) following lenvatinib treatment, based on their respective responses to the therapy. We defined patients who exhibited progressive disease as those with LR. Subsequently, we performed exosomal circRNA sequencing (Fig. 1a) and then intersected the differentially expressed circRNAs with the GEO exosomal circRNA sequencing dataset. This comprehensive analysis ultimately led to the identification of hsa_circ_0007132, which was upregulated in the serum exosomes of both HCC patients and those with LR (Fig. 1b). This finding was corroborated by expression validation in our sequencing cohort (Fig. 1c). Furthermore, an examination of the exoRBase database (http://www.exorbase.org/exoRBaseV2/toIndex) revealed that hsa_circ_0007132 is also notably overexpressed in serum exosomes from patients with HCC and several other types of cancer, while displaying low expression in the serum exosomes of colorectal cancer patients (Fig. 1d and e). Next, we proceeded with the circularization identification of hsa_circ_0007132. Firstly, we designed specific primers for qRT-PCR and confirmed the presence of continuous junction site in the resulting products through Sanger sequencing (Fig. 1f). Additionally, according to relevant information from the circBANK database, hsa_circ_0007132 is derived from exon 8 of the ATXN1 gene (Fig. 1g). Subsequently, RNase R digestion experiments were conducted and indicated that hsa_circ_0007132 exhibited remarkable resistance to RNase R digestion compared to its parental gene, ATXN1 (Fig. 1h). Furthermore, experiments involving actinomycin D corroborated that hsa_circ_0007132 demonstrates greater stability relative to linear ATXN1 mRNA (Fig. 1i). Finally, amplification of hsa_circ_0007132 using divergent and convergent primers revealed that the back splicing junction of hsa_circ_0007132 was exclusively present in cDNA samples, but not in gDNA samples (Fig. 1j).

2.2
Hsa_circ_0007132 is markedly upregulated in HCC tissues and exosomes
In light of the properties of hsa_circ_0007132 derived from exosomes, we initially extracted the culture supernatant from Huh7 and PLC/PRF/5 cells for exosome characterization. Transmission electron microscopy (TEM) and nanoparticle tracking analysis (NTA) unveiled that these exosomes predominantly ranged from 80 to 200 nm in diameter and were enriched in exosomal markers, including Alix, CD63, and TSG101 (Fig. 2a, b, c). Subsequently, qRT-PCR confirmed that hsa_circ_0007132 was highly expressed in HCC cell lines (Fig. 2d). Notably, the expression level in our established LR cell lines was elevated compared to that in the parental cells (Fig. 2e). Furthermore, hsa_circ_0007132 was also expressed at higher levels in the exosomes derived from LR cell lines than in those from the parental cell lines (Fig. 2f), substantiating the remarkable association of hsa_circ_0007132 with LR. Next, we evaluated hsa_circ_0007132 expression in HCC and matched paracancerous tissues from 55 patients. The results indicated that hsa_circ_0007132 was dramatically overexpressed in HCC tissues (Fig. 2g). We subsequently categorized the 55 HCC patients based on hsa_circ_0007132 expression levels, and survival analysis revealed that patients with high hsa_circ_0007132 expression exhibited poorer recurrence-free survival (RFS) compared to those in the low expression group (Fig. 2h). While overall survival (OS) in the high-expression group also showed a trend toward worse outcomes compared to the low-expression group, this difference did not reach statistical significance (Fig. 2i). Moreover, subgroup analyses indicated that hsa_circ_0007132 expression was markedly elevated in HCC patients with macrovascular invasion (MVI) (+) relative to those with MVI (−) (Fig. S1a). Although elevated expression levels were also observed in HCC patients with tumor diameters greater than 5 cm, advanced BCLC stage, and higher TNM stage, these differences did not achieve statistical significance (Figs. S1b, c, d). This suggests that hsa_circ_0007132 may contribute to the invasiveness and metastatic potential of HCC. Additionally, results from RNA fractionation and fluorescence in situ hybridization (FISH) analyses demonstrated that hsa_circ_0007132 is distributed evenly between the nucleus and the cytoplasm (Fig. 2j and k).

2.3
Hsa_circ_0007132 promotes EMT and LR in HCC
Previous findings suggested a close correlation between hsa_circ_0007132 and the process of LR. To explore this further, we ectopically expressed and knocked down hsa_circ_0007132 in Huh7-LR and PLC5-LR cells, with verification accomplished through qRT-PCR (Fig. S1e and f). Firstly, we carried out IC50 assessments for lenvatinib, revealing that the overexpression of hsa_circ_0007132 did not lead to a notable enhancement in the lenvatinib tolerance of LR HCC cells. Conversely, knocking down hsa_circ_0007132 dramatically reduced the IC50 value of LR HCC cells, indicating that the elevated expression level of hsa_circ_0007132 is essential for maintaining LR in HCC cells. Subsequently, we conducted CCK8 and colony formation assays which confirmed that neither the overexpression nor the knockdown of hsa_circ_0007132 affected the proliferative capacity of HCC cells (Fig. S1g and h). Following this, we performed wound healing and Transwell assays on Huh7-LR and PLC5-LR cells with altered hsa_circ_0007132 expression. As illustrated in Fig. 3b–e, the upregulation of hsa_circ_0007132 remarkably enhanced the migration and invasion abilities of HCC cells, while its knockdown exerted an inhibitory effect; this observation was also corroborated at the Western Blot level (Fig. 3f). These experimental results collectively indicate that hsa_circ_0007132 primarily facilitates migration and invasion while sustaining LR in HCC. Considering that hsa_circ_0007132 is overexpressed in exosomes secreted by LR cells and that exosomes play a critical role in intercellular communication, we employed co-culture experiments to ascertain whether exosomal hsa_circ_0007132 is closely associated with cellular processes. We first collected culture supernatants from HCC cells in the control (OE-NC) and hsa_circ_0007132 overexpression (OE-Exo) groups to extract exosomes, which were subsequently co-cultured with Huh7-P and PLC5-P cells. qRT-PCR analysis demonstrated a significant upregulation of hsa_circ_0007132 expression in both Huh7-P and PLC5-P cells following the uptake of exosomes (Fig. 3g). The ensuing IC50 experiments substantiated that the lenvatinib tolerance of HCC cells in the OE-Exo group was markedly enhanced (Fig. 3h). Similarly, the promotional effects on migration and invasion were validated via wound healing and Transwell assays (Fig. 3i and j), indicating that hsa_circ_0007132 derived from exosomes of HCC cells can convey malignant tumor behavior and facilitate LR through exosomal pathways.

2.4
Hsa_circ_0007132 interacts with NONO
To explore the protein partners that interact with hsa_circ_0007132, we employed an RNA pulldown assay using biotinylated full-length probes targeting the splice junctions in Huh7-LR cells, followed by mass spectrometry (MS) analysis. From this investigation, the NONO protein was selected as a target for further research (Fig. 4a). Subsequently, we confirmed the interaction between hsa_circ_0007132 and NONO through RNA pulldown and RNA immunoprecipitation (RIP) experiments conducted in both Huh7 and PLC5 cells (Fig. 4b–e). Moreover, immunofluorescence studies revealed that hsa_circ_0007132 and NONO were co-localized within the nucleus and the cytoplasm, providing additional evidence for their binding relationship (Fig. 4f). To further predict the binding sites, we utilized AlphaFold3 to model the molecular structures of hsa_circ_0007132 and NONO, thereby identifying their potential interaction domains (Fig. 4g). Previous research has demonstrated that NONO forms heterodimers with SFPQ, which are integral to crucial biological activities such as precursor mRNA splicing [12]. However, the RNA pull-down assay failed to detect any interaction between hsa_circ_0007132 and SFPQ (Fig. 4h), suggesting that hsa_circ_0007132 may influence the functionality of NONO through alternative mechanisms.

2.5
Hsa_circ_0007132 inhibits the ubiquitin-mediated degradation of NONO
NONO, as a versatile RBP involved in various cellular processes, has been extensively documented to exhibit upregulation in HCC [13]. Our examination of the TCGA database corroborated the elevated expression of NONO in HCC (Fig. S2a), revealing a significant correlation between high NONO expression and poorer disease free survival (DFS) and OS in HCC patients (Fig. S2b and c). Furthermore, the results of Western blot conducted on our HCC cohort demonstrated that NONO is consistently overexpressed in HCC tissues (Fig. 5a).
Following the confirmation of the interaction between hsa_circ_0007132 and NONO, we further investigated their regulatory relationship. As depicted in Fig. 5b, overexpression of hsa_circ_0007132 in Huh7 and PLC5 cells resulted in a notable upregulation of NONO expression, whereas knockdown of hsa_circ_0007132 led to a decrease, suggesting that hsa_circ_0007132 may be involved in the degradation process of the NONO protein. Subsequently, we treated Huh7 cells with cycloheximide (CHX) to inhibit protein synthesis, and confirmed that overexpression of hsa_circ_0007132 significantly enhanced the stability and half-life of the NONO protein, resulting in its elevated expression (Fig. 5c). Conversely, knockdown of hsacirc_0007132 yielded the opposite result (Fig. S3a). Previous studies have indicated that protein degradation primarily occurs through either the autophagolysosomal pathway or the ubiquitin-proteasome pathway [14]. Therefore, we adopted chloroquine (CQ) and MG132 to block these two degradation pathways. As illustrated in Fig. 5d, the addition of MG132, but not CQ, effectively prevented the degradation of NONO protein induced by the knockdown of hsa_circ_0007132. Notably, subsequent ubiquitination assay also indicated that overexpression of hsa_circ_0007132 markedly restricted the ubiquitination level of NONO (Fig. 5e).
The aforementioned results indicate that hsa_circ_0007132 enhances the protein expression of NONO by inhibiting its ubiquitination. A recent study has demonstrated that NONO can promote HCC progression by mediating the nuclear export of ZEB1 mRNA [11]. As a transcriptional repressor, ZEB1 promotes EMT by binding to the promoter region of E-cadherin and repressing its expression [15,16]. To investigate this relationship, we first conducted a RIP assay to confirm the binding affinity between NONO and ZEB1 mRNA (Fig. 5f). Subsequently, we designed specific shRNAs targeting NONO and evaluated their efficacy, eventually selecting sh-NONO#2 for further phenotypic analyses (Fig. 5g and h). Further, qRT-PCR and Western blot analyses revealed that the knockdown of NONO significantly reduced ZEB1 protein levels without affecting ZEB1 mRNA expression, which is consistent with previous research findings (Fig. 5i and j). Furthermore, the GEPIA analysis revealed that ZEB1 is significantly overexpressed in HCC (Fig. S2d). Although elevated ZEB1 expression did not correlate with DFS, HCC patients exhibiting high ZEB1 expression demonstrated poorer OS (Fig. S2e and f).

2.6
The regulatory role of hsa_circ_0007132 in malignant progression and LR is mediated through NONO/ZEB1
To ascertain whether NONO mediates the tumor-promoting effects and LR associated with hsa_circ_0007132 in HCC, we conducted rescue experiments. The results confirmed that, while overexpression of hsa_circ_0007132 significantly enhanced the lenvatinib tolerance of Huh7-P and PLC5-P cells, this promoting effect was notably reversed upon knockdown of NONO expression (Fig. 6a and b). Similarly, silencing NONO also curtailed the migratory and invasive capabilities of Huh7-P and PLC5-P cells promoted by hsa_circ_0007132 overexpression (Fig. 6c–f). Additionally, the analysis of EMT-associated proteins further substantiated these findings (Fig. 6g). Subsequently, by simultaneously knocking down hsacirc_0007132 and overexpressing NONO in Huh7-LR and PLC/PRF/5-LR cells, we found that the overexpression of NONO effectively counteracted the decreased tolerance to lenvatinib (Fig. S4a and b), along with the impaired migratory and invasive capabilities (Fig. S4c–f), as well as the alterations in EMT-related markers triggered by the knockdown of hsacirc_0007132 (Fig. S4g).
Following this, we conducted a similar intervention to disrupt ZEB1 expression for rescue experiments. The results aligned with those obtained from the knockdown of NONO; the silencing of ZEB1 similarly reversed the pro-EMT and LR effects induced by the overexpression of hsacirc_0007132 in HCC cells (Fig. S5a–f).
In summary, these results underscore that the regulatory function of hsa_circ_0007132 in HCC cells is intricately dependent on the presence of NONO/ZEB1.

2.7
Hsa_circ_0007132 promotes the metastasis of HCC cells in vivo
To further validate the critical role of the hsa_circ_0007132/NONO axis in HCC, we established a lung metastatic model through tail vein injection of HCC cells for in vivo assessments. Firstly, we utilized Huh7-LR cells with or without the silencing of hsa_circ_0007132 to create the lung metastatic model in nude mice, treating them with either DMSO or lenvatinib. Bioluminescence imaging (BLI) results revealed that lenvatinib did not hinder lung metastasis in LR HCC cells while silencing hsa_circ_0007132 effectively suppressed the lung metastasis of Huh7-LR cells (Fig. 7a). This finding was further corroborated by the gross record of the lungs, as well as subsequent H&E staining, which demonstrated a remarkably reduced count of lung metastatic nodules in the hsa_circ_0007132 silenced group compared to the control group treated with DMSO or lenvatinib (Fig. 7b–d). The above findings indicate that elevated expression of hsa_circ_0007132 is extraordinarily critical for the maintenance of LR in HCC cells.
Next, we constructed another lung metastatic model using Huh7-P cells, with or without overexpression of hsa_circ_0007132. The subjects were then grouped based on whether NONO was silenced. BLI results indicated that the overexpression of hsa_circ_0007132 significantly enhanced the lung metastasis of Huh7-P cells, whereas silencing NONO reversed this pro-metastatic effect (Fig. 7e). Furthermore, the photographic analysis of lung specimens, along with subsequent H&E staining, visually corroborated these findings, showing that the silencing of NONO effectively reversed the promotion of HCC cell metastasis induced by hsa_circ_0007132 overexpression (Fig. 7f–h). Collectively, these experimental outcomes underscore the critical role of the hsa_circ_0007132/NONO axis in the maintenance of LR and the metastasis of HCC.

3
Discussion
As a multi-kinase inhibitor, lenvatinib has been clinically utilized since 2018, and extensive research has validated its effectiveness in improving the prognosis of HCC patients [17]. However, the emergence of LR has become one of the principal obstacles in treating patients with advanced HCC [18]. The development of LR involves highly intricate mechanisms, including epigenetic modifications, cancer stem cells (CSCs), and apoptosis [19]. In recent years, numerous studies have provided compelling evidence that circRNAs directly participate in the LR process and play pivotal roles. For example, circMED27 functions as a molecular sponge by competitively interacting with miR-655-3p, thereby blocking the suppressive effect of miR-655-3p on USP28 expression. Then the upregulation of USP28 enhances the lenvatinib tolerance of HCC cells [20]. Similarly, other circRNAs, such as circKCNN2 [21], circPIAS1 [22], and circ_0007386 [23], also regulate the LR process through competing endogenous RNA (ceRNA) mechanisms. Moreover, circPAK1 diminishes the promotion of YAP nuclear export by competitively binding to 14-3-3 ζ, resulting in enhanced YAP nuclear localization and the subsequent inhibition of the Hippo signaling pathway, ultimately leading to LR [24]. Notably, research conducted by Yang et al. demonstrated that downregulation of circCCNY in LR cells promotes the ubiquitination and degradation of E3 ubiquitin ligase SMURF1 through its interaction with HSP60. This results in the liberation of RKIP, thereby inactivating the MAPK signaling pathway and eventually strengthening HCC sensitivity to lenvatinib [25]. Our findings also indicate that high expression levels of hsa_circ_0007132 are fundamental to maintaining lenvatinib tolerance in HCC. In LR HCC cells, silencing hsa_circ_0007132 significantly curtails their EMT abilities, thereby severely attenuating the inherent LR. Mechanistically, hsa_circ_0007132 enhances the protein expression of NONO by binding to it and preventing its ubiquitin-mediated degradation, thereby enhancing NONO-mediated nuclear export of ZEB1 mRNA and increasing ZEB1 protein levels, which fortifies the EMT capacity and consequently contributes to LR. These findings present potential strategies for overcoming LR challenges.
Exosomes, as critical mediators of intercellular material exchange, facilitate the transfer of various ncRNAs and proteins [26]. Recent studies have indicated that ncRNAs within exosomes may play significant roles in the propagation of LR. For example, miR-301a-3p, which is notably upregulated in exosomes derived from LR HCC cells, can be taken up by tumor-associated macrophages (TAMs), leading to the activation of the PTEN/PI3K/GSK3B/Nrf2 signaling pathway. This activation stimulates the polarization of TAMs toward the M2 phenotype, which, in turn, promotes LR through EMT [27]. Furthermore, research conducted by Hao et al. has demonstrated that exosomal circPAK1 secreted by LR HCC cells confers a drug-resistant phenotype to recipient cells upon uptake [24]. In this study, we collected serum exosomes from three patients with PD and three patients with CR following lenvatinib treatment. We performed circRNA sequencing and, in conjunction with the circRNA sequencing dataset from the GEO database pertaining to HCC exosomes, identified hsa_circ_0007132 as being concurrently overexpressed in both HCC exosomes and LR HCC exosomes. Subsequent coculture experiments confirmed that the expression level of hsa_circ_0007132 in recipient cells was upregulated through direct uptake of exosomes enriched with hsa_circ_0007132. Additionally, phenotypic assays directly demonstrated that hsa_circ_0007132 could promote the malignant progression and LR of recipient HCC cells via the exosomal pathway, suggesting that obstructing the transfer of exosomal hsa_circ_0007132 may possess substantial significance.
Numerous research has confirmed that NONO plays a critical role in the initiation and progression of various malignancies [[28], [29], [30]]. For instance, NONO is significantly upregulated in malignant melanoma, where it contributes to the proliferation and metastasis of tumor cells [31]. In esophageal cancer, NONO regulates the proliferation and migration of tumor cells by activating the Akt and Erk1/2 signaling pathways [32]. In HCC, NONO facilitates disease progression through its interaction with ACLY mRNA, thereby enhancing fatty acid biosynthesis through increased expression [33]. These findings underscore the close association of NONO with tumor advancement. Our research also indicates that NONO is aberrantly elevated in HCC, primarily due to the interaction between hsa_circ_0007132 and NONO, which obstructs NONO's ubiquitin-mediated degradation. While previous studies have verified that FBXW7 [34] and RNF8 [35] are involved in the ubiquitin-mediated degradation of NONO and that USP11 mediates its deubiquitination [31], this study did not include such validations. Future research will be necessary to identify or explore novel E3 ligases or deubiquitinating enzymes that participate in the ubiquitination process of NONO.
In summary, our study unveils hsa_circ_0007132 as a novel oncogene that plays a crucial role in mediating HCC progression and LR by interacting with NONO and inhibiting its ubiquitin-mediated degradation. Additionally, as an exosomal circRNA, hsa_circ_0007132 can facilitate the transmission of malignant progression and LR in HCC via the exosomal pathway. Therefore, targeting the hsa_circ_0007132/NONO axis presents a highly promising strategy for reversing LR in HCC.

4
Materials and methods
4.1
Clinical samples
A total of 55 matched HCC and adjacent normal samples were collected from patients at the Third Affiliated Hospital of Sun Yat-sen University. The primary endpoints of the study included RFS and OS. RFS was calculated as the duration from the surgical date to the detection of HCC recurrence, whereas OS was determined as the period from surgery until the patient's death. These endpoints were utilized to evaluate the clinical outcomes within the study cohort.

4.2
Isolation and identification of exosomes
Before collecting exosomes, Huh7 and PLC/PRF/5 cells were inoculated in 100 mm culture dishes. These cells were cultured in a vesicle-depleted medium for 48 h to promote exosome release. After the incubation period, the supernatants were harvested and processed using differential ultracentrifugation to isolate exosomes. The purified exosomes were then resuspended in 50 μL PBS and kept at −80 °C. To characterize the exosomes, their size distribution and concentration were measured using a NanoSight NS300 system with Nanoparticle Tracking Analysis (NTA) 3.0 software.

4.3
Transfection of HCC cells
The overexpression vector (OE-circ) and its corresponding empty vector (OE-NC) were designed and synthesized by Hanbio (Shanghai, China). For gene knockdown experiments, short hairpin RNAs (shRNAs) were procured from GeneCopoeia (Maryland, USA). Viral particles were produced using 293T cells transfected with the respective vectors. Transfections were performed in six-well plates, and stable cell lines were established by selection with puromycin (1.5–2 μg/mL, Solarbio, Beijing, China) to ensure the integration and expression of the constructs. Detailed sequences of all shRNAs are provided in Supplementary Table 3.

4.4
Immunofluorescence (IF) analyses
A 24-well plate containing cell crawlers was prepared, and 5 × 104 cells were seeded into each well. After allowing the cells to adhere, fixation was performed using a 4 % paraformaldehyde solution for 15 min at room temperature. Permeabilization was then carried out with PBS containing 0.4 % Triton X-100, followed by a 1-h blocking step to minimize nonspecific binding. After washing with PBS, the cells were incubated with a primary antibody overnight at 4 °C. On the following day, secondary antibodies (1:400 dilution, Proteintech, China) were applied for 1 h. After thorough PBS washes, nuclei were stained using a DAPI reagent (RiboBio, Guangzhou, China). Finally, coverslips were mounted onto slides, and images were captured using an Ax Imager 2 microscope (Zeiss, Germany).

4.5
Western blot assay and antibodies
Tissues or cells were homogenized in RIPA buffer, and protein levels were measured using the BCA kit (H·Wayen, Shanghai, China). The extracted proteins were resolved by SDS-PAGE and then transferred onto a PVDF membrane for subsequent analysis. Specific primary antibodies, such as anti-NONO, were used to detect target proteins, followed by incubation with appropriate secondary antibodies for 1 h. Enhanced chemiluminescence (ECL) reagents (Advansta, CA, USA) were applied for visualization, and images were acquired using the Tanon™ 5200CE Chemi-Image System (Shanghai, China). A comprehensive list of antibodies used in this study is provided in Supplementary Table 5.

4.6
RNA pull-down
A total of 2 × 107 Huh7-LR cells were collected and lysed using an RNA pull-down kit (Bersinbio, Guangzhou, China). The hsa_circ_0007132 probe was introduced to isolate RNA-protein complexes, and the expression of target proteins was subsequently assessed via Western blot.

4.7
RNA immunoprecipitation
RNA immunoprecipitation (RIP) was conducted using antibodies targeting NONO, and IgG to investigate their binding interactions with hsa_circ_0007132. A total of 2 × 107 HCC cells were harvested, and the RIP procedure was conducted following the instructions provided in the kit (Bersinbio, Guangzhou, China). The enrichment of hsa_circ_0007132 in the immunoprecipitated samples was subsequently quantified using qRT-PCR.

4.8
Co-IP
The procedure was performed following the guidelines provided by the manufacturer for the co-immunoprecipitation (Co-IP) kit (Bersinbio, Guangzhou, China). First, 2 × 107 HCC cells were collected and lysed using IP lysis buffer. The lysate was then incubated with primary antibodies and IgG antibodies to enable specific binding. Protein A/G beads were subsequently added to the mixture to isolate the antigen-antibody complexes. After elution, the captured proteins were subjected to Western blotting to evaluate their interactions.

4.9
Ubiquitination assay
The involvement of hsa_circ_0007132 in NONO ubiquitination and its influence on protein stability was investigated using co-immunoprecipitation (Co-IP) combined with Western blotting. Huh7-LR cells with or without overexpressing hsa_circ_0007132 were cultured and exposed to 10 μM of the proteasome inhibitor MG132 (Selleck, Houston, USA) for 6 h. Cell lysates were harvested and incubated overnight at 4 °C with anti-NONO antibody. Subsequently, 20 μL of protein A/G beads (Bersinbio, Guangzhou, China) were added to the lysates, and incubated for 2 h at 4 °C to allow antibody-bead complex formation. The interactions and stability of the target proteins were analyzed through Western blotting.

4.10
Xenograft in nude mice
BALB/c nude mice, approximately 4 weeks old, were obtained from the Guangdong Provincial Medical Laboratory Animal Center (Guangzhou, China) to create lung metastasis models. Huh7-P or Huh7-LR cells (3 × 106 in 200 μL) from each experimental group were injected into the tail veins of the mice. Lenvatinib treatment was administered based on the assigned experimental groups. One month later, the mice were euthanized, and their lungs were harvested, fixed, and stained with H&E to quantify metastatic nodules.

4.11
Statistical analysis
Data from experiments were analyzed using GraphPad Prism version 8.0. Data are presented as mean ± standard deviation, with all experiments performed in triplicate to ensure reproducibility. Statistical analyses were performed using Student's t-test or one-way ANOVA, depending on the data distribution. Kaplan–Meier analysis was used to generate survival curves, and comparisons were made using the log-rank test.

CRediT authorship contribution statement
Mingbo Cao: Writing – original draft, Validation, Investigation, Data curation, Conceptualization. Yuxuan Li: Writing – original draft, Visualization, Software, Methodology. Xiaorui Su: Validation, Methodology. Yongchang Tang: Formal analysis. Feng Yuan: Formal analysis. Yupeng Ren: Investigation. Meihai Deng: Writing – review & editing, Project administration. Zhicheng Yao: Supervision, Resources, Funding acquisition, Conceptualization.

Institutional review board statement
The research adhered to the principles outlined in the Declaration of Helsinki and received approval from the Ethics Committee of the Third Affiliated Hospital of Sun Yat-sen University (Approval number: A2023-245-01).

Data availability
The data supporting this study are available upon request from the corresponding author.

Funding statement
This work was supported by the 10.13039/501100012245Science and Technology Planning Project of Guangzhou city (202102010171, 2023A03J0211), the National Natural Science Foundation Cultivation Project of the Third Affiliated Hospital of 10.13039/501100002402Sun Yat-sen University (2020GZRPYMS11), the 10.13039/501100021171Guangdong Basic and Applied Basic Research Foundation (2023A1515010522, 2023A1515010135, 2023A1515220090), 10.13039/100018904Beijing Xisike Clinical Oncology Research Foundation (Y-Roche2019/2–0041, Y-HH202102-0044, Y-MSDPU2022-0344).

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.