Hepatitis B virus promotes hepatocellular carcinoma progression by disrupting NSUN2/YBX1-mediated m⁵C modification of PTEN mRNA

Abstract

Hepatitis B virus (HBV) has been implicated in hepatocellular carcinoma (HCC) progression, partly through regulation of the tumor suppressor phosphatase and tensin homolog (PTEN). Although transcriptional regulation of PTEN by HBV is well characterized, its post-transcriptional regulation remains poorly understood. Because RNA 5-methylcytosine (m⁵C) modification influences post-transcriptional gene control and cancer development, we investigated whether HBV modulates PTEN through m⁵C. Methylated RNA immunoprecipitation (MeRIP)–quantitative polymerase chain reaction showed a marked reduction in m⁵C on PTEN mRNA in HBV-producing cells. MeRIP sequencing further identified decreased m⁵C within the PTEN coding sequence region (chr10:89717747–89717771) in HBV-producing HepAD38/tetracycline-off cells, with chr10:89717756 emerging as a critical site where HBV suppresses m⁵C enrichment and PTEN expression. Mechanistically, the m⁵C “writer” NOP2/Sun RNA methyltransferase 2 (NSUN2) and the “reader” Y-box binding protein 1 (YBX1) stabilized PTEN mRNA in an m⁵C-dependent manner. HBV disrupted this pathway, decreasing PTEN mRNA stability via NSUN2- and YBX1-mediated m⁵C. Overexpression of NSUN2 or YBX1 attenuated HBV-driven proliferation, migration, and invasion, and these effects were partially reversed by the PTEN inhibitor VO-Ohpic. The small hepatitis B surface antigen and hepatitis B X protein downregulated NSUN2 and YBX1, linking viral proteins to PTEN suppression. Further, HBV is associated with reduced NSUN2 expression in HBV transgenic (HBV-Tg) mice, HBV-infected primary human hepatocytes as well as HBV-positive clinical HCC specimens, supporting the physiological and clinical relevance of this finding. Together, these findings identify the NSUN2/YBX1/PTEN axis as a potential therapeutic target in HBV-associated HCC.

Highlights

• •HBV reduces PTEN mRNA stability by disrupting site-specific m⁵C modification.
• •NSUN2 and YBX1 are critical for maintaining PTEN mRNA stability via m⁵C modification.
• •HBV viral proteins SHBs and HBx suppress NSUN2 and YBX1 expression.
• •Overexpression of NSUN2 and YBX1 inhibit HBV-driven cell proliferation and invasion.
• •The NSUN2/YBX1/PTEN axis is a potential therapeutic target in HBV-related HCC.

Introduction

Hepatitis B virus (HBV) infection is a significant global health challenge, despite available effective vaccines and antiviral therapies (Nguyen et al., 2020; Wang et al., 2022; Delphin et al., 2024). Chronic HBV infection is a leading cause of hepatocellular carcinoma (HCC), a cancer with high morbidity and mortality (Fung et al., 2022; Hong et al., 2023). The transition from HBV infection to HCC involves complex mechanisms, including HBV genome integration, oncogene activation, and tumor suppressor gene suppression (Revill et al., 2020; Jeng et al., 2023).

Among these suppressors, phosphatase and tensin homolog (PTEN) plays a pivotal role in maintaining cellular homeostasis by antagonizing the oncogenic phosphoinositol-3-kinase (PI3K)/AKT/mTOR pathway (Wang et al., 2012; Worby and Dixon, 2014; Álvarez-Garcia et al., 2019). Mutations or deletions in the PTEN gene are frequently observed across various human cancers including HCC (Pérez-Ramírez et al., 2015). Studies consistently demonstrate that HBV-induced downregulation of PTEN contributes to tumorigenesis by impairing its tumor-suppressive functions (Sun et al., 2020; Liu Yang et al., 2022; Huang F. et al., 2024). Specifically, hepatitis B X protein (HBx), the major factor contributing to HBV's carcinogenic potential, exerts its pro-tumourigenic potential by disrupting p53-mediated PTEN transcription (Chun et al., 2003), highlighting transcriptional regulation of PTEN by HBV. On the post-transcriptional level, miR-520c-3p, upregulated by HBx of HBV, targets the 3′-UTR of PTEN mRNA, promoting hepatoma cell migration and metastasis through the activation of AKT/NF-κB pathway (Liu Yang et al., 2022). Additionally, HBV infection has also been linked to N⁶-methyladenosine (m⁶A) modification of PTEN mRNA. HBV infection promotes m⁶A methylation of PTEN mRNA at the 3′-UTR by inducing the expression of methyltransferase-like 3 (METTL3). This modification leads to PTEN mRNA destabilization, activating the PI3K/AKT signaling pathway and driving HCC progression (Kim et al., 2021). Despite the established link between HBV infection and m⁶A modification of PTEN mRNA, the mechanisms involving other types of RNA modifications remain poorly understood.

RNA 5-methylcytosine (m⁵C) modification is an integral component of RNA epigenetics that regulates RNA stability, splicing, translation efficiency or subcellular localization. This reversible and dynamic process involves three key regulators: m⁵C methyltransferases (“writers”), demethylases (“erasers”), and m⁵C-binding proteins (“readers”), which work together to modulate RNA metabolism (Chen et al., 2024). NOP2/Sun RNA methyltransferase family members 2 (NSUN2) (Courtney et al., 2019) and 6 (NSUN6) (Selmi et al., 2021) are identified as the most common mRNA methyltransferases (“Writers”) for catalyzing m⁵C methylation. The majority of the biological roles of m⁵C modifications are associated with “Reader”, including Aly/REF export factor (ALYREF) (Yang et al., 2017) and Y-box binding protein 1 (YBX1) (Chen et al., 2019). ALYREF recognizes m⁵C-modified mRNA to facilitate mRNA nuclear export, while YBX1 binds directly to m⁵C methylated mRNA to stabilize mRNA (Chen et al., 2021; Guo et al., 2021). Recent studies have revealed that aberrant mRNA m⁵C modification contributes to the pathogenesis of various cancers, including cervical cancer (Chen et al., 2023), ovarian cancer (Liu et al., 2024), esophageal squamous cell carcinoma (Su et al., 2021) and non-small-cell lung cancer (Wang et al., 2023). However, the role and underlying mechanisms of RNA m⁵C modification in HBV-related HCC remain largely unexplored. Whether and how host mRNA m⁵C modifications contribute to HCC progression is still unclear.

Research on m⁵C modification and its machinery in cancers, including HBV-associated HCC, remains limited. The post-transcriptional regulation of PTEN during HBV infection is largely unexplored. This study demonstrates that HBV suppresses PTEN expression by downregulating m⁵C modification of PTEN mRNA. Through m⁵C-MeRIP-qPCR and RIP-qPCR, NSUN2 and YBX1 were identified as key regulators of PTEN mRNA m⁵C modification. Furthermore, HBV infection was shown to decrease the m⁵C modification levels of PTEN mRNA by suppressing NSUN2 and YBX1 expression, leading to PTEN mRNA destabilization and reduced expression, thereby contributing to HBV-associated HCC progression. These findings reveal the NSUN2/YBX1/m⁵C-PTEN axis as a novel regulatory mechanism in HBV-driven HCC and suggest it as a potential therapeutic target.

Results

HBV decreases m⁵C modification of PTEN mRNA at chr10:89717756

To investigate the impact of HBV on the m⁵C methylation of PTEN mRNA, m⁵C-MeRIP-qPCR analysis was performed on HBV-producing cells. The results revealed that the m⁵C levels of PTEN were significantly reduced in HBV-infected HepG2-NTCP cells, HepG2 cells transfected with pREP-1.2 × HBV, and HepAD38/Tet⁻ cells compared to their respective controls (Fig. 1A). To further explore the distribution of m⁵C sites on PTEN mRNA, m⁵C-MeRIP-seq was conducted in HBV-producing (HepAD38/Tet⁻) and non-producing (HepAD38/Tet⁺) cells. Analysis through exomePeak2 and visualization with the Integrative Genomics Viewer (IGV) identified a specific downregulated m⁵C-methylated region in Exon 7 of PTEN mRNA (gene coordinates chr10:89717747–89717771, log2FC = −0.972, p = 0.023) in HepAD38/Tet⁻ cells (Fig. 1B), as confirmed by differential methylation analysis. To pinpoint the exact m⁵C site affected by HBV, three bioinformatics tools (iRNA-m5C, m5Cpred-XS, and RNAm5Cfinder) were employed, predicting six potential m⁵C sites within the PTEN coding sequence (CDS): C89717749, C89717750, C89717752, C89717756, C89717761, and C89717768 (Fig. 1C and Supplementary Table S1). PTEN CDS fragments containing either the wild-type or mutated (cytosine-to-adenine, C-to-A) m⁵C sites were cloned into pmirGLO luciferase reporter vectors (Fig. 1D). Luciferase reporter assays demonstrated a significant reduction in luciferase activity in HepAD38/Tet⁻ cells transfected with wild-type PTEN CDS (PTEN-WT). A reduction was also observed in cells transfected with the all-mutant construct (PTEN-all MUT), but the magnitude of this effect was markedly attenuated compared with that in the PTEN-WT group (Fig. 1E). Notably, this reduction in luciferase activity was reversed upon mutation of the C89717756 site (PTEN-C89717756A) (Fig. 1F), indicating chr10:89717756 as a crucial m⁵C site for HBV-mediated regulation. To validate the role of chr10:89717756, Flag-tagged PTEN expression plasmids carrying either the wild-type (PTEN-WT-Flag) or mutant (PTEN-C89717756A-Flag) version were constructed using the pcDNA3.1/myc-hisA(−) vector (Fig. 1G) and transfected into HepAD38 cells. m⁵C-MeRIP-qPCR analysis demonstrated a substantial reduction in m⁵C enrichment of PTEN-Flag mRNA in HBV-producing HepAD38/Tet⁻ cells compared to the control HepAD38/Tet⁺ cells, however, the C89717756A mutation effectively abolished the decrease in m⁵C modification (Fig. 1H), indicating that HBV specifically targets the m⁵C site at chr10:89717756. To further investigate the functional consequences of m⁵C modification at chr10:89717756, RT-qPCR analysis was performed. The results revealed that PTEN-Flag mRNA levels were markedly reduced in HepAD38/Tet⁻ cells compared to Tet⁺ controls. However, overexpression of the PTEN-C89717756A-Flag mutant partially rescued PTEN mRNA levels, indicating that HBV-mediated repression of PTEN expression is dependent on the intact m⁵C site at chr10:89717756 (Fig. 1I). To directly assess the impact of m⁵C modification on PTEN mRNA stability, an mRNA stability assay was conducted. Following Actinomycin D (ActD) treatment, the half-life of PTEN-WT-Flag transcripts in HepAD38/Tet-cells was significantly reduced (4.79 ± 0.16 h) compared to HepAD38/Tet⁺ cells (7.67 ± 0.57 h), indicative of HBV-induced destabilization. Importantly, the expression of PTEN-C89717756A-Flag partially restored the mRNA half-life (Fig. 1J), further implicating the m⁵C modification at chr10:89717756 in stabilizing PTEN mRNA. Consistent with these observations, Western blot analysis revealed that PTEN-Flag protein levels were significantly diminished in HepAD38/Tet⁻ cells expressing PTEN-WT-Flag, however, this HBV-mediated suppression was notably alleviated when the PTEN-C89717756A-Flag mutant was expressed, further supporting the dependency on the m⁵C site for HBV-induced downregulation (Fig. 1K). Collectively, these findings demonstrate that HBV downregulates PTEN expression by reducing its mRNA stability through decreasing site-specific m⁵C modification at chr10:89717756, highlighting a novel post-transcriptional mechanism of HBV-mediated suppression of tumor suppressor gene expression.

Fig. 1.

HBV reduces the m⁵C level of PTEN mRNA at chr10:89717756. A m⁵C-MeRIP-qPCR analysis of PTEN mRNA m⁵C levels in HepG2-NTCP cells infected with HBV for 10 days, HepG2 cells transfected with pREP-1.2 × HBV, and HepAD38/Tet⁻ cells. B The m⁵C abundances on PTEN mRNA transcripts in HepAD38/Tet⁻ and control HepAD38/Tet ⁺ cells as detected by m⁵C-MeRIP-seq, showing significant differences at regions chr10:89717747–89717771. Blue reads originate from input libraries and red reads originate from m⁵C-IP libraries. C Diagram illustrating six potential m⁵C modification sites within the exon 7 of PTEN coding sequence (CDS) and their corresponding mutations. DPTEN CDS containing either wild type or mutant (C-to-A mutation) m⁵C sites was cloned into the pmirGLO vector downstream of the Firefly luciferase gene. E, F Luciferase reporter assays showing changes in relative luciferase activity for PTEN-WT and PTEN-all MUT (E) or individual mutant (F) in HepAD38/Tet⁺ and HepAD38/Tet⁻ cells. The results were normalized to a value of 1 for control HepAD38/Tet⁺ cells transfected with pmirGLO-PTEN-WT plasmid. G Schematic diagram of wild type PTEN CDS and mutant construct cloned in the pcDNA3.1/myc-his(−)A vector. The paired primers F′ and R′ used for amplifying the plasmid-derived PTEN-Flag mRNA in m⁵C-MeRIP-qPCR, RIP-qPCR, RT-qPCR and RNA stability assays, were listed in Supplementary Table S5 or S6. H m⁵C-MeRIP-qPCR analysis of m⁵C levels of PTEN-Flag mRNA in HepAD38 cells transfected with PTEN-WT-Flag or PTEN-C89717756A-Flag plasmid. I RT-qPCR analysis of PTEN-Flag mRNA levels in HepAD38/Tet⁺ and HepAD38/Tet⁻ cells transfected with PTEN-WT-Flag or PTEN-C89717756A-Flag plasmid. J RNA stability assay displayed the PTEN-Flag mRNA half-life in HepAD38/Tet⁺ and HepAD38/Tet⁻ cells overexpressing PTEN-WT-Flag or PTEN-C89717756A-Flag plasmid by RT-qPCR at 0, 2, 4, and 6 h after treatment with 5 μg/mL actinomycin D (ActD). K Western blot analysis of PTEN protein levels in HepAD38/Tet⁺ and HepAD38/Tet⁻ cells overexpressing PTEN-WT-Flag or PTEN-C89717756A-Flag plasmid. Data are presented as mean ± SD from three independent experiments. ns: not significant; ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

m⁵C modification of PTEN mRNA is regulated by NSUN2 and YBX1

To investigate the regulation of HBV on PTEN m⁵C modification, it is crucial to identify the m⁵C regulatory enzyme of PTEN mRNA. Among these, NSUN2 (Courtney et al., 2019) and NSUN6 (Selmi et al., 2021) are well-established m⁵C “writers” catalyzing the deposition of m⁵C on mRNAs, while ALYREF (Yang et al., 2017) and YBX1 (Chen et al., 2019) act as m⁵C “readers” recognizing and binding to m⁵C-modified transcripts. To assess their impact on PTEN expression, we performed siRNA-mediated knockdown of NSUN2, NSUN6, ALYREF, and YBX1 in HepG2 and Huh7 cells. Western blot analysis revealed that depletion of NSUN2 or YBX1 markedly reduced PTEN protein levels, whereas silencing NSUN6 or ALYREF had no observable effect in either cell line (Fig. 2A). Similarly, RT-qPCR demonstrated a significant decrease in PTEN mRNA levels upon NSUN2 or YBX1 knockdown (Fig. 2B), indicating that NSUN2 and YBX1 are critical for maintaining PTEN expression. To determine whether NSUN2 and YBX1 regulate PTEN expression by stabilizing its mRNA, we conducted RNA stability assays. Silencing of NSUN2 or YBX1 in HepG2 cells resulted in a marked reduction in the half-life of PTEN transcripts (from 6.45 h to 3.81–4.93 h for NSUN2 knockdown and from 5.98 h to 3.00–3.48 h for YBX1 knockdown). Consistent findings were observed in Huh7 cells (Fig. 2C), suggesting that NSUN2 and YBX1 enhance PTEN expression by promoting its mRNA stability. To further confirm the role of NSUN2 in m⁵C modification of PTEN mRNA, m⁵C-MeRIP-qPCR was performed. NSUN2 knockdown significantly reduced the m⁵C enrichment of PTEN mRNA (Fig. 2D), verifying NSUN2 as the primary m⁵C writer for PTEN. To clarify the role of YBX1 in NSUN2-mediated regulation of PTEN mRNA levels, we performed NSUN2 knockdown followed by RNA immunoprecipitation (RIP)-qPCR analysis. Western blot analysis was used to evaluate YBX1 immunoprecipitation efficiency in each group. The results demonstrated that NSUN2 knockdown did not affect the protein levels of YBX1, nor did it alter the IP efficiency of YBX1 across different experimental groups (Fig. 2E, upper panel). This consistency excludes the possibility that variations in RIP-qPCR outcomes were due to experimental discrepancies in IP efficiency or NSUN2-mediated reduction of YBX1 protein levels. Further RIP-qPCR analysis revealed that PTEN mRNA was substantially enriched in YBX1 immunoprecipitates under normal conditions, while NSUN2 depletion significantly impaired YBX1 binding to PTEN mRNA (Fig. 2E, bottom panel). These findings suggest that YBX1's capacity to regulate PTEN mRNA is dependent on NSUN2-mediated m⁵C modification, highlighting the role of YBX1 as an m⁵C reader that stabilizes PTEN transcripts. Supporting these experimental observations, Pearson's correlation analysis using data from The Cancer Genome Atlas-Liver Hepatocellular Carcinoma (TCGA-LIHC) database demonstrated a positive correlation between PTEN and NSUN2 expression, as well as between PTEN and YBX1 transcript levels in HCC tissues (Fig. 2F). These findings together indicate that NSUN2, as an m⁵C writer, collaborates with the m⁵C reader YBX1 to enhance PTEN mRNA stability through m⁵C modification, thereby maintaining PTEN expression.

Fig. 2.

NSUN2 and YBX1 regulate PTEN m⁵C levels. A Western blot analysis of PTEN protein levels in HepG2 and Huh7 cells after knockdown of NSUN2, NSUN6, YBX1, or ALYREF with specific siRNAs. B RT-qPCR analysis of PTEN mRNA levels in the corresponding knockdown cells. C RNA stability assays displayed the PTEN mRNA half-life in NSUN2 or YBX1 knockdown HepG2 and Huh7 cells by RT-qPCR at 0, 2, 4, and 6 h after treatment with 5 μg/mL ActD. D m⁵C-MeRIP-qPCR analysis showing the effect of NSUN2 knockdown on m⁵C modification levels of PTEN mRNA. E Upper panel: IP efficiency of YBX1 was validated using Western blot. GAPDH was used as protein control. Bottom panel: RIP-qPCR displayed the relative enrichment of PTEN mRNA in HepG2 cells upon NSUN2 knockdown precipitated with lgG or YBX1 antibody with the normalization to input. F The correlation between mRNA expression levels for NSUN2 or YBX1 and PTEN in HCC tissues was evaluated by TCGA-LIHC database. Data are presented as mean ± SD from three independent experiments. ns: not significant; ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001.

HBV decreases expression of NSUN2 and YBX1

To investigate the impact of HBV on host m⁵C-regulatory enzymes, we examined the expression of NSUN2/YBX1/PTEN axis in an established HBV transgenic (HBV-Tg) mouse model. Both protein and mRNA levels of NSUN2, YBX1, and PTEN were significantly reduced in liver tissues from HBV-Tg mice compared with wild-type (WT) controls (Fig. 3A and B). To further validate this association in a physiologically relevant human setting, we analyzed primary human hepatocytes (PHHs) following HBV infection. HBV-infected PHHs similarly exhibited marked decreases in NSUN2, YBX1, and PTEN at both the protein (Fig. 3C) and mRNA levels (Fig. 3D). Collectively, these in vivo and primary-cell data indicate that HBV infection is associated with coordinated downregulation of NSUN2, YBX1, and PTEN. We next confirmed these expression changes in HBV-producing cell models. Western blot and RT-qPCR analyses demonstrated that both protein and mRNA levels of NSUN2 and YBX1 were significantly reduced in HepAD38/Tet⁻ cells, HBV-infected HepG2-NTCP cells, and recombinant adenovirus Ad-1.1 × HBV-infected HepG2 and Huh7 cells compared to their respective controls (Fig. 3E and F). To further elucidate the underlying mechanisms, we cloned the promoter regions of NSUN2 and YBX1 and performed luciferase reporter assays. The results indicated a significant reduction in NSUN2 promoter activity in HepAD38/Tet⁻ cells relative to the control cells, consistent with previous findings (Ding et al., 2024), whereas the YBX1 promoter activity remained unchanged between HepAD38/Tet⁺ and HepAD38/Tet⁻ cells (Fig. 3G). This suggests that HBV represses NSUN2 expression at the transcriptional level, while YBX1 expression is not transcriptionally regulated by HBV. Furthermore, RNA stability assays revealed a marked reduction in the mRNA half-lives of both NSUN2 and YBX1 in HepAD38/Tet⁻ cells (Fig. 3H), indicating that HBV-mediated downregulation of NSUN2 occurs through both transcriptional and post-transcriptional mechanisms, while the reduction in YBX1 expression is primarily driven by post-transcriptional regulation. These findings highlight the dual mechanisms by which HBV modulates host m⁵C-regulatory enzymes, contributing to impaired m⁵C modification of PTEN mRNA.

Fig. 3.

HBV decreases the expression of NSUN2 and YBX1. A Western blot analysis of NSUN2, YBX1, and PTEN protein levels in livers from wild-type (WT) controls and HBV-transgenic (HBV-Tg) mice (n = 6 per group). B RT–qPCR analysis of NSUN2, YBX1, and PTEN mRNA levels in livers from WT controls and HBV-Tg mice (n = 6 per group). C Western blot analysis of NSUN2, YBX1, and PTEN protein levels in primary human hepatocytes (PHHs) infected with HBV or mock-infected. D RT–qPCR analysis of NSUN2, YBX1, and PTEN mRNA levels in HBV-infected or mock-infected PHHs. E, F Western blot (E) and RT-qPCR (F) detecting the protein and mRNA expression changes of NSUN2 and YBX1 in HepAD38/Tet⁻ cells, HBV-infected HepG2-NTCP cells, or Ad-1.1 × HBV-infected HepG2 and Huh7 cells. G Luciferase reporter assay showing the relative luciferase activity in HepAD38/Tet⁺ and HepAD38/Tet⁻ cells transfected with NSUN2 or YBX1 promoter reporter plasmids. H RNA stability assays displayed the mRNA half-life of NSUN2 and YBX1 in HepAD38/Tet⁺ and HepAD38/Tet⁻ cells by RT-qPCR at 0, 2, 4, and 6 h after treatment with 5 μg/mL ActD (for NSUN2 transcript) or at 0, 4, 8 and 12 h after treatment with 15 μg/mL ActD (for YBX1 transcript). Data are presented as mean ± SD from three independent experiments. ns: not significant; ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001.

HBV decreases PTEN expression via NSUN2 and YBX1-mediated m⁵C modification

To elucidate the mechanisms involved in the HBV-induced reduction in PTEN m⁵C methylation, PTEN expression levels were examined following NSUN2 or YBX1 overexpression in HepAD38/Tet⁺ or HepAD38/Tet⁻ cells. The results showed that overexpression of NSUN2 or YBX1 mitigated the inhibitory effect of HBV on PTEN expression at both protein and mRNA levels (Fig. 4A and B). Further, RNA stability assays demonstrated that PTEN mRNA stability was significantly reduced in HepAD38/Tet⁻ cells as compared to HepAD38/Tet⁺ cells, as shown by a shorter mRNA half-life. Overexpression of NSUN2 or YBX1 in HepAD38/Tet⁻ cells partially restores PTEN mRNA stability, evidenced by a slower rate of mRNA decay and a significant increase in its half-life. Both NSUN2 and YBX1 overexpression enhanced PTEN mRNA stability in HepAD38/Tet⁻ cells, bringing it closer to the levels observed in HepAD38/Tet⁺ cells (Fig. 4C). These results indicate that HBV modulates PTEN expression through the m⁵C-modification enzymes NSUN2 and YBX1. m⁵C-MeRIP-qPCR analysis showed that HBV-mediated suppression of m⁵C modification on PTEN mRNA was alleviated by overexpression of NSUN2 (Fig. 4D). Western-blot verification confirmed comparable pull-down efficiency of NSUN2 and YBX1 across all RIP samples (Fig. 4E, upper panel). Subsequent RIP-qPCR showed that HBV markedly weakened the association of both NSUN2 and YBX1 with PTEN mRNA in cells expressing the wild-type PTEN-WT-Flag construct, whereas this loss of binding was no longer observed when the m⁵C site was mutated (PTEN-C89717756A-Flag; Fig. 4E, lower panel). Taken together, these findings demonstrate that HBV downregulates the m⁵C abundance of PTEN mRNA via NSUN2 and YBX1, leading to reduced PTEN expression.

Fig. 4.

HBV regulates PTEN mRNA m⁵C modification via NSUN2 and YBX1. A Western blot analysis of PTEN protein levels after overexpression of NSUN2 or YBX1 in HepAD38/Tet⁺ and HepAD38/Tet⁻ cells. B RT-qPCR analysis of PTEN mRNA levels following overexpression of NSUN2 or YBX1. C RNA stability assays displayed the PTEN mRNA half-life in HepAD38/Tet⁺ and HepAD38/Tet⁻ cells upon NSUN2 or YBX1 overexpression by RT-qPCR at 0, 2, 4, and 6 h after treatment with 5 μg/mL ActD. D m⁵C-MeRIP-qPCR analysis of m⁵C modification levels of PTEN mRNA following NSUN2 overexpression. E Upper panel: IP efficiency of NSUN2 or YBX1 was validated using Western blot. GAPDH was used as protein control. Bottom panel: RIP-qPCR displayed the relative enrichment of PTEN-Flag mRNA at the indicated site in HepAD38/Tet⁺ and HepAD38/Tet⁻ cells upon PTEN-WT-Flag or PTEN-C89717756A-Flag overexpression with lgG, NSUN2 or YBX1 antibody with the normalization to input. Data are presented as mean ± SD from three independent experiments. ns: not significant; ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P< 0.001, ∗∗∗∗P< 0.0001.

HBV promotes HCC malignancy by regulating NSUN2 and YBX1-mediated PTEN expression

PTEN is a well-established tumor suppressor and its reduced expression is associated with progression and poor prognosis in HBV-related HCC (Li et al., 2016). To investigate whether NSUN2 and YBX1 contribute to HBV-induced HCC tumorigenesis and progression through the regulation of PTEN expression, we treated HepAD38 cells (HepAD38-NSUN2, HepAD38-YBX1 or HepAD38-Control) with the PTEN inhibitor VO-Ohpic or vehicle in the presence or absence of tetracycline. A CCK-8 assay showed significantly higher cell viability in HepAD38/Tet⁻ cells than in HepAD38/Tet⁺ cells. This HBV-associated pro-proliferative effect was partially attenuated by NSUN2 or YBX1 overexpression whereas VO-Ohpic treatment reversed this attenuation and restored the proliferative capacity of HBV-producing cells (Fig. 5A). Colony formation assays showed a significant increase in colony numbers in HepAD38/Tet⁻ cells compared to HepAD38/Tet⁺ cells. Overexpression of NSUN2 or YBX1 attenuated this effect, leading to a decrease in colony formation. However, the addition of VO-Ohpic reversed the suppressive effects of NSUN2/YBX1, restoring HBV-enhanced cell proliferation (Fig. 5B). Similarly, EdU incorporation assay revealed a marked increase in the percentage of EdU-positive cells in HepAD38/Tet⁻ cells compared with HepAD38/Tet⁺ cells. NSUN2 or YBX1 overexpression partially blunted this increase, and VO-Ohpic again rescued the HBV-driven proliferative phenotype (Fig. 5C), supporting a functional antagonism between the NSUN2/YBX1 axis and PTEN signaling in controlling HBV-mediated proliferation. Wound healing assay (Fig. 5D) demonstrated that HepAD38/Tet⁻ cells exhibited significantly higher migration rates compared to HepAD38/Tet⁺ cells. Overexpression of NSUN2 or YBX1 mitigated this effect. Importantly, treatment with the PTEN inhibitor VO-Ohpic counteracted the inhibitory effects of NSUN2 or YBX1 overexpression, restoring migration rates in HepAD38/Tet⁻ cells to levels comparable to the control group (HepAD38-Control + Vehicle). Transwell assays further revealed that HBV significantly enhanced the migration and invasion of HepAD38 cells. Overexpression of NSUN2 or YBX1 substantially diminished these effects, but VO-Ohpic treatment counteracted the suppression induced by NSUN2/YBX1 overexpression (Fig. 5E). These results confirm that NSUN2 and YBX1 play a critical role in HBV-driven tumorigenesis of HCC by regulating PTEN expression. These findings demonstrate the pivotal role of NSUN2 and YBX1 in HBV-induced tumorigenesis of HCC by modulating PTEN expression, highlighting the intricate interaction between HBV and host m⁵C modification enzymes in driving malignant behaviors.

Fig. 5.

HBV promotes malignant behavior of HCC cells via the NSUN2-YBX1-PTEN axis. A Cell viability assessed by CCK-8 in HepAD38-NSUN2 or HepAD38-YBX1 cells treated with the PTEN inhibitor VO-Ohpic or vehicle, in the presence or absence of tetracycline. B Colony formation assay showing the clonogenecity of the indicated cells. C EdU assay detecting Immunofluorescent staining of the EdU-positive cells in the indicated cells, scale bar = 50 μm. D Wound-healing assay assessing relative motility of the indicated cells. E Transwell assays evaluating migration and invasion of the indicated cells. Data are presented as mean ± SD from three independent experiments. ∗P < 0.05, ∗∗P < 0.01.

SHBs and HBx are key regulators of NSUN2 and YBX1

To identify which HBV viral components are responsible for the downregulation of NSUN2 and YBX1, HepG2 cells were transfected with plasmids encoding the 1.2-unit-length HBV genome (pREP-1.2 × HBV) or its mutant variants lacking HBV surface antigens (HBsAg) [pREP-1.2 × HBV-HBsAg(−)], HBx [pREP-1.2 × HBV-HBx(−)], Pol [pREP-1.2 × HBV-Pol(−)] or HBc [pREP-1.2 × HBV-HBc(−)]. Transfection with the full-length HBV genome (pREP-1.2 × HBV) led to a significant reduction in NSUN2 and YBX1 protein levels, while cells transfected with HBsAg-null or HBx-null replicons exhibited markedly diminished suppression of NSUN2 and YBX1 expression. However, the absence of Pol or HBc did not produce any observable differences in NSUN2 and YBX1 expression compared to the wild-type HBV genome (Fig. 6A). Consistently, RT-qPCR analysis demonstrated that the mRNA levels of NSUN2 and YBX1 were substantially decreased in HepG2 cells transfected with pREP-1.2 × HBV. This reduction was notably attenuated in cells transfected with HBsAg-deficient [pREP-1.2 × HBV-HBsAg(−)] or HBx-deficient [pREP-1.2 × HBV-HBx(−)] mutants, while the depletion of Pol or HBc expression did not alleviate the HBV-induced suppression of NSUN2 and YBX1 (Fig. 6B), indicating the necessity of HBsAg and HBx for NSUN2 and YBX1 suppression. HBsAg comprises three envelope glycoproteins including large (LHBs), middle (MHBs), and small (SHBs) surface proteins. To further dissect the specific roles of these envelope proteins and HBx in regulating NSUN2 and YBX1 expression, HepG2 cells were transfected with Flag-tagged LHBs, MHBs, SHBs, HBx, Pol or HBc plasmids (Fig. 6C). As shown in the upper panel of Fig. 6C, overexpression of Pol or HBc did not alter NSUN2 or YBX1 expression at the protein level. Notably, all four viral components (LHBs, MHBs, SHBs, and HBx) suppressed NSUN2 protein levels, with SHBs and HBx exhibiting the most pronounced inhibitory effect compared to LHBs and MHBs. Similarly, YBX1 protein levels were significantly decreased following the overexpression of SHBs, MHBs, or HBx; however, SHBs and HBx were the primary drivers of YBX1 suppression. These findings highlight SHBs and HBx as principal regulators of NSUN2 and YBX1 expression in HBV-infected cells, suggesting the pivotal role of SHBs and HBx in modulating the m⁵C machinery during HBV-associated hepatocarcinogenesis.

Fig. 6.

SHBs and HBx downregulate NSUN2 and YBX1 expression. A Western blot analysis of NSUN2 and YBX1 protein levels in HepG2 cells transfected with pREP-1.2 × HBV (WT) or mutant constructs lacking HBsAg (containing three forms of LHBs, MHBs and SHBs), HBx, Pol or HBc. Bands corresponding to Pol are indicated by an asterisk. B RT-qPCR analysis of NSUN2 and YBX1 mRNA levels in HepG2 cells transfected with HBV constructs. C Western blot analysis of NSUN2 and YBX1 protein levels in HepG2 cells transfected with Flag-tagged LHBs, MHBs, SHBs, HBx, Pol or HBc expression plasmids, respectively. The bands corresponding to the expected LHBs, MHBs, SHBs or Pol protein are marked with asterisk. Data are presented as mean ± SD from three independent experiments. ns: not significant; ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001.

Discussion

Chronic HBV infection remains a leading cause of hepatocellular carcinoma (HCC) worldwide (Reynolds-Cortez et al., 2024). Despite advancements in oncology research and surgical treatments, the prognosis of HBV-related HCC remains poor. PTEN, a critical tumor suppressor, is frequently dysregulated in various cancers including HCC, where its deletion or reduced expression is observed in nearly half of cases (Wu et al., 2003; Karoui et al., 2004; Nassif et al., 2004; Álvarez-Garcia et al., 2019). In HCC, PTEN expression is significantly lower in tumor tissues compared to adjacent normal tissues, correlating with tumor differentiation, clinical stage, invasion, and metastasis (Zhou and Li, 2018; Akula et al., 2019). HBV-driven silencing of PTEN is central to HBV-related hepatocarcinogenesis and operates through several distinct mechanisms. At the transcriptional level, HBV enhances DNA-methyltransferase (DNMT) activity, producing hyper-methylation of the PTEN promoter and repressing transcription (Guan et al., 2022). In parallel, the viral transactivator HBx binds to and functionally inactivates p53, preventing p53 from engaging the PTEN promoter and further dampening PTEN transcription (Chung et al., 2003). Post-transcriptionally, HBV increases the abundance of PTEN-targeting microRNAs-including miR-181a, miR-382, miR-362 and miR-19a-which bind the PTEN 3′-UTR to inhibit translation or accelerate mRNA decay (Ma et al., 2020; Tang et al., 2020). While the transcriptional regulation of PTEN by HBV has been extensively studied, its post-transcriptional regulation remains largely unexplored. RNA methylation is a critical mechanism of post-transcriptional regulation. Advances in high-throughput sequencing have gradually elucidated the pivotal roles of various RNA methylation modifications, especially the most common m⁶A modification, in viral replication and pathogenesis. For instance, during vesicular stomatitis virus (VSV) infection, RNA helicase dead box protein 46 (DDX46) binds the mitochondrial antiviral signaling protein (MAVS), TNF receptor-associated Factor 3 (TRAF3) and TRAF6 transcripts; recruitment of the m⁶A demethylase alkB homolog 5 (ALKBH5) removes their m⁶A marks, causing nuclear retention of these mRNAs and lowering protein output, which dampens type I interferon production (Zheng et al., 2017). Analogously, Kim et al. showed that HBV-induced m⁶A modification of the PTEN 3′-UTR destabilizes the transcript, activates PI3K/AKT signaling and thereby promotes HCC progression (Kim et al., 2021). Among RNA methylation types, m⁶A and m⁵C are the most prevalent. While the role of m⁶A in HBV infection and HBV-related HCC is increasingly understood, the contribution of m⁵C to these processes is still largely undefined. Clarifying how m⁵C marks on host mRNAs influence HBV-driven hepatocarcinogenesis therefore remains a critical area for future research.

mRNA m⁵C modification is increasingly recognized as a driver of oncogenesis. In the present study, we show that HBV compromises PTEN expression by removing a single m⁵C mark (chr10:89717756) within the PTEN coding region. This CDS-localized modification differs from the better-known m⁶A marks that cluster in the 3′-UTR, indicating a distinct layer of post-transcriptional control. Substituting cytosine 89717756 with adenine (PTEN-C89717756A) abolished the HBV-induced drop in luciferase activity driven by the wild-type PTEN CDS, confirming the site's functional importance. Although this mutation markedly buffered the HBV-dependent loss of PTEN, it did not restore PTEN mRNA stability or protein abundance completely, implying that additional HBV-regulated pathways act in concert with m⁵C erasure to destabilize PTEN transcripts.

mRNA m⁵C marks are deposited primarily by the methyl-transferases NSUN2 and NSUN6, and are recognized by the reader proteins ALYREF and YBX1. Targeted depletion of NSUN2 or YBX1 in HepG2 and Huh7 cells markedly shortened PTEN mRNA half-life and reduced PTEN protein abundance, establishing both factors as pivotal m⁵C regulators of PTEN. Conversely, HBV infection lowered endogenous NSUN2 and YBX1 levels, whereas ectopic expression of either protein partially restored PTEN expression. These findings are consistent with reports that NSUN2, together with YBX1, stabilizes laminin subunit gamma 2 (LAMC2) mRNA in head-and-neck squamous-cell carcinoma and drives EMT-mediated malignancy (Huang S. et al., 2024), and that the same writer–reader pair enhances human growth hormone 1 (HGH1) mRNA stability in breast cancer, where blockade of NSUN2's catalytic activity or disruption of YBX1 binding hampers tumor growth (Zhang X. et al., 2024). A similar NSUN2/YBX1-dependent mechanism has been described for leucine-rich repeat containing 8A (LRRC8A) mRNA in cervical cancer, promoting translation and suppressing apoptosis (Chen et al., 2023). Sub-cellular localization studies further support complementary roles: NSUN2 is predominantly nuclear and facilitates nuclear export of m⁵C-modified transcripts (Zhang R.K. et al., 2024; Gonskikh et al., 2025), whereas YBX1 resides mainly in the cytoplasm, safeguarding mRNA integrity and modulating translation by binding m⁵C sites (Wang et al., 2023; Meng et al., 2024). Although the present study did not dissect their physical interaction, delineating the dynamic cooperation between NSUN2 and YBX1 in controlling PTEN, and possibly other HBV-responsive transcripts, remains an important avenue for future research.

In addition to canonical writer–reader regulation, we acknowledge that m⁵C dynamics may also involve oxidative turnover. Traditionally, ten-eleven translocation (Tet) enzymes are known primarily for inducing oxidation of m⁵C in DNA (He et al., 2011; Ito et al., 2011). Although recent studies have reported that TET proteins (TET1/2/3) can convert m⁵C to 5-hydroxy-methylcytosine (hm⁵C) on mRNA, the systemic investigation of the role and biological relevance of mRNA hm⁵C is still in its infancy (Yang et al., 2022; Li et al., 2023; Fan et al., 2024). Studies have found that the expression level of TET1 was significantly elevated in HBV DNA–integrated HepG2.2.15 cells and HBx–integrated Huh7 cells while TET2 and TET3 expression decreased (Liu et al., 2019). We also assessed TET1/2/3 expression changes in HBV-inducible HepAD38 cells by Western blotting. The results showed that a significant upregulation of TET1 protein level, but an obvious decreasing of TET2 and TET3 protein levels in HepAD38/Tet⁻ cells compared with the control cells (Supplementary Fig. S1A). TET1, and TET2 have been reported to positively regulate the expression of its downstream genes PTEN via DNA demethylation of CpG islands in the PTEN promoter region, and TET3 was identified as the novel ceRNA that may positively modulate PTEN expression by competitive miRNA (Zhang et al., 2019; Jin et al., 2020; Roquid et al., 2020). To better describe the correlation between PTEN and TET proteins in HCC, we used small interfering RNAs (siRNA) to knock down TET1/2/3 respectively in HepG2 and Huh7 cells. In line with the previous literature, we discovered that knockdown of TET1, TET2 and TET3 all led to the reduction of the protein levels of PTEN in both HepG2 and Huh7 cells (Supplementary Fig. S1B), indicating that the expression of PTEN was positively correlated with TET1/2/3 expression. Given that HBV upregulated TET1 expression while TET1 positively regulates PTEN expression, we speculated that the HBV-mediated suppression of PTEN expression was unrelated to the function of TET1 to oxidize m⁵C RNA modification, and the upregulation of TET1 expression in the context of HBV infection may be considered an antagonistic mechanism within cellular microenvironment that counteracts HBV-induced downregulation of PTEN expression. To further verify whether TET2 or TET3 can regulate PTEN mRNA methylation levels, we conducted m⁵C-MeRIP-qPCR to determine the m⁵C level in PTEN transcript. The results demonstrated that knockdown of TET2 had no effect on PTEN mRNA m⁵C levels, and knocking down TET3 markedly reduced the m⁵C abundance in PTEN mRNA (Supplementary Fig. S1C). This implies that the role of TET2 or TET3 in regulating PTEN expression may operate independently of its effects on PTEN mRNA demethylation. Therefore, our data confirmed that TET family members (TET1/2/3) do not function as the “eraser” in the HBV-mediated regulation of PTEN mRNA m⁵C modification.

PTEN is a central tumor suppressor whose loss or inactivation promotes aggressive behaviour in many cancers. In the current work, forced expression of either NSUN2 or YBX1 partly restored PTEN levels in HBV-producing cells and attenuated HBV-driven proliferation, migration and invasion, underscoring the functional importance of the NSUN2/YBX1/m⁵C-PTEN axis. In reconciling our functional data with reports that NSUN2 can promote HCC (Zhai et al., 2021; Song et al., 2023; Xing et al., 2024), we propose a context- and substrate-dependent model. In our HBV-positive systems, NSUN2 and YBX1 constrain malignant behaviors because they stabilize PTEN mRNA via an m⁵C mark within the PTEN coding sequence (chr10:89717756); HBV replication (or HBsAg/HBx expression) lowers NSUN2/YBX1 abundance and binding, erodes this m⁵C, and thereby destabilizes PTEN, which secondarily enhances proliferation, migration, and invasion. These experiments were performed in the HepAD38 Tet-off model (genotype D, 1.1 × HBV genome integrated on a HepG2 background), in which tetracycline withdrawal drives robust pgRNA transcription, core/precore expression, reverse transcription, and Dane particle production, while tetracycline addition suppresses pgRNA yet can allow minimal basal 2.1-kb mRNA transcription, resulted in detectable expression of SHBs even under Tet⁺. Such features, together with differences across studies in viral source (e.g., HepAD38 versus HepG2.2.15 supernatants), viral titer/MOI, infection duration, and cell background (infection versus plasmid transfection), likely contribute to divergent readouts of NSUN2/YBX1 biology. Notably, in our hands, enforced NSUN2/YBX1 expression increases PTEN mRNA stability and protein, mitigates HBV-enhanced colony formation, wound closure, transwell migration, and invasion, and these effects are partly reversed by pharmacologic PTEN blockade—linking the phenotype to the PTEN axis rather than a global antiproliferative effect of NSUN2. We therefore suggest that NSUN2 can act through distinct RNA targets with opposing consequences: in non-viral or early disease settings NSUN2 may favor oncogenic programs, whereas under HBV replication the selective loss of NSUN2/YBX1 function preferentially undermines the tumor-suppressive PTEN pathway. Mapping NSUN2/YBX1 targetomes in HBV-positive versus HBV-negative HCC and testing PTEN-independent outputs in vivo will help define, when and through which substrates, NSUN2 promotes or restrains liver tumorigenesis.

HBV down-regulates this axis mainly through its SHBs and HBx. HBx is a pleiotropic oncoprotein that modulates host transcription; for example, it cooperates with specificity protein 1 (SP1) to activate LRRC8A, fuelling arachidonic-acid metabolism and worsening hepatic lipid dysregulation (Liu et al., 2020). SHBs, the most abundant HBV envelope protein, has likewise been linked to oncogenic reprogramming: it up-regulates C–C chemokine receptor type 9 (CCR9) via vir like m⁶A methyltransferase associated (VIRMA)-dependent m⁶A methylation, thereby accelerating HCC progression and conferring regorafenib resistance (Lv et al., 2024). Prior work shows that HBx can repress NSUN2 transcription by blocking early growth response protein 1 (EGR1) occupancy at the NSUN2 promoter (Ding et al., 2024) and broadly reshapes host chromatin (Zhu et al., 2010; Xie et al., 2014; Wang et al., 2015; Hamamoto et al., 2018). Consistent with a transcriptional component, our promoter-reporter assays indicate HBV lowers NSUN2 promoter activity. We therefore propose a chromatin-based mechanism in which HBx recruits histone modifiers (e.g., histone deacetylases (HDACs), H3K9/H3K27 methyltransferases) (Wang et al., 2015; Hamamoto et al., 2018), and/or DNA methyltransferases (DNMT1/DNMT3A) (Xie et al., 2014; Fu et al., 2016) to the NSUN2 promoter, enforcing a repressive state (loss of H3K9ac; gain of H3K9me3/H3K27me3; CpG hypermethylation). To test this, we will perform NSUN2-promoter bisulfite sequencing and ChIP-qPCR/ChIP-seq for EGR1, DNMT1/DNMT3A, H3K9ac, H3K9me3, and H3K27me3 under HBV-on/off conditions, coupled to EGR1 gain-/loss-of-function rescue of promoter activity. In addition, RNA stability assays demonstrated a sharp decline in the mRNA half-lives of NSUN2 in HepAD38/Tet⁻ cells. Therefore, we hypothesize that HBV might recruit specific mRNA-destabilizing RNA-binding proteins (RBPs) to the NSUN2 transcript or regulate specific miRNAs to target the 3′-UTR of NSUN2 mRNA, promoting its decay. The precise mechanism remains to be defined and constitutes a vital next step in understanding HBV-host interactions. For YBX1, our data support post-transcriptional control (ActD chases shorten YBX1 mRNA half-life without promoter effects), suggesting HBV engages miRNA and/or RBP circuitry to destabilize YBX1 transcripts. We will implement Argonaute 2(AGO2)-RIP-seq/miRNA-seq to identify HBV-induced miRNAs targeting the YBX1 3′-UTR, luciferase-UTR reporters with site-directed mutants for validation, and miRNA mimics/antagomirs to test causality. In parallel, RBP-centric screens [(RIP-seq/CLIP for candidate decay factors such as AU-rich element RNA-binding protein 1 (AUF1), K-homology splicing regulatory protein (KSRP), or zinc finger protein 36 (ZFP36)] will be paired with decay assays after RBP knockdown/overexpression to define effectors of YBX1 mRNA turnover during HBV replication. Together, these orthogonal experiments will mechanistically resolve the transcriptional (NSUN2) and post-transcriptional (NSUN2/YBX1) layers through which HBV down-tunes the NSUN2/YBX1–m⁵C–PTEN axis.

HBV can depress PTEN through multiple routes (promoter DNA methylation, p53 inactivation, miRNAs, and m⁶A) and therefore evaluated whether our observations could be explained indirectly via an upstream “third factor” rather than by an m⁵C mark on PTEN itself. Several convergent observations argue that the predominant effect we observe is direct and m⁵C-dependent. First, RIP-qPCR demonstrates a physical association between YBX1 and PTEN mRNA that is lost when NSUN2 is depleted, consistent with a writer–reader pair acting on the same transcript rather than through an intermediate factor. Second, m⁵C-MeRIP-qPCR shows that NSUN2 overexpression increases m⁵C enrichment on PTEN mRNA and counteracts the HBV-induced loss of this modification. Third, HBV markedly diminishes NSUN2 and YBX1 binding to PTEN mRNA in cells expressing the wild-type PTEN construct, whereas this loss of binding is abolished by a single C to A substitution at chr10:89717756, pinpointing a discrete cis-element required for the effect. Fourth, mRNA-decay kinetics reveal that HBV selectively shortens the half-life of the wild-type PTEN transcript but not the chr10:89717756 mutant; if NSUN2/YBX1 acted primarily through an upstream regulator of DNMT/p53/miRNAs, comparable destabilization of both transcripts would be expected. Epistasis with the m⁶A pathway further supports a direct m⁵C mechanism that operates in parallel to other HBV-driven controls. Moreover, silencing METTL3/14 partially restores PTEN, and combining METTL3/14 knockdown with NSUN2/YBX1 overexpression yields a greater rescue of PTEN mRNA stability than either manipulation alone, indicating cooperative—but mechanistically separable—contributions of m⁶A and m⁵C (data not shown). We acknowledge that the C89717756 mutation does not fully normalize PTEN, implying additional layers of regulation reported by others (promoter methylation, p53, and miRNAs) likely act alongside the m⁵C pathway. Nevertheless, the site-specific loss of m⁵C, the dependency on the NSUN2/YBX1 writer–reader axis, and the selective effect on wild-type but not mutant PTEN collectively support a model in which HBV directly destabilizes PTEN by erasing a critical CDS-embedded m⁵C mark while other pathways provide additive influence.

Recent studies have reported modest increases in hepatic NSUN2 in HBV-positive specimens (Feng et al., 2023; Su et al., 2024), raising the question of whether HBV uniformly upregulates NSUN2 in vivo. However, the biological contexts of these datasets differ substantially from ours. Feng et al. reanalyzed non-tumor liver tissue (GSE83148) (Feng et al., 2023), while Su et al. compared tumor versus matched non-tumor tissue within HBV-positive patients (GSE94660), rather than directly contrasting HBV-positive with HBV-negative HCC (Su et al., 2024). Such designs capture HBV infection in either non-malignant or mixed tissue contexts, where viral RNA modification might benefit from modest NSUN2 upregulation. By contrast, our analyses focus on HBV-related HCC and hepatoma models, where the oncogenic environment and viral proteins impose distinct regulatory pressures.

Indeed, multiple independent lines of evidence support context-dependent downregulation of the NSUN2/YBX1 axis during HBV replication. Public datasets demonstrate that hepatic NSUN2 expression decreases in specific phases of chronic HBV infection (immune-active (IA), inactive carrier (IC); GSE230397) (Supplementary Fig. S2A) and is significantly lower in HBV-related HCC compared with non-HBV-related HCC (Supplementary Fig. S2B). Furthermore, Western blot analysis of freshly resected HBV-positive HCC specimens revealed a strong inverse correlation between NSUN2 and HBx protein abundance (Supplementary Fig. S2C), consistent with our mechanistic data that HBsAg/HBx repress NSUN2 and YBX1 expression. Taken together, these observations suggest that NSUN2 regulation by HBV is highly stage- and context-specific: HBV may harness NSUN2 in non-malignant hepatocytes to facilitate viral replication, but downregulates the NSUN2/YBX1 pathway in tumorigenic settings to erode PTEN m⁵C and drive HCC progression. This interpretation reconciles the apparent discrepancies with prior studies and underscores the importance of disease context, viral load, and tumor stage when evaluating HBV–host interactions at the epitranscriptomic level.

In addition to NSUN2, the regulation of YBX1 by HBV also appears to be context-dependent. In our dataset, both mRNA and protein levels of YBX1 were consistently reduced during HBV replication or infection, and RIP-qPCR assays confirmed that HBV significantly weakened the binding of YBX1 to PTEN mRNA. These findings suggest that under replicative or tumor-associated conditions, HBV may suppress YBX1 expression and function, thereby destabilizing PTEN transcripts. This contrasts with reports where YBX1 does not consistently decrease in HBV-positive systems. For example, Ding et al. (Ding et al., 2024) observed lower YBX1 protein in HepG2.2.15 versus parental HepG2 cells, but no reduction after 7-day infection of HepG2-NTCP cells–differences that likely reflect distinct viral inocula and sources (HepG2.2.15 vs. our HepAD38-derived stocks), viral titers/MOI, infection duration, and in transient-transfection models (e.g., Huh7 with HBV 1.1-mer), unverified HBV protein expression. While YBX1 may be co-opted to maintain viral RNA integrity in some settings, in the oncogenic context HBV appears to attenuate YBX1 expression as part of a broader program to erode host tumor suppressor pathways.

Beyond reconciling literature discrepancies, our findings also highlight broader implications for liver disease and oncology. Epitranscriptomic regulation is increasingly recognized as a dynamic interface between viral infection and host cellular homeostasis. The NSUN2/YBX1-m⁵C-PTEN axis characterized here provides a paradigm for how HBV co-opts RNA modification machinery to disable tumor suppressor networks, complementing previously described mechanisms involving promoter methylation, miRNAs, and m⁶A. Extending this model, NSUN2/YBX1 dysregulation could plausibly influence additional targets relevant to hepatocarcinogenesis, metabolic adaptation, and immune evasion. Moreover, given that NSUN2 and YBX1 are implicated in multiple malignancies, the virus-driven suppression of this pathway in HCC may reflect a more generalizable vulnerability of tumor cells to RNA modification imbalance.

From a translational perspective, these results point to potential therapeutic opportunities. Targeting the NSUN2/YBX1-m⁵C-PTEN axis could restore PTEN expression and counteract HBV-mediated oncogenesis. Pharmacological modulation of RNA modification enzymes is an emerging field, and small-molecule inhibitors or stabilizers of NSUN2/YBX1 function may offer adjunctive strategies in HBV-associated HCC. Furthermore, integrating epitranscriptomic biomarkers such as PTEN m⁵C status into clinical practice could improve patient stratification, predict prognosis, and guide therapeutic decisions, particularly in populations with chronic HBV infection. Finally, because epitranscriptomic dysregulation has been implicated in viral infections beyond HBV, these insights may extend to broader infectious disease contexts where RNA modifications shape host–pathogen interactions.

Conclusions

In summary, as pictorially modelled in Fig. 7, this study identifies PTEN as a direct target of NSUN2/YBX1-mediated m⁵C modification and highlights the NSUN2/YBX1/m⁵C-PTEN signaling axis as a key pathway in HBV-related HCC pathogenesis. These findings offer potential prognostic biomarkers and underscore the therapeutic potential of targeting epitranscriptomic modifications in HBV-related HCC. Further studies are warranted to elucidate the precise molecular mechanisms underlying the interplay of viral and host RNA modifications in HCC progression.

Fig. 7.

Schematic model illustrating how HBV suppresses m⁵C modification on PTEN mRNA, promoting tumor growth and aggressiveness. The left panel depicts the normal process where NSUN2 (writer) catalyzes m⁵C modification at the PTEN mRNA site (chr10:89717756), while YBX1 (reader) recognizes m⁵C-modified PTEN mRNA and enhances its stability. This leads to increased PTEN protein levels, suppressing tumor growth and aggressiveness. The right panel shows the effect of HBV infection, where HBsAg (mainly SHBs) and HBx reduce NSUN2 and YBX1 expression, leading to diminished m⁵C modification, reduced PTEN mRNA stability, and decreased PTEN protein expression, thereby promoting tumor growth and aggressiveness.

Materials and methods

Patients and specimens

Tissue samples were collected from eight patients with HBV-positive hepatocellular carcinoma who underwent surgical resection at Mengchao Hepatobiliary Hospital of Fujian Medical University (Supplementary Table S2). None of the patients received preoperative radiotherapy or chemotherapy. Total protein was extracted from each specimen for Western blot analysis. Written informed consent was obtained from all participants, and the study was approved by the Ethics Committee of Mengchao Hepatobiliary Hospital of Fujian Medical University (Approval No. 2021_100_06).

Animal experiments

Male wild-type C57BL/6 mice were obtained from GemPharmatech Co., Ltd. (Nanjing, China). Age-matched male HBV transgenic (HBV-Tg) mice and wild-type (WT) littermates (6–8 weeks old) were used for all experiments. HBV-Tg mice were kindly provided by Professor Quan Yuan (Xiamen University) and were originally generated by Professor Pei-Jer Chen (National Taiwan University) (Wu et al., 2010). All animal procedures were approved by the Institutional Animal Care and Use Committee of Fujian Medical University (Approval No. IACUC FJMU 2024-0210).

Cell lines and culture conditions

The human embryonic kidney cell lines 293T [(CRL-3216; American Type Culture Collection (ATCC), Manassas, VA)], HEK293 (CRL-1573; ATCC), the human hepatoblastoma cell line HepG2 (HB-8065; ATCC), and the human hepatoma cell line Huh7 (JCRB0403; Japanese Collection of Research Bioresources Cell Bank [JCRB], Japan) were cultured in Dulbecco's Modified Eagle's Medium (DMEM; Catalog no. 12800017; Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS, P30-3302; PAN-Biotech GmbH, Aidenbach, Germany) in a 5% CO₂ incubator at 37°C. Primary human hepatocytes (PHHs) (LV-PHH001; LV-Biotech, Shenzhen, China) were maintained in PHH maintenance medium (LV-WEM001). The hepatoma cell lines HepG2-NTCP (Chung et al., 2024) and HepAD38 (Wang et al., 2016) were generously provided by Prof. QuanYuan (Xiamen University, Fujian, China). HepAD38 cells were maintained in DMEM containing 10% FBS, 3 μg/mL tetracycline (HY–B0474; MedChemExpress, Shanghai, China), and 400 μg/mL G418 (Catalog no. 345810; Sigma-Aldrich, St. Louis, MO, USA). HBV virions were collected from the supernatant of HepAD38 cells after tetracycline withdrawal. HepG2-NTCP cells were cultured in DMEM supplemented with 10% FBS. All transfections were performed when cells reached 80% confluence, using Lipofectamine 3000 Transfection Reagent (L3000015; Thermo Fisher Scientific) according to the manufacturer's instructions.

Plasmid construction

pREP-1.2 × HBV plasmid containing 1.2-unit lengths of the HBV genome (Genotype B, adwsubtype; GenBank accession number: AF100309), negative control plasmid of pREP-Sal, the mutants of pREP-1.2 × HBV-HBsAg(−), pREP-1.2 × HBV-HBc(−), pREP-1.2 × HBV-Pol(−) and pREP-1.2 × HBV-HBx(−) incapable of expressing HBsAg, HBc, Pol and HBx respectively, all of which were described previously (Lin et al., 2010). Synthetic gene fragments encoding LHBs (pre-S1 + pre-S2 + S domains; MHBs and SHBs initiation codons mutated from ATG to ACG), MHBs (pre-S2 + S domains; SHBs start codon mutated from ATG to ACG), SHBs, HBc, Pol and HBx were obtained from General Biosystems and individually cloned into the HindIII/NotI sites of the pcDNA3.1/Hygro(+) vector (V87020; Invitrogen, California, USA), generating pcDNA3.1-LHBs-Flag, pcDNA3.1-MHBs-Flag, pcDNA3.1-SHBs-Flag, pcDNA3.1-HBc-Flag, pcDNA3.1-Pol-Flag and pcDNA3.1-HBx-Flag, respectively. pShuttle-IRES-hrGFP-1.1 × HBV was constructed by inserting chemically synthesized (GenePharma, Shanghai, China) 1.1-unit lengths of the HBV genome into pShuttle-IRES-hrGFP-1 (Catalog no. 240081; Stratagene, California, USA). The full-length of PTEN coding sequence (CDS) (GenBank accession number: NM_000314.8) or cytosine-to-guanine mutants at specific m⁵C sites was chemically synthesized and cloned into the XhoI/SalI sites of the pmirGLO vector (E133A; Promega, Madison, WI, USA), the generated constructs were named pmirGLO-PTEN-WT, pmirGLO-PTEN-C89717749A, pmirGLO-PTEN-C89717750A, pmirGLO-PTEN-C89717752A, pmirGLO-PTEN-C89717756A, pmirGLO-PTEN-C89717761A, pmirGLO-PTEN-C89717768A, and pmirGLO-PTEN-all MUT (with all six sites were mutated). Using the pmirGLO-based construct as the template, Flag-tagged full-length of PTEN CDS or C89717756A mutant was PCR-amplified and cloned into the NotI/KpnI sites of the pcDNA3.1/myc-hisA(−) vector (V80020; Invitrogen). The generated constructs were named pcDNA3.1/myc-his(−)A-PTEN-WT-Flag and pcDNA3.1/myc-his(−)A-PTEN-C89717756A-Flag. The primer sequences were listed in Supplementary Table S3.

The NSUN2 promoter (NCBI ID: NC_000005.10; −2000 to +200 nt relative to the transcription start site) or the YBX1 promoter (NCBI ID: NC_000001.11; −2000 to +200 nt) was chemically synthesized and cloned into the SacI/EcoRV and KpnI/XhoI sites of the pGL4.10 vector (E665A; Promega, Madison, WI, USA), creating pGL4.10-NSUN2 and pGL4.10-YBX1 constructs. Flag-tagged full-length of NSUN2 CDS (GenBank accession number: NM_017755.6) or YBX1 CDS (GenBank accession number: NM_004559.5) was chemically synthesized and cloned into the NheI/KpnI sites of pcDNA3.1/myc-hisA(−) vector, the generated constructs were named pcDNA3.1/myc-hisA(−)-NSUN2-Flag or pcDNA3.1/myc-hisA(−)-YBX1-Flag. pNSUN2 or pYBX1 expressing Flag-Strep-tagged proteins were constructed by inserting of PCR-amplified NSUN2 and YBX1 CDS into the EcoRI/BamHI sites of lentivirus expression vector pCDH-CMV-IRES-sf3-Blast (abbreviated as pCDH-sf3) (Ding, M. et al., 2021), kindly provided by Wang Xing (Fujian Medical University). Templates used for PCR amplification were pcDNA3.1/myc-hisA(−)-NSUN2-Flag or pcDNA3.1/myc-hisA(−)-YBX1-Flag. Sequences of the primer used for plasmid construction were listed in Supplementary Table S3.

Preparation of HBV-producing cells

For transient HBV expression, HepG2 cells were transfected with the pREP-1.2 × HBV plasmid. For induced HBV expression, HepAD38 cells were cultured in tetracycline-free (Tet⁻) medium to initiate HBV replication and virion production. Following a 10-day tetracycline withdrawal, HBeAg secretion in the culture medium was quantified using an enzyme-linked immunosorbent assay (ELISA kit, WANTAIBIO, Beijing, China), according to the manufacturer's instructions. The optical density at 450 nm (OD₄₅₀) was measured until it reached a value of 2. For the HBV infection experiment, virus-containing supernatants from HepAD38/Tet⁻ cells were concentrated using 8% PEG8000 (A100159-0500; Sangon Biotech, Shanghai, China) and resuspended in DMEM at a 100-fold concentration. To infect HepG2-NTCP cells, these cells were first seeded in DMEM supplemented with 10% FBS and 1 μg/mL doxycycline (HY–N0565; MedChemExpress, Shanghai, China). Four days later, the cells were infected with HBV virions at a multiplicity of infection (MOI) of 200 viral genome equivalents (VGE) per cell in the presence of 4% PEG8000. After 24 h, the cells were washed with phosphate-buffered saline (PBS) and cultured in complete medium for 10 days. PHHs were inoculated with HBV virions for 16 h at a MOI of 1000 VGE per cell in the presence of 4% (w/v) PEG8000. After infection, cells were washed three times with PBS, maintained in maintenance medium supplemented with 2% dimethyl sulfoxide (DMSO; Cat. No. D5879, Sigma), and harvested at 1 week post-infection.

Generation of recombinant adenovirus expressing HBV

Recombinant adenoviruses expressing HBV (Ad-1.1 × HBV) and an empty control (Ad-GFP) were created using the AdEasy XL system (Catalog no. 240010; Stratagene, California, USA) as previously described (Wu et al., 2024). Briefly, the plasmid pShuttle-IRES-hrGFP-1.1 × HBV was transformed into BJ5183-AD-1 bacterial cells to enable homologous recombination. The resulting recombinant plasmid was linearized with PacI and transfected into HEK293 cells following the manufacturer's protocol. Recombinant adenoviruses expressing HBV (Ad-1.1 × HBV) were harvested from HEK293 cells after three cycles of freeze-thawing and subsequent centrifugation. The control adenovirus (Ad-GFP) was produced using the empty vector pShuttle-IRES-hrGFP-1, following the same procedure. Recombinant adenoviruses were titered by Quick Titer Adenovirus Titer Elisa Kit (VPK-110; Cell Biolabs, San Diego, CA, USA). HepG2 and Huh7 cells were infected with recombinant adenoviruses at a multiplicity of infection (MOI) of 100 viral genome equivalents (VGE) per cell. This viral dose achieved 80%–100% infection efficiency as confirmed by the expression of green fluorescent protein (GFP) observed under fluorescent microscopy (Axio Observer A1; ZEISS, Germany).

Generation of NSUN2 orYBX1 over-expressing HepAD38 cells

pNSUN2, pYBX1, or an empty vector of pCDH-sf3, along with lentiviral packaging plasmids pMDL, p-VSV-G, and pREV (K4975-00; Invitrogen), were co-transfected into 293T cells at 80% confluence in 6-cm dishes. After 48 h, the supernatant containing lentiviral particles was filtered through 0.45 μm filters and used to infect HepAD38/Tet⁺ cells cultured in 6-cm dishes. Puromycin (2 μg/mL; P9620; Sigma-Aldrich) was added to select stably transduced cells over 7–14 days. Stable NSUN2 or YBX1-overexpressing clones (HepAD38-NSUN2 or HepAD38-YBX1) or empty vector control clones (HepAD38-Contrtol) were used for further study.

m⁵C methylated RNA immunoprecipitation sequencing (m⁵C-MeRIP-seq)

m⁵C-MeRIP-seq was performed by SeqHealth Technology Co., Ltd. (Wuhan, China). Briefly, total RNA from HepAD38/Tet⁺ and HepAD38/Tet⁻ cells (n = 3 per group) were extracted using Trizol reagent (Catalog no. 15596026; Invitrogen) and purified to deplete the ribosomal RNA and avoid DNA contamination. After fragmentation and denaturation, RNA was sheared by sonication into approximately 100-nt fragments. The cleaved RNA fragments were split into two parts: one for input and the other enriched with an m⁵C-specific antibody. Eluted m⁵C-containing fragments (IP) and untreated input control fragments were converted to the final cDNA library. At last, paired-end sequencing (PE150) was conducted on an Illumina Novaseq™ 6000 platform (SeqHealth Technology Co., Ltd.).

Bioinformatics analysis

Raw sequencing data were processed using Trim Galore to remove adapter sequences and generate clean reads. The cleaned reads were then aligned to the GRCh37 human genome reference using Hisat2 (Kim et al., 2015). Peak identification and differential methylation analysis were performed using exomePeak2 software (Liang et al., 2024). For each identified peak region, exomePeak2 calculates the log2 fold change (log2FC) of methylation enrichment between the two conditions and provides a corresponding P value to assess statistical significance. The identified m⁵C peaks were visualized using the Integrative Genomics Viewer (IGV). Potential m⁵C methylation sites were predicted using three computational tools including iRNA-m5C (Lv et al., 2020), m5Cpred-XS (Liu Yibo et al., 2022), and RNAm5Cfinder (Li et al., 2018), all with default parameters. Correlations between NSUN2 or YBX1 and PTEN expression in HCC tissues were assessed using data from The Cancer Genome Atlas-Liver Hepatocellular Carcinoma (TCGA-LIHC) database via GEPIA (http://gepia.cancer-pku.cn/detail.php) (assessed on May 23, 2024). NSUN2 mRNA expression was evaluated in healthy controls and across the clinical phases of chronic HBV infection—immunotolerant (IT), immune-active (IA), inactive carrier (IC), and HBeAg-negative hepatitis (ENEG)—using the GEO dataset GSE230397 via https://www.ncbi.nlm.nih.gov/geo/(assessed on August 18, 2024). For tumor comparisons, NSUN2 expression in HBV-related versus non-HBV-related HCC was analyzed using the TCGA-LIHC cohort via https://www.cancer.gov/ccg/research/genome-sequencing/tcga (assessed on August 15, 2024).

RNA interference

The small interfering RNA (siRNA) specifically targeting NSUN2, NSUN6, YBX1, ALYREF, TET1, TET2 or TET3 was designed and chemically synthesized by Shanghai GenePharma Co. A nontargeting siRNA (NC-siRNA) was used as a negative control. Transfections were performed using Lipofectamine™ RNAiMAX reagent (Catalog No.13778150, Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's protocol. Gene silencing effect was confirmed by Western blot analysis at 48 h post transfection. The sequences of the siRNA oligonucleotides used in this study were provided in Supplementary Table S4.

Western blot analysis

Western blot analysis was performed as described previously (Wu et al., 2021). The primary antibodies used included anti-Flag (F1804; Sigma-Aldrich), anti-HBsAg (DMABT-51328MH; Creative-Diagnostics, Shirley, NY, USA), anti-NSUN2 (20854-1-AP; Proteintech Group, Rosemont, IL, USA), anti-NSUN6 (17240-1-AP; Proteintech), anti-YBX1 (20339-1-AP; Proteintech), anti-ALYREF (ab202894; Abcam, Cambridge, MA, USA), anti-HBx (ab39716; Abcam), anti-PTEN (9188S; Cell Signaling Technology, Beverly, MA, USA), anti-TET1 (ab191698; Abcam), anti-TET2 (21207-1-AP; Proteintech), anti-TET3 (27150-1-AP; Proteintech), anti-Hep B Pol (2C8) (sc-81590; Santa Cruz, California, USA) and anti-HBc generously provided by Prof. QuanYuan (Xiamen University) (Chung et al., 2024). Relative immunoblot bands were compared using the prestained protein marker (MP202; Vazyme, Nanjing, China) and Thermo Scientific PageRuler Prestained Protein Ladder (Catalog No. 26616; Thermo Fisher Scientific).

RNA extraction and reverse transcription quantitative PCR (RT‒qPCR)

Total RNA was extracted from cells using TRIzol reagent following the manufacturer's instructions. RNA was reverse-transcribed into complementary DNA (cDNA) using the Evo M-MLV RT Kit with gDNA Clean for qPCR (AG11705; Accurate Biology, Hunan, China). RT-qPCR was performed in duplicate using 2 × SYBR Green Pro Taq HS Premix II (AG11701; Accurate Biology) on a Gentier 96R Fully Automatic Medical PCR Analysis System (Thermo Fisher Scientific). GAPDH was used as an internal control. Relative mRNA levels were quantified using the threshold cycle (2^−ΔΔCT) method. The primers used for amplification are listed in Supplementary Table S5.

m⁵C-MeRIP-qPCR

m⁵C-MeRIP-qPCR was performed using the GenSeq™ m⁵C-MeRIP Kit (GS-ET-003; Cloudseq, Shanghai, China) according to the manufacturer's instructions. Briefly, 25 μL protein A/G magnetic beads supplied with the kit were pre-washed with the kit IP buffers and incubated with 2 μg anti-m⁵C antibody (MAb-081-100; Diagenode, Belgium) or IgG antibody (ab3735; Abcam) for 1 h at room temperature (2 μg antibody per reaction). Next, total RNA (∼200 μg per reaction) was added to the antibody–bead complex and rotated at 4 °C for 3 h. Beads were then washed three times with 200 μL IP buffers, and methylated RNA was eluted and purified using the kit elution and column system. RT-qPCR was performed following MeRIP to quantify m⁵C methylation levels of target genes. Percent-input normalization was applied (Weng et al., 2018), with IgG values serving as negative controls. The m⁵C IP enrichment levels were calculated as the percentage of the IP group relative to the input. These relative m⁵C enrichment values were then normalized to the corresponding control group. The primers used for m⁵C-MeRIP-qPCR are listed in Supplementary Table S6. For each condition, ≥3 biological replicates were analyzed. Antibody source, catalog number (MAb-081-100), antibody: RNA ratio (2 μg antibody per ∼200 μg RNA), incubation times/temperatures, and control procedures are as specified above to ensure reproducibility.

RIP-qPCR

RNA immunoprecipitation (RIP) assays were performed using the PureBinding®RNA Immunoprecipitation Kit (P0102; Geneseed, Guangdong, China) in accordance with the manufacturer's protocols. Briefly, a total of 1 × 10⁷ cells were harvested and lysed in an appropriate volume of complete RIP Lysis Buffer. RNA-binding proteins were immunoprecipitated using either an anti-NSUN2 antibody or an anti-YBX1 antibody alongside normal rabbit IgG (A7016; Beyotime, shanghai, China) as a negative control. The co-precipitated RNAs were then purified and dissolved in RNase-free water. Subsequently, the RNA targets bound to the immunoprecipitated proteins were analyzed using RT-qPCR. The results were normalized to the input for relative quantification. Enrichment levels were calculated as the percentage of the immunoprecipitated (IP) fraction relative to the input. These relative enrichment values were further normalized against the corresponding control group for comparative analysis. The primers used for RIP-qPCR are listed in Supplementary Table S6.

RNA stability assay

The stability of RNA in cells transfected with various plasmids was assessed by treating the cells with actinomycin D (M4881; Abmole, Shanghai, China) at a final concentration of 5 μg/mL (for NSUN2 or PTEN transcript) or 15 μg/mL (for YBX1 transcript) to inhibit transcription. Cells were harvested at specific time points, and total RNA was then extracted. RNA quantities were determined by RT-qPCR. The RNA degradation rate (k) was calculated using the following equation: N_t/N₀ = e^−kt, where t represents the transcription inhibition time, and N_t and N₀ are the RNA quantities at times t and time 0, respectively. The RNA half-life (t_1/2) can be calculated from the degradation rate using the formula: t_1/2 = ln (2)/k.

Cell counting Kit-8 (CCK-8) assay

Cell viability was measured using Cell Counting Kit-8 reagent (CCK-8; Dojindo Laboratories, Japan) according to the manufacturer's instructions. Briefly, cells were seeded into 96-well plates with 2 × 10³ cells per well (triplicate wells per condition) and cultured for the indicated times. The culture medium was then replaced with fresh medium containing CCK-8 reagent (9:1, v/v), and the cells were incubated for 1 h. Absorbance at 450 nm was measured using a microplate reader (Bio-Tek, Winooski, VT, USA). Absorbance values were proportional to viable cell number.

Colony formation assay

Cells were seeded into 6-well plates at a density of 1000 cells per well and cultured in DMEM supplemented with 10% FBS for two weeks. At the end of the incubation period, the cells were fixed with 4% paraformaldehyde and subsequently stained by crystal violet (C0121; Beyotime). Colonies consisting of more than 50 cells were imaged and counted to assess cell proliferation capacity.

5-Ethynyl-2-deoxyuridine (EdU) incorporation assay

Cells were seeded in 12-well plates at 1.5 × 10⁵ cells per well and cultured for 24 h. EdU labeling was then performed by incubating cells with EdU for 2 h at 37 °C using an EdU Cell Proliferation Imaging Kit (KTA2031; Abbkine, Hubei, China), following the manufacturer's protocol. EdU incorporation and nuclear staining were visualized by fluorescence microscopy (Carl Zeiss, Jena, Germany) using excitation/emission wavelengths of 546/565 nm (EdU) and 360/460 nm (Hoechst), respectively. The percentage of EdU-positive cells (proliferating cells) relative to total nuclei was quantified using ImageJ software (NIH, Bethesda, MA, USA).

Wound-healing assay

Cells (5 × 10⁵ per well) were seeded into 6-well plates and grown to approximately 90% confluence. A wound was created by scratching the cell monolayer with a 200 μL micropipette tip. After washing the wells with PBS to remove detached cells, fresh culture medium was added, and the plates were incubated at 37 °C in a 5% CO₂ atmosphere. Wound closure was photographed at 0 and 48 h using a microscope (Carl Zeiss, Jena, Germany). Cell migration into the wound space was quantified using ImageJ software. The percentage of cell migration was calculated using the formula: migration (%) = [(area at 0 h – area at 48 h)/area at 0 h] × 100%.

Migration and invasion assay

The migration and invasion assays were conducted using 24-well Transwell chambers (Catalog no. 3422; 8-μm pore size; BD Biosciences, San Jose, CA, USA). For the invasion assay, the upper chambers were precoated with 60 μL of 1: 10 diluted LDEV-Free Ceturegel™ Matrix (Catalog no. 40183; Yeasen, Shanghai, China). For both assays, 1 × 10⁵ cells suspended in 300 μL of serum-free DMEM were added to the upper chamber, while the lower chamber was filled with 700 μL of DMEM containing 10% FBS as a chemoattractant. After 48 h, cells on the upper surface of the membrane were gently removed using a cotton swab. Cells that migrated or invaded through to the bottom surface of the membrane were fixed and stained by crystal violet. The number of cells was counted in five randomly selected fields under a microscope equipped with a Qimaging Micropublisher 5.0 RTV camera (Olympus Corporation, Tokyo, Japan).

Luciferase reporter assay

HepAD38 cells were individually transfected with luciferase reporter vectors including pmirGLO-PTEN-WT and cytosine-to-guanine mutants. 48 h post-transfection, the cells were lysed. Thirty micrograms of protein from each sample were used to measure intracellular luciferase activity using the Bright-Glo™ Luciferase Assay System (E2610; Promega), according to the manufacturer's instructions (Yang et al., 2021). Luminescence was measured with a luminometer (Berthold Technologies, Oak Ridge, TN, USA). Relative luciferase activity was calculated by normalizing the luminescence of each sample to that of HepAD38/Tet⁺ cells transfected with pmirGLO-PTEN-WT plasmid, which was set to 1. For the promoter activities reporter assays, HepAD38 cells were seeded into 24-well plates and transfected with 500 ng of either the pGL4.10-NSUN2 or pGL4.10-YBX1. Transfected cells were cultured for 48 h before being harvested for luciferase activity detection using the Bright-Glo™ Luciferase Assay System, following the manufacturer's instructions. Relative luciferase activity was calculated by comparing the luminescence of each sample to that of HepAD38/Tet⁺ cells transfected with the pGL4.10-NSUN2 or pGL4.10-YBX1, which was assigned an arbitrary value of 1. Each transfection was replicated in duplicate to ensure reliability, and the data were reported as the mean ± SD (standard deviation) based on three independent experiments.

Statistical analysis

Statistical analyses were conducted using SPSS software (version 22.0; SPSS Inc., Chicago, IL, USA) and GraphPad Prism (version 8.3.0; GraphPad Software, San Diego, USA). Comparisons between two groups were evaluated using a t-test, while comparisons among multiple groups were analyzed using one-way ANOVA followed by the LSD-t test for post hoc analysis. Data are presented as means ± standard deviation (SD) from triplicate experiments. A p-value of less than 0.05 was considered statistically significant. All experiments were performed at least three times, and representative results are displayed.

Data availability

All data generated in this study are available upon request from the corresponding author.

Ethics statement

This study was approved by the Ethics Committee of Mengchao Hepatobiliary Hospital of Fujian Medical University (Approval No. 2021_100_06) and performed under the guidelines of the Declaration of Helsinki. Written informed consent was obtained from each participant.

Author contributions

Mengxian Lin: Writing-Original Draft, Methodology, Validation, Formal analysis. Kunqi Chen: Methodology, Software, Validation, Formal analysis. Hang Li: Methodology, Software, Validation, Data curation. Xinjian Lin: Writing-Original Draft, Writing-Review and editing, Formal analysis, Investigation. Shuxiang Wu: Funding acquisition, Resources, Project administration, supervision. Xu Lin: Conceptualization, Project administration, Funding acquisition, Supervision. All the authors have read and approved the submitted version of manuscript.

Conflict of interest

The authors declare that they have no conflict of interest.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary Figure S1.

HBV-mediated suppression of PTEN via reduced PTEN mRNA m5C modification is independent of TET1/2/3. (A) Western blot analysis of TET1, TET2, and TET3 protein levels in HepAD38/Tet⁺ and HepAD38/Tet⁻ cells. (B) Western blot analysis of PTEN protein levels in HepG2 and Huh7 cells following knockdown of TET1, TET2, or TET3 with specific siRNAs. (C) m⁵C-MeRIP–qPCR analysis showing the effects of TET2 or TET3 knockdown on m⁵C modification levels of PTEN mRNA in HepG2 cells. Data are presented as mean ± SD from three independent experiments. ns, not significant; ∗∗p < 0.01; ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Supplementary Figure S2.

HBV is associated with reduced NSUN2 expression. (A) NSUN2 mRNA levels in human liver samples across chronic HBV disease phases—immunotolerant (IT), immune-active (IA), inactive carrier (IC), and HBeAg-negative hepatitis (ENEG)—compared with healthy controls (HC) (GEO: GSE230397). (B) NSUN2 mRNA expression in HBV-related HCC (HBV-HCC) versus non-HBV-related HCC from TCGA-LIHC. (C) NSUN2 and HBx protein abundance in eight freshly resected HBV-positive HCC specimens (left, representative immunoblots; right, Pearson correlation of densitometric values). ns, not significant; ∗p < 0.05; ∗∗p < 0.01.