SMYD2 epigenetically activates BMP4/SMAD1/5/8/ID3 axis to enhance cancer stem cell properties and drive sorafenib resistance in hepatocellular carcinoma

Highlights

• •SMYD2 is upregulated in sorafenib-resistant HCC xenografts models and cells.
• •SMYD2 epigenetically activates BMP4 genes via the maintenance of H3K4me1 and H3K36me2 modification in sorafenib-resistant HCC cells.
• •Deletion of BMP4 suppressed HCC stemness, inhibited the activation of SMAD1/5/8, and reduced the expression of ID3 gene.
• •SMYD2 promotes sorafenib-resistant HCC cell metastasis by activating BMP4/R-SMADs/ID3 axis in vivo.
• •HCC patients with positive co-expression of SMYD2/BMP4 or SMYD2/ID3 or SMYD2/BMP4/ID3 exhibited the worst prognosis

Abstract

Background

Drug resistance prominently hampers the effects of sorafenib in hepatocellular carcinoma (HCC). Epigenetics play important roles in drug resistance. However, the contributions of SET And MYND Domain Containing 2 (SMYD2) to sorafenib resistance in HCC remain unknown. This study is aimed at elucidating the role and mechanism of SMYD2 in sorafenib resistance of HCC.

Methods

Using our well-established sorafenib-resistant hepatocellular carcinoma (HCC) cell lines and xenograft mouse models, we evaluated SMYD2 expression levels. To investigate the biological functions of SMYD2, we conducted a series of functional assays in vitro and in vivo. Transcriptomic profiling via RNA sequencing (RNA-seq) was performed to identify downstream targets of SMYD2. Additionally, chromatin immunoprecipitation (ChIP) assays were employed to elucidate the molecular mechanism. Correlating SMYD2 and target gene expression patterns with clinical outcomes in HCC patients was investigated.

Results

SMYD2 expression was significantly elevated in sorafenib-resistant HCC cells compared with parental cells. Knockdown or overexpression of SMYD2 substantially inhibited or enhanced, respectively, HCC stemness and sorafenib resistance. Mechanistically, SMYD2 promoted BMP4 expression via the maintenance of mono-methylation of histone 3 lysine 4 (H3K4me1) and di-methylation of histone 3 lysine 36 (H3K36me2) modification of its promoter. Meanwhile, knockdown or inhibition of BMP4 suppressed the stemness of sorafenib-resistant cells, inhibited the activation of SMAD1/5/8 (R-SMADs), and decreased the expression of inhibitor Of DNA binding 3 (ID3) gene. Moreover, BMP4 addition or ID3 reconstruction can partly reverse the effect caused by repression of SMYD2 or BMP4. HCC patients with positive co-expression of SMYD2/BMP4 or SMYD2/ID3 or SMYD2/BMP4/ID3 exhibited the worst prognosis.

Conclusions

Our study reveals that SMYD2 is an important epigenetic mediator that activates BMP4/R-SMADs/ID3 axis, leading to enhanced stemness and sorafenib resistance. Thus, SMYD2 might represent a potential biomarker and future epigenetic therapeutic target for sorafenib resistance of HCC.

Introduction

Hepatocellular carcinoma (HCC) is the second most common cause of cancer mortality worldwide [1]. Unfortunately, patients with HCC are usually diagnosed at advanced stages [2]. Only three first-line molecular targeted drugs have been approved for advanced HCC, including sorafenib [3,4]. Nonetheless, the clinical benefits of these drugs have been limited, sorafenib can prolong overall survival (OS) by only 2.3–3 months, due to therapeutic resistance and relapse [3,4]. Thus, precisely delineating the complex mechanisms underlying drug resistance in HCC and accurately identifying more superior therapeutic targets are of utmost importance for establishing more effective treatment strategies.

Alterations of epigenetic regulators are associated with liver cancer stem cells (CSCs), epithelial-mesenchymal transition (EMT), and chemoresistance and determine the outcome of HCC [5]. What’s more, the activity of both established and investigational epigenetic therapies within well-defined clinical contexts has provided evidence that this strategy can be highly effective in treating relevant diseases [6]. Therefore, a deep understanding of epigenetic mechanisms needs to be achieved to develop better therapies.

SET and MYND domain containing 2 (SMYD2) is one of the most extensively studied lysine methyltransferases, which catalyzes the lysine methylation of histone 3 lysine 36 (H3K36), histone 3 lysine 4 (H3K4) and histone 4 lysine 20 (H4K20), as well as non-histone proteins such as tumor protein 53 (TP53), protein phosphatase 1 (RB1), heat shock protein 90 (HSP90), estrogen receptor α (ERα), phosphatase and tensin homolog (PTEN), enhancer of zeste 2 polycomb repressive complex 2 subunit (EZH2) and poly(ADP-ribose) polymerase 1 (PARP1) [[7], [8], [9], [10], [11], [12], [13], [14], [15], [16]]. Since the details of the molecular basis of the catalytic ability of SMYD2 were elucidated, several small-molecule inhibitors targeting this pathway have been developed, serving as potential candidates for new anti-cancer drugs [[17], [18], [19]]. However, the mechanism of SMYD2 in drug resistance remain unclear.

Long-term exposure of HCC cells to molecular targeted drugs often leads to resistance, which has been thought to be driven by CSCs [20]. It is CSCs that have the ability of self-renew, show therapeutic resistance, and give rise to relatively differentiated cells. In HCC, several markers, such as epithelial cell adhesion molecule (EpCAM), THY1 (CD90), cluster of differentiation 24 (CD24), Prominin 1 (PROM1, CD133), and NANOG, in addition to a side population fraction, have been identified for the enrichment of liver CSCs, which were distinctive for their high clonogenicity in vitro and high tumorigenicity in vivo [[21], [22], [23], [24], [25]]. Nonetheless, the mechanisms of SMYD2 in liver CSCs and sorafenib resistance remain unclear.

To identify novel and key targets involved in sorafenib resistance, we established sorafenib-resistant HCC cells and xenograft models. Here, we found that SMYD2 was overexpressed in both in vivo and in vitro sorafenib-resistant models, and overexpressed SMYD2 confers sorafenib resistance to hepatoma cells through enhanced stem-cell and metastatic properties. Furthermore, molecular and pharmacological inhibition of SMYD2 decreased the expression of Bone morphogenetic protein 4 (BMP4) gene by decreasing H3K36me2 and H3K4me levels, and consequently weakening self-renewal and metastasis both in vitro and in vivo. What’s more, SMYD2 activates BMP4/R-SMADs/ID3 axis, leading to enhanced stemness and sorafenib resistance. HCC patients with positive co-expression of SMYD2/BMP4 or SMYD2/ID3 or SMYD2/BMP4/ID3 exhibited the worst prognosis. Taken together, these results indicate that targeting SMYD2 represents a promising strategy for overcoming sorafenib resistance in HCC.

Materials and methods

Cell lines and cell culture

Human hepatoma cell lines Huh7 was obtained directly from Shanghai Cell Bank of Chinese Academy of Sciences (Shanghai, China), and SMMC-7721 was purchased from Procell Life Science & Technology Co., Ltd (Wuhan, China). Cells were maintained in Dulbecco’s modified eagle medium (DMEM) supplemented with 10 % fetal bovine serum (FBS) and 1 % penicillin-streptomycin. All cultures were maintained in a humidified incubator at 37 °C and 5 % CO₂. The cell lines have been characterized at the cell bank by DNA fingerprinting analysis using short tandem repeat markers. All cell lines were placed under cryostage after they were obtained from the cell bank and used within 6 months of thawing fresh vials.

Establishment of sorafenib-resistant Huh7 and SMMC-7721 cells

Sorafenib was dissolved in dimethyl sulfoxide. Establishment of sorafenib-resistant cells in vitro was carried. Briefly, Cells were plated in 6-well plates. When cells reached 60 % confluence, they were treated with the appropriate dose of inhibitor at different times. After treatment, cells were collected for further experiments and analyses.

RNA sequencing

Total RNA of Huh7-Res of shControl and SMYD2 knockdown (shSMYD2-#1) groups were extracted using TRIzol Reagent (Life Technologies) according to manufacturer's protocol. The quality of total RNA was checked by Agilent 2100 bioanalyzer (Agilent Technologies Inc.) to have OD260/280 ratio of between 1.8 and 2.0 and RNA integrity number value higher than 8.0. The RNA samples that met the quality assessment were then subjected to Illumina Solexa sequencing using Hiseq 1500 sequencer (Illumina) for performing HiSeq sequencing run (pair-end sequencing of 101 bp). Each sample had an average throughput of 10.8 Gb and a total throughput of 21.5 Gb. An average of 94 % of the bases achieved a quality score of Q30 where Q30 denotes the accuracy of a base call to be 99.9 %. Expression estimation and tests for differential expression were processed by Cufflinks v2.1.1. All data were expressed as fragments per kilobase of exon per million fragments mapped values and fold changes in transcript levels relative to the shControl group from Huh7-Res. RNA-sequencing data are available publicly.

Mice xenograft model

All animal experiments were approved by the Institutional Animal Care and Use Committee of Zhejiang Chinese Medical University (Zhejiang, China) and carried out in accordance with the approved guideline“code of practice: animal experiments in cancer research” (Netherlands Inspectorate for Health Protection, Commodities and Veterinary Public Health, 1999).

Sorafenib-resistant subcutaneous tumors were generated as described previously. Briefly, 1 × 10⁷ Huh7 cells in 200 µl phosphate buffered saline were injected subcutaneously into 3-4 week-old male nude mice. After two weeks, mice were randomly allocated into groups and treated with sorafenib (60 mg/kg/intraperitoneal injection (i.p.), once every other day) for two months. On day 60 after the start of treatment, tumors were removed.

To establish in vivo tumor formation model, 100 to 10,000 Huh7-Res or 7721-Res cells with shControl or shSMYD2-#1, were mixed with Matrigel, and were injected subcutaneously into the flanks of three-four week-old male nude mice. Tumors grew for approximately 4 weeks and were removed.

1 × 10⁷ Huh7-Res or Huh7-Res shSMYD2-#1 cells in 200 µl phosphate buffered saline were injected subcutaneously in 3-4 week-old male nude mice. After two weeks, mice with Huh7-Res cells shControl or shSMYD2-#1 were randomly allocated into 4 groups and orally treated with DMSO or sorafenib (60 mg/kg/day) for 4 weeks.

Tumors were measured twice weekly and volumes. Tumor size was measured with digital calipers and calculated based on the following formula: length $\times$ (width)² $\times \pi$/6. Tumors sections from subcutaneous tumor xenografted male nude mice were H&E stained, immunohistochemically analyzed, and evaluated.

Construction of plasmids, shRNAs, transfection, and lentivirus-mediated gene knockdown

For expression of wide type SMYD2, the cDNA was amplified by PCR and ligated into the correct reading frames of pcDNA3.1-flag vector containing FLAG coding sequences. To generate catalytic dead mutant SMYD2-Y240F expression vectors, SMYD2 cDNA was mutated by the QuikChange® II site-directed mutagenesis kit (Agilent Technologies) before the construction. All plasmid constructs were confirmed by DNA sequencing. Plasmid transfections were performed using Lipofectamine 2000 (Invitrogen, USA). The shRNA against SMYD2 gene shSMYD2s and corresponding nonspecific shRNA (Sigma, USA) were used for RNA interference.

Western blot analysis

Proteins extracted for western blotting assay were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to polyvinylidene fluoride membranes, and blotted with primary and secondary antibodies. Densitometry was performed using ImageJ 1.51h v software (http://rsb.info.nih.gov/ij/).

Immunofluorescence assay

Regarding the immunofluorescence assay, cells were fixed, and then blocked with 1 % bovine serum albumin-phosphate-buffered saline prior to incubation with primary and secondary antibodies.

RNA extraction and quantitative real-time PCR (qRT-PCR)

Total RNA was isolated using TRIzol reagent (Invitrogen, NY) and reverse transcription was performed with 500ng of RNA using PrimeScript™ RT Master Mix (Takara, Japan), according to the manufacturer’s instructions. qPCR was performed with FastStart Universal SYBR® Green Master (Roche, USA) on an Applied Biosystems 7500 Real Time PCR System supplied with analytical software (Applied Biosystems, USA). The average of the technical replicated was normalized to GAPDH levels using the comparative CT method (2^-△△CT). Average and standard deviations of at least 3 experiments are shown in the figures.

Annexin V ⁄ Propidium iodide staining, cell proliferation assay, and sphere formation assay

Annexin V⁄ propidium iodide (PI) staining and flow cytometry analysis were carried out using the Annexin V-FITC Apoptosis Detection Kit (BD Biosciences, USA) according to the manufacturers’ instructions. Cell proliferation assay, colony formation assay and sphere formation assay were carried according to the manufacturers’ instructions.

Chromatin immunoprecipitation (ChIP) and Re-ChIP

Chromatin immunoprecipitation (ChIP) or Re-ChIP was carried out with the ChIP Assay Kit (Millipore, USA) according to the manufacturer’s instructions. The percentage of the bound DNA was quantified against the original DNA input by PCR analysis.

Immunohistochemistry

The tissues were stained with anti-SMYD2 and anti-Ki67 antibodies using an immunohistochemistry (IHC) staining kit (Bios Biological Technology Co., Ltd).

Data availability

In this study, we utilized liver cancer data from the TCGA database, which were downloaded using the R package TCGAbiolinks and analyzed with R version 4.2.3. The TCGA database is a public resource, accessible to anyone for scientific research purposes.

Statistical analysis

All data were analyzed using GraphPad Prism 7.0. All statistical details of the experiments, including exact values of n and what n represents, are mentioned in the figure legends. Statistical significance was determined using the Student’s t-test (when comparing two experimental groups), one-way analysis of variance (ANOVA) or two-way ANOVA as appropriate (when comparing more than two experimental groups). p-values < 0.05 were considered statistically significant. Data were expressed as mean ± standard deviation (SD). ns, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Results

SMYD2 is aberrantly upregulated in sorafenib-resistant tumors and cells

The xenografts model more closely resembles the clinical features of HCC patients and retains the characteristics significantly associated with drug response. To investigate the possible function of SMYD2 in sorafenib-resistant HCC, sorafenib-resistant xenografts model was established using Huh7 cells with a high-dose sorafenib as described previously [26,27]. After tumors reached 5 mm in diameter, mice were orally administered with sorafenib for two months, we found that the treatment led to tumor inhibition among the xenograft (Fig. 1A). However, over time, tumor growth in the sorafenib-treated group ceased to be inhibited and a rapid growth phase occurred, with distant metastases, including liver and lung metastases, suggesting that the xenograft model was resistant to sorafenib (Fig. 1B and 1C). The SMYD2 protein level was detected by western blot and immunohistochemistry analysis in sorafenib-resistant tumors, and it was found that compared with DMSO-treated tumors, the SMYD2 protein level in sorafenib-resistant tumors was significantly increased (Fig. 1D and 1E). Therefore, the significant overexpression of SMYD2 protein was clearly observed in sorafenib-resistant tumors.

Fig. 1.

SMYD2 is overexpressed in sorafenib-resistant tumors and cells. (A-B) Effects of sorafenib treatment on tumor growth in nude mice. (A) Final tumor masses are represented (mean ± SD, n = 4-6/group). (B) Tumor volume was monitored with indicated time (mean ± SD, n = 4-6/group). (C) Representative pictures of livers and lungs with metastatic tumors stained for H&E. (D) Up, western blot analysis of SMYD2 and GAPDH in DMSO-treated and Sorafenib-resistant tumors. Down, Quantification of relative protein expression of SMYD2/GAPDH is shown. (E) Left: IHC staining for Ki67 and SMYD2 expression. Right: Quantification of mean gray value of Ki67 and SMYD2 is shown. (F) Dose-response curve and IC₅₀ of sorafenib-resistant cells treated with sorafenib for 48hr. (G) Left: Apoptosis ability was checked by Annexin V staining and IP staining in parental and sorafenib-resistant cells. Right: Quantification of percentage of apoptosis cells, dead cells, and live cells. (H) Up, western blot analysis of SMYD2 and GAPDH in parental and sorafenib-resistant cells. Down, Quantification of relative protein expression of SMYD2/GAPDH is shown. (I) Left: Immunofluorescence staining for SMYD2 expression in parental and sorafenib-resistant cells. Right: Quantification of mean gray value of SMYD2 protein is shown.

For in vitro selection, through the continuous administration of sorafenib with its concentration gradually increasing over a period of 4 months, sorafenib-resistant HCC cells were successfully developed in vitro by using Huh7 and SMMC-7721 cells. To confirm the successful establishment of sorafenib-resistant cells, 50 % inhibitory concentration (IC₅₀) was measured by cell counting kit-8 (CCK-8) assay, and found that the resistant cells showed higher IC₅₀ values than parental cells (Fig. 1F). Besides, by using Annexin V staining, it was found that, at the two different doses, sorafenib-resistant cells showed less apoptosis and more viability compared to parental cells, which indicated their lower sensitivity to sorafenib-induced apoptosis (Fig. 1G). We next examined the effect on SMYD2 expression. The result of western blot experiment revealed that the expression of SMYD2 increased significantly in sorafenib-resistant HCC cells (Fig. 1H). Immunofluorescence experiments further confirmed that SMYD2 was overexpressed in sorafenib-resistant HCC cells, and that its distribution covered the nucleus as well as the cytoplasm (Fig. 1I).

Taken together, these findings suggest that SMYD2 is up-regulated in sorafenib-resistant HCC cells and tumors.

SMYD2 is required for the resistance to sorafenib in HCC

To test if the level of SMYD2 expression could affect the sensitivity of sorafenib-resistant cells, we carried out loss- and gain-of-function analysis of SMYD2 in vitro. Using lentivirus-mediated-short interfering RNA (shRNA) against SMYD2 (shSMYD2), we firstly achieved the stable knockdown of SMYD2 in sorafenib-resistant HCC cells. Two different shSMYD2, labeled shSMYD2-#1 and shSMYD2-#2, effectively suppressed the expression of SMYD2 protein and mRNA levels (Fig. 2A). Since knockdown of SMYD2 was more prominent with shSMYD2-#1 than shSMYD2-#2, we used shSMYD2-#1 for most of subsequent experiments. We then analyzed whether these cells were sensitive to the inhibition of sorafenib. Measurement of IC₅₀ revealed that shSMYD2 cells were more sensitive than shControl cells to the inhibition of sorafenib in two sorafenib-resistant cells (Fig. 2B). We next treated the shControl and shSMYD2 cells with 10nM sorafenib for 48h to assess apoptosis by Annexin V/PI staining and Flow cytometric analysis. Increased percentage of apoptosis/dead cells and decreased percentage of live cells were observed in both sorafenib-resistant cells with SMYD2 knockdown (Fig. 2C). We next conducted gain-of-function analysis of SMYD2 in vitro in sorafenib-resistant cells with SMYD2 knockdown. We constructed wild-type SMYD2 (SMYD2) in the pcDNA3.1-flag vector, and transferred wild-type SMYD2 to sorafenib-resistant cells with SMYD2 knockdown, and subsequently investigated the effect on sorafenib sensitivity (Fig. 2D). In sorafenib-resistant cells with SMYD2 knockdown, the reconstitution with wide-type SMYD2 significantly restored the resistance to sorafenib by measurement of IC₅₀ and Annexin V/PI staining (Fig. 2E and 2F).

Fig. 2.

Levels of SMYD2 in sorafenib-resistant cells are associated with sensitivity of sorafenib. (A) SMYD2 expression in sorafenib-resistant cells after stable SMYD2 knockdown was assessed by western blot and RT-PCR. (D) Sorafenib-resistant cells were treated with 10μM Sorafenib or 0.12μM (IC₅₀) AZ505 for 48 hours. The SMYD2 expression was assessed by western blot. (G) SMYD2 expression was first silenced by the shSMYD2-#1-expressing lentivirus in sorafenib-resistant cells, and these cells were then transferred with wide-type SMYD2 (pcDNA3.1-flag-SMYD2-widetype, short for “SMYD2”), or the corresponding control vector (pcDNA3.1-flag). SMYD2 expression was assessed by western blot. (B, E, and H) Dose-response curve and IC₅₀ of sorafenib-resistant cells with SMYD2 knockdown (B) or SMYD2 knockdown and exogenously expressed wide-type SMYD2 (E) or SMYD2 inhibition (H) treated with sorafenib for 48hr. (C, F, and I) Up: Apoptosis ability was checked by Annexin V staining and IP staining in sorafenib-resistant cells with SMYD2 knockdown (C) or SMYD2 knockdown and exogenously expressed wide-type SMYD2 (F) or SMYD2 inhibition (I). Down: Quantification of percentage of apoptosis cells, dead cells, and live cells.

Similar results were obtained in sorafenib-resistant HCC cells with SMYD2 inhibition. AZ505 is a highly selective inhibitor of SMYD2 with IC₅₀ of 0.12μM (>600-fold selectivity) [18]. Pharmacological inhibition of SMYD2 with AZ505 also made both sorafenib-resistant cells more sensitive to sorafenib, as measured by IC₅₀ and Annexin V/PI staining (Fig. 2H and 2I). Besides, AZ505 did not cause the SYMD2 reduction in parental cells (data not shown), while the protein levels of SMYD2 were reduced by the treatment of AZ505 in both sorafenib-resistant cells, potentially due to a compensatory effect that is blocked in resistant cells (Fig. 2G).

These results not only imply the possible involvement of SMYD2 in resistant cells response to sorafenib, but also indicate that levels of SMYD2 in resistant cells may inversely be correlated with sensitivity to sorafenib.

Transcriptional analysis revealed that SMYD2 is involved in signaling pathways regulating pluripotency of stem cells

To elucidate the mechanisms underlying the regulation of SMYD2 on sorafenib resistance in HCC, we performed RNA-seq to assay the transcriptomes of SMYD2-knockdown Huh7-Res cells (Fig. 3A). The knockdown of SMYD2 resulted in 1067 genes upregulation and 904 genes downregulation (Fig. S1A and 3C). For a deeper understanding of these genes functions, gene set enrichment analysis (GSEA) was performed, and found that SMYD2-related genes were enriched in signaling pathways regulating pluripotency of stem cells (Fig. 3B and 3D). The significantly differentially expressed genes in signaling pathways regulating pluripotency of stem cells between the control group and SMYD2-knockdown group were shown, and reverse-transcription quantitative Polymerase Chain Reaction (RT-qPCR) was carried out to confirm these results of down- or up-regulated by SMYD2 knockdown in Huh7-Res (Fig. 3C, 3F, and 3G). These data indicate SMYD2 is involved in signaling pathways regulating pluripotency of stem cells.

Fig. 3.

SMYD2 is involved in signaling pathways regulating pluripotency of stem cells. (A) A heatmap depicting down- and up-regulated genes upon SMYD2 depletion in Huh7-Res cells. (B) KEGG pathway analysis of significant genes in Huh7-Res cells with SMYD2 knockdown. Dot plots of top 15 most significant affected pathway were shown. (C) Volcano Plot showing significant genes in Huh7-Res cells with SMYD2 knockdown. The plots labeled with gene names belonging to signaling pathway regulating pluripotency of stem cells. (D) Gene set enrichment analysis (GSEA) of the gene sets in signaling pathway regulating pluripotency of stem cells in the expression profiles of Huh7-res cells expressing SMYD2 shRNA. (E) A Venn diagram comparing depicted 30 genes as potential direct targets of SMYD2 in Huh7-Res cells based on RNA-seq data and TCGA data. (F) The heatmap shows the alterations in genes in the Signal pathway regulating pluripotency of stem cells. (G) qRT-PCR analysis of representative genes of signaling pathway regulating pluripotency of stem cells in Huh7-Res cells upon SMYD2 depletion. Data shown represent the means (±SD) of biological triplicates.

Identification of BMP4 as a target of SMYD2 in sorafenib-resistant cells

To identify putative SMYD2 target genes, we subsequently analyzed the Human Cancer Genome Atlas (TCGA) database to assess the correlation between SMYD2 expression and these differentially expressed genes in signaling pathways regulating pluripotency of stem cells by SMYD2 knockdown in Huh7-Res cells (Fig. 3E and S1B). Bone morphogenetic protein 4 (BMP4), fibroblast growth factor 2 (FGF2), and phosphoinositide-3-kinase regulatory subunit 1 (PIK3R1) were the genes significantly down-regulated by SMYD2 knockdown in Huh7-Res cells, and they were positively correlated with the expression of SMYD2 mRNA in TCGA (Fig. 3E and S1B). BMP4 gene is the highest correlation coefficient with SMYD2 expression in TCGA (Fig. S1B). We hypothesized that BMP4 gene might be a target of SMYD2 in sorafenib resistant cells.

BMP4, a member of the transforming growth factor-β (TGF-β) superfamily, is known to be involved in the regulation of cell proliferation, differentiation, apoptosis, angiogenesis, migration, invasion, EMT, and tumor chemoresistance [[28], [29], [30], [31], [32], [33], [34]]. BMP4 was highly expressed in metastatic HCC cells and is closely associated with poor prognosis of patients [28,31,33]. Endogenous BMP4 or low-dose BMP4 is required for liver CSCs maintenance by up-regulating CD133 expression in HCC [34]. To further analyze the correlation SMYD2 expression with BMP4 in sorafenib-resistant context, we firstly performed western blot and RT-qPCR assay to check the expression of BMP4 in sorafenib-resistant models. We found that BMP4 was overexpressed in sorafenib-resistant tumors and cells (Fig. 4A-4B). What’s more, loss-of-function studies showed that knockdown of SMYD2 expression led to downregulation of both BMP4 protein and mRNA in sorafenib-resistant cells. Interestingly, wild-type SMYD2 but not mutant SMYD2 could restore the reduction of BMP4 expression due to SMYD2 knockdown. (Fig. 4C-4D). Inhibition of SMYD2 by AZ505 alone or in combination with sorafenib also decreased the expression of BMP4 in sorafenib-resistant cells (Fig. 4E). what’s more, BMP4 protein was positively correlated with SMYD2 protein in a variety of treated cells and tumors (Fig. 4A-4E).

Fig. 4.

BMP4 is a target of SMYD2 in sorafenib-resistant cells. (A-E) Left up/Left, western blot analysis of BMP4 and GAPDH in the various cells or tumors indicated. Left down/Middle, Quantification of relative protein expression of BMP4/GAPDH or RT-qPCR analysis of BMP4 gene is shown. Data shown represent the means (±SD) of biological triplicates. Right, the correlation analysis of SMYD2 protein and BMP4 protein in various cells or tumors. (F) ChIP-qPCR assay was indicated the recruitment of SMYD2 and the enrichment of H3K4me3, H3K4me and H3K36me2 on BMP4 loci at Site1-4 in parental and sorafenib-resistant cells. (G) ChIP/re-ChIP experiments were performed in sorafenib-resistant cells with the indicated antibodies at Site2. (H-J) ChIP-qPCR assay was indicated the recruitment of SMYD2 or Flag-SMYD2 or Flag-SMYD2-M, and RNA POLII and the enrichment of H3K4me3, H3K4me and H3K36me2 on BMP4 loci at Site2.

To further evaluate the mechanism of SMYD2 in the regulation of BMP4 gene, chromatin immunoprecipitation (ChIP) assay was performed to detect the recruitment of SMYD2 and the modification of histone H3K4me3, H3K4me1, and H3K36me2 in the promoter regions of BMP4 gene in sorafenib-resistant cells. The results from ChIP followed by high-throughput DNA sequence (ChIP-seq) for H3K4me1 or H3K4me3 modifications of the BMP4 gene on the UCSC webpage were obtained, and four loci from the transcription start site to -2000 bp of BMP4 gene were selected and labeled Site1-4 (Fig. S2A). Compared with parental cells, in sorafenib-resistant cells, the recruitment of SMYD2 and the modifications of histone H3K4me3, H3K4me1, and H3K36me2 on the four sites were significantly increased, except for H3K4me1 in Site4 (Fig. 4F). Site2 was selected as a representative due to its specific characteristics for further research. To further identify the histone substrates of SMYD2 at the promoter of BMP4 gene in sorafenib-resistant cells, sequential ChIP/re-ChIP experiments were performed (Fig. 4G). The results showed that in precipitates, the BMP4 promoter which was immunoprecipitated with the antibody against SMYD2, could be re-immunoprecipitated with the antibody against H3K36me2 and H3K4me1, but not H3K4me3 (Fig. 4G). This suggests that SMYD2 plays a crucial role in the transcription of BMP4 gene via promoting H3K4me1 and H3K36me2 rather than H3K4me3 on the promoter of this gene (Fig. 4G). Next, this was further demonstrated by the loss- and gain-of-function experiments. It was found that the recruitment of SMYD2 and RNA polymerase II in the promoter regions of BMP4 gene was reduced. Moreover, histone H3K4Me3, H3K4me1, and H3K36me2 were also significantly reduced by the knockdown or inhibition of SMYD2 (Fig. 4H and 4J). Secondly, we transferred wild-type SMYD2 and enzyme-mutated SMYD2 into sorafenib-resistant cells with SMYD2 knockdown, and then performed ChIP assay with flag antibody. Subsequently, we found that both wild-type and mutant SMYD2 can bind to the promoter regions of BMP4 gene (Fig. 4I). However, compared with the mutant SMYD2, the recruitment of RNA polymerase II, histone H3K4me1 as well as H3K36me2 at Site2 was significantly increased in cells expressing wild-type SMYD2 (Fig. 4I). These results showed that SMYD2 regulated the expression of BMP4 via its histone methyltransferase activity.

SMYD2 is essential for the maintenance of liver CSC self-renewal in sorafenib-resistant cells

Before determining the role of SMYD2 with the signaling pathways regulating pluripotency of stem cells, we first examined the self-renewal of sorafenib-resistant cells. It has been reported that pluripotent transcription factors (NANOG, CD133, and OCT4), which belong to the signaling pathways regulating pluripotency of stem cells, play important roles in liver CSCs [20,21,32,34]. We performed western blot assay to analyze the expression of these factors in sorafenib-resistant cells, and found that the expression of these factors was increased as previously reported (Fig. 5A) [26]. Sphere formation assay was used to assess the self-renewal of sorafenib-resistant cells, and we found sorafenib-resistant cells exhibited enhanced self-renewal (Fig. 5B). Limiting dilution analysis showed that CSCs were significantly enriched in sorafenib-resistant cells compared with parental cells. (Fig. 5C).

Fig. 5.

SMYD2 is essential for the maintenance for liver CSC self-renewal and tumorigenicity in sorafenib-resistant cells. (A, D, G, and I) Left, western blot analysis of CD133, NANOG, OCT4A, and GAPDH in the sorafenib-resistant cells with indicated treatment. Right, Quantification of relative protein expression is shown. (B, E, H, and J) Spheroid formation assay was performed to assess self-renewal ability in vitro. (C and G) Effects on the tumor-forming frequency of parental or sorafenib-resistant cells (C) or sorafenib-resistant cells with shControl or shSMYD2 (G) in vivo, as determined by limiting dilution assays. CSC frequency was calculated using the ELDA (http://bioinf.wehi.edu.au/software/elda).

We next examined the self-renewal of sorafenib-resistant cells upon SMYD2 knockdown and rescue experiments. We observed that knockdown SMYD2 decreased the expression of pluripotent transcription factors in sorafenib-resistant cells, while wild-type SMYD2 recovered the expression of these factors, but not mutant SMYD2 (Fig. 5D and 5G). We next test if SMYD2 plays a critical role in liver CSCs self-renew. It was found that knockdown of SMYD2 decreased sphere formation in sorafenib-resistant cells, while re-expression of SMYD2 increased sphere formation, but not mutant SMYD2 (Fig. 5E and 5H). In keeping with these findings, limiting dilution analysis showed that CSCs were significantly more enriched in control cell populations of sorafenib-resistant cells than in their SMYD2 knockdown counterparts. (Fig. 5F). Furthermore, we found that the SMYD2 inhibitor AZ505 suppressed the expression of these stem cell transcription factors (Fig. 5I). What’s more, inhibition of SMYD2 also decreased sphere formation in sorafenib-resistant cells (Fig. 5J). Collectively, SMYD2 could promote the self-renewal of liver CSCs in vitro in sorafenib-resistant cells.

BMP4 acts as a downstream effector of SMYD2 to mediate the maintenance of self-renewal of liver CSCs

To evaluate the role of BMP4 in biological function of SMYD2 in sorafenib-resistant cells, BMP4 was knocked down and verified by western blot and qPCR in sorafenib-resistant cells (Fig. 6A and 6B). Measurement of IC₅₀ revealed that shBMP4 cells were more sensitive than shControl cells to the inhibition of sorafenib in sorafenib-resistant cells (Fig. 6C). BMP4 knockdown decreased the expression of stem cell transcription factors in sorafenib-resistant cells (Fig. 6D). Sphere formation assays revealed that the BMP4 knockdown also produced similar changes in self-renewal to that of SMYD2 knockdown (Fig. 6E). Similarly, the replenishment of BMP4 recombinant protein can effectively reverse the physiological changes caused by BMP4 knockdown, including reduced sorafenib resistance, decreased expression of stem cell transcription factors and diminished cell sphere formation ability (Fig. 6G-6I). Besides, treating with BMP4 receptor antagonist (Noggin) alone or combination with sorafenib generated similar changes to that of SMYD2 or BMP4 knockdown in sorafenib-resistant cells (Fig. 6L-6N). Interestingly, when we supplement BMP4 recombinant protein into the SMYD2-knockdown cell lines, it can effectively reverse the reduced sorafenib resistance, the decreased the expression of stem cell transcription factors, and the weakened sphere-forming ability of the resistant cell lines (Fig. 6O-6Q), which were also affected by SMYD2 knockdown. Together, these results demonstrate that SMYD2 promotes liver CSCs self-renewal by enhancing BMP4 expression in sorafenib-resistant cells.

Fig. 6.

BMP4 acts as a downstream effector of SMYD2 to mediate the maintenance of self-renewal of liver CSCs in sorafenib-resistant cells. (A-B) BMP4 expression in sorafenib-resistant cells after stable BMP4 knockdown was assessed by western blot and RT-PCR. (A) Left, western blot analysis of BMP4 and GAPDH. Right, Quantification of relative protein expression of BMP4/GAPDH is shown. (B) qPCR analysis of BMP4 gene. (C, G, L, and O) IC₅₀ of sorafenib treatment was measured by CCK8 assay in the sorafenib-resistant cells with indicated treatment. (D, H, M, and P) Left, western blot analysis of CD133, NANOG, OCT4A, and GAPDH. Right, Quantification of relative protein expression is shown. (E, I, N, and Q) Spheroid formation assay was performed to assess self-renewal ability in vitro.

SMYD2 maintains the self-renewal of liver CSCs through the activation of R-SMADs and the up-regulated expression of ID3 via BMP4

The inhibitor of differentiation 3 (ID3) is a member of the ID protein family, which function as inhibitors of basic helix-loop-helix (bHLH) transcription factors. It has been reported that ID3 is essential for the self-renewal of cancer stem cells. We found that the expression of ID3 gene was down-regulated by SMYD2 knockdown via RNA-seq and RT-PCR (Fig. 3E-3G). ID3 gene is a downstream target of BMP4. After binding to type I and type II receptors, BMP4, belonging to the transforming growth factor-β (TGF-β) superfamily, would directly phosphorylate the transcription factors SMAD1, 5, and 8 (R-SMADs), then couple to SMAD4 and form complexes with Co-SMAD (SMAD4) [37]. This SMAD complex translocases to the nucleus to directly modulate the transcription of downstream target genes, including ID3 [37]. Therefore, we hypothesize that the regulation of stemness of cancer cells by SMYD2 may be achieved through the activation of SMADs and the up-regulation of ID3 expression by BMP4. We firstly investigated the activation of phospho-SMAD1/5/8 and the expression of ID3 gene in sorafenib-resistant cells. It is found that phospho-SMAD1/5/9 and the expression of ID3 gene were increased in both sorafenib-resistant cells (Fig. 7A). We next investigated the role of SMYD2 in the activation of SMAD1/5/9 and the expression of ID3 gene. Loss- and gain-of-function studies showed that SMYD2 was a key factor to modulate the activation of SMAD1/5/9 and the expression of ID3 gene (Fig. 7B-7C), which were further confirmed by the SMYD2 inhibitor AZ505 (Fig. 7D). At last, we investigated the role of BMP4 in SMYD2-mediated activation of SMAD1/5/9 and the expression of ID3 gene. BMP4 was necessary to regulate the activation of R-SMADs and the expression of ID3 gene via loss- and gain-of-function studies (Fig. 7E-7F), which is consistent with previous studies. These results were further confirmed by BMP4 receptor antagonist (Noggin) (Fig. 7G). The addition of recombinant BMP4 protein to both sorafenib-resistant cells with SMYD2 silencing induces reactivation of R-SMADs and restoration of ID3 expression, suggesting that SMYD2 activates R-SMADs signaling pathway and promoted the expression of ID3 genes through BMP4 (Fig. 7H).

Fig. 7.

SMYD2 mediated BMP4 up-regulation promotes stemness maintenance by activating R-SMADs and promoting ID3 expression in sorafenib-resistant cells. (A-H) Left, western blot analysis of P-SMAD1/5/9, SMAD1 and GAPDH in the sorafenib-resistant cells with indicated treatment. Middle, Quantification of relative protein expression is shown. Right, qPCR analysis of the expression of ID3 gene in the sorafenib-resistant cells with indicated treatment. (I and M) The sorafenib-resistant cells with BMP4 (I) or SMYD2 (M) knockdown were then transferred with wide-type ID3 (pcDNA3.1-flag-ID3-widetype, short for “ID3”), or the corresponding control vector (pcDNA3.1-flag). ID3 expression was assessed by RT-qPCR. (J and N) IC₅₀ of sorafenib treatment was measured by CCK8 assay in the sorafenib-resistant cells with indicated treatment. (K and O) Left, western blot analysis of CD133, NANOG, OCT4A, and GAPDH. Right, Quantification of relative protein expression is shown. (L and P) Spheroid formation assay was performed to assess self-renewal ability in vitro.

Interestingly, when we transformed the ID3 plasmid into both sorafenib-resistant cells that silenced BMP4 or SMYD2, we found that re-expression of ID3 was effective in reversing the reduction of resistance to sorafenib, the reduction of the expressions of stem cell markers, and the attenuation of stemness upon BMP4 or SMYD2 silencing (Fig. 7I-7P).

These results suggested that SMYD2 activates R-SMADs and upregulated the expression of ID3 gene through BMP4, thereby maintaining stem cell properties and resistance to sorafenib.

SMYD2 deletion inhibits the growth and metastasis of sorafenib-resistant HCC cells in vivo

Tumor xenograft models were employed to further evaluate the effect of the SMYD2 on tumor growth and metastasis of sorafenib-resistant cells in vivo. We established a subcutaneous tumor xenograft model using sorafenib-resistant cells with SMYD2 knockdown or control cells derived from Huh7 and SMMC-7721 cells in nude mice. To assess the role of SMYD2 in tumor growth, we subcutaneously injected an equal number of sorafenib-resistant cells with SMYD2 knockdown or control shRNA. We found that SMYD2 knockdown dramatically impaired cell viability, tumor growth, and prolonged survival duration (Fig. 8A-8C), which was accompanied by reduced proliferation marker Ki67 by immunohistochemistry (IHC) (Fig. 8E). Besides, Hematoxylin and eosin (H&E) staining showed that SMYD2 knockdown cells developed fewer liver and lung metastasis foci than control cells (Fig. 8D). What’s more, we did detect reductions in SMYD2, BMP4, P-SMAD1/5/9, and ID3 at the protein level or at the mRNA level by IHC or western blot or RT-qPCR in both sorafenib-resistant tumors with SMYD2 silencing than with control shRNA (Fig. 8E-8G).

Fig. 8.

Genetic inhibition of SMYD2 suppresses in vivo tumor growth and liver and lung metastasis. Effects of SMYD2 knockdown on tumor growth and metastasis of sorafenib-resistant cells in a subcutaneous tumorigenic nude mouse model. (A) Tumor volume was monitored with indicated times (mean ± SD, n = 4-6/group). (B) Up: Macroscopic appearance of tumors from indicated treatment. Down: Final tumor masses are represented. (C) Kaplan-Meier survival curve of mice subcutaneous tumors. (D) Left: Representative pictures of livers and lungs stained for H&E. Right: tumor foci number in liver or lung in shown. The percentage of Lung stomata area is shown. (E) Left: Representative pictures of IHC staining for Ki67 and SMYD2 expression. Right: Quantification of mean gray value of Ki67 and SMYD2 is shown. (F) qPCR analysis of the expression of SMYD2, BMP4, and ID3 gene in tumors using sorafenib-resistant cells with shControl or shSMYD2-#1. (G) Left, western blot analysis of BMP4, P-SMAD1/5/9, SMAD1, SMYD2, and GAPDH in tumors using sorafenib-resistant cells with shControl or shSMYD2-#1. Left, Quantification of relative protein expression is shown.

The results suggest targeting SMYD2 suppresses tumor growth and metastasis of sorafenib-resistant HCC cells by BMP4/P-SMADs/ID3 signal pathway.

SMYD2/BMP4/ID3 axis informs poor prognosis of liver cancer patients

To investigate the clinical significance of SMYD2, BMP4, and ID3 in hepatocellular carcinoma (HCC) patients, we analyze 369 HCC samples from the Cancer Genome Atlas (TCGA) database. Based on the optimal cut-off point of the ROC curve for SMYD2 or BMP4 or ID3 expression levels, patients were categorized into high and low SMYD2 or BMP4 or ID3 expression groups (Fig. 9A). Furthermore, gene expression correlation analysis demonstrated that SMYD2 expression is positively correlated with BMP4 expression in HCC, while ID3 expression is not corrected with SMYD2 or BMP4 expression (Fig. 9B). Survival analysis indicated that elevated SMYD2 or BMP4 expression alone was significantly associated with poor prognosis in patients with HCC, while high ID3 expression alone was positively correlated with adverse prognosis but not significantly in HCC patients (Fig. 9C). Subsequently, patients were classified with simultaneous high or low expression of SMYD2 and BMP4, SMYD2 and ID3, BMP4 and ID3, or SMYD2 and BMP4 and ID3, based on the optimal cut-off points of their respective ROC curves. Survival analysis showed that simultaneous high expressions of both or all of them were significantly linked to poor prognosis in liver cancer patients (Fig. 9D). These findings suggest that SMYD2/BMP4/ID3 axis plays a crucial role in the clinical outcomes of HCC.

Fig. 9.

SMYD2/BMP4/ID3 axis informs poor prognosis of liver cancer patients. (A and B) SMYD2, BMP4, and ID3 expression and survival analysis were assessed using the optimal ROC curve cut-off in TCGA. (C) The correlation between SMYD2 and BMP4 in TCGA hepatocellular carcinoma (HCC) samples was evaluated using Pearson's test. (D-E) Survival analysis was conducted on groups with high and low simultaneous expression of SMYD2/BMP4, SMYD2/ID3, BMP4/ID3, or SMYD2/BMP3/ID3 based on the ROC curve cut-off in TCGA.

Discussion

Our data uncover a previously undescribed role of SMYD2 in regulating BMP4-dependant liver CSCs expansion and show that SMYD2 has a critical role in sorafenib-resistant HCC. SMYD2 transcriptionally upregulates BMP4 expression by directly binding to the BMP4 promoter, resulting in the maintenance of H3K4me and H3K36me2 on the promoter regions of BMP4 gene to initiate its transcription. Inhibition of SMYD2 or BMP4 promotes sorafenib-induced growth inhibition in sorafenib-resistant cells. These data provide strong evidence that targeting SMYD2 might be an appealing approach to sensitize sorafenib-resistant HCC cells to sorafenib therapy.

Increasing evidence has revealed that CSCs are a key population of tumor cells that are highly tumorigenic and drug resistant in many cancer types [35]. However, CSCs markers are often nonspecific or even unclear, which pose a considerable challenge to target these cells. Here, our data showed that sorafenib-resistant cells display all the properties of liver CSCs, including an enhanced sphere formation capacity, overexpressed liver CSCs markers, and a strong tumorigenic ability in vivo. Collectively, sorafenib-resistant cells are a good model for study liver CSCs.

SMYD2 has emerged as a highly attractive target based on its elevated expression in HCC, and it’s associated with poor clinical outcomes. In present study, we demonstrated that SMYD2 was upregulated in sorafenib-resistant cells, and sustained the characteristics of liver CSCs in vitro and in vivo. Knockdown or inhibition of SMYD2 in sorafenib-resistant cells reduced sorafenib resistance and self-renewal, whereas ectopic wide-type SMYD2 expression in sorafenib-resistant cells with SMYD2 knockdown recovered these capacities. Collectively, our data showed that SMYD2 played a critical role in regulating CSCs in sorafenib-resistant HCC.

Substrates of SMYD2 vary in different cancer contexts and can be histone and non-histone substrates [[7], [8], [9],[11], [12], [13], [14], [15], [16],36]. So, in the context of sorafenib resistance in HCC, which factor is the enzyme activity substrate of overexpressed SMYD2? Our subsequent ChIP experiments showed that SMYD2 depletion and inhibition reduced the recruitment of SMYD2 and RNA polymerase II as well as histone H3K4me and H3K36me2 modification at the promoter region of BMP4 gene, which could be reversed by complementing wild-type SMYD2. SMYD2 plays a major role in monomethylating H3K4 (H3K4me) and demethylating H3K36 (H3K36me2) on the promoter regions of BMP4 gene. Therefore, the transcriptional activation regulation of BMP4 gene may be mainly achieved through recruiting SMYD2 to its promoter regions, which consequently catalyzes the maintenance of H3K4me and H3K36me2 modification at these sites. However, the involvement of other substrates of SMYD2, such as FGF2, is not excluded here. On the contrary, we initially targeted the lysine methyltransferase SMYD2, aiming to take advantage of its multi-substrate characteristics to achieve better therapeutic outcomes. Moreover, our research on the regulation of other substrates by SMYD2 in the context of sorafenib resistance of HCC, is also in progress.

BMP4 binds to phosphorylation of type II and type I receptors of TGF- ß, and activates SMADs complex to regulate downstream genes, including ID3 gene [37]. Low-dose or endogenous BMP4 upregulated CD133 expression in HCC, and simultaneously promotes the self-renewal, chemotherapeutic resistance, and tumorigenesis of CD133⁺ CSCs [34]. In this study, we found that BMP4, a target of SMYD2, was highly expressed in sorafenib-resistant cells and tumors, and its deletion by shRNA or Noggin inhibited the self-renewal and tumorigenesis of liver CSCs. The deletion of SMYD2 or BMP4 inhibited the activation of R-SMADs and the expression of ID3 gene. Consequently, ID3 which promotes stem cell features could further increases the resistance of CSCs to chemotherapy. Overall, SMYD2 promoted the properties of liver cancer stem cells and increased sorafenib resistance via BMP4/ID3 signaling.

All research ultimately depends on the overall effect in the body. Therefore, it is of great significance to explore what the overall therapeutic effect of targeting SMYD2 in sorafenib-resistant liver cancer. At last, we explored the therapeutic effects of targeting SMYD2 with shSMYD2 and investigated its influence on tumor growth and metastasis of sorafenib-resistant HCC cells in vivo using animal models. We found that targeting SMYD2 by shSMYD2 effectively inhibited sorafenib-resistant tumor growth and distant metastasis in vivo. In addition, we simultaneously detected the expression of BMP4 gene at the protein level and transcription level and the downstream of BMP4 signaling pathway, and found that the results were basically consistent with the previous studies at the cellular level in vitro. A variety of small molecule inhibitors targeting SMYD2 have been developed, but we only tested AZ505 in vitro here. It was found that AZ505 treatment decreased SMYD2 protein at the cellular level in vitro and down-regulated the expression of BMP4 gene. However, the significance of targeting SMYD2 in the treatment of sorafenib resistance was determined in this study, while future research can compare multiple small molecule drugs targeting SMYD2 both in vivo and in vitro, and select the best one to obtain better clinical effects.

Through the TCGA public database, we conducted a comprehensive analysis of 369 liver cancer samples to explore the clinical relevance of SMYD2, BMP4 and ID3 in hepatocellular carcinoma. Analysis of the relationship between SMYD2, BMP4, and ID3 with prognosis in hepatocellular carcinoma (HCC) revealed that, except for ID3, high expression of either SMYD2 or BMP4 is positively correlated with poor prognosis. Additionally, high co-expression of SMYD2 and BMP4 or SMYD2 and ID3, as well as the combined high expression of all three, is significantly associated with worse clinical outcomes. Moreover, a positive correlation was observed between SMYD2 and BMP4 expression, indicating a potential synergistic effect on tumor progression. These findings suggest that SMYD2 plays a critical role in determining clinical outcomes in liver cancer and that its interaction with BMP4 and ID3 may be key factors influencing patient prognosis. Thus, SMYD2, BMP4 and ID3 could serve as potential prognostic biomarkers for liver cancer, offering new insights for the development of personalized therapeutic strategies.

Our results showed that sorafenib resistance in HCC was due to the enrichment of liver CSCs, which appears to be regulated by SMYD2, resulting in enhanced self-renewal property. Active BMP4/ID3 signaling pathway associated with sorafenib resistance is related to epigenetic regulator SMYD2, which up-regulated the expression of BMP4 gene by maintaining H3K4me and H3K36me2 to escape sorafenib stress. We expect our findings to have direct clinical implications as SMYD2 inhibitors are already in advanced clinical development to overcome/delay HCC resistance.

Ethics statement

This animal study was reviewed and approved by the Animal Care and Use Committee of Zhejiang Chinese Medical University (Zhejiang, China).

Availability of data and material

The raw RNA-seq data generated in this study have been deposited in the SRA database under accession number PRJNA1094599 (SRA records will be accessible with the following link after the indicated release date: https://www.ncbi.nlm.nih.gov/sra/PRJNA1094599). The additional data collected during this study are available from the corresponding author upon reasonable request.

Funding statement

This work was supported by grants from National Natural Science Fund of China (Grant Numbers 81402470), Zhejiang Basic Public Welfare Research Project (Grant Numbers LY21H100001), Pre-research Project of Zhejiang University (Grant Numbers ZAYY1 and ZAYY05), Medical Science and Technology Project of Zhejiang Province (Grant Numbers 2017KY128 and 2024KY207), and Hangzhou Science and Technology Development Plan Project (Grant Numbers 20170533B65), and Hangzhou Medical and Health Science and Technology Project (Grant Numbers ZD20210022 and A20220019).

CRediT authorship contribution statement

Shanshan Wang: Project administration, Writing – original draft, Funding acquisition, Visualization, Writing – review & editing. Weicheng Wu: Project administration, Writing – review & editing, Supervision. Zhen Shi: Software, Data curation, Writing – review & editing. Mei Bin: Funding acquisition, Methodology, Writing – review & editing, Data curation, Software, Investigation, Visualization. Fengwei Zhang: Data curation, Writing – review & editing, Funding acquisition, Investigation. Long Cai: Supervision, Writing – review & editing. Kaiqing Lin: Writing – review & editing, Supervision. Zhihui Li: Writing – review & editing, Supervision.

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.