miR-328-3p suppresses hepatocellular carcinoma progression by regulating HMOX1 expression

Abstract

Introduction

Most oncogenic genes contribute to cancer progression, but their role and regulatory mechanisms are not yet fully understood in hepatocellular carcinoma (HCC). This study aimed to explore the role of miR-328-3p and the regulatory relationship between miR-328-3p and HMOX1 in HCC.

Methods

We utilized Cox and LASSO regression to identify a panel of oncogenic genes associated with hepatocellular carcinoma (HCC) progression within the TCGA-LIHC cohort and the GSE104580 dataset. The expression levels of the hub gene, HMOX1, were assessed in HCC cell lines using qPCR. The functional roles of miR-328-3p and HMOX1 were evaluated through a series of in vitro assays, including CCK-8 for proliferation, colony formation, wound healing, and Transwell assays for migration and invasion. The direct interaction between miR-328-3p and HMOX1 was explored using a luciferase reporter assay, Western blot (WB) for protein expression analysis, and functional assays to determine the impact on cell proliferation and migration.

Results

Eight candidate genes (BIRC5, TNSF4, SPP1, HMOX1, ADM, RBP2, IGF1, and LECT2) were screen out. The hub gene HMOX1 among had high expression level in HCC cell lines. High HMOX1 expressing cell line had significantly increased proliferation and migration capacities. Moreover, HMOX1 was identified as a target of miR-328-3p, which regulated the HMOX1 expression in qPCR and WB assays. High miR-328-3p expressing HCC cell had diminished capacities for proliferation and migration. However, concurrent upregulation of HMOX1 expression resulted in enhanced proliferative and migratory abilities in these cells.

Conclusion

Our study has advanced our understanding of the roles of miR-328-3p and HMOX1 in HCC, demonstrating the inhibitory effect of miR-328-3p on the oncogenic activity of HMOX1. Hence, these results revealed the function of miR-328-3p and a novel mechanistic pathway for HCC and suggested the potential therapeutic targeting of miR-328-3p and HMOX1 for HCC intervention strategies.

Instruction

Hepatocellular carcinoma (HCC) accounts for 75–85% of liver cancer cases, making it the third leading cause of cancer death worldwide [1]. Transarterial embolization (TACE) is the recommended first-line treatment for advanced HCC [2]. Although TACE improves survival from 16 to 26 months compared to untreated patients, its clinical benefits are insufficient, with a 5-year survival rate of only 32.4% for TACE-treated patients [3, 4]. Moreover, Patients who underwent TACE had varying prognoses due to the high heterogeneity of HCC, even within the same patient population [5]. Such a phenomenon is partly due to a lack of understanding of the exact molecular mechanisms in HCC. To date, imaging parameters have been associated with TACE treatment response, but these variables are insufficient to explain TACE failure and represent intratumor heterogeneity [6, 7]. To improve patient survival time, it is far-reaching to investigate the risk factors and mechanisms in HCC, allowing it to conduct effective prognostic risk stratification and develop personalized follow-up and treatment strategies.

Treatment failure, such as TACE failure, is the primary cause for dismal prognosis in HCC. TACE can induce immunogenic cell death (ICD), release numerous tumor antigens, and activate anti-tumor response by augmenting the infiltration of tumor-specific CD8 + T cells [8]. In contrast, HCC patients with larger tumor exhibit elevated Th-2 cytokines 2 months following TACE, indicating an immunosuppressive environment [9]. Consistently, TACE promotes the expression of programmed death receptor 1 (PD-1) and programmed death ligand 1 (PD-L1) in HCC, suppressing anti-tumor immunity [10]. Collectively, the contributory factors to poor prognosis for HCC patients were multidimensional and complex. Hence, it is essential to investigate the key gene of associated with HCC progression and treatment failure to improve prognosis.

Most factors such as microRNAs (miRNAs) have been identified to be linked to hepatocarcinogenesis [11]. For example, previous studies suggest that reduced miR-122 levels due to hepatic inflammation may enhance tumor progression [12]. Moreover, miR-223 could inhibit HCC development by hindering hypoxia-induced angiogenesis and immune suppression, showing that it participates in the crosstalk between immunoediting and oncogenic drivers in HCC [13]. Notwithstanding, miRNAs expression aberration is hepatomas’ hallmark [14], yet the mechanisms responsible for these aberrations remain ambiguous. In this present study, we investigated the role in target gene HMOX1, which related to hepatocarcinogenesis. Notably, the expression pattern of HMOX1 was regulated by miR-328-3p in HCC. Our study revealed the newly functional insights between HMOX1 and miR-328-3p, which hold promise for the management and treatment of patients with HCC.

Materials and methods

Data collection and pre-processing

Gene expression profile data and corresponding clinical data of HCC patients, including the TCGA-LIHC cohort (comprising 374 HCC samples and 50 control samples), were downloaded from the Cancer Genome Atlas (TCGA) [15]. We applied exclusion criteria to ensure data integrity: samples with incomplete follow-up information, survival less than 30 days, and duplicate sequencing data were excluded. To obtain the treatment refractoriness–related genes, we used the GSE104580 cohort from Gene Expression Omnibus [16]. Furthermore, the list of the immune gene was obtained from the ImmPort database [17]. These datasets underwent rigorous pre-processing to standardize and prepare the data for subsequent analysis.

Identification of Hub genes contributing to HCC progression

Differentially expressed genes (DEGs) of RNA-Seq and microarray data were conducted using the ‘DESeq2’ and ‘limma’ R packages, respectively in the TCGA and GSE104580 cohort with a threshold (|logFC| > 1 and p < 0.05). Volcano plots of DEGs were generated with the ‘ggplot2’ R package. Additionally, significantly 289 DEGs between TACE-responders and TACE-nonresponders groups were also taken from the previous study [18]. The above DEGs and the immune gene list were overlaid to identify candidate genes using the ‘UpsetR’ R package. Correlation of them was performed using the ‘corplot’ R package and their potential functional/pathways enrichment was further performed using the ‘clusterprofiler’ R package.

In the TCGA cohort, candidate genes with p < 0.05 were identified as genes correlated with overall survival (OS) by univariate Cox regression. To remove redundant genes, LASSO based on 10-fold cross-validation was conducted with the ‘glmnet’ R package. STRING database and Cytoscape were used to identify the hub gene of these candidate genes (http://string-db.org).

Quantitative real-time PCR

To further verify the candidate genes, we performed quantitative real-time PCR (qRT-PCR) to examine the relative expression level of the feature genes in transformed human liver epithelial-2 (THLE2) and HCC cell lines (LM3, Hep3B, and Huh7). Total RNA was isolated from cell lines using TRIzol (Takara Biotechnology, Japan). We synthesized cDNA using a reverse transcription kit (EnzyArtisan, China), followed by PCR. QRT-PCR was carried out with 2 × S6 Universal SYBR qPCR Mix (EnzyArtisan, China). All primer sequences were as follows: HMOX1: F-CCAGGCAGAGAATGCTGAGTTC and R-AAGACTGGGCTCTCCTTGTTGC; GAPDH: F- AAGGTGAAGGTCGGAGTCAAC and R-GGGGTCATTGATGGCAACAATA.

Cell transfection

The siRNA targeting hub gene, HMOX1, was designed by Gene Pharma (Shanghai, China) (sense: 5′-CCAGCAACAAAGUGCAAGATT-3′, antisense: 5′-UCUUGCACUUUGUUGCUGGTT-3′). miR-NC mimic, miR-328-3p mimic, miR-NC inhibitor, and miR-328-3p inhibitor were purchased from RiboBio (China). HMOX1 overexpression plasmid and empty negative control plasmid (pcDNA3.1) were obtained from Obio Technology Corp (China). The siRNA, miRNAs, and plasmids were transfected into cells with Lipofectamine 3000 (Invitrogen, USA). After 48 h, the transfected cells and transfection efficiency were assessed.

Dual-luciferase reporter assay

Dual-luciferase reporter plasmids containing the wild-type (WT) or mutant HMOX1 3’UTR regions were synthesized using the pmirGLO dual-luciferase vector by Obio (Obio Technology, China). HEK293T cells were co-transfected with plasmids and either miR-328-3p mimic or miR-NC mimic using Lipofectamine 3000. The luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega, USA), with Renilla luciferase activity used to normalize the firefly luciferase values.

Western blot analysis

Western blot analysis was performed as previously described [19]. The following primary antibodies were used: anti-HMOX1 (Proteintech, Cat# 10701-1-AP) and anti-β-actin (Proteintech, Cat# 81115-1-RR). The results were semi-quantified using ImageJ software and visualized with GraphPad Prism software. In addition, β-actin was selected as internal reference.

Proliferation assay

To measure cell proliferation, Huh7 cells were seeded in a 96-well plate for CCK8 assay. After 24, 48, and 72 h, the cells in each well were incubated for 1 h with 10 µl CCK8 reagent (NCM Biotech, China), and the absorbance at 450 nm was measured. Also, EdU (5-Ethynyl-2′-deoxyuridine) staining assay was performed by EdU incorporation assay kit (Invitrogen, USA). Additionally, cells were cultured in 6-well plates for colony formation assay and incubated for around 2 weeks until the colony was formed.

Migration assay

The wound healing assay was performed as follows. Equal amounts of cells were cultured in full medium in a 6-well plate until 90% confluency. Cell cultures were scratched using sterile pipette tips. An inverted microscope was used to observe and photograph the cells. For transwell assay, 600 µL Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, USA) containing 10% fetal bovine serum (Gibco, USA) was added to the lower chamber, while 200 µL serum-free DMEM with cells was seeded into the upper transwell chamber. Following 24 h incubation, the cells were preserved with 4% methanol, dyed with 0.1% crystal violet, and photographed in microscopic fields.

Statistical analysis

All statistical analyses for the cohort were conducted in R software (version 4.1.1) and all experimental data were analyzed by GraphPad Prism (version 9.0). Student t-test or Wilcoxon signed-rank test were used for statistical comparisons and Spearman analysis was applied for correlation analysis. All experiments were performed in triplicate. P < 0.05 was regarded as statistically significant.

Result

Landscape of candidate genes in HCC progression

We found 319 DEGs in the GSE104580 cohort and 6847 DEGs in the TCGA cohort. Figure 1A displayed the DEGs expression volcano plots for each cohort. The 289 DEGs identified in the previous study were then integrated and de-duplicated to create the final gene list of treatment failure. Additionally, 1793 immune-related genes were obtained from the ImmPort database. Three gene profiles were intersected to identify 33 potential genes (Fig. 1B). The correlation between these genes is depicted in Fig. 1C. In the analysis of specific functions and biological pathways (Fig. 1D, E), which were primarily associated with the regulation of inflammatory response and the TNF signaling pathway. These enriched pathways demonstrated that TRIGs have the anticipated strong positive correlation with immunity.

Fig. 1.

Landscape of candidate genes in HCC A Volcano map of DEGs in the GSE104580 cohort and TCGA cohort. Blue dots: down-regulated genes; yellow dots: up-regulated genes. B Treatment failure-related genes overlapped with existing, immune genes and DEGs from the TCGA cohort. C Correlation analysis for candidate genes. D Detailed information relating to the biological processes of candidate genes. E Candidate genes-related KEGG pathway enrichment. F Univariate Cox regression of candidate TRIGs. G Lasso coefficients of candidate genes

Figure 1F summarized an exploratory Cox regression examining the impact of candidate genes on OS. 24 genes had the highest correlation with OS. Applying LASSO regression on them yielded an optimal model with eight predictors of OS. The distribution of LASSO coefficients profiles of 8 genes was shown in Fig. 2C, D. HMOX1 served as hub gene of these genes (Fig. 1G).

Fig. 2.

HMOX1 role in HCC. A Protein–protein interaction networks of candidate genes. B The HMOX1 expression in HCC cell lines. C The efficiency of siRNA targeting HMOX1. D, E Colony formation assay and CCK-8 assay for HMOX1 knockdown cell line. F, G Wound healing (scale bar: 250 μm) and transwell assay (scale bar: 100 μm) for HMOX1 knockdown cell line. All experiments were performed in triplicate. ***P < 0.001; **P < 0.01; *P < 0.05

HMOX1 promote liver carcinogenesis

HMOX1 was selected as the hub gene among these candidate genes (Fig. 2A).We adopted qRT–PCR to analyze the relative differential expression of HMOX1 in three human hepatomas and normal THLE2 cell lines (LM3, Hep3B, and Huh7). The results showed that the expression levels of HMOX1 were significantly higher in HCC cell lines (LM3, Hep3B, and Huh7) than that of normal THLE2 cell lines (Fig. 2B). Moreover, we assessed the expression and function of HMOX1. To evaluate the function of HMOX1, siRNA was transfected into Huh7 to generate a knockdown cell line (Fig. 2C). The colony formation assay revealed that Huh7 cells with HMOX1 knockdown had significantly diminished proliferative capacity (Fig. 2D). Similarly, the CCK8 assay demonstrated the same pattern (Fig. 2E). Moreover, compared to the control group, HMOX1 knockdown significantly reduced Huh7 cell line migration (Fig. 2F, G). All of the results indicated that HMOX1 is involved in a wide variety of biological processes in HCC.

HMOX1 related miRNAs

In Fig. 3A and 14 candidate miRNAs might affect HMOX1 expression in HCC. We therefore used qPCR to quantify their expressed pattern in THLE2 and Huh7 (Fig. 3B). Among them, compared the normal liver cell line, miR-328-3p was obviously lowly expressed in HCC cell. Furthermore, compared other group, miR-328-3p mimic significantly reduced HCC cell proliferation in CCK8 (Fig. 3C) and Edu assay (Fig. 3D) as well as their clonogenic ability (Fig. 3E). Consistently, the results of the Transwell experiment indicated that the number of cells crossing the chamber in the miR-328-3p mimic group was lower than that in the control group (Fig. 3F). Increasing the expression of miR-328-3p could inhibit the proliferation and migration of liver cancer cells.

Fig. 3.

HMOX1 Related miRNAs. A miRNAs networks of HMOX1. B miRNAs expression in HCC cell lines. C CCK-8 assay for miR-328-3p mimic. D Edu assay for miR-328-3p mimic, scale bar: 75 μm. E Colony formation assay for miR-328-3p mimic, scale bar: 100 μm. F Transwell assay for miR-328-3p mimic, scale bar: 100 μm. All experiments were performed in triplicate. ***P < 0.001; **P < 0.01; *P < 0.05

The mir-328-3p had a strong impact on HMOX1 in HCC

We then focused on miR-328-3p and used a dual luciferase reporter assay for miR-328-3p. As shown in Fig. 4A, these results indicated that miR-328-3p directly targeted HMOX1. Moreover, the HMOX1 gene expression level in miR-328-3p mimic significantly decreased in qPCR assay (Fig. 4B), and corresponding protein level in WB assay (Fig. 4C, D). Meanwhile, with the decreasing miR-328-3p expressed level, HMOX1 expressed level was clearly elevated (Fig. 4E, F). Overall, these finding revealed that miR-328-3p could regulate the HMOX1 expression in HCC.

Fig. 4.

miR-328-3p targeted HMOX1. A Luciferase assay between HMOX1 and miR-328-3p. B–D HMOX1 expression in overexpressing miR-328-3p HCC cells. E, F HMOX1 expression in inhibiting miR-328-3p HCC cells. All experiments were performed in triplicate. ***P < 0.001; **P < 0.01; *P < 0.05

The relationship between mir-328-3p mimic and OE-HMOX1

HMOX1 overexpression HCC cell line was successfully constructed (Fig. 5A, B). WB results indicated increased HMOX1 expression in miR-328-3p mimic and OE-HMOX1 group compared with miR-328-3p mimic group (Fig. 5C, D). There were no differences between miR-NC mimic and OE-NC groups. In miR-328-3p mimic, the results from cell proliferation assays showed a significantly lower proportion of positive cells compared miR-328-3p mimic and OE-HMOX1 group, indicating that miR-328-3p could affect HCC proliferation activity by regulating HMOX1expression (Fig. 5E, F). Moreover, in Fig. 5G, the finding from Transwell assay suggested that miR-328-3p might play a role in the process of HCC metastasis through regulating HMOX1.

Fig. 5.

The relationship between OE-HMOX1 miR-328-3p. A, B HMOX1 expression in OE-HMOX1 cell lines. C, D HMOX1 expression in different cell lines. E CCK-8 assay for miR-328-3p mimic ± OE-HMOX1. F Colony formation assay for miR-328-3p mimic ± OE-HMOX1, scale bar: 100 μm. G Transwell assay for miR-328-3p mimic ± OE-HMOX1. All experiments were performed in triplicate, scale bar: 100 μm. ***P < 0.001; **P < 0.01; *P < 0.05

Discussion

HCC is characterized by intratumoral heterogeneity as well as high aggressiveness, and is considered an immunogenic tumor with multiple immune cells infiltrating the tumor [20]. Such properties invariably contribute to the refractory nature of HCC in most therapeutic strategies.For instance, TACE is recommended for unresectable HCC (uHCC) at Barcelona Clinic Liver Cancer (BCLC) stages A and B, but the outcomes of patients undergoing TACE remain heterogeneous even in the same intermediate stage [3, 4]. Therefore, it is crucial to identify predictors or models that contribute to the early identification of high risk patient, as more precise risk stratification and optimal intervention may considerably improve prognosis. Moreover, it is reported that TACE has two distinct effects on the immune response of tumors [21]. Most oncogenic gene are intricately linked to tumor development and progression of cancer. In this study, we investigated the role and relationship between HMOX1 and miR-328-3p in HCC.

The role of HMOX1 in various malignant tumors is contentious [22]. Studies indicate that HMOX1 is associated with tumorigenesis, involving anti-apoptotic, proliferative, invasive, and metastatic processes, potentially serving as a therapeutic target for cancer [23]. Evidence suggests a correlation between HMOX1 and tumor growth and metastasis in HCC [24]. In HCC mouse models, downregulation of HMOX1 increases cellular damage and apoptosis, thereby inhibiting tumor growth [25]. However, the specific functions of HMOX1 in the tumor microenvironment (TME) of HCC vary depending on different pathways, such as ferroptosis and drug resistance [26, 27]. Overexpression of HMOX1 can protect HepG-2 cells from the antitumor effects of cisplatin and simultaneously inhibit HCC progression through microRNA pathways [28]. The interactions between HMOX1 and various microRNAs reveal a complex regulatory network [29]. Furthermore, while our study indicates a correlation between elevated HMOX1 expression and poor prognosis in HCC, it is important to acknowledge that conflicting evidence exists within the literature. Several studies have confirmed low expression levels of HMOX1 in HCC, challenging the notion of its consistent upregulation in this context [24]. This divergence in findings underscores the complexity of HMOX1’s role in liver cancer and necessitates a nuanced discussion of potential contributing factors. The regulation of HMOX1 is intricately linked to the Nuclear factor erythroid 2-related factor 2 (Nrf2) signaling pathway, which can be modulated by various factors including age, gender, and viral infections such as HBV [30–32]. For instance, chronic HBV infection may lead to sustained activation of the Nrf2 pathway, inducing HMOX1 expression as a defense mechanism against oxidative stress [33, 34]. Additionally, the heterogeneity of HCC and the diverse genetic alterations that occur during tumorigenesis could contribute to the variable expression patterns of HMOX1. Certain genetic mutations or epigenetic modifications might either upregulate or downregulate HMOX1 expression, depending on the tumor’s specific molecular context [35, 36]. Methodological differences in assessing HMOX1 expression could also account for the discrepancies observed across studies. Variations in the use of immunohistochemistry versus RNA sequencing, the specificity of antibodies, cutoff points for positive staining, and tissue processing methods can all influence the reported expression levels of HMOX1 [37]. Considering these factors, the interpretation of HMOX1’s role in HCC must be contextualized within the specific patient population and the experimental conditions of each study. Future research should aim to clarify the mechanisms governing HMOX1 expression in HCC and to understand how these mechanisms interact with clinical variables to influence patient outcomes. Such understanding could pave the way for targeted therapeutic strategies that leverage HMOX1’s potential as both a prognostic biomarker and a therapeutic target in HCC.

Using data from TCGA set, we utilized the Cox model and LASSO algorithm to identify the candidate genes of HCC progression. These candidate genes played a significant role in immune-related pathways, with a particular emphasis on the regulation of the inflammatory response and TNF signaling pathway. Among them (BIRC5, TNFSF4, SPP1, HMOX1, ADM, RBP2, IGF1, and LECT2), HMOX1 was identified as hub gene in HCC. HMOX1 is overexpressed in HCC, where it promoted cell proliferation and migration. HCC patients with high HMOX1 expression correlated with dismal prognosis.

Notably, in breast cancer, targeting HMOX1 was regarded as a promising therapy, with an activation of antitumor response [38]. The most important factor in determining the efficacy of most treatments is the mobilization of immune cells to identify and eventually eliminate cancer cells [8]. Therefore, our results provided great insights for targeting HMOX1 in HCC management.

Through further mechanistic exploration of regulating HMOX1, we found miR-328-3p involved in HMOX1 expression. However, reports on the role of miR-328-3p in HCC are rare. miR-328-3p had low expression in HCC. Elevated miR-328-3p expression could significantly inhibit proliferation and migration of cancer cells. Moreover, miR-328-3p was able to regulate HMOX1 expression. We noted a negative correlation between the upregulation of miR-328-3p expression and the downregulation of HMOX1 expression. Conversely, suppression of miR-328-3p expression by a specific inhibitor led to an augmentation in HMOX1 protein levels. High miR-328-3p expressing HCC cells had diminished capacities for proliferation and migration. However, concurrent upregulation of HMOX1 expression resulted in enhanced proliferative and migratory abilities in these cells. Hence, these results contribute to the understanding of miR-328-3p’s role and suggest a potential mechanistic pathway in HCC.

Previous studies have reported low miR-328-3p expression with a poor prognosis across multiple tumor, such as lung adenocarcinoma, osteosarcoma and head and neck squamous cell carcinoma (HNSCC) [39–41]. Moreover, it assumes a pivotal role in modulating the intricate network within the TME. For example, osteosarcoma progression can be inhibited by miR-328-3p, which may be restored through the downexpression of MMP-16 [40]. It potentially also exerts inhibitory effects on HNSCC progression through the targeting of H2AFX, thereby modulating the mTOR pathway [41]. In non-small cell lung cancer, upregulation of miR-328-3p enhances the sensitivity of radiotherapy [42]. Collectively, miR-328-3p is a versatile molecule implicated in the modulation of numerous essential cellular signaling pathways in TME. In our study, we unveiled a novel pathway involving miR-328-3p and HMOX1 in HCC, which presents a promising target for the suppression of hepatocarcinogenesis.

While our findings provide valuable insights into the role of miR-328-3p and HMOX1 in HCC and suggest potential avenues for therapeutic intervention, it is important to consider the limitations of this study. Our research, which is primarily based on in vitro experiments and a limited number of in vivo models, may not fully capture the clinical context of HCC. The regulatory effects of miR-328-3p on HMOX1 expression, as observed, necessitate further validation across a broader spectrum of samples and HCC subtypes. Furthermore, our comprehension of the intricate mechanisms governing the interaction between miR-328-3p and HMOX1, including their dynamic changes within the tumor microenvironment and potential interactions with other signaling pathways, remains incomplete. Future research endeavors must endeavor to surpass these limitations by engaging in large-scale clinical studies and conducting a more profound investigation into the molecular underpinnings of these interactions. This will be essential for a comprehensive understanding of the roles played by miR-328-3p and HMOX1 in HCC progression and for evaluating their viability as therapeutic targets.

Conclusion

In conclusion, our investigation has advanced our understanding of the roles of miR-328-3p and HMOX1 in HCC. However, the role and its precise mechanisms of miR-328-3p in the HCC pathogenesis have yet to be fully elucidated. Our finding demonstrated the inhibitory effect of miR-328-3p on the oncogenic activity of HMOX1, suggesting the potential therapeutic targeting of miR-328-3p and HMOX1 for HCC intervention strategies.

Author contributions

Study concept and design: Deguo Wang and Jian Dong; Manuscript revision and manuscript preparation: Weixing Wang and Jun Li; Literature search, formal analysis and validation: Weixing Wang, Jun Li and Changjun Pan; Visualization: Weixing Wang and Jun Li.

Data availability

Please feel free to contact the corresponding author for inquiries regarding access to the datasets generated and/or analyzed in the course of this study.

Declarations

Competing interests

The authors declare no competing interests.