miR-138-5p Inhibits Vascular Mimicry by Targeting the HIF-1α/VEGFA Pathway in Hepatocellular Carcinoma

Tumour vascular mimicry (VM) is the process by which new blood vessels are formed by tumour cells rather than endothelial cells. An increasing number of studies have revealed that the VM process is associated with cancer progression and metastasis. MiR-138-5p has been reported to act as a tumour suppressor in many cancers. However, the role and underlying mechanism of miR-138-5p in hepatocellular carcinoma (HCC) VM remain unclear. In this study, VM density was detected by CD31/periodic acid-Schiff double staining in HCC clinical specimens. We found that miR-138-5p expression correlated strongly and negatively with microvessel density. Additionally, the miR-138-5p mimic or inhibitor decreased or increased, respectively, tube formation capacity in HepG2 and Hep3B cells. Consistent with this finding, miR-138-5p repressed vessel density in vivo. Moreover, miR-138-5p targeted hypoxia-inducible factor 1α (HIF-1α) and regulated the expression of HIF-1α and vascular endothelial growth factor A (VEGFA), which are established classical master regulators for angiogenesis. Consistent with these findings, the HIF-1α inhibitor CAY10585 effectively blocked HCC cell VM and VEGFA expression. In conclusion, miR-138-5p inhibits HepG2 and Hep3B cell VM by blocking the HIF-1α/VEGFA pathway. Therefore, miR-138-5p may serve as a useful therapeutic target for miRNA-based HCC therapy.


Introduction
Hepatocellular carcinoma (HCC) is the most common primary liver malignancy and is the third leading cause of cancer-related death worldwide [1,2]. Accumulated evidence indicates that vascular invasion and metastasis play important roles in the high mortality rate of patients with HCC [3,4]. Tumour growth and metastasis require an adequate blood supply to transport nutrients and oxygen and remove metabolic waste, providing a pathway for tumour metastasis and stimulating growth of the tumour mass.
Along with endothelial vessels (EVs), tumour vascular mimicry (VM) occurs via the generation of a de novo vascular network by tumour cells to supply nutrients for sustaining tumour growth. EVs and VM are both important processes in the tumorigenicity and metastasis of HCC.
MicroRNAs (miRNAs) are noncoding RNAs approximately 22 nucleotides in length that have an important function in many essential physiological and pathological processes. miRNAs regulate target genes at the posttranscriptional level by binding to their 3 ′ -UTR [5,6]. It has been reported that miR-138-5p acts as a tumour suppressor to inhibit HCC growth, metastasis, glycolysis, and oxaliplatin resistance in vivo and in vitro [7][8][9]. Moreover, miR-138-5p inhibits HCC progression by inhibiting hepatitis B virus replication and expression. However, the roles of miR-138-5p in HCC VM remain largely unknown.
Recent reports have demonstrated that miR-138-5p inhibits endothelial progenitor cell proliferation by inhibiting hypoxia-inducible factor 1α (HIF-1α) [10,11]. HIF-1α is an important molecule for EVs and VM. Given that miR-138-5p affects both tumour cells and endothelial cells and regulates the expression level of HIF-1α in endothelial cells, we hypothesized that miR-138-5p might suppress tumour growth by blocking blood nutrient supply. The effects of miR-138-5p on VM have not been reported in any type of cancer. In this study, we found that miR-138-5p inhibited VM and regulated HIF-1α/vascular endothelial growth factor A (VEGFA) expression by targeting HIF-1α in HCC cells.

Tissue Specimens.
Fresh human HCC and corresponding paratumour tissues were obtained from surgical specimens immediately after their resection from patients who underwent routine surgery at Qingyuan Hospital, The Six Affiliated Hospital of Guangzhou Medical University (Qingyuan, China). Patients who received preoperative irradiation, chemotherapy, immunotherapy, or targeted therapy were excluded. All patients were confirmed to have HCC by postoperative pathology. Finally, 23 pairs of fresh human HCC and corresponding paratumour tissues were included in the study. The research protocol was approved by the Ethics Committee of The Six Affiliated Hospital of Guangzhou Medical University (No. QPH-IRB-A0153). All experiments were carried out in accordance with the international ethical guidelines for biomedical research involving human subjects.

RNA Extraction and Quantitative Real-Time Polymerase
Chain Reaction (qRT-PCR). Total RNA was extracted from cultured cells or human tissues by using TRIzol reagent (Takara, Dalian, China). qRT-PCR assays were conducted to determine the mRNA levels by using a SYBR Green PCR Master Mix kit (Takara, Dalian, China). β-Actin was used as an internal control to analyse HIF-1α and VEGFA mRNA levels. U6 expression was used as an internal control for analysing miR-138-5p levels in HCC tissue and HCC cell lines. The 2 -ΔΔCT method was used to assess the relative levels of HIF-1α and VEGFA mRNA levels. All assays were evaluated in at least three repeats, and all the results are shown as the mean ± standard deviation (SD) for analysis. The primers are shown in Supplemental Table 1.The median expression level was used as the cut-off. Low expression was classified as values below the 50th percentile. High miR-138-5p expression was classified as values at or above the 50th percentile.
2.5. RNA Interference and Transfection. Hep3B and HepG2 cells were seeded into a 6-well plate at 30-50% confluence. Twenty-four hours later, the cells were transfected with the HIF-1α inhibitor CAY10585 or transfected with a miR-138-5p mimic (Sangon Biotech, Shanghai, China) or a miR-138-5p inhibitor (Sangon Biotech, Shanghai, China) using Lipofectamine 2000 reagent (Invitrogen, Carlsbad, USA) according to the manufacturer's instructions. Cells transfected with a scramble control siRNA (NC) were used as controls. The cells were harvested 48 h after transfection. The sequences of the siRNA against HIF-1α, miR-138-5p mimic, and miR-138-5p inhibitor are listed in Supplemental Table 1.

Luciferase Assay.
To test the direct binding of miR-138-5p to the target gene HIF-1α, a luciferase reporter assay was performed as previously described. The fragment of HIF-1α 3′UTR containing miR-138-5p binding sites or mutated sites was inserted into the pmirGLO vector (Promega, Madison, USA) to generate HIF-1α 3′UTR-WT or HIF-1α3′ UTR-MUT. Then, the constructed plasmids were transfected into HepG2 cells together with miR-NC or miR-138-5p. Luciferase activity was measured 48 h posttransfection using a dual-luciferase assay kit (Promega, Madison, USA).

2.7.
Immunohistochemistry. We performed IHC assays by using an IHC kit (Boster Bio-Engineering Company, Wuhan, China) to detect the protein levels of VEGFA and HIF-1α. Tumour tissue was embedded in paraffin and cut into 4 μm sections. Sections were deparaffinized with xylene and rehydrated with graded ethanol then immersed in citrate and boiled at 126°C for 1-2 min to recover antigen. Endogenous peroxidase was blocked with H 2 O 2 (3%) for 20 min. Each section was first incubated with 50 μL of goat serum (16210064, Thermo Fisher, Waltham, MA, USA) for 15 min at room temperature, followed by 50 μL of primary antibody against VEGFA (1 : 200) or HIF-1α (1 : 200) in a humidified chamber at 4°C. After removing the primary antibody, each section was incubated with 50 μL of the secondary antibody goat anti-rabbit IgG H&L (HRP) (ab205718, 1 : 2000, Abcam, UK). Positively stained cells HepG2 HIF-1a miR-138-5p 3 Journal of Immunology Research were stained with 1 mL of 0.05% DAB solution (11718096001, Sigma-Aldrich, USA) followed by haematoxylin (H9627, Sigma-Aldrich, USA) for 3 min. We also used CD31/periodic acid-Schiff (PAS) double staining to detect VM density. The staining intensity was represented by scores as follows: 0, negative; 1, weak; 2, medium; and 3, strong. The staining extent scores were as follows: 0 represents <10%, 1 represents 11-25%, 2 represents 26-50%, 3 represents 51-75%, and 4 represents 76-100%. The final protein expression score was calculated as the intensity score × extent score and ranged from 0 to 12.  Figure 1: miR-138-5p is frequently downregulated in HCC tissues and HCC cell lines compared to control, and low miR-138-5p expression is associated with a high VM density and high HIF-1α and VEGFA levels and indicates a poor prognosis in HCC patients. (a) and (b) RT-PCR analysis showed that miR-138-5p was more highly expressed in HCC tissues than in paratumour tissues. (c) RT-PCR analysis showed that miR-138-5p was expressed at lower levels in HCC cell lines than in human normal hepatocytes (MIHA cells).(d) CD31/PAS double staining was used to assess VM formation (red arrow). Immunohistochemical staining was used to assess HIF-1α and VEGFA expression. The results showed that miR-138-5p expression was related to VM density and HIF-1α and VEGFA levels. (e) The correlation between miR-138-5p and HIF-1α mRNA levels in 23 HCC tissues. The ΔCt values were subjected to Pearson correlation analysis. (f) The correlation between miR-138-5p and HIF-1α mRNA levels in the HepG2 cell line (9 repeats). The ΔCt values were subjected to Pearson correlation analysis. (g)-(i) The log-rank (Mantel-Cox) test showed that HCC patients with low miR-138-5p levels and high HIF-1α and VEGFA levels demonstrated worse OS than other patients from the KM-Plotter database. Statistical significance was set as follows: * represents P < 0:05, * * represents P < 0:01, and * * * represents P < 0:001.
2.9. Bioinformatics Analysis. TargetScan 3.1 online software (https://www.targetscan.org/) was used to predict the putative target genes of miR-138-5p. The species was "human," and "miR-138-5p" was entered as the miRNA name.  Four-week-old male BALB/c nude mice were purchased from the Guangdong Experimental Animal Center of the Chinese Academy of Sciences and were bred and maintained in specific pathogen-free conditions. All mice were housed in a pathogen-free barrier facility with food and water ad libitum. Cells from each group were resuspended in serum-free DMEM at a density of 5 × 10 7 cells per mL, and then, 0.1 mL of the suspension was injected into the backs of the nude mice. At the end of the treatment cycle (4 weeks later), all mice were euthanized by CO 2 asphyxiation. Tumour tissues were obtained for CD31/periodic acid-Schiff (PAS) double staining.

Survival Analysis.
We performed Kaplan-Meir survival analysis on "The Cancer Genome Atlas (TCGA)" publicly available datasets using the 'Kaplan-Meir Plotter' database (https://kmplot.com/analysis/). The miR-138-5p data were based on the miRNA subset. The HIF-1α and VEGFA mRNA cohorts were based on the RNA-seq subset. We also split the dataset by autoselecting the best cut-off in the system.

Statistical
Analysis. Data were shown as χ ± SD, and all data were obtained from at least three replicates. Data were analysed by SPSS 13.0 (SPSS, Inc.) and GraphPad Prism (GraphPad Software, Inc.). The chi-square test was used to examine the difference in miR-138-5p expression between paratumour and tumour tissues. Student's t test or one-way ANOVA was used to assess the differences among groups. Statistical significance was set as follows: * represents P < 0:05, * * represents P < 0:01, and * * * represents P < 0:001.

miR-138-5p Was Negatively Associated with VM Density and Poor Prognosis in HCC Patients.
To determine the relationship between miR-138-5p expression and HCC, we conducted qRT-PCR to examine miR-138-5p expression levels. The median expression level was used as the cut-off. Low expression of miR-138-5p in 21 tissues (4 paratumour and 17 tumour tissues) was classified as values below the 50th percentile. High miR-138-5p expression in 25 tissues (19 paratumour and 6 tumour tissue) was classified as values at or above the 50th percentile. As shown in Figures 1(a) and 1(b), the expression level of miR-138-5p in HCC tissues was lower than that in paratumour tissues. Similarly, miR-138-5p was expressed at lower levels in HCC cell lines than in the human normal hepatocyte cell line (MIHA) (Figure 1(c)). Furthermore, we found that low expression of miR-138-5p was associated with a high VM density and high HIF-1α and VEGFA levels (Figure 1(d)), and the expression of miR-138-5p was negatively correlated with HIF-1α mRNA levels (Figures 1(e) and 1(f)). Furthermore, the survival results from the Online Kaplan-Meier Plotter database indicated that low miR-138-5p and high HIF-1α and VEGFA mRNA levels were associated with a poor prognosis in HCC patients (Figures 1(g)-1(i)). Together, these data suggested that downregulation of miR-138-5p was associated with a high VM density and high HIF-1α and VEGFA levels and might be associated with HCC progression.

miR-138-5p
Repressed VM Formation in HCC Cells. Our data show that the lower level of miR-138-5p in HCC tissues indicates a possible high VM density and poor prognosis. We sought to determine whether miR-138-5p participates in the process of HCC VM. To this end, we conducted a tube formation assay to assess the effects of miR-138-5p knockdown/overexpression by a miR-138-5p inhibitor or miR-138-5p mimic. The expression levels of miR-138-5p after cell transfection with the miR-138-5p mimic or inhibitor are shown in Figures 2(a) and 2(b). Compared with the control group, HCC cells transfected with the miR-138-5p inhibitor exhibited a significantly enhanced VM formation ability, whereas this process was significantly repressed in HCC cells transfected with the miR-138-5p mimic (Figures 2(c) and 2(d)). Our results suggested that the dysregulation of miR-138-5p expression in HCC cells might play an important role in HCC VM in vitro.

miR-138-5p
Repressed Vessel Density In Vivo. To determine the effects of miR-138-5p on HCC VM formation, we injected HepG2 cells that were treated with miR-138-5p  Journal of Immunology Research

Discussion
Angiogenesis is an important hallmark of tumours [13]. Tumours need an adequate blood supply to ensure enough nutrition for further development. In addition to secreting angiogenic substances to induce endothelial cells to form tubes to produce more blood vessels [14], tumour cells can form new tube structures to increase blood supply, which is called VM. VM is a recently described mechanism in which a blood supply is provided by tumour cells rather than endothelial cells, and it has been observed in certain highly aggressive tumours. miRNAs, small noncoding RNAs of approximately 22 nucleotides in length, have been suggested to be important modulatory factors in the process of VM in HCC [15].
It has been reported that a lower miR-138-5p level is associated with tumour progression and metastasis and that miR-138-5p acts as a tumour suppressor in most cancers [16][17][18][19]. In HCC research, recent studies have shown that miR-138-5p is frequently downregulated in HCC compared to controls and inhibits the occurrence and development of HCC. For example, Liu et al. [20] indicated that miR-138-5p could target and inhibit SOX9 expression and thus repress cell proliferation and invasion in HCC. Lin et al. [21] found that miR-138-5p suppresses metastasis and tumorigenesis by enhancing vimentin expression and ubiquitination of cyclin E in HCC. However, the role of miR-138-5p in HCC VM has not been determined.
This study represents the first investigation of miR-138-5p in relation to VM in HCC. In this study, we found that miR-138-5p was downregulated in HCC tissues, and low miR-138-5p expression was significantly correlated with a poor prognosis of patients. Moreover, low miR-138-5p expression was related to a high EV density. Given the different basic tube formation abilities in different cell lines, we chose the HepG2 and Hep3B cell lines to carry outgain-or loss-of-function experiments according to the basic expression level of miR-138-5p and the basic tube formation ability. Consistently, in vitro experiments confirmed that miR-138-5p functioned as a tumour suppressor gene by inhibiting HCC VM. miRNAs exert biological functions by regulating target genes via binding to their 3′-UTR. We found that HIF-1α might be a target gene of miR-138-5p by using bioinformatics tools. By conducting a dual-luciferase reporter assay, we demonstrated that miR-138-5p directly binds to the 3′-UTR of HIF-1α mRNA and that miR-138-5p expression is negatively related to HIF-1α mRNA expression in HCC cell lines. HIF-1α is a hypoxia-responsive factor that responds to hypoxia by activating the master regulator of the transcription of many The HIF-1α inhibitor CAY10585 reversed the effects of the miR-138-5p inhibitor on tube formation by HCC cells. Data are shown as the mean ± SD. All experiments were repeated at least 3 times. * Represents P < 0:05; * * Represents P < 0:01; * * * Represents P < 0:001.