miR-381 Reverses Multidrug Resistance by Negative Regulation of the CTNNB1/ABCB1 Pathway in HepG2/Dox Cells,and the Diagnostic and Prognostic Values of CTNNB1 / ABCB1 Are Identified in Patients with LIHC

Abstract

Multidrug resistance (MDR) is the biggest challenge in cancer therapy. In this study, we explored the molecular mechanism of MDR in human liver cancer and explored the related diagnostic and prognostic values of the targeted genes in patients with hepatocellular carcinoma. We constructed a multidrug-resistant liver cancer cell line, HepG2/Dox, using the parental subline HepG2. The (3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide) (MTT) assay was used to test the viability of the liver cancer cells. Western blotting was performed to test the expression of ABCB1, β-catenin, and β-actin. Luciferase assays were performed to confirm the relationship between miR-381 and its target genes. The diagnostic and prognostic values of target genes were analyzed using publicly available data from The Cancer Genome Atlas. The Mann–Whitney U test and logistic regression were performed to evaluate the association between ABCB1 or CTNNB1 expression and clinical features in patients with liver hepatocellular carcinoma (LIHC). Finally, Kaplan-Meier and Cox regression analyses were performed to test the effect of ABCB1 or CTNNB1 expression on the overall survival of patients with LIHC. ABCB1 expression was upregulated in HepG2/Dox cells. ABCB1 was found to be a direct target of hsa-miR-381 and was negatively regulated by has-miR-381. Moreover, hsa-miR-381 directly targeted the CTNNB1 3′ UTR and decreased the luciferase activity of CTNNB1. Transfection with miR-183 partially reversed chemotherapeutic drug resistance by downregulating the expression of ABCB1 and CTNNB1 in HepG2/Dox cells. Spearman's analysis results showed that CTNNB1 and ABCB1 were positively correlated in patients with liver cancer, and increased CTNNB1 and ABCB1 expression occurred in patients with liver cancer. High expression of ABCB1 and CTNNB1 indicated poor prognosis in patients with liver cancer; however, neither ABCB1 nor CTNNB1 expression was an independent diagnostic factor in patients with LIHC. Overexpression of hsa-miR-381 partially reversed the MDR of HepG2 cells by directly targeting and negatively regulating the expression of CTTNB1 and ABCB1. Moreover, high expression of ABCB1 or CTNNB1 indicated poor prognosis in patients with liver cancer.

Introduction

Liver cancer is the second major cause of cancer-related deaths worldwide, with high incidence and mortality rates. More than 700,000 deaths caused by liver cancer have been reported recently worldwide, with nearly 50% of the global population with liver cancer in China (Chaudhary et al., 2019). Radical surgical resection is the gold standard treatment for hepatocellular carcinoma (HCC). However, liver cancer is insidious, and early diagnosis is difficult. Approximately 90% of liver cancer patients are diagnosed in the middle and advanced stages and have passed the opportunity for surgical resection, leading to very poor prognosis. The 5-year survival rate of liver cancer is only 7% (Fernandez-Palanca et al., 2021). Cancer recurrence is a key factor leading to death without further improvement in the clinical cure rate of HCC. This is a major risk factor affecting the quality of life and long-term survival rate of patients with HCC (Lee, 2020). Advanced HCC lacking the opportunity for surgery is mainly treated with a combination of treatments, including radiotherapy, chemotherapy, intervention, and minimally invasive and targeted drugs. Among these, chemotherapy mostly uses transcatheter arterial chemoembolization for local chemotherapy, but its clinical effect is unsatisfactory. Conventional chemotherapeutic drug treatment has serious side effects and cannot significantly alleviate disease progression or prolong the life of patients (Li et al., 2020). Thus, multidrug resistance (MDR) in HCC is a clinically significant problem in chemotherapeutics.

In addition, it is necessary to conduct in-depth research on the pathogenesis of liver cancer and its key regulatory factors to improve the early diagnosis of liver cancer and the risk of recurrence after HCC. Follow-up monitoring and the development of new liver cancer treatment drugs are conducive to improving the efficacy of liver cancer treatment, delaying tumor recurrence and metastasis, and thereby improving the prognosis of patients with HCC.

HCC is not sensitive to chemotherapy because of MDR to chemotherapy drugs (Lage, 2016). ATP-binding cassette (ABC) transporters play an important role in MDR production, and ABC transporter superfamily B member 1 (ABCB1) is the most significant. ABCB1 is also known as multidrug resistance protein 1 (MDR1) or P-glycoprotein (P-gp) (Sinha et al., 2021). ABCB1 is responsible for reducing the intracellular concentrations of chemotherapeutic agents. The high expression of ABCB1 in liver cancer cells is the main reason for MDR. Furthermore, the abnormal activation of cancer-related signaling pathways leads to high expression of ABCB1 and promotes the occurrence of HCC. It has been reported that a noncoding RNA regulates the expression of ABCB1, such as microRNAs (miRNA). It is a noncoding single-stranded RNA molecule with a length of ∼22 nucleotides encoded by endogenous genes, which can directly regulate the expression of ABCB1 by regulating cancer-related signaling pathways. For example, miR-491-3p is inversely associated with that of ABCB1 and Sp3 in HCC cells, contributing to increased drug sensitivity of HCC (Zhao et al., 2017). MDR regulates miR-223 by downregulating the expression of ABCB1 in HCC cells (Yang et al., 2013), also inhibiting the lncRNA HOTAIR reversed drug resistance by suppressing the ATAT3/ABCB1A pathway (Zhou et al., 2017).

The occurrence and development of HCC involve a variety of factors and complex signaling pathways, including the PI3K/AKT (Wang et al., 2019), TGF-beta/Smad (Wan et al., 2019), MAPK/ERK (Zhao et al., 2018), Wnt/beta-catenin (Wang et al., 2018), and AKT/TP53 signaling pathways (Wang et al., 2017). In this study, we used the multidrug-resistant cell line HepG/Dox and its parental subline HepG2 as the cell model and further explored the molecular mechanism of MDR in liver cancer cells. Moreover, the diagnostic and prognostic values of the targeted genes related to MDR were investigated in patients with HCC. In addition, we explored whether ABCB1 and its related oncogenes could be used as independent risk factors and prognostic factors for primary liver cancer incidence and mortality. This research may provide a useful theoretical reference for clinical therapy in favor of patients with liver hepatocellular carcinoma (LIHC).

Materials and Methods

Cell lines and agent

Human liver cancer cell line HepG2 (HB-8065) and human kidney cell line 293 T (CRL-3216) were obtained from the American Type of Cell Collection (ATCC, Manassas, VA). SMC7721 (CL-0216) was obtained from Procell (Wuhan, China). HepG2 and SMMC7721 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37°C in a humidified 5% CO₂ atmosphere. 293T cells were cultured in RPMI 1640 medium supplemented with 10% FBS at 37°C in a humidified 5% CO₂ atmosphere. DMEM and RPMI 1640 were purchased from Gibco Corporation (Gibco BRL, Gaithersburg, MD). CTNNB1 human shRNA lentiviral particles (Locus ID 1499) (Cat. No. TL313664V) and pGFP-C-sh Lenti (TR30023) were purchased from Origene (Beijing, China). This study was approved by our local institutional review board of Southwest Medical University.

(3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide) assay

HepG2 and HepG/Dox cells were treated with increasing concentrations of doxorubicin or cisplatin for 48 h. The concentrations of doxorubicin were 1000, 500, 200, 80, 40, 20, 10, 5, 2.5, and 1.25 nM, and there was a 0.1% dimethyl sulfoxide (DMSO)-treated control. The concentrations of cisplatin were 10000, 5000, 2500, 1250, 625.00, 312.50, 156.25, 78.12, 39.06, and 19.5 nM, and there was a 0.1% DMSO-treated control. Before testing, 10 μL of (3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide) (MTT) (5 mg/mL) was added to the culture medium and cultured for 4 h. DMSO (100 μL) was added to dissolve the purple crystals. The absorbance of each well was measured at 490 nm using a microplate reader.

Dual-luciferase reporter assay

MIR381 human miRNA expression plasmid (MI0000789) and PCMVMIR miRNA expression vector (Cat. No. PCMVMIR) were purchased from Origene. The micrOFF hsa-miR-381-3p inhibitor (Cat. No. miR20000736-1-5) and micrOFF inhibitor NC (Cat. No. miR2N0000001-1-5) were purchased from RibBio Corporation (Guangzhou, China). The wild-type or mutant type 3′ UTR of the ABCB1 sequence containing a putative miR-381 binding site was synthesized and cloned into pGL3 luciferase vectors to produce pGL3-ABCB1 wild-type vector (ABCB1-WT) and pGL3-ABCB1 mutant reporter vector (ABCB1-MUT). To confirm the putative target of miR-381, ABCB1-WT or ABCB1-MUT was co-transfected into 293T cells and HepG2 cells with MIR381 human miRNA expression plasmid/micrOFF hsa-miR-381-3p inhibitor and pGL3 vector, pGL3-ABCB1-WT (miR control or miR381), p GL3 vector, or pGL3-ABCB1-MUT (inhibitor control or miR-381 inhibitor). After transfection for 48 h, luciferase activity was detected using the Dual-Luciferase Reporter Assay Kit (Promega).

Quantitative reverse transcription PCR assay

Total RNA was isolated using the RNApure kit (Bioteke) and was retrotranscribed using MLV-reverse transcriptase (Invitrogen). The expression levels of miR-381 and ABCB1 were detected using quantitative real-time PCR. The following primers were used: β-actin, 5′-AGAGGGAAATCGTGCGTGAC-3′ and 5′-CAATAGTGATGACCTGGCCGT-3′; miR-381 forward, 5′-AGTCTATACAAGGGCAAGCTCTC-3′, miR-381 reverse, 5′-ATCCATGACAGATCCCTACCG-3′; and ABCB1 forward, 5′-CCCATCATTGCAATAGCAGG-3′, ABCB1 reverse, 5′-GTTCAAACTTCTGCTCCTGA-3′.

Western blotting analysis

The samples were lysed using RIPA buffer, and the total proteins were separated by PAGE, as previously described (Schulman et al., 2000; Nishitani et al., 2006; Peng et al., 2010). The primary antibodies used were as follows: β-catenin antibody (Cat. No. #9562) and anti-β-actin antibodies (Cat. No. #4967) were rabbit antibodies against β-catenin and β-actin and purchased from Cell Signaling Technology. Anti-ABCB1 antibody (ab235954) was a rabbit polyclonal antibody against ABCB1/P-gp and was purchased from Abcam.

Bioinformatics analysis

The survival probability in high/low levels of ABCB1 or CTNNB1 in primary LIHC and normal specimens was analyzed using the UALCAN database (http://ualcan.path.uab.edu/index.html). The expression of ABCB1 and CTNNB1 was determined using the data with clinical information from the LIHC project. The RNAseq data were obtained from UCSC XENA (https://xenabrowser.net/datapages/) in TPM format of TCGA and GTEx processed uniformly by the Toil process (Vivian et al., 2017). TCGA HCC (LIHC) data and GTEx corresponding normal tissue data were extracted. The RNAseq data in TPM (transcripts per million reads) format were log2 converted for expression comparison between samples. The Mann–Whitney U test was performed to analyze ABCB1 and CTNNB1 expression in para-cancer (n = 160) and tumor (n = 371).

Statistical analysis

The data were analyzed using SPSS 20.0. The data are presented as the mean ± standard deviation. Student's t-test and one-way analysis of variance were used to calculate statistical differences. *p < 0.05 was considered statistically different.

Results

ABCB1 expression was upregulated in HepG2/Dox cells

MDR is considered a major cause of chemotherapy failure. We constructed a pair of MDR liver cancer cell lines, doxorubicin-sensitive HepG2 cells and paired doxorubicin-resistant HepG2/Dox cells, and evaluated the efficacy of doxorubicin and cisplatin in vitro. First, HepG2 and HepG2/Dox cells were treated with increasing concentrations of doxorubicin or cisplatin for 48 h. Next, an MTT assay was performed to test the inhibitory effects of doxorubicin and cisplatin in liver cancer cells. As shown in Figure 1A, doxorubicin had an increasing antitumor effect on HepG2 and HepG2/Dox cells. The IC₅₀ values were 452.60 and 85.89 nM for HepG2/Dox and HepG2 cells, respectively, treated for 48 h. The ratio of IC₅₀ values was ∼5.27-fold in HepG2/Dox cells compared with HepG2 cells, suggesting that HepG2/Dox had higher doxorubicin resistance than HepG2 cells. Next, the inhibitory effects of cisplatin on HepG2 and HepG2/Dox cells were tested using the MTT assay. The results showed that cisplatin had an antitumor effect with increasing concentrations of cisplatin. The IC₅₀ values of cisplatin were 35541 nM in HepG2/Dox cells and 5222 nM in HepG2 cells after 48 h, respectively (Fig. 1B). The ratio of IC₅₀ values was ∼6.81-fold higher in cisplatin-treated HepG2/Dox cells than in cisplatin-treated HepG2 cells, suggesting that HepG2/Dox also had a higher cisplatin resistance than HepG2 cells. All data revealed that HepG2/Dox cells exhibited MDR characteristics.

FIG. 1.

ABCB1 expression is upregulated in HepG2/Dox cells. (A) Doxorubicin-sensitive HepG2 cells and paired doxorubicin-resistant HepG2/Dox cells were treated with increasing concentrations of doxorubicin for 48 h. Cell viability was determined by MTT assay and the inhibition rate was used to calculated IC50 value. Each treatment was set two replicates. (B) HepG2 and HepG2/Dox cells were treated with increasing concentrations of cisplatin for 48 h. IC50 values were calculated by using the inhibition rate of each group. The experiment was repeated twice and each treatment contained two replicates. (C) The expression of ABCB1 was determined by Western blotting analysis in HepG2 and HepG2/Dox cells. β-actin was used as the internal reference gene. (D) The relative expression of ABCB1 is shown in histogram. **p < 0.01, compared with the N.C. shRNA-treated cells. MTT, (3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide).

ABCB1 (known as P-glycoprotein, P-gp; MDR1, MDR1) is a main multidrug exporter linked to MDR in various human cancers. Next, we detected the expression of ABCB1 in HepG2 cells and HepG2/Dox cells by Western blotting analysis. As shown in Figures 1C and D, the expression of ABCB1 was significantly increased in HepG2/Dox cells compared to HepG2 cells (**p < 0.01), suggesting that the HepG2/Dox cell line is an appropriate cell model to investigate the multidrug effects on liver cancer cells.

ABCB1 is a direct target of hsa-miR-381 and is negatively regulated by has-miR-381

To investigate which miRNAs regulate ABCB1 in human liver cancers and to what extent, a TargetScan 7.1 was used to predict the binding site of ABCB1 and the possible miRNAs, and this was verified by a dual-luciferase reporter assay. As shown in Figure 2A, the position 501–530 of the ABCB1 3′ UTR was the putative miRNA targeting site with hsa-miR-381. The mutation of CTNNB1 3′ UTR was also designed as a control miRNA. The validated binding sites of CTNNB1 and hsa-miR-381 were tested using a luciferase activity assay. The 293T cells were co-transfected with ABCB1-WT and hsa-miR-381 or miRNA control for 48 h. The luciferase activity assay results showed that overexpression of hsa-miR-381 significantly reduced the luciferase activity of ABCB1-WT, but did not affect the luciferase activity of ABCB1-MUT. Moreover, transfection with the miR-381 inhibitor increased the luciferase activity of ABCB1-WT, which did not affect the luciferase activity of ABCB1-MUT (Fig. 2B).

FIG. 2.

ABCB1 is the direct target of has-miR-381. (A) TargetScan software predicted the position 501–530 of ABCB1 3′ UTR was the targeting sequences of has-miR-381. (B) The luciferase activity of 293T cells was co-transfected with ABCB1-WT or ABCB1-MUT and hsa-miR-381 or miR-381 inhibitor for 48 h. The luciferase activity was evaluated in each group. *p < 0.05, **p < 0.01, compared with miR-control group. (C) The miRNA levels of miR-381 and ABCB1 were detected by qRT-PCR in HepG2 cells and HepG2/Dox cells. Data are shown as mean ± SD. **p < 0.01. ABCB1-MUT, ABCB1 mutant reporter vector; ABCB1-WT, ABCB1 wild-type vector; qRT-PCR, quantitative reverse transcription PCR; SD, standard deviation.

Furthermore, the effect of miR-381 on ABCB1 expression was tested and confirmed by quantitative reverse transcription PCR (qRT-PCR). As shown in Figure 2C, HepG2 cells and HepG2/Dox cells were transfected with miR-381 plasmid and miR-381 control plasmid for 48 h, and the results showed that miR-381 levels were significantly upregulated in miR-381 plasmid-transfected HepG2 and HepG2/Dox cells compared with miR-381 control plasmid-transfected cells (**p < 0.01, compared with control plasmid-transfected cells). Importantly, the expression of ABCB1 was significantly decreased in miR-381 plasmid-transfected HepG2 and HepG2/Dox cells compared with miR-381 control plasmid-transfected cells (**p < 0.01). All data revealed that miR-381 directly targeted ABCB1 and negatively regulated the expression of ABCB1 in HepG2 and HepG2/Dox cells.

Hsa-miR-381 directly targets the position 807–835 of CTNNB1 3′ UTR and decreases the luciferase activity of CTNNB1

The TargetScan online software (www.targetscan.org) analysis results showed that has-miR-381 directly targeted at position 807–835 of CTNNB1 3′ UTR. The luciferase activity assay results showed that luciferase activity of HepG2 cells was significantly decreased in the co-transfection group with ABCB1-WT and miR-381 plasmid, compared with that of the ABCB1-WT and miR-control plasmid co-transfection group (**p < 0.01). In addition, the luciferase activity of CTNNB1-MUT was not affected by miR-381. This was consistent with the results obtained for the SMMC-7721 cells (Fig. 3). All the results revealed that position 807–835 of CTNNB1 3′ UTR might be the direct target of miR-381.

FIG. 3.

Has-miR-381 directly targets at position 807–835 of CTNNB1 3′ UTR. (A) TargetScan predicted the position 807–835 of CTNNB1 3′ UTR was the targeting sequences of has-miR-381. The mutation of CTNNB1 3′ UTR was also shown here. (B) The luciferase activity of HepG2 cells and SMMC-7721 cells was co-transfected with ABCB1-WT or ABCB1-MUT and hsa-miR-381 for 48 h. The luciferase activity was evaluated in each group. Data are shown as mean ± SD. **p < 0.01, compared.

Transfection with miR-381 partially reverses chemotherapeutic drug resistance

To investigate whether transfection with miR-381 reversed doxorubicin and cisplatin resistance in HepG2 and HepG2/Dox cells, HepG2 and HepG2/Dox cells were transfected with miR-381 plasmid and treated with increasing concentrations of chemotherapy drugs for 48 h. The IC₅₀ values were calculated and compared with those of HepG2 and HepG2/Dox cells transfected with the miR-control plasmid. The IC₅₀values for doxorubicin were 27.74 and 147.30 nM in miR-381-transfected HepG2 and HepG2/Dox cells (Fig. 4A). Furthermore, the IC₅₀ values of cisplatin in HepG2 cells were 842.8 and 7762 nM in HepG2/Dox cells for 48 h, respectively (Fig. 4B). As shown in Table 1, the IC₅₀ values of doxorubicin and cisplatin were significantly decreased in miR-381-transfected HepG2 cells and HepG2/Dox cells compared with the control group. All the data revealed that transfection of miR-381 could partially reverse chemotherapeutic drug resistance, including doxorubicin and cisplatin, in human liver cancer cells.

FIG. 4.

IC50 value is calculated in miR-183-transfected HepG2 and HepG2/Dox cells. HepG2 cells and HepG2/Dox cells were transfected with miR-381 human miRNA expression plasmid and treated with increasing concentrations of doxorubicin (A) and cisplatin (B) for 48 h. Cell viability was determined by MTT assay and the inhibition rate was used to calculate IC50 value. Each treatment was set two replicates. miRNA, microRNAs.

Table 1.

IC50 Values for Doxorubicin and Cisplatin in HepG2 Cells and HepG2/Dox Cells

IC50 value (nM)	Doxorubicin		Cisplatin
IC50 value (nM)	HepG2	HepG2/Dox	HepG2	HepG2/Dox
Ctrl.	85.89	452.60	5222	35541
miR-381	27.74	147.30	842.8	7762

miR-381/β-catenin regulates the expression of ABCB1 in liver cancer cells

To further explore the molecular mechanism of ABCB1 in chemotherapeutic drug-resistant cells, miR381 human miRNA expression plasmid and control plasmid express microRNA precursors from CMV promoter (pCMVMIR) vector were transfected into HepG2/Dox cells. As shown in Figure 5A, the results showed that the expression of β-catenin and ABCB1 was decreased in miR-381-transfected HepG2/Dox cells compared with that in control vector-transfected HepG2/Dox cells. Conversely, miR-381 anti-miRNA oligonucleotides (AMO) and pCMV-MIR vectors were transfected into HepG2/Dox cells, and the results showed that the expression levels of β-catenin and ABCB1 were upregulated in miR-381 AMO-transfected HepG2/Dox cells, compared with the control vector group (Fig. 5B).

FIG. 5.

The expression of ABCB1 in liver cancer cells was regulated by miR-381/β-catenin. (A) HepG2/Dox cells were transfected with miR381 human miRNA expression plasmid and control plasmid for 48 h. The levels of β-catenin and ABCB1 were determined by Western blotting analysis. (B) HepG2 cells were transfected with miR-381 AMO and pCMV-MIR vector for 48 h. The expression of β-catenin and ABCB1 was determined by Western blotting analysis. (C) HepG2 and SMMC-7721 cells were transfected with CTNNB1 shRNA and control shRNA for 48 h. The expression of β-catenin and ABCB1 was determined by Western blotting analysis in CTNNB1 shRNA- and control shRNA-transfected HepG2 cells and SMMC-7721 cells. β-actin was used as internal reference gene. AMO, anti-miRNA oligonucleotides; pCMV-MIR, express microRNA precursors from CMV promoter.

Next, to confirm whether the expression of ABCB1 is regulated and mediated by the β-catenin-related pathway, we first knocked down the expression of β-catenin in HepG2 and SMMC-7721 cells. The expression of ABCB1 was tested by Western blotting analysis. As shown in Figure 5C, the level of ABCB1 decreased in CTNNB1 shRNA-transfected HepG2 cells and SMMC-7721 cells. All data revealed that miR-381 regulated the expression of ABCB1 and reversed the chemotherapeutic drug resistance, mainly regulated by the expression of the β-catenin-related pathway in liver cancer cells.

CTNNB1 and ABCB1 have a positive correlation in liver cancer patients

We analyzed data from TCGA (https://portal.gdc.cancer.gov/) LIHC (HCC) project in the level 3 HTSeq-(Fragments Per Kilobase per Million) FPKM format RNAseq data. FPKM format RNAseq data were converted into TPM (transcripts per million reads) format and log₂ conversion. The statistical descriptions of CTNNB1 and ABCB1 in patients with LIHC are shown in Table 2. Spearman analysis results of CTNNB1 and ABCB1 showed a Spearman correlation coefficient r = 0.250, p < 0.001 (Fig. 6A), indicating that CTNNB1 and ABCB1 have a statistically positive correlation in liver cancer patients.

FIG. 6.

Higher expression of CTNNB1 and ABCB1 in LIHC patients was associated with shorter survival probability compared with lower expression of CTNNB1 and ABCB1 in patents. (A) Correlation analysis of CTNNB1 and ABCB1. Spearman correlation coefficient r = 0.250, p < 0.001, indicating CTNNB1 and ABCB1 have a positive correlation. (B) The expression of CTNNB1 and ABCB1 in LIHC tumor specimens was significantly higher compared with normal specimens. TCGA's LIHC (HCC) and GTEx corresponding normal tissue data were extracted and compared. Mann–Whitney U test, ***p < 0.001. (C) The expression of ABCB1 in primary LIHC and normal specimens. Normal group: n = 50, primary tumor group: n = 371. The statistical significance of normal-versus-primary p-value was 8.352400E-04. (D) The expression of CTNNB1 in primary LIHC and normal specimens. Normal group: n = 50, primary tumor group: n = 371. The statistical significance of normal-versus-primary p-value was <1E-12. (E) Diagnostic value of ABCB1 and CTNNB2 expression in LIHC patients. ROC curve for ABCB1 and CTNNB2 in LIHC patients. (F) The survival probability of high expression of ABCB1 group (n = 93) and low/medium expression of ABCB1 group (n = 272) of LIHC patients in GEPIA database. p = 0.0064. The survival probability of high expression of CTNNB1 group (n = 91) and low/medium expression of CTNNB1 group (n = 274) of LIHC patients in GEPIA database. p = 0.0041. GEPIA; ROC; TCGA; HCC, hepatocellular carcinoma; LIHC, liver hepatocellular carcinoma. Color images are available online.

Table 2.

The Statistical Description of CTNNB1 and ABCB1 in LIHC Patients

Group	No.	Minimum	Maximum	Median	IQR	Mean	SD	SE
ABCB1	374	0.057	7.461	4.379	1.934	4.078	1.625	0.084
CTNNB1	374	3.529	8.653	6.344	1.001	6.305	0.816	0.042

IQR, interquartile range; LIHC, liver hepatocellular carcinoma; SD, standard deviation; SE, standard error.

Increased expression of CTNNB1 and ABCB1 is found in liver cancer patients

Next, we analyzed the RNAseq data from the TCGA LIHC (https://portal. gdc. cancer. gov/;hepatocellular carcinoma) project. The RNAseq data were obtained from UCSC XENA (https://xenabrowser.net/datapages/) in TPM format of TCGA and GTEx processed uniformly by the Toil process (Vivian et al., 2017). TCGA HCC (LIHC) data and GTEx corresponding normal tissue data were extracted. The data were converted into RNAseq data in TPM (transcripts per million reads) format and log2 conversion for expression comparison between samples. As shown in Figure 6B, the Mann–Whitney U test was performed, and data from para-cancer (n = 160) versus tumor (n = 371) showed that in the ABCB1 group, the average level of normal group was 3.832 ± 0.702, and the average level of tumor group was 4.168 ± 1.573; Mann–Whitney U test (Wilcoxon rank-sum test) results showed that the level of ABCB1 in tumors was higher than that in normal tissues, and the difference was statistically significant (p < 0.001). Moreover, the average level of CTNNB1 in the normal group was 5.037 ± 0.538, and the average level of CTNNB1 in the tumor group was 5.834 ± 0.873; the Mann–Whitney U test (Wilcoxon rank-sum test) results showed that the expression of CTNNB1 in tumors was higher than that in normal tissues, and the difference was statistically significant (p < 0.001). In addition, using the UALCAN and TCGA databases, it was demonstrated that ABCB1 and CTNNB1 were overexpressed in primary LIHC (n = 371), compared with normal specimens (n = 50), with p-values of 8.35 E-04 and 1E-12, respectively (Fig. 6C, D). All data revealed that ABCB1 and CTNNB1 had a significantly positive correlation, and both were significantly increased in hepatocellular liver carcinoma.

Relationship between ABCB1, CTNNB1, and clinicopathological characteristics of liver cancer patients

The association between the expression of ABCB1 and CTNNB1 and clinicopathological characteristics, including gender, age, race, weight, height, BMI, tumor status, pathologic stage, T stage, N stage, M stage, histologic grade, residual tumor, adjacent hepatic tissue inflammation, AFP concentration, albumin concentration, prothrombin time, Child-Pugh grade, Ishak fibrosis score, vascular invasion, overall survival (OS) event, disease-specific survival (DSS) event, and progression-free interval (PFI) event status of LIHC patients, was investigated using the RNAseq data in level 3 HTSeq-FPKM format in TCGA LIHC project. As shown in Table 3, 187 patients were classified as having high ABCB1 expression, and 187 patients were classified as having low ABCB1 expression. As shown in Table 3, the expression of ABCB1 was significantly correlated with gender (p < 0.001), age (p < 0.001), weight (p = 0.010), height (p < 0.001), residual tumor (p = 0.026), and AFP concentration (p < 0.001) in LIHC patients. There was no significant correlation between ABCB1 expression and other clinicopathological factors, including race (p = 0.298), BMI (p = 0.312), tumor status (p = 0.987), T stage (p = 0.298), N stage (p = 0.624), M stage (p = 0.123), pathologic stage (p = 0.098), histologic grade (p = 0.199), adjacent hepatic tissue inflammation (p = 0.723), albumin concentration (p = 0.736), prothrombin time (p = 0.097), Child-Pugh grade (p = 0.070), Ishak fibrosis score (p = 0.204), vascular invasion (p = 0.462), OS events (p = 0.447), DSS events (p = 0.479), and PFI events (p = 1.000). Furthermore, the data in Table 4 showed that the level of CTNNB1 in LIHC patients was significantly correlated with AFP concentration (p = 0.02) and Child-Pugh grade (p = 0.005) and there was no significant correlation with gender (p = 0.825), age (p = 0.232), race (p = 0.170), weight (p = 0.505), height (p = 0.957), BMI (p = 0.528), tumor status (p = 0.336), pathologic stage (p = 0.135), T stage (p = 0.503), N stage (p = 0.623), M stage (p = 0.622), histologic grade (p = 0.583), residual tumor (p = 0.218), adjacent hepatic tissue inflammation (p = 0.980), albumin concentration (p = 1.000), prothrombin time (p = 0.658), Ishak fibrosis score (p = 0.888), vascular invasion (p = 0.978), OS event (p = 0.447), DSS event (p = 1.000), and PFI event (p = 0.836).

Table 4.

Analysis of CTNNB1 Expression and Baseline Data

Characteristic	Low expression of CTNNB1	High expression of CTNNB1	p
n	187	187
Gender, n (%)			0.825
Female	59 (15.8)	62 (16.6)
Male	128 (34.2)	125 (33.4)
Age, n (%)			0.232
≤60	82 (22)	95 (25.5)
>60	104 (27.9)	92 (24.7)
Race, n (%)			0.170
Asian	84 (23.2)	76 (21)
Black or African American	5 (1.4)	12 (3.3)
White	88 (24.3)	97 (26.8)
Weight, n (%)			0.505
≤70	90 (26)	94 (27.2)
>70	86 (24.9)	76 (22)
Height, n (%)			0.957
<170	103 (30.2)	98 (28.7)
≥170	73 (21.4)	67 (19.6)
BMI, n (%)			0.528
≤25	88 (26.1)	89 (26.4)
>25	86 (25.5)	74 (22)
Tumor status, n (%)			0.336
Tumor free	98 (27.6)	104 (29.3)
With tumor	83 (23.4)	70 (19.7)
Pathologic stage, n (%)			0.135
Stage I	90 (25.7)	83 (23.7)
Stage II	50 (14.3)	37 (10.6)
Stage III	36 (10.3)	49 (14)
Stage IV	4 (1.1)	1 (0.3)
T stage, n (%)			0.503
T1	92 (24.8)	91 (24.5)
T2	52 (14)	43 (11.6)
T3	36 (9.7)	44 (11.9)
T4	5 (1.3)	8 (2.2)
N stage, n (%)			0.623
N0	123 (47.7)	131 (50.8)
N1	1 (0.4)	3 (1.2)
M stage, n (%)			0.622
M0	135 (49.6)	133 (48.9)
M1	3 (1.1)	1 (0.4)
Histologic grade, n (%)			0.583
G1	30 (8.1)	25 (6.8)
G2	93 (25.2)	85 (23)
G3	57 (15.4)	67 (18.2)
G4	5 (1.4)	7 (1.9)
Residual tumor, n (%)			0.218
R0	169 (49)	158 (45.8)
R1	6 (1.7)	11 (3.2)
R2	1 (0.3)	0 (0)
Adjacent hepatic tissue inflammation, n (%)			0.980
None	61 (25.7)	57 (24.1)
Mild	51 (21.5)	50 (21.1)
Severe	9 (3.8)	9 (3.8)
AFP (ng/mL), n (%)			0.020
≤400	117 (41.8)	98 (35)
>400	24 (8.6)	41 (14.6)
Albumin (g/dL), n (%)			1.000
<3.5	35 (11.7)	34 (11.3)
≥3.5	119 (39.7)	112 (37.3)
Prothrombin time, n (%)			0.658
≤4	108 (36.4)	100 (33.7)
>4	43 (14.5)	46 (15.5)
Child-Pugh grade, n (%)			0.005
A	123 (51)	96 (39.8)
B	5 (2.1)	16 (6.6)
C	1 (0.4)	0 (0)
Ishak fibrosis score, n (%)			0.888
0	36 (16.7)	39 (18.1)
1/2	17 (7.9)	14 (6.5)
3/4	15 (7)	13 (6)
5/6	43 (20)	38 (17.7)
Vascular invasion, n (%)			0.978
No	106 (33.3)	102 (32.1)
Yes	57 (17.9)	53 (16.7)
OS event, n (%)			0.447
Alive	126 (33.7)	118 (31.6)
Dead	61 (16.3)	69 (18.4)
DSS event, n (%)			1.000
Alive	143 (39.1)	144 (39.3)
Dead	40 (10.9)	39 (10.7)
PFI event, n (%)			0.836
Alive	94 (25.1)	97 (25.9)
Dead	93 (24.9)	90 (24.1)
Age, median (IQR)	62 (52, 69)	60 (51, 69)	0.284

Table 3.

Analysis of ABCB1 Expression and Baseline Data

Characteristic	Low expression of ABCB1	High expression of ABCB1	p
n	187	187
Gender, n (%)			<0.001
Female	81 (21.7)	40 (10.7)
Male	106 (28.3)	147 (39.3)
Age, n (%)			0.001
≤60	105 (28.2)	72 (19.3)
>60	82 (22)	114 (30.6)
Race, n (%)			0.298
Asian	86 (23.8)	74 (20.4)
Black or African American	8 (2.2)	9 (2.5)
White	84 (23.2)	101 (27.9)
Weight, n (%)			0.010
≤70	106 (30.6)	78 (22.5)
>70	70 (20.2)	92 (26.6)
Height, n (%)			<0.001
<170	122 (35.8)	79 (23.2)
≥170	52 (15.2)	88 (25.8)
BMI, n (%)			0.312
≤25	96 (28.5)	81 (24)
>25	77 (22.8)	83 (24.6)
Tumor status, n (%)			0.987
Tumor free	99 (27.9)	103 (29)
With tumor	76 (21.4)	77 (21.7)
Pathologic stage, n (%)			0.098
Stage I	81 (23.1)	92 (26.3)
Stage II	47 (13.4)	40 (11.4)
Stage III	44 (12.6)	41 (11.7)
Stage IV	5 (1.4)	0 (0)
T stage, n (%)			0.298
T1	87 (23.5)	96 (25.9)
T2	51 (13.7)	44 (11.9)
T3	44 (11.9)	36 (9.7)
T4	4 (1.1)	9 (2.4)
N stage, n (%)			0.624
N0	132 (51.2)	122 (47.3)
N1	3 (1.2)	1 (0.4)
M stage, n (%)			0.123
M0	135 (49.6)	133 (48.9)
M1	4 (1.5)	0 (0)
Histologic grade, n (%)			0.199
G1	23 (6.2)	32 (8.7)
G2	92 (24.9)	86 (23.3)
G3	62 (16.8)	62 (16.8)
G4	9 (2.4)	3 (0.8)
Residual tumor, n (%)			0.026
R0	169 (49)	158 (45.8)
R1	4 (1.2)	13 (3.8)
R2	1 (0.3)	0 (0)
Adjacent hepatic tissue inflammation, n (%)			0.723
None	64 (27)	54 (22.8)
Mild	52 (21.9)	49 (20.7)
Severe	8 (3.4)	10 (4.2)
AFP (ng/mL), n (%)			<0.001
≤400	94 (33.6)	121 (43.2)
>400	49 (17.5)	16 (5.7)
Albumin (g/dL), n (%)			0.736
<3.5	36 (12)	33 (11)
≥3.5	113 (37.7)	118 (39.3)
Prothrombin time, n (%)			0.097
≤4	110 (37)	98 (33)
>4	37 (12.5)	52 (17.5)
Child-Pugh grade, n (%)			0.070
A	110 (45.6)	109 (45.2)
B	6 (2.5)	15 (6.2)
C	0 (0)	1 (0.4)
Ishak fibrosis score, n (%)			0.204
0	39 (18.1)	36 (16.7)
1/2	13 (6)	18 (8.4)
3/4	18 (8.4)	10 (4.7)
5/6	35 (16.3)	46 (21.4)
Vascular invasion, n (%)			0.462
No	103 (32.4)	105 (33)
Yes	60 (18.9)	50 (15.7)
OS event, n (%)			0.447
Alive	126 (33.7)	118 (31.6)
Dead	61 (16.3)	69 (18.4)
DSS event, n (%)			0.479
Alive	146 (39.9)	141 (38.5)
Dead	36 (9.8)	43 (11.7)
PFI event, n (%)			1.000
Alive	95 (25.4)	96 (25.7)
Dead	92 (24.6)	91 (24.3)
Age, median (IQR)	58 (50, 67)	65 (54, 70)	<0.001

DSS, disease-specific survival; PFI, progression-free interval; OS, overall survival.

The diagnostic and prognostic value of ABCB1 and CTNNB1 expression in LIHC patients

Next, we tested whether ABCB1 and CTNNB1 could be used as independent prognostic markers in LIHC patients. Univariate Cox regression analysis was based on the TCGA dataset. Univariate analysis revealed that pathologic stage, T stage, M stage, and tumor status were significantly associated with OS. There were no statistically significant factors associated with OS in LIHC patients, including age, sex, race, weight, height, N stage, histologic grade, residual tumor, adjacent hepatic tissue inflammation, Ishak fibrosis score, vascular invasion, prothrombin time, AFP concentration, albumin concentration, and the expression of ABCB1 and CTNNB1.

We conducted ROC curve analysis of ABCB1 and CTNNB2 expression data to evaluate the diagnostic value of these genes. The area of CTNNB1 was 0.798, which revealed that in predicting normal and tumor outcomes, the predictive ability of the variable CTNNB1 had a moderate accuracy (AUC = 0.798, CI = 0.760–0.836). In addition, the area of ABCB1 was 0.658, suggesting that the predictive ability of the variable ABCB1 had a lower accuracy (AUC = 0.658, CI = 0.612–0.703) (Fig. 6E). All data suggested that the expression of ABCB1 or CTNNB1 was not an independent diagnostic factor in LIHC patients (Table 5).

Table 5.

Univariate Analysis in LIHC Patients from TCGA Database

Variable	Characteristics	Total (N)	HR (95% CI) Univariate analysis	p-value Univariate analysis
Age	Age (>60 vs. ≤60)	373	1.205 (0.850–1.708)	0.295
Gender	Gender (Male vs. Female)	373	0.793 (0.557–1.130)	0.200
Race	Race (Black or African American & White vs. Asian)	361	1.341 (0.926–1.942)	0.121
Weight	Weight (>70 vs. ≤70)	345	0.941 (0.657–1.346)	0.738
Height	Height (≥170 vs. <170)	340	1.232 (0.849–1.788)	0.272
Pathologic.stage	Pathologic stage (Stage III & Stage IV vs. Stage I & Stage II)	349	2.504 (1.727–3.631)	<0.001
T.stage	T stage (T3 & T4 vs. T1 & T2)	370	2.598 (1.826–3.697)	<0.001
N.stage	N stage (N1 vs. N0)	258	2.029 (0.497–8.281)	0.324
M.stage	M stage (M1 vs. M0)	272	4.077 (1.281–12.973)	0.017
Tumor.status	Tumor status (With tumor vs. tumor free)	354	2.317 (1.590–3.376)	<0.001
Histologic.grade	Histologic grade (G3 & G4 vs. G1 & G2)	368	1.091 (0.761–1.564)	0.636
Residual.tumor	Residual tumor (R1 & R2 vs. R0)	344	1.604 (0.812–3.169)	0.174
Adjacent.hepatic.tissue.inflammation	Adjacent hepatic tissue inflammation (Mild & Severe vs. none)	236	1.194 (0.734–1.942)	0.475
Fibrosis.ishak.score	Ishak fibrosis score (3/4 & 5/6 vs. 0 & 1/2)	214	0.740 (0.445–1.232)	0.247
Vascular.invasion	Vascular invasion (Yes vs. No)	317	1.344 (0.887–2.035)	0.163
Prothrombin.time	Prothrombin time (>4 vs. ≤4)	296	1.335 (0.881–2.023)	0.174
AFP (ng/mL)	AFP (ng/mL) (>400 vs. ≤400)	279	1.075 (0.658–1.759)	0.772
Albumin (g/dL)	Albumin (g/dL) (≥3.5 vs. <3.5)	299	0.897 (0.549–1.464)	0.662
ABCB1	ABCB1 (High vs. low)	373	1.192 (0.844–1.683)	0.319
CTNNB1	CTNNB1 (High vs. low)	373	1.190 (0.843–1.681)	0.323

HR, hazard ratio.

High expression of ABCB1 and CTNNB1 indicates poor prognosis of liver cancer patients

The relationship between ABCB1 and CTNNB1 expression and clinical characteristics of LIHC patients in the GEPIA database. As shown in Figure 6F, high expression of ABCB1 was associated with a worse survival probability in LIHC patients (p = 0.0064). Similarly, higher levels of CTNNB1 were also associated with poor survival probability in LIHC patients (p = 0.041). All data revealed that higher levels of ABCB1 and CTNNB1 indicated poor prognosis in patients with liver cancer.

Discussion

The incidence of HCC has gradually increased worldwide. MDR is the biggest obstacle in clinical liver cancer treatment (Ramos-Penafiel et al., 2018). Thus, a better understanding of chemotherapeutic drug resistance mechanisms is crucial to improve the early diagnosis of liver cancer, develop novel targets to overcome drug resistance in liver cancer, improve the therapeutic effect of liver cancer, delay tumor recurrence and metastasis, and improve the prognosis of patients (Luo et al., 2020). ABCB1 is the first ABC transporter subtype to be discovered. Under normal circumstances, ABCB1 has a cytoprotective function, which can transport harmful toxins produced by human metabolism out of the cell. However, in chemotherapeutic drug-treated patients, ABCB1 normally acts as an efflux pump that transports chemotherapeutic agents out of cells and contributes to chemotherapeutic drug resistance (Yuan et al., 2017). In this study, we constructed doxorubicin-resistant HepG2/Dox cells with paired doxorubicin-sensitive HepG2 cells. The Western blotting analysis showed that ABCB1 expression was significantly increased in HepG2/Dox cells compared to HepG2 cells.

To determine how ABCB1 expression was regulated in doxorubicin-resistant HepG2 cells, we first predicted the possible miRNAs using TargetScan online tools. The results showed that the position 501–530 of ABCB1 3′ UTR might be the putative miRNA targeting site with hsa-miR-381. The dual-luciferase activity assay results showed that ABCB1 was directly targeted by hsa-miR-381 and negatively regulated by has-miR-381. In HepG2/Dox cells, miR-381 exerted antitumor effects and reversed chemotherapeutic drug resistance. This was consistent with several studies by Qiao et al. (2019), who have found that miR-381 functions as a tumor suppressor in pancreatic cancer; miR-381 and miR-489 were reported to inhibit the proliferation of gastric cancers by targeting CUL4B through Wnt/beta-catenin signaling (Fang et al., 2019). In addition, miR381 was found to regulate cancer progression in colorectal cancer (Han et al., 2021), osteosarcoma cancer (Yang et al., 2020), and ovarian cancer (Yang et al., 2020).

A miRNA can bind to several target genes at the 3′ UTR of mRNA, and multiple miRNAs can also target a single mRNA. Next, we found that miR-381 was also directly targeted at position 807–835 of CTNNB1 3′ UTR and decreased the luciferase activity of CTNNB1. This suggests that miR-381 negatively regulates the Wnt/β-catenin signaling pathway by directly targeting CTNNB1 in doxorubicin-resistant HepG2 cells. The Wnt/β-catenin signaling pathway is involved in chemotherapeutic drug resistance in various human cancers. For example, β-catenin activation is associated with castration resistance in prostate cancer (Patel et al., 2020). Inhibition of the Wnt/β-catenin pathway is a new therapeutic method for drug-resistant pancreatic cancer (Ryu et al., 2021). Moreover, β-catenin has been reported to be involved in the chemosensitivity of drug-resistant breast cancer cells (Alshaer et al., 2019; Alkaraki et al., 2020), prostate cancer (Khurana and Sikka, 2019), gastric cancer (Ryu et al., 2019), and cisplatin-resistant human cervical carcinoma (Hou et al., 2020). In this study, our results showed that miR-381 could directly target the 3′ UTR of targeted genes, including CTNNB1 and ABCB1, and reverse the chemotherapeutic drug resistance of HepG2/Dox cells.

Furthermore, the diagnostic and prognostic values of CTNNB1 and ABCB1 were evaluated in patients with LIHC. The expression of ABCB1 was significantly correlated with gender (p < 0.001), age (p < 0.001), weight (p = 0.010), height (p < 0.001), residual tumor (p = 0.026), and AFP concentration (p < 0.001) in LIHC patients, and the level of CTNNB1 in LIHC patients was significantly correlated with AFP concentration (p = 0.02) and Child-Pugh grade (p = 0.005). ROC curve analysis showed that the predictive ability of the variable CTNNB1 had a moderate accuracy (AUC = 0.798 and CI = 0.760–0.836) and the variable ABCB1 had a lower accuracy in predicting LIHC (AUC = 0.658 and CI = 0.612–0.703). Interestingly, both ABCB1 (p = 0.0064) and CTNNB1 (p = 0.041) were associated with poor survival probability in LIHC patients, suggesting that higher levels of ABCB1 and CTNNB1 indicated poor prognosis in LIHC patients.

Thus, in this study, all data showed that overexpression of hsa-miR-381 partially reversed the MDR of HepG2 cells by directly targeting and negatively regulating the expression of CTTNB1 and ABCB1. Moreover, a high expression of ABCB1 or CTNNB1 indicated poor prognosis in liver cancer patients.

Footnotes

Disclosure Statement

No competing financial interests exist.

Funding Information

The work was supported by Luzhou Science and Technology Bureau-Southwest Medical University Joint Project 2015LZCYD-S01 (2/15) and Sichuan Provincial Department of Science and Technology Project 2020ZHFP0058.

References

Alkaraki

, Alshaer

, Wehaibi

, Gharaibeh

, Abuarqoub

, Alqudah

D.A.

et al. (2020). Enhancing chemosensitivity of wild-type and drug-resistant MDA-MB-231 triple-negative breast cancer cell line to doxorubicin by silencing of STAT 3, Notch-1, and beta-catenin genes. Breast Cancer, 27, 989–998.

Alshaer

, Alqudah

D.A.

, Wehaibi

, Abuarqoub

, Zihlif

, Hatmal

M.M.

, et al. (2019). Downregulation of STAT3, beta-catenin, and notch-1 by single and combinations of siRNA treatment enhance chemosensitivity of wild type and doxorubicin resistant MCF7 breast cancer cells to doxorubicin. Int J Mol Sci, 20, 3696.

Chaudhary

, Poirion

O.B.

, Lu

, Huang

, Ching

, and Garmire

L.X.

(2019). Multimodal meta-analysis of 1,494 hepatocellular carcinoma samples reveals significant impact of consensus driver genes on phenotypes. Clin Cancer Res, 25, 463–472.

Fang

, Zhong

, Wang

, Yuan

, Guo

, Yao

, et al. (2019). miR381 and miR489 suppress cell proliferation and invasion by targeting CUL4B via the Wnt/betacatenin pathway in gastric cancer. Int J Oncol, 54, 733–743.

Fernandez-Palanca

, Mendez-Blanco

, Fondevila

, Tunon

M.J.

, Reiter

R.J.

, Mauriz

J.L.

, et al. (2021). Melatonin as an antitumor agent against liver cancer: an updated systematic review. Antioxidants (Basel), 10, 103.

Han

, Wang

, Mao

, Shen

, and Huang

(2021). lncRNA FLVCR1AS1 drives colorectal cancer progression via modulation of the miR381/RAP2A axis. Mol Med Rep, 23, 139.

Hou

, Kan

, Zhu

, Zhang

, Zhou

, Ren

, et al. (2020). Cinnamolide sesquiterpene lactone suppresses in vitro and in vivo cancer cell growth in cisplatin-resistant human cervical carcinoma cells by inducing mitochondrial mediated apoptosis, caspase activation, loss of MMP and targeting Akt/beta-Catenin signaling pathway. J BUON, 25, 709–715.

Khurana

, and Sikka

S.C.

(2019). Interplay between SOX9, Wnt/beta-Catenin and androgen receptor signaling in castration-resistant prostate cancer. Int J Mol Sci, 20, 2066.

Lage

(2016). Gene therapeutic approaches to overcome ABCB1-mediated drug resistance. Recent Results Cancer Res, 209, 87–94.

10.

Lee

(2020). Associations between physical activity and liver cancer risks and mortality: a systematic review and meta-analysis. Int J Environ Res Public Health, 17, 8943.

11.

, Zhang

, Xu

, Cheng

, and Zhang

(2020). Effectiveness and safety of adjunctive traditional Chinese medicine therapy for primary liver cancer patients: a protocol for systematic review and meta analysis. Medicine (Baltimore), 99, e21281.

12.

Luo

, Teng

Q.X.

, Dong

J.Y.

, Yang

D.H.

, Wang

, Dessie

, et al. (2020). Antimicrobial peptide reverses ABCB1-mediated chemotherapeutic drug resistance. Front Pharmacol, 11, 1208.

13.

Nishitani

, Sugimoto

, Roukos

, Nakanishi

, Saijo

, Obuse

, et al. (2006). Two E3 ubiquitin ligases, SCF-Skp2 and DDB1-Cul4, target human Cdt1 for proteolysis. EMBO J, 25, 1126–1136.

14.

Patel

, Brzezinska

E.A.

, Repiscak

, Ahmad

, Mui

, Gao

, et al. (2020). Activation of beta-Catenin cooperates with loss of Pten to drive AR-independent castration-resistant prostate cancer. Cancer Res, 80, 576–590.

15.

Peng

, Xu

, Zhou

, Yang

, Liang

Z.Q.

, and Zhang

(2010). Effect of rosiglitazone on cells cycle, apoptosis and expression of Skp2 and p27Kip1 in hepatocellular carcinoma cell line. Zhonghua Gan Zang Bing Za Zhi, 18, 148–149.

16.

Qiao

, Li

, Wang

, and Bian

(2019). miR381 functions as a tumor suppressor by targeting ETS1 in pancreatic cancer. Int J Mol Med, 44, 593–607.

17.

Ramos-Penafiel

, Olarte-Carrillo

, Ceron-Maldonado

, Rozen-Fuller

, Kassack-Ipina

J.J.

, Melendez-Mier

, et al. (2018). Effect of metformin on the survival of patients with ALL who express high levels of the ABCB1 drug resistance gene. J Transl Med, 16, 245.

18.

Ryu

W.J.

, Han

, Lee

S.H.

, and Choi

K.Y.

(2021). Suppression of Wnt/beta-catenin and RAS/ERK pathways provides a therapeutic strategy for gemcitabine-resistant pancreatic cancer. Biochem Biophys Res Commun, 549, 40–46.

19.

Ryu

W.J.

, Lee

J.E.

, Cho

Y.H.

, Lee

, Seo

M.K.

, Lee

S.K.

, et al. (2019). A therapeutic strategy for chemotherapy-resistant gastric cancer via destabilization of both beta-catenin and RAS. Cancers (Basel), 11, 496.

20.

Schulman

B.A.

, Carrano

A.C.

, Jeffrey

P.D.

, Bowen

, Kinnucan

E.R.

, Finnin

M.S.

, et al. (2000). Insights into SCF ubiquitin ligases from the structure of the Skp1-Skp2 complex. Nature, 408, 381–386.

21.

Sinha

B.K.

, Perera

, and Cannon

R.E.

(2021). NCX-4040, a unique nitric oxide donor, induces reversal of drug-resistance in both ABCB1- and ABCG2-expressing multidrug human cancer cells. Cancers (Basel), 13, 1680.

22.

Vivian

, Rao

A.A.

, Nothaft

F.A.

, Ketchum

, Armstrong

, Novak

, et al. (2017). Toil enables reproducible, open source, big biomedical data analyses. Nat Biotechnol, 35, 314–316.

23.

Wan

Z.H.

, Ma

Y.H.

, Jiang

T.Y.

, Lin

Y.K.

, Shi

Y.Y.

, Tan

Y.X.

, et al. (2019). Six2 is negatively correlated with prognosis and facilitates epithelial-mesenchymal transition via TGF-beta/Smad signal pathway in hepatocellular carcinoma. Hepatobiliary Pancreat Dis Int, 18, 525–531.

24.

Wang

, Liu

, Wu

, Zheng

, et al. (2017). Overexpression of centromere protein K (CENP-K) gene in hepatocellular carcinoma promote cell proliferation by activating AKT/TP53 signal pathway. Oncotarget, 8, 73994–74005.

25.

Wang

, Yang

, Wang

, Shi

, and Li

(2019). Rosmarinic acid inhibits proliferation and invasion of hepatocellular carcinoma cells SMMC 7721 via PI3K/AKT/mTOR signal pathway. Biomed Pharmacother, 120, 109443.

26.

Wang

, Sun

, Wang

, Liu

, Li

, Yao

, et al. (2018). Long non-coding RNA DSCR8 acts as a molecular sponge for miR-485-5p to activate Wnt/beta-catenin signal pathway in hepatocellular carcinoma. Cell Death Dis, 9, 851.

27.

Yang

, He

, Duan

, Shen

, and Dong

(2020). lncRNA FLVCR-AS1 promotes osteosarcoma growth by targeting miR381-3p/CCND1. Onco Targets Ther, 13, 163–172.

28.

Yang

, Zheng

Z.M.

, Li

X.N.

, Li

Z.F.

, Wang

, Geng

Y.F.

, et al. (2013). MiR-223 modulates multidrug resistance via downregulation of ABCB1 in hepatocellular carcinoma cells. Exp Biol Med (Maywood), 238, 1024–1032.

29.

Yuan

Z.T.

, Shi

X.J.

, Yuan

Y.X.

, Qiu

Y.Y.

, Zou

, Liu

, et al. (2017). Bufalin reverses ABCB1-mediated drug resistance in colorectal cancer. Oncotarget, 8, 48012–48026.

30.

Zhao

, Gong

, Liu

, Lv

, Hu

, and Zhang

(2018). Downregulation of Nedd4L predicts poor prognosis, promotes tumor growth and inhibits MAPK/ERK signal pathway in hepatocellular carcinoma. Biochem Biophys Res Commun, 495, 1136–1143.

31.

Zhao

, Qi

, Chen

, Wei

, Yu

, Yan

, et al. (2017). The miR-491-3p/Sp3/ABCB1 axis attenuates multidrug resistance of hepatocellular carcinoma. Cancer Lett, 408, 102–111.

32.

Zhou

J.J.

, Cheng

, He

X.Y.

, Meng

, Ye

H.L.

, and Chen

R.F.

(2017). Knockdown of long non-coding RNA HOTAIR sensitizes hepatocellular carcinoma cell to cisplatin by suppressing the STAT3/ABCB1 signaling pathway. Oncol Lett, 14, 7986–7992.