MicroRNA-93 activates c-Met/PI3K/Akt pathway activity in hepatocellular carcinoma by directly inhibiting PTEN and CDKN1A

Abstract

To assess the role of microRNAs (miR) in hepatocellular carcinoma (HCC), we performed comprehensive microRNA expression profiling using HCC cell lines and identified miR-93 as a novel target associated with HCC. We further verified miR-93 expression levels in advanced HCC tumors (n=47) by a direct PCR assay and found that elevated miR-93 expression level is significantly correlated with poor prognosis. Elevated miR-93 expression significantly stimulated in vitro cell proliferation, migration and invasion, and additionally inhibited apoptosis. We confirmed that miR-93 directly bound with the 3' untranslated regions of the tumor-suppressor genes PTEN and CDKN1A, respectively,and inhibited their expression. As a result of this inhibition, the c-Met/PI3K/Akt pathway activity was enhanced. IHC analysis of HCC tumors showed significant correlation between c-Met protein expression levels and miR-93 expression levels. Knockdown of c-Met inhibited the activation of the c-Met/PI3K/Akt pathway regardless of hepatocyte growth factor (HGF) treatment, and furthermore reduced the expression of miR-93 in these HCC cells. miR-93 also rendered cells to be more sensitive to sorafenib and tivantinib treatment. We concluded that miR-93 stimulated cell proliferation, migration, and invasion through the oncogenic c-Met/PI3K/Akt pathway and also inhibited apoptosis by directly inhibiting PTEN and CDKN1A expression in human HCC.

Figure 1. Identification of miR candidates in HCC

(A) Heat map of miR sequences used to identify miRs whose relative expression levels changed by a factor of 10-fold based on a logarithmic scale. (B) Verification of miR-93 in six HCC cell lines and a normal hepatocyte cell line referenced by miR-181a.

Figure 2. miR-93 is enhanced in HCC tumor tissues assessed in PEAT

HCC clinical specimens and non-cancer liver specimens were assessed to confirm the significance of miR-93. (A) Expression of miR-93 in HCC tumors (n=47) and adjacent non-HCC tissues (n=40). (B) Expression of miR-93 was significantly higher in HCC tumors (n=13) with vascular invasion versus HCC tumors without vascular invasion (n=33). (C) Expression of miR-93 in normal liver autopsy tissues (n=16), cirrhotic liver tissues (n=8), and normal tumor-free liver tissue removed during resection of hepatic melanoma metastases (n=10). (D) Comparison of high and low expressing miR-93 in a disease-free survival curve using Kaplan Meier methods (Log-rank test, p=0.035). * p< 0.05, * *p< 0.01.

Figure 3. Effect of miR-93 knockdown on HCC cell function in vitro

Clinical HCC specimens showed the relevance of miR-93, therefore we confirmed the significance in vitro. (A) Cell proliferation assay comparing anti-miR-93 (red), mimic-miR-93 (green) and a control-miR (blue) for 8 days. (B) Cell migration assay with anti-miR and control-miR in HepG2 cells for 24 hrs. Scale bar = 200μm. (C) Cell migration assay with mimic-miR and control-miR in SNU449 cells for 24 hrs. Scale bar = 200μm. (D) Invasion assays comparing anti-miR-93, mimic-miR-93, and control-miR. Scale bar = 300μm. (E) Invasion rate of HepG2 cells for 72 hrs. (F) Invasion rate of SNU449 cells comparing anti-miR-93 (red), mimic-miR-93 (green) and a control-miR (blue). (G) Apoptosis status monitored by propidium iodide (PI) showing a difference between control (blue) and anti-miR-93 (red) exposed cells. (H) Annexin V assay comparing HCC specimens with mock transfected cells. (I) 10 day incubation in a 3D culture with anti-miR-93, mimic-miR-93, and control-miR. Scale bar = 100μm. (J) Growth curve of tumorigenicity after treatment with anti-miR-93 and mimic-miR-93. * p< 0.05, * *p< 0.01.

Figure 4. miR-93 targets PTEN and CDKN1A in HCC

Pathway analysis to confirm the target of miR-93. (A) Tumor-related genes were selected using a DIANA microT v3.0 algorithm. The database showed that the potential targets of miR-93 are the PTEN and CDKN1A 3′UTR. (B) Luciferase activity in the PTEN-3′UTR was significantly higher in anti-miR-93 transfected HepG2 cells versus control transfected or mimic-miR-93 transfected cells. (C) CDKN1A-3′UTR luciferase activity. (D) Western blot analysis of protein levels of PTEN and CDKN1A upon treatment with anti-miR for 48 hrs referenced by β-actin. (E) c-Met/PI3K/Akt pathway analysis in HepG2 and SNU449. (F) Biostatistical analysis of miR-93 expression in 47 HCC tumor specimens with AFP mRNA status positive expression group (n=25) and negative expression group (n=22). Statistical analysis showed stimulation of miR-93 in the AFP positive group. p=0.048, Wilcoxon test. * p< 0.05.

Figure 5. miR-93 expression correlates with HGF and c-Met IHC intensity

We performed c-Met IHC staining in PEAT and examined the relationship with miR-93 expression. (A) Photomicrographs of 3-μm liver tissue sections showing strong (3+) staining for c-Met in HCC tissue and (B) negative (0) staining for c-Met in non-HCC tissue. Scale bar = 100μm. (C) Expression of miR-93 in 45 HCC tumor specimens with c-Met IHC staining status ranging from 0 to 3+. Biostatistical analysis showed significant increasing of miR-93 going from the c-Met negative group (0) to the c-Met staining group (1+, 2+ and 3+). p=0.002, Wilcoxon test. (D) Assessment of in vitro overexpression of miR-93 using HGF (50 ng/mL) administered for 24 hrs to 4 HCC cell lines (SNU449, Hep3B, PLC/RPF/5, and HepG2) to monitor miR-93 expression. (E) Akt pathway activity in cells with siMet and HGF treatments versus controls. (F-G) Expression of miR-93 in siMet transfected cells. * p< 0.05, * *p< 0.01.

Figure 6. Anti-miR-93 enhanced drug-sensitivity to tyrosine kinase inhibitors

Growth inhibition curves for high (HepG2) and low (SNU449) miR-93 expressing HCC cell lines in the presence of tyrosine kinase inhibitors for 24-72 hrs. (A-a) Results of anti-miR-93 and control-miR transfected HepG2 cell lines exposed to sorafenib for 24 hrs. (A-b) HepG2 cells exposed to sorafenib for 72 hrs. (A-c) Cell viability curve for HepG2 cells exposed to 40nM of sorafenib for 24-72 hrs. (A-d) HepG2 cells exposed to tivantinib for 24 hrs. (A-e) HepG2 cells exposed to tivantinib for 72 hrs. (A-f) Cell viability curve for HepG2 cells exposed to 40nM of tivantinib for 24-72 hrs. (B-a, b) SNU449 cells after transfection of anti-miR-93 in sorafenib for 24hrs and 72hrs, compared with control-miR. (B-c) Cell viability curve for SNU449 cells exposed to 40nM of sorafenib for 24-72 hrs. (B-d, e) SNU449 cells after transfection in tivantinib for 24hrs and 72hrs. (B-f) Cell viability curve for SNU449 cells exposed to 200nM of tivantinib for 24-72 hrs. * p< 0.05.

Figure 6. Anti-miR-93 enhanced drug-sensitivity to tyrosine kinase inhibitors

Growth inhibition curves for high (HepG2) and low (SNU449) miR-93 expressing HCC cell lines in the presence of tyrosine kinase inhibitors for 24-72 hrs. (A-a) Results of anti-miR-93 and control-miR transfected HepG2 cell lines exposed to sorafenib for 24 hrs. (A-b) HepG2 cells exposed to sorafenib for 72 hrs. (A-c) Cell viability curve for HepG2 cells exposed to 40nM of sorafenib for 24-72 hrs. (A-d) HepG2 cells exposed to tivantinib for 24 hrs. (A-e) HepG2 cells exposed to tivantinib for 72 hrs. (A-f) Cell viability curve for HepG2 cells exposed to 40nM of tivantinib for 24-72 hrs. (B-a, b) SNU449 cells after transfection of anti-miR-93 in sorafenib for 24hrs and 72hrs, compared with control-miR. (B-c) Cell viability curve for SNU449 cells exposed to 40nM of sorafenib for 24-72 hrs. (B-d, e) SNU449 cells after transfection in tivantinib for 24hrs and 72hrs. (B-f) Cell viability curve for SNU449 cells exposed to 200nM of tivantinib for 24-72 hrs. * p< 0.05.

Figure 7. Schematic representation of miR-93 with the oncogenic c-Met/PI3K/Akt pathway miR-93 directly binds to the PTEN and CDKN1A 3′UTRs