miR-133b suppresses metastasis by targeting HOXA9 in human colorectal cancer

Abstract

Functions and mechanisms of microRNA (miRNA) involved in colorectal cancer (CRC) metastasis are largely unknown. Here, a miRNA microarray analysis was performed in CRC primary tissues and metastatic hepatic tissues to disclose crucial miRNA involved in CRC metastasis. MiR-133b was decreased and negatively correlated with metastasis in CRC. Overexpression of miR-133b significantly suppressed metastasis of CRC in vitro and in vivo. HOXA9 was identified as a direct and functional target of miR-133b. In addition, HOXA9 was negatively correlated with miR-133b and promoted CRC malignant progress. Moreover, miR-133b decreased HOXA9 expression, and subsequently downregulated ZEB1 and upregulated E-cadherin expression. Intriguingly, lower miR-133b and higher HOXA9 expression significantly contributed to poorer outcomes in CRC patients. Multivariate analysis indicated that miR-133b was an independent and significant predictor of CRC patient overall survival. In conclusion, we newly determined that miR-133b targeted the HOXA9/ZEB1 pathway to promote tumor metastasis in CRC cells. This axis provided insights into the mechanism underlying miRNA regulation of CRC metastasis and a novel therapeutic target for CRC treatment.

Keywords: HOXA9; colorectal cancer; metastasis; miR-133b.

Figure 1. MiR-133b is downregulated in CRC tumor samples and cell lines

(A) Microarray analysis of miRNA expression in CRC tissues from CRC primary cites (P) and hepatic metastatic sites of CRC (LM). (B) Relative expression of miR-133b in TCGA data, which contained 8 normal samples, 34 stage I, 95 stage II, 71 stage III and 31 stage IV samples. (C) qPCR analysis of miR-133b expression in 66 adjacent normal control tissues (N), 56 CRC without metastasis (T), and 10 CRC with metastasis (M). (D) qPCR analysis of miR-133b expression in normal cell lines and different CRC tumor cell lines.

Figure 2. Recovery miR-133b expression in CRC tumor cells inhibits cell proliferation and migration in vitro

MiR-133b was overexpressed in CRC cell lines by using miR-133b mimic (scramble or miR-133b). (A) qPCR analysis of miR-133b levels in CRC cell lines after transfection. (B) Cell colony formation assay. (C) Cell growth was analyzed at different time points after transfection. (D) Cell invasion was analyzed by way of gel invasion assay after virus infection. (E and F). wound closure assay was detected after transfection. *P<0.05 versus scramble control.

Figure 3. MiR-133b suppresses CRC tumorigenesis and metastasis in vivo

HCT8 and HCT116 cells were transfected with miR-133b-mimic or scramble and then subcutaneously injected into nude mice. (A) in vivo subcutaneous tumor cell transplantation, the gross morphology of tumors are represented. (B, C) The statistical data for tumor volume (B) and tumor weight (C) are shown. (D) Representative image of hepatic metastatic sites. (E) Haptic metastatic tumor volume are shown. *P < 0.05, versus scramble control.

Figure 4. HOXA9 was identified as miR-133b targets in CRC

(A) miR-133b targeting genes and the corresponding mutations. (B) examination of luciferase activity. Co-transfection of a wildtype or a mutant HOXA9 3′UTR with miR-133b mimics into HEK 293 cells. Firefly luciferase activity was measured and standardized by Renilla luciferase activity. (C) Left, HCT8 cells were infected with miR-133b mimics or scramble control, and protein levels of HOXA9 was analyzed with western blot. Right, the expression of HOXA9 in normal control (N), CRC without metastasis (T) and CRC with metastasis(M). 3 pair samples were used and a representative sample is shown. (D) Relative expression of HOXA9 in TCGA of CRC, which contained 40 normal samples, 45 stage I, 111 stage II, 83 stage III and 39 stage IV samples. (E) Screenshot of the profiles of the human gene HOXA9 within the BioGPS online portal. All data used for this study are available through the BioGPS database (biogps.org). (F) the expression correlation analysis for miR-133b and HOXA9 in 66 CRC tumor samples. *P < 0.05 versus scramble or healthy controls.

Figure 5. Knockdown of HOXA9 suppresses CRC tumorigenesis and metastasis

HCT8 cells were infected with a lentivirus expressing specific shRNA of HOXA9 genes (HOXA9-KD) or shRNA scramble controls (Lv-scramble). (A) qPCR analysis of the knockdown effect in HCT8 and HCT116 cells infected with HOXA9-KD or Lv-scramble. (B) in vitro cell proliferation was detected with CCK8 assay. (C) in vitro cell proliferation was detected with colony formation assay. (D, E) In vitro cell migration was analyzed with a gel invasion assay (D) and a wound-healing assay (E). (F) Knock down of HOXA9 decreased ZEB1 while increased E-cadherin expression. (G) Overexpression of miR133-b decreased HOXA9 and ZEB1, while increased E-cadherin expression. *P < 0.05, versus scramble control.

Figure 6. Rescue experiments validated the miR-133b - HOXA9 /ZEB1 pathway

(A) Restoration of HOXA9 expression in cells expressing miR-133b mimic blocked the miR-133b-induced suppression of migration and invasion. (B) Knockdown of HOXA9 abolished migration and invasion elevation in miR-133b inhibited cells. (C, D) Re-expression of HOXA9 in cells expressing miR-133b reversed ZEB1 and E-cadherin expression levels alteration induced by miR-133b and knockdown of HOXA9 after miR-133b inhibitor transfection also abrogated protein change of ZEB1 (C) and E-cadherin (D).

Figure 7. Low level of miR-133b and high level of HOXA9 predicted poor prognosis in colorectal cancer patients

(A) overall survival (OS) rates of miR-133b in 66 patients by Kaplan–Meier analysis with log-rank test (**P < 0.001). (B) overall survival (OS) rates of HOXA9 in 66 patients by Kaplan–Meier analysis with log-rank test (*P = 0.041). (C) Overall survival (OS) rates of three groups based on the expression of miR-133b and HOXA9: miR-133b+/HOXA9- (14 specimens); miR-133b+/HOXA9+, or miR-133b-/HOXA9- (19 specimens); miR-133b-/HOXA9+ (33 specimens).