EGF up-regulates miR-31 through the C/EBPβ signal cascade in oral carcinoma

Abstract

Oral squamous cell carcinoma (OSCC) is one of the most prevalent carcinomas worldwide. MicroRNAs (miRNAs) are short, non-coding RNAs that regulate gene expression and modulate physiological or pathological processes including OSCC carcinogenesis. miR-31 has been found to be up-regulated in OSCC and to act as an oncogenic miRNA. However, the molecular mechanism underlying miR-31 up-regulation in OSCC is still obscure. The activation of epidermal growth factor receptor (EGFR) signaling axis plays key roles in driving oral carcinogenesis. Our screening identified that there is up-regulation of miR-31, miR-181b and miR-222 in OSCC cells following EGF treatment. Subsequent analysis showed that EGF treatment led to AKT activation, which then resulted in miR-31 up-regulation. Moreover, EGF treatment and the AKT activation induced by exogenous expression up-regulated C/EBPβ expression. The miR-31 up-regulation induced by EGF was abrogated by AKT inhibition or by the knockdown of C/EBPβ expression. In OSCC cell subclones stably overexpressing the functional isoform of C/EBPβ, miR-31 expression was up-regulated. Curcumin is a natural ingredient exhibiting anti-cancer potential. It was found that curcumin attenuated AKT activation and the up-regulation of C/EBPβ and miR-31 caused by EGF stimulation in OSCC cells. Lastly, concordance across the expression of EGFR, the expression of C/EBPβ and the expression of miR-31 in OSCC tissues was found. This study describes a novel scenario where the up-regulation of miR-31 expression in OSCC is, at least in part, a consequence of EGFR oncogenic activation. Although the AKT activation and C/EBPβ expression after EGF treatment might not be directly linked, both events are the crucial mediators underlying miR-31 up-regulation in the EGFR signaling axis.

Figure 1. EGF up-regulates miR-31 expression by virtue of AKT activation in OSCC cells.

(A) The Graphic algorithm was used to illustrate the expression profile of 14 miRNAs following EGF treatment in SAS and HSC-3 cells. Green, up-regulated; red, down-regulated. (B–E) SAS cells. (B, D, E) qRT-PCR analysis. (C) Western blot analysis. (B) Inhibition of potential EGF downstream signals by pretreatment with LY294002 and U0126. LY294002 decreased endogenous miR-31 expression as well as EGF induced miR-31 expression. (C) Exogenous AKT expression and AKT activation mediated by plasmid transfection. (D) Up-regulation of miR-31 expression induced by AKT activation. (E) LY294002 administration blocked both endogenous miR-31 expression and AKT induced miR-31 expression. VA, vector alone. Numbers below pictures are normalized values. Data in (A) are from duplicate experiments. Other data are the means ± SE from at least triplicate analysis. ns, not significant; *, p<0.05; **, p<0.01; ***, p<0.001; un-paired t-test.

Figure 2. EGF induces C/EBPβ expression in SAS cells.

(A; D, Upper; E, Lower) qRT-PCR analysis. Others, Western blot analysis. (A) EGF treatment up-regulated C/EBPβ mRNA expression at 24 h, 48 h and 72 h. (B) After EGF treatment, AKT activation peaked as early as 1 h and then decreased at 6 h to reach its basal level after 12 h, while C/EBPβ increased at 6 h and lasted after 24 h. (C) Activation of AKT by plasmid transfection for 24 h (Upper) and treatment with LY294002 for 24 h (Lower) increased and decreased C/EBPß expression, respectively. (D) Transfection with 100 nM si-C/EBPβ oligonucleotide down-regulated C/EBPβ mRNA expression at 48 h and 72 h (Upper); it also down-regulated C/EBPβ protein expression at 48 h (Lower). (E) EGF treatment for 48 h up-regulated C/EBPβ protein expression (Upper) and miR-31 expression (Lower). This up-regulation was attenuated after transfecting with si-C/EBPβ oligonucleotide. Numbers below Western blot pictures are normalized values. #, quantitation unavailable due to faint image signals. Data are the means ± SE from at least triplicate analysis. ns, not significant; **, p<0.01, ***, p<0.001; un-paired t-test.

Figure 3. OSCC cell subclones overexpressing C/EBPβ show miR-31 up-regulation.

(A, C, D) qRT-PCR analysis. (B) Western blot analysis. Stable OECM-1 and SAS cell subclones exhibit (A) exogenous C/EBPβ mRNA expression, and (B) protein expression. Exogenous C/EBPβ protein expression results in significant (C) up-regulation of miR-31 and (D) down-regulation of FIH mRNA expression. The numbers below the pictures are normalized values. Data are the means ± SE from at least triplicate analysis. **, p<0.01, ***, p<0.001; un-paired t-test.

Figure 4. Down-regulation of miR-31 expression following treatment with curcumin in oral keratinocytes.

qRT-PCR analysis. (A) SAS cell. (B) OECM-1 cells. (C) HSC-3 cells. (D) NOK primary culture cells. Data are the means ± SE from triplicate analysis. **, p<0.01; ***, p<0.001; un-paired t-test.

Figure 5. Curcumin down-regulates miR-31 expression via EGF downstream signals in OSCC cells.

(A) Western blot analysis. (B, C) qRT-PCR analysis. (A) SAS cells. (B, C) SAS cells (Upper) and HSC-3 cells (Lower). (A) The analysis shows the attenuation of EGF induced AKT activation and C/EBPβ up-regulation after treatment with 12 µM curcumin for 24 h. EGF induced ERK activation was not obviously attenuated by curcumin. Treatment with EGF or curcumin had little effect on the expression of β-catenin or Bcl2. (B) Curcumin treatment down-regulates endogenous C/EBPβ mRNA expression and EGF induced C/EBPβ mRNA expression in SAS cells and HSC-3 cells. (C) Curcumin attenuates endogenous miR-31 expression and EGF induced miR-31 expression in both types of cell. The numbers below the pictures are normalized values. #, quantification unavailable due to faint image signals. Data are the means ± SE from at least triplicate analysis. *, p<0.05; **, p<0.01; ***, p<0.001; un-paired t-test.

Figure 6. Immunoreactivity of EGFR and C/EBPβ, together with staining of miR-31 in representative TMA tissues.

(A), a NCOM tissue. (B–D), three individual OSCC tissues. a to d were consecutive TMA sections from the same sample. (a, b) Immunohistochemistry of EGFR and C/EBPβ, respectively. (c, d) In situ hybridization of miR-31 probe and scramble probe, respectively. Cytosolic and/or membranous brown-red EGFR immunoreactivity, nuclear and cytosolic brown-red C/EBPβ immunoreactivity, and cytosolic and nuclear bluish miR-31 are considered positive signals. Arrows in (A) indicate the representative positive signals in NCOM. The digital scores for the tumor samples were obtained by pixel analysis quantification and are shown in right lower corner of each picture. 100x magnification.

Figure 7. Linear regression analysis of pixel scores of EGFR, C/EBPβ, and miR-31, and immunofluorescence analysis.

(A–C), Correlation between the scores of C/EBPβ and EGFR, between the scores of C/EBPβ and miR-31, and between the scores of EGFR and miR-31, respectively. Significant correlations were found between the scores of each pair of molecules. (D) Immunofluorescence of a representative OSCC tumor. a, phosphorylated (p)-EGFR, red fluorescence; b, C/EBPβ, green fluorescence. c, overlapping of p-EGFR and C/EBPβ. d, overlapping of picture c and the staining of DAPI. The images reveal the presence of p-EGFR and nuclear C/EBPβ in the vast majority of tumor cells. In addition, the presence of tumor cells exhibiting the absence of both p-EGFR and nuclear C/EBPβ expression are indicated by arrows. Furthermore, the presence of tumor cells exhibiting p-EGFR expression when C/EBPβ expression is absent are indicated by arrow heads. Bars, 10 µm.