IGF-1-mediated PKM2/β-catenin/miR-152 regulatory circuit in breast cancer

Dysregulation of miRNAs is important in breast cancer initiation and malignant progression. Recently we showed that miR-152 downregulation is associated with breast cancer development, yet the underlying mechanism of miR-152 remains to be well elucidated. In this study, we identified β-catenin as a new direct target of miR-152. MiR-152 inhibited cell proliferation by targeting and inhibiting both β-catenin and PKM2 expression. We found that miR-152 expression sensitized the breast cancer cells to paclitaxel treatment by inhibiting β-catenin and PKM2 expression. Intriguingly, IGF-1 induced β-catenin and PKM2 expression and enhanced β-catenin and PKM2 interaction. Subsequently, IGF-1-induced β-catenin and PKM2 complex translocated into the nucleus, which in turn activated expression of miR-152. These results suggested a regulatory circuit between miR-152, β-catenin and PKM2 in breast cancer. By using human clinical specimens, we also showed that miR-152 expression levels were negatively correlated with β-catenin and PKM2 levels in breast cancer tissues. Our findings provide new insights into a mechanism of miR-152 involved in β-catenin and PKM2 inhibition which would have clinical implication for the cancer development and new treatment option in the future.

Breast cancer is the most frequent type of cancer in women 1 . Although the 5-year relative survival rate for female breast cancer patients has improved due to both improvements in breast cancer treatment and early detection, breast cancer is still the second leading cause of cancer deaths 2 . Therefore, further understanding of molecular mechanisms in breast cancer cells is important to develop new biomarkers and treatment options for breast cancer.
MiRNAs are a class of small endogenous non-coding RNAs composed of 17-24 nucleotides that act as post-transcriptional regulators through directly binding to the 3′-untranslated region (3′ UTR) of their target mRNAs, resulting in the degradation or translational inhibition of the mRNAs 3,4 . MicroRNA-152 (miR-152) contains two different mature miR-152 sequences, namely miR-152-5p and miR-152-3p. MiR-152-3p, the 3′ arm of the hairpin precursor, was highly conserved in evolution and has been well investigated in human cancers than miR-152-5p 5 . Recently, decreased expression of miR-152-3p (here referred as miR-152) has been observed in various types of human cancer cell lines and tumor tissues, such as ovarian 6 , gastrointestinal cancer 7 , hepatocellular carcinoma 8 , endometrial 9 and breast cancer 10 . It was reported that the miR-152 directly targeted DNMT1 (DNA methyltransferase 1) in malignant cholangiocytes, leading to significant reduction of DNMT1 expression at both mRNA and protein levels 11 . This finding was further confirmed in subsequent studies on hepatitis B virus-related hepatocellular carcinoma 8 , ovarian cancer 6 , breast cancer 10 , pancreatic cancer 12 and prostate cancer 13 . Except for DNMT1, accumulating evidence indicates that miR-152 targets on multiple oncogenes like PKM2, IRS-1 and IGF-1R in human breast cancer, and inhibits a variety of cellular functions, including proliferation, angiogenesis and migration, suggesting that miR-152 may potentially function as a tumor suppressor in breast cancer 8,10,14 . β-catenin, the downstream molecule of IGF-1 15,16 , is originally identified as an important junctional component at the cell membrane, where it serves to link cadherin to the actin cytoskeleton via binding of α-catenin 17 . Alternatively, accumulation of β-catenin in the cytoplasm followed by its translocation and activation in the nucleus has also been well characterized as the central event in the progression of canonical WNT/β-catenin signaling 18 . In recent years, dysregulation of Wnt/β-catenin signaling pathway has been recognized as one of the hallmarks of breast cancer initiation and progression, mainly due to the abnormal excessive expression and the activating mutations of β-catenin 19,20 . Recent studies have illustrated the participation of miRNAs in the post-transcriptional regulation of WNT/β-catenin pathway, such as miRNA-720, miRNA-141 and miRNA-208a 21 . The objective of this study was to reveal the molecular mechanism of miR-152 and β-catenin in breast cancer. Previous study indicated that miR-152 directly targeted and inhibited PKM2 in breast cancer 14 . PKM2 is the key member of pyruvate kinase (PK) that catalyzes the final step in glycolysis by transferring the phosphate from phosphoenolpyruvate (PEP) to ADP, thereby generating pyruvate and ATP 22 . In recent years, numerous studies have indicated that PKM2 is preferentially expressed in malignant cancer, playing a vital role in cancer cell proliferation and tumor growth 23,24 .
The IGF signaling contains a dynamic network of proteins including ligands (insulin, IGF-1, IGF-2), their associated receptors (IGF-1R and IGF-2R) and several IGF binding proteins (IGFBPs) that participate in the regulation of human cancer development 25 . Of particular interest, IGF-1 has been most strongly implicated in breast cancer progression based on its mitogenic and anti-apoptotic activities 26 . Earlier studies have identified the association of IGF-1 with the increased risk of breast cancer development 27 . In addition, IGF-1 could bind with ER (estrogen receptor) or PR (progesterone receptor) to promote tumorigenesis and tumor growth in breast cancer 28,29 . Although multiple miRNAs, such as miR-122 30 , miR-18b 31 , miR-515-5p 32 , miR-148a and miR-152 10,14 , are involved in IGF-1 regulation pathway by directly targeting IGF-1, IGF-1R, IRS-1 or FOXO3a in breast cancer, the effects of IGF-1 induced signaling cascades on miRNA expression in breast cancer has yet to be evaluated.
In the present study, we showed that miR-152 was downregulated in breast cancer. We aim to address the following questions: (1) Whether β-catenin is direct target of miR-152 in breast cancer; (2) whether miR-152 overexpression inhibits cell proliferation by inhibiting both β-catenin and PKM2 expression; (3) what is role of miR-152 in breast cancer resistance to paclitaxel treatment; (4) whether miR-152 is involved in IGF-1-induced β-catenin and PKM2 expression. The answer of these questions would provide new insights into a better understanding of the role of miR-152 in breast cancer development.

MiR-152 is downregulated in breast cancer tissues and cell lines.
To test miR-152 expression levels in breast cancer tissues, quantitative real time PCR (qRT-PCR) analysis was performed to detect miR-152 expression levels in 18 pairs of breast cancer tissues and matched adjacent normal breast tissues. The results showed that the expression levels of miR-152 were significantly decreased in breast cancer tissues than those in noncancerous tissues (Fig. 1A). Similarly, the suppression of miR-152 expression also occurred in breast cancer cell lines MCF7, T47D, MDA-MB-231, MDA-MB-453 when compared with immortalized breast epithelial MCF-10A cells by qRT-PCR (Fig. 1B). Our results confirmed that miR-152 was downregulated in both breast cancer tissues and cancer cell lines.

MiR-152 directly targets β-catenin in breast cancer cells.
To further understand molecular mechanism of miR-152 in breast cancer cells, several well-developed miRNA algorithms, such as TargetScan, miR-NApath, and TargetSearch, were employed to obtain a list of possible mRNA targets of miR-152. As a potential miR-152 binding site was predicted within the 3′-UTR of β-catenin mRNA, we constructed the 3′UTR reporters of β-catenin containing the miR-152 binding sites (WT) and corresponding mutant constructs (Mut) with 3 base pair substitution downstream of the pMIR-GLO luciferase reporters. HEK-293T (293 T) cells were co-transfected with miR-NC or miR-152 mimics, and reporter constructs, then the luciferase activity assay was performed after transfection for 48 hours. As shown in Fig. 1C, the luciferase activity of β-catenin WT 3′UTR was reduced by miR-152 overexpression, whereas the Mut constructs activity was not affected. In addition, to test whether miR-152 downregulates the expression of β-catenin, we transfected negative control (miR-NC) or miR-152 mimics into MCF7 and MDA-MB-231 cells, then detected β-catenin expression at the protein level. Immunoblotting results showed that miR-152 overexpression was sufficient to inhibit β-catenin protein expression (Fig. 1D). To further examine the reduction of β-catenin transactivation by miR-152, TCF/LEF-1 luciferase reporter analyses showed that force expression of miR-152 significantly inhibited β-catenin transactivation while downregulation of miR-152 by using anti-miR-152 promoted β-catenin transactivation (Fig. 1E). Thus, the suppression of β-catenin by miR-152 was functionally validated, suggesting that miR-152 specifically targeted β-catenin by binding its seed sequence to β-catenin 3′UTR region. Moreover, miR-152 significantly inhibited protein expression and transactivation of β-catenin.
MiR-152 represses cell growth by both targeting β-catenin and PKM2. To assess whether miR-152 suppressed β-catenin expression by directly targeting its 3′UTR region, we established MCF7 and MDA-MB-231 cell lines stably expressing miR-152 or miR-NC using lentiviral transduction. The expression levels of miR-152 were analyzed in these stable cell lines using qRT-PCR ( Fig. 2A). Then, the expression levels of β-catenin were determined by Immunoblotting. Cells overexpressed miR-152 showed low expression levels of β-catenin protein when compared with those of negative control cells. Moreover, miR-152-suppressed β-catenin expression was reversed by forced expression of β-catenin (Fig. 2B).
Importantly, cell proliferation suppressed by miR-152 was rescued by overexpression of β-catenin cDNA without 3′UTR in MCF7 and MDA-MB-231 cells when compared with control group (Fig. 2C). Then, colony formation assay was performed to test whether β-catenin forced expression is sufficient to restore miR-152-inhibited cell transformation. Our results showed that forced expression of β-catenin restored miR-152-inhibited colony formation (Fig. 2D).
Our previous studies explored that miR-152 inhibited the Warburg effect by directly targeting PKM2 14 ; while β-catenin had been reported to be an interacting protein of PKM2 in various cancer cells dependent on EGF stimulation. We predict that miR-152 simultaneously targets both PKM2 and β-catenin in breast cancer cells. Therefore, we performed studies to identify the known binding site of miR-152 in PKM2 3′UTR region. The luciferase activity of PKM2 3′UTR WT reporter, but not Mut reporter was significantly reduced in the cells transfected with miR-152 (Fig. 2E). Additionally, we observed that expression levels of PKM2 were downregulated in miR-152 stable expressing cells, and were restored in cells with forced expression of PKM2 without 3′UTR (Fig. 2F). Similarly, cell proliferation and colony formation were used to examine the involvement of PKM2 in miR-152-inhibited effects in the two different breast cancer cell lines. The results demonstrated that overexpression of PKM2 rescued miR-152-inhibited cell proliferation (Fig. 2G) and colony formation (Fig. 2H), confirming that PKM2 was also an important target of miR-152 for breast cancer cell proliferation and colony formation. Taken together, these data provide evidence that miR-152 may act as tumor suppressor in breast cancer cells by inhibiting cell growth and colony formation via inhibiting both β-catenin and PKM2.

MiR-152 sensitizes the breast cancer cells to paclitaxel treatment by decreasing β-catenin
and PKM2 expression. Aberrant miRNA expression has been implicated to affect the response to various chemotherapy in breast cancer. Paclitaxel remains a major drug for breast cancer treatment, especially for recurrent or metastatic breast cancer 36 . Based on the potential roles of β-catenin and PKM2 in paclitaxel resistance, we conducted CCK8 assay to measure cell viability of MCF7 and MDA-MB-231 stable cells treated with different concentrations of paclitaxel. It was observed that overexpression of miR-152 significantly sensitized the breast cancer cells to paclitaxel treatment; however, the chemotherapy sensitivity due to miR-152 overexpression was significantly reversed by overexpression of PKM2. Additionally, overexpression of β-catenin in these cells also partially reversed the paclitaxel sensitivity, demonstrating that both of PKM2 and β-catenin are involved in miR-152-induced paclitaxel treatment response and PKM2 played a more sensitive role in paclitaxel resistance of breast cancer (Fig. 3).

IGF-1 induces β-catenin and PKM2 expression and nuclear accumulation. Earlier study has
shown that EGF stimulation induced PKM2-β-catenin interaction and translocation into the nucleus in multiple types of cancer cells 37 . Given the important role of IGF-1 system in breast cancer cells and the ensure response of both MCF7 and MDA-MB-231 cells to IGF-1, we treated these cells with IGF-1 for 0, 8, and 24 hours. As shown in Fig. 4A, protein expression levels of β-catenin and cyclin D1 were increased after IGF-1 treatment. In addition, IGF-1 exposure for 24 hours also caused the significant induction of total PKM2 expression levels in both MCF7 and MDA-MB-231 cells (Fig. 4B). Notably, accumulation of β-catenin often results in β-catenin   Figure S1). Moreover, cell fractionation analysis was also used to assess the subcellular localization of β-catenin and PKM2 in MCF7 and MDA-MB-231 cells with IGF-1 treatment for 24 hours (Fig. 4C and D). These results indicated that IGF-1 induced the overexpression of β-catenin and PKM2 and promoted their nuclear translocation in breast cancer cells.

IGF-1 enhanced the binding of β-catenin and PKM2.
Above results showed that IGF-1 increased expression levels of PKM2 and β-catenin. We performed co-immunoprecipitation (Co-IP) analysis to further explore the interplay between β-catenin and PKM2 in response to IGF-1 stimulation. Our results showed a stronger interaction between β-catenin and PKM2 by IGF-1 treatment compared with the solvent (Fig. 5A). These results suggested that IGF-1 enhanced the expression levels of β-catenin and PKM2 and promoted their interaction, possibly leading to their nuclear translocation. In agreement with our earlier data, we hypothesize that IGF-1-enhanced binding of β-catenin and PKM2 may be responsible for β-catenin and PKM2 overexpression. To further dissect whether the IGF-1 dependent effects on β-catenin in MCF7 cells were mediated by PKM2, we performed Immunoblotting analysis to detect IGF-1-stimulated protein levels of β-catenin with PKM2 knockdown in MCF7 cells. MCF7 cells were transfected with lentivirus carrying sh-PKM2 and selected by puromycin to construct PKM2 knockdown cells. The Immunoblotting analysis results showed that PKM2 depletion significantly blocked IGF-1-induced β-catenin expression, suggesting that IGF-1-triggered expression of PKM2 could play an important role in IGF-1 associated β-catenin enrichment (Fig. 5B). In addition, TCF/LEF-1 luciferase reporter analyses also indicated the overexpression of PKM2 remarkably enhanced IGF-caused β-catenin transactivation (Fig. 5C). Overexpression of miR-152 significantly repressed PKM2, β-catenin, IGF-1R and IRS-1 expressions, but not GSK3β, which was not a target of miR-152 (Fig. 5D). In addition, IGF-1 incubation significantly induced expression levels of miR-152 in MCF7 cells, but not in MDA-MB-231 cells; while depletion of PKM2 or β-catenin partly abrogated the IGF-1-induced miR-152 upregulation in MCF7 cells, but not in MDA-MB-231 cells (Fig. 5E). We also interfered IGF-IR in breast cancer cell lines and the results showed that miR-152 expression level was inhibited by IGF-1R knockdown (Supplementary Figure S2).

Correlation of miR-152 suppression, β-catenin and PKM2 overexpression in clinical breast cancer tissues. β-catenin and PKM2 expression levels in 10 pairs of breast cancer tissues were detected by
Immunoblotting. Consistent with previous results in breast cancer cells, protein levels of β-catenin and PKM2 were much higher in most cases of cancer tissues than those in their matched adjacent normal tissues. The representative images was shown in Fig. 6A. Next, we examined β-catenin, PKM2 and miR-152 expression levels in 25 pairs of breast cancer specimens and their matched adjacent normal breast tissues (NBT), including 9 triple negative breast cancer (TNBC) tissues and 16 triple positive breast cancer (TPBC) tissues by using In situ hybridization and Immunohistochemical staining (Fig. 6B). Overexpression of β-catenin and PKM2 was observed both in TNBC and TPBC tissues, but not in NBT tissues, while miR-152 levels were much higher in NBT, and lower or non-expressed in TPBC and TNBC tissues, confirming that expression levels of miR-152 were inversely correlated with the β-catenin and PKM2 expressions levels in these tissues (Table 1). Pearson Correlation analysis showed the significant negative correlations between miR-152 and PKM2 (correlation = 0.−315, p = 0.026), and β-catenin (correlation = 0.−286, p = 0.044), respectively. Similarly, the correlation of PKM2 with β-catenin (correlation = 0.410, p = 0.003) was also significantly different (Fig. 5C,D,E). These results indicated the highly negative correlations between expression levels of miR-152 and its target proteins: β-catenin and PKM2 in human breast cancer tissues.

Discussion
Recent studies have indicated the involvement of different miRNAs in breast cancer proliferation, differentiation, apoptosis and chemotherapy by regulating the post-transcriptional expression of mRNAs 38,39 . MiR-152 downregulation is associated with breast cancer development. Overexpression of miR-152 significantly inhibited cell proliferation, colony formation and tumor angiogenesis by targeting IGF-1R and IRS1 in breast cancer. Furthermore, DNA methyltransferase 1 (DNMT1), a target of miR-152, mediated CpG island hypermethylation of miR-152 gene promoter that may suppress miR-152 expression as a feedback regulatory mechanism 10 . In addition, miR-152 plays an important role in breast cancer cell proliferation and angiogenesis via PKM2/NF-κB/EGR1/miR-152 complex in respond to IGF-1R activation 14 . In this study, our results identified β-catenin as a new target of miR-152.
Recently, increasing evidence showed that miRNAs can affect sensitivity of breast cancer to paclitaxel treatment. For example, miR-101 could increases paclitaxel sensitivity by suppressing MCL-1 expression in human triple-negative breast cancer 40 . MiR-520h is involved in paclitaxel chemosensitivity by modulating DAPK2 expression 41 . Both β-catenin and PKM2 were implicated to play the important roles in paclitaxel resistance of breast cancer [42][43][44] . Our data showed that miR-152 sensitized breast cancer cells to paclitaxel treatment by targeting β-catenin and PKM2 repression, suggesting a novel role of miR-152 in breast cancer paclitaxel resistance. Thus, the combination treatment of miR-152 with paclitaxel could provide a new strategy to overcome paclitaxel resistance in breast cancer in the future.
Although IGF-1 has been well demonstrated to enhance breast cancer cell proliferation, only a small subset of miRNAs are shown to be significantly regulated by IGF-1 treatment in breast cancer 45 . Recent study has suggested that IGF-1 induced PKM2 phosphorylation, thus resulting in PKM2 overexpression and nuclear translocation 46 . Our results showed that IGF-1 significantly increased the expression levels of β-catenin and PKM2, enhanced their interaction, and promoted their nuclear accumulation, thus leading to miR-152 transcriptional activation. Moreover, downregulation of PKM2 or β-catenin abrogated the IGF-1-induced miR-152 expression, suggesting that miR-152 was involved in a new IGF-1-mediated miR-152/PKM2/β-catenin regulatory circuit in breast cancer. However, IGF-1-induced upregulation of miR-152 expression was not observed in MDA-MB-231 cells. It is possible that the hypermethylation of CpG islands in miR-152 gene promoters were found in MDA-MB-231 cells, which may abrogate the binding of many transcription factors, such as EGR1 14 . Additionally, the lower miR-152 expression level in MDA-MB-231 cells may be related to the hypermethylation of miR-152 gene promoter.
Taken together, our results reveal the vital role of miR-152 as a tumor suppressor in human breast cancer by inhibiting both β-catenin and PKM2 expression levels at the post-transcriptional level, and point out the potential clinical application of miRNA-analog for the cancer treatment.   Table 1. The expression of miR-152, PKM2 and β-catenin in breast cancer patients. "−" means negative staining; the positive staining was gradually increased from "+" to "+++".

Materials and Methods
cancer cells were maintained in high glucose DMEM (Invitrogen) supplemented with 10% fetal bovine serum, 100 units/mL penicillin and 100 μg/mL streptomycin. MCF-10A was cultured in DMEM/Ham's F-12 medium supplemented with growth factors and 100 ng/ml Cholera Toxin. All cells were incubated in a 5% CO 2 at 37 °C. β-catenin antibodies were from Cell signaling technology (CST, Beverly, MA, USA); β-actin antibody were from sigma (St. Louis, MO, USA); GSK3β and PKM2 antibodies were from Signalway Antibody (SAB, College Park, MD, USA); laminB antibodies were from Epitomics (Burlingame, CA, USA). The IGF-1 cytokine was purchased from sigma.

Tissue samples. Tissues and pathological slides of breast cancer and matched adjacent normal breast tissues
were obtained from our tissue bank in Nanjing medical university. The tissue samples were collected immediately after the surgical removal, and snap-frozen in liquid nitrogen. The specimens had been fixed in formalin and embedded in paraffin before they were archived. None of the patients received preoperative chemotherapy.
Ethics statements. All  In situ hybridization. Slides were treated and hybridized with 10 pmol probe (LNA-modified and DIG labeled oligonucleotide) complementary to miR-152, according to the manufacturer's instructions. After incubation with anti-DIG-HRP Fab fragments conjugated to horseradish peroxidase, the hybridized probes were detected by incubating with 3,3′-diaminobenzidine (DAB) solution, and nuclei were counterstained with Carazzi's Haematoxylin. The assessment of signals was performed by two experienced pathologists in a blind manner.

Immunoprecipitation (Co-IP).
Proteins from 293 T cells transfected with PKM2 and β-catenin were extracted by RIPA lysis buffer. The protein extraction was incubated with indicated antibody (5-10 μg) for 16 hours at 4 °C followed by incubation with protein A/G agarose. The beads were washed five times with 1 ml of IP buffer. The samples were detected using Immunoblotting with indicated antibodies.
Luciferase Reporter Assay. The 3′-UTR regions of β-catenin and PKM2 containing predicted miR-152 seed-matching sites (wide type, WT; mutant type, Mut) were inserted into pMIR-GLO vector (Ambion, CA, USA). 293 T cells were seeded into a 24-well plate for luciferase assay. After cultured overnight, cells were cotransfected with the WT or Mut plasmid and equal amounts of miR-152 or miR-NC using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instruction. Luciferase assays were performed 24 hours after transfection using the Dual Luciferase Reporter Assay System (Promega, WI, USA).
Immunohistochemistry assay. For Immunohistochemistry (IHC), 4 μm sections of formalin-fixed and paraffin breast cancer tissues were divided into three groups (NBT, TNBT and TPBC) according to clinical diagnoses. After incubation with hydrogen peroxide, the sections were blocked for 1 hours with 10% goat serum and incubated with the indicated antibodies at 4 °C for 16 hours. After washing, the slides were incubated with HRP-conjugated secondary antibodies for 40 minutes (Proteintech). The antibody signals were detected using DAB regent. Sections incubated with the preimmune IgG were used as the negative control. Stained sections were examined under a light microscope (Zeiss).
Immunofluorescence assay. MCF7 and MDA-MB-231 cells were incubated with or without IGF-1. After incubation for 24 hours, Cells were fixed with 4% formaldehyde at 4 °C, and blocked with 1% bovine serum albumin (BSA) in PBST (PBS containing 0.05% Tween-20). Cells were immunostained with antibodies against PKM2 or β-catenin overnight. The antigen-primary antibody complex was detected using fluorescence isothiocyanate (FITC) secondary antibodies (Proteintech). Microscopic observation was performed under a fluorescence microscope (OLYMPUS IMAGING).
Statistical analysis. The data were analyzed by SPSS 16.0 and presented as the mean ± SEM. Student's test and ANOVA were performed to compare the differences. Pearson's correlation analysis was used to calculate the correlation between two groups. Differences were considered to be statistically significant at p < 0.05.