Doxycycline inhibits breast cancer EMT and metastasis through PAR-1/NF-κB/miR-17/E-cadherin pathway

Doxycycline displays high efficiency for cancer therapy. However, the molecular mechanism is poorly understood. In our previous study, doxycycline was found to suppress tumor progression by directly targeting proteinase-activated receptor 1 (PAR1). In this study, microRNAs were found to be involved in PAR1-mediated anti-tumor effects of doxycycline. Among these miRNAs, miR-17 was found to promote breast cancer cell metastasis both in vivo and in vitro. Moreover, miR-17 could reverse partial doxycycline inhibition effects on breast cancer. Employing luciferase and chromatin immunoprecipitation assays, nuclear factor-kappaB (NF-κB) was found to bind miR-17 promoters. Furthermore, E-cadherin was identified as the target gene of miR-17. These results showed that miR-17 can resist the inhibitory effects of doxycycline on breast cancer epithelial–mesenchymal transformation (EMT) by targeting E-cadherin.


INTRODUCTION
Breast cancer is a prevalent malignancy and the leading cause of cancer death among women worldwide [1]. Various therapies are used for breast cancer treatments; these interventions include surgery, chemotherapy and targeted therapy [2][3][4][5][6]. In addition to its antimicrobial activity, doxycycline was reported to display strong efficacy for cancer therapy as an antimicrobial drug [7,8].
Successful breast cancer treatment relies on better understanding of molecular mechanisms involved in cancer progression. Protease-activated receptor 1 (PAR1) is a G protein-coupled receptor involved in metastatic and invasive cancer processes [12][13][14][15][16]. Our previous studies indicated that in PAR1-dependent manner, doxycycline possesses higher inhibition ability in lung cancer and breast cancer [8,11].
miRNAs comprise a class of small noncoding RNAs modulating gene expression at the post-transcriptional level by repressing translation and promoting destabilization of target mRNAs [17][18][19][20][21]. In the past ten years, thousands of miRNAs were identified [22]. Importance in tumor genesis and progression of miRNAs attracted the attention of scientists and considered for funding by the government. In the present studies, doxycycline repressed miR-17 expression [23][24][25][26]. Ectopic expression of miR-17 partially reverses inhibitory effect of doxycycline on breast cancer invasion and tumor genesis. Nuclear factor NF-κB mediates regulation of miR-17 by doxycycline [13,[27][28][29][30]. Luciferase activity assay indicated that miR-17 directly targets E-cadherin and represses its expression. Therefore, doxycycline inhibits breast cancer invasion and tumor genesis through PAR1/ NF-κB/miR-17/E-cadherin pathway. Such results provide deeper understanding of doxycycline and importance of miRNA in cancer therapy.

PAR1 is highly expressed in breast cancer
To verify PAR1 expression in breast cancer, 10 pairs of clinical specimens and cell lines were used. Western blot analysis showed that PAR1 protein levels were higher in tumor tissues than those in adjacent tissues ( Figure 1A). PAR1 expression in normal breast tissues and breast cancer specimens was investigated and the results showed that the expression levels of PAR1 in breast cancer tissues were higher than those in non-tumor tissues. Images were taken from the Human Protein Atlas (http://www. proteinatlas.org) online database ( Figure 1B) [31,32]. PAR1 mRNA expression levels in normal vs tumor tissues of the breast were detected and results were shown in Figure 1C. Quantitative reverse transcription Polymerase Chain Reaction (qRT-PCR) was applied to test PAR1 expression in breast cancer cell lines. Results showed that PAR1 was expressed the highest in MDA-MB-231 cells ( Figure 1D). These high PAR1 expression cell lines were more sensitive to doxycycline compared with low PAR1 expression cells ( Figure 1E).

Doxycycline inhibits breast cancer cell epithelialmesenchymal transition (EMT) through PAR1/ NF-κB signaling pathway
To verify whether doxycycline plays its inhibitory function in breast cancer cells through PAR1, thrombin was used to activate PAR1 in MCF-7 cells, E-Cadherin (a classical and most well-studied member of the cadherin superfamily, acting as a canonical epithelial marker) and Vimentin were detected with immunofluorescence assay [33]. SEM and immunofluorescence results indicated that doxycycline could promote cell junction and inhibit EMT, whereas thrombin can counteract this inhibitory effect in MCF-7 cells (Figures 2A and 2B). Figure 2A showed that doxycycline treated cells formed a reduction in lamellipodia and filopodia compared with the control group. In doxycycline and thrombin treated group, the cell morphology was similar as control group, cells change from epithelial phenotype to mesenchymal phenotype. To analyze the mechanism of doxycycline during EMT progression in breast cancer, downstream signaling pathways of PAR1 were analyzed ( Figure 2C). Western blot analysis revealed that PAR1/AKT/NF-κB signal pathway changes after doxycycline and thrombin treatments.

NF-κB directly promotes expression of miR-17
Using microarray expression profiles, we obtained differentially expressed microRNAs in MCF-7 cells after treatment with doxycycline alone or doxycycline with thrombin ( Figure 3A). Then, miRNA-specific qRT-PCR was performed to validate candidate miRNAs. Results revealed three miRNAs that are more sensitive to doxycycline, miR-17, miR-10a, and miR-569 ( Figure 3B). Transwell and wound healing assay results showed that all three miRNAs can promote cell migration and invasion; miR-17 presented the most remarkable effect ( Figures 3C  and 3D). To further explore whether this doxycycline effect is mediated by miR-17, a luciferase reporter assay was performed to evaluate effects of NF-κB on miR-17 promoter ( Figures 3E and 3F). Results showed that NF-κB directly binds to promoter of miR-17 and unregulated its transcription ( Figure 3G).

miR-17 counteracts partial doxycycline inhibitory effect on breast cancer cells
To verify whether doxycycline suppressor function in breast cancer happens through miR-17, a rescue experiment was performed. MCF-7 and MDA-MB-231 cells were treated with doxycycline alone or doxycycline with miRNA mimics. In miR-17 transfected group, the relative miR-17 expression level was higher than that in control group, both in MCF-7 and MDA-MB-231 cells, this result indicated that miR-17 expression was up-regulated after miR-17 mimic transfection ( Figure 4A). In SEM (scanning electron microscopy) analysis, miR-17 up-regulation drove cell morphology to change from epithelial phenotype to mesenchymal phenotype and then attenuated cell junction in breast cancer cells ( Figure 4B). Wound healing and transwell assays indicated that miR-17 upregulation can restore doxycycline-inhibited cell invasion and migration ( Figures 4C and 4D). In doxycycline-pretreated cells, miR-17 implication can also upregulate Vimentin and downregulate E-cadherin expression levels. Therefore, miR-17 can restore doxycycline-inhibited EMT in breast cancer cells ( Figure 4E).

miR-17 directly targets E-cadherin
To further test the exact mechanism of miR-17 in doxycycline regulation of breast cancer, downstream targets of miR-17 were predicted and verified. In luciferase reporter assay, miR-17 knock-down reduced miRNA binding to E-cadherin untranslated region (UTR) and increased luciferase activity ( Figure 5A). qRT-PCR and Western blot results suggested that ectopic E-cadherin expression can reverse miR-17-induced repression of E-cadherin expression ( Figures 5B and 5C).
Wound healing and transwell assays confirmed that ectopic E-cadherin expression counteracts the effect of miR-17 on breast cancer cell invasion and migration ( Figures 5D and 5E). Immunofluorescence results revealed that E-cadherin reversed the effect of miR-17 on breast cancer cell EMT ( Figure 5F).

miR-17 reverses inhibition effect of doxycycline on tumor progression in vivo
To investigate the role of miR-17 in vivo, MCF-7 cells were inoculated subcutaneously in mouse axillae. Mice were sacrificed after treatment with doxycycline alone or doxycycline with miR-17 for 28 days. In mice treated with doxycycline alone, tumor volumes were smaller than those in control group, whereas tumor volumes rebounded in mice treated with miR-17 and doxycycline ( Figures 6A and 6B). Doxycycline inhibit lung metastasis, while miR-17 restoration can promote lung metastasis of breast cancer ( Figure 6C). qRT-PCR and IHC were performed to analyze expressions of miR-17 and EMT markers in solid tumors. Results indicated that miR-17 expressions were reduced and E-Cadherin expressions were up-regulated after doxycycline treatment ( Figure 6D and 6E). Immunohistochemical staining for E-cadherin results shown that the E-cadherin were highly expressive in doxycycline treated group compared with the untreated group. miR-17 can promote EMT progression, up-regulate Vimentin and reduce E-cadherin expressions ( Figure 6E). These results imply that miR-17 restoration can promote      Compared with the mock group, in Doxycycline treatment groups, NF-κB and Vimentin were stainted lighter, whereas E-Cadherin was stained darker. In the Doxycycline/miR-17 treatment group, however, NF-κB and Vimentin were stained slightly lighter than those in mock group but darker than those in Doxycycline treatment groups while E-Cadherin was stained darker than that in mock group and similar with or slightly lighter than those in Doxycycline treatment groups. IHC results showed that miR-17 can promote EMT progression, upregulate Vimentin, and repress E-cadherin expression. (F) Mechanism of doxycycline inhibits tumor progression. Results show that doxycycline inhibits tumor progression through PAR-1/NF-κB/miR-17/E-cadherin pathway. The HE images were 100×, IHC images were 400×. Results were obtained from three independent experiments, with each experiment performed in triplicate; error bars represent standard deviation ( * P < 0.05, ** P < 0.01). Data are represented as mean ± SEM. breast cancer cell metastasis through EMT regulation. Figure 6F displays the mechanism of doxycycline inhibits tumor progression. Doxycycline inhibits PAR1 and downstream NF-κB/miR-17/E-cadherin pathway to suppress tumor progression.

DISCUSSION
As a member of the tetracycline family, doxycycline affects angiogenesis and metastasis in tumor progression [34,35]. Previous study implied that doxycycline could induce EMT by regulating TMPRSS2 in LNCaP cells. While other researches demonstrated that doxycycline could inhibit the progression of cancer stem cell phenotype and EMT through FOXM1 and IκB/NF-κB signal pathway. To explore the molecular mechanism of anti-tumor effects of doxycycline, PAR1 was identified as a direct target in our previous study. PAR1 belongs to G protein-coupled receptors, which are involved in metastatic and invasive cancer processes [36]. Critical effects of PAR1 in tumor occur through multiple signal pathways. Malik supported that NF-κB activation is induced by PAR1 [13]. NF-κB is a well-known transcriptional factor that is important in tumor progression and regulates various gene expressions, including those of protein-coding genes and non-coding RNAs [13,37]. We observed that doxycycline inhibits NF-κB via PAR1.
miR-17 belongs to NF-κB-regulated miRNAs driving key physiological responses during tumor genesis and proliferation. In breast cancer, miR-17 is overexpressed and is associated with cell metastasis and proliferation [24]. In the present study, miR-17 was highly expressed in breast cancer cells and resisted doxycycline inhibition of cell migration and invasion. miR-17 also directly targets E-cadherin and represses its expression at post-transcription level [38]. E-cadherin dysregulation is correlated with EMT progression. Down-regulation of E-cadherin contributes to cancer progression by increasing proliferation and invasion. In the present work, miR-17 directly decreased E-cadherin and contributed to migration and invasion of breast cancer cells.
In summary, the current work indicated that inhibitory effects of doxycycline on breast cancer cell occur through a miR-17-dependent pathway. Both in vitro and in vivo experiments showed that binding of doxycycline to PAR1 leads to NF-κB inactivation, which subsequently down-regulates miRNA and up-regulates E-cadherin. Up-regulation of E-cadherin inhibits EMT progression and decreases invasion ability of breast cancer cells. These results support a model, in which repression of PAR1 by doxycycline upregulates E-cadherin expression by inactivating NF-κB and miR-17, thereby inhibiting the invasion of breast cancer cells. The present study provides a deeper understanding of the molecular mechanism of doxycycline in tumor regulation and may assist cancer therapies using tetracycline.

Scanning electron microscopy (SEM)
Cells were grown on climbing films and treated with doxycycline (1 μM). The same volume of ddH2O containing the same percent of DMSO was added in the control group. After 24 h, cells were fixed and dehydrated in acetone/isoamyl acetate (1:1) and dried with a gradient concentration of acetonitrile. After coating with gold, cells were photographed using a scanning electron microscope (LEO 1530 VP, Germany).

Matrigel invasion assay
A total of 5 × 10 4 cells were seeded in top-chamber inserts coated with matrigel (BD Biosciences) with RPMI-1640 containing 2% FBS. The bottom chamber was filled with 500μL Dulbecco's modified Eagle's medium (DMEM)containing 10% FBS. Doxycycline (1 μM) were added to the upper chamber. After 24 h of incubation, inserts were washed three times with 1×phosphatebuffered saline (PBS) gently and the cells transferred through the filter membrane of the chambers were fixed in 4% paraformaldehyde (precooled at 4°C) and stained with 0.1% crystal violet. The passed cells were counted under a microscope (Nikon, Japan).

Wound healing assay
Cells were grown in 24-well culture plates at a density of 2×10 5 cells/well. 24 h later, a wound was made www.impactjournals.com/oncotarget in the center of the well. After treatment with doxycycline (1μM) and mimics, wound images were photographed after 24 and 48 h using a light microscope (Nikon, Japan).

Murine Xenograft Model
Six-week-old female BALB/c-null mice were used to test the effects of doxycycline and miR-17 on breast cancer. Animal experiments were conducted in accordance with National Institutes of Health Animal Use Guidelines. All experimental protocols were approved by the Institutional Animal Care and Use Committee at Tianjin International Joint Academy of Biomedicine accordance. A total of 1 × 10 7 cells were injected subcutaneously to nude mice. When tumor volume reached approximately 50 mm 3 , mice were treated with 30 mg/kg of doxycycline and miR-17 mimics. The nude mice were monitored at three-day intervals for tumor appearance. Tumor size was calculated using the following equation: tumor size = width 2 × length × 0.5. All mice were euthanized after four weeks of treatments.

Immunohistochemical analysis
Tumor tissues from mice were fixed in 4% paraformaldehyde and embedded in paraffin. Thick slices were cut into 4 μm samples. Tissues were deparaffinized with xylene and dehydrated with ethanol of decreasing concentrations. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide. Microwave antigen repair technique was utilized to retrieve antigens. After blocking, primary antibodies were incubated overnight at 4 °C. After incubation for 1 h, tissue sections were washed with PBS and incubated with biotinylated goat anti-mouse IgG antibody (Zhongshan Biology Technology Co., Ltd., Beijing, China) at 37 °C for 30 min. Following staining with 3,3′-diaminobenzidine/H 2 O 2 and hematoxylin and eosin, sections were cleared and mounted for observation and analysis.

Chromatin immunoprecipitation assay
A total of 1 × 10 7 cells were plated in 15 cm culture dish. After treatment, cells were cross-linked using 1% formaldehyde (Sigma-Aldrich). Fixed cells were lysed, and chromatin was sheared using sonication. Chromatin fraction was incubated with NK-κB antibody overnight at 4 °C. DNA was extracted and used in polymerase chain reaction (PCR) amplification with miR-17-specific primers.

Luciferase activity assay
Cells were seeded in 96-well plates at a final concentration of 1 × 10 3 cells per well and maintained at 37 °C in 5% CO 2 , 24 h later, miRNA mimics were cotransfected in luciferase reporter plasmids containing miR-17 binding sites. 48 h after transfection, the medium was removed and the cells were lysed directly in the wells. Then 30μL cell lysate was added into the well of a white 96-well plate, followed by the addition of the firefly luciferase and the firefly luciferase activity was measured with a Dual-Glo ® luciferase assay system (Promega, Madison, WI, USA). Afterwards, stop reagent and the Renilla luciferase were added and the Renilla luciferase activity was determined by the same instrument. The relative luciferase activity was the ratio of the firefly luciferase activity to the Renilla luciferase activity. Mutations of promoter region of miR-17 were reference literature [39].

Quantitative reverse transcription PCR (qRT-PCR) analysis
After treatment, cells were collected, and total RNAs were isolated using TRIzol (Sigma, St. Louis, MO, USA) per manufacturer's instructions. Approximately 2 μg RNAs were used in RT using miRNA-specific RT primers. cDNAs were then used to examine miR-17 expression. U6 small nuclear RNA was used as control. miR-17 expression levels were quantified using 2 -ΔΔCt methods. Each experiment was performed in triplicate.

Statistical analysis
All results were presented as mean ± standard deviation. Values were analyzed using Student's t-test, ANOVA, and multivariate statistical analysis. Level of statistical significance was set at P < 0.05.

CONFLICTS OF INTEREST
There are no conflicts of interest in this study.