NEDD9 Inhibition by miR-25-5p Activation Is Critically Involved in Co-Treatment of Melatonin- and Pterostilbene-Induced Apoptosis in Colorectal Cancer Cells

The underlying interaction between melatonin (MLT) and daily fruit intake still remains unclear to date, despite multibiological effects of MLT. Herein, the apoptotic mechanism by co-treatment of MLT and pterostilbene (Ptero) contained mainly in grape and blueberries was elucidated in colorectal cancers (CRCs). MLT and Ptero co-treatment (MLT+Ptero) showed synergistic cytotoxicity compared with MLT or Ptero alone, reduced the number of colonies and Ki67 expression, and also increased terminal deoxynucleotidyl transferase dUTP nick end labeling- (TUNEL) positive cells and reactive oxygen species (ROS) production in CRCs. Consistently, MLT+Ptero cleaved caspase 3 and poly (ADP-ribose) polymerase (PARP), activated sex-determining region Y-Box10 (SOX10), and also attenuated the expression of Bcl-xL, neural precursor cell expressed developmentally downregulated protein 9 (NEDD9), and SOX9 in CRCs. Additionally, MLT+Ptero induced differentially expressed microRNAs (upregulation: miR-25-5p, miR-542-5p, miR-711, miR-4725-3p, and miR-4484; downregulation: miR-4504, miR-668-3p, miR-3121-5p, miR-195-3p, and miR-5194) in HT29 cells. Consistently, MLT +Ptero upregulated miR-25-5p at mRNA level and conversely NEDD9 overexpression or miR-25-5p inhibitor reversed the ability of MLT+Ptero to increase cytotoxicity, suppress colony formation, and cleave PARP in CRCs. Furthermore, immunofluorescence confirmed miR-25-5p inhibitor reversed the reduced fluorescence of NEDD9 and increased SOX10 by MLT+Ptero in HT29 cells. Taken together, our findings provided evidence that MLT+Ptero enhances apoptosis via miR-25-5p mediated NEDD9 inhibition in colon cancer cells as a potent strategy for colorectal cancer therapy.


MLT+Ptero Increased the Number of TUNEL-Positive Cells and ROS Production and Suppressed Ki67-Positive Cells in CRCs
To confirm whether the cytotoxicity of MLT+Ptero was induced by apoptosis, Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay was conducted in HT29 and SW480 cells, since TUNEL-positive staining implies a feature of apoptosis. The numbers of TUNEL-positive cells were significantly increased by MLT+Ptero in HT29 and SW480 cells compared to MLT or Ptero alone ( Figure 3A). Furthermore, MLT+Ptero attenuated the expression of Ki67 as a biomarker of cell proliferation in HT29 cells compared to the untreated control ( Figure 3B). Consistently, the apoptotic features, such as apoptotic bodies and cell shrinkages, were observed in Ptero-(40 µM) and MLT-(1mM) treated three CRCs ( Figure 3C). Also, MLT+Ptero significantly increased reactive oxygen species (ROS) production in HT29 or SW480 cells compared to MLT or Ptero alone ( Figure 3D,E).  HT29 and SW480 cells were exposed to MLT (1 mM) and/or Ptero (40 µM) for TUNEL staining. The fluorescent signals from fragmented DNA (green) and 4 ,6-diamidino-2-phenylindole. (DAPI) (blue) were visualized and photographed by a FLUOVIEW FV10i confocal microscopy. Magnification bar = 50 µm. Bar graphs represent quantification of TUNEL-positive cells (%). Data represent means ± SEM of triplicate samples. * p < 0.05, *** p < 0.001 vs. untreated control. (B) Effect of MLT+Ptero on Ki67 expression in HT29 cells. Immunofluorescence staining of proliferation marker Ki67 in HT29 cells. Nuclei were stained by DAPI (blue) stain and anti-rabbit Alexa Fluor 546 (red). ** p < 0.01, *** p < 0.001 vs. untreated control by one-way ANOVA test. (C) Effect of MLT+Ptero on apoptotic morphological changes in HT29, SW480, and HCT116 cells. Following exposure to MLT and/or Ptero in three colon cancer cells for 24 h, apoptotic morphology of the cells was observed in the cells under phase contrast microscope. (D,E) Effect of MLT+Ptero on ROS production in HT29 or SW480 cells. HT29 or SW480 cells were treated with MLT (1 mM) and/or Ptero (40 µM) for 24 h and then 10 µM Dichloro-dihydro-fluorescein diacetate (DCFH-DA) for 30 min at 37 • C. Fluorescence intensity was measured by Dichloro-dihydro-fluorescein diacetate (FACS) Calibur. Bar graphs showed quantification of ROS generation. Data represent means ± SD. * p < 0.05 versus untreated control (n = 2, one-way ANOVA, Tukey's test).

Inhibition of miR-25-5p Reduced the Antiproliferative and Apoptotic Effects of MLT+Ptero in CRCs
To validate the critical role of miR-25-5p in cytotoxicity and apoptosis by MLT+Ptero in CRCs, MTT assay, colony formation assay, and Western blotting were conducted in HT29, SW480, and HCT116 cells transfected with miR-25-5p inhibitor plasmid. Here, inhibition of miR-25-5p reversed the reduced number of colonies by MLT+Ptero in HT29 and HCT116 cells 2 week after MLT+Ptero co-treatment by colony formation assay ( Figure 7A,B). Consistently, inhibition of miR-25-5p blocked cytotoxicity by MLT+Ptero in HT29, SW480, and HCT116 cells transfected with miR-25-5p inhibitor plasmid compared to MLT or Ptero alone by MTT assay ( Figure 7C). Furthermore, Western blotting showed that suppression of miR-25-5p reduced PARP cleavage and the expression of Bcl-xL by MLT+Ptero in HT29 and SW480 cells ( Figure 7D).

Discussion
In the current work, the underlying apoptotic mechanisms of MLT and Ptero were investigated in HT29, SW480, and HCT116 cells in association with miR-25-5p mediated NEDD9 signaling. Herein, MLT+Ptero revealed synergistic cytotoxicity with combination index below 1 compared to MLT or Ptero alone in three CRCs, implying the potent synergy of MLT+Ptero. Furthermore, MLT+Ptero decreased the numbers of colonies of HT29 and SW480 cells and decreased expression of proliferation marker Ki67 in HT29 cells, demonstrating the antiproliferative potential of MLT+Ptero co-treatment compared to MLT or Ptero alone. Also, MLT+Ptero increased the number of TUNEL-positive cells and ROS production, cleaved PARP, and caspase 3 as features of apoptosis [44] compared to MLT or Ptero alone in CRCs, indicating the synergistic apoptotic effect of MLT+Ptero.
It is well documented that SOX10 suppresses Wnt/β-catenin signaling and reduces epithelial-mesenchymal transition (EMT) migration and invasion of tumor cells, and enhances apoptosis as a tumor suppressor [45], whereas SOX9 enhances the growth of lung adenocarcinoma [46] with poor prognosis of patients with colorectal cancer [47]. Likewise, overexpression of NEDD9 suggests poor prognosis in patients with colorectal cancer [43] and promotes migration and progression of colon cancer cells through Wnt signaling [41], while SOX9 is known to mediate NEDD9 in melanomas. Here MLT+Ptero activated SOX10, reduced the expression of SOX9 and NEDD9 in HT29 and SW480 cells, and also completely attenuated fluorescence of NEDD9 overexpression in HT29 cells. Conversely, NEDD9 overexpression reversed cytotoxicity and cleavages of PARP and caspase 3 by MLT+Ptero, implying the pivotal role of NEDD9 inhibition in the apoptotic effect of MLT+Ptero.
Interestingly, miR-25-5p binds directly to the 3 -UTR sequence of NEDD9, and inhibition of miR-25-5p reversed the decreased expression of NEDD9 by MLT+Ptero by immufluorescence as well as rescued downregulation of NEDD at mRNA level by MLT+Ptero in HT29 cells by RT-PCR, strongly indicating close interaction between miR-25-5p and NEDD9 signaling, which should be confirmed by further study in vivo and in vitro in the near future.

MTT Assay
Cell viability of Ptero and/or MLT was assessed in SW480, HT29, and HCT116 cells by MTT assay. Briefly, colon cancer cells (1 × 104 cells per well) were distributed onto 96-well microplate and exposed to various concentrations of MLT (0, 0.125, 0.25, 0.5, 0.75, 1, 1.25, 1.5 mM) and/or Ptero (0, 10, 20, 40, 60, 80, 100 µM) for one day. The cells were incubated with MTT (1 mg/mL) for 2 h, and then were exposed to MTT lysis solution overnight. Thereafter, optical density was measured by using a microplate reader (Molecular Devices Co., Silicon Valley, CA, USA) at 570 nm wavelength and the cell viability was calculated as a percentage of viable cells in Ptero-and/or MLT-treated group versus untreated control.

Colony Formation Assay
HCT116 and HT29 cells were seeded onto 6-well plates at a density of 103 cells per well in RPMI 1640 including 10% FBS for 24 h and then exposed to Ptero (40 µM) and/or MLT (1 mM) for 2 weeks. Then, colonies were stained with Diff-Quick solution (Sysmex, Kobe, Japan), washed once with PBS, and then the fields were photographed under a fluorescence microscope (AXIO observer A1, ZEISS, Oberkochen, Germany).

Observation of Apoptotic Morphological Features
Apoptotic morphology was observed one day after exposure to Ptero and/or MLT in HT29, SW480, and HCT116 cells under phase contrast microscope.

TUNEL Assay
The DeadEndTM TUNEL system kit was applied to detect cell death based on Roche's instructions (Roche Molecular Biochemicals, Mannheim, Germany). Briefly, SW480 or HT29 cells exposed to Ptero and/or MLT for 24 h were washed with cold PBS and fixed with 4% paraformaldehyde for 30 min. Fixed cells in permeabilization solution (0.1% sodium citrate and 0.1%Triton X-100) were incubated with TUNEL assay mixture for 1 h. Then TUNEL-stained cells were visualized by a FLUOVIEW FV10i confocal microscopy (Olympus, Tokyo, Japan).

Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR)
Total RNAs from HT29 colon cells were isolated using the QIAzol (Invitrogen) and 1 µg of total RNA was applied for making cDNA by Superscript Reverse Transcriptase and then was amplified by Platinum Taq polymerase with Superscript One Step RT-PCR kit. Primers sequences used were synthesized by Bioneer (Daejeon, Korea). The primers for NEDD9 and GAPDH cDNA detection are as follows: NEDD9, 5 -CGCTGCCGAAATGAATAT-3 , and 5 -CCCTGTGTTCTGCTCTATGACG-3 ; GAPDH, 5 -GCACCGTCAAGGCTGAGAAC-3 , and 5 -GGATCTCGCTCCTGGAAGATG-3 . PCR amplification steps were conducted based on Jung's paper [6]. The amplified products were separated on 2% agarose gel and RT-qPCR was carried out with the LightCycler TM instrument (Roche Applied Sciences, Indianapolis, IN, USA).

Immunofluorescence Assay
Colon cancer cells treated by Ptero (40 µM) and/or MLT (1 mM) for 24 h were fixed with 4% formaldehyde and were permeabilized in 0.1% Triton X-100, according to the paper [13]. The fixed cells were washed with 1X PBS, blocked with 2% BSA in 1X PBS for 30 min at room temperature (RT), and incubated with the specific antibodies of NEDD9, SOX10, and Ki67 (1:1000; Abcam, Cambridge, UK) overnight at 4 • C. After washing, the cells were incubated with Alex Fluor 489 goat mouse-IgG antibody (Invitrogen) and Alexa Fluor 546 goat rabbit-IgG antibody (1:1000) for 1 h at RT. After washing twice, the nuclei of the cells were stained with 4,6-diamidino-2-phenylindole (DAPI; Sigma) and then were visualized under a FLUOVIEW FV10i confocal microscope (Olympus). Images of NEDD9and SOX10-stained cells were taken by a Delta Vision imaging system (Applied Precision, Issaquah, WA, USA).

MicroRNA Microarray and Data Analysis
Based on the paper of [13], for control and test RNAs isolated from HT29 cells exposed to Ptero and/or MLT for 24 h, the syntheses of target miRNA probes and hybridization were conducted by using Agilent's miRNA Labeling Reagent and Hybridization kit. Total RNAs (100 ng each) were dephosphorylated, denatured, and incubated for 10 min at 100 • C, ligated with pCp-Cy3 mononucleotide and purified with MicroBioSpin 6 columns (Bio-rad, Hercules, CA, USA). After purification, denatured labelled probes were transferred onto assembled Agilent Human miRNA Microarray (Human miRNA Microarray Release, AXBK) and hybridized for 20 h at 55 • C in an Agilent Hybridization oven (Agilent Technologies, Santa Clara, CA, USA). The hybridized microarrays were washed, and hybridized images were scanned using Agilent's DNA microarray scanner and quantified with Feature Extraction Software (Agilent Technologies). All data were normalized (set measurements less than 0.01 to 0.01) and fold-changed probes (not less than 2.0-fold between test and control) were chosen and analyzed by using GeneSpringGX 7.3 (Agilent Technologies).

Statistical Analysis
The data values were expressed as means ± SD from at least three independent experiments. Student's t-test was performed for two group comparison, while the one-way analysis of variance (ANOVA) followed by a Turkey post-hoc test was carried out for multi-group comparison using GraphPad Prism software (Version 5.0, San Diego, CA, USA). The statistical difference between groups was determined, only when p-value was less than 0.05.

Conclusions
MLT+Ptero exhibited significant cytotoxicity, suppressed the proliferative activity and Ki67 expression, increased TUNEL-positive cells and ROS production, cleaved PARP and caspase 3, attenuated the expression of Bcl-xL, NEDD9, and SOX9, and also activated SOX10 compared to MLT or Ptero alone in colon cancer cells. Also, MLT+Ptero induced differentially expressed microRNAs and also upregulated miR-25-5p at mRNA level in HT29 cells. Conversely, miR-25-5p inhibitor or NEDD9 overexpression reversed cytotoxicity, decreased colony formation, and PARP cleavage by MLT+Ptero in CRCs. Taken together, our findings provide new insights that co-treatment of MLT and Ptero synergistically enhanced apoptotic effect via miR-25-5p mediated NEDD9 signaling in CRCs as a potent therapeutic strategy for colorectal cancer prevention or treatment.