Stimulation of MMP-9 of oral epithelial cells by areca nut extract is related to TGF-β/Smad2-dependent and –independent pathways and prevented by betel leaf extract, hydroxychavicol and melatonin

Background: There are 200-600 million betel quid (BQ) chewers in the world. BQ increases oral cancer risk. Matrix metalloproteinase-9 (MMP-9) is responsible for matrix degradation, cancer invasion and metastasis. Whether areca nut extract (ANE), a BQ component, stimulates MMP-9 secretion, and the related signaling pathways awaits investigation. Results: ANE (but not arecoline) stimulated MMP-9 production of gingival keratinocytes and SAS cancer epithelial cells. ANE stimulated TGF-β1, p-Smad2, and p-TAK1 protein expression. ANE-induced MMP-9 production/expression in SAS cells can be attenuated by SB431542 (ALK5/Smad2 inhibitor), 5Z-7-Oxozeaenol (TAK1 inhibitor), catalase, PD153035 (EGFR tyrosine kinase inhibitor), AG490 (JAK inhibitor), U0126 (MEK/ERK inhibitor), LY294002 (PI3K/Akt inhibitor), betel leaf (PBL) extract, and hydroxychavicol (HC, a PBL component), and melatonin, but not by aspirin. Conclusions: AN components contribute to oral carcinogenesis by stimulating MMP-9 secretion, thus enhancing tumor invasion/metastasis. These events are related to reactive oxygen species, TGF-β1, Smad2-dependent and –independent signaling, but not COX. These signaling molecules can be biomarkers of BQ carcinogenesis. PBL, HC and melatonin and other targeting therapy can be used for oral cancer treatment. Methods: ANE-induced MMP-9 expression/secretion of oral epithelial cells and related TGF-β1, Smad-dependent and –independent signaling were studied by MTT assay, RT-PCR, western blotting, immunofluorescent staining, and ELISA.

AGING polyphenols (flavonol, tannin), minerals (Cu, Fe etc.), carbohydrate, fat, protein, crude fibers etc. [1,2]. ANE, arecoline, reactive oxygen species generated during oxidation of ANE, and the AN-derived nitrosamines are considered to be the possibly carcinogens. They exhibit genotoxicity, mutagenicity and cell transformation capacities in different assay systems [1,2]. Clinically, BQ chewing increases the risk of oral leukoplakia, oral lichenoid lesions, oral submucous fibrosis (OSF) and oral squamous cell carcinoma (OSCC) [1,2]. BQ ingredients are involved in the initiation and promotion of oral cancer by induction of DNA damage, chromosomal aberration, tissue inflammation, fibrosis and malignant transformation [1,3]. However, limited information is known about the BQ components in tumor invasion, metastasis and progression.
Matrix metalloproteinases (MMPs) play important roles in tissue inflammation, tumor invasion and metastasis, by degradation of extracellular matrix [4,5]. OSCC expresses higher level of MMP-2 and MMP-9 [6]. It is intriguing to know whether BQ components may affect MMPs expression/production and contribute to oral carcinogenesis. Recently, areca nut extract (ANE) activates MMP-9, but not MMP-2 expression in gingival epithelial cells, that can be inhibited by NF-kB inhibitor and curcumin [7]. ANE also stimulates MMP-9, but decreases tissue inhibitor metalloproteinase-1 (TIMP-1) and TIMP-2 secretion of SAS tongue cancer epithelial cells [8]. Salivary MMP-9 levels and MMP-2 and MMP-9 mRNA expression in OSCC are markedly increased and related to lymph node metastasis [9]. All the above reveal the importance of MMPs in oral carcinogenesis.
To know whether BQ chewing and AN components can promote cancer progression, invasion and metastasis, it is interesting to know whether AN components may stimulate MMP-9 expression in oral epithelial cells and the role of TGF-β1/Smad2-dependent and Smadindependent (TAK1 and other related signal transduction) pathways. Moreover, one clinically critical question is whether including of PBL into BQ may enhance or decrease its carcinogenicity that is important for development of health policy and disease prevention for the country. PBL contains chemicals mainly hydroxychavicol (HC), eugenol, chavicol and carotene [2], and are shown to exhibit potential anti-mutagenic and anticarcinogenic effect [2]. PBL extract and HC are found to have anti-oxidant, anti-inflammatory and anti-platelet effect possibly via scavenging ROS and inhibition of cyclooxygenase (COX) [21,22]. Moreover, melatonin has been shown to have anti-cancer effects by mitigating the initiation, progression and metastasis of cancer development and growth possibly via receptor-dependent and -independent manners [23,24]. Melatonin is shown to scavenge reactive oxygen species (ROS)/redox regulation, inhibition of signaling molecules, increase sensitivity of cancer cells to chemotherapeutic drugs, modulating of non-coding RNA, control of angiogenesis and apoptosis etc. [23,24]. Since melatonin is present as natural hormone in human body and used clinically for insomnia/sleep improvement in different conditions [24], its preventive effect on ANE-induced toxicity is useful.
Since ROS and COX are associated with ANE toxicity and tissue inflammation is consistently found in tissues of OSF and OSCC [1-3, 10, 11], clarifying the preventive efficacy of betel leaf (piper betle leaf [PBL], hydroxychavicol (HC), aspirin and melatonin, on the ANE-induced alterations in oral mucosal cells is valuable for cancer prevention and treatment. Results of this study can be helpful for targeting therapy of OSCC in the future.

Effect of ANE and arecoline on MMP-9 mRNA expression of SAS cells
We recently found the stimulation of MMP-9 production of oral epithelial cells by ANE at non-cytotoxic concentrations as analyzed by enzyme-linked immunosorbent assay (ELISA). On the other hand, arecoline showed no marked stimulatory effect on MMP-9 production of oral epithelial cells, even decrease MMP-9 production possibly due to cytotoxicity at concentrations of 0.4 and 0.8 mM [8]. Accordingly, ANE also stimulated MMP-9 secretion of gingival keratinocytes (GK) (data not shown). ANE also induced MMP-9 mRNA expression of oral epithelial cells as analyzed by reverse transcriptionpolymerase chain reaction (RT-PCR) ( Figure 1A). But arecoline showed no marked stimulatory effect on MMP-9 expression ( Figure 1B). Accordingly, ANE stimulated the MMP-9 secretion of SAS cells after 3 days of exposure ( Figure 1C), whereas arecoline showed slight decrease of MMP-9 secretion ( Figure 1D). At tested concentrations, ANE showed no evident changes on viability of SAS cells ( Figure 1E). Arecoline suppressed the viability of SAS cells at concentrations higher than 0.4 mM ( Figure 1F).

Role of TGF-β1 and Smad2 signaling in ANEinduced MMP-9 secretion of GK and SAS cells
ANE stimulated TGF-β1 protein expression and this event can be inhibited by SB431542, an ALK5/Smad2/3 signaling inhibitor (Figure 2A). ANE induced the Smad2 phosphorylation in both SAS cells and primary  Figure 2B, 2C). SB431542 markedly attenuated the ANE-induced Smad2 phosphorylation in SAS cells and GK ( Figure 2D, 2E). Intriguingly, SB431542 also prevented the ANE-induced MMP-9 secretion of both SAS cells and GK ( Figure 2F, 2G). In this experiment condition, SB431542 showed little effect on the viability of SAS cells ( Figure 2H).

Effect of Catalase, PD153035 and AG490 on ANEinduced MMP-9 production
ANE has been shown to stimulate ROS production and glutathione (GSH) depletion [10], we therefore tested and found that catalase attenuated the ANEinduced MMP-9 production of oral epithelial cells ( Figure 4A). Since ROS has been shown to be important signaling molecules, we further tested the possible EGFR and JAK signaling pathways in this event. Intriguingly, we found that PD153035 (a EGFR antagonist) evidently attenuated the ANEinduced MMP-9 production of SAS cells ( Figure 4B). In addition, AG490 (a JAK inhibitor) was also effective to inhibit the ANE-induced MMP-9 production of SAS cells ( Figure 4C). Accordingly, PD153035 and AG490 also attenuated the ANEinduced MMP-9 mRNA expression of SAS cells ( Figure 4D, 4E). At these experimental conditions, catalase, PS153035 and AG490 showed no marked effect on the viability of SAS cells ( Figure 4F, 4G, 4H), as analyzed by 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) assay (a mitochondrial enzyme activity marker).

Effect of betel leaf, hydroxychavicol and aspirin on ANE-induced MMP-9 production
Betel leaf (piper betle leaf, PBL) extract and hydroxychavicol (HC) has been shown to exhibit antiplatelet, antioxidant and even anti-inflammatory properties and inhibit cyclooxygenase (COX) enzyme activities [21,22]. Three PBL components, hydroxychavicol (HC), catechol and catechin showed no marked stimulatory effect on MMP-9 production of oral epithelial cells (data not shown). Interestingly PBL extract and HC effectively inhibited the ANE-induced MMP-9 production of SAS epithelial cells ( Figure 6A, 6B). But aspirin (a cyclooxygenase [COX] inhibitor) showed little effect on ANE-induced MMP-9 production of SAS cells ( Figure 6C). Accordingly, PBL extract attenuated the ANE-induced MMP-9 mRNA expression of SAS cells ( Figure 6D). In this experimental conditions, PBL and HC showed little cytotoxicity to SAS cells ( Figure 6E, 6F)

Effect of melatonin on ANE-induced MMP-9 mRNA expression and protein secretion
Pretreatment and co-incubation by melatonin (100 and 250 μg/ml) effectively prevented the ANE-induced MMP-9 mRNA expression and MMP-9 secretion into culture medium ( Figure 7A, 7B). Melatonin showed little effect on ANE-induced changes of cell viability of SAS cells ( Figure 7C).

DISCUSSION
OSCC has been ranked as top 5 prevalent cancer in male of Taiwan. This can be due to oral habits such as BQ chewing, smoking and alcohol consumption in oral cancer patients. The 5-year survival rate of single primary OSCC patients is about 68% in Taiwan, a BQ chewing endemic area [25]. MMPs are involved in tumor invasion and metastasis and thus the prognosis/outcome of OSCC patients. Increased MMP-9 mRNA and protein expression are also reported in OSCC tissues [6,9]. All these results suggest the contributory role of MMP-9 in oral carcinogenesis, tumor invasion and metastasis. Extracellular matrix homeostasis is important for wound healing, modulating keratinocyte behaviors, promoting keratinocyte migration, granulation tissue remodeling, and cancer [26]. Extracellular matrix turnover is tightly controlled by matrix metalloproteinases (MMP-2, MMP-9, collagenases etc.) and tissue inhibitors of metalloproteinases (TIMPs). TIMP-1 and TIMP-2 may inhibit the enzyme activity of MMP-2 and MMP-9 [27]. Degradation of basement membrane and interstitial connective tissues that contains mainly fibrillar collagens, and other ECM proteins such as AGING  proteoglycans, elastin, laminin, fibronectin etc. by enzymatic proteases around the tumor cells is an essential step for cancer invasion and metastasis [27,28]. Over-expression of MMPs is associated with tumor growth, relapse, metastasis and overall survival in breast and liver cancers [27]. Since BQ chewing is one major risk factor of chewer's mucosa, leukoedema, oral leukoplakia, ulcer and oral carcinogenesis [1,29], it is interesting to know whether BQ components may stimulate MMP-9 expression and secretion in oral mucosal cells. In addition to stimulation of MMP-9 secretion in SAS cancer epithelial cells [8], ANE, but not arecoline, also stimulates MMP-9 mRNA expression of SAS cells and MMP-9 secretion of #denotes statistically significant difference when compared with ANE-treated group. AGING primary oral keratinocytes. Stimulation of MMP-9 production and expression by ANE indicates that BQ chewing possibly may contribute to the pathogenesis of chewer's mucosa, leukoedema, ulcer, oral carcinogenesis and cancer progression by induction of MMP-9 and thus promote tumor invasion and metastasis. Since arecoline does not stimulate MMP-9 production, presence of other components in ANE that may stimulate MMP-9 production should be further clarified.
Recently AN components are reported to stimulate TGF-β, IL-1α and IL-1β [12,13,15,16], that may potentially induce downstream transforming growth factor β-activated kinase-1 (TAK1) signaling. Interestingly we found that ANE increased TGF-β1  protein expression as well as the downstream p-Smad2 and p-TAK1 protein expression and signaling. SB431542 (an ALK5/Smad2/3 signaling inhibitor) attenuated the ANE-induced p-Smad2 and TGF-β1 protein expression as well as MMP-9 secretion. Moreover, 5Z-7-Oxozeaenol (a selective TAK1 inhibitor) prevented the ANE-induced MMP-9 mRNA expression and protein secretion of SAS and GK. These results suggest the involvement of both Smaddependent and TAK1 signaling in ANE-induced MMP-9 activity of oral mucosal cells. Similarly,  found the regulation of MMP-9 by TAK1 in gastric cancer cells [30]. Targeting and knockdown of TAK1 inhibits chemokine receptor 7, MMP-9, IL-8, COX-2, tumor growth and metastasis of breast cancer [31]. Similarly, TGF-β1 is shown to stimulate MMPs in hepatocellular carcinoma cells via ALK5/Smad and PI3K/Akt signaling [32,33]. These results raise the possible targeting TAK1 and ALK5/Smad molecules in future prevention and treatment of oral cancer.
Reactive oxygen species (ROS) production shows reciprocal interaction with TAK1 activation. Whereas ROS activate TAK1/AMPK signaling, TAK1 activation also stimulates further ROS production in different kind of cells [17,34]. Interactions and crosstalk between TGF-β/TAK1 and EGFR and JAK signaling are reported [19,35]. Interestingly we also found the crosstalk between TAK1 and Smad2 signaling in ANEinduced events in oral epithelial cells (data not shown). BQ components may stimulate ROS production of oral mucosal cells [10,13,22]. Antioxidants such as Nacetylcysteine (NAC) and catalase are able to prevent the ANE-and arecoline-induced toxic events such as COX-2 expression, PGE2, IL-1α, cytotoxicity etc. [13,22] as well as MMP-9 secretion of SAS cancer epithelial cells in this study. ANE can activate EGFR [12], and inhibition of EGFR and JAK signaling by PD153035 and AG490 also suppressed the ANEinduced MMP-9 production of SAS cancer cells. Similarly, ROS may stimulate COX-2 expression of synovial fibroblasts via TAK1 signaling [17]. Moreover, inhibition of EGFR and JAK may potentially prevent the ANE-induced ADAM17, IL-1α, PGE2 production, COX-2 expression, ADAM9 maturation, and the ANE-induced decline in keratin 5 and 14 [12,13]. EGFR overexpression in head/neck SCC is reported and associated with tumor stages and tumorfree survival [36]. Further, alteration of JAK2/Stat3 related genes' expression in late stage OSCC is also noted [37]. These indicate that ROS scavengers, EGFR and JAK targeting molecules can be potentially used for future therapy of OSCC.
PBL is a popularly-used additive in BQ. Whether addition of PBL into BQ is beneficial or harmful to oral and systemic health is an intriguing health issue. PBL components and HC are demonstrated to exhibit hepatoprotective, immunomodulatory, anti-platelet, antiinflammatory and antioxidative effects [12,21,22]. In addition, PBL and HC have potential cancer treatment effects toward various cancers [41,42]. However, the anticancer effect of PBL is not clear. In this study, PBL extract and one major PBL component -HC (at non-toxic concentrations) prevent the ANE-induced MMP-9 production in oral cancer epithelial cells, suggesting attenuation effects to cancer invasion and metastasis. This is possibly due to antioxidant or other effects, but not through inhibition of COX enzyme activity, because aspirin showed little effect on ANE-induced MMP-9 production in this study. More studies are necessary to further clarify the pharmacological effects of PBL and HC.
Melatonin is secreted from pineal gland and has physiological regulation of sleep and circadian rhythms [43]. It is suggested for use in prevention and treatment of various cancer including head and neck SCC, possibly via effects on angiogenesis, cancer cell proliferation, apoptosis, matrix turnover, matrix metalloproteinase activity and epithelial-mesenchymal transition that are important in primary tumor and cancer metastasis [43,44]. To further clarify the anticarcinogenic effect of melatonin, we found that melatonin prevents the ANE-induced MMP-9 mRNA expression/secretion of oral cancer cells. Why melatonin inhibits ANE-induced MMP-9 awaits further investigation. The potential mechanisms are blocking of signaling transduction molecules, its antioxidant property and prevent free radicals-mediated cell damage.

CONCLUSIONS
In conclusion, these results indicate that ANE components (BQ chewing) may enhance tumor invasion and metastasis via stimulation of MMP-9 mRNA expression and secretion. This event is associated with ROS, TGF-β1/Smad-dependent (Smad2), TAK1 and Smad-independent (EGFR, JAK, PI3K/Akt, MEK/ ERK) signaling pathways, but not COX activation. ROS scavengers and various small molecules such as PBL, HC, melatonin or antibody for target therapy of these signal transduction pathways can be used for prevention and treatment of BQ chewing-related oral cancer and other diseases (Figure 8).

Culture of GK and SAS cancer epithelial cells
GK were cultured as previously in KGM-SFM with supplements [3,11], by the approval of Ethics Committee, National Taiwan University Hospital (Ethical Approval Number: 201012090RC, 201112153 RID, 201512152RINC). SAS tongue cancer epithelial cells were cultured in DMEM containing 10% FBS and penicillin/streptomycin. They were regularly subcultured before confluence for use.

Cytotoxicity of ANE and arecoline to SAS cells
Briefly 5 × 10 5 cells were inoculated onto 6-well culture wells in 2 ml medium. After 24 hours, medium was changed with fresh medium containing various concentrations of ANE (50-800 μg/ml) or arecoline (0.05-0.8 mM) for 3 days. Medium was collected for ELISA analysis. Cells were washed with PBS and then incubated in medium containing 0.5 mg/ml MTT for 3 hours. Medium was removed and the insoluble formazan generated was dissolved in dimethylsulfoxide (DMSO) and read against blank (DMSO) at OD540 by a Microplate Spectrophotometer.

Effect of ANE and arecoline on MMP-9 gene expression of SAS Cells: Role of some inhibitors
Briefly 5 x 10 5 cells were inoculated onto 6-well culture wells in 2 ml medium. After 24 hours, medium was changed with fresh medium containing various concentrations of ANE (100-800 μg/ml) or arecoline (0.05-0.8 mM) for 24 hours. Culture medium was collected for ELISA analysis of MMP-9 secretion as before [8,45,46]. Total RNA was isolated for semiquantitative RT-PCR.
In some experiments, cells were exposed to different concentrations of some inhibitors (PD153035, AG490, 5Z-7-Oxozeaenol, LY294002, PBL, melatonin) for 30 min before the addition of ANE (800 µg/ml). Cells were further co-incubated for 3 days. Medium was collected for ELISA analysis. Cell layers were used for MTT assay for estimation of cytotoxicity. Cell layers were used for RNA isolation.

Semi-quantitative reverse transcription-polymerase chain reaction
Reverse transcription-polymerase chain reaction (RT-PCR) was done as described before [12]. In short, 3 μg of total RNA was reverse transcribed in a total volume of 44.5 μl reaction solution comprising 8 μl of deoxyribonucleotide triphosphate (dNTP) (2.5 mM), 4 μl of random primer (500 μg/μl), 1 μl of RNase inhibitor (40 U/μl), 4.5 μl of 10x RT buffer, and 0.5 μl of RT (21 U/μl) at 42°C for 90 minutes. Thereafter, 4 μl of complementary deoxyribonucleic acid (cDNA) was taken for PCR amplification in a reaction mixture of 50 μl containing 4 μl of dNTP (2.5 mM), 5 μl of 10x Super Taq buffer, 0.2 μl of Super Taq enzyme (2 U/μl) and 1 μl of each specific primer. The PCR program was set at 94°C for 5 minutes for the first cycle, and then further amplified for 20-35 cycles at 94°C for 1 min, 55°C for 1 min, and then 72°C for 2 min by a thermal cycler (Perkin Elmer 4800, PE Applied Biosystems, Foster City, CA, USA). Finally, the reaction was completed at 72°C for an additional 10 min. The specific primer pairs for β-actin (218 bp) was described previously and used as control [12]. The MMP-9 primer sequence was GTGGTAGCGCACCAGAGGCG and ACCGCCAACTACGA CCGGGA with amplified product of 179 bp. The PCR-generated products were loaded into 1.8% agarose gel electrophoresis with reference of molecular weight markers ranging 100-1200 bp. The gels were finally stained with ethidium bromide and photographed for presentation.
Effect of ANE on TGF-β1, p-Smad2, and p-TAK1 protein expression

Western blotting analysis
Briefly near confluent GK (in 6-well) or SAS cells (1 × 10 6 cells) were cultured on 10-cm culture dishes. They were then exposed to different concentrations of ANE (100-800 μg/ml) for 24 hours. Protein from cell lysates were collected and subjected to western blotting analysis as described previously [11,12], except that TGF-β1 and p-Smad2 antibodies were used as primary antibodies. In some experiments, SB431542 was used for pretreatment and co-incubation with ANE (800 μg/ml) for 24 hours before cell harvesting.

Immunofluorescent staining analysis
Briefly SAS cells (1 × 10 5 cells/well) were cultured on coverslips in 24-well culture. They were then exposed to ANE (800 μg/ml) for 5-120 min. Immunofluorescence staining of p-TAK1 expression was performed as before [45]. Medium was decanted, and cells were rinsed by PBS and fixed by paraformaldehyde (4%) for 20 min. Cells were then rinsed by PBS, permeabilized by Triton X-100 (2%), incubated in 0.3% v/v H2O2 for 20 min. After washed by PBS, bovine serum albumin (BSA, 5%) was applied for blocking of non-specific binding for 1 h, and cells were further incubated at room temperature to primary antibodies (p-TAK1) for overnight. Cells were washed with PBS and then incubated in TRITC-conjugated secondary antibodies (red fluorescence) in the dark for 1 h and counterstained by DAPI (1:1000) of nucleus staining for 30 min. Cells in coverslips were finally mounted and photographed/observed under an inverted microscope and DP Controller/Manager software (Olympus IX71, Olympus Corporation)

Statistical analysis
At least four or more independent experiments were conducted. If necessary, the results of cell study were analyzed by One-way ANOVA and post-hoc Turkey test.

Ethics approval
This study is approved by the Ethics Committee, National Taiwan University Hospital.

CONFLICTS OF INTEREST
All of the authors declare no conflict of interest for this submitted paper.