Areca nut components stimulate ADAM17, IL-1α, PGE2 and 8-isoprostane production in oral keratinocyte: role of reactive oxygen species, EGF and JAK signaling

Betel quid (BQ) chewing is an etiologic factor of oral submucous fibrosis (OSF) and oral cancer. There are 600 million BQ chewers worldwide. The mechanisms for the toxic and inflammatory responses of BQ are unclear. In this study, both areca nut (AN) extract (ANE) and arecoline stimulated epidermal growth factor (EGF) and interleukin-1α (IL-1α) production of gingival keratinocytes (GKs), whereas only ANE can stimulate a disintegrin and metalloproteinase 17 (ADAM17), prostaglandin E2 (PGE2) and 8-isoprostane production. ANE-induced EGF production was inhibited by catalase. Addition of anti-EGF neutralizing antibody attenuated ANE-induced cyclooxygenase-2 (COX-2), mature ADAM9 expression and PGE2 and 8-isoprostane production. ANE-induced IL-1α production was inhibited by catalase, anti-EGF antibody, PD153035 (EGF receptor antagonist) and U0126 (MEK inhibitor) but not by α-naphthoflavone (cytochrome p450-1A1 inhibitor). ANE-induced ADAM17 production was inhibited by pp2 (Src inhibitor), U0126, α-naphthoflavone and aspirin. AG490 (JAK inhibitor) prevented ANE-stimulated ADAM17, IL-1α, PGE2 production, COX-2 expression, ADAM9 maturation, and the ANE-induced decline in keratin 5 and 14, but showed little effect on cdc2 expression and EGF production. Moreover, ANE-induced 8-isoprostane production by GKs was inhibited by catalase, anti-EGF antibody, AG490, pp2, U0126, α-naphthoflavone, Zinc protoporphyrin (ZnPP) and aspirin. These results indicate that AN components may involve in BQ-induced oral cancer by induction of reactive oxygen species, EGF/EGFR, IL-1α, ADAMs, JAK, Src, MEK/ERK, CYP1A1, and COX signaling pathways, and the aberration of cell cycle and differentiation. Various blockers against ROS, EGF, IL-1α, ADAM, JAK, Src, MEK, CYP1A1, and COX can be used for prevention or treatment of BQ chewing-related diseases.


INTRODUCTION
Chewing betel quid (BQ) is popular in Taiwan, India and many Southeast Asian countries [1][2][3]. This habit increases the risk of oral leukoplakia, oral submucous fibrosis (OSF) and oral cancer. There are approximately 2-2.8 million BQ chewers in Taiwan [4] and 600 million BQ chewers worldwide [1]. BQ contains areca nut (AN), lime and inflorescence Piper betle with or without Piper betle leaf. However, the mechanisms and signaling transduction pathways of BQ chemical carcinogenesis are not clear. The induction of reactive oxygen species (ROS), damage to cellular targets (DNA, protein, lipid) after metabolic activation of BQ components by phase 1 enzymes (e.g., cytochrome P450s) [5], the cytotoxic effects of BQ constituents, keratinocyte inflammation and oncogene activation are suggested to be the contributing factors. ROS may be involved in the initiation, promotion and progression of cancer. During BQ chewing, ROS generation is confirmed by both in vitro [6,7] and in vivo (in saliva) studies [8] and may induce oral squamous cell carcinoma (OSCC) in Papua New Guinea and other countries [2,9], via auto-oxidation or metabolic activation by cytochrome p450 (CYP) enzymes [10]. The roles of ROS production by BQ components and the related upstream/downstream signaling in mediating cytotoxicity, aberrant differentiation and prostanoid production/tissue inflammation are crucial in BQ carcinogenesis.
Clinical studies have found the increased expression of a disintegrin and metalloproteinases (ADAMs) in OSCC of Taiwan and other country [11,12]. Overexpression of epidermal growth factor (EGF) and EGF receptor (EGFR) is also noted in head and neck squamous cell carcinoma (HNSCC) [13]. EGFR can be activated by EGF, heparinbinding (HB)-EGF, transforming growth factor-α (TGF-α) and amphiregulin, as well as by ROS [14]. EGFR (HER1, erbB1) is a receptor tyrosine kinase (RTK) that modulates cell proliferation and differentiation via Janus kinase (JAK), Src and Ras/mitogen-activated protein kinases (MAPKs) signaling. Recently, the elevated expression of EGFR and MAPKs is crucial in the pathogenesis of oral cancer [15,16]. Src is a non-receptor tyrosine kinase that may be activated by metals, ROS and ultraviolet (UV) irradiation [17]. Src kinase activity is necessary for EGF and other HER ligand signaling to signal transducer and activator of transcription (STAT) and MAPK pathways in various cancers [16][17][18].
EGF/EGFR, tumor necrosis factor-α (TNF-α) and IL-1α may be involved in the sequential stages of carcinogenesis and tissue fibrosis. These effects occur via activation of receptors, ADAMs and TAK1 to cleave and release EGF [28]. An increased expression of cyclooxygenase-2 (COX-2) in different stages of oral cancer and marked inflammatory cell infiltration in OSF tissues may play a crucial role in the multi-step chemical carcinogenesis [29,30]. Previous reports have found the induction of COX-2 and PGE 2 production in GK by ANE via the activation of ROS, CYP1A1, EGFR, Ras, Src, (HO-1 and MEK/ERK [25,31,32]. It is intriguing to determine whether EGF, IL-1α, and ADAMs are activated by BQ components to induce the release of oxidative stress markers and inflammatory mediatorse.g., 8-isoprostane and PGE 2 production-in oral mucosal cells. Moreover, signal transduction pathways such as ROS, JAK (a downstream molecule of EGFR), and MEK that mediate these cellular responses should be clarified. We hypothesized that BQ chewing may induce tissue inflammation, leading to OSF and oral cancer via stimulation of ROS, EGF/EGFR, JAK, IL-1α and ADAM17 (also called TNF-α converting enzyme, TACE) to impair differentiation and cell cycle progression, as well as the production of 8-isoprostane and PGE 2 production in oral keratinocytes. These complex cross-talk events among EGF, EGFR, IL-1α, ADAM, JAK, Src and other signaling molecules may play an important role in BQ chewingrelated diseases (e.g., cancer, OSF, and atherosclerosis). The results of this study may highlight our development of methods (small molecule inhibitors, antibodies etc.) for prevention and targeting therapy of BQ chewing-related diseases.

Effect of ANE and arecoline on EGF and IL-1α production by GKs
At concentrations of 400 and 800 μg/ml, ANE stimulated EGF secretion of GKs to 1.8 and 3.3-folds of control, respectively ( Figure 1A). Interestingly, arecoline at concentrations of 0.2-0.8 mM also induced EGF secretion of GKs to 1.4-2.8 folds of control ( Figure  1B). Similarly, ANE (400 and 800 μg/ml) induced IL-1α production of GKs by 1.7 to 5.4-folds of control, whereas www.impactjournals.com/oncotarget ANE inhibited IL-1α production by GKs at concentrations of 50-200 μg/ml ( Figure 1C). On the other hand, arecoline stimulated IL-1α production by GKs at a concentration of 0.8 mM, whereas it slightly inhibited IL-1α secretion by GKs at a concentration of 0.05 mM ( Figure 1D).

Signaling for ANE-induced EGF production by GKs
To determine the upstream signaling molecules responsible for ANE-induced EGF production, we found that anti-EGF antibody (aby) effectively decreased the useful EGF content in the culture medium of GKs ( Figure  3A). Pretreatment and co-incubation of catalase effectively prevented the ANE-induced EGF production by GKs ( Figure 3B). On the other hand, GM6001 (an inhibitor of metalloproteinases), anti-TNFα neutralizing aby, pp2 (a Stimulation of IL-1α production of GK by ANE (50-800 μg/ml) (n=14). D. Effect of arecoline on IL-1α production of GK (n=20). *denotes significant difference when compared with control (P < 0.05).

Upstream signaling for ANE-induced ADAM17 production by GKs
To reveal the upstream signaling molecules responsible for ANE-induced ADAM17 production, we found that pretreatment and co-incubation by anti-EGF neutralizing aby slightly decreased ANE-induced ADAM17 production by GKs (P > 0.05) ( Figure 5A). Pretreatment and co-incubation by pp2 and U0126 inhibited ANE-induced ADAM17 production by GKs ( Figure 5B, 5C). Moreover, pretreatment and co-incubation by α-naphthoflavone and aspirin also attenuated ANEinduced ADAM17 production by GKs ( Figure 5D, 5E).

Role of JAK signaling in ANE-induced effects on GKs
Because JAKs are important signaling molecules responsible for EGFR-mediated events, we further tested and found that AG490 (a JAK inhibitor) could not prevent ANE-induced EGF production by GKs ( Figure 6A). By contrast, AG490 attenuated ANE-induced ADAM17 and IL-1α production by GKs ( Figure 6B, 6C). Accordingly, ANE inhibited keratin 5, keratin 14, cdc2 protein expression, whereas ANE stimulated the protein expression of mature ADAM9 (84 KD) but had no marked effect on precursor ADAM9 (105 KD). ( Figure 6D). AG490 may prevent the inhibitory effect of ANE on keratin 5 and keratin 14. Additionally, AG490 suppressed the stimulatory effect of ANE on the protein expression of mature ADAM9, with an increase in precursor ADAM9 expression ( Figure 6D).

Role of EGF and JAK on ANE-induced COX-2 expression and PGE2 production by GKs
To understand the role of EGF and JAK in mediating ANE-induced COX-2 expression and PGE 2 production, anti-EGF aby and AG490 were used to suppress the effect of EGF/EGFR and JAK signaling. Interestingly anti-EGF aby effectively inhibited ANE-induced PGE 2 production and COX-2 expression in GKs ( Figure 7A, 7C). Similarly, AG490 also markedly suppressed ANE-induced PGE 2 production and COX-2 expression in GKs ( Figure 7B, 7D).
Effect of catalase, anti-EGF antibody, IL-1 receptor associated kinase (IRAK) inhibitor, AG490, pp2, U0126, α-naphthoflavone, ZnPP and aspirin on ANE-induced 8-isoprostane production by GKs Generally, 8-isoprostane is considered an oxidative stress marker and product. In this study, catalase effectively prevented ANE-induced 8-isoprostane production by GKs ( Figure 8A). We further tested whether  co-incubation by catalase on ANE-induced EGF production in GK (n=12). C. Pretreatment and co-incubation by GM6001 on ANEinduced EGF production in GK (n=5). D. Pretreatment and co-incubation by anti-TNFα neutralizing aby on ANE-induced EGF production in GK (n=3). E. Pretreatment and co-incubation by pp2 on ANE-induced EGF production in GK (n=6). F. Pretreatment and co-incubation by α-naphthoflavone on ANE-induced EGF production in GK (n=27). G. Pretreatment and co-incubation by ZnPP on ANE-induced EGF production in GK (n=18). H. Pretreatment and co-incubation by aspirin on ANE-induced EGF production in GK (n=10). *denotes significant difference when compared with solvent control. #denotes statistically significant difference when compared with ANE-treated group (P < 0.05). I. Effect of anti-EGF neutralizing aby on the ANE-induced alterations of ADAM9, keratin 5, keratin 14, cdc2 and GAPDH (control) protein expression as analyzed by western blotting. One representative western blot picture was shown. the induction of EGF by ANE is important for this event. Anti-EGF neutralizing aby evidently attenuated the ANEinduced 8-isoprostane production ( Figure 8B). However, IRAK inhibitor (inhibitor of IL-1) showed little preventive effects on ANE-induced 8-isoprostane production ( Figure  8C). To elucidate the role of JAK (a downstream molecule of EGF/EGFR) signaling, AG490 pretreatment and co-incubation almost completely inhibited ANE-induced 8-isoprostane production by GKs ( Figure 8D). Similar inhibitory effects of pp2 ( Figure 8E) and U0126 ( Figure  8F) on ANE-induced 8-isoprostane production were also noted. Moreover, to clarify the role of various metabolic enzymes in 8-isoprostane production, α-naphthoflavone could attenuate ANE-induced 8-isoprostane production Figure 4: A. Pretreatment and co-incubation by catalase on ANE-induced IL-1α production in GK (n=11). B. Pretreatment and co-incubation by anti-EGF neutralizing aby on ANE-induced IL-1α production in GK (n=4). C. Pretreatment and co-incubation by PD153035 on ANEinduced IL-1α production in GK (n=11). D. Pretreatment and co-incubation by U0126 on ANE-induced IL-1α production in GK (n=8). E. Pretreatment and co-incubation by α-naphthoflavone on ANE-induced IL-1α production in GK (n=21). *denotes significant difference when compared with solvent control. #denotes statistically significant difference when compared with ANE-treated group (P < 0.05). Pretreatment and co-incubation by pp2 on ANE-induced ADAM17 production in GK (n=4). C. Pretreatment and co-incubation by U0126 on ANE-induced ADAM17 production in GK (n=5). D. Pretreatment and co-incubation by α-naphthoflavone on ANE-induced ADAM17 production in GK (n=8). E. Pretreatment and co-incubation by aspirin on ANE-induced ADAM17 production in GK (n=10). *denotes significant difference when compared with solvent control. #denotes statistically significant difference when compared with ANE-treated group (P < 0.05). www.impactjournals.com/oncotarget B. Pretreatment and co-incubation by AG490 on ANE-induced ADAM17 production in GK (n=5). C. Pretreatment and co-incubation by AG490 on ANE-induced IL-1α production in GK (n=47). *denotes significant difference when compared with solvent control. #denotes statistically significant difference when compared with ANE-treated group (P < 0.05). D. Effect of AG490 on ANE-induced alterations of ADAM9, keratin 5, keratin 14, cdc2 and GAPDH (control) protein expression as analyzed by western blotting. One representative western blot picture was shown.
Recently, we have found the stimulation of EGFR phosphorylation and activation by ANE [25], possibly due to the induction of EGF production by ANE as found in this study. This event is inhibited by catalase, but not by GM6001, anti-TNFα aby, pp2, α-naphthoflavone, ZnPP and aspirin, suggesting that ANE-induced EGF production is correlated to ROS, but not by TNFα production, proteinase cleavage, Src, CYP1A1, HO-1 and B. Pretreatment and co-incubation by AG490 (15 and 30 μM) on ANE-induced PGE2 production in GK. Results were expressed as Mean ± SE. *denotes significant difference when compared with solvent control. #denotes statistically significant difference when compared with ANE-treated group (P < 0.05). C. Pretreatment and co-incubation by anti-EGF neutralizing aby on ANE-induced COX-2 protein expression of GK. D. Pretreatment and co-incubation by AG490 on ANE-induced COX-2 protein expression of GK. One representative western blotting picture was shown.
COX. EGF can be an early response signaling molecule for ANE-induced cellular events in GKs. Moreover, anti-EGF aby attenuates ANE-induced ADAM9 maturation, but not the ANE-induced decline of cytokeratin 5, 14 and cdc2, indicating the presence of differential signaling pathways responsible for different downstream effective molecules. Anti-EGF aby and AG490 suppress the ANE-induced COX-2 expression, PGE 2 and 8-isoprostane production, but not cdc2 expression of GK. During BQ chewing, ROS may be generated by auto-oxidation of BQ components in saliva or via their intracellular metabolic activation [1,2]. BQ-induced ROS overproduction is correlated to DNA/cell damage, tissue inflammation, cell cycle regulation, apoptosis and gene expression with incubation by anti-EGF neutralizing aby on ANE-induced 8-isoprostane production in GK. C. Pretreatment and co-incubation by IRAK inhibitor on ANE-induced 8-isoprostane production in GK. D. Pretreatment and co-incubation by AG490 on ANE-induced 8-isoprostane production in GK. E. Pretreatment and co-incubation by pp2 on ANE-induced 8-isoprostane production in GK. F. Pretreatment and coincubation by U0126 on ANE-induced 8-isoprostane production in GK. G. Pretreatment and co-incubation by α-naphthoflavone on ANEinduced 8-isoprostane production in GK. H. Pretreatment and co-incubation by ZnPP on ANE-induced 8-isoprostane production in GK. I. Pretreatment and co-incubation by aspirin on ANE-induced 8-isoprostane production in GK. *denotes significant difference when compared with solvent control. #denotes statistically significant difference when compared with ANE-treated group (P < 0.05). associated lipid peroxidation, protein modification and DNA damage. Recently, we have found the activation of ROS, CYP1A1, EGFR, Ras, Src and HO-1 signaling by ANE to induce COX-2 expression/PGE 2 production in GK [25]. Moreover, EGF can activate EGFR to stimulate cell proliferation, differentiation, invasion and metastasis via stimulation of downstream JAK, Src, Ras/ MAPKs and PI3K/Akt signaling [14,[16][17][18]. GW2974, a dual inhibitor of EGFR and ErbB2 tyrosine kinase, may attenuate the 7,12-dimethylbenz[a]anthracene (DMBA)induced hamster cheek pouch tumor with concomitant reduction of tissue PGE 2 , indicating the presence of crosstalk between EGFR and arachidonic acid metabolism [45]. Studies also reveal the upregulation of COX-2 and EGFR in oral leukoplakia and oral carcinogenesis [46]. In this study, ROS-EGF/EGFR-and JAKs-COX-2 signaling pathways are shown to contribute to oral mucosal inflammation and carcinogenesis in BQ chewers. ANE-induced ADAM9 maturation and decrease of cytokeratin expression are correlated to JAK. ANE has been shown to PI3K/Akt, EGFR and COX signaling and contribute to BQ carcinogenesis [25,47,48]. However, additional signaling molecules are present to downregulate cdc2 by ANE.
ROS are critical molecules for stimulation of ANE-induced PGE 2 production in GK [25,31]. To know more about the role of ROS in BQ carcinogenesis, we interestingly found that ROS is necessary for the ANEinduced EGF, IL-1α, and 8-isoprostane production. However, ANE at lower concentrations partly inhibited the IL-1α and 8-isoprostane production, possibly because ANE also contains some anti-oxidative components. IL-1α is involved in tissue inflammation, immune modulation and carcinogenesis via binding to IL-1 receptor to trigger signal transduction pathways such as IL-1 receptor (IL-1R)-associated kinase (IRAK) and TGFβ-activated kinase-1 (TAK1) [49,50]. 8-Isoprostane has been used as a disease marker for obesity, ischemiareperfusion injury, and cancer [51]. It may activate thromboxane receptors in response to oxidative injury [52]. Exposure to ANE may stimulate ROS and thereby downstream signaling pathways such as EGF/EGFR, IL-1α/IL-1R and 8-isoprostane/receptor to stimulate oral carcinogenesis. This may explain why ROS may activate receptors, receptor-activated protein kinases and nuclear transcription factors, including growth factor receptors, JAK, Src kinase, Ras signaling, MAPKs, PI3K/Akt pathway, NF-kB [16][17][18]. In addition to catalase, the ANEinduced IL-1α production is prevented by anti-EGF aby, PD153035 and U0126, but enhanced by α-naphthoflavone. These results suggest that ANE-induced IL-1α production of GK is mediated by ROS, EGF/EGFR and MEK/ ERK activation. Similarly IL-1α production and nuclear localization are correlated to ROS levels, EGFR activation and MEK/ERK in fibrosarcoma, skin keratinocytes and in cerebral ischemia injury [53,54]. Furthermore, IL-1α and TNF-α are important mediators involved in carcinogenesis and fibrosis of many organs via activation of receptor activation/TAK1 signaling [55,56]. GK expressed various types of CYP enzymes mainly CYP1A1, 2C8/19, 2E1, and 3A3/3A4, and may involve in ANE-induced COX-2 expression and PGE 2 production in GK [25]. Interestingly α-naphthoflavone by itself stimulates IL-1α production. This may partly explain the inhibition of CYP1A1/ CYP1A2 by α-naphthoflavone enhanced the ANE-induced IL-1α production. The involvement of CYP1A1/CYP1A2 and its inhibition by α-naphthoflavone on ANE-induced events suggest that possibly metabolic activation of ANE components is necessary for some of the ANE-induced carcinogenic events [25] and increase the risk of OSF and oral cancer [57,58].
Since EGFR ligands can be shed from plasma membrane by metalloproteinases and sheddases -a disintegrin and metalloproteinases (ADAMs). ADAM10, 12, 17 are the major sheddases of EGFR ligands in response to stimuli such as G-protein coupled receptors, growth factors, cytokines, wounding and phorbol ester etc. [59]. Over-expression of ADAMs (ADAM9, 10, 12, and 17 etc.) is popularly noted in epithelial inflammation and carcinogenesis [60] and increased expression of certain ADAMs may enhance tumor cells invasion, proliferation in vitro and promote tumor formation in vivo. ADAM17 may enhance the invasion of oral cancer [43]. An increased expression of ADAM10 is found in OSCC of Taiwan [11] and expression ADAM17 in head/neck SCC in Germany [12]. MMP2 and MMP-9 also contribute to BQ-related oral carcinogenesis by promotion of cancer invasion and metastasis [26,27]. In this study, we further found the stimulation of ADAM9 maturation and ADAM17 secretion by ANE, suggesting the involvement of ADAM9 and ADAM17 in BQ carcinogenesis. ANE-induced ADAM17 secretion can be suppressed by pp2, U0126, α-naphthoflavone and aspirin, indicating this event is associated with Src, MEK/ERK, CYP1A1 and COX signaling. Src is a non-receptor tyrosine kinase that is activated by metals, ROS and UV irradiation [17]. Src overexpression has been found in head/neck cancers. Activated Src may induce downstream signaling of MAPKs, NF-kB and PI3K. Moreover, BQ components can stimulate Src activation and ERK to promote cancer cells' migration and motility [61].
Previous studies show the association between tissue inflammation and cancer/fibrosis with an elevation of COX-2 expression and prostanoid production in oral cancer and precancer [30]. AN components may induce tissue injury and inflammation, COX-2 expression and PGE 2 production in GK via ROS, EGFR, Src, and MEK/ERK signaling [25,31,32]. In this study, ANE is further found to induce 8-isoprostane production. PGE 2 is involved in oral carcinogenesis by induction of sustaining epithelial hyperplasia, angiogenesis, immunosuppression and tumor metastasis. The 8-isoprostane has been suggested as an oxidative stress marker during chemical carcinogenesis and may induce vasoconstriction but inhibit angiogenesis [62][63][64]. 8-Isoprostane levels in serum, urine and exhaled breath condensate are used as the disease marker in tissue fibrosis, prostate and lung cancer [65][66][67], suggesting the potential use of 8-isoprostane as a marker of oral cancer and OSF. ANEinduced PGE 2 production is related to ROS, EGFR, Src, MEK/ERK, CYP1A1 and HO-1 [25], and in addition, EGF and JAK signaling in this study. Interestingly, ANEinduced 8-isoprostane production of GK is prevented by catalase, anti-EGF aby, IRAK inhibitor, AG490, pp2, U0126, α-naphthoflavone, ZnPP and aspirin. These results demonstrate that ANE-induced 8-isoprostane production in GK is related to ROS, EGF and IL-1 production and the downstream signaling via IRAK, JAK, Src, MEK/ERK. CYP1A1, HO-1 and aspirin are also associated with these processes.
Based on this study and other prior reports [1,2,6,25,32], we conclude that AN components play crucial roles in the pathogenesis of BQ-induced oral cancer and OSF possibly via induction of ROS, EGF/EGFR, JAK, Src, MEK/ERK, IL-1α, ADAMs, CYP1A1, HO-1 and COX signaling pathways, as well as the aberration in cell cycle-and differentiation-related proteins of oral keratinocytes (Figure 9). Auto-oxidation or metabolic activation of ANE components by CYP1A1 may generate ROS and reactive intermediates. ROS may induce multiple signaling pathways such as EGF/EGFR,

Culture of gingival keratinocytes (GKs)
GKs were cultured as described previously [25,32]. With the approval of the Ethics Committee of National Taiwan University Hospital, human gingiva (with a gingivitis index < 1) was obtained during clinical crown-lengthening procedures with proper written informed consent by the patients. Most of the subepithelial connective tissue of the gingiva was first removed using a surgical knife, and then tissues were cut into small pieces, placed onto culture dishes and cultured in KGM-SFM with supplements. The cell passages of GKs ranging from 1 to 3 were used through this study.
Effect of ANE and arecoline on 8-isoprostane, EGF, IL-1α, and ADAM17 production by GKs Near-confluent GKs in 6-well culture plates were exposed to 2 ml of fresh medium containing various concentrations of ANE and arecoline. Cells were further incubated for 24 h. Culture medium was collected for the analysis of 8-isoprostane, EGF, IL-1α, and ADAM17 levels by ELISA.

Statistical analysis
Four or more separate experiments were performed. The results were expressed as the mean ± SE and analyzed by paired Student's t-test. A P value < 0.05 was considered to indicate a statistically significant difference between 2 study groups.