Arecoline, the major alkaloid of areca nut, has been shown to cause strong genotoxicity and is considered a potential carcinogen. However, the detailed mechanism for arecoline-induced carcinogenesis remains obscure. In this study, we noticed that the levels of p21 and p27 increased in two oral squamous cell carcinoma cell lines with high confluence. Furthermore, when treated with arecoline, elevated levels of p21 and p27 could be downregulated through the reactive oxygen species/mTOR complex 1 (ROS/mTORC1) pathway. Although arecoline decreased the activity of mTORC1, the amounts of autophagosome-like vacuoles or type II LC3 remained unchanged, suggesting that the downregulation of p21 and p27 was independent of autophagy-mediated protein destruction. Arecoline also caused DNA damage through ROS, indicating that the reduced levels of p21 and p27 might facilitate G 1/S transition of the cell cycle and subsequently lead to error-prone DNA replication. In conclusion, these data have provided a possible mechanism for arecoline-induced carcinogenesis in subcytolytic doses in vivo. (Cancer Sci 2012; 103: 1221–1229)
Oral cancer is one of the most common cancers, resulting in 128 000 deaths worldwide in 2008. Oral squamous cell carcinoma (OSCC) is frequently found on the tongue and buccal, as well as gingival areas and accounts for more than 90% of oral cancer incidence. In addition to cigarettes and alcohol, betel quid is a risk factor of oral cancer. Betel quid is composed of areca nut, lime, and Piper betle leaf. Being a major alkaloid in areca nut, arecoline has long been considered a potential carcinogen. Several reports showed that arecoline may increase reactive oxygen species (ROS), retard the cell cycle, and induce apoptosis.[3, 4] Arecoline was also proved to enhance unscheduled DNA synthesis that caused strong genotoxicity in mouse germ cells. In addition, arecoline impeded DNA repair and mitotic spindle assembly.[6-8] Recently, arecoline was suggested to regulate the expression of certain genes through epigenetic control.
Cell cycle progression is dependent on sequential coordination among multiple proteins. In particular, cyclin levels can oscillate in different stages and interact with cyclin-dependent kinases (CDKs), forming the cyclin/CDK complexes for promoting cell cycle progression. In contrast, p21 and p27 belong to the kinase inhibitor protein family and regulate the cell cycle in response to various stresses, such as DNA damage, hypoxia, and confluence stress.[10-13] For example, once activated by DNA damage sensors like ataxia telangiectasia mutated, p53 enhances p21 transcription to arrest cell cycle before G1/S transition for DNA repair.
Reactive oxygen species consists of various radicals that are triggered by a range of agents and may exert different effects on downstream signaling pathways, including the levels of p21 and p27.[14, 15] Deregulated ROS levels are strongly implicated in cancers. Reactive oxygen species can induce DNA damage that results in genomic instability and contributes to carcinogenesis.[17, 18]
The mTOR complex 1 (mTORC1) is composed of mTOR, raptor, and GβL and is controlled by multiple pathways, such as PI3K-Akt signaling. Activated mTORC1 facilitates protein synthesis through controlling various downstream factors including 4E-BP1, p70S6K, and eEF2K, thus allowing release of eIF4E for translation initiation and activation of elongation factor 2 (eEF2) for translation elongation. Another important role of mTORC1 is to regulate autophagy. In conditions like energetic stress, mTORC1 may be inhibited by AMP-activated protein kinase (AMPK) to allow autophagy induction.
This study was initiated from an interesting observation: upregulation of p21 and p27 in OSCC cells with high confluence could be downregulated by arecoline treatment. In 2003, the International Agency for Research on Cancer announced that betel quid and areca nut chewing were carcinogenic to humans. Arecoline is one of the most abundant alkaloids of areca nut, suggesting its potential role in carcinogenesis. Here, we show the mechanism behind arecoline-induced downregulation of p21 and p27, which may be important for betel quid chewing oral carcinogenesis.
Materials and Methods
Two OSCC cell lines, OCSL and OC2, derived from two Taiwanese men with habits of drinking alcohol, smoking, and areca nut chewing, were maintained in RPMI-1640 medium supplemented with 10% FBS and 1% penicillin–streptomycin. Cells from a Japanese oral cancer cell line, Ca9-22, were maintained in DMEM/F12 with similar supplements. Cells were routinely cultured in a 37°C incubator supplied with 5% CO2. 12–16 h after seeding, experiments were carried out when cell confluence reached 95% (high) or approximately 30–40% (low).
Reagents and antibodies
Arecoline, N-acetylcysteine (NAC), nocodazole, 2′-7′-dichlorofluorescin diacetate (DCFDA), NH4Cl, and DMSO were purchased from Sigma (St. Louis, MO, USA). Antibodies against phospho-eEF2 (T56), eEF2K, phospho-eEF2K (S366), phospho-p70S6K (T389), phospho-S6 (S235/236), LC3, and rapamycin were from Cell Signaling Technology (Danvers, MA, USA). Antibodies against p21, p53, and phospho-histone H3 (S10) were from Epitomics (Burlingame, CA, USA). Antibody against phospho-4E-BP1 (T45) was from Abcam (Cambridge, MA, USA). Antibody against p27 was from BD Biosciences (Franklin Lakes, NJ, USA), and antibody against γH2AX was from Abnova (Walnut, CA, USA).
Areca nut extract (ANE) was prepared from fresh nuts. In brief, the nuts were chopped into 0.5–1 cm3 pieces by a blender and the water-soluble ingredients were extracted at 4°C overnight. The supernatant was harvested and concentrated by −70°C lyophilisation. The powder derived from water extract was weighed, redissolved in ddH2O, and stored at −20°C before experimental use.
Cell lysate preparation and Western blot analysis
Cells in 24-well plates were washed twice with PBS and lysed with 70 μL 4× Laemmli loading buffer, followed by boiling for 10 min. Equal amounts of samples were run on SDS-PAGE gels and transferred to PVDF membranes. Expression of individual proteins was detected using corresponding antibodies, followed by the secondary antibody conjugated with HRP. After incubation with ECL, the membranes were exposed to X-ray film (Kodak, Rochester, NY, USA).
Cell proliferation assay
Cells with 30–40% or more than 95% confluence were treated with the indicated reagents. One or 2 days later, MTT reagent (Sigma) with 1 mg/mL final concentration was added to each well. Plates were swirled gently for 5 s and cultured continuously for 3 h. After incubation, medium was removed. Cells were washed twice with PBS and MTT metabolic product was resuspended in 500 μL DMSO. After swirling for seconds, 50 μL supernatant from each well was transferred to optical plates for detection at 595 nm.
Detection of ROS
The OSCC cells were treated with 0.5 mM arecoline. After 6 or 12 h, medium was removed and cells were incubated in PBS containing 10 μM DCFDA for approximately 30 min. Cells were observed using a fluorescence microscope and photographed. For quantifying the level of ROS, cells cultured in 24-well plates were treated with arecoline as indicated, and 50 μM DCFDA was added 30 min before harvesting cells. Cells were finally washed once with PBS and dissolved in 200 μL DMSO containing 1 mM NAC for quenching reaction. After swirling for 5–10 s, 50 μL supernatant was transferred for fluorescence evaluation.
Cells were harvested for RNA extraction using TriPure reagent (Roche, Mannheim, Germany) 24 h after treatment with 0.5 mM arecoline. After cDNA synthesis, the reaction was carried out as previously described.
Cell cycle analysis
Cells seeded in six-well plates were treated with or without 0.5 mM arecoline for 24 h. The detailed procedures of cell cycle analysis were as described.
Incorporation of BrdU
After attachment, cells were starved in serum-free medium for 1 day. Subsequently, cells were cultured in fresh medium containing 10% FBS with or without 0.5 mM arecoline for 24 or 30 h. Twelve hours after addition of FBS, BrdU was added at final concentration 10 μM. Cells were then washed twice with PBS and fixed in ice-cold 4% formaldehyde/PBS for 20 min. After washing with PBS again, cells were incubated in 1 N HCl for 30 min, followed by washing with Tris buffer (pH 7.4) three times, and stained with anti-BrdU antibodies at 4°C with gentle swirling overnight. After PBS washing, cells were then incubated with second antibodies conjugated with Alexa 555 at room temperature for 3–4 h. After final washing, incorporation of BrdU was observed using a fluorescence microscope and photographed. For quantification, treated cells were harvested by trypsinization and fixed with 70% ice-cold ethanol, followed by the procedures above except that different secondary antibodies were used. Cells were finally sent for flow cytometry analysis after propidium iodide staining.
Arecoline downregulates p21 and p27 in OSCC cells with high confluence
Two OSCC cell lines, OCSL and OC2, were established from patients from middle Taiwan. In attempt to evaluate the effect of arecoline on these two cell lines, we noticed that the protein levels of p21 and p27 fluctuated dramatically depending on the confluence of harvested cells (Fig. 1A). Arecoline (0.5 mM) downregulated p21 and p27 in high-confluence OSCC cells (Fig. 1B). Interestingly, the inhibitory effect of arecoline was not significant at a lower confluence (Fig. S1). Cell cycle regulators such as CDK inhibitors p21 and p27 in cultured cells at high density are usually increased to induce cell cycle arrest contact inhibition. In contrast, reduction of p21 and/or p27 may decrease the sensitivity to contact inhibition and facilitate continuous cell growth.[23-26] We therefore speculated that downregulation of p21 and/or p27 could play a role in carcinogenesis of betel quid-dependent oral cancers. However, 0.5 mM arecoline also caused detectable cell death (data not shown). Thus, we adopted lower doses of arecoline to rule out possible cytotoxic effects. Arecoline at lower concentrations had no effect on cell growth but still decreased p21 and p27 in OSCC cells with high confluence (Fig. 1C,D, Fig. S1).
Arecoline enhances ROS in OSCC cells
To ask how arecoline could downregulate p21 and p27, we first checked the endogenous level of p53. Interestingly, OCSL showed weak expression of a truncated form of p53 and OC2 cells possessed mutant p53 (Fig. 1B), suggesting that the downregulation of p21 and p27 was mediated through a p53-independent pathway.
Arecoline had been proven to positively induce ROS in various cell types,[4, 28] so we hypothesized that arecoline regulated the levels of p21 and p27 through a ROS-dependent pathway. In order to prove this hypothesis, we examined arecoline-induced ROS in these two OSCC cells. Production of ROS in both OCSL and OC2 cells was enhanced in the presence of arecoline using the fluorescent ROS detector DCFDA (Fig. 2A). Treating the cells with different concentrations of arecoline induced ROS in a dose- and time-dependent manner, with the exception for the highest dose (2.5 mM) that caused cytotoxicity (Fig. 2B,C). To confirm the effect of arecoline on ROS, NAC (a ROS quencher) was used to block the function of ROS. As shown in Figure 2D, the growth-inhibitory effect of arecoline could be significantly counteracted in the presence of NAC.
Arecoline strongly inhibits mTORC1 signaling
Previous studies also indicated that synthesis and activity of p21 and p27 were regulated by mTORC1.[29-32] To test if arecoline downregulated p21 and p27 through the mTORC1 pathway, we examined the effect of arecoline on mTORC1 activity. Arecoline strongly inactivated mTORC1 activity in all three tested OSCC cell lines as evidenced by the decreased amounts of phosphorylated 4E-BP1 and S6 (Fig. 3A). Downregulation of other mTORC1 downstream factors including p70S6K, eEF2, and eEF2K further confirmed the inhibitory effect of arecoline on mTORC1 activity (Fig. S2). To verify whether this inhibitory effect was linked to the ROS pathway, mTORC1 activities were examined in OCSL cells cotreated with arecoline and NAC. The NAC completely abolished the inhibitory effect of arecoline and rescued mTORC1 activity (Fig. 3B, lane 2 vs lane 4), highly suggesting that arecoline inhibited mTORC1 through a ROS-dependent pathway. We then used the mTORC1 inhibitor, rapamycin, to confirm that p21 and p27 expression could be regulated by mTORC1 signaling in OSCC cells. Indeed, rapamycin alone was sufficient to decrease p21 and p27 (Fig. S3A), whereas arecoline failed to downregulate p21 or p27 after treatment with NAC (Fig. 3C, lane 2 vs lane 4). Interestingly, treatment using arecoline did not decrease transcripts of p21 but increased transcripts of p27 significantly, suggesting that arecoline regulated p21 and p27 in a post-transcriptional way (Fig. S3B). In addition, PI3K-Akt and AMPK signaling are two well-known upstream regulators of mTORC1. Unexpectedly, arecoline had no effect on PI3K-AKT activation despite the strong inhibition in S6 phosphorylation (Fig. S4). Unlike PI3K-Akt, activated AMPK was reported to inhibit mTORC1. However, arecoline inactivated the activity of AMPK as evidenced by the decrease of AMPK phosphorylation (Fig. S5A). Inhibition of AMPK by compound C also caused no significant influence on p21 and p27 (Fig. S5B). Taken together, these data indicate that arecoline-induced downregulation of p21 and p27 in OSCC cells was mediated through a novel, ROS/mTORC1-dependent, but Akt- and AMPK-independent pathway.
Arecoline-induced downregulation of p21 and p27 might be autophagy-independent
mTORC1 has been shown to promote protein translation but inhibit autophagy. Autophagy was reported to play roles in degradation of cytosolic proteins, organelles, or pathogens in a non-selective manner. Accumulating evidence suggests that autophagy could be involved in the selective degradation of proteins through p62-mediated ubiquitination.[35, 36] Areca nut extracts have been reported to induce autophagy in OSCC cells,[37, 38] suggesting that downregulation of p21 and p27 might be through autophagy-mediated protein degradation. Thus, it was critical to understand whether arecoline-induced downregulation of p21 and p27 involved the autophagy pathway. Surprisingly, as shown in Figure 4(A), ANE but not arecoline strongly induced autophagosome-like vacuoles in OC2 cells. Arecoline-induced autophagosome-like vacuoles were not significant even under the blockade of autophagic flow by NH4Cl (Fig. 4B). Therefore, we further investigated whether arecoline or ANE could induce production of type II LC3 (LC3-II), a common indication of autophagy induction. Consistently, accumulation of LC3-II was only observed in cells treated with ANE (Fig. 4C, lanes 4 and 5). However, arecoline but not ANE efficiently reduced phosphorylation of S6 (Fig. 4C, lane 3 vs lane 5). This result was confirmed by examining the phosphorylations of 4E-BP and S6 in another oral cancer cell line Ca9-22 (Fig. S6). Therefore, we concluded that arecoline-induced downregulation of p21 and p27 was not mainly through mTORC1-dependent autophagy.
Arecoline-treated cells bypass G 1/S checkpoint in spite of DNA damage
Both p21 and p27 play important roles in maintaining genome integrity during cell cycle progress. It is known that p21 and p27 can be upregulated to arrest cells mainly at the G1/S checkpoint for genome repair. As ROS is thought to cause DNA damage, we speculated that arecoline-induced p21 and p27 downregulation may result in a less stringent G1/S checkpoint, and thus increased the possibility of error-prone DNA replication, leading to arecoline-induced carcinogenesis. Indeed, arecoline induced DNA damage as indicated in the increase of γH2AX (Fig. 3C, lane 1 vs lane 2) and enhanced cell populations at the G2/M stage in unsynchronized OC2 cells (Fig. 5A). Interestingly, arecoline also increased the cell population in S stage in addition to the G2/M stage. This increase is also indirect evidence for enhanced G1/S transition. However, using BrdU incorporation assay, we found that DNA replication in synchronized OC2 cells was not obviously affected by arecoline as compared to the serum starvation control (Fig. 5B). Under such conditions, the strong increase of γH2AX indicated that arecoline treatment induced DNA damage but still allowed cell transition into G2/M as evidenced by the increase of histone H3 phosphorylation (Fig. 5C, lane 1 vs lane 2). Consistently, synchronized cells treated with arecoline for a longer time clearly increased the population retarded at G2/M with completed DNA replication, as evidenced by a higher percentage of incorporated BrdU and phosphorylated H3 (Fig. 5D). These results suggested that genome replication still proceeded in spite of DNA damage.
In this study, we showed an interesting observation that the levels of p21 and p27 were obviously upregulated, particularly in high confluence of OSCC cells. Several studies have indicated that cell confluence affects protein expression or function in malignant cells. For example, accumulation of HSP27 during high confluence increases resistance against anticancer drugs in colon cancer cells. CD26 also has a confluence-dependent expression pattern in colon adenocarcinoma cell lines HCT-116 and HCT-15. Therefore, the arecoline-induced downregulation of p21 and p27 in confluent cells might play an important role in regulating the mechanism of oral carcinogenesis. Here, we illustrated the potential roles of p21 and p27 in arecoline-induced oral carcinogenesis.
Several factors could affect the cellular levels of p21 and p27, including gene expression, stability, and intracellular localization.[42, 43] For example, active p53 increases the expression of p21 and p27 to arrest cells.[44, 45] In human epithelial cells, arecoline was reported to decrease p53–p21 signaling and repress DNA repair. Given that p53 was either weakly expressed or mutated in the two oral cancer cell lines, it is conceivable that p53 was not the major factor accounting for arecoline-induced downregulation of p21 and p27 (Fig. 1B). Instead, we showed that mTORC1 played an important role in this biological effect.
However, the detailed mechanism by which mTORC1 regulates the levels of p21 and p27 remains unclear. One possible mechanism is through the autophagy degradation pathway. However, 10 nM rapamycin is sufficient to downregulate p21 and p27 but fails to increase autophagosome-like vacuoles or LC3-II simultaneously (data not shown and Fig. 4C, lane 8), highly suggesting that downregulation of p21 and p27 might not be through the mTORC1-dependent autophagy degradation pathway. The other possible mechanism is through protein synthesis, as arecoline affected translational machinery like S6 (Fig. 3A,B) and eEF2, which is known to mediate GTP-dependent translocation of synthesizing peptide chain from A-site to P-site in ribosomes. In addition, mTORC1 was reported to phosphorylate SGK1 and modulate the function of p27.
As important regulators at cell cycle checkpoints, p21 and p27 are frequently considered as tumor suppressors. Reduction of p21 and/or p27 possibly facilitates deregulated proliferation or insensitive to contact inhibition.[23, 26, 47] Several studies have indicated the prognostic potential of p27 in various cancers, including oral cancers.[48-53] It was shown that p27 could induce senescence in cancer cells and was considered a checkpoint for limiting the progression of prostatic intraepithelial neoplasia to invasive cancer. Unlike p27, the role of p21 remains controversial.[55, 56] Previous studies regarding the prognostic application of p21 in different stages of OSCC were inconsistent.[48, 57-60] It remains unclear whether these inconsistencies result from p21 polymorphism, which may be associated with oral malignancy.
Long-term exposure of betel quid, which is composed of areca nut, lime, and Piper betle leaf, is frequently associated with the induction of carcinogenesis in the oral cavity. However, the detailed mechanism behind betel quid chewing carcinogenesis is still elusive. Arecoline is the major alkaloid in areca nut. According to previous studies, the detected concentration of arecoline in betel quid chewers' saliva was approximately 0.3 mM or 0.1–10 μg/mL, with the sudden peak concentration of approximately 100 μg/mL during betel quid chewing.[6, 62] This urged us to propose a model as shown in Figure 6: the arecoline concentration in the oral cavity of betel quid chewers may fluctuate dramatically in vivo. The sudden impact of high concentrations of arecoline may induce DNA damage through the ROS pathway and prevent increases of p21 and p27 through the ROS/mTORC1 pathway simultaneously, leading to a bypass of the G1/S checkpoint. Once betel quid is removed, cells are allowed to continue proliferation through G2/M presumably without genomic integrity. Finally, in combination with other unidentified factors, such as p53 mutation, arecoline leads to the deregulated growth of surviving cells, accounting for the mechanism of arecoline-induced carcinogenesis.
The authors would like to thank Mr. Chun-Jen Wang for technical assistance. This work was supported by the National Science Council of Taiwan (97-2311-B-194-001-MY3 and NSC-99-2314-B-705-002-MY2).