2-O-Methylmagnolol upregulates the long non-coding RNA, GAS5, and enhances apoptosis in skin cancer cells

Magnolol, a hydroxylated biphenol compound isolated from the bark of Magnolia officinalis, has been shown to exhibit anti-proliferative effect in various cancer cells, including skin cancer cells. Methoxylation of magnolol appears to improve its anti-inflammatory activity, yet the effect of this modification on the agent's antitumor activity remains unknown. In this work, we report that 2-O-methylmagnolol (MM1) displays improved antitumor activity against skin cancer cells compared to magnolol both in vitro and in vivo. The increased antitumor activity of MM1 appears to correlate with its increased ability to induce apoptosis. DNA microarray and network pathway analyses suggest that MM1 affects certain key factors involved in regulating apoptosis and programmed cell death. Interestingly, the level of the long non-coding (lnc) RNA of growth arrest-specific 5 (GAS5) was increased in MM1-treated cells, and inhibition of lncRNA GAS5 inhibited MM1-induced apoptosis. Conversely, overexpression of lncRNA GAS5 inhibited cell proliferation and promoted cell apoptosis in skin cancer cells. The expression of lncRNA GAS5 in the skin cancer tissues was found to be lower than that in the adjacent normal tissues in a majority of patients. Taken together, our findings suggest that MM1 has improved antitumor activity in skin cancer cells, and that this is due, at least in part, to the upregulation of lncRNA GAS5 and the enhancement of apoptosis.

Skin cancer represents a major and growing public health problem. The three most common forms of skin cancer are basal cell carcinoma, squamous cell carcinoma, and melanoma. Basal cell carcinoma and squamous cell carcinoma are together known as non-melanoma skin cancers and are both derived from keratinocytes, whereas melanomas are derived from melanocytes. 1 Squamous cell carcinomas can undergo metastasis, basal cell carcinomas rarely show this progression, and melanomas can be highly metastatic.
Although surgical removal remains the primary treatment for skin cancer, herbal therapy is becoming increasingly popular among physicians and patients. 2 Magnolia officinalis is widely used in traditional Chinese medicine because of its various pharmacological activities. [3][4][5] Magnolia bark is rich in the extensively investigated biphenol compound, magnolol (5,5′diallyi-2,2′-dihydroxybiphenyl, C 18 H 18 O 2 ). 6 The cytotoxic and anti-angiogenic activities of magnolol appear to reflect the presence of hydroxyl and allylic groups on a biphenolic moiety. 7 An early study showed that magnolol had remarkable inhibitory effects on mouse skin tumor promotion in an in vivo carcinogenesis test. 8 More recently, magnolol has been reported to exert antitumor activity by inhibiting proliferation and inducing apoptosis, [9][10][11][12][13] countering metastasis, [14][15][16] and suppressing angiogenesis. 17,18 Magnolol induces apoptosis in the cells of many human cancers, including gallbladder cancer, non-small cell lung cancer, prostate cancer, human breast cancer, bladder cancer, colon cancer, and skin cancer. [9][10][11][12]14,[18][19][20][21][22][23][24] Therefore, magnolol has been suggested as a potential apoptosistargeting drug. 25 Together, the results from the existing in vivo and in vitro studies have indicated that magnolol is a promising candidate for the development of new strategies for preventing and/or treating skin cancers in humans.
Previous studies demonstrated the methoxylation of honokiol enhanced the anti-inflammatory effect. 26 In an attempt to optimize the anti-inflammatory ability of magnolol, we previously synthesized two derivatives of magnolol bearing one or two methoxy moieties, and designated these agents '2-Omethylmagnolol' (MM1) and 'dimethylmagnolol' (M2M), respectively. 27 However, although we showed that MM1 and M2M exhibited improved anti-inflammatory activity, the cytotoxic mechanisms and in vivo anti-carcinogenic effects of these compounds remained largely unknown. In the present study, we compared the antitumor activity of magnolol, M2M and MM1, against skin cancer cells in vitro and in vivo, and explored the underlying cytotoxic mechanisms.

Results
Cytotoxic effects of magnolol, M2M and MM1 in skin cancer cells. To optimize the anti-inflammatory ability of magnolol, we previously synthesized methoxylated derivatives of magnolol ( Figure 1a). We found that whereas magnolol displayed a moderate cytotoxic effect against keratinocytes at 10 μg/ml, M2M and MM1 displayed little or no such effect. 27 Thus, this methoxylated compound appeared to have lower cytotoxicity against normal keratinocytes compared to magnolol. Here we set out to evaluate the antitumor effects of magnolol, M2M and MM1. We first  examined the cytotoxic effects of these compounds in the skin cancer cell lines, A375 (melanoma) and A431 (squamous carcinoma). As shown in Figure 1b, MM1 displayed a greater cytotoxic effect against A375 and A431 cells compared to magnolol and M2M using sulforhodamine B (SRB) assay. Further examination showed that MM1 more significantly inhibited cell proliferation than magnolol and M2M using MTT assay (Figure 1c). To elucidate the effect of MM1 on cell proliferation, cells treated with 75 μM of magnolol or MM1 for 24 h were analyzed by flow cytometry using propidium iodide staining, As shown in Figure 1d, the percentage of cells in the sub-G1 region (M1) was greatly increased in the cells treated with MM1 when compared to the cells treated with magnolol or mock control, suggesting that the induction of apoptosis by MM1 accounts for its potent anti-proliferation activity. Taken together, these data showed that the structural features contributing to the greatest cytotoxic effects on skin cancer cells are a hydroxyl with a methoxy moiety, as found in MM1.
Effects of magnolol and MM1 on the clonogenic and anchorage-independent growth ability of skin cancer cells. Next, we examined the effects of magnolol and the highly potent compound MM1 on the clonogenic and anchorage-independent growth ability of A375 and A431 cells. As shown in Figure 2a, MM1 displayed significant greater suppression of clonogenic ability than magnolol when treated at 50 μM. Similarly, MM1 displayed a significantly greater suppression of anchorage-independent growth ability than magnolol when treated at 75 μM in A375 cells ( Figure 2b). A431 cells are incapable of anchorageindependent growth, and thus were not tested in this set of experiments. These results indicated that MM1 has a significant inhibitory effect than magnolol in the colony formation and anchorage-independent growth of skin cancer cells.
Effects of magnolol and MM1 on the induction of apoptosis in melanoma cells. To investigate whether MM1-induced cell death is associated with the induction of apoptosis, we analyzed the kinetics of caspase activation in A375 cells. As shown in Figure 3a, the cleavages of caspase-7, caspase-8, caspase-9, and PARP were readily detected as early as 24 h after cells were exposed to 75 μM MM1. Morphological examination using phase-contrast microscopy indicated that while the untreated control cells displayed a normal morphology, the MM1-treated cells displayed a loss of adherence, condensation of cytoplasm, and the formation of apoptotic bodies, which are all indicative of apoptosis ( Figure 3b). These results indicate that MM1 displays a greater ability than magnolol to induce apoptosis in A375 cells. Treatment of A375 cells with the broad-spectrum caspase inhibitor, z-VAD-fmk, effectively inhibited the MM1triggered inductions of apoptotic phenotypes ( Figure 3b) and caspase activity (Figure 3c), indicating that apoptosis plays a critical role in the cytotoxic mechanism of MM1.
Antitumor effects of magnolol and MM1 in vivo. The in vivo antitumor activities of magnolol and MM1 were evaluated using A375 cell xenografts in nude mice. When the A375 cell xenografts reached about 50 mm 3 , we administered magnolol or MM1 0.1 μmol in 100 μl of acetone by intra-peritoneal (i.p.) injection three times per week. Tumor volumes were measured at the indicated times. As shown in Figure 4a, the i.p. administration of MM1 significantly reduced tumor growth, whereas magnolol had little such effect. Under our experimental conditions, there was no noticeable change in body weight or overt sign of toxicity in the treated mice ( Figure 4b). The histological appearance of the excised tumor xenograft by hematoxylin-eosin staining revealed that the entire tumor mass consisted of melanoma cells (Figure 4c; upper panel). Immunohistochemistry (IHC) staining of the excised tumor xenografts revealed no staining of activated caspase-3 in untreated tumor xenografts, whereas weak and heavy staining of activated caspase-3 were detected in magnolol-and MM1-treated tumors, respectively ( Figure 4c; lower panel). In agreement with the results of our in vitro studies, these results indicate that MM1 displays greater antitumor activity than magnolol against skin cancer cells in vivo.
Transcriptomic and pathway analyses of MM1-treated cells. To explore the molecular actions of MM1 on A375 cells, we used Affymetrix human GeneChips to compare the gene expression profiles of A375 cells treated with or without MM1. A total of 1270 genes displayed differential expression over 1.5-fold between treated and untreated A375 cells ( Figure 5a). We subjected these genes to pathway analyses using the DAVID Database web server, and performed gene ontology prediction. Our results suggested that MM1 could affect certain key factors involved in regulating apoptosis and programmed cell death ( Figure 5b). Since recent studies have indicated that long non-coding RNAs (lncRNAs) are also involved in regulating various physiologic functions, including proliferation, migration, and apoptosis, [28][29][30][31][32] we examined whether any of the differentially expressed genes represented lncRNAs. We found that the differentially expressed genes included 64 lncRNAs. The 20 putative target lncRNAs with the highest scores are listed in Table 1.
They include lncRNA growth arrest-specific 5 (GAS5), which is a known tumor suppressor that promotes growth arrest and/or apoptosis in multiple cell types. 33,34 Overexpression of lncRNA GAS5 was shown to reduce invasion in human melanoma cells, 34 while knockdown of lncRNA GAS5 was reported to abolish cell cycle arrest at G1 phase in stomach cancer. 35 We thus selected lncRNA GAS5 for further study.
To address the potential importance of lncRNA GAS5 upregulation for the antitumor activity of MM1, we first examined whether MM1 more effectively induced lncRNA GAS5 compared to magnolol. As shown in Figure 5c, quantitative real-time RT-PCR revealed that lncRNA GAS5 was increased to a higher level in MM1-treated cells compared to magnolol-treated cells. lncRNA GAS5 is downregulated in skin cancer tissues.
To explore the possibility that the expression of lncRNA GAS5 may be relatively low in human skin cancer and thus amenable to the treatment by MM1, we employed quantitative RT-PCR to examine the relative expression levels of lncRNA GAS5 in the skin cancer tissues (T) and adjacent normal skin tissues (N) from 25 skin cancer patients. The ratio of relative expression (T/N) was smaller than 0.5 in 18 out of the 25 patients entered in this study ( Figure 7a). Therefore, lower expression of lncRNA GAS5 appears to prevail in human skin cancer and strategy of targeting lncRNA GAS5 could be considered as another approach for the treatment of skin cancer.

Discussion
We previously reported that MM1, a new derivative of magnolol bearing a methoxy moiety, exhibits significantly improved anti-inflammatory activity compared to magnolol, yet has negligible skin toxicity. 27 Here we report that MM1 also displays better antitumor activity than magnolol against  (Figures 1,2) and in a xenograft animal model in vivo (Figure 4). In addition, similar results were also observed in oral cancer cell lines (Supplementary Figure 1). These findings suggest that MM1 could be more therapeutically relevant than magnolol for clinical applications.
Regarding the molecular basis underlying the improved antitumor activity of MM1, we found that MM1 appears to have  (Table 1), which are now recognized as a major component of the human transcriptome, and have been shown to regulate many key biological processes. 36 The differentially expressed lncRNAs included GAS5 (upregulated), which is a tumor suppressor lncRNA that promotes growth arrest and/or apoptosis in multiple cell types. 33,34 Our real-time RT-PCR assay confirmed that lncRNA GAS5 is indeed upregulated in MM1-treated cells (Figure 5c), suggesting that this lncRNA may contribute to the enhanced apoptosis induction and improved antitumor and anti-metastatic activities of MM1. Indeed, knockdown of lncRNA GAS5 reduced MM1-mediated apoptosis in A375 cells (Figure 6a). In contrast, overexpression of lncRNA GAS5 decreases cell viability and promotes the apoptosis of skin cancer cells (Figures 6b and c). Therefore, MM1-induced upregulation of lncRNA GAS5 can promote apoptosis and inhibit cell proliferation in skin cancer cells (Figure 7b). Our finding is consistent with the known activity of lncRNA GAS5 to induce apoptosis in several cancer cells. 34 As yet, the molecular basis of lncRNA GAS5-mediated induction of apoptosis remains poorly understood at present. Further investigations on the mechanisms of lncRNA GAS5-induced apoptosis are urgently needed.
In this context, it is interesting to note that the expression of lncRNA GAS5 is relatively low in most of the human skin cancer tissues when compared to their adjacent normal tissues (Figure 7a). Low expression of lncRNA GAS5 was also observed in several other cancers, including breast and lung cancers. 37,38 Cancer patients with low expression of lncRNA GAS5 may be amenable to the treatment by MM1. Further test of MM1 for its potential clinical application against human cancers is warranted.

Materials and Methods
Culture media and antibodies. Culture media and fetal bovine serum were purchased from Life Technologies (Grand Island, NY, USA). Antibodies against cleaved PARP (Asp214) and caspase-7, -8, and -9 were purchased from Cell Signaling (Temecula, CA, USA). The antibody against β-actin was purchased from Sigma (St. Louis, MO, USA).
Compounds. Magnolol was isolated from Magnolia cortex, 27 and MM1 was prepared as described by Lin et al. 27 Tissues. The human skin cancer and corresponding non-cancerous normal tissues used in this study were obtained from 25 skin cancer patients who underwent surgical resection at Lin-Kou Chang Gung Memorial Hospital. This study was approved by the Ethics Committee of Chang Gung Memorial Hospital, and written informed consent was obtained from each patient.
Cell culture. A375 (melanoma) and A431 (squamous cell carcinoma) cells were obtained from the American Type Culture Collection (Manassas, VA, USA). SAS, a high-grade tumorigenic human tongue squamous cell carcinoma, was obtained from the Japanese Collection of Research Bioresources (Tokyo, Japan). 39 The cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 0.5 mM sodium pyruvate, 1.2 g/l sodium bicarbonate, and 2.5 mM L-glutamine.
SRB colorimetric assay for cytotoxicity screening. Cells were seeded in 96-well tissue plates at a density of 4 × 10 4 cells in 200 μl medium per well. Compounds were added at the indicated concentrations or wells were left untreated as controls, and the plates were incubated for 24 h. Viable cells were detected using the SRB colorimetric method. Briefly, 150 μl 50% trichloroacetic acid (Sigma) was added to each well and the plates were incubated at 4°C for 1 h. The plates were then washed five times with ddH 2 O, and then air-dried. SRB solution (50 μl/well; Sigma) was added, and the plates were incubated for 30 min at room temperature. The plates were washed five times with 1% acetic acid, air dried, and treated with 10 mM Tris base (pH 10). Absorbance was measured using a microplate reader at 565 nm.
Assays for cell proliferation, clonogenic ability, and anchorageindependent growth. Assays for cell proliferation (MTT), clonogenic ability, and anchorage-independent growth were performed as described previously.
Flow cytometric analysis of cell cycle. Cells were fixed in − 20°C absolute ethanol for 4 h and resuspended in 1 ml of PBS containing 20 μg/ml DNase-free, RNaseA. After incubating at 37 o C for 30 min, the cells were treated Western blotting. Western blotting was performed as described previously. 39 Detection of lncRNA GAS5 by quantitative real time. RNA from skin cancer cells was extracted as previously described. 40 RNA from each tissue was isolated using an RNeasy mini kit (QIAGEN, Gaithersburg, MD, USA), followed by treatment with RQ1 RNase-free DNase (Promega, Madison, WI, USA) according to the manufacturer's instructions. Two micrograms of treated RNA sample was subjected to reverse transcription (RT). 41 lncRNA GAS5 was detected by quantitative real time RT-PCR using the TaqMan non-coding RNA expression assay (product number Hs03464472 m1; Applied Biosystems) as described previously. 42 GAPDH was used as an internal control.
RNA interference (RNAi). Target genes were downregulated by RNAimediated inhibition of RNA expression, using a mixture of four siRNAs for each target gene (ON-TARGETplus SMARTpool; Dharmacon, Lafayette, CO, USA) as previously described. 40 The siGENOME nontargeting siRNA pool (Dharmacon) was used as the control. The selected siRNA sequences were submitted to BLAST searches to ensure that each siRNA targeted only one human gene. The four siRNAs targeting human lncRNA GAS5 (GenBank accession no. NR_002578.2) covered the following: nucleotides 385-403 from the start codon (lncRNA GAS5-1: Annexin V-FITC and propidium iodide staining assay. A375 overexpressed with pCDNA3.1-lncRNA GAS5 or vector control was plated in 6-well plates. After 48 h incubation, the cells were collected and subjected to Annexin V and propidium iodide staining by using Vybrant Apoptosis Assay Kit 2 (Invitrogen) according to the manufacturer's protocol. After staining, flow cytometry was performed to quantify apoptotic cells.
Assay for antitumor activity in vivo. A xenograft mouse model was used to examine the antitumor activity of the tested compounds against human skin cancer cells. Six-week-old male BALB/c nu/nu mice (National Laboratory Animal Center, Taipei City, Taiwan) were maintained in an SPF (specific pathogen free) environment. Animals (n = 6 per group) were inoculated subcutaneously in the right flank with A375 tumor cells (2 × 10 6 ) in a volume of 100 μl on day 0. Treatment was initiated when the tumors were 50 mm 3 . Mice were randomized to three groups: treatment with magnolol, MM1, or vehicle (control). The tested agents were administered by i.p. injection three times per week. Mice were randomly divided into three groups when the tumors were 50 mm 3 and treated with acetone as control, magnolol, or MM1 0.1 μmol in 100 μl of acetone by i.p. injection three times per week. Tumor width (W) and length (L) were measured once a week with a caliper, and the tumor volume (V) was calculated according to the formula: V = 0.5 × W 2 × L. All animal experiments were performed in accordance with the guidelines for the Animal Care Ethics Commission of Chang Gung University under an approved animal protocol. 42 IHC and hematoxylin-eosin staining. IHC and hematoxylin-eosin staining were performed as described previously. 42 The primary antibody used for IHC staining targeted activated cleaved-caspase 3 (9546; Cell Signaling).
Statistics. Statistical differences were evaluated using the Student's t-test, and were considered significant at Po0.05. The presented results are representative of three independent experiments with similar results.

Conflict of Interest
The authors declare no conflict of interest.