(+)-2-(1-Hydroxyl-4-Oxocyclohexyl) Ethyl Caffeate Suppresses Solar UV-Induced Skin Carcinogenesis by Targeting PI3K, ERK1/2, and p38

Skin cancer is one of the most common types of cancers in the United States (1). Epidemiologic studies on cancer incidence show that in the last decade, skin cancer in men has increased, which is in contrast to the decreases observed in 3 major cancers, prostate, lung, and colon (2, 3). In any type of skin cancer, the major etiologic factor is solar ultraviolet (solar UV, i.e., sunlight) irradiation (4, 5). Solar UV consists of UVA, comprising the largest portion (95%), which has a relatively weaker effect on causing skin cancer compared with UVB. UVB comprises a smaller portion (5%) of solar UV, but is considered to be a complete carcinogen (6). UVA and UVB cause DNA damage (7) and activate various signaling pathways for skin tumor promotion (8). Recent reports suggest that PI3K (9, 10) and mitogen-activated protein kinase (MAPK) family members, including ERKs (11–15) and p38 (14–18), are major factors in skin carcinogenesis. This conclusion is based on research findings showing that these kinases activate oncogenic transcription factors, activator protein-1 (AP1) and nuclear factor-κB (NF-κB; refs. 19 and 20). Thus, an effective strategy might be to directly suppress the activities of these kinases or their common regulators (21–24) for chemoprevention of skin cancer.

Elucidating the anticancer effects of plant-derived compounds might be a successful approach for chemoprevention (24). (+)-2-(1-Hydroxyl-4-oxocyclohexyl) ethyl caffeate (HOEC; Fig. 1) is found in Incarvillea mairei var. grandiflora (Wehrhahn) Grierson. Although the plants of the Incarvillea genus are widely grown for ornamental purposes, especially in China, Incarvillea delavyi, one of the Incarvillea spices, recently received attention for its anti-nociceptive effects (25). In particular, HOEC isolated from I. mairei is reported to exhibit anti-inflammatory effects on inflammation-induced mice (26). However, the potential anticancer effects and molecular targets of the plant extract or its lead compound have not yet been studied.

Herein, we report that HOEC directly targets and inhibits the kinase activities of PI3K, ERK1/2, and p38 and downregulates their downstream signaling pathways in solar UV–treated JB6 Cl41 mouse epidermal cells. The inhibition of these pathways leads to a reduction of AP1 and NF-κB transcriptional activities. In an in vivo model of solar UV–induced skin carcinogenesis, HOEC significantly reduces tumor volume and tumor number by inhibiting the PI3K, ERK1/2, and p38 signaling pathways.

HOEC is a natural compound isolated from the whole plants of I. mairei and the compound used in this study was synthesized according to previous reports (27). The purity of HOEC was assessed by high-performance liquid chromatography (HPLC) and found to be greater than 97%. Active p110 (PI3K), active ERK1, active ERK2, active p38α, inactive RSK2, and inactive ATF2 recombinant proteins for kinase assays were purchased from Millipore. Antibodies to detect phosphorylated tyrosines (p-Tyr, i.e., p-p110), phosphorylated Akt (p-Akt S473), phosphorylated ERK1/2 (p-ERK1/2 Thr202/Tyr204), phosphorylated p38 (p-p38 Thr180/Tyr182), phosphorylated RSK2 (p-RSK2 Ser380), phosphorylated MSK1 (p-MSK1 Ser376), phosphorylated ATF2 (p-ATF2 Thr69/71), phosphorylated S6 ribosomal protein (p-S6 ribosomal protein Ser235/236), phosphorylated c-Fos (p-c-Fos Ser32), phosphorylated c-Jun (p-c-Jun Ser63), total p110, total ERKs, total RSK, total Akt, total ATF2, total MSK, total S6 ribosomal protein, total c-Fos, and total c-Jun were purchased from Cell Signaling Technology. The antibody against β-actin was from Santa Cruz Biotechnology. CNBr-sepharose 4B beads were obtained from Amersham Pharmacia Biotech. The luciferase assay substrate and the Cell Titer 96 Aqueous One Solution Reagent [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (MTS)] Kit for the cell proliferation assay were from Promega.

The JB6 Cl41 murine epidermal cell line (a promotion-sensitive clone of the JB6 P+ cell line) was cultured in Eagle's minimum essential medium (MEM) and HaCaT human keratinocytes were cultured in Dulbecco's modified Eagle medium (DMEM)/high glucose containing penicillin (100 units/mL), streptomycin (100 μg/mL), and 4% or 10% fetal bovine serum (FBS; Gemini Bio-Products), respectively. Cells were maintained in a 5% CO₂, 37°C humidified incubator. Cells were cytogenetically tested and authenticated before being frozen. Each vial of frozen cells was thawed and maintained in culture for a maximum of 8 weeks.

The kinase assay was performed according to the instructions provided by Millipore. Briefly, for ERK1, ERK2, or p38α activity analysis, the relevant active protein (100 ng) was incubated with HOEC (0, 10, or 20 μmol/L) for 30 minutes at 30°C. Then each reaction mixture was mixed with isotope-unlabeled ATP and 10 μCi [γ-³²P] ATP with each compound in 10 μL of reaction buffer containing 20 mmol/L HEPES (pH 7.4), 10 mmol/L MgCl₂, 10 mmol/L MnCl₂, and 1 mmol/L dithiothreitol (DTT). After incubation at 30°C for 30 minutes, the reaction was stopped by adding 5 μL protein loading buffer and the mixture was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). For measuring PI3K activity, an active p110α protein (100 ng) was incubated with HOEC (0, 10, or 20 μmol/L) for 30 minutes at 30°C. Then, 20 μL of 0.5 mg/mL phosphatidylinositol (Avanti Polar Lipids) were added, and the mixture was incubated for 5 minutes at room temperature. Reaction buffer (100 mmol/L HEPES, pH 7.6, 50 mmol/L MgCl₂, 250 μmol/L ATP) containing 10 μCi [γ-³²P] ATP was added, and the mixture was incubated for an additional 30 minutes at 30°C. The reaction was stopped by the addition of 15 μL of 4 N HCl and 130 μL of chloroform:methanol (1:1) and mixed by vortexing. The lower chloroform phase (30 μL) was spotted onto a 1% potassium oxalate–coated silica-gel plate (previously activated for 1 hour at 110°C) and subjected to thin-layer chromatography and autoradiography to visualize the ³²P-labeled phosphatidylinositol 3-phosphate (PIP) product. Each experiment was repeated twice and the relative amounts of incorporated radioactivity were assessed by autoradiography.

Recombinant p110α, ERK1, ERK2, or p38α (200 ng) proteins or solar UV–treated JB6 Cl41 cell lysates (500 μg) were incubated with HOEC-sepharose 4B (or sepharose 4B only as a control) beads (50 μL 50% slurry) in reaction buffer (50 mmol/L Tris, pH 7.5, 5 mmol/L EDTA, 150 mmol/L NaCl, 1 mmol/L DTT, 0.01% NP-40, and 2 mg/mL bovine serum albumin). After incubation with gentle rocking overnight at 4°C, the beads were washed 3 times with buffer (50 mmol/L Tris, pH 7.5, 5 mmol/L EDTA, 150 mmol/L NaCl, 1 mmol/L DTT, and 0.01% NP-40) and binding was visualized by Western blotting.

To determine whether HOEC inhibits a protein's kinase activity by interacting with the protein's ATP-binding pocket, we performed an ATP and HOEC competition assay. ATP (0, 1, 10, or 100 μmol/L) was mixed with active p110α, ERK1, ERK2, or p38α (200 ng) for 30 minutes. After incubation, HOEC-sepharose 4B (or sepharose 4B only as a control) beads (50 μL 50% slurry) were added in reaction buffer and kept at 4°C overnight. The beads were washed and proteins (p110, ERK1, ERK2, or p38) were detected by Western blot analysis.

The solar UV source comprised UVA-340 lamps purchased from Q-Lab Corporation. The UVA-340 lamps provide the best possible simulation of sunlight in the critical short wavelength region from 365 nm down to the solar cutoff of 295 nm with a peak emission of 340 nm (28). The percentage of UVA and UVB emitted from UVA-340 lamps was measured by a UV radiometer and was recorded as 94.5% and 5.5%, respectively. Specifically, uncovered dishes of serum-starved cells were exposed to UV irradiation under UVA-340 lamps. The dose of solar UV radiation was 48 kJ/m² UVA/2.9 kJ/m² UVB, which was the same dose used in the animal study.

Cells were disrupted on ice for 40 minutes in cell lysis buffer [20 mmol/L Tris pH 7.5, 150 mmol/L NaCl, 1 mmol/L Na₂EDTA, 1 mmol/L EGTA, 1% Triton X-100, 2.5 mmol/L sodium pyrophosphate, 1 mmol/L β-glycerophosphate, 1 mmol/L sodium vanadate, and 1 mmol/L phenylmethylsulfonyl fluoride (PMSF)]. After centrifugation at 12,000 rpm for 10 minutes, the supernatant fraction was harvested as the total cellular protein extract. The protein concentration was determined using the Bio-Rad protein assay reagent. Total cellular protein extracts were separated by SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membranes in 20 mmol/L Tris-HCl (pH 8.0), containing 150 mmol/L glycine and 20% (v/v) methanol. Membranes were blocked with 5% nonfat dry milk in 1 × TBS containing 0.05% Tween 20 (TBS-T) and incubated with antibodies against p-Tyr (p-p110), p110α, p-Akt, Akt, p-ERK1/2, ERK1/2, p-RSK, RSK, p-p38, p38, p-ATF2, ATF2, p-S6 ribosomal protein (p-S6), S6 ribosomal protein (S6), p-MSK1, MSK1, p-c-Fos, c-Fos, p-c-Jun, c-Jun, or β-actin. Blots were washed 3 times in 1× TBS-T buffer, followed by incubation with an appropriate horseradish peroxidase (HRP)–linked IgG. The specific proteins in the blots were visualized using the enhanced chemiluminescence (ECL) detection reagent.

To identify the potential binding proteins of HOEC, a shape similarity method (28) using the PHASE module of Schrödinger's molecular modeling software package was performed to search for biological targets of HOEC based on its structure. The atom-type parameter for volume was set at pharmacophore, which means the queries were used not only to consider shape similarity but also to align potential pharmacophore points between the queries and the targets. The target library was obtained from the Protein Data Bank (29) and our in-house library. To provide more structural orientations for possible alignment, we set the maximum number of conformers per molecule to be generated to 100, while retaining up to 10 conformers per rotatable bond. We filtered out conformers with similarity below 0.75. We then obtained the PDB ID associated with each aligned target molecule and used these PDB IDs to search the online Protein Data Bank to identify the protein. Based on the shape similarity results, ERK1/2, PI3K and p38α were identified as potential protein targets of HOEC. To study the HOEC interaction with PI2K, ERK1/2, and p38α, computational docking and modeling studies were performed. First, an X-ray diffraction structure of p110 with a resolution of 2.8 Å (PDB ID 3HHM, Chain A; ref. 30), an X-ray diffraction structure of ERK1 with a resolution of 2.39 Å (PDB ID 2ZOQ; ref. 31), an X-ray diffraction structure of ERK2 with a resolution of 2.4 Å (PDB ID 2OJJ; ref. 32) and an X-ray diffraction structure of p38 with a resolution of 2.9 Å (PDB ID 3DS6; ref. 33) were downloaded. Hydrogen atoms were added consistent with a pH of 7 and all water molecules removed. Finally, an ATP-binding site–based receptor grid for docking with each respective kinase was generated. HOEC was prepared for docking by default parameters using the LigPrep program in Schrödinger. Then HOEC-protein docking was accomplished with the program Glide using default parameters under the extra precision (XP) mode allowing the acquisition of the best-docked representative structure.

JB6 Cl41 cells stably transfected with an AP1 or NF-κB luciferase reporter plasmid were seeded (1 × 10⁴ viable cells/well) into a 24-well plate. Cells were incubated overnight at 37°C in a humidified atmosphere of 5% CO₂ and were starved in serum-free medium for another 24 hours. The cells were exposed to solar UV (48 kJ/m² UVA/2.9 kJ/m²) and then treated with HOEC (0, 10, or 20 μmol/L). After 6 hours, cells were harvested and disrupted with 100 μL of lysis buffer [0.1 mmol/L potassium phosphate pH 7.8, 1% Triton X-100, 1 μmol/L dithiothreitol (DTT), and 2 μmol/L EDTA]. Then firefly luciferase activities were measured by a luminometer (Luminoskan Ascent, Thermo Fisher Scientific) using substrates provided in the reporter assay system (Promega).

Female SKH-1 hairless mice, 6 weeks old, were purchased from Charles River and maintained under conditions based on the guidelines established by Research Animal Resources and the University of Minnesota Institutional Animal Care and Use Committee (IACUC). Skin carcinogenesis in mice was induced using UVA-340 lamps. Mice were divided into 5 groups. In the control groups, the dorsal skin was topically treated with acetone as vehicle (n = 9) or HOEC (0.5 mg/mouse, n = 12) in 200 μL acetone. In the solar UV–treated group (n = 21), the dorsal skin was topically treated with 200 μL acetone after exposure to solar UV. The mice in 1 group (0.5 mg/mouse, n = 20) received topical application of HOEC for a total of 30 weeks three times a week. The UV absorption spectra of HOEC indicated that HOEC absorbed UV in the solar UV range between 250 and 380 nm, which means that HOEC could act as a sun-blocking agent. Therefore, to eliminate the UV blocking effect of HOEC, we treated mice with HOEC for 12 weeks after solar UV exposure and for an additional 18 weeks without UV exposure.

The HOEC (0.5 mg/mouse) was dissolved in 200 μL of acetone, which is sufficient to cover all of the dorsal skin area of each mouse, which measured approximately 2 × 4.5 cm². The dose of solar UV was gradually increased by 10% each week for 6 weeks to acclimate the mice to the protocol and prevent sunburn and hyperplasia while maintaining the tumor-promoting effect of solar UV. At week 1, the solar UV dose was 30 kJ/m² UVA/1.8 kJ/m² UVB given twice/week. At week 6, the dose was 48 kJ/m² UVA/2.9 kJ/m² UVB and this dose was maintained for weeks 6 to 12 and then UV treatment was discontinued and tumor growth was monitored for an additional 18 weeks. A tumor was defined as an outgrowth of at least 1 mm in diameter that persisted for 2 weeks or more. Mouse weight and tumor number and volume were recorded every week until the end of the experiment (30 weeks). Tumors and skin were frozen for Western blot analysis. Tumor volume was calculated according to the following formula: tumor volume (mm³) = length × width × height × 0.52.

All quantitative results are expressed as mean values ± SD. Statistically significant differences were obtained using the Student t test or by one-way ANOVA. A P-value of <0.05 was considered to be statistically significant.

To better understand how HOEC interacts with p110α (Fig. 2Aa-b), ERK1 (Fig. 2Ba-b), ERK2 (Fig. 2Ba, c), or p38α (Fig. 2Ca-b), we performed a computational docking study using the Glide docking program included in Schrödinger Suite 2012. In the docked models, HOEC binds well at the ATP-binding pockets of all 4 kinases (Fig. 2). Taken together with the in vitro and ex vivo pull-down assay results (Supplementary Fig. S1A–S1C), our data confirmed that HOEC binds to the respective ATP-binding site and inhibits the kinase activity of p110α, ERK1, ERK2, or p38α. The images were generated using the UCSF Chimera program (34).

To rule out the possibility that the inhibitory effect of HOEC was because of toxicity, we conducted a cell toxicity assay and results indicated that HOEC had no cytotoxic effect on normal JB6 Cl41 murine epidermal cells or HaCaT human keratinocytes (data not shown). To identify direct target(s) of HOEC, we conducted several in vitro kinase assays. HOEC selectively inhibited PI3K (10 μmol/L, 48%; 20 μmol/L, 97%; Supplementary Fig. S1A-a), ERK1 (10 μmol/L, 6%; 20 μmol/L, 16%; Supplementary Fig. S1A-b), ERK2 (10 μmol/L, 8%; 20 μmol/L, 54%; Supplementary Fig. S1A-c), and p38α (10 μmol/L, 53%; 20 μmol/L, 68%; Supplementary Fig. S1A-d) kinase activities in a concentration-dependent manner, but had no effect on EGFR or JNKs kinase activity (data not shown). To confirm and identify additional kinase targets of HOEC, we used a commercial in vitro kinase profiling system (Kinase Profiler, Millipore). The Kinase Profiler results showed that all other kinases tested were not affected by HOEC (Supplementary Table S1). To determine whether HOEC interacts directly with p110α, ERK1, ERK2, or p38α, we performed in vitro (Supplementary Fig. S1B-a) and ex vivo (Supplementary Fig. S1B-b) pull-down binding assays. Recombinant p110α, ERK1, ERK2, and p38α strongly interacted with HOEC-conjugated sepharose 4B beads (Supplementary Fig. S1B-a). In addition, these proteins in solar UV–treated JB6 Cl41 cell lysates were pulled down with HOEC-sepharose 4B beads (Supplementary Fig. S1B-b). To determine whether the binding of HOEC to p110α, ERK, or p38α occurs in an ATP-competitive manner, we performed an ATP (0, 1, 10, or 100 μmol/L) competitive kinase pull-down assay. Results showed that the binding of HOEC to p110α, ERK1, ERK2, or p38α decreased in an ATP-competitive manner (Supplementary Fig. S1C). These results suggest that PI3K, ERK1, ERK2, and p38α are major targets of HOEC.

Our data showed that HOEC binds to the ATP-binding site of p110α, ERK1, ERK2, or p38α (Fig. 2) and suppresses their respective activity (Supplementary Fig. S1A-a–S1A-d). Therefore, the inhibitory effects of HOEC likely suppress the respective downstream signaling pathways. The PI3K/Akt and MAPKs pathways are highly activated in UV-induced skin cancers (10, 11, 16). HOEC decreased solar UV–induced phosphorylation of PI3K downstream proteins, including Akt (10 μmol/L, 38%; 20 μmol/L, 62%) and S6 ribosomal protein (S6; 10 μmol/L, 5%; 20 μmol/L, 62%) compared with the same kinases in solar UV–induced HOEC, untreated JB6 Cl41 cells (Fig. 3A). ERK1/2 and p38 are MAPK family members and HOEC affects these MAPK signaling pathways. RSK2 and MSK1 are ERKs downstream signaling proteins and ATF2, c-Jun, and c-Fos act downstream of p38. Results indicated that HOEC (20 μmol/L) suppressed solar UV–induced phosphorylation of RSK2 by 85%, MSK1 by 64% (Fig. 3B), c-Fos by 29%, and c-Jun by 64% (Fig. 3C). These findings suggest that HOEC suppresses solar UV–induced activation of PI3K/Akt and MAPKs signaling pathways mediated by their well-known upstream kinases, PI3K, ERK1, ERK2, or p38.

After stimulation with UV, the transcriptional activities of AP1 and NF-κB are induced through the PI3K/Akt or MAPKs signaling pathways (8, 35). To examine the effect of HOEC on the transcriptional activity of AP1 and NF-κB, we exposed JB6 Cl41 cells stably transfected with the AP1 or NF-κB luciferase reporter plasmid to solar UV (48 kJ/m² UVA/2.9 kJ/m² UVB) and then treated cells with HOEC for another 6 hours. HOEC (20 μmol/L) suppressed solar UV–induced transactivation of AP1 by 47% and NF-κB by 80% (Fig. 4).

To determine the chemopreventive effects of HOEC in vivo, we utilized a solar UV–induced skin carcinogenesis mouse model. We treated mice with solar UV and then topically applied vehicle or HOEC to the dorsal surface of SKH-1 hairless mice. Volume (Fig. 5A) and number (Fig. 5B) of tumors in mice treated with HOEC (0.5 mg/mouse) after solar UV exposure were significantly inhibited by HOEC (P < 0.05). At week 30, all mice were sacrificed and mouse skin and tumor tissues were collected for Western blot analysis. Phosphorylation of p110, Akt, S6K, RSK, MSK, p38, ATF2, c-Jun, and c-Fos was induced by solar UV and HOEC strongly suppressed the phosphorylation (Fig. 5C and Supplementary Fig. S2). The phosphorylation level of HOEC-targeted kinases was quantified by densitometry (Fig. 5C). RSK2, MSK1, S6K40, and c-Jun are well-known substrates of ERK1/2, p110, and p38, respectively. The phosphorylation levels of RSK2, MSK1, S6K40, and c-Jun were significantly induced by solar UV treatment and strongly inhibited by HOEC (40%, 50%, 77%, and 77% inhibition, respectively) in skin tissues (Fig. 5C). In addition, phosphorylated Akt, ATF2, and c-Fos levels were also decreased by HOEC treatment (Supplementary Fig. S2). PCNA and the cleavage status of caspase-7 or PARP are representative markers of cell growth or cell death, respectively. Moreover, HOEC substantially decreased PCNA levels in skin tissues. However, HOEC did not affect caspase-7 or PARP cleavage. Note that HOEC was topically applied after UV exposure to exclude the possibility of any sunblock effect of the HOEC compound (Supplementary Fig. S3). To eliminate the possibility that residual HOEC could remain on the skin between UV exposures, we examined the amount of HOEC remaining in the dorsal skin tissues (500 mg) from mice sacrificed at 0, 24, and 48 hours after HOEC treatment. Mass spectrometry (LC/MS) results indicated that skin tissues harvested at 24 or 48 hours after HOEC treatment showed decreased HOEC levels of 50% or 90%, respectively, compared with skin treated with HOEC and harvested immediately (data not shown). This result supports the idea that residual HOEC does not block penetration of subsequent UV irradiation because most (90%) of the HOEC had disappeared within 48 hours. Overall, our data show that HOEC binds its target kinases and inhibits their respective activities, blocking signal transduction to their downstream molecules resulting in decreased cell growth but not necessarily cell death. These results suggest that HOEC is a robust multitargeted chemopreventive agent against solar UV–induced skin carcinogenesis that acts through mechanisms other than, but not entirely excluding, a sunblock effect.

Cancer chemoprevention with natural compounds is a promising strategy because of the ability of these compounds to inhibit multiple signaling pathways activated by carcinogens (24) and tumor promotors. Solar UV, comprising both UVA and UVB, is a major factor in the increased incidence of skin cancers, including nonmelanoma skin cancer and melanoma (36, 37). Moreover, multiple signaling pathways are activated by UV exposure and lead to the acquisition of gene alterations in skin cells (38, 39). Therefore, identifying chemopreventive compounds with multiple molecular targets against solar UV–induced skin carcinogenesis is crucial because of the continued increasing incidence of this cancer. Several groups have reported the skin photoprotective effects of various phytochemicals (11, 22, 40). We used the natural compound, HOEC, which is a caffeic ester from I. mairei var. grandiflora (Wehrhahn) Grierson (27, 41). HOEC was revealed to act as an anti-inflammatory agent targeting 5-LOX, which confirmed a key role of 5-LOX in the pathogenesis of inflammation, especially rheumatoid arthritis (26). HOEC may be a prodrug of caffeic acid because of the chemical relevance and biofunctional similarity. However, the mechanism of the anticancer effects and direct targets of HOEC have not yet been elucidated.

In this study, we found that HOEC directly binds with PI3K, ERK1/2, and p38 and inhibits their kinases activities. PI3K, ERK1/2, and p38 are known to play critical roles in several cancers, including skin cancer (9, 12, 16, 39). The PI3K/Akt signaling pathways are activated by UV irradiation and can induce apoptosis in human keratinocytes (42). ERK1/2 signaling pathways are also stimulated by UV irradiation and UVB irradiation of keratinocytes results in the activation of ERK1/2 and p38 MAPK pathways (8, 14, 16, 39). Animal study results also revealed that treatment of mouse skin with HOEC inhibits solar UV–induced skin carcinogenesis by attenuating PI3K, ERK1/2, and p38 activation leading to decreased downstream signalings. The effects of HOEC on UV-induced oxidative stress or UV-induced inflammation have not yet been elucidated. However, HOEC is a novel compound that suppresses solar UV–induced skin carcinogenesis by directly binding and inhibiting the activity of multiple kinases, resulting in the attenuation of UV-induced inflammation and skin carcinogenesis (14). Thus, we suggest that HOEC is a novel compound that functions as a multikinase inhibitor of solar UV–induced skin cancer.

The pursuit of molecular protein targets of natural compounds in skin cancer or other cancers is challenging. HOEC is a novel and unique compound because it targets at least 3 important proteins that play significant roles in solar UV–induced skin cancer, which makes it very potent in suppressing solar UV–induced skin carcinogenesis. Compared with taxifolin (43) or norathyriol (11), which each effectively suppressed solar UV–induced skin carcinogenesis in mice, HOEC has stronger effects on its target kinases' activities and downstream transcription activities. For example, the effect of HOEC against PI3K (IC₅₀ = 10.3 μmol/L), is 7.6-fold greater than that of taxifolin (IC₅₀ = 78.4 μmol/L). The inhibition of ERK2 and p38 by HOEC (IC₅₀ = 20.9 and 12.9 μmol/L, respectively) is 1.4- and 2.3-fold greater than the inhibition of EGFR by taxifolin (IC₅₀ = 29.0 μmol/L). Moreover, the inhibition of AP1 and NF-κB transcription activity by HOEC was 1.4- and 2.8-fold greater than the effect of norathyriol (11).

UV-induced activation of these kinases leads to increased AP1 and NF-κB transcription activities (43, 44). With UV irradiation, AP1 is stimulated in human keratinocytes and promotes tumor development and UVB-induced photocarcinogenesis is reduced by repression of AP1 and NF-κB activation (45, 46). Because AP1 and NF-κB activate many oncogenes in carcinogenesis, the inhibitory effect of HOEC on AP1 and NF-κB transactivation activity might account for its ability to suppress EGF-induced anchorage-independent growth of JB6 Cl41 P+ murine epidermal cells and HaCaT human keratinocytes.

This result is also supported by our in silico identification and computational docking model. The theory of shape-similarity screening is derived from the idea that molecules possessing similar shape and electrostatic capabilities might exhibit analogous biological activity. According to our in silico computer screening results, HOEC is similar in shape to well-known kinase inhibitors of PI3K, ERK2, and p38. The specific similarity score of HOEC to each kinase is 0.83 with (Z)-5-((5-(4-fluoro-2-hydroxyphenyl)furan-2-yl)methylene)thiazolidine-2,4-dione (47), which is a known ERK2 inhibitor; 0.83 with 1-(4-bromophenyl)-2-(5-(furan-2-yl)-4H-1,2,4-triazol-3-ylthio)ethanone (48), which is a known PI3K α inhibitor; and 0.81 with 1-(5-tetra-butyl-2-methyl-2H-pyrazol-3-YL)-3-(4-chloro-phenyl)-urea (BMU; ref. 49), which is a known p38 inhibitor, respectively. Based on these results, ERK1/2, PI3K, and p38 are potential protein targets of HOEC.

Most studies focusing on UV-induced signaling have involved mainly UVB irradiation. However, in this study, we used solar UV treatment, which comprises both UVA and UVB, making it similar to our natural environment. An effective strategy to suppress solar UV–induced skin carcinogenesis might include the use of a multitargeted inhibitor because solar UV activates multiple signaling pathways. We and others have found that determining the most important target among many signals activated by UV is highly challenging and to attempt to suppress only one signaling molecule is probably not feasible (24). Notably, the chemopreventive effects of HOEC are not because of its absorption of UV (as described in the Materials and Methods), but because of a direct inhibition of key signaling pathways.

In summary, HOEC attenuated solar UV–induced PI3K, ERK1/2, and p38 signaling pathways in JB6 Cl41 cells and decreased skin tumor formation in SKH-1 hairless mice exposed to solar UV. The chemopreventive effect of HOEC on skin carcinogenesis in vitro and in vivo is because of direct binding and inhibition of PI3K, ERK1/2, and p38 activation, resulting in suppression of AP1 and NF-κB transactivation. Treatment with HOEC resulted in decreased cell growth but did not necessarily cause cell death in preventing skin carcinogenesis. Based on our results, we conclude that the identification of HOEC as a chemopreventive agent against skin cancer might contribute to the further development of novel multitargeted agents to prevent solar UV–induced skin carcinogenesis.

No potential conflicts of interest were disclosed.

Conception and design: D.Y. Lim, H. Li, W.-D. Zhang, Z. Dong

Development of methodology: M.-H. Lee, W.-D. Zhang, Z. Dong

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D.Y. Lim, S.H. Shin, L. Shan, W.-D. Zhang, Z. Dong

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D.Y. Lim, M.-H. Lee, S.H. Shin, H. Chen, J. Ryu, L. Shan, A.M. Bode, W.-D. Zhang, Z. Dong

Writing, review, and or revision of the manuscript: D.Y. Lim, M.-H. Lee, S.H. Shin, H. Chen, L. Shan, A.M. Bode, Z. Dong

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M.-H. Lee, W.-D. Zhang, Z. Dong

Study supervision: A.M. Bode, W.-D. Zhang, Z. Dong

The authors thank T. Schuster and N. Oi for assistance with experiments and N. Brickman for assisting with article submission.

This work was supported by The Hormel Foundation and NIH grants CA166011, CA172457, R37 CA081064, ES016548, and CA027502 (to Z. Dong).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.