The flowers and buds of C. indicum were purchased from Qingping Chinese herbal medicine market of Guangzhou, Guangdong, China, and were identified by Professor Zi-Ren Su of Guangzhou University of Chinese medicine (Guangdong, China). The extraction and identification procedures of CI_SCFE were based on previously published literature (27). Briefly, the flowers and buds of C. indicum were loaded into the extraction vessel of the 532 Supercritical Fluid Extraction Equipment (Applied Separations). CI_SCFE was obtained using an extraction time of 4 h, a pressure of 25 MPa, a temperature of 45˚C and a flow rate of 20 l/h. High-performance liquid chromatography-pulsed amperometric detection (HPLC-PAD) and gas chromatography-mass spectrometry (GC-MS) were used to analyze the chemical composition of CI_SCFE (Table SI and Fig. S1). HPLC analysis was performed on a Shimadzu LC40 HPLC system. The separation was performed on a ACE Excel 5 Super C18 column (4.6x250 mm, 5 µm; cat. no. EXL-1211-2546U; Advanced Chromatography Technologies) with a flow rate of 1.0 ml/min, column temperature at 30˚C, and injection volume of 10 µl. The mobile phase consisting of acetonitrile (solvent A) and 0.1% aqueous formic acid (solvent B) was used to elute the targets with the gradient mode (0-5 min: 5% A→25% A; 5-15 min: 25% A→25% A; 15-25 min: 25% A→45% A; 25-35 min: 45% A→55% A; 35-40 min: 55% A→70% A). Luteolin-7-glucoside (cat. no. B20887), luteolin (cat. no. B20888), linarin (cat. no. B20860) and chlorogenic acid (cat. no. B20782) standards were purchased from Shanghai Yuanye Biotechnology Co., Ltd. The content of these compounds was quantitatively analyzed with peak areas under the standard curves at 334 nm. GC-MS analysis was performed on an Agilent 6890-5975 GC-MS system (Agilent Technologies, Inc.). The oven temperature was initially set at 60˚C, then ramped up to 100˚C at a gradient of 10˚C/min (held for 1 min), then to 110˚C at a rate of 1˚C/min (held for 1 min); then to 150˚C at a rate of 3˚C/min (held for 1 min) and finally to 260˚C at a rate of 10˚C/min (held for 5 min). Split injection (0.5 µl) was conducted with a split ratio of 60:1 and helium was used as carrier gas of 1.0 ml/min flow rate. The spectrometer was set to electron impact (EI) mode with an ionization energy of 70 eV, a scanning range of 40-400 amu, and a scanning rate of 0.34 sec/scan. The temperatures of the inlet and ionization source were 230 and 250˚C, respectively. Ultimately, CI_SCFE at a concentration of 0.48 mg/cm²/mouse (low concentration CI_SCFE) or 1.6 mg/cm²/mouse (high concentration CI_SCFE) in 10% Tween 80 were used for subsequent animal experiments.

A total of 75 specific-pathogen free male Kunming mice (22-24 g, age 8 weeks) were obtained from the Animal Experiment Center, Guangzhou University of Chinese Medicine (Animal Quality Certificate No. 44005800007154). The mice were housed in an environment conforming to the prescribed humidity (50±5%) and temperature (23±2˚C), with food and water ad libitum and were maintained under a 12 h light/dark cycle. The laboratory animal license number was SCXK (Yue) 2018-0085 and the ethics certification number was 20190304024. Under the supervision of authorized researchers, all experiments in the present study were approved by the Animal Care and Use Committee of Guangzhou University of Chinese Medicine (approval no. 20190304024; Guangzhou, China) based on the Guidelines for the ethical review of laboratory animal welfare People's Republic of China National Standard GB/T 35892-2018(29). According to additional markers that may constitute humane endpoints in tumor research, the experimental endpoint for tumor size took into account the fact that UV exposure on the back of mice causes damage to skin tumors. The experiment was halted when the volume of any of the skin lesions >1,000 mm³, the maximum diameter >10 mm or when the diameter ≤10 mm but interfered with animal feeding or hindered animal movement. According to the humane endpoint guidance of the present animal experiments, two mice were euthanized, one from the CI_SCFE-L group and one from NAA group (30).

The skin area on the back (2.5x3.0 cm²) of mice was depilated with a shaver (FS607; FLYCO) (Fig. 1). In our previous study, the concentrations of 0.48 mg/cm²/mouse and 1.6 mg/cm²/mouse of CI_SCFE reduced skin damage caused by UV exposure (28). Therefore, the mice were randomly allocated to five treatment groups: Sham (no medication or UV radiation), model (only UV radiation without drug treatment), low concentration CI_SCFE (0.48 mg/cm²/mouse), high concentration CI_SCFE (1.6 mg/cm²/mouse) and positive control nicotinamide (NAA; 0.65 mg/cm²/mouse; Sigma-Aldrich; Merck KGaA) groups. CI_SCFE and NAA were applied daily to the shaved area of the skin on the back of the mouse.

Ultra-Vitalux light bulbs were used as a source of UV light (UVA:UVB=93:7; Osram). Mice were positioned 30 cm away from the UV lamp at 23˚C and were irradiated on Mondays, Wednesdays, Fridays and Sundays for 31 weeks. According to our previous study (31), the minimum erythema dose (MED) for mice was 100 mJ/cm². UV intensity was 1 MED for the first week and was increased by 1 MED/week until 400 mJ/cm² in the 4th week, which was maintained until the end of the experiment. The experiment continued for 31 weeks, with daily observations of mice health and behavior. At 31 weeks, isoflurane (induction, 5%; maintenance, 2%) was used for anesthesia and cervical dislocation was performed on the mice after photographic documentation of the dorsal skin status. Finally, skin tissue from the irradiated area was extracted for subsequent experiments.

The dorsal skin was fixed in 10% formalin at 26˚C for 48 h. Tissue dehydration was then performed with the following procedure: Soaking in 10% formalin at 26˚C for 1 h, soaked in 75% ethanol at 26˚C for 1.5 h, soaked in 85% ethanol at 26˚C for 1.5 h, soaking in 95% ethanol at 26˚C for 1.5 h, soaking in anhydrous ethanol at 26˚C for 1 h three times, soaking in TO Bio-permeable agent (cat. no. AYA0150; Shanghai Acmec Biochemical Technology Co., Ltd.) at 26˚C for 1.5 h twice and soaking in paraffin wax at 60˚C for 2 h. Subsequently, the tissue was embedded in melted paraffin at 65˚C and sectioned (4 µm). Paraffin sections were stained using Gomori aldehyde fuchsin at 26˚C for 10 min (GAF; cat. no. G1593; Beijing Solarbio Science & Technology Co., Ltd.), hematoxylin and eosin at 26˚C for 20 min (H&E; cat. no. 0619A19; Beijing Leagene Biotechnology Co., Ltd.) and Sirius red staining at 26˚C for 1 h (cat. no. DC0041-100; Beijing Leagene Biotechnology Co. Ltd.). The histological changes in the skin were examined using a light microscope (BX53; Olympus Corporation).

The procedure for fixation, embedding and sectioning of skin tissue is the same as in the Histological analysis section. Skin sections (5 µm) were subsequently deparaffinized with the following procedure: Placed in xylene at 26˚C for 10 min three times, in anhydrous ethanol at 26˚C for 5 min and three times, in 95% ethanol at 26˚C for 3 min, in 85% ethanol at 26˚C for 3 min and in 75% ethanol at 26˚C for 3 min. The skin sections were heated in EDTA for 15 min at 95˚C and then the sections were soaked in 3% hydrogen peroxide at 26˚C for 20 min. Tissue sections were then sealed with 10% goat serum (cat. no. SL038; Beijing Solarbio Science & Technology Co., Ltd.) for 30 min at 37˚C. Subsequently, sections were incubated with Ki-67 (1:300; cat. no. ab15580; Abcam) and CD11b (1:4,000; cat. no. ab133357; Abcam) antibodies diluted in PBS at 4˚C overnight. Samples were incubated with anti-rabbit IgG H&L (HRP; 1:1,000; cat. no. ab6721; Abcam) at 37˚C for 1 h, followed by incubation with 3,3-diaminobenzidine at 26˚C for 1 min (cat. no. ZLI-9017; ZSGB-BIO) and counterstaining with hematoxylin at 26˚C for 3 min. Sections were imaged using a light microscope (Olympus Co.; BX53). Samples were semi-quantified using ImageJ software (version 1.53e; National Institutes of Health).

At week 31, the UV-irradiated dorsal skin of mice was removed and analyzed using a ROS assay. The mouse skin was encapsulated in optimal cutting temperature encapsulant (cat. no. 4583; Sakura Finetek USA, Inc.) and frozen sections (-80˚C; thickness, 8 µm) were incubated with DCFH-DA (cat. no. BB18081; Bestbio) at 37˚C for 30 min. The samples were measured at a wavelength of 525 nm using a fluorescence microscope (BX53; Olympus Corporation). The average fluorescence intensity of ROS was analyzed using ImageJ software (version 1.53e; National Institutes of Health).

Skin tissue was homogenized by adding 9 times the volume of saline (g/ml) and the supernatant was collected after centrifugation at 3,000 x g for 10 min at 4˚C. The protein concentration of the supernatant was measured using a BCA kit (cat. no. P0012; Beyotime Institute of Biotechnology). CAT and SOD levels were measured according to the manufacturer instructions of the CAT (cat. no. A007-1; Nanjing Jiancheng Bioengineering Institute) and SOD (cat. no. A001-3; Nanjing Jiancheng Bioengineering Institute) assay kits.

Skin tissues were ground in PBS with a homogenizer (KZ-III-FP; Wuhan Servicebio Technology Co., Ltd.) at 60 Hz for 6 min at 4˚C and subsequently centrifuged at 4˚C and 3,000 x g for 20 min. The supernatant was collected and ELISA kits were used accordingly to the manufacturer's instructions to measure the levels of TNF-α (cat. no. 430904; BioLegend, Inc.), IL-6 (cat. no. 431304; BioLegend, Inc.) and 8-OHdG (cat. no. MM-0221M1; Jiangsu Meimian Industrial Co., Ltd.).

Skin samples were homogenized in RIPA lysis solution (cat. no. P0013B; Beyotime Institute of Biotechnology). Samples were centrifuged at 4˚C and 14,000 x g for 10 min and the protein content was determined using a BCA kit (cat. no. P0010; Beyotime Institute of Biotechnology). The proteins (40 µg/lane) were electrophoresed using a 10% SDS-polyacrylamide gel and transferred to PVDF membranes. PVDF membranes were blocked with 5% skimmed milk for 1 h at 26˚C and incubated overnight at 4˚C with heme oxygenase 1 (HO-1; 1:1,000; cat. no. ab13248; Abcam), CD11b (1:1,000; cat. no. ab133357; Abcam), VEGF (1:1,000; cat. no. SC7269; Santa Cruz Biotechnology, Inc.), c-Myc (1:1,000; cat. no. SC40; Santa Cruz Biotechnology, Inc.), p65 (1:1,000; cat. no. 8242S; Cell Signaling Technology, Inc.), PTEN (1:1,000; cat. no. ET1606-43; HUABIO), phosphorylated (p)-p62 (1:1,000; cat. no. ab211324; Abcam), NAD-dependent protein deacetylase sirtuin-1 (SIRT1; 1:1,000; cat. no. ab110304; Abcam), p-p65 (1:1,000; cat. no. ab76302; Abcam), p-IκBα (1:1,000; cat. no. ab133462; Abcam), acetyl-p65 (1:250; cat. no. ab19870; Abcam, IκBα (1:1,000; cat. no. ab32518; Abcam), Nrf2 (1:1,000; cat. no. ab137550; Abcam), NAD(P)H dehydrogenase [quinone] 1 (NQO1; 1:50,000; cat. no. ab80588; Abcam), GAPDH (1:1,000; cat. no. 5174; Cell Signaling Technology, Inc.), lamin B1 (1:2,000; cat. no. AF5161; Affinity Biosciences) and Keap1 (1:2,500; cat. no. ab139729; Abcam) antibodies. Subsequently, membranes were incubated with anti-mouse IgG H&L (1:5,000; cat. no. LK2003; Tianjin Sungene Biotech Co., Ltd.) or anti-rabbit IgG H&L (1:5,000; cat. no. ab6721; Abcam) antibodies for 1 h at 26˚C. Blots were visualized using ECL reagents (cat. no. FD8000; Hangzhou Fude Biotechnology Co., Ltd.) and the blot densities were quantified using ImageJ software (version 1.53e; National Institutes of Health). GAPDH or Lamin B1 were used as loading controls (32).

Data were expressed as mean ± standard deviation. Data were analyzed using a one-way ANOVA followed by Tukey's post hoc test. P<0.05 was considered to indicate a statistically significant difference. Data were visualized and analyzed using GraphPad software (version 8.3.0; Dotmatics).

GC-MS analysis and HPLC analysis were used to detect the chemical constituents of CI_SCFE. As shown in Table SI, CI_SCFE mainly contains d-Camphor, Caryophyllene oxide, Endo-Borneol, α-Curcumene, Cis-verbenol, β-Caryophyllene, Eucalyptol, Thymol as detected by GC-MS. In addition, HPLC analysis detected four compounds in CI_SCFE (Fig. S1), which were Chlorogenic acid, Luteolin-7-glucoside, Linarin and Luteolin.

Over the course of the present study, it was demonstrated that the skin of the model group showed shallow wrinkles, erythema and a leathery appearance after 9 weeks of UV exposure compared with the sham group (Fig. 2A). The skin in the low dose and high dose CI_SCFE groups and the NAA group exhibited no erythema and showed few wrinkles compared with the model group. After 24 weeks of UV irradiation, papular lesions and broken crusts were observed on the skin of model mice. At week 31, ulcerative papules, adhesive scales, recurrent local bleeding and crusting on the skin was observed on model mice. However, after pretreatment of mice with low or high dose CI_SCFE and NAA, there were no papules and deep wrinkles only appeared on the skin at 31 weeks of UV irradiation. Moreover, the dermal vessels in the model group were expanded and proliferated compared with the sham group after 31 weeks of UV irradiation. After pretreatment with CI_SCFE, at both doses tested, vasodilation and hyperplasia in the dermis caused by UV were markedly reduced (Fig. 2B).

To observe the histopathologic changes of mouse skin after UV irradiation, GAF, Sirius red and H&E staining were used (Fig. 3A). H&E staining showed that the skin of model mice exhibited abnormal proliferation of the epidermis, keratinocytes extending into the dermis and extensive infiltration of inflammatory cells at week 31. Sirius red and GAF staining of the mouse skin showed that UV irradiation damaged elastic and collagen fibers and reduced their density. Nevertheless, compared with the model group, low and high dose CI_SCFE and NAA treatment significantly reduced inflammatory cell infiltration and the deformation and degradation of elastin fibers and collagen fibers and markedly reduced UV-induced abnormal epidermal hyperplasia (Fig. 3B; P<0.01).

To examine the effects of CI_SCFE on proliferation, angiogenesis and the expression levels of cancer-related proteins, immunohistochemical assays and western blotting were used to detect Ki-67, VEGF, c-Myc and PTEN expression levels. After 31 weeks of UV irradiation, the epidermis of the model mice exhibited increased Ki-67 expression compared with the sham group (Fig. 4A and B; P<0.001). The epidermal layer in the CI_SCFE and NAA groups demonstrated a significant reduction in Ki-67 expression compared with the model group (P<0.001). VEGF and c-Myc protein expression levels were significantly increased and protein expression levels of PTEN were significantly decreased in the model mice compared with the sham group (Fig. 4C-F; P<0.001). After pretreatment with CI_SCFE or NAA, VEGF and c-Myc protein expression levels were significantly reduced while PTEN protein expression levels were significantly increased compared with the model group (P<0.05).

To investigate whether the inhibition of skin cancer progression by CI_SCFE is related to its antioxidant effects, the levels of ROS, SOD, CAT and 8-OHdG were assayed. An increase in ROS accumulation was observed in the model group compared with the sham group (Fig. 5A and B). However, compared with the model group, CI_SCFE and NAA treatment significantly reduced UV-induced ROS overexpression (P<0.05). The activity levels of SOD and CAT were significantly decreased and 8-OHdG was significantly increased in the skin of the model mice at 31 weeks compared with the sham group (Fig. 5C-E; P<0.05). In mice treated with CI_SCFE and NAA, the levels of SOD and CAT activity were significantly increased (P<0.05), while 8-OHdG levels were significantly decreased in mice treated with low dose CI_SCFE compared with those in the model group (P<0.01).

To explore the level of inflammation in irradiated mouse skin, the expression levels of CD11b, IL-6 and TNF-α were assayed. After 31 weeks of UV irradiation, an increase in CD11b expression was observed in the dermis of model mice compared with the sham group (Fig. 5A and F; P<0.001). The protein expression level of CD11b in the skin of mice treated with CI_SCFE and NAA was significantly reduced compared with the model group (Fig. 5G and H; P<0.01). The expression levels of IL-6 and TNF-α in the model group were significantly higher compared with the sham group (Fig. 5I and J; P<0.05). Compared with the model group, topical application of CI_SCFE and NAA significantly reduced the expression levels of IL-6 and TNF-α (P<0.05).

High expression of Nrf2 in tumor cells promotes tumor development (15). Furthermore, NFκB is an important inflammatory and oncogenic transcription factor, and activation of SIRT1 inhibits the NF-κB pathway and suppresses the inflammatory response (21). To explore the potential mechanism of action of CI_SCFE against skin cancer in mice, the p62/Keap1-Nrf2 and SIRT1/NF-κB pathway-related proteins were examined. The results showed that the protein expression levels of p62, p-p62, Nrf2, HO-1 and NQO1 were significantly increased, while that of Keap1 significantly decreased in the model group compared with the sham group (Fig. 6; P<0.05). Furthermore, CI_SCFE treatment significantly decreased the expression levels of p62, p-p62, Nrf2, HO-1 and NQO1 and significantly increased that of Keap1 compared with the model group (P<0.05). However, the p-p62/p62 ratio was not significantly altered in any treatment condition.

Furthermore, there was a significant increase in the protein expression levels of p65, p-p65, acetyl-p65 and p-IκBα in the skin of the model group compared with the sham group (Fig. 7; P<0.05). However, the aforementioned protein expression levels were significantly lower in the CI_SCFE group compared with the model group (P<0.05). Additionally, the expression levels of IκBα and SIRT1 were significantly lower in the model group after 31 weeks of UV irradiation compared with the sham group (P<0.01). Topical pretreatment of mice with CI_SCFE significantly increased the protein expression level of IκBα compared with the model group (P<0.05). High dose treatment of mice with CI_SCFE significantly increased protein expression levels of SIRT1 compared with the model group (P<0.05). Furthermore, p-IκBα/IκBα and p-p65/p65 expression ratios were significantly increased in the model group compared with the sham group (P<0.01). Additionally, p-IκBα/IκBα and p-p65/p65 expression ratios in the skin of CI_SCFE mice were significantly decreased compared with the model group (P<0.01).