Male ICR mice (8 weeks old) were obtained from the Laboratory Animal Center of Xi'an Jiaotong University (Xi'an, China). The animals were housed under controlled temperature (23±2°C), humidity (55±5%) and light (12-h light/dark cycle) with free access to standard diet and water. After acclimatization for 1 week, 72 mice were randomly allocated into two groups: non-irradiated group (n=8) and UV-irradiated group (n=64). The dorsal skin of mice (2×3 cm²) was shaved using an electric razor, and this operation was repeated before UV irradiation and postirradiation treatment. All animal experiments were approved by the Laboratory Animal Administration Committee of Xi'an Jiaotong University and performed according to the Guidelines for Animal Experimentation of Xi'an Jiaotong University and the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication no. 85–23, revised 2011).

UV irradiation of mice was performed using the UV light source provided by 2 UVA lamps (315–400 nm, peak wavelength: 365 nm) and 4 UVB lamps (280–315 nm, peak wavelength: 312 nm) (both from Beijing Lighting Research Institute, Beijing, China). The distance from the lamps to the animals' backs was 30 cm. The minimal erythema dose (MED) was preliminarily measured with a UV meter (Lutron UV-340A, Taipei, Taiwan), and 1,200 mJ/cm² of UVA and 180 mJ/cm² of UVB were assembled 1 MED in this study. Mice were irradiated 3 times a week (Monday, Wednesday and Friday) for 12 weeks. The irradiation dose was increased weekly by 1 MED from 1 MED up to 4 MED and then maintained at 4 MED for the rest weeks. The non-irradiated group was treated identically with the lamps power off.

After 12 weeks of UV irradiation, mice in the UV-irradiated group were randomly reallocated into eight groups with 8 mice per group: UV-irradiated plus vehicle (ethanol: propylene glycol, 7:3 v/v)-treated group (UV control group), UV-irradiated plus ATRA (Sigma-Aldrich, St. Louis, MO, USA)-treated group (ATRA group), UV-irradiated plus ATRA and AGN193109 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA)-treated group (ATRA + RAR antagonist group), UV-irradiated plus ATRA and HX531 (Tocris Bioscience, Bristol, UK)-treated group (ATRA + RXR antagonist group), UV-irradiated plus TTNPB (Sigma-Aldrich)-treated group (RAR agonist group), UV-irradiated plus TTNPB and AGN193109-treated group (RAR agonist + RAR antagonist group), UV-irradiated plus SR11237 (Sigma-Aldrich)-treated group (RXR agonist group) and UV-irradiated plus SR11237 and HX531-treated group (RXR agonist + RXR antagonist group). ATRA and retinoid receptor-specific agonists and antagonists were applied topically 5 times a week in 100 µl vehicle per treatment for 8 weeks. According to previous studies (16–19), these agonists and antagonists were applied at the following concentrations with slight modification: ATRA, 160 nM; TTNPB, 160 nM; AGN193109, 400 nM; SR11237, 160 nM; HX531, 400 nM. Eight mice in the non-irradiated group were used as the normal group and treated with vehicle alone.

Twenty-four hour after the final treatment, mice were sacrificed by cervical dislocation under anesthesia, and dorsal skin was quickly removed. The skin samples were fixed in 10% neutral buffered formalin for 24 h, embedded in paraffin and sectioned at 5 µm. Masson's trichrome staining was performed to examine the skin collagen fibers. The stained sections were captured under an Olympus BX51 light microscope equipped with a DP70 digital camera (Olympus, Tokyo, Japan).

Hydroxyproline (Hyp) can be converted to the equivalent of collagen by multiplying the factor 7.46, considering Hyp is the almost exclusive amino acid of collagen and accounts for 13.4±0.24% of mammalian collagen in previous studies (20,21). Hence, in this study, total Hyp content in the skin was measured using the Hyp assay kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer's instruction for the determination of collagen content.

Fresh skin samples were homogenized in ice-cold RIPA lysis buffer (Heart Biological Technology, Co., Ltd., Xi'an, China). Lysates were centrifuged at 14,000 × g for 10 min at 4°C, and the supernatants were collected as the total proteins. Protein concentration was measured using the BCA protein assay kit (Beyotime Institute of Biotechnology, Shanghai, China). Each sample was subsequently denatured by boiling in Laemmli loading buffer for 5 min. Equal amounts of protein were separated by 8–10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to PVDF membranes (Millipore, Billerica, MA, USA). After blocking with 5% non-fat dried milk for 2 h at 37°C, the membranes were incubated overnight at 4°C with the following primary antibodies: rabbit anti-MMP-3 monoclonal antibody (Cat. no. ab52915; 1:1,000 dilution), rabbit anti-MMP-13 polyclonal antibody (Cat. no. ab39012; 1:1,000 dilution) (both from Abcam, Cambridge, MA, USA), rabbit anti-c-Jun monoclonal antibody (Cat. no. 9165; 1:1,000 dilution), rabbit anti-c-Fos monoclonal antibody (Cat. no. 2250; 1:1,000 dilution) (both from Cell Signaling Technology, Inc., Danvers, MA, USA), goat anti-type I procollagen polyclonal antibody (Cat. no. sc-8787; 1:500 dilution) and mouse anti-β-actin monoclonal antibody (Cat. no. sc-47,778; 1:1,000 dilution) (both from Santa Cruz Biotechnology, Inc.). After washing 3 times, the membranes were incubated with the appropriate horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 h at 37°C, followed by enhanced chemiluminescence (Millipore). The signals were captured, and the intensity of the protein bands was quantified using Image J software (NIH, Bethesda, MD, USA) (22).

All data are presented as the means ± SD. Statistical analysis was performed using SPSS 19.0 software (SPSS, Inc., Chicago, IL, USA). Data were analyzed using one-way ANOVA or a Student's t-test to perform comparisons between two groups. A P-value of <0.05 was considered to indicate a statistically significant difference.

Histological sections of the dorsal skin were subjected to Masson's trichrome staining to visualize the changes in collagen fibers in the dermis. As shown in Fig. 1, the normal group displayed regularly arranged collagen fibers. Compared with the normal group, UV irradiation caused large amounts of abnormal, fragmented and disorganized collagen fibers in the UV control group. ATRA and RAR agonist improved the UV-induced damage to collagen fibers, whereas these effects were markedly inhibited by RAR antagonist. Similar changes in collagen fibers were observed either between the ATRA and ATRA + RXR antagonist groups or between the RXR agonist and UV control groups. These results indicated that ATRA and RAR agonist could ameliorate the UV-induced damage to skin collagen fibers through RAR.

The collagen content was determined through measuring the amount of Hyp. As shown in Fig. 2, UV irradiation induced a 34.86% decrease in the collagen content (P<0.01, vs. normal group). ATRA and RAR agonist significantly increased the collagen content by 29.83 and 25.00%, respectively, as compared with the UV control group (all P<0.01, vs. UV control group). However, these effects were significantly suppressed by RAR antagonist (all P<0.01, vs. ATRA and RAR agonist groups, respectively). There was no significant difference in the collagen content either between the ATRA and ATRA + RXR antagonist groups or between the RXR agonist and UV control groups. These results were consistent with those of Masson's trichrome staining.

UV irradiation stimulates collagen breakdown and inhibits procollagen synthesis (1). MMP-3 and MMP-13 are the key regulators of collagen degradation in photoaged mouse skin. To investigate the mechanism underlying the beneficial histological effects of ATRA on photoaged skin, the protein expression of MMP-3, MMP-13 and type I procollagen was detected by western blotting. As shown in Fig. 3, UV irradiation significantly increased the protein levels of MMP-3 and MMP-13 and decreased the protein level of type I procollagen (all P<0.01, vs. normal group). ATRA significantly reduced MMP-3 and MMP-13 protein levels and increased type I procollagen protein level (all P<0.01, vs. UV control group), while these effects were markedly inhibited by RAR antagonist (all P<0.01, vs. ATRA group). There was no significant difference in the protein levels of MMP-3, MMP-13 and type I procollagen between the ATRA and ATRA + RXR antagonist groups. These data indicated that ATRA could stimulate the protein expression of type I procollagen and inhibit the protein expression of MMP-3 and MMP-13 through RAR in photoaged skin.

We also examined the effects of RAR and RXR agonists on the protein expression of MMP-3, MMP-13 and type I procollagen in photoaged skin. As shown in Figs. 4 and 5, UV irradiation resulted in a significant increase in the protein levels of MMP-3 and MMP-13 and a significant decrease in the protein level of type I procollagen (all P<0.01, vs. normal group). RAR agonist prominently reduced MMP-3 and MMP-13 protein levels and increased type I procollagen protein level (all P<0.01, vs. UV control group), whereas these effects were significantly suppressed by RAR antagonist (all P<0.01, vs. RAR agonist group). There was no significant difference in the protein levels of MMP-3, MMP-13 and type I procollagen between the RXR agonist and UV control groups. These data demonstrated that RAR agonist rather than RXR agonist could stimulate the protein expression of type I procollagen and inhibit the protein expression of MMP-3 and MMP-13 through RAR in photoaged skin.

Transcription factor AP-1 is mainly composed of c-Jun and c-Fos proteins and plays an important role in the mechanism of photoaging (3). To investigate the effect of ATRA on AP-1 expression in photoaged skin, the protein expression of c-Jun and c-Fos was detected by western blotting. As shown in Fig. 6, UV irradiation significantly increased the protein level of c-Jun (P<0.01, vs. normal group). ATRA significantly reduced c-Jun protein level (P<0.01, vs. UV control group), while this effect was markedly inhibited by RAR antagonist (P<0.01, vs. ATRA group). There was no significant difference in the protein level of c-Jun between the ATRA and ATRA + RXR antagonist groups. Moreover, no significant difference in the c-Fos protein level was observed between all groups. These results indicated that ATRA could down-regulate the protein expression of c-Jun through RAR in photoaged skin.

We also examined the effects of RAR and RXR agonists on the protein expression of c-Jun and c-Fos in photoaged skin. As shown in Figs. 7 and 8, UV irradiation resulted in a significant increase in the protein level of c-Jun (P<0.01, vs. normal group). RAR agonist prominently reduced c-Jun protein level (P<0.01, vs. UV control group), whereas this effect was significantly suppressed by RAR antagonist (P<0.01, vs. RAR agonist group). There was no significant difference in the protein level of c-Jun between the RXR agonist and UV control groups. Furthermore, RAR and RXR agonists had no significant effect on the c-Fos protein level in photoaged skin. These results demonstrated that RAR agonist rather than RXR agonist could down-regulate the protein expression of c-Jun through RAR in photoaged skin.