PLR, puerarin and daidzein were purchased from Biopurify Phytochemicals (Chengdu, China). PLR was validated by Prof. Baojian Wu at Guangzhou University of Chinese Medicine. Vitamin C (VC) was obtained from Yuanye Biotechnology (Shanghai, China). SR8278 was obtained from MedChemExpress (Monmouth Junction, NJ). Biochemical kits for SOD, MDA and HYP were purchased from Jiancheng Bioengineering Institute (Nanjing, China). Biochemical kit for ROS was obtained from Beyotime Biotechnology (Shanghai, China).

200 g of PLR was soaked in 1,000 ml of 80% ethanol solution for 30 min. After reflux and suction filtration, the filtrate was concentrated to 100 ml in vacuum, and then freeze-dried to powder. The extraction yield was 10.5%. The contents of puerarin and daidzein in PLR extract powder were estimated to be 20.8% and 3.16%, respectively, according to a classic LC-MS/MS method (Supplementary Table S1; Supplementary Figure S1 in supplementary material). For cell experiments, PLR extract powder was made into a stock solution (50 mg/ml in DMSO). For animal experiments, PLR extract powder was made into an emulsion (20 mg/ml) as previously described (Ali et al., 2019).

C57BL/6 mice (male, eight-week-old) were purchased from Guangzhou Carrot Biotechnology (Guangzhou, China). Rev-erbα ^−/− mice (C57BL/6 background) have been described (Zhang et al., 2019). All mice were maintained on a standard 12 h/12 h light/dark cycle (light on from 07:00 a.m. to 19:00 p.m. and light off from 19:00 p.m. to 07:00 a.m.) with ad libitum intake of food and water. All animal experiments were conducted with the approval of the Animal Ethics Committee of Guangzhou University of Chinese Medicine (License number: 20210225033, approved on 01/03/2021).

Mice with shaved back were grouped randomly as follows: (1) control group (n = 6); (2) UVB group (n = 6): mouse back skin was irradiated with UVB (120 mJ/cm²) daily and blank emulsion was used to smear the back after irradiating; (3) VC group (n = 6): the emulsion containing VC (40 mg/kg) was used to smear the back after irradiating; and (4) PLR groups, which were divided into three subgroups: low dose group (n = 6, 25 mg/kg), middle dose group (n = 6, 100 mg/kg) and high dose group (n = 6, 400 mg/kg). Mice in PLR groups were smeared with the emulsion containing PLR extract after irradiating. The source of radiation was a narrow band UVB bulb (Philips model PL-9 9W/01/2P) emitting photons with wavelengths between 306 and 316 nm. The radiation flux was measured by using an UV irradiance meter (LS125, Linshang Technology, Shenzhen, China). All mice were treated for 45 days and sacrificed at 14:00 p.m. to collect skin samples for further assays.

L929 mouse fibroblast cells (Procell Biotechnology, Wuhan, China) were used to assess the protective effects of PLR and VC on UVB-induced cell senescence. Cells were cultured in MEM (Procell Biotechnology, PM150410) supplemented with 100 μg/ml streptomycin, 100 U/mL penicillin and 10% FBS (Procell Biotechnology, 164210-50). Once reaching a confluence of ∼90%, cells were subjected to UVB irradiation (100 mJ/cm²) and then treated with VC (6 μg/ml) or PLR extract (0.1, 0.5 or 1 μg/ml). 24 h later, L929 cells were collected for further assays.

Briefly, skin samples were fixed, dehydrated, embedded, and cut into four-μm-thick slices. For histopathological evaluations, skin slices were stained with hematoxylin and eosin and photographed with a Leica DM750 microscope (Leica Microsystems, Switzerland).

A SA-β-gal staining kit (Beyotime, Shanghai, China) was used to detect the β-galactosidase activity. Following VC or PLR treatment, L929 cells were fixed with β-galactosidase fixation solution. After washing twice with PBS, cells were incubated with 1 ml staining solution at 37°C overnight. Staining images were pictured by a Leica DM750 microscope (Leica Microsystems, Switzerland).

Skin samples were homogenized in lysis buffer using a homogenizer (Jingxin Technologies, Shanghai, China). Tissue lysate was centrifuged (10000 ×g) for 10 min. The supernatant was collected to determine the contents of HYP, SOD and MDA using their respective assay kits.

The cell viability was detected by CCK-8 assay (Beyotime, Shanghai, China). Following VC or PLR treatment, L929 cells were cultured with 10 µL CCK-8 solution and 100 µL fresh medium for 30 min. The absorbance at 450 nm was monitored by a BioTek Synergy H1 Multi-Mode Microplate Reader (BioTek Instruments, Bad Friedrichshall, Germany).

DCFH-DA was used as a probe to measure intracellular ROS. In brief, following VC or PLR treatment, L929 cells were loaded with DCFH-DA (10 μM in serum-free medium) and incubated at 37°C for 20 min. Real-time fluorescence was monitored using a BioTek Microplate Reader at 488 nm excitation and 525 nm emission wavelengths.

Gal4 co-transfection assay was performed as described in our previous publication (Wang et al., 2019). Briefly, pGL4.35-Luc reporter (100 ng), Gal4-REV-ERBα-LBD plasmid (200 ng) and pRL-TK vector (10 ng) were co-transfected into L929 cells using Lipofectamine 3,000 (Thermo Fisher Scientific, Waltham, United States). 18 h after transfection, cells were treated with SR8278 (10 µM) or PLR (1 μg/ml) or vehicle for another 24 h. A reporter assay system (Promega, Walldorf, Germany) was used to detect the cellular luciferase activity.

pRL-TK vector (10 ng) and Bmal1-Luc reporter plasmid (250 ng) were co-transfected into L929 cells. The luciferase activities were measured as described in the “Gal4 co-transfection assay” section.

Total RNAs in skin and cell samples were extracted using RNAex Pro Reagent (Accurate Biology, Guangzhou, China). Reverse transcription was performed using the Evo M-MLV RT kit (Accurate Biology, Guangzhou, China). qPCR reaction, with a total reaction volume of 20 μL and 5 ng cDNA, was performed using the 2X SYBR Green Pro Taq HS Premix (Accurate Biology, Guangzhou, China). The conditions of quantitative PCR were as follows: 40 cycles at 95°C for 5 s, 60°C for 15 s and 72°C for 10 s. Gapdh was used as an internal standard. The primers used are listed in Table 1.

Lysed proteins from skin tissues and cells were subjected to SDS-polyacrylamide gel electrophoresis and then transferred onto a polyvinylidene difluoride membranes (Millipore, Bedford, MA). The membranes were incubated with primary antibodies (anti-BMAL1, 1:2000, proteintech, 14268-1-AP) overnight at 4°C, then with secondary antibodies (1:10,000, Lablead, RN0100) for 1 h at room temperature. Immunoreactive bands were visualized by using an enhanced luminescence kit (Merck-Millipore, Bedford, MA) and a Tanon Imaging System (Tanon, Shanghai, China), and quantified with ImageJ software (NIH, Bethesda, MD). GAPDH was used as an internal standard.

Data are expressed as the means ± standard deviation. Statistical differences between the means of two groups were analyzed by Student’s t-test. p < 0.05 (*) was considered to be significantly different.

We first established a skin aging model by irradiating the back of mice with UVB. After 45 days, we found that the mice irradiated with UVB showed aging signs on the back skin, such as wrinkling, roughness, melanosis, and epidermal thickening (Figures 1A,B). HYP is an important component of fibrillar collagen and MMP-1 degrades collagen fibers in extracellular matrix, both of which are closely related to skin aging (Li and Wu, 2018; Gao et al., 2021). We determined the levels of HYP and MMP-1 in the skin of control and UVB-treated mice. We found that the level of HYP was decreased and the expression of Mmp-1 was increased in UVB-treated mice (Figure 1C). In addition, the expression levels of p21 and p53, two genes closely related to skin aging (Chen Q. et al., 2020), were markedly increased in the skin of UVB-treated mice (Figure 1D). These results indicated that the mouse skin aging model was successfully established.

Next, we tested the PLR efficacy against skin aging induced by UVB in mice. We found that skin wrinkles, epidermis thickness, and expression levels of Mmp-1/p21/p53 were decreased, while the level of HYP was increased after treatment with PLR or VC (used as a positive control) in mice with UVB-induced skin aging (Figures 2A–D). Moreover, we determined whether SOD activity and MDA content (two indicators related to oxidation) were altered in mouse skin. We found that UVB inhibited SOD activity and increased the MDA level (Figure 2E). However, PLR dose-dependently promoted SOD activity and decreased the MDA level (Figure 2E). These results indicated dose-dependent protective effects of PLR on photoaging in mice.

To test whether PLR can relieve UVB-induced cell senescence in vitro, L929 cells were treated with VC (positive control) or various concentrations of PLR (0.1, 0.5 and 1 μg/ml) after exposure to UVB. In line with our expectations, UVB decreased cell viability, whereas PLR (in a dose-dependent manner) and VC significantly alleviated the inhibitory effect of UVB on cell viability (Figure 3A). In addition, PLR attenuated the induction of β-galactosidase activity by UVB (Figure 3B). RT-qPCR assay was applied to assess the regulatory effects of PLR on cell senescence-associated genes. Consistent with previous in vivo studies, the expression of Mmp-1/p21/p53 was remarkedly induced after cell exposure to UVB, however, PLR (in a concentration-dependent manner) and VC significantly attenuated the induction effects of UVB on the expression of these genes (Figures 3C,D). Furthermore, we found that PLR and VC suppressed the induction of ROS in UVB-treated cells (Figure 3E). These data clearly indicated that PLR can inhibit UVB-induced cell senescence.

Because previous studies have identified BMAL1 as a key regulator of skin aging (Kondratov et al., 2006), we wondered whether PLR can modulate the expression of BMAL1 to alleviate UVB-induced skin aging. Compared with control mice, BMAL1 mRNA and protein levels were markedly decreased in mouse skin exposed to UVB (Figures 4A,B). In line with this, the expression levels of its target genes such as Cry1, Cry2, Per1, and Per2 were decreased by UVB (Figures 4C,D). Interestingly, PLR treatment increased the expression of BMAL1 and its target genes in aged skin, thereby attenuating the inhibitory effects of UVB on BMAL1 (Figure 4). Likewise, PLR enhanced the expression of BMAL1 and its targets (Cry1, Cry2, Per1 and Per2) in UVB-treated L929 cells (Figure 5). Altogether, these data indicated that PLR can up-regulate the expression of Bmal1, potentially contributing to its protective effect on UVB-induced aging.

BMAL1 regulation of aging is attained mainly through modulation of oxidative stress and ROS homeostasis (Töbelmann and Dittmar, 2021; Zhang et al., 2022). BMAL1 promotes oxidative defense by up-regulating the expression of NRF2 (a major regulator of antioxidant enzymes) and of the major antioxidant enzymes (Early et al., 2018; Chhunchha et al., 2020). Given that PLR increases Bmal1 expression, we next tested whether PLR also up-regulates the expression of NRF2 and major antioxidant enzymes. We found that the expression levels of Nrf2, Prdx6, Sod1, and Sod2 were decreased in the skin of UVB-treated mice (Figure 6A). However, PLR treatment rescued the inhibitory effects of UVB on Nrf2 and the antioxidant genes in mice with skin aging (Figure 6A). Likewise, PLR treatment led to increased mRNA levels of Nrf2, Prdx6, Sod1, and Sod2 in UVB-treated L929 cells (Figure 6B). These results demonstrated that PLR increased the expression of Nrf2 and antioxidant enzymes, consistent with its promoting effect on Bmal1 expression.

REV-ERBα is a known negative regulator of BMAL1 and a druggable target (Burris, 2008; Patke et al., 2020). We wondered whether PLR regulates the expression of Bmal1 through REV-ERBα. Like SR8278 (a known REV-ERBα antagonist), PLR increased the luciferase activities in the Gal4-REV-ERBα-LBD chimeric assays, indicating PLR as an antagonist of REV-ERBα (Figure 7A). Furthermore, PLR enhanced the Bmal1 promoter activity according to a luciferase reporter assay in L929 cells (Figure 7B). Moreover, we found that PLR dose-dependently increased the expression of Bmal1 and the genes involved in antioxidant responses such as Nrf2, Prdx6, Sod1, and Sod2 (Figures 7C,D). Notably, PLR did not alter the viability and total RNA level of normal L929 cells (Supplementary Figure S2). Therefore, PLR was identified as a REV-ERBα antagonist which can induce Bmal1 expression and promote antioxidant genes.

Lastly, we tested whether REV-ERBα is indeed necessary for the pharmacological effects of PLR on skin aging. We generated Rev-erbα-deficient (Rev-erbα ^−/−) mice, and tested the PLR effect using this mouse line. As expected, Rev-erbα mRNA was absent and Bmal1 was significantly increased in the skin of the knockouts (Figure 8A). Rev-erbα deficiency attenuated the UVB-induced skin aging phenotype due to enhanced Bmal1 expression, as evidenced by reduced skin wrinkles and epidermal thickness as well as decreased Mmp-1, p21 and p53 expression but an increased HYP level (Figures 8B–E). In the meantime, Rev-erbα deficiency increased SOD activity and reduced the MDA level (Figure 9A) and caused elevations in mRNA levels of Nrf2, Prdx6, Sod1, and Sod2 in mouse skin after UVB treatment (Figure 9B). In contrast, these skin properties did not change in Rev-erbα-deficient mice under a non-irradiated condition (Supplementary Figure S3). Remarkably, PLR showed no effect on UVB-induced skin aging in Rev-erbα ^−/− mice based on histological examinations and measurements of aging- and ROS-related markers as well as antioxidant factors (Figures 8, 9). Collectively, these results indicated that REV-ERBα as a drug target was required for the protective effect of PLR on UVB-induced skin aging.