Dimethylsulfoxide (DMSO) and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) were purchased from Solarbio (Beijing, China). Fenofibrate and the FABP4 inhibitor BMS309403 were obtained from Sigma-Aldrich (St. Louis, MO, USA). 4’-6-diamidino-2-phenylindole (DAPI) and Hoechst stains were purchased from Beyotime Biotech (Nantong, China). A SmartFlare uptake control probe (positive control) and FABP4 mRNA-specific SmartFlare probe were obtained from Millipore (Billerica, MA, USA). BODIPY fluorophore 493/503 for lipid droplets was obtained from Molecular Probes (Eugene, OR, USA). Adenoviruses (Ad-NC and Ad-FABP4) were obtained from HanBio (Shanghai, China).

Protocols for experiments involving animals were approved by the Animal Experimentation Ethics Committee at Soochow University (Suzhou, China). Male Sprague-Dawley (SD) rats (4 weeks old) and male C57 mice (4 weeks old) were purchased from the Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai, China). For irradiation, the rats were anesthetized with an intraperitoneal injection of ketamine (75 mg/kg) and xylazine (10 mg/kg), and the hair on the rat buttocks was shaved using a razor. A 3-cm-thick piece of lead was used to shield the rats and localize the radiation field (3 cm $\times$ 4 cm). A single dose of 45 Gy irradiation [23, 24, 25] was administered to the hindlimb region of the SD rats at a dose rate of 750 cGy/min using a 6-MeV electron beam accelerator (Clinac 2100EX, Varian Medical Systems, Inc., CA, USA). This dose was selected because it can significantly induce skin injury [23, 24, 25, 26]. For the treatment, the rats were then randomly assigned to receive treatments by subcutaneous injection of DMSO, fenofibrate, or adenovirus [26].

Total RNA was extracted from cells and tissues with Trizol reagent (Invitrogen, Carlsbad, CA, USA). PPAR$\alpha$ and FABP4 mRNA levels were quantified by quantitative real-time PCR as reported previously [27]. The primers used are listed in Supplementary Table 1.

Human skin samples were obtained from a victim of an iridium radiation accident as reported previously [28]. The skin samples were obtained 160 days after irradiation from the right limb, which was exposed to iridium-192 (${}^{192}$Ir) metal chain (with an activity of 966.4 GBq or 26.1 Ci). Normal skin tissues were obtained when performing skin grafting from the dorsal myocutaneous flap. Informed consent for sample collection was obtained from the patient.

Skin tissues from mouse, rat and monkey were obtained as reported previously [24]. Mouse skin tissues were irradiated with 35 Gy electron beam. Rat skin tissues were irradiated with 45 Gy electron beam. The skin tissues of monkeys were treated with 0 or 20 Gy irradiation. Skin tissues were fixed in 10% neutral-buffered formalin and embedded in paraffin. Three-micrometer paraffin sections were deparaffinized and heat treated with citrate buffer (pH 6.0) for 7 min following an epitope retrieval protocol. Three-micrometer paraffin sections were incubated with a rabbit anti-PPAR$\alpha$ antibody (Abcam, Cambridge, MA, USA, #ab 8934) at 4 °C overnight, followed by incubation with an anti-rabbit biotinylated secondary antibody (Beyotime, Nantong, China), diaminobenzidine substrate detection, washing, hematoxylin staining, dehydration, and mounting.

Tissue MDA levels were determined using thiobarbituric acid (TBA) assays as reported previously [24]. MDA levels were normalized to those of the control group.

ROS levels were determined using the ROS-sensitive dye 2,7-dichlorofluorescein diacetate (DCF-DA) (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). The cells were washed with PBS and incubated with DCF-DA (10 $\mu$M) for 30 min. Skin tissues were trypsinized into single cell suspension according to the manufacturer’s instructions. The level of DCF fluorescence, which reflects the ROS concentration, was observed with a fluorescence microscope. DCF fluorescence levels in skin cells and tissues were quantified at 488 nm using a 96-well plate reader.

Human keratinocyte HaCaT cells, human fibroblast WS1 cells [24, 25, 26] and primary skin fibroblasts were maintained in Dulbecco’s modified Eagle’s medium (DMEM). All culture media was supplemented with 10% FBS (Gibco, Grand Island, NY, USA). Cells were grown at 37 °C in 5% CO${}_2$ incubators. The cells were exposed to different dosages of ionizing radiation using an X-ray linear accelerator (Rad Source, Suwanee, GA, USA) and a fixed dose rate of 1.15 Gy/min.

Cells were incubated with DMSO or fenofibrate. Cell viability was measured using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assays. The cells were incubated for 4 h with 200 $\mu$g/mL MTT (Sigma, St Louis, MO, USA). The reagent was dissolved in DMSO (Solon, OH, USA). The absorbance values were measured at 490 nm using a 96-well plate reader. The experiments were performed in triplicate.

Cells were fixed in 4% paraformaldehyde, washed with PBS, and permeabilized with 1% Triton X-100 in PBS. The cells were then blocked with blocking buffer (PBS, 1% Triton X-100, and 5% BSA) and incubated at 4 °C with a PPAR$\alpha$ (Abcam, #ab 8934) or $\gamma$H2AX antibody (Abcam; #ab 81299) overnight. Next, a rhodamine-conjugated goat anti-rabbit antibody (1:100) was added for 30 min at room temperature. The nuclei were counterstained with DAPI.

The luciferase reporter with four PPREs in luciferase promoter was a kind gift from Dr. Zengpeng Li (Third Institute of Oceanography, State Oceanic Administration, Xiamen, China). The plasmid was verified by sequencing. Cells were transfected with the constructed vectors using Fugene HD transfection reagent (Promega, Madison, WI, USA). For each transfection, 50 ng pRL-TK (Promega) was used to enhance the transfection efficiency. Measurement of luciferase activity using the dual luciferase reporter assay system (Promega). Promoter activity was expressed as the ratio of firefly luciferase activity to Renilla luciferase activity.

Detailed descriptions are given as previously described [25]. Briefly, the membranes were blotted with antibodies against PPAR$\alpha$ (Abcam, Cambridge, MA, USA, #ab227074), GAPDH (Abcam, #ab181602), and FABP4 (Abcam, #ab 92501).

Cells were pretreated with DMSO or fenofibrate and then exposed to irradiation. Apoptosis was measured using a 7-AAD/Annexin-V double staining apoptosis kit (BD Biosciences, Franklin Lakes, NJ, USA) and flow cytometry (BD Biosciences, CA, USA). The Annexin-V+/7-AAD- cells indicated early apoptosis, and the Annexin-V+/7-AAD+ cells indicated late apoptosis. The percentages of both types of cells were counted.

WS1 cell nuclear protein was extracted using a nuclear protein isolation kit (Beyotime). The sequences for the double-stranded oligonucleotide probes (Supplementary Table 2) were synthesized and labeled with biotin by Shanghai Sangon Biotech Co. Ltd. (Shanghai, China). EMSAs were performed according to the LightShift EMSA Kit instructions (Pierce, Rockford, IL, USA).

The data are expressed as the mean $\pm$ SEM of at least three independent experiments. The results were evaluated via one-way ANOVA to determine statistical significance. The statistical analyses were performed using Prism 8 (GraphPad software, San Diego, CA). The differences were considered significant at p 0.05.

We firs0074 analyzed the response of PPAR$\alpha$ to ionizing radiation in multiple animal models. Rats were irradiated with a 45 Gy electron beam as reported previously [26, 27]. The real-time PCR analysis results showed that PPAR$\alpha$ mRNA levels in the irradiated skin tissues were 26.62% of those in the nonirradiated skin tissues. This result is consistent with our RNA-Seq data (GEO database accession number GSE86252) [28]. Next, we attempted to confirm the expression of PPAR$\alpha$ in skin tissue after irradiation by immunohistochemistry in different animal models. The results showed that the expression of PPAR$\alpha$ in the skin tissues of mice, rats, and monkeys after irradiation was significantly lower than that of the nonirradiated control group (Fig. 1B). Moreover, in the irradiated epidermis of a human patient, the expression of PPAR$\alpha$ was decreased, with pronounced distribution from the nucleus to the cytosol (Fig. 1C), indicating PPAR$\alpha$ inactivation in irradiated skin cells. In addition, ionizing radiation downregulated PPAR$\alpha$ protein levels in a dose-dependent manner in human skin fibroblast WS1 and human keratinocyte HaCaT cells (Fig. 1D).

Fig. 1.

Ionizing radiation decreases the expression of PPAR$\mathrm{\alpha }$expression in skin tissues. (A) PPAR$\alpha$ mRNA levels in irradiated and nonirradiated skin tissues of rats (n = 5). PPAR$\alpha$ mRNA expression was measured by real-time PCR. The data are shown as the mean $\pm$ SEM. (B) The expression of PPAR$\alpha$ in irradiated and nonirradiated skin tissues of mouse, rat and monkey. Skin tissues were collected three days after indicated radiation doses. PPAR$\alpha$ expression was measured by IHC as described in the Materials and Methods section. (C) The expression of PPAR$\alpha$ in nonirradiated and irradiated human skin tissues. (D) Western blotting analyses of PPAR$\alpha$ expression at different doses of radiation in WS1 and HaCaT cells. (E) Quantitative analysis of Western blotting assay. Data are depicted as the mean $\pm$ SD from three independent experiments. * p 0.05; ** p 0.01, compared with the control group.

Because fenofibrate is a synthetic fibrate ligand of PPAR$\alpha$, we next explored its effect on PPAR$\alpha$ activation and its influence on the radiosensitivity of cultured skin cells. HaCaT cells were exposed to fenofibrate and then subjected to immunofluorescence for PPAR$\alpha$ detection. The results indicated that fenofibrate induces the translocation of PPAR$\alpha$ into the nucleus (Fig. 2A). Moreover, the activity of the PPRE harboring luciferase reporter was significantly increased after fenofibrate addition; this result confirmed PPAR$\alpha$ activation by fenofibrate (Fig. 2B).

Fig. 2.

Fenofibrate-meditated PPAR$\alpha$ activation protects skin cells against radiation. (A) HaCaT cells were treated with 25 and 50 $\mu$M fenofibrate, and immunofluorescence was performed to investigate PPAR$\alpha$ translocation. (B) The effect of fenofibrate on PPAR$\alpha$ activity was measured with a PPRE luciferase reporter. Luciferase activity was assayed 24 h after transfection. The firefly luciferase activity of each sample was normalized to the Renilla luciferase activity. The final luciferase activity was normalized to that of the control group. (C) HaCaT and (D) WS1 cells were pretreated with fenofibrate and subjected to 0 or 20 Gy irradiation. One hour later, the cellular ROS levels of each group of cells were determined using a DCF-DA probe. Cellular fluorescence was observed using a fluorescence microscope. ROS levels were quantified with a microplate reader. (E) HaCaT cells were pretreated with DMSO or fenofibrate and then irradiated. Mitochondrial membrane potential was evaluated using JC-1 staining. (F) HaCaT cells were pretreated with 25 and 50 $\mu$M fenofibrate. Then, the cells were mock irradiated or irradiated with 20-Gy X-rays. Cell apoptosis rates were detected with Annexin-V/7-AAD staining at (F) 48 h and (G) 72 h after irradiation. The data are shown as the mean $\pm$ SEM of three independent experiments. (H) HaCaT cells were treated with DMSO or fenofibrate. Representative photomicrographs of BODIPY fluorophore 493/503 staining for lipid droplets. The cells were observed with a confocal microscope (Olympus, Tokyo, Japan). (I) Quantification of the ratio of JC-1 aggregate to JC-1 monomer and ATP contents. (J) Quantification of BODIPY fluorophore 493/503 staining for lipid droplets. Data are depicted as the mean $\pm$ SD from three independent experiments. *p 0.05; ** p 0.01, compared with the control group.

Because ionizing radiation elicits cutaneous free radical reactions [3, 4], we examined whether PPAR$\alpha$ activation confers protection against radiation-induced oxidative damage. Fenofibrate concentrations of up to 50 $\mu$M did not significantly affect viability in HaCaT and WS1 cells (Supplementary Fig. 1). We first measured fenofibrate effects on cellular ROS elimination in human HaCaT keratinocytes, WS1 fibroblasts and primary human fibroblasts. HaCaT cells pretreated with 25 or 50 $\mu$M Fenofibrate significantly reduced radiation-induced ROS levels (Fig. 1C). Similar results were obtained in WS1 cells and human primary fibroblasts; in these cells, 50 $\mu$M fenofibrate exhibited the strongest antioxidative activity (Fig. 2D and Supplementary Fig. 2).

Mitochondrial functional failure, involving mitochondrial membrane potential changes is considered to be one of the most important factors leading to cell death [29, 30]. Nonirradiated HaCaT cells were stained with JC-1 to show red fluorescence, while a large number of cells switched to green fluorescence after irradiation. These results indicate a decrease in mitochondrial membrane potential. HaCaT cells treated with fenofibrate showed less fluorescence from red to green, suggesting that fenofibrate can maintain mitochondrial membrane potential after ionizing radiation (Fig. 2E). These results demonstrated that fenofibrate protects mitochondria from ionizing radiation.

We next explored whether fenofibrate was associated with decreased apoptosis in skin cells. As shown in Fig. 2E,F, fenofibrate did not affect apoptosis in HaCaT cells that were not exposed to irradiation. In comparison, treatment with both 25 and 50 $\mu$M fenofibrate significantly decreased apoptosis in HaCaT cells that were exposed to 20 Gy irradiation (Fig. 2G). These results demonstrated that fenofibrate-mediated PPAR$\alpha$ activation reduces apoptotic cell death caused by irradiation in skin cells.

Because epidermal lipids and free fatty acids play important roles in cell growth, differentiation and permeability barrier function [31, 32], we investigated whether fenofibrate-mediated PPAR$\alpha$ activation modulated lipid accumulation in human keratinocytes. The results revealed that PPAR$\alpha$ activation by fenofibrate increased cytoplasmic lipid accumulation in HaCaT cells (Fig. 2H).

Next, we sought to investigate whether fenofibrate could mitigate the progression of radiation-induced skin injury in animal models. A radiation-induced rat skin injury model (45 Gy electron beam irradiation) [25, 26] was used to evaluate the role of clinically approved fenofibrate in oxidative damage. After exposure to 45 Gy of irradiation, rat skin was injected subcutaneously with DMSO or fenofibrate. To test whether fenofibrate affects radiation-induced lipid peroxidation, we measured ROS and MDA concentrations in skin tissues three days after 45 Gy of irradiation. As shown in Fig. 3A,B, both cellular ROS and MDA levels were significantly lower in fenofibrate-injected tissues than in DMSO-injected tissues. This result indicated that fenofibrate attenuated radiation-induced ROS and consequent lipid peroxidation.

Fig. 3.

Fenofibrate ameliorates radiation-induced skin injury in a rat model. Rat gluteal skin was unexposed or irradiated with a single dose from a 45-Gy electron beam. The rats were then randomly assigned to receive one of the following treatments (n = 4): (1) subcutaneous 110 $\mu$L DMSO injection (in 890 $\mu$L PBS); (2) subcutaneous 400 $\mu$g fenofibrate injection (in 110 $\mu$L DMSO and 890 $\mu$L PBS). (A) Relative ROS levels in the rat skin. Three days after irradiation, skin ROS levels were determined as described in the Materials and Methods section. (B) MDA concentration levels in rat skin from different groups at three days after irradiation. p 0.05, compared with the control group.

Our results showed that fenofibrate-mediated PPAR$\alpha$ activation promoted lipid accumulation in skin cells. This finding indicates a potential relationship between skin cells and lipid metabolism. Free fatty acids, which are relatively insoluble and potentially toxic, can be transported to other cells by noncatalytic binding proteins [33]. FABPs belong to a family of intracellular proteins and exhibit a high affinity for non-covalent binding to long-chain fatty acids [34]. We, therefore, hypothesized that FABPs may be involved in the radioprotective role of fenofibrate. Among the 12 identified members of the human FABP family, three putative binding sites for PPAR$\alpha$ (PPRE) in the proximal promoter of FABP4 (Fig. 4A) were predicted by bioinformatics analysis. This result suggested transcriptional regulation by PPAR$\alpha$. FABP4 is an intracellular lipid-binding protein responsible for fatty acid transportation [35] and we have recently shown that FABP4-mediated the radioprotection of adipocytes [26]. We next performed EMSAs to investigate the binding of potential transcriptional factors. EMSAs revealed that oligonucleotides representing the predicted PPAR$\alpha$ binding sites all formed a specific complex when incubated with WS1 nuclear extracts (Fig. 4B). Western blotting analyses showed that fenofibrate increased FABP4 protein levels in both HaCaT and WS1 cells (Fig. 4C). Using real-time PCR analyses, we found that fenofibrate increased FABP4 mRNA levels in a dose-dependent manner (Fig. 4D). A FABP4 mRNA-based fluorescent probe, but not a SmartFlare uptake control probe, confirmed that fenofibrate upregulated FABP4 transcripts specifically in WS1 and HaCaT cells (Fig. 4E and Supplementary Fig. 3). Taken together, these results clearly indicated that FABP4 is positively regulated by the PPAR$\alpha$ agonist fenofibrate in skin cells.

Fig. 4.

Fenofibrate activates FABP4 expression in skin cells. (A) Bioinformatics analysis predicted three putative binding sites in the proximal promoter of FABP4. (B) EMSA using nuclear proteins from WS1 cells and oligonucleotides carrying the indicated probes. Lanes 1, 3 and 5 contain the probes without nuclear extracts. Lanes 2, 4 and 6 contain the oligonucleotide probes 1, 2 and 3, respectively. (C) HaCaT and WS1 cells were treated with 10–50 $\mu$M fenofibrate. FABP4 expression was measured by Western blotting analyses. (D) WS1 cells were treated with 10–50 $\mu$M fenofibrate, and FABP4 mRNA was quantified by real-time PCR. (E) Quantitative analysis of Western blotting assay. (F) Quantification of FABP4-specific SmartFlare probe fluorophore microscope (G) HaCaT and WS1 cells were treated with 50 $\mu$M fenofibrate for 24 h, and FABP4 mRNA was detected with a FABP4-specific SmartFlare probe (Millipore, Billerica, MA, USA). Fluorescent signals reflecting the FABP4 mRNA levels were observed using a confocal microscope. p 0.05 and ** p 0.01, compared with the control group.

Next, we sought to investigate whether increased FABP4 expression could modulate radiation-induced damage in skin cells. Skin cells were pre-infected with a control adenovirus (Ad-NC) or FABP4 overexpression adenovirus (Ad-FABP4) and subjected to X-ray irradiation (Fig. 5A,B). The results showed that FABP4 overexpression reduced radiation-induced ROS levels (Fig. 5C). Moreover, FABP4 overexpression increased cellular lipid accumulation in HaCaT cells, which mimics the effect of fenofibrate (Fig. 5D). Immunofluorescence assays for $\gamma$H2AX foci showed that fewer foci were present in irradiated WS1 cells with FABP4 overexpression than in control cells (Fig. 5E). These data suggested that FABP4 facilitated the repair of radiation-induced DNA damage.

Fig. 5.

FABP4 confers radioprotection to skin cells. WS1 cells were infected with the indicated adenoviruses. (A) FABP4 expression was measured by Western blotting analyses. (B) FABP4 expression was measured by Western blotting. (C) WS1 cells were infected with the indicated adenovirus, followed by 0 or 20 Gy irradiation. Cellular ROS levels for each group of cells were determined 1 h after radiation using a DCF-DA probe and quantified with a microplate reader. (D) Quantification of BODIPY fluorophore 493/503 staining for lipid droplets. (E) Quantification of Western blotting assay. (F) Quantification of nuclear $\gamma$H2AX foci fluorescent signals. (G) The effect of FABP4 overexpression on lipid accumulation in HaCaT cells. Representative photomicrographs of BODIPY fluorophore 493/503 staining for lipid droplets. (H) WS1 cells were infected with Ad-NC or Ad-FABP4, and the dynamic repair process of DNA double-strand breaks (DSBs) was measured by detecting nuclear $\gamma$H2AX foci after X-ray irradiation. p 0.05 and ** p 0.01, compared with the control group.

To investigate whether FABP4 mediated the radioprotective role of fenofibrate, FABP4 inhibitor BMS309403 [36] was used. The results showed that the addition of BMS309403 exacerbated radiation-induced ROS in human skin fibroblasts. Moreover, the ROS-eliminating activity of fenofibrate was abrogated by BMS309403 (Fig. 6A). These results indicated that FABP4 was involved in antioxidant response and that FABP4 mediated the ROS eliminating the role of fenofibrate. Moreover, combined treatment with fenofibrate and BMS309403 abrogated lipid accumulation activity of fenofibrate in keratinocytes, which suggested that FABP4 mediated fenofibrate-induced lipid accumulation (Fig. 6B). Taken together, these above results indicated that FABP4 was likely to mediate the radioprotective role of fenofibrate.

Fig. 6.

FABP4 mediates the radioprotective role of fenofibrate. (A) WS1 cells were infected with the indicated adenovirus or treated with fenofibrate and/or BMS309403, followed by 0 or 20 Gy irradiation. Cellular ROS levels for each group of cells were determined 1 h after radiation using a DCF-DA probe and quantified with a microplate reader. (B) The FABP4 inhibitor BMS309403 abrogated fenofibrate-induced lipid accumulation in HaCaT cells. Representative photomicrographs of BODIPY fluorophore 493/503 staining for lipid droplets. (C) Quantification of BODIPY fluorophore 493/503 staining for lipid droplets. p 0.05, compared with the control group.

Next, we investigated whether FABP4 overexpression could reduce radiation-induced skin damage in vivo. The buttock region of rats was irradiated with a 45 Gy electron beam to model the irradiation-induced skin injury in rats. Irradiation at 45 Gy significantly increased skin ROS levels at three days after treatment, as shown in Fig. 7A, ROS levels were significantly lower in tissues infected with Ad-FABP4 than in the control tissues. Moreover, FABP4 overexpression also reduced radiation-induced MDA levels (Fig. 7B). These results indicate that FABP4 overexpression attenuates lipid peroxidation resulting from radiation-induced oxidation.

Fig. 7.

FABP4 attenuates radiation-induced skin injury in a rat model. Rat gluteal skin was unexposed or irradiated with a single dose from a 45-Gy electron beam, followed by subcutaneous injection of Ad-NC (5 $\times$ 10${}^9$ genomic copies of Ad-NC in a 200-$\mu$L volume) or Ad-FABP4 (5 $\times$ 10${}^9$ genomic copies of Ad-FABP4 in a 200-$\mu$L volume) (n = 4). (A) Relative ROS levels in rat skin. Three days after irradiation, skin ROS levels were determined as described in the Materials and Methods section. (B) MDA concentration levels in rat skin from different groups at three days after irradiation. p 0.05, compared with the control group.