Erythropoietin alleviates lung ischemia-reperfusion injury by activating the FGF23/FGFR4/ERK signaling pathway

Clinical sample

A total of 40 blood samples from patients with LIRI at our hospital were collected in the present study. Inclusion criteria were as follows: (1) all patients were diagnosed with LIRI for the first time; (2) there were no significant abnormalities in liver and renal function in all patients; and (3) the clinical data of each patient was complete. Exclusion criteria were as follows: (1) patients with a history of chemotherapy or radiation; (2) patients who cannot cooperate; and (3) patients with infectious diseases or malignant tumors. Also, samples were taken from 40 healthy volunteers as control. Table 1 describes the distribution of age and sex between two groups. Each participant and their families were informed and agreed to participate in this study, and they signed a consent form. The human study was approved by the ethics committee of Tongde Hospital of Zhejiang Province (2022-072 (K)) and conducted according to the Declaration of Helsinki.

Animal model and pharmacological treatment

Sixty six male SD rats (8–10 weeks, weight 200–250 g) were purchased from Beijing SiPeiFu biotechnology Co., LTD. The rats received humane care in accordance with the Institutional Animal Care and Use Committee. All rats were kept in cages with controlled temperature (25 °C, approx.) and a 12-hour light/dark cycle, with food and water ad libitum. Before the operation, the rats were fasted for 12 h and water was forbidden for 4 h. The rats were randomly divided into two groups, with nine rats in each group: control group and the left LIRI model group. The specific methods are as follows: (1) SD rats of the experimental group were anesthetized by intraperitoneal injection of 100 mg/kg ketamine +10 mg/kg toluene thiazide; (2) after anesthesia, the rats were placed on a table and placed in a supine position, and the limbs of the rats were fixed on the rat operating table; (3) an anterior median incision was made on the neck to expose and separate the right trachea and carotid artery. The trachea was carefully incised for intubation, and mechanical ventilation was performed by the external small animal ventilator. (4) The chest cavity and the left pulmonary hilus were exposed along the left sternum, and the blocking band then passed under the left sternum. After resting for 5 min, the left pulmonary hilus was blocked with a blocking line at the end of the expiratory breath. Reperfusion was performed 1 h later.

Different treatments were performed in these rats (n = 8 in each group). EPO dosage (3,000 U/kg) was established with reference to a previous study (He et al., 2018). To explore the dose–response relationship, a low-dose group (1,000 U/kg) has been added. The rats were grouped as follows: control group, EPO single administration group (3,000 U/kg), LIRI model group, LIRI model + EPO prevention group (3,000 U/kg), LIRI model + EPO low dose treatment group (1,000 U/kg), LIRI model + EPO high dose treatment group (3,000 U/kg). The EPO pretreatment group was given caudal vein administration 15 min in advance, the treatment group was given caudal vein administration at the same time of reperfusion, and both the prevention group and the treatment group were given caudal vein administration every day after modeling. Three days post-reperfusion, the rats were euthanized by CO₂ inhalation, and blood, alveolar lavage fluid and lung tissue were collected for follow-up studies. The grouping for in vivo experiments is illustrated in Fig. S1. The animal experiment was approved by the Committee of Experimental Animals of Zhejiang Academy of Traditional Chinese Medicine (KTSC2021416).

Cell culture and treatment

BEAS-2B cells were purchased from the ATCC (Manassas, VA, USA) and cultured in RPMI 1640 medium (Invitrogen, Carlsbad, CA, USA) supplemented with 5% heat-inactivated fetal bovine serum (Gibco, Waltham, MA, USA), two mmol/l l-glutamine (Gibco, Waltham, MA, USA), 1% penicillin-streptomycin (Gibco, Waltham, MA, USA) at 37 ^∘C and 5% CO₂ in humidified air. Small interfering RNA (siRNA) oligonucleotides targeting FGF23 (siFGF23), siFGFR4, FGF23 overexpression vector (OE-FGF23), OE-FGFR4 and control vector were bought from GenePharma (Shanghai, China). Transfection was performed by using Lipofectamine 2000 Reagent (Invitrogen, Carlsbad, CA, USA) according to the manual. Tert-butyl hydroperoxide (TBHP)-treated BEAS-2B cells were selected as a biological model to establish the I/R cell model.

Histological analysis

Morphological changes in the lobar tissue of the left lung were examined using hematoxylin and eosin (HE). Lung tissues collected from the rats with different treatments were fixed in 4% paraformaldehyde overnight at 4 °C and embedded in paraffin for 50–60 min at room temperature. Subsequently, the samples were cut into 3–5 µm slices and stained with HE and the tissue morphology was analyzed. The histological changes were examined blindly via a light microscope (BA210T; MOTIC, Xiamen, China) by two independent researchers.

Immunohistochemistry

The expression level of FGF23 in lung tissue was detected by immunohisochemistry (IHC). In short, 3% hydrogen peroxide was employed to block lung tissues at room temperature for 1 h. Antigen retrieval was performed by microwave irradiation in citrate retrieval solution (PH 6.0). Secondarily, each slice was incubated overnight at 4 °C with rabbit monoclonal anti-FGF23 (ab192497, 1: 1000; Abcam, Shanghai, China). The sections were washed and subsequently incubated with a secondary antibody (ab205718, 1: 2000; Abcam, Shanghai, China) at 37 °C for 30 min. Then the sections were stained with DAB until the stain developed, and nuclei counterstaining was performed with hematoxylin. Finally, the slices were observed under a microscope.

Enzyme-linked immunosorbent assay

Enzyme-linked immunosorbent assay (ELISA) was employed to determine the release of FGF23 in both the serum and BALF. The collected blood samples from the patients were centrifuged at 1,000 rpm at 4 °C for 10 min to obtain the serum supernatant. The supernatant of the tissue homogenate was collected for further analysis. ELISA kits were used to assess the concentration of EPO (PE230, Beyotime, Suzhou, China), FGF23 (PF300, Beyotime, Suzhou, China) and FGFR4 (SEKH-0178, Solarbio, Beijing, China) in each sample as per the manufacturer’s instructions. The optical absorbance at 450 nm was recorded using a microplate reader (Thermo Labsystems, Philadelphia, PA, USA).

Western blot and co-IP

RIPA lysis buffer (P0013B, Beyotime, Shanghai, China) plus phosphatase inhibitors (35657, Sigma, Burlington, MA, USA) was employed to isolate the total protein of each group, following the manufacturer’s instructions. Subsequently, proteins in cell lysates were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred onto a polyvinylidene difluoride membrane (Merck-Millipore, Darmstadt, Germany), in addition to incubating with a primary antibody. Afterward, the membrane was incubated with a secondary antibody (Cell Signaling Technology, Danvers, MA, USA) at room temperature for 2 h. Finally, the blots were identified using enhanced chemiluminescence (Pierce, Rockford, IL, USA) in agreement with the manufacturer’s instructions. Primary antibodies were as follows: FGF23 (ab307420, 1:3000), FGFR4 (ab178396, 1:3000), p-ERK1/2 (ab184699, 1:1000). GAPDH (ab8245, 1:5000) was used as loading control. For co-IP, lysates of 1 × 10⁷ BEAS-2B cells were immunoprecipitated with IP buffer containing IP antibody-coupled agarose beads, and protein-protein complexes were later subjected to Western blot.

qPCR analysis

qPCR was employed to assess the expression of genes BAX, BCL-2 and caspase-3. Total RNA was extracted from tissue specimens and treated cells using the TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA). The RNA concentration was determined by Nano Drop spectrophotometer (DS-11FX, DeNovix, Wilmington, DE, USA). Then, total RNA was further reverse transcribed into complementary DNA through the PrimeScript RT reagent kit (TaKaRa, Shiga, Japan). qPCR was performed using TB Green Premix Ex Taq (TaKaRa, Wilmington, DE, USA) and LightCycler 96 real-time PCR instrument (Roche, Basel, Switzerland). The amplification cycling reactions (45 cycles) were as follows: 7 s at 95 °C, 10 s at 57 °C and 15 s at 72 °C. Relative quantification values of the target genes were standardized according to the comparative threshold cycle (2^−ΔΔCT). GAPDH was applied as a HOUSEKEEPING GENE. Primer sequences are displayed in Table 2.

TUNEL assay

TUNEL assay (C1086, Beyotime, Shanghai, China) was implemented to rate the effect of EPO on TBHP-induced BEAS-2B cell apoptosis under the manufacturer’s protocol. In brief, 4% paraformaldehyde was applied to fix the treated BEAS-2B cells for 30 min at room temperature and then cultured with 0.3% Triton X-100 for 5 min. Secondarily, 50 µl TUNEL reaction reagent was added to each sample at 37 °C for 1 h. After staining with 0.5 µg/ml of DAPI for 10 min at room temperature. Finally, TUNEL-positive cells were observed with the help of a fluorescence microscope (Beyotime, Shanghai, China).

Flow cytometry apoptosis analysis

Annexin V-FITC/PI apoptosis detection kit (556547, BD Biosciences, San Jose, CA, USA) was implemented to determine cell apoptosis, as instructed by the manufacturer. In short, resuspended cells were stained with 10 µL Annexin V-FITC and PI in the dark for 15 min. Subsequently, the apoptotic cells were analyzed by flow cytometry (CytoFLEX, Beckman Coulter, Brea, CA, USA).

Statistical analysis

All the above assays were carried out independently in triplicates. The data were analyzed with the help of GraphPad Prism 7.04 software and displayed as mean ± SD. Statistical differences were performed via One-way ANOVA with Tukey’s post hoc t-test for multiple comparisons. P < 0.05 was considered significant. *P < 0.05, **P < 0.01; ***P < 0.001.

Effect of I/R injury on FGF23 levels and apoptosis

Fibroblast growth factors (FGFs) are critical biological regulators which are associated with pathophysiological processes, such as LIRI. Likewise, it has been evinced that exogenous FGF ligands could protect against I/R, which offers potential therapeutic options for patients with LIRI (Deng et al., 2020). In order to investigate the role of FGF23 in LIRI, the LIRI rat model was established. As shown in Fig. 1A, alveolar and interstitial inflammation and pulmonary fibrosis were observed in the lung tissues of rats in the I/R group (model), confirming the successful construction of the LIRI rat model. Next, IHC and ELISA were performed to analyze the level of FGF23 in the lung tissue, the serum, and the BALF sample in vivo. It was observed that the FGF23 levels were decreased in lung tissue in the I/R group, in comparison with the control group (Fig. 1B). Also, the data of the ELISA validated similar results in the serum and BALF samples (Figs. 1C & 1D). In addition, Western blot demonstrated that the expression of FGF23, FGFR and p-ERK1/2 was down-regulated in lung tissue in the experimental group compared to the control group (Fig. 1E). Of note, based on the evidence that apoptosis is one of the important phenotypes in the process of ischemia-reperfusion injury (Bi, Li & Zhang, 2023), apoptosis-related gene expression and cell apoptosis was detected by qPCR and TUNEL assay. It was indicated that the expression level of Bcl-2 was decreased in the experimental group than in the control group, however, the expression level of BAX and caspase-3 was increased in the group (Fig. 1F). Equally importantly, TUNEL-positive cells were higher in rats with IR injury compared to those of the control group (Fig. 1G).

EPO attenuates I/R-induced lung injury in vivo

It has been demonstrated that EPO exerts a protective effect on the multiple I/R model (Kartal et al., 2023; Kittur et al., 2023). However, additional studies and clinical trials are required to assess the role of EPO in LIRI. The results of HE revealed that I/R treatment significantly induced lung injury in comparison with the sham group, which could be partially weakened by pre- or post-treatment of EPO (Fig. 2A). Interestingly, the results of IHC and ELISA showed that FGF23 expression level in lung tissue (Fig. 2B) and content in serum (Fig. 2C), in addition to FGF23 BALF (Fig. 2D) was significantly increased in the EPO administration group compared with the model group. Further, this increased trend appears in a dose-dependent manner (Figs. 2B–2D). In addition, in comparison with the model group, the number of TUNEL-positive cells was decreased in the EPO treatment group (Fig. 2E). Likewise, the qPCR analysis showed that pro-apoptosis related genes BAX and caspase-3 were significantly downregulated along with a decrease in the levels of anti-apoptosis related protein BCL-2 after EPO administartion. (Fig. 2F). In addition, I/R treatment significantly suppressed the expression of FGF23, FGFR4 and p-ERK1/2, which could be abrogated by EPO pretreatment or posttreatment (Fig. 2G). Additionally, in comparison with the control group, the EPO group did not cause damage to lung tissue and inhibited apoptosis, indicating EPO had no side effect on the rat lung (Figs. 2A–2G). Taken collectively, EPO attenuated I/R-induced lung injury via regulating the FGF23/FGFR4/p-ERK1/2 pathway.

EPO attenuates the injury of TBHP-induced BEAS-2B cells by activating FGF23/FGFR4/p-ERK1/2 pathway

Since EPO has a pronounced promotive activity against injury induced by TBHP in rat lungs, BEAS-2B cells were cultured with different concentrations of TBHP to mimic an I/R injury. Cell apoptosis in the 100 µM TBHP group, 200 µM TBHP group and 400 µM TBHP group was significantly increased compared with the control group (Figs. 3A & 3B). In addition, the results of ELISA demonstrated that TBHP diminished the FGF23 content in cell supernatant in a dose-dependent manner compared with the control group (Fig. 3C). Also, the mRNA expressions of caspase-3, BAX, and Bcl-2 were detected by qPCR, as described (Fig. 3D). The results indicate that TBHP treatment effectively reduced the expression of BCL-2, and increased the expression of caspase-3 and BAX. The expression levels of FGF23, FGFR4 and p-ERK1/2 were also examined by Western blotting (Fig. 3E). It was observed that TBHP remarkably decreased the expression of FGF23, FGFR4 and p-ERK1/2 in BEAS-2B cells (Fig. 3E). In addition, the relationship between FGF23 and FGFR4 was verified by CO-IP assay, which revealed that both FGF23 and FGFR4 proteins were identified by western blot in the immune complexes obtained with anti-FGF23 or anti-FGFR4 antibodies, indicating an interaction between FGF23 and FGFR4 could be disturbed by TBHP treatment. Because the cell model constructed at 400 µM TBHP worked best, it was used in a series of subsequent experiments. In order to verify the protective effect of EPO on TBHP-induced apoptosis of lung epithelial cells, TBHP (400 μM) was added to cells with EPO (0, 5, 10 U/mL) 30 min later, and cells were collected at different time points for FITC-Annexin V/PI. As displayed in Figs. 4A and 4B, EPO blocked the development of cell apoptosis in comparison to the control group. Similar to the results of the flow cytometry, EPO significantly reduced the expression of BAX and caspase-3 as well as increased the expression of BCL-2 mRNA in BEAS-2B cells treated with TBHP (Fig. 4C). Moreover, EPO could not only enhance the expression of FGF23 and FGFR4 but also strengthen the interaction between them (Figs. 4E & 4F). In addition, EPO could induce the expression of FGF23 in BEAS-2B cells treated with TBHP (Fig. 4E).

FGF23 or FGFR4 inhibited TBHP-induced apoptosis

Afterward, BEAS-2B cells with FGF23 were either knocked down or overexpressed to investigate the function of FGF23 in cell apoptosis and the transfection outcome was assessed by Western blot (Figs. S2A & Fig. S2B). The FGF23 expression in siFGF23 (BEAS-2B cells transfected with siFGF23) group was strikingly down-regulated compared to the siNC group (BEAS-2B cells transfected with control vector). Flow cytometry analysis showed that transfection with siFGF23 significantly triggered apoptosis compared with transfection with siNC (Figs. 5A& 5B). Furthermore, the results of ELISA demonstrated that knockdown of FGF23 restricted the FGF23 content in cell supernatant (Fig. 5C). Also, the data of qPCR depicted that FGF23 silencing effectively reduced the expression of Bcl-2, and increased the expression of caspase-3 and BAX (Fig. 5D). As can be easily understood, compared with OE-NC-TBHP group, overexpression of FGF23 significantly inhibited apoptosis and up-regulated the content of FGF23 in cell supernatant (Figs. 5A–5D). Interestingly, knocking down FGFR4 or overexpressing FGFR4 has a similar effect on apoptosis as knocking down FGF23 or overexpressing FGF23 (Figs. S2A & S2B, Figs. 5A–5D). It is worth noting that FGF23 positively regulates the expression of FGFR4 in BEAS-2B cells or cell supernatant; however, regulating the expression of FGFR4 has no significant effect on the expression of FGF23 (Fig. S2).

EPO exerts an anti-damage effect depending on the regulation of the FGF23 signaling pathway

To validate the interaction between EPO and FGF23 signaling pathway in TBHP-induced lung cells, siFGF23, siFGFR4, FGF23 overexpression plasmid (OE-FGF23), OE-FGFR4 and their control vector were transfected into BEAS-2B cells, respectively. Transfection efficiency was then examined by Western blot (Fig. S3A). Knocking down the expression of FGF23 or FGFR4 promoted apoptosis, enhanced the expression of BAX and caspase-3, and inhibited the expression of Bcl-2 in TBHP-induced BEAS cells in the background of EPO compared with NC group (Figs. 6A–6C). Conversely, overexpression of FGF23 or FGFR4 produced protective effects similar to EPO in TBHP-induced models (Figs. 6A–6C). Furthermore, Western blot results demonstrated that FGF23 or FGFR4 positively regulated the expression of p-ERK1/2 (Fig. S3A). However, up-regulating or down-regulating FGFR4 expression did not significantly affect FGF23 expression, either in BEAS-2B cells or in supernatant (Fig. S3B).

The key factors of expression level in patients with LIRI

Furthermore, samples from 40 patients with LIRI and 40 healthy controls were collected in the present study to substantiate the role of the above key factors in LIRI. Initially, we compared the difference in EPO levels between the control group and the LIRI group. As shown in Fig. 6A, the results of ELISA revealed that the EPO content in serum from the patients with LIRI was lower than the patients from the control group. Also, the FGF23 and FGFR4 content were decreased compared with the control group (Figs. 7B & 7C). As can be easily understood, the results of Western blot showed that the expression levels of EPO, FGF23 and FGFR4 in LIRI serum were significantly lower than those in healthy controls (Fig. 7D). In addition, the expression levels of caspase-3 and BAX were elevated in comparison with the control; however, the BCL-2 expression level was lower than in the healthy sample (Fig. 7E).