The SH-SY5Y cell line, which was produced by the American Type Culture Collection (ATCC, Manassas, VA, United States), was cultured in Dulbecco’s modified Eagle’s high glucose medium (DMEM, 11965, Life Technologies, Rockville, MD, United States), supplemented with 10% fetal bovine serum (10099, Invitrogen, CA, United States) in a humidified incubator with 5% CO₂ at 37°C. We prepared rotenone (R8875, Sigma-Aldrich, St Louis, MO, United States) in dimethyl sulfoxide (DMSO; D2650, Sigma-Aldrich) at a stock concentration of 50 mM. It was then diluted with DMEM to a concentration of 1 μM and incubated with cells for different amounts of time (0, 12, 24, 36, and 48 h) according to the corresponding experiments. The vehicle was used as the control. The PI3K/Akt specific inhibitor LY294002 (A8250, APExBIO, Houston, TX, United States) was diluted with DMEM to a concentration of 0–10 μM and incubated with cells for 3 h to choose the appropriate stimulus concentration.

GDF15 was overexpressed by transfecting cells with pcDNA3.1-GDF15. cDNA encoding full-length GDF15 was cloned into the pcDNA vector, and the sequence was confirmed using Sanger sequencing. Empty plasmids were used as the mock control. All plasmids were verified using DNA sequencing. Lipofectamine 2000 (11668, Invitrogen, Waltham, MA, United States) was used for cell transfections following the manufacturer’s instructions.

The cell culture supernatant was collected after 0–48 h of treatment with rotenone (1 μM) and stored at –80°C until analysis. GDF15 concentrations were measured using a high-sensitivity human enzyme-linked immunosorbent assay (ELISA) kit (ab155432, Abcam, Cambridge, MA, United States) as per the manufacturer’s instruction.

Levels of protein expression were measured using western blotting analysis, which was performed using standard methods with commercially available antibodies. After treating cells with vehicles or 1 μM rotenone for 24 h, the whole-cell lysate was prepared using lysis buffer, protease inhibitors, and phenylmethylsulphonyl fluoride. The protein concentration was measured using a Bicinchoninic Acid protein assay kit (23227, Thermo Scientific, Waltham, MA, United States). Soluble proteins (30 μg) were separated using 8–12% SDS-polyacrylamide gels and then transferred onto polyvinylidene difluoride membranes (200 mA, 60 min), which were then blocked in 5% non-fat milk or bovine serum albumin in tris-buffered saline with tween 20 and incubated with primary antibodies overnight. The primary antibodies used were as follows: anti-GDF15 (1:5000, ab39999, Abcam), anti-Bcl-2 (1:2000, ab182858, Abcam), anti-Bax (1:1000, ab182734, Abcam), anti-p53 (1:1000, ab32049, Abcam), anti-PGC-1α (1:1000, ab106814, Abcam), anti- tyrosine hydrolase (TH) (1:200, ab112, Abcam), anti-mTOR (1:1000, 2983S, Cell Signaling Technology, Danvers, MA, United States), anti-phospho-mTOR (1:1000, 5536S, Cell Signaling Technology), anti-Akt (1:1000, 4685S, Cell Signaling Technology), anti-phospho-Akt (1:1000, 4060S, Cell Signaling Technology), and anti-β-actin (1:5000, BE3212-100, EASYBIO, Seoul, South Korea), used as an internal control. Then, secondary HRP-linked antibodies were used to detect primary antibodies. The immunoreactive signal was detected using an ECL substrate kit (Millipore, Billerica, MA, United States). Gray values of the protein bands were analyzed using ImageJ 1.8.0r software (National Institutes of Health, Bethesda, MD, United States).

RNA was isolated from cells following the manufacturer’s protocol using TRIzol (Invitrogen, 15596-026), and reverse transcription was used to convert 1 μg RNA to cDNA using a cDNA Synthesis kit (Vazyme, R323-01). Quantitative real-time polymerase chain reaction (qRT-PCR) analysis was performed using the Super Real PreMix Plus SYBR Green (Vazyme, Q711-02). The relative mRNA expression levels were normalized to those of GAPDH and calculated using the 2^–ΔΔCT method. The primer sequences for qRT-PCR were as follows: PPAR-γ coactivator-1-alpha (PGC1-α): forward: 5′-TCCTCACACCAAACCCACAG-3′ and reverse: 5′-TGGGGTCATTTGGTGACTCTG-3′, α-synuclein: forward: 5′-AATGAAGAAGGAGCCCCACAG-3′ and reverse: 5′-AGGCTTCAGGTTCGTAGTCTT-3′, bcl-2: forward: 5′-GA ACTGGGGGAGGATTGTGG-3′ and reverse: 5′-GCCGGTwTC AGGTACTCAGTC-3′, bax: forward: 5′-TGAGCGAGTGT CTCAAGCG-3′ and reverse: 5′- CTGCCACTCGGAAAAAG ACC-3′, p53: forward: 5′-GGCCCACTTCACCGTACTAA-3′ and reverse: 5′-GTGGTTTCAAGGCCAGATGT-3′, GAPDH: forward:5′-GCACCGTCAAGGCTGAGAAC-3′ and reverse: 5′-TGGTGAAGACGCCAGTGGA-3′.

Cells (5 × 10⁴ cells/well) were seeded in a 96-well plate and grown overnight. Next, cells were treated with vehicle, rotenone (1 μM), and 0–100 ng/ml recombinant human GDF-15 (57-GD-025, R&D Systems, United States). Cell viability was detected using a Cell Counting Kit (CCK) 8 assay (A311-02, Vazyme, Nanjing, China). Briefly, after 24 h, 10 μL of CCK8 solution was added to each well and the incubation continued for another 4 h at 37°C. Absorbance was measured using a multimode microplate reader (Thermo Fisher Scientific Inc., MA, United States) at 450 nm. Cell viability was expressed as a percentage relative to the absorbance of control cells.

Cell apoptosis induced by rotenone treatment was examined using an Annexin V-PE/7-ADD apoptosis kit (Vazyme, A213) following the manufacturer’s protocol. Briefly, we collected the cells and washed them twice in phosphate-buffered saline. After that, cells were labeled with 500 μL 1 × binding buffer containing 10 μL Annexin V-PE and 10 μL 7-ADD in the dark for 10 min. Subsequently, we used a flow cytometer (BD Falcon; BD Biosciences, Franklin Lakes, NJ, United States) to detect cell apoptosis, and the Annexin V-PE and 7-ADD values were set as the horizontal and vertical axes, respectively, for plot construction. Mechanically damaged, late apoptotic, dual negative/normal, and early apoptotic cells were located in the upper left, upper right, lower left and lower right quadrants of the flow cytometric dot plot, respectively.

The mitochondrial functions of cells were measured using a Seahorse XF cell Mito Stress test kit (103015-100, Agilent Technologies, Santa Clara, CA, United States) and a Seahorse XF24 Extracellular Flux Analyzer. To measure the oxygen consumption rate, mitochondrial complex inhibitors (oligomycin [1 μM], FCCP [1 μM], and rotenone/antimycin A [0.5 μM]) were successively added to the cell culture microplate to measure key parameters of mitochondrial function using the Seahorse XF24 Analyzer. Each sample was assayed based on a minimum of three replicates, and the data were normalized to the protein content in each well.

For apoptosis detection, paraffin-embedded sections were stained using the Terminal dUTP nick-end labeling (TUNEL) technique and an in situ apoptosis detection kit (KGA7061, KeyGEN BioTECH, Jiangsu, China) following the manufacturer’s protocols. Briefly, cells were incubated with TUNEL reaction mixture and DAPI for 30 min in a dark room at 37°C after being fixed, blocked, and permeabilized. The images were scanned and analyzed with a fluorescence microscope (Olympus IX73, Japan). The apoptotic ratio was calculated as the ratio of TUNEL-positive cells to total cells.

To evaluate the mitochondrial membrane potential (MMP), we used a JC-10 assay kit (ab112134, Abcam). Following the commercial protocols, cells were cultured in a 96-well black plate with a transparent bottom (Costar, Corning, NY, United States). After the experimental procedures, the cell mitochondria were stained with JC-10 under standard conditions for 30 min and kept away from light. Later, the buffer was added to each well, and the plates were read using a fluorescence microplate reader (Spark 10M, TECAN, Zurich, Switzerland). The green fluorescence was read at 490 nm (excitation) and 520 nm (emission). The orange fluorescence was read at 515 nm (excitation) and 590 nm (emission). The data were expressed as the orange/green fluorescence ratio.

Treated cells were incubated with DCFH-DA (CA1410, Solarbio, United States) at a concentration of 10 μmol/l for 30 min at 37°C to detect intracellular reactive oxygen species (ROS) levels. Fluorescence was recorded at 488 nm (excitation) and 525 nm (emission) using a flow cytometer (BD Falcon; BD Biosciences, Franklin Lakes, NJ, United States). The mitochondrial ROS levels in different groups of SH-SY5Y cells were determined by performing Mito-Sox staining. Briefly, SH-SY5Y cells were cultured in a fluorescent chamber and subjected to different treatments. Next, SH-SY5Y cells were incubated with 5 μM Mito-Sox (M36008, Invitrogen) for 15 min at 37°C in the dark. The morphology of SH-SY5Y cells was detected with MitoTracker Green FM (Invitrogen, Carlsbad, CA, United States). SH-SY5Y cells subjected to different treatments were incubated for 30 min with DMEM supplemented with 20 nM MitoTracker Green FM. After washing with PBS three times, cells were mounted with DAPI and photographed using a confocal microscope equipped with a camera (Leica DMI8, Germany). At last, the sample was randomly photographed, and the fluorescence intensity was analyzed from five different view fields in each group using ImageJ software based on three independent experiments.

Statistical analysis was performed using GraphPad Prism 5.01 software (GraphPad, Inc., La Jolla, CA, United States). All analyses were performed depending on the distribution (parametric or non-parametric) of the data. The Shapiro–Wilk’s test for normality was used for this purpose. The results are presented as the mean ± S.E. (standard error). If the data followed a normal distribution, they were analyzed using one-way analysis of variance, whereas two-group comparisons were performed using the Student’s t-test. P-values < 0.05 were considered statistically significant.

We used a human ELISA kit to measure the secretion levels of endogenous GDF15 protein in the cell culture supernatant following rotenone treatment, and the level of GDF15 released to the presented as a time-dependent increase, especially after 24 h (Figure 1A). In addition, we used western blotting to detect GDF15 and TH expression (Figure 1B). GDF15 concentrations increased significantly (P < 0.01) after treatment with rotenone in SH-SY5Y cells, suggesting a time-dependent change. The levels of TH, which is the enzyme that limits the level for dopamine biosynthesis, were downregulated in SH-SY5Y cells treated with rotenone.

To detect the protective effect of GDF15 on rotenone-treated SH-SY5Y cells, a cell viability test was performed using a CCK8 assay in cells treated with different rhGDF15 concentrations (0–100 ng/mL) after exposure to 1 μM rotenone or DMSO for 24 h. Compared with that in control cells (not treated with rotenone or rhGDF15), cell viability was evidently decreased after treatment with 1 μM of rotenone for 24 h, and a concentration-dependent elevation of cell viability was observed after rhGDF15 treatment (Figure 1C). This indicated that rotenone treatment resulted in an elevation of GDF15 expression and rhGDF15 had a protective effect on SH-SY5Y cells treated with rotenone depending on the dose and time.

To explore the downstream reactions following treatment with GDF15, the expression levels of several proteins involved in apoptosis were detected using western blotting, and the mRNA levels were quantified using qRT-PCR. Western blotting analyses GDF15 overexpression had no evident effect on the expression of TH and Bcl-2/Bax under normal conditions. It showed that 24-h treatment with 1 μM rotenone causes a marked and significant increase of p53 protein as compared to control group cells, and it causes a decrease of PGC-1α, Bcl-2/Bax, and TH expressions. However, the overexpression of GDF15 proved to attenuate or reduce the effect caused by rotenone treatment (Figures 1D–F). mRNA levels codifying for Bcl-2/Bax, PGC-1α, p53, and α-syn were quantified in SH-SY5Y cells by real-time PCR. The overexpression of GDF15 had no evident effect on the expression of α-syn under normal conditions. After treatment with rotenone, it could be found that the expression of α-syn and p53 have increased, while Bcl-2/Bax and PGC-1α was decreased, whereas GDF15 overexpression could efficiently alleviate these changes (Figures 1G,H). These results indicated that rotenone-induced toxicity might cause apoptosis by regulating Bcl-2/Bax and p53, whereas GDF15 could inhibit this harmful effect.

Apoptosis was quantified using flow cytometry after exposing cells to rotenone for 24 h. Flow cytometry showed an increased percentage of apoptotic cells after treatment with rotenone (P < 0.05), whereas cells overexpressing GDF15 showed lower levels of apoptosis in comparison with cells only treated with rotenone (P < 0.05; Figures 2A,B). After rotenone exposure, SH-SY5Y cells exhibited the typical apoptotic characteristics of Hoechst-stained nuclei, including obvious nuclear shrinking with increased density, condensation, and fracture. We further examined SH-SY5Y cell apoptosis by TUNEL assays (Figures 2C,D). The results of TUNEL revealed that rotenone caused the severe apoptosis of SH-SY5Y cells. Although the percent positivity with rotenone increased significantly in comparison with that in the control group (P < 0.05), GDF15 overexpression significantly reduced the level of rotenone-induced apoptosis (P < 0.05). These results indicated that GDF15 overexpression could suppress cell apoptosis caused by rotenone-induced toxicity.

DCFH-DA and Mito-Sox were used to detect the level of intracellular ROS and mitochondrial ROS, respectively, in different groups of SH-SY5Y cells. As illustrated in Figures 3A–D, the ROS level was significantly increased in cells not overexpressing GDF15 compared with that in normal control and GDF15-overexpressing cells. Collectively, our results reveal that GDF15 reduces SH-SY5Y cell apoptosis by decreasing ROS generation. MMP was assessed by JC-10 assays, and this was significantly decreased by rotenone treatment, whereas GDF15 could mitigate the toxic effect of rotenone (Figure 4A). These results showed that GDF15 overexpression could prevent rotenone-induced mitochondrial damage.

The mitochondrial function of SH-SY5Y cells was evaluated by measuring the oxygen consumption rate using a Seahorse XF24 Extracellular Flux Analyzer. The oxygen consumption rates of rotenone-treated SH-SY5Y cells were significantly lower than those of control and GDF15-overexpressing cells, as evidenced by the basal respiration, maximal respiration, and ATP production. And the level of basal respiration, maximal respiration, and ATP production was found an improvement in rotenone treated GDF15 overexpression group compared with rotenone group (Figures 4B,C). These data suggested that GDF15 impairs oxidative phosphorylation and mitochondrial function.

SH-SY5Y cells were treated with thee PI3K/Akt-specific inhibitor LY294002 (10 μM) 3 h before transfection with the GDF15 overexpression plasmid according to western blotting analysis (Figures 4D,E). As shown in Figures 4F–H, increased GDF15 expression after transfection with the GDF15 overexpression plasmid increased the phosphorylation of both Akt and mTOR. Moreover, inhibition of the PI3K/AKT pathway using LY294002 upregulated the expression levels of p53 and eliminated the protective effect of GDF15, which suggested that the effect of GDF15 expression was related to Akt/mTOR-dependent phosphorylation.