This article is mentioned in:

Introduction

Aging is a normal physiological process that not only increases susceptibility to diseases, but also gradually reduces physiological functions (1). Age-related hearing loss (AHL), also referred to as presbycusis, is a type of sensorineural deafness with bilateral high-frequency hearing loss that progressively declines with age (2). Aging-related changes in the central auditory system have a negative impact on auditory perception and verbal communication, or both (3). However, the mechanisms involved in central presbycusis remain unclear.

A number of aging pathways are closely related to cellular protein homeostasis, suggesting that maintaining a healthy protein homeostasis is crucial for longevity (4). Accumulating lines of evidence have indicated that the unfolded protein response (UPR) is the link between aging and proteostasis in a variety of vital organisms (4,5). Hearing loss caused by both environmental and genetic factors is improved by regulating the UPR (6). The continuous aggravation of the endoplasmic reticulum (ER) and mitochondrial (MT) stress participates in the pathogenesis of AHL. Heat shock protein (HSP)- or stress-related chaperones have been identified as markers of the UPR, and a number of studies have been conducted mainly on the UPR signaling pathways (7-9). However, the specific mechanisms and proteostatic UPR signaling pathways involved in age-related diseases remain unknown.

Metformin, an AMP-activated protein kinase (AMPK) activator, is a medication used in the first-line treatment of type 2 diabetes. Metformin readily crosses the blood-brain barrier and is distributed to various regions in the brain (10), and recently, it has been reported to promote neurogenesis, thereby providing potential neuroprotection (11). However, whether metformin interacts with the UPR and the exact role of metformin in the UPR remain unclear. It is well known that AMPK/ERK1/2 are growth factor signals and important metabolic regulators; however, a link between changes in the auditory system and the UPR has not yet been reported, at least to the best of our knowledge. Similarly, few studies have reported the influence of metformin on ERK1/2 (12,13), which directs various metabolic processes.

In this study, we investigated the molecular mechanisms of the UPR and AMPK/ERK1/2 pathway in the aging process and the effects of metformin on the process of age-related neurodegeneration. The findings of this study may provide a novel strategy for the protection of the brain against aging, specifically by treatment with metformin.

Materials and methods

Animal procedures

A total of 200 male Sprague-Dawley (SD) rats (weighing 80-100 g, 3 weeks old) were purchased from the Experimental Animal Center of Tongji Medical College, Huazhong University of Science and Technology. The rats were provided with free access to standard food and water and were housed in cages with an ambient temperature of 25°C, a controlled humidity of 50% and a 12/12-h light/dark cycle. Prior to treatment, all rats were allowed to acclimatize to their environment for 1 week and were then randomly divided into 3 groups as follows (n=65-67 per group): The control group, the D-galactose (D-gal) group and the D-gal + metformin group (Met group). The rats in the D-gal and Met groups were treated with D-gal (dissolved in 0.9% saline, 500 mg/kg, Sigma) via subcutaneous injection for 8 weeks, while the rats in the control group were administered an equal volume of normal saline on the same schedule. The rats in the Met group were subjected to the daily administration of D-gal (subcutaneously) and metformin (intragastrically; metformin was dissolved in drinking water, 150 mg/kg/day, Dongrui Chemical Co.) for the same 8 weeks. After the 8-week establishment of the model, the rats in the 3 groups were further randomly divided into the 3-month-old, 9-month-old (6 months post-injection) and 15-month-old (12 months post-injection) subgroups. Metformin was intragastrically administered to the rats in the Met groups (the 9-month-old Met group and the 15-month-old Met group) for 8 weeks before each sacrifice time point. The rats in the control and D-gal groups received the same dose of water without metformin on the same schedule. More detailed information on the animal procedures is provided in Fig. S1. All experimental procedures involving the care of animals were performed in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (14). The protocol was under the supervision of the Committee on the Ethics of Animal Experiments of Huazhong University of Science and Technology (Wuhan, China; permit no. IACUC S2219).

Auditory functions were examined by the auditory brain-stem response (ABR) just before each sacrifice time point. When the rats were sacrificed, blood was harvested, and the brain tissues were immediately used in experiments or stored at −80°C for later use.

Cell culture and treatment

Highly differentiated rat pheochromocytoma cells (PC12 cells) were obtained from the Ministry of Education Key Laboratory for Environment and Health of Tongji Medical College (HUST). The cells were cultured in RPMI-1640 medium with 10% fetal bovine serum (Gibco, Thermo Fisher Scientific) and 100 units/ml penicillin at 37°C in an incubator containing 5% CO2 and humidified air. At 24 h after seeding, the cells were pre-treated with the ERK1/2 inhibitor, U0126 (Sigma), and the AMPK inhibitor, compound C (Sigma), for 30 min. The cells were then treated with or without metformin (1,1-dimethylbiguanide hydrochloride, 100 µmol/ml, Sigma) and cultured for 24 h as previously described (15), followed by incubation with D-gal (15 mg/ml) at 37°C for an additional 48 h. Detailed information on the treatment of each group of PC12 cells is described in Fig. S2.

PC12 senescence model

PC12 cells were treated with increasing concentrations of D-gal (3, 6, 9, 12, 15, 18, 21 and 24 mg/ml) for 48 h; there was a significant decrease in cell number from 18 mg/ml (Fig. S3). The Cell Counting kit-8 (CCK-8) assay (Dojindo Laboratories) was used to determine the cell viability and cell proliferation according to the manufacturer's instructions and as previously described (16). D-gal was used to induce senescence in the PC12 cells, which were pre-treated with compound C (40 µmol/l), U0126 (10 µmol/l) or metformin (100 µmol/l) for 24 h for further experiments. The selected concentrations of the chemical inhibitors, D-gal and metformin did not affect cell viability (Fig. S3). The β-galactosidase (β-Gal) assay, which measures the activity of β-Gal by immunohistochemistry, is a simple and effective method for measuring age-related changes in vitro and in vivo (17). The PC12 cells were stained using the β-Galactosidase Staining Kit (C0602, Beyotime) according to the manufacturer's instructions.

Flow cytometry

Annexin V/PI analysis

The Annexin V-FITC Apoptosis Detection kit (KeyGen Biotech) was used to detect cell apoptosis. The PC12 cells were harvested with EDTA-free 0.25% trypsin at 37°C in a cell incubator, washed twice with phosphate-buffered saline (PBS) and resuspended in 500 µl of binding buffer. Subsequently, 5 µl of Annexin V-FITC and PI were added to the PC12 cells for 15 min in the dark at room temperature and analyzed using a flow cytometer (Olympus).

Measurement of mitochondrial membrane potential (ΔΨm)

A Mitochondrial Membrane Potential Assay kit with JC-1 (Beyotime) was used to detect changes in ΔΨm according to the manufacturer's instructoins. After each treatment, changes in the ΔΨm of the PC12 cells were detected using the JC-1 assay.

Measurement of ATP production

The ATP content in the cultured PC12 cells was determined using an adenosine 5′-triphosphate assay kit (A095, Nanjing) according to the manufacturer's instructions.

Hearing test: ABR

The auditory threshold was examined by evoking the ABR 4 times during the entire experiment (at 1, 3, 9 and 15 months of age). The rats (n=10 per group) were anesthetized with ketamine (100 mg/kg) and chlorpromazine (5 mg/kg) via intraperitoneal injection (18), placed in a double-walled radiofrequency and electrically shielded booth, and an insert earphone was set. Stainless steel needle electrodes were fixed subcutaneously ventrolateral to the right and left ears and the vertex for recording. The detailed procedures of the ABR test have been previously reported (1).

Morphological observation: Nissl staining analysis

The rat brains were fixed in 10% neutral formaldehyde solution, embedded in paraffin, and then sectioned at 5 µm. The sections were deparaffinized following standard procedures. The brain slices were soaked in 0.3% toluidine blue solution for 40 min at room temperature. The sections were then dehydrated using 95% ethanol for 5 sec, slightly counterstained with 0.5% eosin for 5 sec, cleared using xylene, placed under cover slips and analyzed under a light microscope (Leica).

Measurement of antioxidant enzyme activity

The levels of malondialdehyde (MDA), which results from lipid peroxidation and the activities of glutathione (GSH), superoxide dismutase (SOD) and catalase (CAT), were measured in vitro and in vivo using the appropriate kits (Nanjing Jiancheng Institute of Biological Engineering) according to the manufacturer's instructions and as previously described (19).

Western blot analysis

Tissues from both sides of the auditory cortex were dissected carefully on ice, and total protein was then extracted immediately from the tissues and from the PC12 cells with an ice-cold radioimmunoprecipitation assay (RIPA) lysis solution (Beyotime). The protein concentration of the samples from both the PC12 cells and the SD rats was quantified using a BCA Protein Assay kit (Beyotime). Equal amounts of protein from each sample were separated by 10-12% sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) and electrotransferred onto poly-vinylidene difluoride (PVDF) membranes. The membranes were blocked with 5% BSA for 1 h and incubated overnight at 4°C with an appropriate dilution of specific primary antibodies. The following antibodies were used: Anti-extracellular signal-regulated kinase (ERK)1/2 and anti-p-ERK1/2 (diluted 1:1,000, 4695P and 4370P, Cell Signaling Technology); anti-forkhead transcription factor (Foxo)3 (diluted 1:1,000, NBP2-24579, Novus Biologicals); anti-p-Foxo3 (diluted 1:1,000, ARG51655, Arigo Biolaboratories); anti-HSP90α (diluted 1:2,000, GTX109753, GeneTex); anti-HSP60 (diluted 1:1,000, 4870, Sigma); anti-AMPKα1/2 (diluted 1:1,000, 21191, Signalway); anti-p-AMPKα1/2 (diluted 1:600, sc-33524, Santa Cruz Biotechnology); anti-GRP78 (diluted 1:2,000, GTX102580, GeneTex); anti-GADD153/CHOP (diluted 1:3,000, NBP2-13172, Novus Biologicals); anti-caspase-3 and anti-cleaved caspase-3 (diluted 1:1,000, 9662 and 9661, Cell Signaling Technology); anti-p53 (1:200, sc-98, Santa Cruz Biotechnology); anti-α-tubulin (diluted 1:1,000, ab18251, Abcam) and anti-β-actin (diluted 1:3,000, Mab1445, Lianke). The membranes were washed 4 times before being incubated with appropriate horseradish peroxidase (HRP)-conjugated secondary antibodies (diluted 1:4,000, ANT019 and ANT020, AntGene) for 1 h at room temperature and developed using ECL Plus reagents (Beyotime) to visualize the membranes. Protein quantification was performed using Quantity One 4.6.2 Software (Bio-Rad).

Quantitative polymerase chain reaction (qPCR)

Total DNA was extracted from the tissues from both sides of the auditory cortex and from 107 cultured neurons using a Genomic DNA Purification kit (Tiangen Biotech Co., Ltd.). The purification and concentration of the DNA was evaluated with a Gene Quant Pro DNA/RNA Calculator (BioChrom). Primers and probes specifically designed for mitochondrial DNA (mtDNA) CD and mtDNA D-loop have been described previously (20). The LC-480 Real-time PCR system (Roche Diagnostics Ltd.) was used for performing the PCR amplification. TaqMan PCR mix (2X, 10 µl, Takara), a probe (0.2 µl for each, 10 mM), primers (0.4 µl of each reverse and forward, 10 mM), distilled water (5 µl), and sample DNA (4 µl, 10 ng/ml) comprised the 20-µl reaction mixture. The cycling conditions were as follows: An initial phase at 95°C for 30 sec, then 40 cycles at 95°C for 5 and 30 sec at 60°C. ΔCq (Cqdeletion−CqD-loop) was used to calculate the abundance of the mtDNA 4,834 bp deletion. The relative expression indicates the factorial difference in the deletions between the subgroups, and it was calculated using the 2−ΔΔCq method (21), where ΔΔCq=ΔCqmtDNA deletion in D-gal group−ΔCqmtDNA deletion in control group or ΔΔCq=ΔCqmtDNA deletion in Met group−ΔCqmtDNA deletion in D-gal group. qPCR and the detailed method of calculation were performed as reported previously by our laboratory (22).

Electron microscopy

Transmission electron microscopy (TEM) was used to examine the ultrastructure of the auditory cortex and the cultured PC12 cells. After the PC12 cells were treated with the appropriate drugs, the cells were fixed in 2.5% glutaraldehyde (Sigma) at 4°C for 24 h. A total of 27 (n=3 from each subgroup) were used for the TEM assay. The extraction of the specimens and the preparation, conditions, and procedures were performed as previously described (22,23).

TUNEL assay

A TUNEL Apoptosis Detection kit (FITC; Yeasen) was used to detect PC12 cell apoptosis. At the same time, the nuclei were counterstained with a DAPI staining solution (1 µg/ml; Beyotime) at room temperature for 3-5 min. A laser-scanning confocal microscope (Nikon) was used to examine the cells.

Statistical analysis

The data are expressed as the means ± standard deviation (SD). Statistical analyses were conducted using GraphPad Prism 5.0 software and Microsoft Excel. A two-tailed, unpaired Student's t-test were used to determine statistical significance when comparing 2 groups, and one-way analysis of variance (ANOVA) followed by a Dunnett's multiple comparisons test was used to compare differences between >2 groups. Differences with a P-value <0.05 were considered to indicate statistically significant differences. All experiments were repeated independently at least 3 times.

Results

Metformin treatment alleviates age-related hearing loss

The ABR threshold was significantly elevated in the 15-month-old rats compared with the 3-month-old rats. There was a significant increase in the average ABR threshold in the D-gal group compared with the control group, while metformin significantly suppressed this elevation, particularly in the 15-month-old rats (Fig. 1). This finding indicated that D-gal accelerated AHL and that metformin exerted a protective effect on D-gal-induced aging in rats.

Figure 1 ABR thresholds in the different groups of rats. The data are expressed as the means ± SD of 10 rats per group. *P<0.05, **P<0.01; #P<0.05, ##P<0.01. ABR, auditory brainstem response; dB SPL, decibels of sound pressure level; m, months; D-gal, D-galactose; Met, metformin.

Metformin attenuates the neurodegeneration induced by D-gal in vivo and in vitro

TEM was used to examine the effects of metformin on neurodegeneration at the ultrastructural level in the auditory cortex and in cultured PC12 cells. The sections from the PC12 cells were analyzed at ×1,700 and ×5,000 magnification (Fig. 4). The sections from the auditory cortex were analyzed at ×1,500, ×5,000, and ×6,800 magnification, respectively focusing on the nuclei, organelles and the myelin. In all rats in the 3-month-old groups (Fig. 2A-C), the nuclear membrane was intact, and the nuclei (N) were round in shape. The chromatin in the 9-month-old control group (Fig. 2D) was uniform, and the mitochondria (white arrows) were normal. In the 9-month-old D-gal group (Fig. 2E), some irregular nuclei and condensed chromatin were noted; moreover, these changes were more significant in the rats in the 15-month-old D-gal group (Fig. 2H). Visible ultrastructural degeneration, such as condensed chromatin, vacuolated mitochondria (black arrows), and irregular nuclei, were found in the 15-month-old rats (Fig. 2G-I). Additionally, the myelin was swollen and disrupted (white arrowhead), and lipofuscin was abundant (black arrowhead). Metformin treatment alleviated these changes induced by D-gal, as the ultrastructure of the organelles appeared healthier (Fig. 2C, F and I).

Figure 4 Electron photomicrographs of PC12 cells. Abundant normal mitochondria (white arrows), an intact nuclear membrane, and normal-shaped nuclei were found in PC12 cells. However, the mitochondria were swollen and vacuolated in the D-gal group, and pretreatment with metformin alleviated the mitochondrial damage. The images on the bottom panels are a larger magnification of the area depicted in the white boxes on the top panels. D-gal, D-galactose; Met, metformin.

Figure 2 (A-I) Ultrastructural morphology of the auditory cortex. The nuclei (N) and the mitochondria (white arrows) were normal in the control groups. Vacuolated mitochondria (black arrows), and irregular nuclei were found in the D-gal groups. The myelin was swollen and disrupted (white arrowhead), and lipofuscin was abundant (black arrowheads) in the 15-month-old rats. Metformin attenuated the D-gal-induced neurodegeneration in the auditory cortex at different ages. m, months; D-gal, D-galactose; Met, metformin.

In addition, β-galactosidase staining was performed to detect the apoptosis of PC12 cells. A significant accumulation of positively stained cells was observed in the D-gal (15 mg/ml) group, while pre-treatment with metformin reduced the number of positive cells (Fig. 3A and B). Considering these results, our model of senescent PC12 cells, which was induced by 15 mg/ml D-gal, was successful and was used in our subsequent experiments.

Figure 3 Determination of senescence in PC12 cells. (A) Senescence was evaluated in PC12 cells treated with 15 mg/ml of D-gal by the Senescence β-Galactosidase Staining kit. Scale bar, 25 µm. (B) Cells positive for β-galactosidase staining were counted and are shown in panel, n=6 for each subgroup. (C) Western blot analysis of (a-c) p53 and (d-f) caspase-3 expression in the auditory cortex of rats. *P<0.05, **P<0.01, ***P<0.001. The data are presented as the means ± SD; n=6 for each subgroup. D-gal, D-galactose; Met, metformin.

In our in vitro experiments, abundant normal mitochondria (white arrows), an intact nuclear membrane and normal-shaped nuclei were observed in the PC12 cells. However, in the D-gal group, the mitochondria were swollen and vacuolated (white arrows); however, pre-treatment with metformin alleviated the mitochondrial damage. However, the inhibition of ERK1/2 and AMPK aggravated these changes (Fig. 4).

Toluidine blue staining was used to determine neuronal loss in the auditory cortex region, the sections were analyzed at ×200 and ×400 magnification (Fig. 5). The results demonstrated that the number of neurons in the auditory cortex regions was significantly reduced in the D-gal group compared with the control group, and these alterations were significantly attenuated by metformin treatment. In the rats in the same age groups, the rats in the control group and Met group had numerous neurons compared with those in the D-gal group, and the neurons in the rats in the control group and Met group were arranged tightly and were morphologically intact. An irregular structure, a decreased number of pyramidal cells, and degenerated pyramidal cells were visible in the auditory cortex region of the brains of 15-month-old rats (Fig. 5A and B).

Figure 5 (A) Sections from the auditory cortex of rats were stained with toluidine blue and used to count surviving neurons. The sections from the auditory cortex were analyzed at ×200 and ×400 magnification, (a-c) 3-Month-old rats; (d-f) 9-month-old rats; (g-i) 15-month-old rats. The top images were analyzed at ×200 magnification, and the bottom images were analyzed at ×400 magnification (n=6 for each subgroup); each paraffin-embedded rat brain sample was cut into several slices; the images in the top and bottom panels are the same mouse brain, but from different slices. (B) The corresponding statistical chart is displayed. (C) qPCR analysis of CD 4,834 levels in the 3-, 9-, and 15-month-old control, D-gal and Met groups. (D) qPCR analysis of CD 4,834 levels in PC12 cells. *P<0.05, **P<0.01; #P<0.05. The data are displayed as the means ± SD; n=6 for each subgroup. m, months; D-gal, D-galactose; Met, metformin; CD, common deletion.

Metformin alleviates the diminished function of the antioxidant system via the AMPK/ERK1/2 signaling pathways

We examined the activity of antioxidants and investigated the mechanisms underlying the effects of metformin on antioxidants. The activities of the antioxidants, GSH, SOD and CAT, were decreased in the blood serum of the rats in the 3-, 9- and 15-month-old D-gal groups, respectively, compared with those in the age-matched control groups, while concurrent metformin treatment significantly attenuated this decrease (Fig. 6A-C). In addition, we also found that the levels of CAT, GSH and SOD were reduced with age. These alterations in enzyme activity were also detected in the PC12 cells, suggesting that GSH, SOD and CAT enzyme activities were markedly decreased in the senescent aging model and increased upon treatment with metformin (Fig. 6E-G). MDA is a product of lipid peroxidation mediated by oxygen free radicals and is widely known as a marker of oxidative stress (24). The data indicated that the MDA levels in the D-gal group increased significantly compared with those in the age-matched control group. In the metformin group, the MDA level was decreased compared with the age-matched D-gal group. In addition, the levels of MDA increased with aging (Fig. 6D and H). All these changes caused by D-gal were attenuated by metformin treatment, while the effect was partially suppressed when the PC12 cells were treated with the AMPK and ERK1/2 inhibitors (Fig. 6E-H). The results also implied that the improved antioxidant capacity induced by metformin occurred through the AMPK and ERK1/2 signaling pathways to a certain extent.

Figure 6 Analysis of antioxidants in vitro and in vivo. Metformin treatment significantly increased the (A) GSH, (B) SOD and (C) CAT levels in the auditory cortex of D-gal-injected rats. Metformin treatment significantly decreased the (D) MDA level compared with that in the respective D-gal groups. The activities of these enzymes were detected in PC12 cells, suggesting that (E) GSH, (F) SOD, (G) CAT and (H) MDA enzyme activities were markedly decreased in the senescent aging model and increased upon treatment with metformin; however, this upregulation was prevented significantly by inhibitors of AMPK and ERK1/2. The data are presented as the means ± SD (n=6). *P<0.05, **P<0.01, ***P<0.001, #P<0.05, ##P<0.01, ###P<0.001. D-gal, D-galactose; Met, metformin; SOD, superoxide dismutase; GSH, glutathione; MDA, malondialdehyde.

Metformin protects the mitochondria through the AMPK/ERK1/2 pathways

The function of the mitochondria was determined through the evaluation of mitochondrial membrane permeabilization and ATP production. The flow cytometry data revealed a decreased mitochondrial membrane permeabilization in the D-gal group, which indicated that the number of early apoptotic mitochondria was increased (Fig. 7B and D). Annexin V/PI staining and flow cytometry analysis of cell death were repeated 4 times and similar results were obtained (Fig. 7A). The average values of the percentages of apoptotic cells from repeated experiments are shown in Fig. 7C. The number of apoptotic mitochondria significantly decreased with metformin treatment, while the protective effect was suppressed when the PC12 cells were treated with AMPK and ERK1/2 inhibitors (Fig. 7). In addition, we detected the ATP levels in the PC12 cells. ATP, which is produced by mitochondria, is the most intuitive indicator of MT function. The levels of ATP decreased in the D-gal group, but increased with metformin treatment; however, when the PC12 cells were simultaneously treated with AMPK or ERK1/2 inhibitors, metformin could not increase the ATP levels (Fig. 7E). Taken together, these results indicated that metformin protected the mitochondria through the AMPK and ERK1/2 signaling pathways to a certain extent.

Figure 7 (A) Annexin V/PI staining and flow cytometry analysis of cell death. LL (lower left) indicates viable PC12 cells. UR (upper right) and LR (low right) represent the late apoptotic cells and early apoptotic cells respectively. UR and LR panels were summed as the total percentage of apoptotic cells in this experiment. This experiment was repeated four times and similar results were obtained. (C) The average value of the percentages of apoptotic cells from repeated experiments are shown. (B) Following treatment, the cells were stained with JC-1 and analyzed by flow cytometry. A decreased mitochondrial membrane permeabilization is a hallmark of early apoptosis. (D) D-gal increased the expression level of abnormal mitochondrial membrane potential, and pre-treatment with metformin reduced it. (E) ATP production in PC12 cells from each group was determined. *P<0.05, **P<0.01, ***P<0.001. The data are presented as the means ± SD; n=6 for each subgroup. D-gal, D-galactose; Met, metformin; U, U0126; CC, compound C.

Metformin treatment decreased the accumulation of the mtDNA common deletion (CD) in vitro and in vivo (Fig. 5C and D). Previous studies have confirmed that the mitochondrial 4,834-bp deletion is a biomarker of AHL, and the 4,834-bp deletion gradually deteriorates with age (25,26). In this study, compared with age-matched control groups, the D-gal groups exhibited a markedly elevated CD level (Fig. 5C); CD accumulated with age, and metformin decreased the CD level throughout the process. The same phenomenon was observed in the PC12 cells (Fig. 5D).

Metformin prevents apoptosis via the AMPK/ERK1/2 signaling pathways

To investigate the stress-related apoptotic factors, caspase-3 and p53, and to examine the effects of metformin on these proteins, changes in protein expression were examined by western blot analysis. In the auditory cortex tissues, significant increases in p53 and cleaved caspase-3 expression were noted after the injection of D-gal. However, we found a significantly decreased expression of p53 and cleaved caspase-3 following treatment with metformin (Fig. 3C). We also analyzed apoptosis by counting the number of cells stained by TUNEL assay. Apoptosis Detection kit (FITC) staining and Annexin V/PI staining were performed on the PC12 cells. The right upper quadrant and right lower quadrant represent the late and early apoptotic cells, respectively. The number of apoptotic cells in the Q2 and Q3 panels (upper and lower right) was calculated, and the percentages of cells undergoing apoptosis was determined (Fig. 7A). Significant increases in the number of TUNEL-positive cells were noted after the injection of D-gal; however, treatment with metformin decreased the percentage of apoptotic cells (Fig. 8). Moreover, treatment with AMPK and ERK1/2 inhibitors partially reversed the protective effects of metformin. Thus, metformin alleviated the apoptosis of the PC12 cells through the AMPK and ERK1/2 pathways.

Figure 8 (A and B) The number of apoptotic cells was determined by TUNEL staining in PC12 cells. *P<0.05, **P<0.01. The data are presented as the means ± SD; n=6 for each subgroup. D-gal, D-galactose; Met, metformin; U, U0126; CC, compound C.

Metformin regulates the UPR to maintain protein homeostasis in vivo and in vitro

Protein misfolding and aggregation are common pathological features of a number of neurodegenerative diseases (27). In this study, significant increases in the levels of UPR-related proteins were observed after the injection of D-gal, which is consistent with the findings of previous studies (28,29). The ER stress markers, GRP78 and CHOP, are also known as major modulators of the ER-UPR. HSP90 and HSP60 are chaperone proteins that assist other proteins in folding properly and are also modulators of the MT-UPR. In our in vivo experiments, we found that the expression levels of HSP90, HSP60, GRP78 and CHOP were upregulated in the 3-month-old D-gal groups; however, metformin significantly decreased these expression levels. The levels of the MT-UPR chaperone proteins, HSP90 and HSP60, were still upregulated in the 15-month-old D-gal groups, and metformin significantly decreased these expression levels. However, the expression levels of the ER-UPR chaperone proteins, GRP78 and CHOP, decreased in the 15-month-old D-gal group compared with the control group, but increased in the 15-month-old Met group compared with the D-gal group (Fig. 9). In our in vitro experiments, HSP90 expression was upregulated following treatment with D-gal, and its expression increased further following additional treatment with metformin. Treatment with AMPK and ERK1/2 inhibitors significantly reversed the effects of metformin (Fig. 10C). A significant elevation in GRP78 expression was also observed in the PC12 cells following treatment with D-gal; however, treatment with metformin led to a significant decline in GRP78 expression. Conversely, in the groups pre-treated with U0126, metformin did not decrease GRP78 expression, but when the groups were pre-treated with compound C, changes were not evident (Fig. 10D). These findings focus on the UPR pathways, suggesting the role of metformin in the maintenance of cellular homeostasis.

Figure 9 Regulation of ER-UPR and MT-UPR in vitro and in vivo. Western blot analysis of (A-C) HSP90, (D-F) HSP60, (G-I) GRP78 and (J-L) CHOP expression in each group of rats. *P<0.05, **P<0.01, ***P<0.001. The data are displayed as the means ± SD; n=6 for each subgroup. D-gal, D-galactose; Met, metformin.

Figure 10 Western blot analysis of the effects of metformin on the (A and B) ERK1/2/AMPK and the (C and D) UPR in PC12 cells. *P<0.05, **P<0.01, ***P<0.001. The data are presented as the means ± SD; n=6 for each subgroup. D-gal, D-galactose; Met, metformin; U, U0126; CC, compound C.

Metformin regulates the AMPK/ERK1/2 signaling pathways in vivo and in vitro during aging

AMPK and ERK1/2 act independently, and the crosstalk between them is frequent. Notably, the crosstalk is highly variable; both cross-inhibition and cross-activation can occur depending on the cell type (30). In this study, metformin treatment significantly activated p-AMPK and p-Foxo3α in the 3-, 9-, and 15-month-old rats (Fig. 11A and B). A significant decrease in the phosphorylation levels of Foxo3α and ERK1/2 was noted after the injection of D-gal. By contrast, we observed an increase in the Foxo3α and ERK1/2 phosphorylation levels in the 3-month-old metformin group (Fig. 11B and C); however, the ERK1/2 phosphorylation levels were decreased in the 15-month-old metformin group (Fig. 11C). In our in vitro experiment, a significant elevation in the phosphorylation levels of ERK1/2 and AMPK was noted following treatment with D-gal, and a further elevation in the levels of these proteins was found following co-treatment with metformin. In addition, the inhibition of ERK1/2 blocked the effects mediated by metformin. However, surprisingly, the inhibition of AMPK slightly increased the phosphorylation level of ERK1/2. Thus, there was some association between ERK1/2 and AMPK, indicating that the AMPK pathway can inhibit the ERK1/2 pathway in PC12 cells to a certain extent (Fig. 10A and B).

Figure 11 Western blot analysis of the effects of metformin on the AMPK/Foxo3/ERK1/2 signaling pathways [(A-C) AMPK, (D-F) Foxo3 and (G-I) ERK1/2] in aging rats. *P<0.05, **P<0.01. The data are presented as the means ± SD; n=6 for each subgroup.

Discussion

D-gal accelerates the process of age-related neurodegeneration

Mice treated with D-gal exhibited oxidant damage in the blood, brain and liver, and D-gal treatment in mice was confirmed as a successful mimetic aging model (31). Our previous study confirmed that D-gal injection induced a mimetic aging effect in the inner ear of Wistar rats (32). Eight weeks of D-gal treatment successfully established an ideal mimetic SD rat model of presbycusis in our previous study (33).

In this study, a significant decline in the levels of antioxidants was noted after the injection of D-gal, as observed in several neurodegenerative disorders. ROS production increases with age, and it is known that a lack of proteostasis, oxidative stress and the dysfunction of organelles are critical factors leading to the onset of age-related diseases (34-36). The mitochondria play a central role in a number of metabolic tasks, such as the regulation of cellular metabolism, cellular proliferation regulation, calcium signaling and hormonal signaling; however, the function of the mitochondria decreases with age (37-39). Studies have suggested that the MT-UPR is triggered when mitochondrial DNA and nuclear DNA are not balanced (40,41). Multiple studies have indicated that there is a strong associatoin between the UPR and longevity, physiology, and age-related diseases (40,42,43). The ROS signal is a major modulator in the process of aging (44); however, the effect of the MT-UPR on longevity does not depend on the ROS signal (40). The 4,834-bp mtDNA deletion in rats is similar to the 3,867-bp deletion in mice and the 4,977-bp deletion in humans; these deletions are also known as CDs, and they have been reported to be the most common age-related mtDNA damage and are used as biomarkers of aging (45-47). Animal models of aging established by our laboratory can trigger oxidative stress and increase the accumulation of the CD in the peripheral and central auditory system (19,48). In this study, we also found the same results in PC12 cells, and the CD accumulated more readily in aged rats. The mitochondrial ultrastructure was damaged, and flow cytometry data revealed that the decreased mitochondrial membrane permeabilization after the injection of D-gal, which indicated that the number of early apoptotic mitochondria was increased. Both mtDNA mutations and mitochondrial ultrastructural damage in vitro and in vivo weaken mitochondrial function. Herein, significant decreases in the number of neurons and severe apoptosis were observed after the injection of D-gal. At the same time, the p53- and caspase-3-dependent apoptosis pathways were activated by D-gal.

The UPR protects organelle protein homeostasis in the aging brain

Recently, a number of researchers have attempted to clarify the mechanism sof the UPR in maintaining homeostasis (49). Although it has been reported that the UPR is involved in age-related neurodegenerative diseases (27); however, the specific role and function of the UPR in various diseases is fairly complex and controversial.

This study indicated an increase in UPR in the D-gal groups. The accumulation of unfolded or misfolded proteins induced by chronic D-gal injection resulted in the elevation in the levels of the stress response-related proteins, HSP90, HSP60, GRP78 and CHOP. In some cases, the aberrant accumulation of UPR-related proteins can be a dangerous signal to activate the inflammatory response (50) and shorten lifespan (40). In this study, hearing ability, the morphology of organelles and multiple functions were improved by achieving a balance of the UPR by metformin treatment. Notably, in the 15-month-old D-gal group, GRP78 and CHOP expression decreased significantly, and greater apoptosis and significant neuronal death were observed. This finding may illustrate that aged rats cannot deal with chronic and excessive stress from D-gal and that ER cannot trigger a sufficient UPR to remove or degrade the unfolded or misfolded proteins. Our results are in accordance with those of other studies that the proper activation and mild stimulation of the UPR can ameliorate metabolic and age-related diseases, yet the improper or excessive activation of the UPR can release harmful signals (50-52). These results indicate that ER-UPR in aged rats cannot cope with long-term persistent oxidative stress or trigger adequate UPR signals to combat the stress of unfolded proteins. Our data also indicate that the UPR is time-dependent and that metformin suppresses the overexpressed MT-UPR in rats following exposure to D-gal, whereas metformin activates the ER-UPR, which is beneficial for healthy brain metabolism, in aged rat brains.

Metformin protects neurons via the UPR through the AMPK/ERK1/2 signaling pathways

In this study, metformin significantly increased antioxidant capability in vitro and in vivo. Other researchers also reported that metformin exerts antioxidant effects on the brain during oxidative stress (53). In this study, a significant decrease in the occurrence of CD was observed in rats treated with metformin and in senescent PC12 cells. Additionally, metformin alleviated the level of apoptosis and changes in ultrastructural morphology through the AMPK and ERK1/2 pathways. Metformin significantly decreased neuronal apoptosis and reversed abnormal neuronal damage; thus, this study indicated that metformin afforded a useful benefit, maintaining a healthy state and delaying the aging process of the auditory system.

Our data not only confirmed that UPR pathways are potential regulating points for the effects of metformin on proteostasis, but also indicated that the inhibition of AMPK or ERK1/2 can influence the UPR. Our investigations revealed long-term changes in AMPK and ERK1/2 signaling in the auditory cortex and in a cellular model of senescence.

ERK1/2 is a member of the MAPK family and integrates external signals into signaling events, promoting a large amount of cell growth and proliferation (54). In many cases, the phosphorylation of ERK1/2 results in the activation of its kinase activity and leads to prosurvival factors (55,56). By contrast, others have reported that ERK1/2 is often related to cell death and apoptosis (57-60). The level of phosphorylated ERK1/2 and its specific function vary with different stimuli, stimulus methods and stimulus duration. In fact, the role of ERK1/2 in the aging process pf the brain remains unknown. A few studies have indicated that the phosphorylation of ERK1/2 in brain tissue decreases with age (61,62). This study found that oxidative stress or unfolded protein stress triggered by D-gal deactivated ERK1/2 in 3-month-old rats, yet metformin recovered the phosphorylation levels to regulate cell proliferation and differentiation in rat brains. However, in 15-month-old rats, the results revealed the overexpression of p-ERK1/2 in aged D-gal-treated rats. This finding was consistent with the changes observed in the morphological and functional experiments. The persistent activation of ERK1/2 contributes to oxidative toxicity and induces cell death (63), and metformin markedly reverses phosphorylation to protect neurons from apoptosis or death. In our in vitro experimetns, D-gal and metformin activated ERK1/2 and AMPK phosphorylation. To our surprise, AMPK inhibition slightly increased the phosphorylation level of ERK1/2, indicating that the AMPK pathway can partially inhibit the ERK1/2 pathway in PC12 cells.

The crosstalk between the AKT signaling pathway and the ERK1/2 signaling pathway is dynamic. It has been reported that ERK1/2 activation is accompanied by a gradual decrease in both AKT levels and apoptosis (64), and phosphorylated AKT inhibits the activation of ERK1/2 (65). The findings of this study that increased levels of AMPK in rat brains were accompanied by a decrease in ERK1/2 and that an increase in ERK1/2 was accompanied by a decrease in AMPK. However, whether metformin suppresses ERK1/2 activity in aged rats through the regulation of AMPK or through UPR signaling requires further investigation. Our findings also indicated that Foxo3α may be a regulator of the protection of neuronal survival against D-gal by metformin.

In conclusion, the present data suggest that D-gal injection induces multiple molecular and functional changes, oxidative stress, mitochondrial dysfunction and abnormal apoptosis in the brain similar to natural aging. Metformin regulates the UPR to promote neuronal survival and cellular homeostasis, ultimately resulting in the enhanced hearing ability of D-gal-treated rats. The protective effect of metformin involves the regulation of the UPR through the AMPK/ERK1/2 pathways. Our studies demonstrated a crosstalk between the AMPK/ERK1/2 and UPR signaling pathways. Metformin-induced cell protection in vitro and in vivo may provide novel therapeutic targets for postponing the aging process of the brain and suggest that metformin may be a potential therapeutic treatment for age-related diseases.

Supplementary Data

Funding

This study was supported by research grants from the National Natural Science Foundations of China (nos. 81873700 and 81230021).

Availability of data and materials

The analyzed datasets generated during the present study are available from the corresponding author on reasonable request.

Authors' contributions

HC, WK and WJK conceived and designed the experiments; HC, BH, YH, XZ, ZH, XC, HS, JY, YL and XY performed the experiments; HC, WK and WJK analyzed the data; HC, BH, YH, WK and WJK wrote the manuscript. All the authors have read and approved the final version of this manuscript.

Ethics approval and consent to participate

All experimental procedures involving the care of animals were performed in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (14). The protocol was under the supervision of the Committee on the Ethics of Animal Experiments of Huazhong University of Science and Technology (Wuhan, China; permit no. IACUC S2219).

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Abbreviations:

Acknowledgments

Not applicable.

References