Meldonium was synthesised and provided by Yuanye Biological Technology Co., Ltd. (Shanghai, China). It was diluted in sterilised water for in vivo experiments, dissolved in normal saline, and diluted with RPMI-1640 medium for in vitro experiments. Acetazolamide was purchased from Selleck Chemicals (Batch No. S4506. Houston, United States) and was dissolved in sterilised water containing 40% cyclodextrins and 1% dimethyl sulfoxide (Sinopharm, Shanghai, China) and diluted with RPMI-1640 medium. Biotin-labelled meldonium was prepared by Wayen Biological Technology Co., Ltd. (Shanghai, China). Biotin was obtained from Invitrogen Co., Ltd. (Massachusetts, United States). Biotin-labelled meldonium and biotin were diluted in RPMI-1640 medium.

Adult male Swiss mice weighing 20 ± 2 g were provided by Beijing SiPeifu Biotech Co., Ltd. (Beijing, China). Mice were treated humanely and maintained in a temperature-controlled room (22 ± 2°C) with a 12-h light/dark cycle (lights on at 07:00 and off at 19:00); food and water were available ad libitum. All animal experiments were conducted in accordance with the national legislation and approved by the Institutional Animal Care and Use Committee (IACUC number: IACUC-DWZX-2021-762, Laboratory Animal Center of the Academy of Military Medical Science, Beijing, China).

Mice were divided randomly into six groups: 1) Control group: physiological saline, 2) Hypoxia group: Mice were pre-treated with physiological saline before being placed into a hypobaric hypoxia experimental chamber to simulate acute exposure to high altitude (Guizhou Fenglei aero armoury Co., Ltd., Guizhou, China). The chamber conditions were as follows: equivalent height above sea level, 8,000 m; operating time, 24 h; and simulated lifting speed, 30 m/s. (3–5) Meldonium groups: Mice were pre-treated with different concentrations of meldonium (50, 100, or 200 mg/kg) before exposure to hypobaric hypoxia. 6) Acetazolamide group: Mice were pre-treated with acetazolamide (50 mg/kg), which was used as a positive control before exposure to hypobaric hypoxia. Meldonium and acetazolamide were administered once daily for 3 days. Following the induction of hypoxia for 24 h, mice were sacrificed by euthanasia for experiments. This study was performed in accordance with the guidelines for the care and use of laboratory animals, of the National Institute of Health, United States (Guide for the Care and Use of Laboratory Animals, 2011).

Twenty-four hours after hypoxia exposure, the animals were anaesthetised with ketamine (90 mg/kg) and xylazine (5 mg/kg) (intraperitoneal injection). Blood was collected from the abdominal aorta of mice and analysed using an automated blood gas analyser (Radiometer, Copenhagen, Denmark). Blood pH, partial pressure of carbon dioxide (pCO₂), and bicarbonate ion concentration (cHCO₃ ⁻) were recorded.

After anaesthetisation, mice were perfused with 4% paraformaldehyde. Lung tissue was collected and fixed in 4% paraformaldehyde overnight and cut into 4 μm thick sections for haematoxylin and eosin (H&E) staining. The pathological changes were evaluated by three pathologists, who were blinded to the design, using an inverted microscope (IX71, Olympus, Tokyo, Japan). Lung injury was detected using the Smith scoring criteria: pulmonary hyaline membrane formation, alveolar cavity enlargement, alveolar wall thickening, haemorrhage, necrosis, and inflammatory cell infiltration. Lung injury severity was graded from 0 to 4.

Lung tissue (30 mg) was mixed with cold normal saline at a ratio of 1:10 (weight to volume), and homogenised by using a high-speed cryogenic tissue grinder (Servicebio, Wuhan, China). SOD activity and MDA content in the lung were measured using commercial kits (Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer’s instructions. MDA content was measured at 532 nm, while SOD activity was 450 nm.

Rat alveolar type II epithelial RLE-6TN cells (CRL-2300, ATCC, Virginia, United States) were cultured in RPMI-1640 medium (Gibco, California, United States) supplemented with 10% foetal bovine serum (HyClone, Logan, Utah, United States). Cells were seeded into cell culture plates at a density of 5 × 10⁴ cells/mL. The cells were divided into six groups: 1) Control group and 2) Hypoxia group, cells were cultured with normal complete culture medium; (3–5) Meldonium groups, complete culture medium containing 10, 20, or 40 μM meldonium, respectively; 6) Acetazolamide group, complete culture medium containing 10 μM acetazolamide. After culturing in a 5% CO₂ incubator at 37°C with saturated humidity for 24 h, except of the cells in the control group, other groups were placed in a humidified 37°C hypoxia incubator (Baker Ruskinn, United States) for 24 h. The conditions of the hypoxia incubator were as follows: 0.5% O₂, 5% CO₂, supplemented with N₂.

Cell viability was assayed using the Cell Counting Kit-8 (CCK-8, Dojindo, Kumamoto Prefecture, Japan) according to the manufacturer’s protocol. The optical density of the samples was measured at 450 nm using a microplate reader (Molecular Devices, California, United States).

The lactate concentration in the supernatant of the cell culture medium was quantified using a Lactate Assay Kit (I256, Dojindo, Kumamoto Prefecture, Japan) according to the manufacturer’s protocol. The absorbance of the sample was measured at 450 nm.

Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) represent the functions of mitochondria and glycolysis. Metabolic changes in the cells were analysed using a Seahorse XFe96 Extracellular Flux Analyser (Agilent Technologies, California, United States). 4×10³ cells per well were seeded into an XF 96 cell culture microplate (Agilent Technologies, California, United States). OCR and ECAR were assayed using the Seahorse XF Cell Mito Stress Test Kit (Agilent Technologies, California, United States, 103015–100) and Seahorse XF Glycolysis Stress Test Kit (Agilent Technologies, California, United States, 103020–100). Inhibitors and activators were used per well at the following concentrations via ports A–C: 1) ECAR including glucose (10 mM), oligomycin (1 μM) and 2-DG (50 mM); 2) OCR including oligomycin (1.5 μM), FCCP (0.5 μM) and Rot/AA (0.5 μM). Data were assessed using the Seahorse XF96 Wave software (Agilent Technologies, California, United States).

The HuProt™ human proteome microarray, composed of approximately 20,000 purified human full-length proteins with N-terminal glutathione S-transferase tags, was provided by Wayen Biotechnology (Shanghai, China) to investigate the target proteins of meldonium. First, the microarray was incubated with blocking buffer (5% bovine serum albumin in 1× PBST) at 4°C in the dark for 1.5 h. Microarrays were then incubated with 1 μM biotin-meldonium or 1 μM biotin at 37°C for 1 h. After washing three times with PBST and twice with distilled water, Cy5-Streptavidin solution (1:1,000 dilution) was added and incubated at 37°C for 20 min. Then, the microarray was washed with PBST 0.1 × PBS and subsequently distilled water and spun dry at 1,000 rpm for 2 min. A GenePix 4000 microarray scanner (Axon Instruments, California, United States) was used to visualise and record the results, and GenePix™ Pro v6.0 (Axon Instruments, California, United States) was used for data analysis.

The candidate positive proteins were identified by the Z-Score of protein spot (more than 2.8 in the biotin-meldonium microarray and less than 2.8 in the biotin microarray). Next, the candidate positive protein was determined as the final meldonium target protein when the Imean-Ratio was more than 1.4, obtained by calculating the ratio of Z-Score between biotin-meldonium microarray and biotin microarray. The enrichment analysis of target protein candidates, including the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway, biological process, molecular function, and cellular component, was performed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) v6.7.

Molecular docking was employed to verify the binding energy between meldonium and the target protein PFKP. The protein crystal structure and PDB format file of PFKP with PDBID 4XYK were downloaded from The PDB Protein Database (http://www.rcsb.org/pdb/home.do). The structure of meldonium was downloaded from The Database of The Chemical Book (http://www.chemicalbook.cn). The target protein and compound molecule were hydrogenated, charged, and optimised using MOE software (Molecular Operating Environment, 10.2008, chemical computing group). After structural optimisation, the ligand interacts with the protein receptor. Ligand-receptor binding was evaluated using the ASE function score value. Low binding energy represents a strong ligand-receptor affinity.

RLE-6TN cells were seeded in a plate with a glass bottom (Cellvis, Maryland, United States) and divided into four groups: 1) Negative control group: normal complete culture medium; 2) Meldonium group: complete culture medium containing 100 μM meldonium; 3) Biotin-labelled meldonium group, complete culture medium containing 100 μM biotin-labelled meldonium; 4) Biotin group: complete culture medium containing biotin. After incubation for 24 h, the cells were fixed with 4% paraformaldehyde for 30 min and washed with cold PBS three times. Cells were permeabilised with 0.2% Triton X-100 at 4°C for 10 min and blocked with 5% fat-free milk for 30 min at room temperature, followed by incubation with the primary antibody PFKP (1:300, Abcam, Cambridge, United Kingdom) at 4°C overnight. The samples were then incubated with Alexa Fluor^® 488 donkey anti-rabbit IgG secondary antibody (1:1,000, ab150061, Abcam, Cambridge, United Kingdom) at room temperature for 1 h in the dark. The cells were incubated with streptomycin (1:1,000, Thermo Fisher Scientific, Massachusetts, United States) at room temperature for 2 h, and then stained with Hoechst 33342 (1:1,000, H3570, Invitrogen, California, United States) at room temperature for 5 min in the dark. Microphotographs were observed using ImageXpress Micro Confocal (Molecular Devices, California, United States) by a colleague blinded to the experimental design. The cells co-localised with biotin-labelled meldonium (red) and anti-PFKP antibody-labelled PFKP (green) appeared yellow or orange.

RLE-6TN cells in 6-well cell culture plates at a density of 2×10⁷ cells/ml were collected and lysed using RIPA buffer (Thermo Fisher Scientific, Massachusetts, United States) containing a protease inhibitor cocktail (Bimake, Houston, United States), centrifuged at 14,000 × g for 15 min at 4°C. The supernatant was added to centrifugal filters (Millipore, Massachusetts, United States) and centrifuged at 4,000 × g for 40 min at 4°C. Biotin-meldonium or biotin (200 μl, 10 μM) was added to the supernatant. To allow meldonium to bind the protein, the supernatant was rotated overnight at 4°C. Streptavidin-agarose beads (150 μl, Invitrogen, California, United States) were added, and the supernatant was rotated for 1 h at room temperature. Agarose beads bind to meldonium through the biotin modification and thereby retains proteins complexed with meldonium. After centrifugation at 1,000 × g for 3 min and the supernatant was discarded, RIPA buffer (1 ml) was added to the tubes and then rotated for 5 min. Repeated three times. The remaining beads were mixed with the same volume of loading buffer and boiled for 5 min. The eluted proteins were detected by western blotting.

PFK activity was measured using a PFK activity colorimetric assay kit (Abcam, Cambridge, United Kingdom) according to the manufacturer’s protocol. Absorbance was measured at 450 nm every 3 min using a microplate reader (Molecular Devices, California, United States). PFK activity is expressed as nmol/min/ml. Experiments were performed at least three times.

Total RNA was extracted from RLE-6TN cells using TRIzol reagent (Invitrogen, Carlsbad, CA), and purified using chloroform, isopropanol, and 75% ethanol. Reverse transcription of cDNA was carried out using a PrimeScript™ RT Reagent Kit with gDNA Eraser (Takara, Osaka, Japan). Quantitative real-time PCR analysis was conducted using a TB Green^® Premix Ex Taq™ II (Takara, Osaka, Japan) on a Bio-Rad CFX96 Real-Time PCR Detection System (Bio-Rad Laboratories Ltd., California, United States). Amplification of the products was conducted as follows: 95°C for 30 s, 40 cycles of 95°C for 15 s, 60°C for 30 s, and 72°C for 30 s. A melting curve was delineated using 0.5°C temperature increments every second from 65 to 95°C. The PCR primers were designed using Primer 5.0 software (Premier, Canada), and synthesised by Beijing TsingKe Biotechnology Co., Ltd. (Beijing, China). The primer sequences used were as follows (5ʹ-3ʹ): LDHA Forward: GCCCGCCTGAAGAAGAG, Reverse: GCAGCCTGGACAGTGAA; PFKP Forward: CTCCGCTGTTCGGGTTG, Reverse: GGGTAGGGTGCGTTTCG; PKM2 Forward: ACC​TGG​TGA​CAG​AAG​TGG​A, Reverse: CGGATGAAAGACGCAAA; PDK1 Forward: AAC​TAA​ATG​CGA​AAT​CAC​C, Reverse: TAAACGCCTTTGTCTGC; β-actin Forward: CGC​GAG​TAC​AAC​CTT​CTT​GC, Reverse: CCT​TCT​GAC​CCA​TAC​CCA​CC. Three replicates were used for each sample. Normalisation was performed for β-actin. Relative gene expression was determined using the 2^−ΔΔCt method.

Cytoplasmic and nuclear proteins in the lung or RLE-6TN cells were extracted using the Cytoplasmic and Nuclear Protein Extraction Reagents Kit (KeyGEN, Nanjing, China) and quantified with a BCA protein reagent kit (KeyGEN, Nanjing, China). Equal amounts of denatured protein were separated by 8% or 10% SDS-PAGE and transferred to PVDF membranes (Millipore, Massachusetts, United States). The membrane was blocked with 5% skim milk at room temperature for 2 h and then incubated with the corresponding primary antibody diluted in blocking buffer at 4°C overnight. Antibodies against PFKP (1:2,000, ab204131), OPA1 (1:2,000, ab42364), and DRP1 (1:2,000, ab184247) were obtained from Abcam (Cambridge, United Kingdom). Antibodies against PKM2 (1:2,000, #4053), LDHA (1:2,000, #2012), PDH (1:2,000, #3205), and β-actin (1:2,000, #4970) were obtained from Cell Signalling Technology (Danvers, MA, United States). The antibodies to Nrf2 (1:2,000, A0674), MFN1 (1:2,000, A9880), MFN2 (1:2,000, A19678), FIS1 (1:2,000, A19666), AQP1 (1:1,000, A15030), AQP4 (1:1,000, A2887), PDK1 (1:2,000, A8930), and lamin B1 (1:2,000, A16909) were obtained from ABclonal (Wuhan, China). Horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (H + L) (1:1,000, zb2301) was purchased from ZSGB-BIO (Beijing, China). After incubation with HRP-conjugated secondary antibodies at room temperature for 1 h, a Pro-light HRP Chemiluminescent Kit (TianGen Biotech, Beijing, China) was used to detect the target protein. β-actin and lamin B1 were used as internal controls to normalise the relative expression of each protein. The optical densities of the bands were quantified using the Image Lab analysis software (Bio-Rad, California, United States).

Data are expressed as the mean ± standard error of the mean. Experimental data were analysed using SPSS Statistics software (version 19.0; IBM Corporation, Armonk, NY, United States) and GraphPad Prism software (Version 8.0, San Diego, CA). Differences among multiple groups were analysed via one-way analysis of variance followed by Fisher’s least significant difference test. Statistical significance was set at p < 0.05.

Lung tissue sections stained with H&E (Figure 1A) showed that, compared with the control group, severe lung injury had occurred in mice exposed to hypoxic condition, including thickening of the alveolar walls, infiltration of inflammatory cells, and substantial haemorrhage, indicating that the model of acute exposure to hypoxic conditions was successfully established. Meldonium (50 and 100 mg/kg) modestly attenuated these histological changes. In particular, the meldonium (200 mg/kg) group showed a thin wall structure in the alveolar wall and a nearly complete structure of the alveolar cavity. Furthermore, the lung injury score in vivo in the hypoxia group was significantly higher than that in the control group (Figure 1B, p < 0.001). However, meldonium (50, 100, and 200 mg/kg) decreased the lung injury score compared with the hypoxia group (p < 0.001). Moreover, the expression of aquaporin 1 and 4 (AQP1 and AQP4) was significantly increased compared with the control group in vivo (Figure 1C, p < 0.05), whereas meldonium (100 mg/kg) markedly decreased AQP1 expression compared with the hypoxia group (100 mg/kg, p < 0.01). CCK-8 assay results in vitro showed that cell viability was markedly decreased in the hypoxia group compared with the control group (Figure 1D, p < 0.001), whereas meldonium (10 and 40 μM) treatment increased cell viability (p < 0.01 or p < 0.001). Blood gas analysis (Figure 1E) showed that pH, partial pressure of carbon dioxide (pCO₂), and bicarbonate ion concentration (cHCO₃ ^-) were significantly decreased in the hypoxia group when compared with the control group (p < 0.001 or p < 0.05), indicating that hypoxia-induced lung injury in mice was accompanied by metabolic acidosis in vivo. However, meldonium significantly ameliorated metabolic acidosis by decreasing the level of pCO₂ (100 and 200 mg/kg, p < 0.001) and increasing the level of cHCO₃ ^- (200 mg/kg, p < 0.001) compared to the hypoxia group. Taken together, these findings demonstrate that hypoxia-induced lung injury in vivo and in vitro can be significantly ameliorated by meldonium pre-treatment, with pharmacodynamic effects similar to acetazolamide.

The lactate concentration data showed that hypoxia increased lactate concentration compared with the control group in vitro (Figure 1F, p < 0.001), however, meldonium (10, 20, and 40 μM) significantly decreased glycolysis, as indicated by the lower lactate concentration in the cell supernatant compared with the hypoxia group (p < 0.001 or p < 0.01). Accordingly, the extracellular acidification rate (ECAR) in vitro was increased in the hypoxia group compared with the control group, while meldonium reduced the ECAR (Figure 1G). The results of ECAR were analysed in terms of glycolysis, representing the glycolytic ability of cells under basal conditions which was significantly increased in the hypoxia group compared with the control group (p < 0.01) and decreased by meldonium (40 μM, p < 0.001). Moreover, glycolytic capacity, representing the ability of cells utilising glycolysis to achieve maximum productivity, was significantly increased in the hypoxia group (p < 0.001), and decreased by meldonium compared with the hypoxia group (40 μM, p < 0.001). Taken together, these findings suggest that meldonium has the ability to regulate glycolysis in vitro.

To confirm the target protein of meldonium, HuProt™ human proteome microarrays assay was performed (Figure 2A). Eighty candidate meldonium target proteins were identified (Figure 2B). The results of the KEGG pathway (Figure 2C) showed that the enriched pathways were mainly involved in energy metabolism, including carbon, glycine, serine, pyruvate and threonine metabolism, glycolysis/gluconeogenesis, and central carbon metabolism. The cellular component indicated that most meldonium target proteins localised to the cytoplasm (Figure 2D). The target proteins were involved in metabolic processes (Figure 2E). Finally, the main function of the target protein was catalysis (Figure 2F). Altogether, the target proteins of meldonium are related to intracellular proteins that have catalytic activity in the metabolic pathways in human cells and require further study to identify.

Consistent with meldonium regulating glycolysis, we found from the human proteome microarray assay that major potential target proteins of meldonium are related to glycolysis (Figure 3A), including PFKP (phosphofructokinase, platelet-type), PFKL (phosphofructokinase, liver type), PKLR (pyruvate kinase, liver, and RBC). The main glycolytic enzymes, including PFKP and M2 isoform of pyruvate kinase (PKM2), catalyse the conversion of glucose to pyruvate and then to lactate via the A isozyme of lactate dehydrogenase (LDHA). Pyruvate can also be converted to acetyl coenzyme A by pyruvate dehydrogenase (PDH) in mitochondria, which was regulated by pyruvate dehydrogenase kinase 1 (PDK1), and then participates in aerobic oxidation in mitochondria. These enzymes are involved in glycolysis and pyruvate metabolism. Thus, we first investigated the mRNA expression of enzymes involved in glycolysis and pyruvate metabolism in vitro. The mRNA levels of PFKP, PKM2, LDHA, and PDK1 were higher in the hypoxia group than that of the control group (Figure 3B, p < 0.001), while meldonium could significantly decrease the mRNA expression of PFKP, PDK1, and PKM2 compared with the hypoxia group (10 or 40 μM, p < 0.05).

Accordingly, western blot assay (Figure 3C) showed that the protein expression of PFKP, PKM2, LDHA, and PDK1 was markedly increased after hypoxia exposure in vitro (p < 0.01 or p < 0.05 or p < 0.001). But meldonium significantly reduced the protein expression of PFKP, PKM2, and LDHA (10, 20 or 40 μM, p < 0.05 or p < 0.01) compared with the hypoxia group, while PDK1 was decreased without a significant difference (p > 0.05). There were decreases without significant difference (p > 0.05) compared with the control group in the protein expression of PDH after exposure to hypoxia, which was significantly increased (p < 0.05, p < 0.01) by meldonium (10 and 40 μM) compared with the hypoxia group. Additionally, the protein expression of PFKP, LDHA, and PDK1 was significantly enhanced compared with the control group in vivo (Figure 3D, p < 0.05 or p < 0.01), while PKM2 did not show a significant difference (p > 0.05). However, meldonium significantly decreased the expression of PFKP, LDHA and PDK1 (10, 20 or 40 μM, p < 0.05), while PKM2 did not show a significant difference (p > 0.05) compared with the hypoxia group. Western blot results suggested that hypoxia-induced reduction of PDH and increase in PDK1 protein expression were restored by meldonium pre-treatment, which reflected the recovery of aerobic oxidation in mitochondrial energy metabolism by meldonium pre-treatment. Taken together, these results indicate that meldonium has the ability to regulate glycolytic enzymes and mitochondrial metabolic enzymes, which promotes pyruvate metabolism to aerobic oxidation in mitochondria.

PFKP is the rate-limiting enzyme of glycolysis and plays a pivotal role in modulating energy metabolism. Consequently, we examined the interaction of meldonium with PFKP protein. First, molecular docking of PFKP with meldonium was performed using MOE software (v2021.09). The binding mode and stability of meldonium with PFKP were analysed by comparing the hydrogen bonds, ionic bonds, and binding energies. The molecular docking results (Figure 4A) show that meldonium forms a hydrogen bond in Asp564 of PFKP and two hydrogen bonds in Asp561 of PFKP. The binding energy was −4.32 kJ/mol, suggesting that meldonium can interact with PFKP protein. After pre-treatment with biotin-labelled meldonium, immunofluorescence experiments showed meldonium co-localisation with PFKP in vitro (Figure 4B) in the cytoplasm only. The pull-down assay in vitro (Figure 4C) indicated that PFKP protein combined with biotin-labelled meldonium more than the biotin group, suggesting that meldonium has a stable interaction with the PFKP protein. Furthermore, the increase of PFK activity after hypoxia exposure in vitro (Figure 4D, p < 0.001) was restored by meldonium (10, 20, and 40 μM; p < 0.001) compared with the hypoxia group. Collectively, meldonium not only has the potential to bind to PFKP but also regulates PFKP activity. Thus, these results suggest that meldonium regulates glycolysis by targeting PFKP.

We first investigated the expression of Nrf2 in the cytoplasm and nucleus in vivo (Figure 4E). There was a notable decrease in Nrf2 protein expression in the cytoplasm after hypoxia compared with the control group in vivo (p < 0.05), while meldonium (200 mg/kg) significantly increased the expression of Nrf2 compared with the hypoxia group (p < 0.01). Furthermore, although not significantly (p > 0.05), the protein expression of Nrf2 in the nucleus was increased after hypoxia, which is beneficial to protect against hypoxia-induced oxidative stress. Compared with the hypoxia group, pre-treatment with meldonium (200 mg/kg) significantly increased the protein expression of Nrf2 in the nucleus (p > 0.05). These data indicate that meldonium promotes Nrf2 transfer from the cytoplasm to the nucleus in vivo.

The changes in oxidative stress-related substances are shown in Figures 4F,G. The level of MDA reflects the degree of lipid peroxidation and SOD, representing the antioxidant capacity. Both indirectly reflect the damage induced by oxidative stress after hypoxia. There was a notable increase in the concentration of MDA (p < 0.05), whereas SOD viability was significantly decreased (p < 0.05) in vivo. However, meldonium significantly decreased MDA content (200 mg/kg, p < 0.05) and increased SOD viability (100 mg/kg, p < 0.05). Therefore, hypoxia-induced oxidative stress in vivo could be ameliorated by meldonium pre-treatment. Overall, these results imply that meldonium promotes the translocation of Nrf2 to the nucleus and resists oxidative stress induced by hypoxia in vivo.

We performed the following experiments to determine whether meldonium has a protective effect on mitochondria after hypoxia in vitro. The cell mitochondria stress test in vitro (Figure 5A) suggested that cell mitochondrial function was severely affected by hypoxia. The OCR evidently decreased after hypoxia exposure, while meldonium increased OCR compared with the hypoxia group but was lower than the control group. Additionally, various mitochondrial functions were improved. Basal respiration, reflecting the demand for energy production in mitochondria, was obviously decreased after hypoxia compared with the control group (p < 0.001), and significantly increased by meldonium compared with the hypoxia group (10, 20, and 40 μM; p < 0.01 or p < 0.05); Additionally, proton leak, an important parameter of mitochondrial damage, was markedly decreased after hypoxia compared with the control group (p < 0.001), while meldonium clearly restored it compared with the hypoxia group (40 μM; p < 0.05); Non-mitochondrial oxygen consumption, which reflects the entire condition of cells, was also greatly decreased in the hypoxia group compared with the control group (p < 0.001), and obviously increased by meldonium compared with the hypoxia group (10, 20, and 40 μM; p < 0.05, p < 0.01, and p < 0.001).

The homeostasis of mitochondrial fission and fusion that represents the normal state and function of mitochondria is regulated by five proteins. As shown in Figure 5B, after hypoxia in vitro, the protein expression of fission protein 1 (FIS1) and optic atrophy 1 (OPA1) significant increased (p < 0.05), mitofusins 2 (MFN2) and dynamin-related protein 1 (DRP1) expressions were markedly decreased (p < 0.05, p < 0.01), while mitofusins 1 (MFN1) levels decreased without any difference (p > 0.05) compared with the control group. In contrast, compared with the hypoxia group, meldonium significantly decreased the protein expression of FIS1 and OPA1 (20 or 40 μM, p < 0.05), and increased MFN2 (20 μM, p < 0.05) and DRP1 expression (20 and 40 μM, p < 0.01, p < 0.05), whereas MFN1 level increased without significant difference (p > 0.05). Thus, we conclude that the damage and dysfunction of mitochondria and imbalance of mitochondrial fission and fusion homeostasis induced by oxidative stress in vitro are ameliorated by pre-treatment with meldonium.