The male Wistar rats aged 3 months (young, Y) and 24 months (old, O) were purchased from the Animal Center of Harbin Medical University. All procedures conformed to the Guide for Care and Use of Laboratory Animals published by the China National Institutes of Health, and the study protocols involving the use of animals were approved by the Institutional Animal Center of Harbin Medical University. The old rats were randomly divided into old control (O, n = 8), spermine (Spm, n = 8), and spermidine (Spd, n = 8) groups, they were injected intraperitoneally with saline, spermine (2.5 mg/kg), and spermidine (10 mg/kg) for 6 weeks, respectively, young rats injected with saline were assigned to young control (Y, n = 8). Spm and Spd were purchased from Sigma-Aldrich (St. Louis, MO, United States), and the optimal concentration used was performed as previously described (Zhang et al., 2017; Chai et al., 2019; Wang et al., 2020a; Wang et al., 2020b). From each experimental group we euthanized six rats to collect hearts. All animals were housed under conditions of constant temperature and humidity, a standard light/dark cycle, and free access to standard rodent chow and water.

Transthoracic echocardiography with Doppler ultrasound techniques was used to assess left ventricular systolic (LVIDs) and diastolic (LVIDd) parameters and function. Briefly, rats were anaesthetized by inhalation of 1.5%–2% isoflurane. The measurement was performed with Acuson CV-70 (Siemens) using standard imaging planes by an observer blinded to the treatment. The parameters were obtained in M-mode tracings at the papillary muscle level or pulse doppler evaluation of the bicuspid valve and averaged using at least five continuous cardiac cycles.

The IHC, Masson, and TUNEL staining were performed on formalin-fixed and paraffin-embedded cardiac tissue sections. For IHC staining of β-galactosidase, the tissue sections were deparaffinized, rehydrated, and antigen retrieval, and were incubated with anti-β-galactosidase (Santa Cruz, United States) overnight, and then were incubated with biotinylated secondary antibody (Zhongshan, China) followed by avidin-biotin peroxidase complex according to the manufacturer’s instructions. Finally, tissue sections were incubated with Diaminobenzidine (Sigma) until a brown color was developed. For Masson staining, the sections were visualized at 40× magnification, and volume fraction of collagen (VFC) fibril structures were quantified from three sections per heart. Results were averaged from five high power random fields from each section. The apoptotic myocytes were detected by TdT mediated dUTP nick end-labeling (TUNEL) assay using a Cell Death Detection Kit (Roche, Germany) according to the manufacturer’s instructions. Briefly, the cardiac tissue sections were incubated with a TUNEL reaction mixture containing TdT and fluorescein-dUTP after permeability treatment. The TUNEL signal was detected by an anti-fluorescent antibody conjugated with alkaline phosphatase, a reporter enzyme, which catalytically generates a colored product. Three slides from each block were evaluated for percentage of apoptotic cells. Four slide fields were randomly examined with 200× magnification. A total of 100 cells were counted in each field.

Rat heart mitochondria were isolated at 4°C by differential centrifugation using a mitochondrial isolation kit (Beyotime Biotechnology, Inc.). Briefly, the rat left ventricular tissue was washed and added 10 times the isolation buffer A. The tissue was finely minced with scissors and then homogenized in the ice-cold buffer, the following processing we did as previously described (Wang et al., 2014). At last, the homogenate was centrifuged at 10,000 g and the pellet was resuspended in ice-cold isolation buffer B prior to the ultrastructural analysis and mitochondrial respiration function assessment.

Isolated mitochondria: The cardiac mitochondrial homogenate was centrifuged at 3000 g and the pellet was fixed in 3% glutaraldehyde for 2–4 h at 4°C, then washed, dehydrated, embedded in resin, and sectioned. The sections were stained with uranyl acetate and lead nitrate and then visualized with an electron microscope (S4800 Hitachi, Tokyo, Japan).

Mitochondrial oxygen consumption was measured using a Clark-type oxygen electrode (Hansatech Instruments, Norfolk, United Kingdom) in mitochondrial respiration buffer. Pyruvate (5 mM) and malate (5 mM) were used as substrates for complex I-containing mitochondria at a final concentration of 500 µg protein/mL. Mitochondrial State 3 respiration, State 4 respiration, respiratory control ratio (RCR), ADP/O ratio, and proton leakage were measured as previously described (Wang et al., 2014).

Superoxide dismutase (SOD) activity in cardiac tissue was measured using commercial kits (SOD: A001-3-1; Jiancheng Bio. Institute, Nanjing, China) with a spectrophotometer (Perkin- Elmer, Norwalk, CT, United States). Protein concentration was measured using the bicinchoninic acid method (Pierce, Rockford, United States) with bovine serum albumin (BSA) as a standard.

ROS generation was measured by dihydroethidium (DHE) staining assay (Cat No. S0063, Beyotime, China). Briefly, frozen tissue sections of cardiac tissues were incubated with 5 μmol/L DHE at 37°C for 30 min, then washed with PBS three times and analyzed using a fluorescence microscope. Images were taken with an Olympus FluoView FV1000 fluorescence microscope (Olympus Optical Co., Ltd., Takachiho, Japan) at an excitation wavelength of 535 nm, and the maximum emission wavelength is 610 nm.

The mitochondria extracted in rat cardiomyocytes were divided into Y, O, Spm, and Spd. Each sample was mixed by multiple parallel samples, and each group of technologies was repeated once. All the reagents and buffers needed for iTRAQ labeling and cleaning were purchased from Applied Biosystems (Foster City, CA, United States). The iTRAQ labeling protocol was referred to the manufacturer’s instructions. In brief, mitochondrial proteins were extracted by 8 M Urea supplemented with 10 mM DTT, pH 8.5 (Amesco St. Louis, MO, United States) and concentration was determined by the Bradford assay. Proteins were dissolved, denatured, alkylated, and digested with trypsin (Promega) at 37°C overnight. To label peptides with iTRAQ reagent, 1 unit of label (defined as the amount of reagent required to label 100 μg protein) was thawed and reconstituted in 150 μl of isopropanol. The digestions from CSM₁₀₀ and SBM were labeled with 114 and 116 iTRAQ reagents, respectively, and this reaction was repeated with 115 and 117 iTRAQ reagents again to guarantee the accuracy of quantitation. Moreover, a strong cation exchange (SCX) column was used for (Applied Biosystems) separating the mixed peptides.

The fractionated peptides were analyzed by a Q-Exactive mass spectrometer fitted with a nano-liquid chromatography system. The eluent was introduced directly to a Q-Exactive mass spectrometer via EASY-Spray ion source. As for data processing, the Proteome Discoverer software 1.3 was employed to interpret raw data and protein quantitation. The rat protein database was downloaded from Uniprot database and combined with the reversed sequences and sequences of widely spread contaminants, such as human keratins for false discovery rate (FDR) control. Finally, the FDR was set to 0.01 to get high confidence results.

The mitochondrial protein concentration was measured by BCA Protein Assay Reagent Kit (Beyotime, Nantong, China). Samples containing total mitochondrial protein were separated by 10% (w/v) SDS-PAGE and transferred onto a PVDF membrane (Millipore, Bedford, MA, United States) (details are in Western Blotting Assay section). The following antibodies were used: Antibodies for PDK4 (Cat# 12949-1-AP), HADHA (Cat# 10758-1-AP), Annexin 6 (Cat# 12542-1-AP), NNT (Cat# 13442-2-AP), and COX-IV (Cat# 66110-1-Ig) were obtained from Proteintech (Wuhan, Hubei, China). The secondary antibody (alkaline phosphatase-conjugated anti-rabbit IgG) was from Promega Corporation (Madison, WI, United States). The intensities of the protein bands were quantified using a Bio-Rad ChemiDocTM EQ densitometer and Bio-Rad Quantity One software (Bio-Rad, Hercules, United States).

Gene Ontology (GO) provides a controlled vocabulary to describe the attributes of genes and gene products in any organism and is applied for cardiovascular research (Lovering et al., 2009). Combining the GO and Uniprot database, the function and localization was annotated for the differentially expressed proteins. The protein-protein interaction networks (PPI) were analyzed in STRING online database v11.5 (http://www.string-db.org/).

Neonatal rat cardiomyocytes (NRCMs) were isolated and cultured by standard methods as previously described (Wei et al., 2016). Three days after being seeded, these cardiomyocytes were divided randomly into the following. Control group: cells cultured under normal incubation conditions. H₂O₂ group (H₂O₂-induced cell senescence group): cardiomyocytes were pre-treated with 40 μM H₂O₂ for 4 h, then cultured for 24 h the same as the control group. H₂O₂-Spm, H₂O₂-Spd, and H₂O₂-DAC groups: cardiomyocytes were pre-treated with 40 μM H₂O₂ for 4 h, then incubated with 20 μM Spm, 10 μM Spd, or 1 mM dichloroacetate (DCA) for 24 h, respectively. H₂O₂-Spd-DFMO group: before the cells were treated with 40 μM H₂O₂, they were pre-incubated with the 5 mM difluoromethylornithine (DFMO, an inhibitor of ODC) for 1 h. The followed treatment was the same as the H₂O₂-Spd group. Both DFMO and DCA were purchased from Sigma-Aldrich (St. Louis, MO, United States), and the optimal concentration applied was performed as previously described (Wang et al., 2020a).

Cellular senescence was determined by SA-β-gal staining. Staining was performed using SA-β-gal staining kit (Beyotime, Haimen, China) according to the manufacturer’s guidelines. Positive staining was evaluated after 12–16 h incubation at 37°C in a CO₂-free atmosphere. The blue stained cells from 10 different fields were counted with results presented as a percentage of positive cells.

JC-1 (Beyotime Biotech, C2006), a sensitive fluorescent probe for ΔΨm, was used to measure the ΔΨm of cardiomyocytes. In healthy cells with high mitochondrial ΔΨm, JC-1 forms aggregates with intense red fluorescence; in apoptotic or unhealthy cells with low ΔΨm, JC-1 remains in the monomeric form, which shows green fluorescence. According to the manufacturer’s directions, the cells were incubated with JC-1 staining solution (5 μg/ml) for 20 min at 37°C and then washed twice with Tyrode standard solution. The images were viewed and scanned under laser confocal microscopy (OLYMPUS, FV1000, Japan) at 488 nm excitation and 530 nm emission for green, and at 543 nm excitation and 590 nm emission for red. The ratio of red to green fluorescence was calculated as ΔΨm. Mitochondrial depolarization is indicated by a reduction in the red/green fluorescence intensity ratio. Image processing and analysis were performed with ImageJ software.

All statistical analyses were performed using SPSS 17.0 (SPSS, Chicago, IL, United States). Values are expressed as mean ± SEM. Mann-Whitney U Test was used for comparing between two groups for non-normally distributed data. Comparisons between the mean of three or more groups were done by one-way ANOVA. Correlation analysis was performed by linear regression. Differences were considered significant at p < 0.05.

To investigate the effect of polyamines on age-related cardiac changes, we first examine the cardiac function between young (3-months-old) and old (24-month-old) rat hearts with or without Spm and Spd supplementation. Echocardiographic evaluation revealed a significant increase in the left ventricular internal diameter at end-diastole (LVIDd) and end-systole (LVIDs), as well as a decrease in the shortening fraction (FS) and ejection fraction (EF) in old rats compared with young rat hearts. Spm and Spd injection reversed aging-induced unfavorable changes in cardiac function (Figures 1A–D). Furthermore, SA-β-gal was stained in the cytosol of the cardiomyocyte, the aged heart showed significantly higher β-gal staining when compared to the young heart. Spm and Spd supplementation overtly decreased aging-induced β-gal staining elevation in old hearts (Figures 1E,F). Disordered collagen fiber network was also observed in old rat hearts, and collagen volume fraction (CVF) was higher than young rat hearts. After Spm or Spd complement, CVF of old rat hearts was decreased to a greater extent (Figures 1G,H). Similarly, a significant increase of p21 was seen in old hearts compared with the young, and the age-related expression of p21 was significantly attenuated by Spm and Spd treatment in the aged heart (Figures 1I,J). These findings suggested that Spm and Spd treatments can effectively inhibit the decrease of age-related heart function and cardiomyocytes aging.

To evaluate the mitochondrial function under polyamines treatment, transmission electron microscope (TEM) was applied to observe the changes of mitochondrial morphology in cardiomyocytes. Mitochondria are longitudinally arranged in the spaces between myofibrils, and mitochondrial cristae were clearly observed in young cardiomyocytes. However, the myocardial fiber structure was disordered, mitochondria were swollen, and mitochondrial matrix density was decreased in aged rats. In contrast, the polyamines treatment heart showed clear cardiac sarcomeres and a dense mitochondrial matrix, only a few mitochondria were showed mildly swollen (Figure 2A). Furthermore, we isolated cardiac mitochondria and measured respiratory function in rats, including the respiratory rates of State 3 and State 4, the respiratory control rates (RCR), ADP/O ratio, and proton leakage using pyruvate/malate as substrates. Proton leakage (H⁺) was measured as oligomycin-inhibited State 3 respiration, which was normalized to young hearts. We found that the State 3 respiratory rate, RCR, and ADP/O ratio were significantly decreased, and proton leakage was increased in the old myocardial mitochondria compared with the young group (p < 0.05 for all). State 3 respiratory rate, RCR, and ADP/O ratio were higher, and the oxidative phosphorylation independent proton leakage was lower in cardiac mitochondria of polyamines treated old rat hearts than that in normal old hearts. However, there were no significant differences in the mitochondrial State 4 respiratory rate among any of the experimental groups (Figures 2B–E). As we know, mitochondria play a significant role in energy metabolism, redox balance, and apoptosis regulation (Jang et al., 2018). Thus, we evaluated the effect of polyamines on apoptosis in aging hearts by TUNEL assay. The results showed a significant increase in the number of TUNEL-positive cardiomyocytes in cardiac tissue from old hearts compared with young hearts, and polyamines supplementation inhibited such an increase (Figures 2F,G). Furthermore, Spm and Spd treatment significantly attenuated oxidative stress in aging myocardium, evidenced by decrease in SOD activity (Figure 2H) and reduction in the fluorescent intensity of DHE (Figure 2I). These data suggest that polyamines improve mitochondrial respiration function through inhibiting oxidative stress and preventing cell apoptosis in myocardium of the aged rat.

To evaluate potential correlations between polyamines and mitochondrial, we isolated mitochondria from four different groups, including Y, O, Spm, and Spd. We first observed the ultra-structure changes of isolated mitochondria with TEM, and found the mitochondria isolated from young hearts showed intact mitochondrial inner and outer membranes, cristae arranged neatly and matrix dense, while the mitochondria from old hearts showed matrix loose, ruptured inner and outer membrane, cristae decreasing or disappearance, and some of mitochondria changed to vacuole. In contrast, Spm and Spd complement significantly retarded the ultra-structure changes of the aged heart mitochondria. Intact mitochondrial membranes, dense matrix, and more clear mitochondrial structure were also observed after polyamines treatment (Figure 3A). Then, the isolated mitochondria were labeled with iTRAQ regent and analyzed by Q-Elite. Principal component analysis (PCA) of the processed datasets was performed to visualize the clustering trends among the samples selected. In the score plot of PCA, the different groups Y, O, Spm, and Spd were well differentiated (Figure 3B). In total, 36,251 tandem mass spectra were captured and searched against the Uniprot rat protein database using stringent matching criteria, resulting in the confident identification of 1355 gene products. According to the report ions, 364 differentially expressed proteins were identified, of which 95 proteins with more than 1.5 folds were selected (Supplementary Table S1). These proteins represented the difference among Y vs. O, Spm vs. O, and Spd vs. O. There were 53 upregulated (ratio < 1.50) proteins and 19 downregulated (ratio < 0.67) proteins between young and old rats. As for Spm injection group, there were 6 upregulated proteins and 18 downregulated proteins. Whereas 9 upregulated proteins and 13 downregulated proteins were identified in the Spd injection group. Additionally, some differentially expressed proteins were identified in both Spm or Spd groups, such as pyruvate dehydrogenase kinase (PDK4), protein Hbb-b1, and protein RT1-A1. Owing to contamination during mitochondria isolation, the localization information of all the differentially expressed proteins should be annotated. To do so, GO, a well-established knowledge database, was employed to give systematic annotation for these proteins. As a result, there are 75 proteins localized in mitochondria, which is about 20% in all 364 differentially expressed proteins (Figure 3C; Supplementary Table S2). While proteins localized in cytoplasm and extracellular space may be contaminants in mitochondria isolation. The function of identified proteins was also annotated by GO, which includes metabolism, translation, transport, apoptosis, and oxidative phosphorylation, and most of these items are tightly related with mitochondrial physiological function (Figure 3D). Among all these 75 mitochondrial proteins, 13 proteins were differentially expressed more than 1.5-fold, Venn diagrams represent the distribution of protein identifications for each group compared to the old (Figures 3E,F). Furthermore, heatmap analysis of the 75 mitochondrial proteins revealed notable differences between Y, O, Spm, and Spd, and some proteins expressed in the polyamines treatment group were remarkably close to the young group (Figure 3G). In summary, our results suggested that polyamines treatment might have an effect on alleviating cardiac aging.

Next, the MetaCore tool was used to study canonical pathways differentially regulated by Spm and Spd treatment in aged rat mitochondria, respectively. Aging gave the most alteration on the metabolic pathway, while the biosynthesis process and oxidative phosphorylation were also altered when comparing young and old rat mitochondria (Figure 4A). Crucial polyamines-altered biological processes include metabolic pathway, oxidative phosphorylation, and aging-related diseases (Figures 4B,C), which means polyamines treatment has an effect on aging-related metabolic pathways and diseases. To gain more insight into the biological functions and molecular characteristics of the differentially expressed proteins, protein–protein interactions (PPI) networks were constructed by the STRING database, which has a collection of known and predicted PPI, including direct (physical) and indirect (functional) associations. We examined 53 mitochondrial proteins differentially expressed in young compared to old, 22 proteins followed the Spd treatment and 13 proteins followed the Spm treatment vs. old (Figures 4D–F). More interestingly, there were conspicuous PPI among polyamines and aging. Thus, we speculate that polyamines treatment may play a significant role in regulating the mitochondrial-related metabolic mechanism and aging progress. Specially, PDK4 was obviously downregulated both in Y vs. O (Y:O = 0.50104) and Spm vs. O (Spm:O = 0.490375). PDK4 also has a decreasing trend after Spd treatment (Spd:O = 0.736708). Of note, there are certain differences in the canonical pathway and PPI networks of mitochondria proteome after Spm and Spd treatment. Spd treatment detected more differentially expressed mitochondrial proteins, indicating that Spd has a stronger regulatory effect on mitochondria in aging rats. Furthermore, we analyzed the interaction network of PDK4 from the String database, the members of which include RPS6KB1, RPS6KB2, DLD, DLAT, PDHX, PDHB, PRKCG, AKT1, AKT2, and AKT3 (Figure 4G). Some of them have already been reported to regulate the aging process, but mitochondrial PDK4 regulation by polyamines stimulation in aging is still not clear. Thus, we paid our attention to these mitochondrial proteins for further biological verification.

To confirm the result of the iTRAQ experiment, 4 differentially expressed proteins, PDK4, mitochondrial trifunctional enzyme subunit α (HADHA), Annexin6, and nicotinamide nucleotide transhydrogenase (NNT) were chosen for validation. In addition to PDK4 and NNT, Annexin6 is closely associated with the cristae in the inner membrane of mitochondria (Rainteau et al., 1995) and HADHA catalyzes the last three steps of mitochondrial beta-oxidation of long-chain fatty acids (Liang et al., 2018; Adeva-Andany et al., 2019). Results showed the relative protein level of PDK4, NNT, Annexin6, and HADHA, as normalized to COX-IV. As expected, the results of Western blotting were in line with our bioinformatic analysis of mitochondrial proteome. Compared with the old hearts, PDK4 was downregulated and NNT was upregulated in both Spm and Spd treated groups, and PDK4 differed significantly (Figures 5A,B). However, the HADHA level was upregulated and Annexin 6 was downregulated in only the Spd treated group, but no response to Spm supplementation (Figures 5C,D). Hence, our results suggested that the protected effects of polyamines in aging cell mitochondria were associated with mitochondrial PDK4 protein.

Following our observation that PDK4 level reduced after polyamines supplementation, we further investigated the correlation between PDK4 and polyamines in aging cardiomyocytes. Primary culture neonatal rat cardiomyocytes (NRCMs) were induced aging by H₂O₂. Cell aging was identified by SA-β-gal staining and p21 expression level. We found that aging cardiomyocytes showed enlarged volume, flattened morphology, and more SA-β-gal positive cells. The expression of p21 was increased obviously more than that in the control group, while Spm and Spd treated groups showed a decrease in SA-β-gal positive cells and p21 expression. Difluoromethylornithine (DFMO, an inhibitor of ODC) treatment abolished the Spd-mediated decrease (Figures 6A–D). We further observed a significant increase in the expression of PDK4 after H₂O₂ induction, and polyamines treatment lowered the expression. In addition, DFMO abolished the Spd-induced decrease of PDK4 expression (Figures 6C,D). These data suggested that polyamines retard cell aging partly through inhibiting PDK4 in NRCMs. As we know, a hallmark of cardiac aging is total number of cardiomyocytes decreased by apoptosis (Broughton et al., 2018). To elucidate the role of Spm and Spd in H₂O₂-induced cell apoptosis, we evaluated the loss of Δψm by JC-1 staining (Figures 6E,F) and assessed Cytochrome-C (Cyt-C) and BCL2 expression by Western blotting (Figures 6G,H). JC-1 assay showed that compared to the control group, Δψm in the H₂O₂ group was notably compromised, presenting as a large area of green fluorescence. Regarding protein expression, the expression level of Cyt-C was elevated accompanying declining levels of BCL2 expression after H₂O₂ pre-treatment. As expounded in Figure 6, pre-treating with 10 μM Spm or 20 μM Spd before the cell incubation with H₂O₂ protected the cells against apoptosis, showing Δψm and BCL2 level elevation and Cyt-C expression reduction. DCA, an inhibitor of PDK4, was utilized to assess the influence of PDK4 on H₂O₂-induced apoptosis. Accordingly, DCA treatment got the same trend with Spm and Spd treated groups (Figures 6E–H). Hence, the effect of polyamines acted as a PDK4 inhibitor to markedly prevent cardiomyocytes apoptosis after H₂O₂-induction.