Abstract
Aims:
This study aims to explore the efficacy of punicalagin (PG) on diabetic cardiomyopathy (DCM), with a specific focus on the mechanisms underlying the effects of PG on mitochondrial fusion/fission dynamics.
Results:
Cardiac structural and functional abnormalities were ameliorated in diabetic rats receiving PG administration as evidenced by increased ejection fraction, and attenuated myocardial fibrosis and hypertrophy. PG enhanced mitochondrial function and inhibited mitochondria-derived oxidative stress by promoting Opa1-mediated mitochondrial fusion. The benefits of PG could be abrogated by knockdown of Opa1 in vivo and in vitro. Inhibitor screening and chromatin immunoprecipitation analysis showed that Stat3 directly regulated the transcriptional expression of Opa1 by binding to its promoter and was responsible for PG-induced Opa1-mediated mitochondrial fusion. Moreover, pharmmapper screening and molecular docking studies revealed that PG embedded into the activity pocket of PTP1B and inhibited the activity of PTP1B. Overexpression of PTP1B blocked the promoting effect of PG on Stat3 phosphorylation and Opa1-mediated mitochondrial fusion, whereas knockdown of PTP1B mimicked the benefits of PG in high-glucose-treated cardiomyocytes.
Innovation:
Our study is the first to identify PG as a novel mitochondrial fusion promoter against hyperglycemia-induced mitochondrial oxidative injury and cardiomyopathy by upregulating Opa1 via regulating PTP1B-Stat3 pathway.
Conclusion:
PG protects against DCM by promoting Opa1-mediated mitochondrial fusion, a process in which PG interacts with PTP1B and inhibits its activity, which in turn increases Stat3 phosphorylation and then enhances the transcriptional expression of Opa1. These results suggest that PG might be a promising new therapeutic approach against diabetic cardiac complication. Antioxid. Redox Signal. 35, 618–641.
Introduction
The prevalence of diabetes mellitus (DM) has rapidly increased during the past years, presenting a massive public health challenge worldwide (16). Cardiovascular complication is the leading cause of mortality in patients with diabetes, regardless of whether it is type 1 or type 2 (22, 55). Despite not accelerating the development of diabetes, diabetic cardiomyopathy (DCM) is a primary complication of diabetes that contributes to increased mortality (3). It is relatively clear that DCM is characterized by abnormal cardiac function and structure, including myocardial hypertrophy and fibrosis (38). Although various mechanisms including impaired insulin signaling pathways, glucotoxicity, lipotoxicity, and abnormal endoplasmic reticulum stress have been demonstrated to be involved in the development of DCM (38), current therapeutic agents are still limited.
Innovation
Punicalagin is the prominent nutrition responsible for the anti-oxidative effects of pomegranate juice. This study demonstrates that punicalagin restores Opa1-mediated mitochondrial fusion and then inhibits mitochondrial oxidative stress, which is responsible for the beneficial effects of punicalagin in diabetic cardiomyopathy. This work identifies punicalagin as a novel mitochondrial fusion promoter against hyperglycemia-induced mitochondrial oxidative injury and cardiomyopathy by upregulating Opa1 via regulating PTP1B-Stat3 pathway.
Mitochondria are primary energy suppliers and principal sources of reactive oxygen species (ROS) in the cells (24). It is estimated that mitochondria provide ∼90% energy in the cardiomyocytes (51). In diabetic hearts, cardiac dysfunction is associated with a decreased capacity of ATP synthesis in mitochondria. Moreover, impaired mitochondrial oxidative-respiratory chain results in increased ROS production (24, 57). Accumulating evidence has indicated that increased mitochondrial ROS production and resultant mitochondrial dysfunction are intimately involved in the pathogenesis of DCM (39, 40). Emerging data suggest that even cardiac mitochondria can form a nearly continuous network of tubules, and undergo mitochondrial fusion and fission (25, 54). Imbalanced mitochondrial fusion/fission dynamics has been highlighted as the initial cause for the adverse events (52). Inhibited mitochondrial fusion and excessive mitochondrial fission are observed in diabetic hearts. Hyperglycemia-induced mitochondrial fission increases mitochondrial ROS overproduction and accelerates mitochondrial dysfunction (18). In contrast, the promotion of mitochondrial fusion inhibits mitochondria-derived oxidative stress and improves mitochondrial function (19). Hence, the discovery of effective and promising pharmacotherapy targeting mitochondrial dynamics balance could be a potential strategy to protect against DCM.
Punicalagin (PG), a major ellagitannin in pomegranate, is demonstrated to be the prominent component responsible for the protective effects of pomegranate juice (10). Previous studies have shown that PG reduces myocardial ischemia/reperfusion injury, and prevents hypoxic pulmonary hypertension via antioxidant action (21, 62). Moreover, it has been shown that PG prevents high-fat diet-induced myocardial damage and ameliorates hyperlipidemia-induced endothelial dysfunction by preserving mitochondrial function (8, 48). Given that mitochondrial dynamics balance plays a critical role in maintaining normal mitochondrial oxidative phosphorylation, investigating whether PG is efficient to prevent DCM by regulating mitochondrial dynamics is worthwhile.
Here, using in vivo diabetic rats and in vitro hyperglycemia-treated cardiomyocytes, we explored the effects of PG on DCM, with a specific focus on the mechanisms underlying the effects of PG on mitochondrial fusion/fission dynamics. Our results have revealed that PG protects against DCM by promoting Opa1-mediated mitochondrial fusion via the PTP1B-Stat3 signaling.
Results
Basic characteristics of animals
As shown in Table 1, no significant increases in blood glucose, serum total cholesterol (TC), triacylglycerol (TG), and body weight were observed between the control rats and the control rats receiving PG treatment (30 or 90 mg/kg/day). The diabetic rats exhibited increased blood glucose and serum lipids as well as decreased body weight when compared with the control rats (p < 0.05 or p < 0.01). PG at the dose of 90 mg/kg/day significantly reduced serum TC levels in the diabetic rats (p < 0.05).
Basic Characteristics of Rats
Values presented are mean ± SEM.
N = 8. * p < 0.05, ** p < 0.01 versus Con; # p < 0.05 versus DM.
DM, diabetic mellitus rats; PG(30), punicalagin at the dose of 30 mg/kg/day; PG(90), punicalagin at the dose of 90 mg/kg/day; SEM, standard error of the mean; TC, total cholesterol; TG, triacylglycerol.
Administration of PG protected against DCM in rats
When compared with these values in the control rats, left ventricle ejection fraction (LVEF) was decreased, whereas left ventricle end-systolic diameter (LVESD) was increased in the diabetic rats (Fig. 1a–c). There were no significant differences in left ventricular end-diastolic diameter (LVEDD) among all the groups (Fig. 1a, d). The diabetic rats receiving PG supplement only at the dose of 90 mg/kg/day showed increased LVEF and decreased LVESD (Fig. 1a–c). Hence, 90 mg/kg/day was considered as an efficient dosage, and was indicated as Con+PG or DM+PG group for the following experiments.
Ventricular hypertrophy and myocardial fibrosis are major characteristics of diabetic hearts in DCM (38). Compared with the control rats, significant cardiac hypertrophy and myocardial fibrosis were induced in the diabetic rats, as evidenced by the increases in cardiomyocyte cross-sectional area and collagen volume fraction (Fig. 1e–g). Treatment with PG inhibited the increases in cardiac hypertrophy and myocardial fibrosis in diabetic rats. Compared with the control hearts, there was increased myocardial apoptosis in diabetic hearts as shown by elevated apoptosis index and caspase-3 activity (Fig. 1h–j). Moreover, the diabetic hearts showed increased ROS generation and decreased antioxidant capacity as evidenced by elevated dihydroethidium (DHE) fluorescence density, increased malondialdehyde (MDA) content, decreased activity of total superoxide dismutase (SOD), cellular copper-zinc-superoxide dismutase (CuZnSOD), mitochondrial manganese superoxide dismutase (MnSOD), total antioxidant capacity (T-AOC), and glutathione peroxidase (Gpx) (Fig. 1k–p). Administration of PG protected cardiomyocytes against apoptosis, inhibited ROS generation, and increased antioxidant capacity in the diabetic hearts (Fig. 1h–p). These results indicated that PG ameliorated cardiac structural and functional abnormalities in DCM.
PG promoted mitochondrial fusion and increased mitochondrial function in diabetic hearts and high-glucose-treated cardiomyocytes
The effects of PG on mitochondrial dynamics were analyzed in the hearts. There were no significant differences in the numbers of mitochondria per μm2 among the experimental groups (Fig. 2a, b), whereas mean mitochondrial size was significantly smaller in the diabetic hearts than in the control hearts (Fig. 2a, c). Moreover, in the diabetic hearts, the percentage of mitochondria whose size was <0.6 μm2 was markedly higher, and the percentage of mitochondria whose size was >1 μm2 was significantly lower than in the control hearts (Fig. 2d). Of the mitochondrial fission and fusion proteins (fission proteins: Drp1 and Fis1; fusion proteins: Mfn1, Mfn2, and Opa1), it was revealed that only Opa1 expression was significantly lower in the diabetic hearts than in the control hearts (Fig. 2f, g), suggesting that Opa1-mediated mitochondrial fusion was inhibited in the diabetic hearts. The mRNA expression of Opa1 was also decreased in the diabetic hearts (Fig. 2h). Treatment with PG restored mitochondrial fusion and Opa1 expression, and reversed the decreases in ATP levels in the diabetic hearts (Fig. 2a–h). Diabetes or PG treatment induced no significant changes in the mRNA expressions of PGC1α, TFAM, and TFB2M and the protein expression of PGC1α in the hearts, suggesting that PG had no significant effects on PGC1α-mediated mitochondrial biogenesis in the diabetic hearts (Fig. 2h–j). The expressions of mtDNA, complexes Ⅰ and V were significantly reduced in the diabetic hearts, while no significant changes were observed in the expression of the other complex proteins (Fig. 2h–j). PG upregulated the expression of mtDNA, complexes I and V in the diabetic hearts (Fig. 2h–j). Moreover, the activities of complexes I, IV, and V were reduced in the diabetic hearts, all of which were enhanced by PG treatment (Fig. 2k).
Previous reports have demonstrated that PG could be directly absorbed from the gastrointestinal tract into the circulation after oral gavage (9, 79). Zou et al. have confirmed the presence of PG in the circulation after oral gavage by using ultraperformance liquid chromatography. It was found that serum PG concentration was detected at micromolar concentrations (79). Therefore, the effects of PG on mitochondrial dynamics and function were further explored in primary cardiomyocytes in vitro. The cells were subjected to normal glucose (5.5 mM, NG) or high-glucose (33 mM, HG) challenge for 48 h with PG at 5, 10, 20, and 50 μM according to the previous studies (48, 72). As shown in Supplementary Figure S1, CCK8 assay showed that HG had no significant effect on cell viability in the cardiomyocytes, whereas PG at 50 μM slightly inhibited cell viability in the cardiomyocytes cultured in both NG and HG medium. After consulting the results of previous studies (48, 72), 10 μM PG treatment was then used to reach potential protection and avoid cell toxicity. As shown in Figure 3a, the mitochondria mainly exhibited elongated tubules with highly interconnected networks in NG-cultured cardiomyocytes. When cultured in HG medium, the mitochondria became smaller and shorter. Compared with NG-cultured cardiomyocytes, HG-cultured cardiomyocytes showed decreased mitochondrial volume and increased number of mitochondria per cell (Fig. 3a, c, d), indicating inhibition of mitochondrial fusion. HG culture led to a significant increase in cellular ROS (green fluorescence) and mitochondrial ROS (red fluorescence) in the cardiomyocytes (Fig. 3b, e, f). The costaining results (merged yellow fluorescence) showed that most cellular ROS have a mitochondrial origin. MDA content was increased, while the activity of T-AOC, total SOD, cellular CuZnSOD, and mitochondrial MnSOD were decreased in HG-cultured cardiomyocytes (Fig. 3g–i). Moreover, HG-treated cardiomyocytes showed increased caspase-3 activity, suppressed mitochondrial respiratory capacity including maximal respiration and spare respiratory capacity, and reduced Opa1 expression (Fig. 3j–n). There were no significant changes in mitochondrial morphology, mitochondrial ROS, and Opa1 expression in the cardiomyocytes cultured in osmotic control (27.5 mM mannitol plus 5.5 mM glucose, the same osmolality as that of HG) compared with those cultured in NG, indicating that osmolality had no significant effects on Opa1 expression and mitochondrial performance (Supplementary Fig. S2). Treatment with PG promoted mitochondrial fusion, reduced caspase-3 activity, improved mitochondrial respiratory capacity, and inhibited oxidative stress by suppressing ROS generation and enhancing the ROS scavenging capacity in both mitochondria and cytoplasm in HG-treated cardiomyocytes (Fig. 3a–l). Consistent with the finding in diabetic hearts in vivo, PG reversed the decrease in Opa1 protein expression in HG-cultured cardiomyocytes (Fig. 3m). PG also attenuated the decrease in Opa1 mRNA expression in HG-cultured cardiomyocytes (Fig. 3n), suggesting that the level of Opa1 may be regulated at the transcriptional level. The above results collectively indicated that PG treatment increased transcription of Opa1, and restored mitochondrial fusion and function in diabetic hearts and HG-treated cardiomyocytes.
PG-induced Opa1 upregulation was responsible for promoting mitochondrial fusion and function and improving cardiac function
It was subsequently explored whether the upregulation of Opa1 expression was responsible for promoting mitochondrial fusion and function after PG treatment. As shown in Figure 4a–e, knockdown of Opa1 with small interfering (si)RNA blocked the mitochondrial fusion-promoting effect and the mitochondrial oxidative stress-inhibitory effect of PG in HG-treated cardiomyocytes. Opa1 siRNA also blunted the mitochondrial respiratory capacity-improving effects of PG (Fig. 4f, g). To further confirm the role of Opa1 in the protective effects of PG on diabetic hearts, adeno-associated virus 9-harboring Opa1 miRNA backbone-based shRNA (AAV-Opa1 shRNA) construct was used to knockdown Opa1 in vivo. Since virus transfection efficiency is relatively low in rat hearts, we used mice hearts for Opa1 gene silencing experiments. As shown in Figure 4h–l, the cardiac function-improving effect and the antioxidative stress effect of PG in diabetic hearts were largely abrogated by the knockdown of Opa1 by transfecting with AAV-Opa1 shRNA. On the contrary, in HG-treated cardiomyocytes, upregulation of Opa1 by transfecting with the adenovirus encoding Opa1 (Ad-Opa1) alone was enough to promote mitochondrial fusion and inhibit mitochondrial ROS generation and improve mitochondrial respiratory capacity (Supplementary Fig. S3). These data indicated that PG protected against HG-induced inhibition of mitochondrial fusion and improved cardiac function in diabetic hearts by upregulating Opa1 expression.
PG-induced upregulation of Opa1 was mediated by Stat3
To identify the signaling pathway involved in the upregulation of Opa1, several specific inhibitors were used to block the potential signaling pathway downstream of PG. As shown in Figure 5a, the cardiomyocytes in HG+PG group were pretreated with the following agents: Bisindolylmaleimide I (Bis, a PKC inhibitor, 10 μM; MedChem Express) (43); Ruxolitinib (Ruxo, a JAK inhibitor, 1 μM; MedChem Express) (32, 65); PD98059 (a MEK inhibitor, 20 μM; MedChem Express) (60); Wortmannin (a PI3K inhibitor, 0.1 μM; MedChem Express) (28, 44); and Stattic (a Stat3 inhibitor, 10 μM; MedChem Express) (45, 69). It was found that PG-induced upregulation of Opa1 in HG-cultured cardiomyocytes was largely abrogated by pretreatment with Stattic, whereas it was unaffected by Bis, Ruxo, PD98059, or Wortmannin, indicating that the activation of Stat3 was involved in PG-induced upregulation of Opa1 expression.
siRNA was used to further confirm that Stat3 participates in regulating Opa1-mediated mitochondrial fusion in response to PG treatment. PG increased the expression of phosphorylated Stat3 in HG-treated cardiomyocytes but not in NG-treated cardiomyocytes (Fig. 5b, c and Supplementary Fig. S4). In NG-treated cardiomyocytes, knockdown of Stat3 resulted in reduced Opa1 expression, inhibited mitochondrial fusion, increased mitochondrial ROS, and suppressed mitochondrial oxidative capacity (Supplementary Fig. S5). In HG-treated cardiomyocytes, knockdown of Stat3 with siRNA blocked the promoting effect of PG on Opa1 protein and mRNA and mitochondrial fusion, indicating that Stat3 modulated the expression of Opa1 at the transcriptional level (Fig. 5b–i). Consistent with the finding in diabetic hearts in vivo, the activities of complexes I, IV, and V were reduced in HG-treated cardiomyocytes, while PG treatment enhanced the activities of complexes Ⅰ, Ⅳ, and V (Supplementary Fig. S6a). Stat3 or Opa1 siRNA blunted the upregulating effects of PG on the activity of complexes I, IV, and V (Supplementary Fig. S6b) and blocked the inhibitory effect of PG on caspase-3 activity (Supplementary Fig. S6c), suggesting that PG protected against HG-induced mitochondrial dysfunction and cardiomyocytes apoptosis via Stat3-Opa1 signaling pathway. Chromatin immunoprecipitation (ChIP) and real-time polymerase chain reaction (PCR) analysis revealed that Stat3 was directly bound to the promoter of Opa1 (Fig. 5j). The Opa1 promoter that bound to Stat3 was significantly decreased under HG condition, which was partly reversed by PG treatment (Fig. 5j). These data imply that Stat3 was directly responsible for the promoting effects of PG on Opa1-mediated mitochondrial fusion.
PG increased Stat3 phosphorylation and promoted Opa1-mediated mitochondrial fusion by directly inhibiting PTP1B
Since Stat3 is known to be activated by phosphorylation at Tyr705, the cardiomyocytes in HG+PG group were additionally pretreated with AG-490 (a tyrosine kinase inhibitor, 20 μM; MedChem Express) (17). Stat3 inhibitor Stattic was used as a positive control. PG-induced increase in Stat3 phosphorylation and Opa1-mediated mitochondrial fusion in HG-cultured cardiomyocytes were largely blunted by Stattic, whereas it was unaffected by AG-490 (Fig. 6a–f). The results suggested that the increase in Stat3 phosphorylation induced by PG was not via activating classical tyrosine kinase. Notably, protein phosphorylation is regulated not only by kinases but also by phosphatases reversely. Therefore, phosphatases but not kinases may be involved in regulating Stat3 tyrosine phosphorylation after PG treatment.
PG is soluble in both water and organic solvents such as ethanol and dimethyl sulfoxide. A recent study has shown that PG was able to permeate several biological membranes (33). It is then speculated that PG may traverse the cell membrane and act directly on some adaptor molecules. To reveal the direct target of PG that increased Stat3 phosphorylation, Mol2 structure of PG was submitted to Pharmmapper selecting Druggable Pharmacophore Models, and 300 potential candidates of 16159 were listed and sorted according to the fit score (Job ID: 200620093452). Based on the results above, we searched for “phosphatase” from the potential candidates listed and two targets were identified, which were “Tyrosine-protein phosphatase nonreceptor type 1” (Rank 1) and “Serine/threonine-protein phosphatase PP1-gamma catalytic subunit” (Rank 124) (Fig. 7a). Since activated Stat3 is dephosphorylated at tyrosine, PTP1B (encoded by “Tyrosine-protein phosphatase nonreceptor type 1”) seemed to be the most likely candidate. The interaction between PG and PTP1B was further investigated using computation docking. Two-dimensional and 3D structure of PG are exhibited in Figure 7b. Data illustrated that PG embedded into the activity pocket of PTP1B (Fig. 7c, d). The maximum binding affinity between PG and the PTP1B was predicted to be −8.2 kcal/mol. In this PG-PTP1B complex, PG was surrounded by many residues such as Arg-221 and Cys-215 in hydrophobic interactions. Moreover, five key hydrogen bond interactions shown as yellow dotted lines were observed between PG and PTP1B (Fig. 7d). The cellular experimental study showed that PG inhibited the activity of PTP1B in both NG-treated and HG-treated cardiomyocytes (Fig. 7e). PG also inhibited the activity of phosphatase in HG-treated cardiomyocytes (Supplementary Fig. S7). In addition, PG inhibited the increase in PTP1B expression in HG-treated cardiomyocytes (Fig. 7f). Furthermore, the animal experimental study validated that PG inhibited PTP1B activity and expression, and increased Stat3 phosphorylation in diabetic hearts (Supplementary Fig. S8).
At last, we determined whether PTP1B was responsible for PG-induced enhancement of Stat3 phosphorylation and Opa1-mediated mitochondrial fusion. It is shown in Figure 7g–n that overexpression of PTP1B with Ad-PTP1B blocked the promoting effect of PG on Stat3 phosphorylation and Opa1-mediated mitochondrial fusion in HG-treated cardiomyocytes. PTP1B overexpression also blocked the inhibitory effect of PG on caspase-3 activity (Fig. 7o). On the contrary, as shown in Figure 8, knockdown of PTP1B with siRNA, which mimicked the effect of PG, increased Stat3 phosphorylation and promoted Opa1-mediated mitochondrial fusion in the cardiomyocytes cultured in HG medium. These results above suggested that PTP1B served as a direct target of PG that regulated Stat3 phosphorylation and Opa1-mediated mitochondrial fusion.
Discussion
In this study, we demonstrate for the first time that PG administration protects against DCM in diabetic rats. Inhibited Opa1-mediated mitochondrial fusion and resultant mitochondrial dysfunction are observed in diabetic hearts and HG-treated cardiomyocytes. PG restores Opa1-mediated mitochondrial fusion and then preserves mitochondrial function in vivo and in vitro, which is responsible for the observed beneficial effects of PG in DCM. Mechanistically, using in vitro and in silico (pharmmapper screening and molecular docking) studies, it is verified that PG embeds into the activity pocket of PTP1B and inactivates PTP1B, which in turn phosphorylate Stat3, which then binds to the promoter of Opa1 and enhances Opa1-mediated mitochondrial fusion (Fig. 9). Our study thus uncovers the detailed molecular mechanism on how PG promotes Opa1-mediated mitochondrial fusion under hyperglycemic condition.
Mitochondrial dysfunction and excessive mitochondrial ROS production have been recognized to be involved in the development of DCM (24). Our previous study has demonstrated that inhibited mitochondrial fusion contributes to mitochondrial dysfunction in DCM (19). Therefore, agents with the ability to promote mitochondrial fusion have great potential in the prevention of mitochondrial ROS and dysfunction in DCM. In this study, the effects of PG administration on mitochondrial fusion/fission dynamics were investigated in the diabetic hearts and HG-treated cardiomyocytes by using transmission electron microscopy (TEM) and MitoTracker Red probe, respectively. This study has found that PG restored mitochondrial fusion and function in vivo and in vitro, suggesting that PG is an active compound to promote mitochondrial fusion. Moreover, PG prevented the development of DCM at the dose of 90 mg/kg/day as evidenced by increased ejection fraction and inhibited myocardial fibrosis and hypertrophy. To the best of our knowledge, our study is the first to identify PG as a novel mitochondrial fusion promoter against hyperglycemia-induced mitochondrial and nonmitochondrial oxidative injury and cardiomyopathy. Our previous study has demonstrated that PG at the dose of 30 mg/kg/day significantly reduced myocardial ischemia/reperfusion injury and improved cardiac function in nondiabetic rats (21). In contrast, there was only a trend toward improved cardiac function, but the differences did not reach statistical significance in the diabetic rats receiving an equivalent dose of PG. This is possibly due to lower bioavailability of PG under diabetic condition, since the metabolism and elimination of many drugs are enhanced in the presence of polyuria in type 1 diabetes (2, 34). Thus, larger dose of PG (90 mg/kg/day) is needed to exert its protective effects in diabetes. To be mentioned, Cerda et al. have evaluated the metabolism of PG in the rats, and found that most of the ingested PG is transformed to nondetected metabolites (9), so it is difficult to evaluate the potential effects of PG metabolites at present. Nevertheless, it is possible that the metabolites of PG may contribute to the beneficial effect in DCM.
Enhanced ROS production in DCM arises not only from mitochondria but also from nonmitochondrial enzymes, including nicotinamide adenine dinucleotide phosphate oxidases, xanthine oxidase, and uncoupled nitric oxide synthase. An intense crosstalk between mitochondrial and nonmitochondrial ROS sources is likely to exist since many studies report that inhibition of mitochondrial ROS inhibits nonmitochondrial ROS and prevents the development of DCM (5, 40, 53). Our study has found that both mitochondrial and cytosol ROS were increased under HG condition, while mitochondria were the major source of ROS in HG-treated cardiomyocytes. PG suppressed HG-induced ROS generation in both mitochondria and cytoplasm. The other pomegranate polyphenols may confer protective effects against DCM similar to PG, since the other polyphenolic compounds and PG have exhibited comparable strong antioxidant activities (64). Pomegranate polyphenols such as punicalin and pedunculagin that have chemical structures analogous to PG may share a similar PTP1B-Stat3-Opa1 signaling mechanism.
Opa1 is located in the inner mitochondrial membrane, which can be proteolytically cleaved to yield long forms (L-Opa1) and short forms of Opa1 (S-Opa1). Song et al. have shown that L-Opa1 and S-Opa1 functionally complement each other, and both of them are necessary for mitochondrial fusion (63). Nevertheless, other groups have reported that L-Opa1 alone is adequate for fusion between mitochondria, whereas S-Opa1 alone does not confer fusion competence and is competent for maintaining mitochondrial oxidative phosphorylation function (36, 42). Although there are some controversies about the relative contributions of the two different isoforms of Opa1, increasing evidence has confirmed that elevating the expression of Opa1 by pharmacological or genetic intervention could improve the function and morphology of mitochondria and attenuate cell damage (31, 67). In contrast, knockdown or knockout of Opa1 induces mitochondrial fragmentation and aggravates myocardial injury (41, 76). In this study, we found that only Opa1 expression was decreased in the diabetic hearts among the fusion proteins (Mfn1, Mfn2, and Opa1). The reason for this may be the deficiency of insulin signaling in the diabetic hearts, since insulin is proved to increase Opa1 protein levels in cardiomyocytes without affecting the other mitochondrial fission and fusion proteins (56). In mitochondria, ROS is mainly produced in the electron transport chain (ETC) that is located in the inner mitochondrial membrane. Opa1-mediated mitochondrial fusion may change structural organization and arrangement of ETC components within the inner mitochondrial membrane, which lead to enhanced ETC activity and inhibited mitochondrial ROS production. Meanwhile, mitochondrial fusion may affect structural and spatial organization of ATP synthase (58). Enhanced activities of ETC and ATP synthase are then coupled with increased ATP production. Moreover, Opa1 is proved to be essential for mtDNA maintenance as it mediates mitochondrial fusion, which results in the exchange of intermitochondrial contents and maintenance of a homogeneous and balanced pool of mitochondrial proteins, including the enzymes needed for mtDNA synthesis (26). Hence, the altered mtDNA in our study may be due to the change of Opa1-mediated mitochondrial fusion. In addition to its effects of mitochondrial fusion promotion, Opa1 has been proved to be involved in regulating the morphology of mitochondrial cristae, and that this facilitates transfer of electrons and enhances mitochondrial respiratory efficiency (15, 54). Thus, the protective effects of Opa1 in DCM are not only through the fusion of mitochondria but also through cristae remodeling.
The protective effects of PG against HG-induced inhibition of mitochondrial fusion and diabetes-induced cardiac dysfunction were blunted by Opa1 knockdown, suggesting that Opa1-mediated mitochondrial fusion is required for the cardioprotection of PG against DCM. Moreover, the mechanisms underlying the regulation of Opa1 were further explored. It was found that both mRNA and protein levels of Opa1 were simultaneously reduced in HG-treated cardiomyocytes, suggesting that the expression of Opa1 is modulated at the level of transcription. Notably, our inhibitor screening and siRNA experiment revealed that Stat3 was responsible for PG-induced transcriptional upregulation of Opa1. In the cardiomyocytes cultured in NG medium, knockdown of Stat3 or Opa1 significantly resulted in inhibited mitochondrial fusion, increased mitochondrial ROS, and suppressed mitochondrial oxidative capacity. These results further support the involvement of Stat3-Opa1 signaling in the regulation of mitochondrial homeostasis. In contrast, in HG-treated cardiomyocytes, knockdown of Stat3 or Opa1 led to a slight trend toward inhibited mitochondrial fusion, but the differences did not reach statistical significance. This is possibly due to the overlapped effects of HG and knockdown of Stat3 or Opa1 on disturbed mitochondrial homeostasis, since the expression of Stat3 or Opa1 was already significantly reduced under HG condition. Stat3-mediated cardioprotective effects are not only via canonical action as a transcription factor but also via nontranscriptional regulation as a modulator of ETC activity (77). Although our study has demonstrated that Stat3 was directly bound to the promoter of Opa1, whether mitochondria-localized Stat3 could regulate Opa1 in the inner mitochondrial membrane is still an interesting question that needs further investigation.
Stat3 is generally phosphorylated by JAKs or receptor tyrosine kinases on Tyr705, and then translocates to the nucleus to induce specific gene expression (59). Unexpectedly, we found that PG-induced Stat3 phosphorylation in HG-cultured cardiomyocytes was unaffected by Ruxolitinib, a JAK inhibitor and AG-490, a tyrosine kinase inhibitor. This is an interesting finding since inhibition of JAK with Ruxolitinib or inhibition of tyrosine kinases with AG-490 efficiently suppresses Stat3 phosphorylation in many cases (7). It seems that PG-induced Stat3 phosphorylation is not via the classical protein kinase pathway, and protein phosphatase may be involved in this process. Our latest pharmmapper screening showed that PTP1B is the only one belonging to the family of protein tyrosine phosphatases and the candidate with the highest fit score. This result is consistent when compared with a previous pharmacophore mapping prediction using the pharmmapper server in 2017 (70). Molecular docking showed that PG interacted with PTP1B via hydrophobic interaction with many amino acids, including Arg-221 and Cys-215, some of which are key residues for the biological activity of PTP1B (4, 61). Concordant with the predicting results obtained from bioinformatics, PG was found to efficiently inhibit the activity of PTP1B and also slightly inhibit the expression of PTP1B.
PTP1B has been shown to play an important role in the development of metabolic diseases associated with insulin resistance, such as diabetes and obesity. The regulation of PTP1B is complex, and occurs at transcriptional and post-translational levels. It has been shown that inflammation increases the transcriptional expression of PTP1B via the activation of NFκB (73). HG enhanced the transcription of PTP1B through PKC-mediated Sp1 activation (35). Hence, HG and elevated inflammation could be important factors responsible for the increased expression and/or activity of PTP1B in DCM. Excessive ROS and reactive nitrogen species (RNS) have been well known to be involved in the development of DCM (11). The role of ROS and RNS in the post-translational regulation of PTP1B activity is different and even opposite. ROS-induced oxidized form of PTP1B is inactive, while RNS-mediated S-nitrosylation of the Cys-215 residue protects PTP1B from subsequent oxidation-induced inactivation (14). Interestingly, Cys-215 is an important interaction site between PG and PTP1B in the molecular docking. It is then speculated that the interaction between PG and PTP1B in Cys-215 residue may affect its S-nitrosylation and help inhibit the activity of PTP1B. Nevertheless, further study is needed to validate this hypothesis.
PTP1B-knockout mice exhibit stronger insulin sensitivity and obesity resistance on a high-fat diet (27). Metabolic disorders and cardiovascular diseases are closely related and often share similar pathways. Recently, emerging evidence reveals that inhibition of PTP1B has protective effects on the cardiovascular system. Genetic deletion or pharmacological inhibition of PTP1B ameliorates cardiac dysfunction induced by myocardial infarction or aging (6, 30). These beneficial effects are largely due to the improvement of endothelial function (66). Intriguingly, our study found that knockdown of PTP1B with siRNA promoted Opa1-mediated mitochondrial fusion in cardiomyocytes in vitro, suggesting that PTP1B inhibition could exert its direct cardioprotective effects besides general metabolic improvement and endothelial function preservation. Therefore, inhibition of PTPIB may be an interesting treatment approach for diabetic patients with cardiac complications via multifarious mechanisms. Nevertheless, it needs to be noted that the PTP family is a very closely related family having a highly conserved catalytic domain among family members. Inhibition of PTP1B may result in the inhibition of other family members (1). Several PTP1B inhibitor drugs have been evaluated in clinical trials, but none of them get FDA approval now due to the low selectivity and undesirable off-target side effects (23). Similar problems may apply to PG. Although toxicity evaluation indicated that oral PG administration to rats at the dose of 4.8 g/kg/day over 37 days did not provoke toxic effects (9), more studies may be needed to evaluate the efficacy and safety of PG as a treatment for diabetic cardiac complications.
It should be noted that our study still has several limitations. First, the effects of PG on DCM were exclusively studied in streptozotocin (STZ)-induced diabetic animals, which may not entirely represent the broad spectrum of diabetes-related changes. Whether PG could protect against DCM in other diabetic conditions needs further investigation. Second, the role of PTP1B in mitochondrial fusion was mainly explored in vitro by using Ad-PTP1B and PTP1B siRNA. We have conceived of administering pharmacological inhibitor of PTP1B to the diabetic animals, but it is hard to exclude that the improvement in cardiac mitochondrial performance is linked to general metabolic improvement or preserved endothelial function as mentioned above. The use of cardiac-specific PTP1B knockout and knock-in animals will be very helpful in further clarifying the role of PTP1B in Opa1-mediated mitochondrial fusion. PTP1B knockout animals are available in previous studies (6, 30), while PTP1B knock-in animals may need to be generated in further study. Despite these limitations, it is believed that our work has provided new important information for revealing the protective effects of PG on DCM and understanding the underlying mechanisms of PG on mitochondrial fusion.
In summary, our study has provided novel findings, indicating that PG protects against DCM and promotes mitochondrial fusion by upregulating Opa1 expression, a process in which PG interacts with PTP1B and inhibits its phosphatase activity, which in turn increases Stat3 phosphorylation and enhances Opa1 expression at the transcriptional level. These findings suggest that PG-modulated mitochondrial fusion might be a new potential option for the treatment of diabetic cardiac complications.
Materials and Methods
Animal experimental design and treatment
All animal experimental procedures were conducted in accordance with the National Institutes of Health guidelines, and the study was approved by the Fourth Military Medical University Ethics Committee. Male 6- to 8-week-old Sprague-Dawley rats were obtained and maintained under controlled temperature, humidity, and light conditions. Diabetes was induced in fasted rats by a single intraperitoneal injection of freshly prepared STZ (65 mg/kg; Sigma-Aldrich, Shanghai, China) dissolved in 0.1 M citrate buffer (pH 4.5) (50). Animals served as the control group received an intraperitoneal injection of vehicle (citrate buffer) only. One week after STZ injection, the rats with fasting glycemia levels ≧11.1 mM were considered diabetic and then administered with the vehicle or PG, respectively, for an additional 12 weeks. PG (purity >85%, the maximum extract concentration available on the market) in animal experiments was obtained by DeSiTe Bio-technology (Chengdu, China). PG was dissolved in sterile saline and was administered to the animals by oral gastric gavage at 30 or 90 mg/kg/day. The low dose of 30 mg/kg/day has been used in our previous study (21) and one other study (71). According to pharmacology guide and some other PG studies (48, 79), low and high doses are usually amplified three- to fourfold. Thus, the dose of 90 mg/kg/day was used as a high dose. Fasting blood glucose and lipids including TC and TG levels were determined using an automatic biochemical analyzer (Chemray 800; Rayto, China).
Measurement of cardiac function
Echocardiography was performed on anesthetized animals at the indicated time point using a VEVO 2100 ultrasound system (Visual Sonics, Toronto, Canada) as described previously (29). For the evaluation, the animals were anesthetized with 2.5% isoflurane. The M-mode images of LVESD and LVEDD were obtained. LVEF was calculated automatically by the system.
Histological analysis
The cross-sectional area of the cardiomyocytes was determined using fluorescein isothiocyanate-labeled WGA staining (Servicebio, Wuhan, China). Myocardial interstitial fibrosis was measured using Masson's trichrome staining (Servicebio). The collagen volume fraction was calculated as the ratio of the area of fibrous tissue to the total area.
Quantification of myocardial apoptosis and ROS production
TUNEL staining (Sigma-Aldrich) and caspase-3 activity assays (Biovision, CA) were utilized for the detection of cardiomyocyte apoptosis as we described previously (37). The apoptosis index was expressed as follows: the number of apoptotic (TUNEL-positive) cells/the total number of nucleated (4′,6-diamino-2-phenylindole-stained) cells × 100%. Intracellular superoxide anion levels were detected using DHE staining (Beyotime Biotechnology, Nantong, China) in fresh heart tissues. The activity of total SOD, cellular CuZnSOD, mitochondrial MnSOD, T-AOC, and the content of Gpx, and malondialdehyde (MDA) were measured to further assess oxidative stress level. All the oxidative stress kits were obtained from Beyotime Biotechnology.
Measurement of mitochondrial morphology
Tissue samples obtained from the anterior wall of LV were fixed with 2.5% glutaraldehyde for 24 h at 4°C and processed as previously described (74). TEM images were obtained using TEM (JEM-1230; JEOL Ltd., Tokyo, Japan) at 300 kV. The number and size of the mitochondria were analyzed using Image-Pro Plus 6.0 software. A minimum of 300 mitochondria from ∼8 images per heart was included in the morphometric analysis as previously described (19).
Mitochondrial complex activities and ATP content
Mitochondrial complex (I–V) activities were determined by complex enzyme activity microplate assay kits (Genmed, Shanghai, China) according to manufacturer's instructions. Intracellular ATP content was determined by using an ATP bioluminescent assay kit (Biovision).
Quantitative real-time PCR
Quantitative real-time PCR was used to assess the relative Opa1 mRNA levels. Total RNA was extracted from heart tissue or cardiomyocytes, and was then reverse transcribed to cDNA with reverse transcriptase. Real-time PCRs were performed on the Bio-Rad Real-Time PCR System. Each reaction contained DNA template, forward and reverse primers, and Power SYBR Green PCR master mix. The levels of Opa1 mRNA were normalized to that of β-actin using the 2−ΔΔCt method. The primer sequences used were as follows: Opa1 forward ATTTCGCTCCTGACCTGGAC, reverse GGTGTACCCGCAGTGAAGAA; PGC1α forward TGGAGTGACATAGAGTGTGCTG, reverse TATGTTCGCGGGCTCATTGT; TFAM forward GGATGAGTCAGCTCAGGGGA, reverse ACGGATGAGATCACTTCGCC; TFB2M forward AGTCAAAGCACTCGGAATCA, reverse TTGCGTTTCCCAAAGCAGTG; mtDNA forward GGTTCTTACTTCAGGGCCATCA, reverse TGATTAGACCCGTTACCATCGA; β-actin forward ACGGTCAGGTCATCACTATCGG, reverse CCACAGGATTCCATACCCAGGAAG.
Cardiomyocyte culture and treatment
Primary neonatal cardiomyocytes were isolated from 1- to 2-day-old neonatal Sprague-Dawley rats. The cardiomyocytes were subjected to NG (5.5 mM) or HG (33 mM) challenge for 48 h with the vehicle or various concentrations of PG supplement. PG (Catalog No: MB6504, purity >98%) in cellular experiments were purchased from Dalian Meilun Biology Technology (Dalian, China).
Assessment of mitochondrial morphology and ROS in cardiomyocytes
The mitochondria in primary cardiomyocytes were stained with MitoTrackerRed CMXRos probe (100 mM; Invitrogen, Carlsbad) and captured with a Nikon A1R MP+ confocal laser-scanning microscope. The number and individual volume of each observed object (mitochondria) were analyzed and quantified as described previously (20). Cellular ROS and mitochondrial ROS levels in cardiomyocytes were assessed using DCFH-DA staining (Beyotime Biotechnology) and MitoSOX staining (Invitrogen).
Measurement of the mitochondrial oxygen consumption rate
The oxygen consumption rate (OCR) was recorded by using an XF24 Extracellular Flux Analyser (Agilent SeaHorse Bioscience) as described previously (12). The cells were sequentially stimulated with 1 μM oligomycin, 0.5 μM FCCP [Carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone], and 1 μM antimycin A to calculate ATP-related OCR, maximal OCR, and nonmitochondrial OCR, respectively. All values of OCR were normalized to the cell counts determined by a well-by-well hemocytometer.
siRNA transfection
The cardiomyocytes were isolated and transfected with Opa1 siRNA (sense: CCAGCAAGGUUAGCUGCAATT; antisense: UUGCAGCUAACCUUGCUGGTT), STAT3 siRNA (#sc-270027; Santa Cruz Biotechnology), PTP1B siRNA (#sc-270525; Santa Cruz Biotechnology), or negative control siRNA (#sc-37007; Santa Cruz Biotechnology) using Lipofectamine RNAiMAX reagent (Invitrogen) according to the manufacturer's instructions. After 48 h of transfection, the cells were subjected to HG with vehicle or PG for another 48 h.
Adenoviral transfection
The cardiomyocytes were isolated and transfected with adenoviruses harboring an empty vector, PTP1B or Opa1 (Hanbio Technology, Shanghai, China). Adenoviral titers used in this study were ∼1.2 × 1010 PFU/mL, and the infection was performed at a multiplicity of infection of 20 as described previously (78). After 48 h of adenoviral transfection, the cardiomyocytes were then incubated with NG or HG for another 48 h.
Knockdown of Opa1 by adeno-associated virus transfection in vivo
Adeno-associated virus 9 (AAV9)-harboring Opa1 miRNA backbone-based shRNA (AAV-Opa1 shRNA) and control shRNA virus (AAV-con shRNA) were constructed by Hanbio Co., Ltd. (Shanghai, China). Male mice were anesthetized using inhaled 2% isoflurane. AAV vectors were diluted to 2.5 × 1011 particles per milliliter and delivered into the free wall of the left ventricle using a 30.5G Hamilton syringe as previously reported (75). One week after the transfection, the mice were injected with STZ (50 mg/kg/day) intraperitoneally for five consecutive days to establish diabetes model as previously described (18), and then administered with the vehicle or PG (90 mg/kg, once daily, orally by gastric gavage), respectively, for another 12 weeks.
Western blot analysis
Total protein was extracted from rat hearts and cardiomyocytes by using RIPA buffer containing protease inhibitor cocktail. Standard Western blotting method was used as described previously (37). The primary antibodies used in this experiment were as follows: Drp1 (#14647; Cell Signaling Technology), Fis1 (#GTX111010; GeneTex), Mfn1 (#13798-1-AP; Proteintech), Mfn2 (#9482; Cell Signaling Technology), Opa1 (#612606; BD Biosciences), PGC1α (#66369-1-Ig; Proteintech), CⅠ-NDUFB8 (#14794-1-AP; Proteintech), CⅡ-SDHB (#10620-1-AP; Proteintech), CⅢ-QUCRC2 (#14742-1-AP; Proteintech), CⅣ-MTCO2 (#55070-1-AP; Proteintech), CⅤ-ATP5A1 (#14676-1-AP; Proteintech), phosphorylated Stat3 Tyr705 (p-Stat3, #9145; Cell Signaling Technology), Stat3 (#9139; Cell Signaling Technology), PTP1B (#ab244207; Abcam), and β-tubulin (#66240-1-Ig; Proteintech). The secondary antibodies used were horseradish peroxidase-conjugated Goat antirabbit secondary antibody (A0208; Beyotime Biotechnology) and horseradish peroxidase-labeled Goat antimouse secondary antibody (A0216; Beyotime Biotechnology).
Chromatin immunoprecipitation analysis
Chromatin immunoprecipitation (ChIP) was conducted using the SimpleChIP Plus Enzymatic Chromatin IP kit (#9005; Cell Signaling Technology) according to the manufacturer's instructions as described previously (13). In brief, the primary cardiomyocytes were crosslinked with 1% formaldehyde and subsequently homogenized in lysis buffer. The enzymatic sheared chromatins were incubated with CHIP-grade antibodies against Stat3 (Cell Signaling) and protein G magnetic beads. DNA was released from the precipitation and then subjected to real-time PCR analysis. The sequences of the Rat Opa1 promoter region primers were as follows: forward: 5′-TCCAGTTAGGTTTTGGGCCTT-3′, reverse: 5′-TCCTTTTATGAGCCCCAATTTCCTT-3′. IgG was used as the negative control.
Prediction of potential targets of PG
The molecular formula of PG was listed at the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP,
Molecular docking
Autodock Vina 1.1.2 was employed to explore the binding affinity as well as binding sites between PG and PTP1B. In brief, the Mol2 file of PG was downloaded from TCMSP (
Statistical analysis
All the data were presented as the means ± standard error of the means. Data were analyzed using one-way ANOVA followed by Bonferroni's post hoc test using GraphPad Prism 6.0 software. Differences with p values <0.05 were deemed statistically significant.
Footnotes
Authors' Contributions
M.D. and J.P. conceived and designed the study. F.F., R.S., and Y.D. performed the animal experiments. C.L., M.L., M.Z., and J.L. carried out the cellular experiments. F.F., C.L., M.L., Y.D., and Q.W. performed the molecular biology experiments. F.F. and C.L. analyzed the data. F.F. drafted the article. M.D., J.P., and G.W. revised and edited the article. All authors have read and approved the final article.
Author Disclosure Statement
No competing financial interests exist.
Funding Information
This study was supported by the grants from National Natural Science Foundation of China (No. 81970316, No. 82070387, No. 81670354, No. 82070051), Innovation Capability Support Program of Shaanxi (No. 2019KJXX-084), and the Fundamental Research Funds for the Central Universities (No. xzy012019115).
Supplementary Material
Supplementary Figure S1
Supplementary Figure S2
Supplementary Figure S3
Supplementary Figure S4
Supplementary Figure S5
Supplementary Figure S6
Supplementary Figure S7
Supplementary Figure S8
Abbreviations Used
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
