Introduction
Diabetes mellitus is one of the worldwide leading causes of death and long-term disability, resulting in huge social and economic burden. Vascular abnormalities are the major contributor to the progression of diabetes and its associated complications [
1]. Because the endothelium is an important component of vascular homeostasis and the primary target of hyperglycemia and hyperlipidemia, endothelial dysfunction occurs in both animal models of diabetes and diabetic patients [
2‐
5] and it has been implicated in diabetic vascular complications [
1,
2]. Although the pathophysiology of diabetic endothelial dysfunction is incompletely characterized, it appears to be multifactorial. Amongst various proposed mechanisms, oxidative damage induced by reactive oxygen species (ROS) has been critical in this disorder [
6]. Exposure of endothelial cells to high glucose induces ROS production
in vitro[
7,
8]. Clinical and animal studies also demonstrate an increase in vascular ROS formation in diabetes [
9,
10]. Importantly, treatment with antioxidants improves endothelial-dependent vascular relaxation in animal models of diabetes [
11], supporting a central role of ROS in the development of diabetic endothelial dysfunction. Multiple mechanisms have been proposed to be responsible for ROS production in diabetes. In addition to enhanced glucose auto-oxidation, increased substrate flux through the polyol pathway and stimulation of eicosanoid metabolism, ROS is mainly produced by mitochondria, NADPH oxidase, xanthine oxidase and un-coupled nitric oxide (NO) synthase (NOS) [
12‐
14]. Both xanthine oxidase and NADPH oxidase have been reported to induce ROS production in diabetic vessels, which significantly contributes to endothelial dysfunction [
15‐
17]. Hyperglycemia-induced mitochondrial respiratory chain deficiencies are postulated to be another critical and unifying source of ROS generation [
15]. Mitochondria-derived superoxide production may be the initiator for a vicious cycle of oxidative stress in diabetes [
18]. Excessive ROS production also induces a dysfunctional eNOS, or referred as eNOS uncoupling, which generates superoxide instead of NO [
3,
15]. Increased production of superoxide in endothelial cells reacts directly with NO to form a more harmful molecule peroxynitrite (ONOOˉ), thereby reducing NO bioavailability [
2,
3]. Elevated ROS production and reduced NO bioavailability, together with the intermediate product peroxynitrite, significantly account for apoptosis in endothelial cells and endothelial dysfunction in diabetes [
19,
20]. However, the regulation of ROS generation has not been fully addressed in diabetes.
Calpains belong to a family of calcium-dependent thiol-proteases. They have been involved in a wide variety of cellular processes including remodeling of cytoskeletal, caspases activation/apoptosis and acute inflammation [
21]. Two major isoforms of calpain, calpain-1 and calpain-2, are ubiquitously expressed, and calpastatin is an endogenous inhibitor for calpain-1 and calpain-2 [
21]. Over-expression of calpastatin is shown to inhibit calpain activity
in vitro and in transgenic mice [
22]. Calpain activity is increased in endothelial cells under diabetic conditions [
23,
24]. An early study showed that inhibition of calpain increased NO production from eNOS and reduced leukocyte-endothelium interactions in microcirculation during hyperglycemia [
25]. These effects of calpain inhibition were further confirmed in a genetic rat model of type 2 diabetes [
24,
26]. It has been also suggested that calpain activation contributes to microvascular albumin leakage in diabetes [
26]. Nevertheless, the role of calpain in diabetic vascular complications has not been fully characterized. Particularly, the functional significance of calpain remains to be determined and whether calpain plays a role in regulating ROS production has never been reported in diabetic endothelial dysfunction.
In the present study, we employed an in vitro model of endothelial cells stimulated with high glucose and multiple in vivo models of diabetes to investigate the role of calpain in ROS generation and endothelial dysfunction.
Discussion
The current study investigated the role of calpain in endothelial ROS production and endothelial dysfunction during diabetes. Herein, we demonstrate for the first time that inhibition of calpain reduces mitochondrial superoxide generation, intracellular ROS production and apoptotic cell death in high glucose-stimulated endothelial cells. The effects of calpain inhibition correlate with an elevation of NO production and selectively scavenging mitochondrial ROS increases NO production during high glucose stimulation. In mouse models of type-1 and type-2 diabetes, transgenic over-expression of calpastatin reduces vascular ROS production, inhibits peroxynitrite formation, and attenuates the dysfunction of endothelium-dependent relaxation. Thus, this study reveals a novel role of calpain activation in endothelial ROS generation under diabetic conditions. Excessive ROS possibly produced by mitochondria quenches NO, thereby generating toxic oxidant species peroxynitrite and reducing NO bio-availability and thus, induces endothelial dysfunction in diabetes.
Studies have demonstrated that high glucose or diabetes increases cytosolic Ca
2+ concentration and thus induces calpain activation in cultured endothelial cells and in micro/macro-vascular tissues of type-1 and type-2 diabetes [
23,
24,
38]. To further confirm these previous findings, the present study also revealed that calpain activity was markedly increased in HUVECs stimulated with high glucose and in aortas of both STZ-induced type-1 diabetic mice and transgenic type-1 diabetic mice (OVE26 mice). The up-regulation of calpain activity in endothelial cells has been implicated in vascular inflammation and endothelial leakage in diabetes [
24,
26]. The present study provides additional functional evidence demonstrating that calpain activation contributes to endothelium-dependent dysfunction in diabetes. In the present study, different mouse models of diabetes including STZ-induced, genetically modified type-1 and db/db type-2 diabetes were employed. Endothelium-dependent vasodilation was impaired in all diabetic mice, confirming previous findings [
9‐
11]. However, transgenic over-expression of calpastatin rescued the endothelium-dependent vascular relaxation in response to Ach in diabetic aortic rings. There was no alteration of vascular relaxation in response to NO donor SNP in both diabetic mice and transgenic mice with calpastatin over-expression as compared to non-diabetic and wild-type mice, respectively, suggesting that the function of smooth muscle relaxation in aortic rings of diabetic mice is preserved. It has been known that the release of prostacyclin, in particular PGI
2, also induces endothelium-dependent relaxation in large arteries [
39,
40]. To exclude the involvement of those NO independent factors in calpain-mediated dysfunction of endothelium-dependent relaxation, we demonstrated that deletion of eNOS abrogated the beneficial effect of calpastatin over-expression on endothelium-dependent relaxation in diabetic aortic rings. Thus, it is highly possible that the improvement of vascular relaxation in response to Ach in diabetic calpastatin transgenic mice results from an increase in NO bioavailability. It is worthwhile to mention that a previous study has demonstrated that the predominant agonist-induced endothelium-dependent vasodilation is mediated by endothelium-derived hyperpolarizing factor (EDHF), not by NO in murine resistance vessels [
41]. In the small arteries of diabetic rats, it has been shown that that NO-dependent vasorelaxation is preserved, whereas EDHF-dependent response is impaired [
42]. Thus, it seems that big arteries (such as aortas) and small arteries may respond differently to diabetes in terms of vascular relaxation. Whether calpastatin over-expression could also improve EDHF-dependent response in diabetic resistance arteries needs future investigation for clarification. In addition, calpain has been shown to positively regulate eNOS activation and NO production in endothelial cells in response to VEGF [
43], suggesting that the role of calpain in NO production may be dependent on distinct stimuli.
An important finding is that calpain activation mediates ROS production in vasculatures of diabetes. Intriguingly, inhibition of calpain attenuated mitochondrial superoxide generation in high glucose-stimulated endothelial cells, suggesting that calpain may play a role in mitochondrial ROS generation in diabetes. Recent findings suggest that over-generation of ROS through mitochondrial electron transport chain contributes to the diabetic vascular injury [
14,
18]. Uncoupling protein 2, a critical regulator of mitochondrial-derived ROS release, has been shown to attenuate the endothelial dysfunction by increasing NO bioavailability and inhibiting ROS production in diabetic mice [
14,
44]. In addition to the electron transport system, mitochondria also produce ROS through monoamine oxidase-dependent pathway in diabetes [
13]. However, it is currently unknown how calpain activation contributes to mitochondrial ROS generation in endothelial cells. Study has shown that hyperhomocysteinemia induces the translocation of active calpain-1 from cytosol to mitochondria, leading to increased intramitochondrial oxidative stress in cultured rat heart microvascular endothelial cells [
45]. In this regard, some important mitochondrial proteins have been identified as substrates of calpain-1, such as ATP5A1 [
46], optic atrophy-1 [
47], apoptosis-inducing factor [
48], etc. Disruption of these mitochondrial proteins may induce mitochondrial dysfunction and excessive ROS generation. Nevertheless, further investigations are needed to determine the mechanisms by which calpain induces mitochondrial ROS generation in endothelial cells. Since there may be cross-talks between main ROS sources (mitochondria, NADPH oxidase, xanthine oxidase or un-coupled NOS) [
49], it is very hard to determine the relative importance of individual ROS source in diabetes.
There is a general consensus that increased ROS production in the vascular wall, particularly within endothelial cells, contributes largely to the diabetic endothelial injury [
1,
2]. In addition to its role in promoting vascular inflammation [
50,
51], excessive ROS production can induce apoptotic cell death in endothelial cells [
8,
20]. In support of this view, we showed that inhibition of calpain prevented high glucose-induced apoptosis in endothelial cells. ROS, in particular superoxide anion reacts very rapidly and efficiently with NO to generate the extremely detrimental species peroxynitrite, reducing the bio-availability of NO production [
19]. In this regard, we found that NO production was reduced in high glucose-stimulated endothelial cells and peroxynitrite formation was increased in vascular walls of diabetic mice. This is also supported by previous reports [
7‐
9,
52]. Importantly, inhibition of calpain elevated NO production in endothelial cells and attenuated the formation of peroxynitrite in aortas of diabetic mice. These data suggest that calpain activation mediates ROS generation, which in turn quenches NO, and thereby reducing its bio-availability in endothelial cells during diabetes. It is important to mention that a reduction in NO production may also result from eNOS dysfunction [
3]. However, the present study found that incubation with high glucose did not change the protein levels of eNOS and phosphorylated eNOS, and the formation of eNOS dimers in endothelial cells, indicating that calpain may not directly disrupt the function of eNOS in producing NO. Indeed, previous reports have demonstrated that diabetes reduced NO bio-availability without altering eNOS protein and its dimer formation in endothelial cells [
53,
54]. Further evidence in support of our conclusion was that calpain cleaved eNOS protein without affecting the eNOS activity [
43,
55]. However, our current and these previous findings are different from a recent report which showed that calpain activation may induce the disruption of eNOS in producing NO during oxidized LDL stimulation [
56]. This discrepancy may be due to different stimuli: hyperglycemia versus oxidized LDL. Therefore, we conclude that inhibition of calpain increases NO bio-availability, at least in part, by reducing ROS formation and improving eNOS function in endothelial cells under pathological conditions.
While the present study investigated the role of calpain in ROS generation in endothelial dysfunction during diabetes, it is important to mention that multiple mechanisms may be involved in calpain activation-mediated diabetic vascular complications. Calpain has been shown to target and cleave IκB, and activate NF-κB signaling, leading to inflammatory responses and apoptosis in vasculatures of diabetes [
25]. Calpain may also mediate apoptosis in endothelial cells by directly targeting Bid and apoptosis inducible factor [
57]. Thus, further studies will be needed to fully address the role of calpain in diabetic vascular complications.
In conclusion, we have demonstrated an important role of calpain in endothelial ROS production during hyperglycemia/diabetes, which is associated with apoptosis and a reduction in NO generation in endothelial cells. Genetic inhibition of calpain through over-expression of calpastatin reduces vascular ROS production and peroxynitrite formation, and improves endothelium-dependent relaxation in diabetic mice. Given that calpain has been implicated in vascular inflammation [
24,
25] and endothelial leakage [
26], this study provides further evidence to support the view that calpain may serve as a potential therapeutic target for diabetic cardiovascular complications.
Competing interest
The authors declare that they have no competing interest.
Authors’ contributions
BC, FT and IC carried out in vitro studies. QZ measured calpain activity and ROS formation in aortas. RN measured mitochondrial superoxide generation in HUVECs. BC and LS assessed vascular relaxation and analyzed the data. TP designed the study. GC, WW and PS contributed to research materials. BC, GC, WW, PS and TP wrote the manuscript. All authors read and approved the final manuscript.