The Calcium-Dependent Protease Calpain Causes Endothelial Dysfunction in Type 2 Diabetes

This study was performed in accordance with the National Institutes of Health and Thomas Jefferson University institutional animal care and use committee guidelines for the use of experimental animals. We used ZDF rats (Charles River Laboratories, Noblesville, IN) of 10–14 weeks of age, the age range at which they develop overt diabetes with hypoinsulinemia and permanent hyperglycemia (12). ZDF rats have a homozygous mutation in their leptin receptor gene (fa/fa) that causes them to develop a metabolic disorder similar to type 2 diabetes in humans (12). Zucker lean (ZL) rats are either homozygous wild-type or heterozygous at the leptin receptor gene (+/?), do not develop diabetes, and are used as control animals because of the common genetic background. Rats were divided into one of three groups: 1) control ZL rats injected with vehicle, 2) ZDF rats injected with vehicle, and 3) ZDF rats injected with 27 μg/kg i.p. calpain inhibitor benzyloxycarbonyl-leucyl-leucinal (ZLLal) once a day for 5 consecutive days. The microcirculation of the mesentery was used for these studies because it is highly amenable to intravital microscopy and because it expresses a typical inflammatory phenotype in response to both hyperglycemia (13) and diabetes (14). In a preliminary study, administration of ZLLal 60 min before intravital microscopy, as well as 120 min local superfusion of ZLLal during intravital microscopy experiments, failed to inhibit leukocyte-endothelium interaction in the diabetic microcirculation (data not shown).

Mesenteric segments with dense microvascular networks were dissected under microscopy. Sections were snap-frozen in liquid nitrogen and homogenized as previously described (15). Calpain activity was studied by Western blot analysis using polyclonal antibodies against either the NH₂ terminus domain (RP1 calpain-1 or RP2 calpain-2) or domain IV (RP3 calpain-1 and RP3 calpain-2; Triple Point Biologics, Portland, OR) of the large subunit of m- and μ-calpain, respectively. Once activated, m- and μ-calpains autolytically cleave their NH₂-terminal ends, respectively, resulting in the loss of NH₂ terminus antibody recognition, which can be used as a measure of calpain activation. Quantification of the stable domain IV was used to measure total m- and μ-calpain content in mesenteric extracts. Proteins were detected by chemiluminescence (Supersignal West Pico; Pierce, Rockford, IL) and quantified by laser densitometry (personal densitometer; Molecular Dynamics, Piscataway, NJ).

Rats were anesthetized with sodium pentobarbital (60 mg/kg i.p.) and prepared for intravital microscopy as previously described (13). Leukocyte rolling, adhesion, and transmigration were studied in mesenteric postcapillary venules. Leukocyte rolling is defined as the number of leukocytes rolling past a fixed point per minute; leukocyte adherence is defined as the number of leukocytes firmly adhered to a 100-μm length of endothelium for at least 30 s. The number of leukocytes extravasated within a 5-μm perivascular area were counted and normalized with respect to the area. Erythrocyte velocity was determined online using an optical Doppler velocimeter obtained from the Microcirculation Research Institute (College Station, TX). Erythrocyte velocity (V_rbc) and venular diameter (D) were used to calculate the venular wall shear rate (g) using the formula: g = 8(V_mean/D), where V_mean = V_rbc/1.6.

At the completion of intravital microscopy experiments, sections of mesentery and ileum were fixed in vivo, dehydrated using graded acetone washes, and embedded in plastic as previously described (13). Immunohistochemical localization of vascular cell adhesion molecule-1 (VCAM-1) and intracellular adhesion molecule-1 (ICAM-1) was accomplished using monoclonal antibodies (BD Transduction Laboratories, San Diego, CA) and the avidin/biotin immunoperoxidase technique. The percentage of positively staining venules was determined by examining 5–10 tissue sections per rat, with 50 venules analyzed per tissue section.

NO levels in mesenteric postcapillary venules were measured during intravital microscopy experiments, using an ISO-NO microelectrode (100-nm diameter at the tip) connected to an ISO-NO Mark-II NO meter (World Precision Instruments, Sarasota, FL), according to a previously described technique (13,16). Briefly, the microelectrode was positioned directly over a mesenteric postcapillary venule at a distance of <5 μm from the venular wall. To calculate the amount of NO released by the vascular endothelium, background currents obtained at a distance of 0.5–1 cm from the venular wall were subtracted from currents measured at the venular wall. Endothelial NO values were normalized to the size of the blood vessel studied and reported as nanomolar NO/1,000 μm² vessel area.

The association of endothelial NO synthase (eNOS) and heat shock protein 90 (hsp90) in the mesentery of all experimental groups of rats was studied by immunoprecipitation and Western blotting according to a technique described previously (15). Briefly, mesentery extracts containing ∼750 μg protein were incubated with anti-eNOS monoclonal antibody in immunoprecipitation buffer for 1 h followed by incubation with prewashed protein G-agarose (Sigma) for 2 h. The resulting pellet was washed three times in PBS, boiled in SDS sample buffer, and resolved by SDS-PAGE. Immunoblot analysis was performed using primary antibodies against eNOS (BD Transduction) and hsp90 (Santa Cruz Biotechnology, Santa Cruz, Ca).

Densely vascularized segments of mesentery were isolated. Mesenteric microvascular endothelial cells (MMECs) were then isolated using magnetic microbeads (Miltenyi Biotech, Auburn, CA) ligated with anti-rat platelet endothelial cell adhesion molecule-1 (PECAM-1) antibody (17). MMECs were phenotyped by fluorescence-activated cell sorting analysis of PECAM-1 and by functional uptake of acetylated LDL. For fluorescence-activated cell sorting studies, MMECs were incubated with monoclonal CD31 antibody conjugated with phycoerythrin (BD Pharmingen) for 30 min at 4°C and analyzed using a Becton Dickinson flow cytometer (17). Uptake of labeled acetylated LDL (10 μg/ml; Biomedical Technologies, Stoughton, MA) by MMECs was visualized by fluorescent microscopy.

Rat MMECs were cultured at 37°C with 5% CO₂ in Dulbecco’s modified Eagle’s medium supplemented with 150 μg/ml endothelial cell growth supplement (Sigma, St. Louis, MO) on gelatin-coated tissue culture dishes. A twofold approach was used to minimize changes in the diabetic phenotype of MMECs. First, we only used primary or second-passage MMEC cultures. Second, glucose concentrations in the culture medium were adjusted to match the average glucose levels detected in the plasma of donor ZDF rats (Table 1). Thus, MMECs from ZL rats were cultured in 5.5 mmol/l d-glucose (normal glucose) medium, and MMECs from ZDF rats were cultured in 20 mmol/l d-glucose. In additional control experiments, MMECs from nondiabetic ZL rats were incubated with 20 mmol/l d-glucose.

Calpain activity in intact MMECs was determined with the highly sensitive fluorescent calpain substrate t-BOC-Leu-Met-CMAC (Molecular Probes, Eugene, OR), as described previously (18). Mean fluorescence signals from MMECs grown to confluence in a 24-well plate were measured using a CytoFluor plate reader (excitation λ = 360 nm, emission λ = 460 nm). Specificity of the t-BOC for calpain was confirmed by both pharmacological inhibition and antisense depletion of calpain activity. Sense (ATGCTTCTCGGGCACAAT) and antisense (ATTGTGCCCGAGAAGCAT) thioated μ-calpain oligonucleotides were obtained from MWG (High Point, NC).

The calpain inhibitor ZLLal (Biomol Research Laboratories, Plymouth Meeting, PA) was used because, in the concentration range used in this study, it has highest selectivity for calpains than for other proteosomal enzymes (19). ZLLal was dissolved in a final ethanol concentration of 0.01%. In control experiments, intraperitoneal injection of saline (0.9% NaCl) containing 0.01% ethanol in rats did not affect baseline leukocyte rolling, adherence, or transmigration (data not shown).

Data are presented as the means ± SE and compared by ANOVA with post hoc analysis by Fisher’s corrected t test. Probabilities of ≤0.05 were considered statistically significant.

Compared with nondiabetic ZL rats, diabetic ZDF rats displayed fasting hyperglycemia that was not lowered by calpain treatment (Table 1). Mean arterial blood pressures were not significantly different among the three groups of rats studied (Table 1). However, venular shear rate values in postcapillary venules of ZDF rat mesenteries were significantly decreased, despite remaining within physiological range (Table 1). This finding is in agreement with previous studies demonstrating decreased shear rates (14) and increased vascular permeability (20) in the microcirculation of diabetic rats. Thus, 10- to 14-week-old ZDF rats develop overt hyperglycemia and hemodynamic perturbations in the microcirculation that are highly relevant to diabetic vascular disease.

To investigate the effect of diabetes on calpain activity in vivo, we measured autoproteolysis of m- and μ-calpain, a well-established indicator of calpain proteolytic activity in biological systems (21). On activation, calpains undergo autoproteolysis with removal of the NH₂ 9–14 amino acids from the large 80-kDa subunit (22,23). Accordingly, evaluation of m- and μ-calpain autoproteolysis in freshly isolated vascular segments of mesenteric tissue from ZL and ZDF rats was assessed by immunoblot analysis, using primary antibodies to the NH₂ terminus domains of the m- and μ-calpain large subunits (Fig. 1A). We found increased proteolytic activity of μ-calpain in ZDF rats, as demonstrated by 75% loss in NH₂ terminus antibody binding (Fig. 1). Administration of 27 μg/kg i.p. of the calpain inhibitor ZLLal to ZDF rats for 5 consecutive days completely blocked μ-calpain autoproteolysis in the rat mesentery (Fig. 1). Total μ-calpain content was quantified using a primary antibody against domain IV of the large subunit, which recognizes both unautolyzed and autolyzed μ-calpain. No significant differences in total μ-calpain protein expression were found in ZDF rats (Fig. 1A).

In contrast, the proteolytic activity of m-calpain was not increased in ZDF rat mesenteries, as demonstrated by a lack of changes in the m-calpain NH₂ binding domain expression level (Fig. 1). No significant changes were observed in total m-calpain expression, as detected using a primary antibody against m-calpain domain IV (data not shown). These findings demonstrate that type 2 diabetes selectively increases the proteolytic activity of μ-calpain in densely vascularized tissues.

Nondiabetic ZL rats exhibited a low baseline leukocyte rolling value of 24.1 ± 2.4 cells/min in mesenteric postcapillary venules (Fig. 2A). In contrast, ZDF rats had a threefold increase in leukocyte rolling (P < 0.05 vs. ZL). Injection of 27 μg/kg i.p. ZLLal once a day for 5 consecutive days to ZDF rats attenuated leukocyte rolling to 47.8 ± 10.0 cells/min, a value that did not reach statistical significance.

There were a low number of leukocytes firmly adhering to the vascular endothelium of mesenteric postcapillary venules in ZL rats (Fig. 2B). Leukocyte adherence was increased fivefold in ZDF rats (P < 0.01 vs. ZL rats). Admin-istration of 27 μg/kg i.p. ZLLal once a day for 5 days significantly attenuated the number of adherent leukocytes in mesenteric postcapillary venules of ZDF rats (Fig. 2B).

The number of leukocytes that had transmigrated into the perivascular area was also studied. Few leukocytes were found in the perivascular area of postcapillary venules in nondiabetic ZL rats (Fig. 2C). In contrast, increased numbers of transmigrated leukocytes were observed within 5 μm of the postcapillary venular wall in ZDF rats (P < 0.01 vs. ZL rats). Inhibition of calpain activity with ZLLal greatly reduced the number of leukocytes extravasated in the mesentery of ZDF rats (Fig. 2C).

These data demonstrate that the diabetic microcirculation of ZDF rats experiences chronic inflammatory signals with increased leukocyte trafficking and that inhibition of calpain activity acutely attenuates vascular inflammation in the face of chronic hyperglycemia.

To investigate the molecular mechanisms of increased leukocyte trafficking interactions in the diabetic microcirculation, we studied endothelial cell surface expression of ICAM-1 and VCAM-1 (Fig. 3), two cell adhesion molecules relevant to vascular inflammation and atherogenesis. Compared with nondiabetic ZL rats, expression of ICAM-1 increased 2.5-fold in the vascular endothelium of ZDF rats (P < 0.01), and it was attenuated to control values by ZLLaL treatment (P < 0.01 vs. untreated ZDF rats). Similarly, we found evidence of a 2.8-fold increase in VCAM-1 expression in the vascular endothelium of ZDF rats. Treatment of ZDF rats with the calpain inhibitor ZLLal attenuated VCAM-1 expression levels to control values (P < 0.01 vs. untreated ZDF rats). These data demonstrate that increased calpain activity plays a role in the upregulation of endothelial cell adhesion molecules in diabetes.

Endothelial NO suppresses VCAM-1 expression at the molecular level (24). Accordingly, we measured levels of endothelial NO in mesenteric postcapillary venules to further study the molecular mechanisms of vascular dysfunction in diabetes. We found that levels of NO decreased by 2.4-fold in the inflamed postcapillary venules of ZDF rats (Fig. 6). Inhibition of calpain activity with ZLLal significantly restored endothelial NO levels in mesenteric venules of ZDF rats (P < 0.05 vs. untreated ZDF rats) (Fig. 4). Thus, loss of endothelial NO in the diabetic vasculature is at least in part calpain dependent.

Previous studies have indicated that increased calpain activity impairs NO production by decreasing the association of the regulatory protein hsp90 with the eNOS complex (25). Accordingly, we measured total hsp90 content in the mesentery of ZL and ZDF rats using Western blot analyses. We also performed immunoprecipitation studies to determine the extent of eNOS/hsp90 association in vascular segments of the ZDF rat mesentery. Densitometric analysis revealed a 32 ± 1.7% reduction in hsp90 expression levels in ZDF rat mesenteries, compared with nondiabetic ZL rats (P < 0.05). Furthermore, less hsp90 was associated with eNOS in the vasculature of ZDF rats as compared with ZL rats (Fig. 5). Inhibition of calpain activity significantly attenuated the loss of hsp90 from the eNOS complex (Fig. 5). These data demonstrate that increased calpain activity impairs the posttranslational assembly of eNOS in the diabetic vasculature, and they also correlate with the evidence of calpain-dependent loss of NO illustrated in Fig. 4.

Measuring of calpain activity in mesenteric tissue homogenates does not provide absolute evidence of increased calpain activity in the vascular endothelium. We isolated MMECs from ZL and ZDF rats to further confirm that diabetes does increase calpain activity in the endothelium of the microcirculation. Microvascular endothelial cells were phenotyped by uptake of acetylated LDL and flow cytometry analyses of PECAM-1 (Fig. 6). Calpain activity in normal and diabetic MMECs was measured with the use of the fluorescent probe t-BOC (Fig. 7 upper panel). We detected a 70% increase in basal calpain activity in MMECs isolated from ZDF rats, compared with MMECs from nondiabetic ZL rats (Fig. 7, lower panel). In additional control experiments, a comparable degree of calpain activation in the range of 1.6 ± 0.19-fold change was detected in nondiabetic MMECs incubated with 20 mmol/l d-glucose for 48 h (P < 0.05 vs. nondiabetic MMECs cultured under normal glucose conditions). This result points to hyperglycemia as the leading cause of increased calpain activity in the diabetic vasculature.

Treatment of diabetic MMECs with 10 μmol/l antisense oligonucleotides to μ-calpain for 48 h attenuated calpain activity to values found in nondiabetic control MMECs (Fig. 7). A comparable degree of calpain activity attenuation was also obtained after treatment of diabetic MMECs with ZLLal for 24 h (Fig. 7), a result that strengthens the significance of our in vivo data. Taken together with the biochemical results reported in Fig. 1, these data provide strong evidence that hyperglycemia and diabetes increase calpain activity in the vascular endothelium.

The present study demonstrates a novel role for the calcium-dependent protease calpain in the pathophysiology of diabetic vascular disease. We provide novel evidence implicating μ-calpain in the endothelial dysfunction and vascular inflammation of the ZDF rat, a genetically occurring animal model of type 2 diabetes.

Oxidative stress, upregulation of endothelial cell adhesion molecules, and leukocyte trafficking has been linked with damage to blood vessels and organ tissue in diabetes (8,26,27). In vivo studies have demonstrated that accumulation of adherent leukocytes is increased in the aortic endothelium of alloxan-induced diabetic rabbits (28) and in the retinal microcirculation of diabetic rats (27), where capillary occlusions by leukocytes appears to precede destruction of the retinal capillary bed (29). The relationship between diabetes and cell surface expression of adhesion molecules has been and continues to be the subject of several in vivo and in vitro studies. Upregulation of cell adhesion molecules occurs in spontaneously hyperglycemic mice (30). Circulating levels of soluble ICAM-1 and VCAM-1 are increased in diabetic patients (31) and in response to transient hyperglycemia (6). Monocytes isolated from diabetic patients are more adhesive to cultured human endothelium (32), and they are in a proinflammatory state (33). Overall, the proinflammatory effects of diabetes on endothelial cell adhesion molecules in humans is associated with glucose-driven oxidative stress (34) and reduced release of NO (2,8).

It is now well appreciated that inflammation of the microcirculation can increase the risk of developing cardiovascular disease and atherosclerosis (35). Thus, inflammatory microangiopathy is considered an additional contributing factor to accelerated atherogenesis in diabetes. In fact, infiltration of leukocytes through the microvascular network of vasa vasorum may actually favor development of atherosclerosis (36).

Indeed, the cellular mechanisms responsible for the upregulation of leukocyte-endothelium interactions in the hyperglycemic vasculature remain elusive, which limits therapeutic interventions in diabetic patients. In the present study we provide novel evidence linking the calcium-dependent cysteine protease calpain with increased leukocyte trafficking in a relevant animal model of type 2 diabetes.

The calpains are a family of calcium-dependent proteases that act independently of the proteosome pathway and cleave a number of cellular substrates, including kinases, phosphatases, transcription factors, and cytoskeletal proteins (37). We focused our research on the well-characterized, ubiquitously expressed μ- and m-calpain isoforms, which have been isolated in a protein form. We found that diabetes selectively increases μ-calpain activity in vascular sections of the rat mesentery as well as in microvascular endothelial cells. The difference in isoform activation observed in the current study may be related to selective regulation of calpains under different pathological conditions. In support of this view, it has been suggested that whereas m-calpain is activated in response to large calcium fluxes, such as during apoptosis, μ-calpain may be more important in the setting of cell signaling under physiological and pathological conditions (21). Future studies addressing the signaling cascade(s) responsible for calpain activation in states of insulin resistance and diabetes may offer additional insight on the contribution of calpains to endothelial dysfunction and vascular disease.

Our study demonstrates that increased levels of endothelial NO and attenuated expression of ICAM-1 and VCAM-1 are key mechanisms of the anti-inflammatory action of calpain inhibition in the diabetic vasculature in vivo. Of note, it has been established in the literature that obesity, insulin resistance, and type 2 diabetes cause vascular inflammation with endothelial dysfunction (8), loss of endothelial NO (2), and activation of circulating leukocytes (27,33). Activated leukocytes adhere to the vascular endothelium, where they quench NO, increase oxidative stress, and release proinflammatory cytokines. Given the important role played by oxidative stress in the cardiovascular complications of obesity and diabetes (33,34), further studies are needed to better understand the impact of calpain inhibition on the production of oxidant species and inflammatory cytokines in the vascular wall during states of hyperglycemia and insulin resistance.

There is a general consensus in the literature that endothelial NO exerts potent anti-inflammatory effects in the cardiovascular system. Accordingly, endothelial dysfunction with impaired release of NO has been demonstrated in experimental animal models of hyperglycemia (4,13) and diabetes (14,30), as well as in diabetic humans (2,5). Notable to diabetes, insulin exerts important anti-inflammatory effects during acute experimental hyperglycemia of the microcirculation in the rat (13). In human endothelial cells, insulin increases release of NO (38), induces eNOS expression (39), and suppresses levels of ICAM-1 (40). Interestingly, the NO-potentiating effect of insulin is inhibited by tumor necrosis factor-α, a leukocyte-derived proinflammatory cytokine elevated in states of insulin resistance (41). Thus, it is likely that recurrent transient inflammatory events of the vascular wall initiated by primary loss of NO may exacerbate insulin resistance and cardiovascular disease. Interestingly, statins have been recently proven to exert cardiovascular protective actions, which are independent of their cholesterol-lowering activity because of direct anti-inflammatory and NO-potentiating effects (42). Recent studies have demonstrated reduced expression of eNOS in cultured endothelial cells exposed to elevated ambient glucose (43) or dysfunctional eNOS under hyperglycemic conditions (44). The anti-inflammatory action of NO is largely based on its inhibitory effect on leukocyte-endothelium interactions. At the molecular level, NO suppresses leukocyte-endothelium interactions by modulating the activity of nuclear factor-κB, a nuclear transcription factor that increases transcription of proinflammatory cell adhesion molecules such as VCAM-1 (24). Loss of basal release of NO in postcapillary venules induces inflammatory responses in the microcirculation characterized by increased leukocyte-endothelium interactions and upregulation of endothelial cell adhesion molecules (42). Thus, it is likely that after calpain inhibition, the attenuation in VCAM-1 expression levels observed in the present study were a consequence of increased availability of NO in the vascular endothelium.

The production of NO by eNOS is dependent, among other factors and regulatory proteins (45), on the molecular chaperone hsp90 (46). There is compelling evidence that hsp90 plays a key role in activating eNOS, as suggested by a recent study showing that the interaction of hsp90 and eNOS permits hsp90 to serve as a docking site for Akt-dependent phosphorylation of eNOS (47). Thus, decreased association of hsp90 with the eNOS complex decreases eNOS activity and endothelial NO release (46). Interestingly, it has been shown that the neutral cysteine protease calpain degrades hsp90 in several cell systems (48,49). Recently, a study by Su and Block (25) demonstrated that calpain-mediated degradation of hsp90 decreases NO production in pulmonary artery endothelial cells. These observations agree with our finding of reduced availability of endothelial NO and decreased hsp90/eNOS association in the microcirculation of diabetic rats. Indeed, inhibition of calpain activity in ZDF rats restored the association of eNOS with hsp90 along with increased NO release and attenuation of leukocyte trafficking in the microcirculation.

In conclusion, we have demonstrated that the activity of the calcium-dependent protease μ-calpain is increased in the diabetic vasculature and that inhibition of calpain activity attenuates the vascular dysfunction associated with chronic diabetes. These findings uncover a role for the calcium-dependent protease calpain in the pathophysiology of diabetic vascular disease.

T.J.S. was supported by National Institutes of Health Training Grant 5T32HL07599. This work was supported by National Institutes of Health Grant 1R01DK064344-01A1 (to R.S.).