Calpain Inhibitor SNJ-1945 Attenuates Events Prior to Angiogenesis in Cultured Human Retinal Endothelial Cells

Abstract

Purpose

Vascular endothelial growth factor (VEGF) is an important regulator of angiogenesis and microvascular permeability. VEGF-induced cytoskeletal reorganization plays a crucial role in angiogenesis. Cytoskeletal organization in endothelial cells is regulated by calpain proteases (EC 3.4.22.17). Calpains are a family of 14 calcium-regulated, intracellular cysteine proteases, which modulate cellular functions by limited, specific proteolysis. Calpain 1 (μ-calpain) and calpain 2 (m-calpain) are the 2 major typical calpain isoforms and are responsible for most calpain activity in endothelial cells. The purpose of the present study was to determine if an orally available form of calpain inhibitor, SNJ-1945, prevented angiogenesis induced by VEGF in cultured retinal endothelial cells.

Methods

Human retinal microvascular endothelial cells (HRMEC) were incubated with VEGF (60–100 ng/mL) for 24 h. Calcium uptake was measured with Fluo8. Total calpain activity was measured using fluorescent-labeled casein substrate, and separate activities for calpains 1 and 2 were assessed by casein zymography. Proteolysis of endogenous calpain substrate α-spectrin in situ was analyzed by immunoblotting. Angiogenesis in vitro was evaluated by measuring cell migration and tube formation into Matrigel.

Results

Incubation of HRMEC with VEGF resulted in calcium uptake, increased activity of mainly calpain 2, and increased calpain proteolysis of α-spectrin. Treatment of endothelial cells with calpain inhibitor SNJ-1945 reversed VEGF-mediated tube formation and cell motility.

Conclusions

Inhibition of angiogenesis by specific calpain inhibitor in the presence of VEGF supported our hypothesis that calpains may be involved in VEGF-mediated angiogenesis in retinal endothelial cells. Therefore, manipulating calpain activity by calpain inhibitor SNJ-1945 might provide a promising therapy for management of pathological angiogenesis, such as that occurring in proliferative retinopathy and age-related macular degeneration with neovascularization.

Introduction

Excess angiogenesis is a pathological hallmark in proliferative diabetic retinopathy (PDR) and age-related macular degeneration with neovascularization (NVAMD or “wet” AMD). Among US Medicare beneficiaries 65 years and older, the prevalence of these potentially blinding diseases was found to be 2.1% (PDR) and 0.5% (NVAMD).¹ Vascular endothelial growth factor (VEGF) is an important regulator of angiogenesis.^2,3 Indeed, the limited treatment options for NVAMD include several VEGF inhibitors,⁴ but these are not totally effective, suggesting the need for additional drug development.

VEGF-induced cytoskeletal reorganization plays a crucial role in angiogenesis. Cytoskeletal organization in endothelial cells is regulated by calpain cysteine proteases,^5,6 suggesting the development of antiangiogenesis drugs for retina, based on calpain inhibitors. Calpain inhibitors have been found to be partially effective in preventing calpain activation in several animal models of retinal degeneration.⁷ Calpains are a family of 14 calcium-regulated intracellular proteases, which modulate cellular function by limited, specific proteolysis.^8,9 Pertinent to ocular pathology, calpain inhibitors, leupeptin and SJA6017, reduced basic fibroblast growth factor (bFGF)-induced angiogenesis in the cornea of guinea pig.¹⁰ Su and colleagues¹¹ also reported that overexpression of endogenous calpain inhibitor calpastatin prevented VEGF-induced angiogenesis in human pulmonary microvascular endothelial cells (PMEC).

Although essential for further development of ocular drugs based on calpain inhibitors, no data exists in retinal cells testing if calpains are involved in VEGF-induced angiogenesis. Therefore, in the present study, we examined the effect of the chemically optimized calpain inhibitor SNJ-1945¹² on VEGF-induced angiogenesis in cultured, human retinal microvascular endothelial cells (HRMEC). Our data suggested that calpain is involved in VEGF-induced angiogenesis in retinal tissue. They provide the biochemical rationale for further testing calpain inhibitors with in vivo animal models of pathologic retinal neoangiogenesis.

Methods

HRMEC were purchased from Cell Systems Corporation (Kirkland, WA) and were cultured according to the instructions provided by the supplier. Only passages 3–6 were used for experiments. Cultures were checked often for purity by immunostaining with a rabbit polyclonal antibody against von Willebrand factor (Sigma-Aldrich, St. Louis, MO).

To measure Ca²⁺ mobilization in HRMEC, cells were seeded on 3.5 cm, glass bottom plates (Iwaki, Japan) and cultured overnight in a CO₂ incubator. Medium was replaced with 80 μL NW-Fluo8 containing loading buffer (ABD Bioquest, Sunnyvale, CA), and the cells were then incubated for 15 min at room temperature. The calcium response induced by 100 ng/mL VEGF was visualized under a laser scanning microscope (LSM710, Zeiss) at Em = 488nm/Ex = 530 nm. Cells were pretreated for 30 min with 4 mM EGTA (Sigma) as a negative control or with 250 μM SNJ-1945 to inhibit calpains before induction by VEGF. SNJ-1945 is ((1s)-1-(((1s)-1-benzyl-3-cyclopropylamino-2,3-di-oxopropyl)amino) carbonyl)-3-methylbutyl) carbamic acid 5-methoxy-3-oxapentyl ester and was obtained from Senju Pharmaceutical Co., Ltd (Kobe, Japan).

Calpain activities were assayed in vitro by detecting calcium-dependent cleavage of fluorescent BODIPY FL casein as previously reported¹³ and by casein zymography. To detect calpain activity with BODIPY FL-labeled casein, HRMEC were scraped and sonicated in buffer A containing 20 mM Tris (pH 7.5), 0.5 mM EGTA, and 2 mM dithioerythritol (DTE). Soluble proteins were obtained by centrifugation at 13,000g for 20 min at 4°C. Thirty μg soluble proteins from cell lysates were mixed with 10 μL BODIPY FL-labeled casein substrate solutions in buffer B (buffer A plus 3 mM CaCl₂) with a final volume of 200 μL. Five mM EDTA was substituted for Ca²⁺ in controls. After incubation for 30 min at 37°C, the reaction was terminated by adding 2 μL 1 M EDTA. The fluorescence was measured by spectrophotometry (excitation 485 nm, emission 528 nm). Calpain activity was calculated by subtracting the fluorescence units of samples with EDTA from those with CaCl₂. Protein concentrations were measured using Bio-Rad Protein Assay (Bio-Rad, Hercules, CA) with bovine serum albumin (BSA) as the standard. All steps were performed at 4°C.

Calpain activity in intact cells was measured by detecting calpain-specific spectrin breakdown products using antibody against αII-spectrin as reported by Newcomb and colleagues.¹⁴

Casein zymography was performed using the method of Racer and colleagues with some modifications.¹⁵ Ten percent acrylamide gels, copolymerized with 0.05% casein, were prerun with buffer containing 25 mM Tris (pH 8.3), 192 mM glycine, 1 mM EGTA, and 1 mM DTT for 15 min at 4°C. Eighty μg soluble proteins were then loaded and electrophoresed at 125 volts (constant) for 3 h at 4°C. After electrophoresis, the gels were incubated overnight with gentle shaking at room temperature in buffer containing 20 mM Tris (pH 7.4), 2 mM CaCl₂, and 10 mM DTT. Gels were stained with SimpleBlue™ SafeStain (Invitrogen, Carlsbad, CA). They were then digitized on a flat bed-scanner. Image analysis was performed by Image J1.37v software (NIH, Bethesda, MD). Bands of caseinolysis appeared white. Our previous studies using native-PAGE followed by immunoblotting identified the migration positions of 2 calcium-activated caseinolytic bands in zymograms of rat retina.¹⁶ Based on this, the present report used the upper band in zymograms as calpain 1 and the lower band as calpain 2.

For immunoblot analysis, SDS-PAGE of soluble proteins was first performed on 3%–8% Tris-acetate gels (NuPAGE; Invitrogen) with the Tris-acetate buffer system (Invitrogen). Immunoblots were performed by electrotransferring proteins from NuPAGE gels to polyvinylidene fluoride (PVDF) membranes (Millipore, Bedford, MA) at 30 volts (constant) for 60 min at ice-cold temperatures using NuPAGE transfer buffer (Invitrogen). The membranes were incubated in blocking solution containing Tris-buffered saline supplemented with 5% skim milk and 0.05% Tween 20 at room temperature for 1 h. The mouse monoclonal antibody against αII-spectrin (clone AA6, Affiniti Research Product, Exeter, UK) was used at 1:500 dilution. Immunoreactivity was visualized with alkaline phosphatase-conjugated to the secondary antibody and BCIP/NBT (Bio-Rad). For quantitative analysis, the staining intensity of the bands was determined by use of densitometric image analysis with Image J 1.37v software. To eliminate variability of staining between individual gels, the densities of bands for truncated fragments were normalized to the density of their normal parent proteins run on the same gel.

Endothelial cell migration was evaluated by measuring wound repair in an endothelial monolayer as previously reported with some modifications.¹¹ Briefly, cells were seeded in 24-well plate with a square-shaped seal in the center of each well and were allowed to grow to confluence. A wound area was created by removing the seal from the plate with sterile forceps. The plate was rinsed with Opti-DME containing 0.5% FBS to remove detached cells. The plate was then incubated in the same medium containing 100 ng VEGF/mL alone or with addition of 50 μM CBO-P11 (inhibitor of VEGF receptor) or 250 μM SNJ-1945 (inhibitor of calpain). Wound closure was monitored with an inverted phase-contrast microscopy (Olympus, San Diego, CA). Digital images were obtained at 48 h later with an Olympus digital camera attached to the microscope and were used for quantitative assessment of migration. The original wound area was measured at 0 h and compared to the area in treated wells at the conclusion of the experiment. Inhibition was tested in 3 independent experiments. Means for each group were calculated, and data were expressed as a percentage of the repaired area compared to the area at 0 h.

A tube formation assay was performed as previously reported,¹¹ with some modifications. Plates (96-well) were coated with 100 μL growth factor-reduced Matrigel matrix (BD Biosciences Discovery Labware, Bedford, MA) per well and allowed to polymerize for 30 min at 37°C. Approximately 2.5 × 10⁴ HRMEC in 100 μL volume per well were seeded on coated plates in Opti-DME containing 0.5% FBS at 37°C. VEGF, CBO-P11, and SNJ-1945 were added to a final concentration of 100 ng/mL, 100 μM, and 250 μM, respectively. Capillary-like tube structures formed by HRMEC on Matrigel were photographed at 8 h at 100× magnification with the digital camera (Olympus). Total tube length in mm/mm² in each well was measured. Each condition was tested in 4 independent experiments. Data were expressed as a percentage of the mean tube length compared to controls.

For statistical analysis, experimental and control endothelial cells in each experiment were matched for cell batch, seeding density, and number of passages to avoid variation in tissue culture factors that can influence measurements of calpain activity and angiogenesis. Results were shown as means ± SD for independent experiments. Statistical differences were determined by a Student's t-test (2 groups) or ANOVA analysis followed by Dunnett's multiple comparison test (>2 groups). A value of P < 0.05 was considered as significant.

Results

Since increased intracellular calcium is a requirement for calpain activation,⁸ calcium uptake was first visualized in HRMEC after stimulation by VEGF. Thirty seconds after stimulation, intracellular free calcium increased (Fig. 1B vs. 1A). Calpain inhibitor SNJ-1945 did not prevent Ca²⁺ uptake (D), while EGTA chelator prevented Ca²⁺ uptake (C). This indicated that at least the bulk VEGF-stimulated Ca²⁺ uptake was due to influx from the medium.

FIG.1.

Vascular endothelial growth factor (VEGF)-induced Ca⁺² influx in cultured primary human retinal microvascular endothelial cells (HRMEC). Calcium was visualized by Fluo8 in passage 4 cells, without (A) or with VEGF (100 ng VEGF/mL) (B), VEGF + EGTA (4 mM) (C), and VEGF + calpain inhibitor SNJ-1945 (250 μM) (D). Results were similar in 3 separate experiments.

To determine if the VEGF-stimulated Ca²⁺ uptake activated calpains, HRMEC were incubated with or without VEGF for 24 h, and calpain activities were measured by 2 independent methods. Using fluorescent-labeled BODIPY labeled casein to measure total activity of all calpain isoforms in HRMEC, VEGF stimulation was found to cause a statistically significant, 1.9-fold increase in calpain activities (Fig. 2A), findings that agree with published results.¹¹ Zymography using casein-loaded gels determined that calpains 1 and 2 were the major typical calpains responsible for most calpain activity in our HRMEC (Fig. 2B), as is found in other types of endothelial cells.¹⁷ In HRMEC, VEGF increased mainly calpain 2 activity (1.6-fold increase in calpain 2 vs. 1.2-fold increase in calpain, Figs. 2B and 1C). These data using exogenous casein substrate showed that at 24 h potentially active calpain 2 was increased by VEGF stimulation.

FIG. 2.

Calpain activities in control and vascular endothelial growth factor (VEGF)-treated human retinal microvascular endothelial cells (HRMEC) at 24 h measured by homogenizing cells and using assays based on calcium-dependent cleavage of fluorescent BODIPY FL casein substrate (A) or by zymography in casein-loaded gels (B). Insert: representative zymogram showing bands of lytic activities, previously characterized as calpains 1 (upper) and 2 (lower); the zymogram was then scanned for the bar graph. Grey bars indicate calpain activities in control cells, and black bars indicate calpain activities in VEGF-treated cells. Results are means ± SD; n = 4 experiments, ^*P < 0.05 vs. control.

Within cells, activated calpains hydrolyze only certain proteins at substrate-specific sites, frequently liberating bioactive fragments.⁸ A 150 kDa breakdown fragment (SBDP) of endogenous, intact, 280 kDa α-spectrin is one such biomarker for calpain activation within cells.⁷ In our intact HRMEC, an antibody to the calpain-specific SBDP detected increased production of this 150 kDa marker band after stimulation with VEGF (Fig. 3). This indicated that the VEGF induced influx of calcium and upregulation of calpain noted above increased calpain activation, resulting in calpain-specific proteolysis of the cytoskeletal protein α-spectrin.

FIG. 3.

Effect of vascular endothelial growth factor (VEGF) on endogenous calpain activity against α-spectrin at 24 h. Representative immunoblot showing the intact 280 and 150 kDa fragment of α-spectrin (A). Bar graph depicting the changes in the ratio of 150/280 kDa bands of α-spectrin (B). Results are expressed as mean ± SD; n = 4 experiments, ^*P < 0.05 vs. control.

Endothelial cell tube formation and angiogenesis require cell motility. HRMEC were, therefore, analyzed for the migration into a wound area produced by mechanically denuding a square area in the midst of confluent cells on culture plates. VEGF significantly increased cell motility into the denuded area compared to untreated cells (Fig. 4), and cocultured with a competitor (CBO-P11) of the VEGF receptor inhibited this VEGF effect. SNJ-1945 also blocked VEGF-induced cell motility (Fig. 4). SNJ-1945 inhibition of VEGF-induced cell motility occurred in a dose-dependent manner. These data strongly suggested that calpain could be involved in VEGF-induced cell migration during tube formation by retinal endothelial cells.

FIG. 4.

Effect of calpain inhibitor SNJ-1945 on endothelial cell migration. Human retinal microvascular endothelial cells (HRMEC) were incubated in the absence of added vascular endothelial growth factor (VEGF) (control), with 100 ng/mL VEGF, or with VEGF plus 50 μM CBO-P11 or 100/250 μM SNJ-1945. Cell migration after 48 h was evaluated by measuring repair of a denuded area in the endothelial monolayer. Results are expressed as mean ± SD; n = 3 experiments, ^#P < 0.05 vs. control (without VEGF) and ^*P < 0.05 vs. VEGF alone.

Since cytoskeletal rearrangement is vital to angiogenesis and VEGF caused cleavage of the cytoskeletal protein α-spectrin, our VEGF-stimulated HRMEC were analyzed for their ability to form initial structures in blood vessel formation, cord-like tubes. Even unstimulated HRMEC were able to form tubes into Matrigel (Fig. 5A, “Control”). Addition of VEGF significantly enhanced this tube formation (Fig. 5A and 5B, “VEGF”). Coculture with a competitor (CBO-P11) for the VEGF-receptor, inhibited VEGF-induced angiogenesis (Fig. 5B), showing that our tube formation was VEGF dependent. Calpain specific inhibitor SNJ-1945 also inhibited tube formation in the presence of exogenous VEGF, suggesting that calpain is an intracellular mediator in the chain of events leading to angiogenesis in retinal cells.

FIG. 5.

Effect of calpain inhibitor SNJ-1945 on retinal endothelial cell tube formation at 8 h. Human retinal microvascular endothelial cells (HRMEC) were seeded on Matrigel in the absence (control) or presence of 100 ng/mL vascular endothelial growth factor (VEGF) or with VEGF plus 100 μM CBO-P11 or 250 μM SNJ-1945 (A). Images shown are representative from 4 independent experiments (amplification: ×100). Bar graph depicting the changes in tube length (B). Results are expressed as mean ± SD; n = 4 experiments, ^#P < 0.05 vs. control and ^*P < 0.05 vs. VEGF alone.

Discussion

Su and colleagues¹¹ have elegantly demonstrated the role of calpain 2 in downstream events leading to VEGF-induced angiogenesis in vitro (cultured pulmonary endothelial cells) and in vivo (subcutaneous Matrigel plugs in mice). The present report was specific for cultured human primary retinal cells and, importantly, demonstrated that the conditions for calpain activation also exist in VEGF-induced retinal cell angiogenesis. These conditions included increased intracellular calcium (Fig. 1), upregulation of calpain 2 activity (Fig. 2), and calpain-specific hydrolysis of cytoskeletal α-spectrin (Fig. 3). Addition of calpain inhibitor SNJ-1945 did not prevent calcium influx (Fig. 1C) because the inhibitor mechanism involves covalent attachment of the α-keto carbonyl to the active site cysteine in calpain, not regulation of calcium. Indeed, molecular modeling studies based on x-ray crystallography data on the active site of calpain with calcium show that docking of SNJ-1945 produces a more open active site cleft.¹⁸ Thus, SNJ-1945 might bind and still be useful in retina even after pathologic increases in calcium, providing a “therapeutic window” for treatment.

Calpastatin (CS) is the endogenous, protein inhibitor of calpains.⁸ While CS is very specific for calpains, its molecular mass of native calpastatin of ~120 kDa,⁸ and even the 27-mer inhibitory consensus sequence peptide, make them impractical for direct use in drug therapy. Small aldehyde calpain inhibitors such as E64, leupeptin and SJA6017 are available, and leupeptin and SJA6017 reduced basic fibroblast growth factor-induced angiogenesis in the cornea of guinea pig.¹⁰ Inhibition of calpain 2 in fibroblasts by molecular “freezing” of calpain in an inactive conformation resulted in inhibition of EGF-induced calpain 2 activity and reduction in productive cell motility.¹⁹ In a rat model of ocular hypertension, oral SNJ-1945 ameliorated loss of cells in the retinal ganglion cell layer and strongly inhibited calpain-specific breakdown of α-spectrin in the retinal soluble proteins.²⁰ The present study is the first study to demonstrate that the synthetic, second generation, calpain inhibitor SNJ-1945 attenuates VEGF-induced human retinal endothelial cell migration (Fig. 4) and tube formation in vitro (Fig. 5). SNJ-1945 is a membrane permeable, calpain-specific inhibitor reacting with the active site of calpain.⁷ SNJ-1945 contains a α-ketoamide warhead, which is masked by a cyclopropane residue. This masked warhead makes SNJ-1945 less prone to react nonspecifically with various biologic amino and thiol groups, a problem with previous peptidyl inhibitors such as leupeptin and SJA6017.⁷ Further, SNJ-1945 is absorbed after oral administration and is taken up by the retina.¹²

Our finding that calpain 2 was the specific isoform of calpain activated in VEGF stimulated human retinal cells was important because the human genome codes for 13 other calpain genes.⁹ Transcripts for calpain 3 splice variants are found in both human and monkey retinas.²¹ The rodent genome, also codes for a retina-specific splice variant of calpain 3 called Rt88.²² Although Rt88 is not expressed in man because of a stop codon, young rodents models of cataract provide an illustration of the potential problems for drug development caused by calpain isoforms. Activation of another calpain 3 splice variant called Lp82 (lens-specific) is a major mechanism in young rodent cataractogenesis.²³ Lp82 is even resistant to endogenous inhibitor calpastatin.²⁴ Thus, the present study showing that calpain 2 in human endothelial retina cells is the correct target for inhibitors in man will be useful for designing future drug studies.

Franco and colleagues reported that calpain 2, but not calpain 1, was required for proteolysis of the cytoskeletal and focal adhesion proteins FAK, paxillin, spectrin, and talin.²⁵ This suggested functions for specific isoforms of calpains in substrate cleavage and regulation of cell migration. Calpains play an important role in signal transduction, leading to cell migration, differentiation, and proliferation in a variety of cells, including endothelial cells.^26,27 The timing for our first observations of various events associated with VEGF action varied. After VEGF treatment, calcium was increased by 30 s. Activation of calpain may start at this time. This activation may cause breakdown of substrates such as spectrin leading to tube formation, which was observed 8 h after VEGF treatment. Thus, calpain activity may have initially decreased when calcium increased because activation causes autolysis of calpain as well as proteolysis of spectrin. After that, production of calpain may have increased. As a consequence of calpain production, increased calpain activities were observed 24 h after VEGF treatment. The exact cause and effect mechanism for such changes on cell migration and tube formation are not known but may be related to calpainmediated alteration in the cell architecture affecting the cytoskeletal components and cell adhesion.^28–30

Regulation of angiogenesis occurs by balancing angiogenic and angiostatic factors; dysregulation of any of these factors can lead to pathological conditions, such as occurring in proliferative diabetic retinopathy and neovascular AMD. The present study provides the rationale for testing calpain inhibitors as drugs in animal models of retinal angiogenesis. The data are intriguing because they suggest that calpain inhibitor drugs may be adjunct treatment in antiangiogenesis therapy.

Acknowledgment

We thank Dr. Masatoshi Maki (Nagoya University, Japan) for his suggestion in preparing the manuscript.

Author Disclosure Statement

T.R.S. is a paid consultant for Senju Pharmaceutical Co., Ltd., a company that may have a commercial interest in the results of this research and technology. M.A. is an employee of Senju Pharmaceutical Co., Ltd. These potential conflicts of interest have been reviewed and managed by the OHSU.