PKCδ Impaired Vessel Formation and Angiogenic Factor Expression in Diabetic Ischemic Limbs

Primary antibodies for immunoblotting were obtained from commercial sources, including actin (horseradish peroxidase [HRP]; I-19), SHP-1 (C19), VEGF (147), PKCα (C-20), PKCβ (C-18), PKCε (C-15), and nitric oxide synthase (NOS) 3 (C-20) antibodies from Santa Cruz Biotechnology Inc.; phospho (p)-tyrosine, p-PKCδ (Thr 505), PKCδ, p-VEGF receptor 2 (VEGFR2; Y1175), VEGFR2, p-PDGF receptor-β (PDGFR-β; Tyr 1009), and PDGFR-β antibodies from Cell Signaling; anti-α smooth muscle actin from Abcam; polyclonal antibody against protein-tyrosine phosphatase 1B (PTP1B) and CD31 from BD Bioscience; SHP-2 and SHP-1 antibodies from Millipore; and rabbit and mouse peroxidase-conjugated secondary antibody from GE Healthcare Bio-Sciences. All other reagents used, including EDTA, leupeptin, phenylmethylsulfonyl fluoride, aprotinin, and Na₃VO₄, were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise stated.

C57BL/6J mice (6 weeks old) were purchased from The Jackson Laboratory (Bar Harbor, ME) and bred in our animal facility. Prkcd^−/− mice, described previously and provided by Dr. Michael Leitges (35), were generated by the insertion of a LacZ/neo cassette into the first transcribed exon of the PKCδ gene. This insertion abolished the transcription of PKCδ, leading to a null allele. Prkcd^−/− mice with mixed background of 129SV and C57BL/6J strains were crossbred for 10 generations (F12) with wild-type C57BL/6J background from The Jackson Laboratory. Animals were rendered diabetic for a 2-month period by intraperitoneal streptozotocin injection (50 mg/kg in 0.05 mol/L citrate buffer, pH 4.5; Sigma) on 5 consecutive days after an overnight fast; control mice were injected with citrate buffer. Blood glucose was measured by Glucometer (Contour, Bayer Inc.). Throughout the study period, animals were provided with free access to water and standard rodent chow (Harlan Teklad, Madison, WI). All experiments were conducted in accordance with the Canadian Council of Animal Care and University of Sherbrooke guidelines.

We assessed blood flow in nondiabetic and Prkcd^+/+ and Prkcd^−/− mice diabetic for 2 months. Animals were anesthetized, and the entire lower extremity of each mouse was shaved. A small incision was made along the thigh all the way to inguinal ligament and extending superiorly toward the mouse abdomen. The femoral artery was isolated from the femoral nerve and vein and ligated distally to the origin of the arteria profunda femoris. The incision was closed by interrupted 5-0 sutures (Syneture).

Hind limb blood flow was measured using PIM3 laser Doppler perfusion imaging (Perimed Inc.). Consecutive perfusion measurements were obtained by scanning the region of interest (hind limb and foot) of anesthetized animals. Measurements were performed before and after artery ligation and on postoperative days 7, 14, 21, and 28. To account for variables that affect blood flow temporally, the results at any given time were expressed as a ratio against simultaneously obtained perfusion measurements of the right (ligated) and left (nonligated) limb. Tissue necrosis was scored to assess mice that had to be killed during the course of the experiment due to necrosis/loss of toes.

Right and left abductor muscles from Prkcd^+/+ and Prkcd^−/− mice were harvested for pathological examination. Sections were fixed in 4% paraformaldehyde (Sigma-Aldrich) for 18 h and then transferred to 90% ethanol for light microscopy and immunohistochemistry. Paraformaldehyde-fixed tissue was embedded in paraffin, and 6-µm sections were stained with hematoxylin and eosin (Sigma). Apoptotic cells were detected using the TACS 2 Tdt-Fluor in situ apoptosis detection kit (Trevigen, Gaithersburg, MD) according to the manufacturer’s instructions.

Adductor muscles were blocked with 10% goat serum for 1 h and exposed in sequence to primary antibodies (CD31 and α-smooth muscle actin, 1:100) overnight, followed by incubation with secondary antibodies Alexa-647 conjugated anti-rabbit IgG and Alexa-594 conjugated anti-mouse (1:500; Jackson ImmunoResearch Laboratories). Confocal images were captured on a Zeiss LSM 410 microscope; images of one experiment were taken at the same time under identical settings and handled in Adobe Photoshop similarly across all images.

Adductor muscles were lysed in Laemmli buffer (50 mmol/L Tris [pH 6.8], 2% SDS, and 10% glycerol) containing protease inhibitors (1 mmol/L phenylmethylsulfonyl fluoride, 2 μg/mL aprotinin, 10 μg/mL leupeptin, 1 mmol/L Na₃VO₄; Sigma). Protein amount was measured with a BCA kit (Bio-Rad). The lysates (10–20 µg protein) were separated by SDS-PAGE, transferred to polyvinylidene fluoride membrane, and blocked with 5% skim milk. Antigens were detected using anti-rabbit HRP-conjugated antibody for other Western blotting and detected with the ECL system (Pierce Thermo Fisher, Piscataway, NJ). Protein content quantification was performed using computer-assisted densitometry with Image J software (National Institutes of Health).

Real-time PCR was performed to evaluate mRNA expressions of PKCα, PKCβ, PKCδ, PKCε, VEGF, PDGF, KDR/Flk-1, PDGFR-β, endothelial NOS (eNOS), SDF-1, FGF2, SHP-1, SHP-2, and PTP1B of nonischemic and ischemic limbs. Total RNA was extracted from adductor muscles with TRI-REAGENT, as described by the manufacturer and RNeasy mini kit (Qiagen, Valencia, CA). The RNA was treated with DNase I (Invitrogen) to remove any genomic DNA contamination. Approximately 1 μg RNA was used to generate cDNA using SuperScript III reverse transcriptase and random hexamers (Invitrogen) at 50°C for 60 min. PCR primers and probes are listed in Supplementary Table 1. Glyceraldehyde-3-phosphate dehydrogenase and 18S ribosomal RNA expression were used for normalization. PCR products were gel purified, subcloned using a QIA quick PCR Purification kit (Qiagen), and sequenced in both directions to confirm identity.

Adductor muscles were lysed and nuclear-specific proteins isolated using the NucBuster Protein Extraction Kit (Novagen, Madison, WI) according to the manufacturer’s instructions. Detection of hypoxia-inducible factor-1α (HIF-1α) in the nucleus was quantified using the nonradioactive transcription factor assay kit (Cayman, Ann Arbor, MI). Briefly, nuclear protein (20 µg) was incubated for 24 h in a 96-well plate containing immobilized specific double-stranded DNA consensus sequence of the HIF-1α response element. HIF-1α contained in the nuclear extract was linked specifically to the HIF-1α response element. Wells were washed five times, and the HIF transcription factor complex was detected by addition of a specific primary antibody directed against HIF-1α and incubated for 1 h. Wells were washed five times and exposed with secondary antibody conjugated to HRP for 1 h. Wells were then washed five times, and developing agent was added to provide a sensitive colorimetric readout at 450 nm (Infinite M200; Tecan Group Ltd., Männedorf, Switzerland) to quantify nuclear HIF-1α levels.

The data are shown as mean ± SD for each group. Statistical analysis was performed by unpaired t test or by one-way ANOVA, followed by the Tukey test correction for multiple comparisons. All results were considered statistically significant at P < 0.05.

Nondiabetic and diabetic male Prkcd^−/− mice and control littermates (Prkcd^+/+) underwent unilateral right femoral artery ligation. Body weight and fasting glucose levels were measured when mice were killed (Supplementary Table 2). Blood flow reperfusion was assessed using the PIM 3 laser Doppler imaging system (Fig. 1A). Diabetic Prkcd^+/+ mice exhibited reduced blood reperfusion of the ischemic limb compared with nondiabetic Prkcd^+/+ mice (P = 0.0046). In contrast, reperfusion of blood flow of diabetic Prkcd^−/− mice was significantly improved (P = 0.0003) compared with diabetic Prkcd^+/+ mice and similar to nondiabetic Prkcd^+/+ and Prkcd^−/− mice 28 days after the ligation (Fig. 1B). Because diabetic patients are at high risk of lower limb amputation, we assessed limb necrosis and apoptosis. Impaired reperfusion in ischemic limbs of diabetic Prkcd^+/+ mice was associated with elevated tissue necrosis, amputation (Fig. 1C), and apoptosis (Fig. 2) compared with nondiabetic counterparts.

One main effect of hypoxia is to induce angiogenesis and to promote new capillary formation. To test whether activation of PKCδ is responsible for poor collateral vessel formation in diabetes, we measured capillary density and capillary diameter in the ischemic adductor muscles. Figure 3 demonstrated that the adductor muscles of diabetic Prkcd^+/+ mice displayed a 31% vascular density reduction compared with nondiabetic Prkcd^+/+ mice. The decline of capillary density was accompanied with a 50% reduction in number of vessels with a diameter of 50 µm or less. Interestingly, diabetic Prkcd^−/− mice showed a significant increase in capillary density and number of vessels with a diameter of less than 50 µm compared with diabetic Prkcd^+/+ mice (Fig. 3D).

Hyperglycemia is known to activate multiple PKC isoforms, preferably the β and δ isoforms, in vascular cells. Expression of various isoforms of PKC was assessed by quantitative PCR in muscle tissues (Fig. 4). Compared with nondiabetic Prkcd^+/+ mice, mRNA expression of PKCβ and δ was modestly increased in adductor muscles of diabetic Prkcd^+/+ mice and unchanged in Prkcd^−/− mice (Fig. 4B and D). There was no significant difference in the mRNA expression of PKCα and -ε (Fig. 4A and C). Although diabetic Prkcd^+/+ mice did not exhibit higher levels of protein expression of PKCα, -β2, -ε, or -δ isoforms, adductor muscles of Prkcd^+/+ mice showed a significant and impressive increase of PKCδ phosphorylation (Thr 505; P = 0.0040), as a marker of PKCδ activation, 28 days after unilateral femoral artery ligation compared with nondiabetic Prkcd^+/+ mice (Fig. 5).

To explain how the absence of PKCδ improved reperfusion in diabetic ischemic limbs, we performed a wide analysis of the gene and protein expression of angiogenic-related factors and their receptors. Quantitative gene expression analyses by real-time PCR indicated that VEGF-A, PDGF-B, and PDGFR-β mRNA expression was significantly decreased in the adductor muscles of diabetic mice by 46, 30, and 63%, respectively, compared with nondiabetic Prkcd^+/+ mice (Fig. 6A, C, and D). The reduction of these genes in diabetic Prkcd^+/+ mice was not observed in diabetic Prkcd^−/− mice. Moreover, mRNA expression of VEGFR2 (KDR/Flk-1), PDGF-B, and PDGFR-β was significantly upregulated in diabetic Prkcd^−/− compared with diabetic Prkcd^+/+ mice (Fig. 6B–D). These results suggest that impaired PDGF and VEGF expression by PKCδ activation might be the contributing factor for poor collateral vessel formation in diabetes. Expression of other angiogenic factors, such as SDF-1, FGF-2, and eNOS, as well as transcriptional factor activity of HIF-1α, was unchanged within all groups of mice (Fig. 6E–H and Supplementary Fig. 1). In contrast to 4 weeks after femoral artery ligation, transcriptional factor activity and mRNA levels of HIF-1α were significantly decreased in diabetic Prkcd^+/+ mice compared with nondiabetic Prkcd^+/+ and diabetic Prkcd^−/− mice (Supplementary Figs. 2 and 3).

To further investigate the mechanisms of impaired angiogenic response to restore blood flow in diabetes, the expression, activation, and signaling pathway of VEGF-A and PDGF-B and their respective receptors (VEGFR2 and PDGFR-β) were examined. Protein expression of PDGF-B was significantly decreased in diabetic versus nondiabetic adductor muscles of wild-type animals. In contrast, VEGF-A and PDGF-B protein expression was elevated in the ischemic limb of the diabetic PKCδ null mice (Fig. 7A and B). Phosphorylation of VEGFR2 and PDGFR-β was inhibited in ischemic adductor muscles of diabetic mice compared with nondiabetic Prkcd^+/+ mice. However, activation of Src was elevated in adductor muscles of diabetic Prkcd^+/+ mice compared with nondiabetc Prkcd^+/+ and Prkcd^−/− mice (Fig. 7B). Interestingly, tyrosine phosphorylation of VEGFR2 and PDGFRβ, as well as PLCγ1, Akt, and ERK phosphorylation, was greatly enhanced in Prkcd^−/− mice compared with diabetic Prkcd^+/+ mice (Fig. 7A and B). We did not observe any changes in the eNOS protein expression among experimental groups (Fig. 7A).

We have previously shown that activation of PKCδ leads to increased expression of SHP-1, which inhibits the PDGF-signaling pathway and promotes retinal pericyte apoptosis in diabetic animals. To determine whether SHP-1 is implicated in PKCδ-induced VEGFR2 and PDGFR-β dephosphorylation in diabetic ischemic adductor muscles, we measured SHP-1 expression by quantitative PCR and immunoblot analysis. Figure 8A and B indicates that mRNA expression of SHP-1, but not SHP-2 or PTP1B, is elevated in diabetic Prkcd^+/+ mice, whereas SHP-1 is clearly downregulated in Prkcd^−/− mice. We confirmed through immunoblot analysis that SHP-1 protein expression was elevated by 2.3-fold in ischemic adductor muscles of diabetic Prkcd^+/+ mice compared with nondiabetic Prkcd^+/+ mice. The increase expression of SHP-1 was not observed in diabetic Prkcd^−/− mice (Fig. 8C). No change was detected in the protein expression of SHP-2 and PTP1B within all groups of mice (Fig. 8D).

Diabetes is associated with the progression of vascular complications, such as peripheral arterial disease, and is a major risk factor for lower limb amputations (4). In the current study, we have demonstrated that activation of PKCδ diminishes the expression of VEGF and PDGF, two critical proangiogenic factors contributing to poor capillary formation and blood flow reperfusion of the ischemic limbs. In addition to reducing expression of VEGF and PDGF, phosphorylation of VEGF and PDGF receptors was abrogated in diabetic ischemic muscles compared with nondiabetic ischemic muscles. The inhibition of growth factor receptor phosphorylation was associated with the upregulation of SHP-1 expression, which has been reported to deactivate tyrosine kinase receptors such as VEGF and PDGF receptors. Overall, deletion of PKCδ prevents the reduction of VEGF and PDGF expression and re-establishes KDR/Flk-1 and PDGFR-β phosphorylation, favoring new capillary formation and blood flow reperfusion.

Wound healing is a complex, well-orchestrated, and dynamic process that involves a coordinated and precise interaction of various cell types and mediators. Given the fundamental contribution of VEGF and PDGF to the angiogenic process, the mechanism by which activation of PKCδ isoform prevents growth factors expression and signaling actions may provide a better understanding of how diabetes reduces collateral vessel formation in the ischemic limb. In this study, we demonstrated that PKCδ is activated in diabetic ischemic muscles and reduced blood flow reperfusion, contributing to tissue necrosis, amputation, and apoptosis. Previous studies have reported that PKCδ is involved in vascular cell apoptosis. PKCδ activates p-38, mitogen-activated protein kinase, p53, and caspase-3 cleavage to favor endothelial (36) and smooth muscle cell apoptosis (37,38). Therefore, deletion of PKCδ may enhance vascular cell migration and proliferation, two significant steps in the formation of new blood vessels.

Total expression of PKC isoform in ischemic muscles was slightly affected by diabetes, probably because mRNA and protein analyses were performed 28 days after femoral artery ligation. However, phosphorylation of PKCδ on threonine 505, a phosphorylation site within the activation loop, clearly suggests that PKCδ is activated in the muscles of diabetic ischemic limbs compared with nondiabetic muscles. Previous data showed that the inhibition of PKCδ, using an isozyme-specific peptide, improved the number of microvessels and cerebral blood flow after acute focal ischemia in normotensive rats (39). Our data demonstrate that deletion of PKCδ restores blood flow perfusion in diabetic ischemic muscles by promoting the number of capillaries and reducing tissue apoptosis.

The reduction of VEGF and PDGF receptor expression and the downstream signaling pathway is associated with impaired angiogenesis process in diabetic foot ulcer and ischemic diseases. Our results indicate that diabetes-induced PKCδ activation decreases VEGF, PDGF, KDR/Flk-1, and PDGFR-β mRNA expression in the ischemic limb, which is completely restored in PKCδ-null mice. Interestingly, impaired angiogenic response in ischemic arterial diseases of type 1 and type 2 diabetes is associated with VEGF inhibition in endothelial cells and monocytes (13,40). It is possible that the ablation of PKCδ may also affect VEGF signaling in monocytes, which may contribute to vessel formation abnormalities. However, this assumption will need further investigation.

HIF-1α is a master regulator of angiogenic factors in response to tissue hypoxia. Previous study showed that HIF-1α gene transfer increased recovery of limb perfusion, increasing eNOS activation and vessel density (41). In our study, however, the increase in the expression of VEGF in muscles of PKCδ-deficient mice may not have been entirely due to upregulation of HIF-1α. Because protein extraction was performed 4 weeks after the femoral artery ligation, it is possible that the expression of HIF-1α could have returned to basal levels. This hypothesis is supported by results obtained 2 weeks after the surgery. Our data demonstrated that HIF-1α transcriptional factor activity and mRNA expression were increased in nondiabetic and diabetic PKCδ-null mice 2 weeks only after surgery (Supplementary Figs. 2 and 3). Besides VEGF and PDGF expression, our data suggest that PKCδ activation disrupts VEGF and PDGF signaling, whereas in PKCδ-deficient mice, the activity of VEGFR2, PDGFR-β, PLCγ1, Akt, and ERK is enhanced. Surprisingly, Src phosphorylation was increased in the ischemic muscles of diabetic wild-type mice even if PDGFR-β activity was reduced. However, a previous study reported that reactive oxygen species (ROS) production induced Src phosphorylation (42). Because ROS are massively produced in ischemic and hyperglycemic conditions, it is probable that ROS production is responsible for the Src phosphorylation seen in diabetic wild-type mice.

There is strong evidence that progenitor cell recruitment and homing participate in angiogenesis and wound repair, which are guided by SDF-1 (43). Although the number of progenitor cells is reduced in diabetic mice, inadequate progenitor cell mobilization has been proposed as one potential mechanism of impaired angiogenesis (44). However, our results did not observe any change in SDF-1 expression in PKCδ-null mice, suggesting that mobilization and local trafficking of progenitor cells to the ischemic site was not affected by the PKCδ isoform.

Despite advances in revascularization techniques, limb salvage and pain relief cannot be achieved in many diabetic patients with diffuse peripheral vascular disease. VEGF-mediated gene therapy has shown promising results as an innovative method in the treatment of severe cardiovascular diseases. However, a randomized study of gene therapy failed to meet the primary objective of significant amputation reduction (45). During the 10-year follow-up period, no significant differences were detected in the number of amputations or causes of death with the use of transient VEGF-A–mediated gene therapy. One reason for this lack of improvement is perhaps because neovascularization requires the interaction of multiple growth factors that can promote, in a synergic manner, new and mature blood vessels. Enhancing the responsiveness of diabetic vascular cells to proangiogenic factors may offer a potential new approach to treat peripheral arterial diseases. Protein tyrosine phosphatase is a group of proteins that is critical in abating cell response to growth factors by inhibiting tyrosine kinase phosphorylation. Our results demonstrated that SHP-1 expression was increased in diabetic ischemic muscles and was responsible for VEGF and PDGR receptor dephosphorylation.

Although not significant, a slight rise in SHP-2 (18%) and PTP1B (37%) expression was observed in diabetic PKCδ-null mice. Previous studies have shown that PDGF activation enhanced SHP-2 and PTP1B activity (46,47), which may explain our results. We have reported that activation of PKCδ induces the expression of SHP-1 in cultured pericytes exposed to high glucose concentrations and inhibits the PDGF signaling pathway contributing to pericyte apoptosis (23). Others studies have also shown that SHP-1 is a negative regulator of VEGF signal transduction and inhibits endothelial cell proliferation (48,49). Interestingly, silencing SHP-1 increased phosphorylation of KDR/Flk-1 and markedly enhanced capillary density in a nondiabetic hind limb ischemia model (50). However, our current study does not provide a direct link between SHP-1 expression and reduced angiogenesis, which will require further investigations. Nevertheless, our findings have identified PKCδ, and potentially SHP-1, as potential therapeutic targets for the treatment of diabetic peripheral arterial diseases and cardiovascular complications.

In summary, we have provided evidence that PKCδ is activated by diabetes in ischemic muscles and induced SHP-1 expression, contributing to VEGF and PDGF unresponsiveness and poor angiogenesis response. Although various therapies are partly successful in restoring blood flow to the affected tissues, there is no effective strategy to specifically produce new functional vessels to dismiss diabetic ischemic stress. Our data enhance our understanding of the mechanisms underlying poor collateral vessel formation induced by PKC activation and may offer potential novel targets to regulate angiogenesis therapeutically in patients with diabetes.

This study was supported by grants from the Canadian Diabetes Association, Fonds de Recherche du Québec–Santé, and Diabète Québec to P.G. and was performed at the Centre de Recherche Clinique Étienne-Le Bel, a research center funded by the Fonds de Recherche du Québec–Santé. P.G. is currently the recipient of a Scholarship Award from the Canadian Diabetes Association and the Canadian Research Chair in Vascular Complications of Diabetes.

No potential conflicts of interest relevant to this article were reported.

F.L., M.P., B.D., and P.G. performed experiments and analyzed the data. M.L. provided the Prkcd-deficient mice. A.G. performed animal care and researched data. F.L. and P.G. wrote the manuscript. P.G. is the guarantor of this work, and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

The authors gratefully acknowledge Marie-Élaine Clavet (Montreal Heart Institute) for her assistance with histochemistry technics.