Primary antibodies for immunoblotting were purchased from commercial sources: GAPDH horseradish peroxidase (V-18), endothelial nitric oxide synthase (eNOS) (C-20) from Santa Cruz Biotechnology Inc (Dallas, TX, USA); protein kinase B (Akt) (9272S), phospho-Akt (193H12), AMP-activated protein kinase α (AMPKα) (2532S), p-AMPKα (2535S), p-eNOS Ser1177 (9571S), Ras homolog family member A (RhoA) (2117S), rho-associated coiled-coil-containing protein kinase (ROCK) 1 (4035S) and ROCK-2 (8236S), and secondary antibody of anti-rabbit (7,074 V) and anti-mouse (7076S) peroxidase-conjugated from Cell Signaling (Danver, MA, USA); anti-α smooth muscle actin (ab5694) and p-ROCK2 (ab228008) from Abcam (Toronto, ON, CA); anti-CD31 (558,736) and p-eNOS Ser633 (612,664) from BD Bioscience (Mississauga, ON, CA). Secondary antibodies for immunofluorescence FITC conjugated anti-rat IgG (712-095-153) and Alexa Fluor 647 conjugated anti-rabbit IgG (711-606-152) were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA, USA). Halt Protease Inhibitor Cocktail (78,430) was purchased from Thermo Fisher Scientific (Waltham, MA, USA). SEP-COLUMN (RK-SEPCOL-1) for plasma peptides extraction and Apelin-12 (Human, Rat, Mouse, Bovine) Enzyme Immunoassay (EIA) (EK-057-23) were purchased from Phoenix Pharmaceuticals Inc (Burlingame, CA, USA). Fetal bovine serum (FBS, 080-150), penicillin-streptomycin (P/S, 450-201-EL) and phosphate-buffered saline (PBS, 311-410-CL) were purchased from Wisent Bioproducts (Saint-Jean Baptiste, QC, CA). Dulbecco's Modified Eagle Medium (DMEM) low glucose (31,600–034) were obtained from Invitrogen (Burlington, ON, CA). VEGF-A₁₆₅ was purchased from R & D (293-VE-010, Minneapolis, MN, USA). Pyr-apelin-13 (pyr-Ape-13) was synthesized and generously provided by Dr. Boudreault's laboratory from the Institut de Pharmacologie de Sherbrooke. D-mannitol (M-120), sodium citrate (BP327.1) and citric acid (A940–500) were purchased from Fisher scientific (Hampton, NH, USA). All other reagents used, including streptozotocin (S0130), ethylenediaminetetraacetic acid (EDTA, E5134), bovine serum albumin (BSA, A7906) d-glucose (158,968), leupeptin (L2884), phenylmethylsulfonyl fluoride (P7626), aprotinin (A6279), NaF (S7920) and Na₃VO₄ (S6508) were purchased from Sigma-Aldrich (St. Louis, MO, USA).

Bovine aortic endothelial cells (BAECs) were isolated from fresh harvested aorta as previously described (7). Cells from passages 2 to 7 were trypsinized and cultured in DMEM 2.5% FBS and 1% penicillin-streptomycin. For all the in vitro experiments, cells were cultured in DMEM 0.1% FBS containing normal (NG; 5.6 mmol/L + 19.4 mmol/L of Mannitol) or high glucose (HG; 25 mmol/L) levels for up to 48 h and exposed to hypoxia for the last 16 h (1% O₂) to reproduce the ischemic state of PAD. For immunoblotting and cell signaling experiments, BAECs were stimulated with either VEGF-A 10 ng/ml for 5 min (7) or Pyr-apelin-13 200 nM for 1 h, or 24 h prior to RNA extraction and quantitative PCR analysis. For migration, proliferation and tube formation assays BAECs were stimulated with either VEGF-A 25 ng/ml (7) or Pyr-apelin-13 100 nM. Concentrations used for Pyr-apelin-13 stimulation were based on previous dose dependent experiments (data not shown).

BAECs were trypsinized, counted and seeded at 20 000 cells inside the 2 wells of Ibidi's culture insert (80,209, Ibidi, Fitchburg, WI, USA) placed into a 12-well plate. After cell adhesion (4 h later), BAECs were exposed to NG or HG concentrations for 24 h. Following the 24 h treatment, each culture insert was removed, and the wells were filled with 1 ml of either NG or HG medium. Cells were then stimulated, or not, with either VEGF-A (25 ng/ml) or Pyr-apelin-13 (100 nM) and placed into the hypoxic incubator (1% O₂) for 16 h. Cell migration was evaluated by taking images under Nikon eclipse Ti microscope at 10 × magnification immediately after removing the culture insert and at the end of the experiment (16 h later). Analysis was performed with Image J software by measuring the difference in occupied area immediately after insert removal and following 16 h stimulation in NG or HG condition. Results were reported as a % of cell migration for analysis.

BAECs were trypsinized, counted and seeded in a 96-well plate at 5 000 cells per well. After cell adhesion (4 h later), cells were exposed to NG or HG concentrations for 24 h, stimulated with either VEGF-A (25 ng/ml) or Pyr-apelin-13 (100 nM) and placed into the hypoxic incubator for 16 h. Cells were then fixed in 4% paraformaldehyde for 5 min, rinsed twice in PBS and incubated with the nuclear counterstain DAPI (D9542, Sigma-Aldrich, St. Louis, MO, USA) at 0.001 mg/ml for 10 min. Fluorescence microscopy and the NIS-Elements software of Nikon eclipse Ti microscope were used to visualize and count cells, reported as the number of cells/mm² for analysis.

BAECs were cultured in 100 mm petri dish and exposed to NG or HG concentrations for 24 h and then placed into the hypoxic incubator (1% O₂) for 16 h. On the day of the assay, 10 μl of Matrigel Matrix (Growth Factor Reduced, 356,230, Corning, Glendale, AZ, USA) was applied into each well of Ibidi's μ-slide rack (81,506, Ibidi, Fitchburg, WI, USA) and incubated for 30 min at 37°C to allow polymerization. Cells were then trypsinized, counted, and 600 000 cells were seeded per well on top of the Matrigel. After seeding, cells were immediately stimulated, or not, with VEGF-A (25 ng/ml) or Pyr-apelin-13 (100 nM) and the μ-slide rack was placed into the hypoxic incubator for 4 h. At the end of the experiment (4 h later), the lumen formation was visualized, and images were taken under Nikon eclipse Ti microscope at 4 × magnification. Endothelial cell tube formation ability after NG or HG treatment and angiogenic factor stimulation was measured using Image J software by counting the number of formed lumens and normalized on the NG condition to present it as a fold increase.

Adductor muscles or endothelial cells were lysed in RIPA buffer containing protease inhibitors (1 mmol/L phenylmethylsulfonyl fluoride, 2 µg/ml aprotinin, 10 µg/ml leupeptin, 1 mmol/L Na3VO4, 1 mmol/L NaF) and protein concentrations were measured by DC kit (5,000,116, BioRad, Mississauga, ON, CA). The lysates (20 μg for BAECs and 50 μg for adductor muscles) were separated by SDS-PAGE and transferred to PVDF membranes, which were blocked with 5% skim milk for 1 h at room temperature. Membranes were incubated overnight with primary antibodies at a 1:1,000 dilution in 5% skim milk (or in 5% BSA for phospho-ROCK-2 and p-AMPKα), followed by the corresponding secondary antibodies conjugated with horseradish peroxidase at a 1:10 000 dilution in 5% skim milk (or 1:5 000 for phospho-ROCK-2 and p-AMPKα). Immobilon Forte Western HRP substrate (WBLUF0100, millipore, Etobicoke, ON, CA) was used to visualize immunoreactive bands and protein content was quantified using Computer-assisted densitometry ImageLab imaging software (Chemidoc, BioRad).

C57Bl/6 male mice were purchased from Charles River (strain 027, Wilmington, MA, USA). At 8 weeks of age, the mice were rendered diabetic by intraperitoneal streptozotocin injection (50 mg/kg in 0.05 mol/L citrate buffer, pH 4.5) on 5 consecutive days after overnight fasting. Control mice were injected with citrate buffer. Blood glucose was measured with a glucometer (Contour, Bayer, Inc.) one week after the injections to confirm diabetes. All experiments were conducted in accordance with the Canadian Council of Animal Care and were approved by the Animal care and Use Committees of the Université de Sherbrooke, according to the NIH Guide for the Care and Use of Laboratory Animals.

After 2 months of diabetes, blood flow was measured in 16-week-old nondiabetic (NDM) and diabetic (DM) mice. Animals were anesthetized by inhalation of Isoflurane USP (1-chloro-2,2,2-trifluoroethyl difluoromethyl ether) at a concentration of 5% (initiation) and then maintained at 1%–2% during the whole surgical procedure (approximately 20 min). To mimic the ischemic condition in PAD, unilateral hindlimb ischemia was induced by ligation of the femoral artery as we previously described (24). Directly after the hindlimb surgery, osmotic pumps (model 1,004; Alzet Osmotic pumps, Cupertino, CA, USA) with Pyr-apelin-13 (in PBS) were implanted subcutaneously, according to the manufacturer's instructions, to allow 4-week infusions. Dose dependent experiments of Pyr-apelin-13 (0.36, 1 and 2 mg/kg/day) were performed in DM mice to select the proper concentration. The blood flow recovery following hindlimb ischemia was assessed in NDM, DM and DM + pyr-Ape-13.

The hindlimb blood flow was measured using a laser Doppler perfusion imaging (PIMIII) system (Perimed Inc., Las Vegas, NV, USA). Consecutive perfusion measurements were obtained by scanning the region of interest (hindlimb and foot) of anesthetized animals. Measurements were performed pre-artery and post-artery ligation to ensure the success of the surgery, and on post-operative 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 value of the ischemic (right) and nonischemic (left) hindlimb. Following laser Doppler perfusion imaging on day 28, mice were euthanized by exsanguination via the left ventricle under deep anesthesia (Isoflurane USP, inhalation at a concentration of 5%) and the ischemic adductor muscles were harvested.

Mice were housed individually for 5 consecutive days between day 23 and 27 to measure their activity using voluntary exercise wheel (Scurry mouse activity wheel, Model 80820FS, Lafayette Instrument Neuroscience, Lafayette, IN, USA). A magnetic sensor (Scurry mouse activity counter, Model 86,110) attached to the wheel recorded the number of revolutions and running distance using the Scurry activity monitoring software (Lafayette Instrument Neuroscience). Results are shown as cumulative distance of the 5 days.

Mice were euthanized by exsanguination and blood was immediately centrifuged to isolate the plasma. Halt Protease Inhibitor Cocktail was added to the plasma to prevent apelin degradation and samples were frozen in liquid nitrogen. Peptides from the plasma were extracted on SEP-COLUMN containing 200 mg of C18 following the manufacturer's instructions. Briefly, the plasma was acidified using the Buffer A provided in the extraction kit and loaded into the pretreated C-18 SEP-COLUMNs. Columns were washed twice with Buffer A then the peptides were eluted with Buffer B. Eluants were evaporated to dryness in a centrifugal concentrator and then rehydrated in 125 μl of Assay Buffer provided in the EIA kit. Apelin-12 EIA kit was used according to the manufacturer's procedures to detect apelin-12, 13 and 36 isoforms, with a sensitivity level in the range of ng/ml, based on the principle of competitive EIA.

Ischemic adductor muscles from NDM, DM and DM + pyr-Ape-13 were harvested for pathological examination and sections were fixed in 4% paraformaldehyde (89,370–094, VWR, Radnor, PA, USA) for 18 h and then transferred to 70% ethanol. Fixed tissues were embedded in paraffin and 4 µm sections were stained with hematoxylin & eosin (H & E) or used for immunofluorescence assay. H & E-stained cross section images were captured with the Nikon eclipse Ti microscope, and 4 regions of 0.25 mm² were randomly selected on the entire ischemic muscle. Within these areas, the diameter of 50 myofibers were randomly measured, using the Image J software, to obtain the mean myofiber diameter per mouse. All images were taken at the same time under identical settings.

Cross-sections of ischemic adductor muscle were blocked at room temperature for an hour with 10% goat serum and then incubated overnight with primary antibodies (CD31 (1:50) and α-smooth muscle actin (α-SMA), 1:200), followed by 1 h incubation with the secondary antibody (1:400). Images of the entire ischemic muscle were taken with the Nikon eclipse Ti microscope. Vessels with only CD31-positive signal and a lumen diameter less than 10 µm (capillaries) were separately counted from vessels with CD31/α-SMA staining ranging from 10 to 30 µm (arterioles). Vascular density of capillaries and arterioles was normalized by the entire muscle fiber density of each mouse. All images were taken at the same time under identical settings and similarly handled in Adobe Photoshop and Image J software.

Quantitative PCR was performed to evaluate mRNA expression of genes in the ischemic adductor muscles and BAECs as we previously described (7). The expression of the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used to normalized data. The specific primer sequences are listed in Table 1.

In vivo and in vitro data are shown as mean ± SD for each group except for blood flow measurements which are presented as mean ± SEM. Statistical analysis was performed by unpaired Kruskal–Wallis followed by Dunn's multiple comparisons test (Figures 1B,C,E,F, 4A–4D, 5G, 6A, 7G) or by unpaired One-Way ANOVA followed by Tukey's multiple comparisons test (Figures 2B–D, 3C–E, 4E,F, 5B–F,H–J, 6, 7A–F). Data in each group were checked for normal distribution using D'Agostino and Pearson normality test based on α = 0.05. All results were considered statistically significant at P < 0.05.

We have previously demonstrated that crucial factors involved in the angiogenic processes, such as VEGF, are affected in diabetes (7, 25). To determine and understand if apelin can improve endothelial function under a hyperglycemic milieu, we have evaluated its role on critical steps involved in blood vessel formation and compared its response to VEGF-A. Under NG conditions, hypoxia exposure promoted endothelial cell migration by 38% coverage of the surface area as compared to the surface area initially created by Ibidi's insert. As expected, under the same conditions, VEGF-A and Pyr-apelin-13 stimulation enhanced cell migration by 80% (P = 0.0018) and 116% (P < 0.0001), respectively (Figures 1A,B). As previously reported under HG and hypoxic conditions, VEGF-A-stimulated cell migration was inhibited by 92% (P = 0.0100), while Pyr-apelin-13 promoted endothelial cell migration by 93% (P = 0.0284; Figures 1A,B). Cell proliferation is also an important step in the angiogenic process. Our data demonstrated that the mitogenic effect of VEGF-A (1.4-fold; P = 0.0486) on endothelial cell proliferation under NG and hypoxic conditions were blunted in HG level exposure (Figure 1C). In contrast, Pyr-apelin-13-induced endothelial cell proliferation was conserved in HG conditions (1.9-fold; P = 0.004; Figure 1C). Then, we investigated the capacity of these growth factors to induce vascular lumen formation on Matrigel. We observed that VEGF and apelin treatment increased similarly endothelial cell lumen formation when exposed to NG condition and hypoxia by 1.6-fold (P = 0.0094) and 1.5-fold (P = 0.0063), respectively (Figures 1D,E). In line with the other angiogenic processes, the capacity of VEGF to induce lumen formation was abrogated in HG level exposure, which was not observed with Pyr-apelin-13. Treatment with Pyr-apelin-13 enhanced lumen formation in both NG (1.5-fold; P = 0.0063) and HG (1.7-fold; P = 0.0056) conditions (Figures 1D,E). Inhibition of VEGF-induced proangiogenic actions in diabetes has been well documented by us and others, and associated with impaired activation of VEGFR-2 and downstream effector Akt (7, 26). As we reiterated in our study, under NG and hypoxic conditions, VEGF stimulation led to a 2.8-fold increase (P = 0.0008) in Akt phosphorylation, whereas under HG exposure, Akt activation was reduced by 56% (P = 0.0068; Figure 1F). In contrast to VEGF-A, Pyr-apelin-13 stimulated Akt phosphorylation by 2-fold (P = 0.0423) and 3-fold (P = 0.0181) under hypoxia and NG or HG conditions, respectively (Figure 1F). Taking together, these results clearly demonstrate that HG conditions caused inhibition of VEGF actions on endothelial cell proangiogenic function, a phenomenon that was not observed under the same conditions following Pyr-apelin-13 stimulation.

Since in vitro experiences in endothelial cells exposed to HG levels indicated that treatment with apelin remained effective to induce proangiogenic actions, we investigated its potential beneficial effects in a diabetic hindlimb ischemia mouse model. Apelin has been shown to influence glucose and lipid metabolism as well as insulin resistance in animal models of type 2 diabetes (27). Here, treatment with apelin osmotic pumps did not influence systemic blood glucose nor affect weight in diabetic mice (Table 2). Because apelin's capacity to improve blood flow reperfusion following hindlimb ischemia in normoglycemic animals has already been reported (19, 22), but never in diabetic animals, we decided to only investigate its actions on diabetic mice. After 2 months of diabetes, femoral artery ligation was performed in nondiabetic and diabetic mice receiving or not systemic delivery of Pyr-apelin-13 with subcutaneous osmotic pumps. Blood flow reperfusion of the lower limb was measured once a week over a period of 4 weeks using laser Doppler imaging (Figure 2A and Supplementary Figure S1A). We selected 3 doses of Pyr-apelin-13 (0.36, 1 and 2 mg/kg/day) to determine which doses would induce the strongest blood flow recovery in diabetic mice. Dose dependent experiments indicated a highest blood flow reperfusion in diabetic mice receiving Pyr-apelin-13 at 1 mg/kg/day (Supplementary Figure S1B). Therefore, this concentration was used for further analysis. Following 28 days post-ligation, diabetic mice exhibited a 36% blood flow recovery compared to 75% in nondiabetic mice (P = 0.0018; Figure 2B). Interestingly, systemic administration of Pyr-apelin-13 in diabetic animals improved blood flow reperfusion to 65% (P = 0.0224; Figure 2B).

A common symptom in patients with PAD is the development of fatigue or pain in the legs leading to reduced walking distance. Thus, we have measured the functional recovery by placing the mice in individual cages equipped with a voluntary exercise wheel, an environment without constraint and human intervention, during 5 consecutive days. Three weeks following limb ischemia, untreated diabetic mice exhibited a significant 90% reduction (P < 0.0001) in running distance as compared to nondiabetic mice (Figure 2C). Interestingly, diabetic mice receiving Pyr-apelin-13 improved by 4.4-fold (P = 0.0076) compared to untreated diabetic mice (Figure 2C). These results suggested an impact of diabetes on the functional recovery of the hindlimb following ischemia, that can be partially recovered by the administration of Pyr-apelin-13 in diabetic mice.

Apelin levels in the plasma were measured using Phoenix Pharmaceutical EIA kit, reported to cross-react with apelin-12, 13 and 36 isoforms. Several studies reported variations in apelin plasma concentrations in the context of diabetes (27, 28). We observed a 42% reduction (P = 0.0026) in apelin plasma levels in diabetic mice in response to limb ischemia compared to nondiabetic mice (Figure 2D). Implantation of Alzet osmotic pumps immediately after limb surgery ensured a continuous systemic delivery of apelin for a period of 28 days, which enhanced plasma apelin concentrations in diabetic mice by 1.8-fold (P < 0.0001) compared to untreated diabetic mice (Figure 2D).

To support the blood flow reperfusion data, we evaluated the ability of Pyr-apelin-13 to promote collateral vessel formation following ischemia by measuring vascular density on cross-sections of the ischemic adductor and calf muscle of all groups of mice. Diabetes reduced the formation of arterioles (vessels <30 μm) and capillaries (CD31-positive vessels <10 μm) in response to tissue ischemia by 36% (P = 0.0007; Figures 3A,C) and 38% (P = 0.0021; Figures 3A,D), respectively, compared to normoglycemic littermate controls. However, delivery of Pyr-apelin-13 in diabetic mice significantly enhanced the amount of arterioles, from 9.71 to 17.00 vessels/mm² (1.8-fold, P < 0.0001; Figures 3A,C), as well as the formation of capillaries, raised from 32.55 to 62.26 (1.9-fold, P < 0.0001; Figures 3A,D), suggesting that Pyr-apelin-13 treatment can promote angiogenesis/arteriogenesis in response to ischemia despite being exposed to diabetes. Hindlimb ischemia causes structural disorganization and degeneration of the muscle fiber (29), and muscle fiber atrophy has been observed in biopsy from patients with PAD (30). We measured myofiber diameter on H & E staining sections of the ischemic muscle as an indicator of muscle fiber atrophy. As a result of persistent ischemia in diabetic mice hindlimb, we observed a 39% reduction (P = 0.0001; Figures 3B,E) in mean myofiber size compared to nondiabetic mice. Interestingly, apelin delivery enhanced mean myofiber diameter by 1.4-fold (P = 0.0143; Figures 3B,E), which was associated with improvement of blood flow reperfusion in diabetic mice.

Apelin and its receptor are known to be upregulated in the muscle and the artery by hypoxia in nondiabetic animal models (21, 22). Gene expression of the apelinergic system is influenced by diabetes in different tissues (27, 31). Following 28 days of femoral artery ligation, we observed that diabetes reduced mRNA expression of both APJ and apelin in the ischemic muscle by 60% (P = 0.0207; Figure 4A) and 67% (P = 0.0055; Figure 4B), respectively. In contrast, Pyr-apelin-13 administration increased APJ mRNA expression by 5-fold (P = 0.0017; Figure 4A) and its own expression by 3.6-fold (P = 0.0055; Figure 4B) in diabetic mice. As we previously reported, many other growth factors are downregulated by diabetes in the ischemic muscle, including VEGF-A, Flk-1, PDGF-B and PDGFR-β (7, 8, 24). Increased capillary density in the ischemic hindlimb of diabetic mice receiving Pyr-apelin-13 led us to investigate whether apelin influenced the expression pattern of well-known proangiogenic growth factors. In our study, Flk-1 (P = 0.0005; Figure 4C), VEGF-A (P = 0.0131; Figure 4D) and PDGF-B (P = 0.0110; Figure 4F) mRNA expression in the ischemic muscle were significantly decreased by diabetes as compared to nondiabetic controls. Our results indicated that the administration of Pyr-apelin-13 significantly increased the gene expression of Flk-1 (P = 0.0247; Figure 4C), VEGF-A (P = 0.0746; Figure 4D), PDGFR-β (P = 0.0367; Figure 4E) and PDGF-B (P = 0.0202; Figure 4F) in diabetic mice.

The signaling pathways by which apelin may promote angiogenesis under hyperglycemia were investigated in the ischemic adductor muscle. The binding of apelin to its receptor APJ leads to the recruitment of the small Gα_i protein which subsequently activates the Akt signaling pathway, promoting cell survival and angiogenesis. As expected, diabetes markedly reduced Akt phosphorylation in the ischemic muscle by 42% (P = 0.0271) compared to nondiabetic mice (Figures 5A,B). In contrast, diabetic mice who received Pyr-apelin-13 exhibited a 2.4-fold (P = 0.0054) increase Akt phosphorylation (Figures 5A,B). APJ activation may also induce the phosphorylation and activation of AMPK via the Gα_i and Gα_q subunits (32). AMPK is mainly known as a fuel-sensing enzyme and for its metabolic function, but it also promotes endothelial cell proliferation and migration (33). Our results showed that apelin delivery raised AMPKα1/2 phosphorylation by 3.1-fold (P = 0.0279; reported to total AMPK) to 3.7-fold (P = 0.0124; reported to GAPDH) (Figures 5A,C,D) and total protein expression by 2.2-fold (P = 0.0369; Figures 5A,E) in the ischemic muscle of diabetic mice compared to untreated diabetic mice. Then, we assessed eNOS as a common downstream target of Akt and AMPK. Once phosphorylated, eNOS produces NO, a major regulator of endothelial cell growth and angiogenesis (34). It has been shown that Ser 633 and Ser1177 of eNOS can be phosphorylated by AMPK while Akt only phosphorylates eNOS at Ser1177 (35, 36). As expected, diabetes reduced Ser1177 phosphorylation of eNOS by 57% (P = 0.0257; reported to GAPDH) and 70% (P = 0.0448; reported to total eNOS) (Figures 5A,F,G) compared to normoglycemic animals but had little impact on eNOS Ser633 activation. Interestingly, the administration of Pyr-apelin-13 in diabetic mice increased both Ser1177 phosphorylation of eNOS by 3.7-fold (P = 0.0010; reported to GAPDH) or 4.3-fold (P = 0.0224; reported to total eNOS) (Figures 5A,F,G) and Ser633 by 2.5-fold (P = 0.0168; reported to GAPDH) or 2.1-fold (P = 0.0058; reported to total eNOS) (Figures 5A,H,I). Furthermore, apelin delivery in diabetic mice elevated total eNOS protein expression in the ischemic muscle by 2.7-fold (P = 0.0244) compared to untreated diabetic mice (Figures 5A,J). We also explored the RhoA/ROCK signaling pathway, which is known to be recruited by the small Gα_12/13 protein following GPCRs activation (37), and can mediate cellular morphological processes such as cell contractility, actin cytoskeleton organization, cell migration and proliferation (38). Although no difference in ROCK-2 phosphorylation and protein expression was observed in the ischemic muscle of diabetic mice compared to nondiabetic mice, apelin administration raised ROCK2 phosphorylation by 2.2-fold (P = 0.0196; Figure 6A) and ROCK-2 total protein expression by 10.7-fold (P = 0.0292; Figure 6B) compared to untreated diabetic mice. We did not observe any difference in phospho-ROCK2 protein expression when reported on ROCK2 expression (Supplementary Figure S2A), due to significant increased expression of ROCK-2 in diabetic mice receiving Pyr-apelin-13.

To confirm that the signaling pathways enhanced by the treatment with Pyr-apelin-13 in ischemic muscle occurred in endothelial cells, we investigated their activation in cultured endothelial cells exposed to hypoxia and NG or HG conditions. Either under NG and HG exposure, 1 h stimulation with Pyr-apelin-13 similarly raised AMPKα1/2 (Figure 7A) and eNOS phosphorylation at Ser1177 (Figure 7B) and Ser633 (Figure 7C). In addition, eNOS mRNA expression in endothelial cells was increased by 2.6-fold (P = 0.0008) and 1.8-fold (P = 0.0410) following 24 h stimulation of apelin under NG and HG conditions, respectively (Figure 7D). As for the in vivo results, we observed a very similar pattern for RhoA/ROCK2 signaling pathway, apelin treatment increased ROCK-2 phosphorylation (Figure 7E) and total protein expression (Figure 7F) as well as RhoA mRNA expression (Figure 7G) in both NG and HG conditions. However, no significant difference in ROCK-2 phosphorylation following apelin stimulation when reported on ROCK-2 protein expression was noted (Supplementary Figure S2B). These data demonstrate that apelin stimulation, through APJ receptor, promotes signaling pathways endothelial cells that are not affected by hyperglycemia.