Ongoing efforts to remove pathological inflammatory stimuli are crucial for the protection of endothelial cells in diabetes. Nerve injury-induced protein 1 (Ninj1) is an adhesion molecule that not only contributes to inflammation but also regulates the apoptosis of endothelial cells. In the present study, Ninj1 was found highly expressed in endothelial cells in Type 2 diabetic mice and increased in high-glucose (HG) cultured HUVECs. Furthermore, we found that Ninj1 levels are up-regulated in endothelial cells in clinical specimens of diabetic patients when compared with nondiabetic tissues, indicating a biological correlation between Ninj1 and endothelial pathophysiology in diabetic condition. Functional blocking of Ninj1 promoted endothelial tube formation and eNOS phosphorylation in the HG condition. Additionally, blocking Ninj1 inhibited the activation of caspase-3 and increased the Bcl-2/Bax ratio, thus inhibiting HUVECs apoptosis induced by HG. HG-induced ROS overproduction, p38 MAPK and NF-κB activation, and the overexpression of VCAM-1, ICAM-1, MCP-1, and IL-6 genes were ameliorated after Ninj1 was blocked. Using the signaling pathway inhibitor LY294002, we found that Bcl-2 expression and eNOS phosphorylation after Ninj1 blockade were regulated via PI3K/Akt signaling pathway. The in vivo endothelial contents, α-SMA+PECAM-1+ vascular numbers, and blood perfusion in the hindlimb were markedly up-regulated after Ninj1 was blocked. According to our findings, functional blocking of Ninj1 shows protective effects on diabetic endothelial cells both in vitro and in vivo. Thus, we consider Ninj1 to be a potential therapeutic target for preventing endothelial dysfunction in diabetes mellitus.

Background

Diabetes mellitus (DM) leads to numerous hyperglycemia-related crippling or fatal vascular complications, including cardiovascular disease, peripheral vascular disease [13], and microvascular damage [4,5]. Hyperglycemia-induced endothelial dysfunction is the harbinger and initiating factor of diabetic vascular complications in DM patients [1,6,7]. Endothelial cells cultured in high-glucose (HG) conditions in vitro were observed to have delayed reproduction and excessive cell apoptosis [811]. In vivo animal studies have demonstrated that DM can cause damage to endothelial cells and vascular endothelium [1012].

Signaling pathways involving endothelial dysfunction are complicated, and the most important pathways include eNOS activation-related PI3K/Akt pathway and inflammation-related JNK and MAPK pathway, among others [1,8]. Till now, the prevention and treatment of diabetic endothelial dysfunction remain a challenge due to the complexity of the pathogenesis. The efficiency of preventing vascular complications through intensive lowering of blood glucose is still the subject of debate [13,14]. Though much interest has been directed toward the role of inflammation in DM [1517], treatment of single inflammatory factors makes it difficult to protect the function of endothelial cells. Thus, further identification of upstream factors that mediate endothelial dysfunction in DM is crucial and might reveal new therapeutic programs.

A novel candidate involved in HG-induced endothelial dysfunction is an adhesion molecule called nerve injury-induced protein 1 (Ninjurin1, Ninj1). Ninj1 is initially identified in dorsal root ganglion (DRG) neurons and Schwann cells after nerve injury [19,20] and is also basally expressed on macrophages and endothelial cells [21,22]. Recently, more studies have focused on the role of Ninj1 outside the nervous system [2130]. It is involved in inflammatory diseases [2123], cancer progress [2427], and erectile dysfunction in mice [28,29]. Jennewein et al. [23] study showed that Ninj1 contributes to inflammation by mediating leukocyte migration and modulating inflammation mediators, while blocking Ninj1 attenuated LPS-triggered inflammation by modulating p38 MAPK/AP-1 activation. In another study, down-regulation of Ninj1 in capillary pericytes (cPCs) strengthened cPCs-mediated angiogenic effects, whereas overexpression of Ninj1 decreased this effect [31,32]. What’s more interesting is that target blocking of Ninj1 in Type 1 diabetic mice restores erectile function through dual angiogenic and neurotrophic effects [29]. However, the role of Ninj1 in endothelial cells of Type 2 diabetes and relating peripheral vascular complication is still poorly described and need further study.

In the present study, we hypothesized that Ninj1 contributes to endothelial dysfunction in DB. We demonstrate that up-regulated Ninj1 is involved in HG-induced endothelial dysfunction both in vitro and in vivo. Ninj1 inhibition with blocking antibody hNinj123–37in vitro ameliorated HG-induced inflammation by suppressing ROS production, p38 MAPK/NF-κB activation, and their downstream proinflammatory genes. In addition, blocking Ninj1 inhibited caspase-3 activation and increased Bcl-2 expression, thus inhibiting apoptosis induced by HG. Moreover, activation of PI3K/Akt signaling pathway after Ninj1 was blocked restored tube formation capacity of endothelial cells under HG condition. Finally, in vivo administration of mNinj126–37 protected vascular numbers and endothelium integrity and blood perfusion in Type 2 DM mice, at least partially through a direct effect of Ninj1 blocking antibody on endothelial cells, suggesting that Ninj1 is a potential therapeutic target in protecting endothelial cell function in diabetes.

Research design and methods

Human tissue specimens and preservation

Diabetic muscle specimens were obtained from 11 Type 2 diabetic patients undergoing lower limb amputation due to ischemic diabetic foot complications. Nondiabetic specimens were normal muscle tissues which adjacent to ischemic muscles that were obtained from eight nondiabetic patients undergoing lower limb amputation due to acute arterial embolism induced by shedding of atrial fibrillation thrombosis. Specimens were snap-frozen in liquid nitrogen and stored at −80 °C until the preparation of protein extracts or formalin-fixing. The present study was approved by the Ethics Committee of Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine and carried out according to the recommendations of the Declaration of Helsinki. Written informed consent was obtained from each patient involved in the present study.

Animals

Four - totwelve-week-old male Type 2 diabetic mice C57BL/KsJ-db/db (db/db) and C57BL/KsJ-db/+ (db/+) were purchased from the Shanghai Research Center for Model Organisms (Shanghai, China) and housed under specific pathogen-free conditions. All animal experiments were approved by the Animal Ethics Committee of Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine.

Ninj1 blocking antibody

The Ninj1 blocking antibody was designed as previously reported [21,23] with some modifications. The functional blocking antibody of mouse Ninj1 (amino acids 26–37:PPRWGLRNRPIN; mNinj126–37), human Ninj1 (amino acids 23–37: DASPARWGWRHGPIN; hNinj123–37; three amino acid sequences were extended to increase the hydrophilicity of blocking antibody), and a control antibody (amino acids: PPRAGLRNRPIN; Ninj1ctr) were synthesized from the antibody service of WuXi AppTec (Shanghai, China).

Cell culture and high glucose treatment

Primary Human umbilical vein endothelial cells (HUVECs) with a certificate of authenticity were purchased from the ScienCell Research Laboratories (Carlsbad, CA, U.S.A.) and similarly cultured as previously described [8]. Cells were seeded in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 5% fetal bovine serum (Gibco, Waltham, MA, U.S.A.), 5 ng/ml VEGF and bFGF (R&D, Minneapolis, MN, U.S.A.), and 1% antibiotic/antimycotic solution (Gibco). The cultures were maintained at 37°C in a humidified 5% CO2 atmosphere. Cells were treated with either 5.5 mmol/l D-glucose (normal glucose, NG, Sigma-Aldrich, St. Louis, MO, U.S.A.) or 33.3 mmol/l D-glucose (high glucose, HG), or 5.5 mmol/l D-glucose plus 27.8 mmol/l L-glucose (L-Glu, used as an osmotic control, Sigma-Aldrich) for indicated times and subsequently used for further experimental studies. In some experiments, the Ninj1 blocking antibody and/or isoform nonselective PI3K/Akt inhibitor LY294002 (10 μM, Selleck, TX, U.S.A.) was used.

Detection of Ninj1 expression with immunofluorescence

For immunofluorescence, muscle tissue was fixed in 4% paraformaldehyde for 24 h at 4°C, embedded in optimal cutting temperature (OCT) compound and snap-frozen, and then cut into 10 μm thick serial tissue sections. Then, they were incubated with mouse monoclonal antibody (mAb) against ninjurin1 (Ninj1, BD biosciences, San Diego, CA, U.S.A.) and rabbit mAb against PECAM-1 (Abcam, Cambridge, MA, U.S.A.) overnight. HUVECs were also fixed and incubated overnight with anti-Ninj1 antibody at 4°C. The antibody concentration used was listed in Supplementary Table S1. The secondary antibody was applied for 1 h at room temperature. For control purposes, negative staining was also performed with secondary antibody alone to exclude a possibility of nonspecific staining. Both tissues and cells were examined with inverted fluorescence microscopy (Olympus IX81, Tokyo, Japan).

Tube formation assay

HG cultured HUVECs with or without Ninj1 blocking antibody were plated on Matrigel (Corning, NY, U.S.A.) in 24-well plates at 1 × 104 cells per well. NG and L-glucose cultured HUVECs were used as controls. Tube formation was photographed through an optical microscope 16 h later. Randomly chosen pictures were analyzed with Angiogenesis Analyzer for ImageJ.

Cell proliferation analysis

HUVECs were cultured in either HG or NG conditions for 7 days with or without the addition of Ninj1 blocking antibody or control antibodies. The cells were normally passaged, the concentration of Ninj1 blocking antibody was maintained as described above, and the total number of cell numbers was counted. The proliferating cells were identified by an EdU (5-ethynyl-2′-deoxyuridine) Imaging Kit (Thermo Fisher Scientific, Waltham, MA, U.S.A.) according to the manufacturer’s instructions and examined with inverted fluorescence microscopy (Olympus IX81, Tokyo, Japan).

Apoptosis analysis

HUVECs were seeded at 1 × 105 cells per well in six-well plates in HG conditions for 48 h with or without low-dose or high-dose Ninj1 blocking antibody or control antibodies. The cells were collected by trypsinization, washed twice with cold PBS, and then centrifuged. After the supernatants were discarded, the cells were resuspended in 1× annexin-binding buffer and stained with FITC-Annexin V and propidium iodide (PI, BD biosciences, San Diego, CA, U.S.A.). The apoptosis rate was detected by flow cytometry (Becton-Dickinson Company, U.S.A.).

Detection of ROS production

The intracellular ROS levels of NG- and HG-cultured HUVECs, with or without Ninj1 blocking antibody treatment, were determined by labeling with 2′,7′-dichlorofluorescin diacetate (DCFH-DA) using the Reactive Oxygen Species Assay Kit (Beyotime, Shanghai, China) according to the manufacturer’s instructions. Briefly, adherent HUVECs were incubated with DCFH-DA at a final concentration of 5 mM for 20 min at 37°C and washed three times with PBS buffer to remove free DCFH-DA. The ROS levels in HUVECs were analyzed immediately with inverted fluorescence microscopy (Olympus IX81, Tokyo, Japan).

RNA isolation and real-time quantitative PCR

Several most studied proinflammatory genes in diabetic endothelial dysfunction and diabetic complications were chosen to be investigated. As previously described [33], total RNA was extracted from HUVECs using Trizol reagent (Invitrogen, Carlsbad, CA, U.S.A.), and real-time quantitative PCR (qPCR) analysis was performed with a Mx3000PTM QPCR System (Stratagene, CA, U.S.A.) using Power SYBR Green PCR master mix (2×) (Applied Biosystems, CA, U.S.A.) with 1.5 μg of RNA to determine gene expression levels. The primers for qPCR analysis were as follows: MCP-1 sense, 5′-CAGCCAGATGCAATCAATGCC-3′, and anti-sense 5′-TGGAATCCTGAACCCACTTCT-3′; IL-6 sense, 5′-CCTGAACCTTCCAAAGATGGC-3′, and anti-sense 5′-TTCACCAGGCAAGTCTCCTCA-3′; ICAM-1 sense, 5′-AAGTTGTTGGGCATAGAGAC-3′ and anti-sense 5′-CATCAGGGCAGTTTGAATA-3′; VCAM-1 sense, 5′-CATCCCTACCATTGAAGATAC-3′ and anti-sense 5′-GACATAAAGTGTTTGCGTACTC- 3′. The experiments were repeated at least in triplicate.

Western blot

Total protein was isolated from HUVECs and from human specimens and from hindlimb muscle tissues of both db/db and db/+ mice as previously described [29]. The protein in tissue samples or cell lysates was quantified, electrophoresed, and transferred to PVDF membranes and then incubated overnight at 4°C with appropriate antibodies as follows: mouse monoclonal anti-Ninj1, rabbit monoclonal anti-PECAM-1, eNOS, p-eNOS, active-caspase-3, Bax, rabbit polyclonal anti-Bcl-2 (Abcam, Cambridge, U.K.), rabbit monoclonal p38 MAPK, p-MAPK, anti-NF-κB p65, p-NF-κB p65, PI3K, p-PI3K p85 (Tyr458)/p55(Tyr199), Akt, p-Akt (Ser473), and β-actin (CST, Danvers, MA, U.S.A.). The antibody concentration used was listed in Supplementary Table S1. HRP-conjugated secondary antibody and a chemiluminescence kit (Roche, Basel, Switzerland) were used to detect signals, and then, images were acquired using an LAS3000 machine (GE Healthcare Life Sciences).

Animal studies

Four-week-old mice were divided into three groups: age-matched db/+ controls (n=6), db/db mice with right hindlimbs receiving i.m. injection of Ninj1ctr (1 mg/kg/BW, n=6) once a week for 8 weeks, and db/db mice with right hindlimbs receiving i.m. injection of mNinj126-37 (1 mg/kg/BW, n=6) once a week for 8 weeks.

Laser doppler imaging analysis

As described previously [34], Doppler imaging analysis was performed with a laser Doppler perfusion imager (moorFLPI; Moor Instruments, Devon, U.K.) to noninvasively evaluate the blood flow in the hindlimbs. The color-coded digital images of the region from the toe to the knee of the right hindlimbs were analyzed. Regional blood flow values from db/+ mice were used as baseline values. The relative blood perfusion ratio of the mNinj126–37- or Ninj1ctr-treated right hindlimb of db/db mice was normalized with that of db/+ mice.

Statistical analysis

Data from all experiments are reported as the mean ± standard deviation. Statistical significance (P<0.05) was determined using Student’s t-test or one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test, as appropriate. All experiments were performed with at least three replicates per group.

Results

Ninj1 protein was up-regulated in the endothelial cells of the hindlimb muscle of db/db mice and the primary HUVECs exposed to the HG condition

Immunofluorescence analysis revealed that the expression of Ninj1 protein was significantly higher in the hindlimb muscles of the 8-week-old db/db mice than in the age-matched db/+ mice (Figure 1A and B). The contents of endothelial cells were significantly lower in the hindlimb muscles of db/db mice than in controls (Figure 1A and B). Nonspecific staining of both Ninj1 and PECAM-1 was excluded by control staining (Supplementary Figure S1A). Notably, Ninj1 protein was expressed not only in endothelial cells but also in some other cell types that require further identification. Western blot analysis showed that Ninj1 protein gradually increased with time, especially from 6 to 12 weeks, whereas the endothelial content in the hindlimb muscles of db/db mice was significantly decreased from 6 weeks and was negatively correlated with Ninj1 expression (Figure 1C–E). Basic data of the mice, which were displayed as mean and standard deviation analyzed by Student’s t-test, are listed in Supplementary Table S2.

Ninj1 was increased in endothelial cells in Type 2 diabetic mice and in in vitro cultured HUVECs exposed to HG condition

Figure 1
Ninj1 was increased in endothelial cells in Type 2 diabetic mice and in in vitro cultured HUVECs exposed to HG condition

(A) Representative PECAM-1 (green) and Ninj1 (red) double-immunostaining in hindlimb muscle tissue from 8-week-old db/+ control and db/db mice. Nuclei were labeled with DAPI (blue); scale bar = 100 μm. (B) The Ninj1 positive area or PECAM-1 positive area in hindlimb muscle tissue was quantified by ImageJ (n=8); ****P<0.0001 compared with db/+ group. (C) Ninj1 and PECAM-1 protein expression in hindlimb muscle tissue of 12-week-old db/+ mice and 4–12-week-old db/db mice was measured by Western blot. (D and E) Quantitative analysis of Ninj1 and PECAM-1 protein based on normalized band intensity values by ImageJ (n=4); ****P<0.0001, ***P<0.001, **P<0.01, and *P<0.05. (F) Representative Ninj1 (green) immunostaining in HUVEC. Nuclei were labeled with DAPI (blue). Cells were treated with normal glucose (NG, 5.6 mmol/l) or high glucose (HG, 33.6 mmol/l) for 24 h, scale bar: 25 μm. (G) Relative Ninj1 fluorescence intensity was semiquantified by ImageJ (n=8); **P<0.01 compared with NG group. (H) HUVECs were treated with HG for 0, 12, 24, 48, 96, or 192 h. Ninj1 protein expression was measured by Western blot and quantified by ImageJ based on normalized band intensity values (n=4); ****P<0.0001 and ***P<0.001.

Figure 1
Ninj1 was increased in endothelial cells in Type 2 diabetic mice and in in vitro cultured HUVECs exposed to HG condition

(A) Representative PECAM-1 (green) and Ninj1 (red) double-immunostaining in hindlimb muscle tissue from 8-week-old db/+ control and db/db mice. Nuclei were labeled with DAPI (blue); scale bar = 100 μm. (B) The Ninj1 positive area or PECAM-1 positive area in hindlimb muscle tissue was quantified by ImageJ (n=8); ****P<0.0001 compared with db/+ group. (C) Ninj1 and PECAM-1 protein expression in hindlimb muscle tissue of 12-week-old db/+ mice and 4–12-week-old db/db mice was measured by Western blot. (D and E) Quantitative analysis of Ninj1 and PECAM-1 protein based on normalized band intensity values by ImageJ (n=4); ****P<0.0001, ***P<0.001, **P<0.01, and *P<0.05. (F) Representative Ninj1 (green) immunostaining in HUVEC. Nuclei were labeled with DAPI (blue). Cells were treated with normal glucose (NG, 5.6 mmol/l) or high glucose (HG, 33.6 mmol/l) for 24 h, scale bar: 25 μm. (G) Relative Ninj1 fluorescence intensity was semiquantified by ImageJ (n=8); **P<0.01 compared with NG group. (H) HUVECs were treated with HG for 0, 12, 24, 48, 96, or 192 h. Ninj1 protein expression was measured by Western blot and quantified by ImageJ based on normalized band intensity values (n=4); ****P<0.0001 and ***P<0.001.

In the in vitro cultured HUVECs exposed to HG conditions for 24 h, significantly higher expression of Ninj1 protein was also found by immunofluorescence staining (Figure 1F and G). Nonspecific staining was excluded by control staining (Supplementary Figure S1B). Western blot analysis revealed that the increased expression of Ninj1 in HUVECs exposed to HG conditions was time-dependent before 48 h (Figure 1H). Our findings are consistent with previous studies [28,29]. These previous studies and our results were the basis for us to explore the relationship between Ninj1 expression and endothelial cell function in DM.

Ninj1 expression was also up-regulated in the endothelial cells of human specimens obtained from Type 2 diabetic patients

Basic data of the patients, including fasting blood glucose, blood pressure, cholesterol, triglyceride, low-density lipoprotein, high-density lipoprotein, and free fatty acid, which were displayed as mean and standard deviation analyzed by Student’s t-test, are listed in Supplementary Table S3. The fasting blood glucose of Type 2 diabetic patients was significantly higher than nondiabetic patients, while other data (including blood pressure, cholesterol, triglyceride, low-density lipoprotein, high-density lipoprotein, and free fatty acid) showed no significant difference between the two groups. The immunofluorescence analysis revealed that the expression of Ninj1 protein was also significantly higher in the endothelial cells of clinical specimens obtained from Type 2 diabetic patients as compared with nondiabetic cases (Figure 2A–G) and control staining (Supplementary Figure S2), indicating there may exist an interaction between Ninj1 expression and endothelial function in diabetes. Additional results of total levels of Ninj1 in muscle specimens was also significantly higher in Type 2 diabetic patients than nondiabetic patients which confirmed by Western blot (Figure 2H and I).

Proof of high level Ninj1 expression in endothelial cells of clinical specimens obtained from Type 2 diabetic patients

Figure 2
Proof of high level Ninj1 expression in endothelial cells of clinical specimens obtained from Type 2 diabetic patients

(AC) Representative PECAM-1 (green) and Ninj1 (red) double-immunostaining in lower extremity muscle tissue from Type 2 diabetic patients. Nuclei were labeled with DAPI (blue); scale bar = 100 μm. (DF) Representative PECAM-1 (green) and Ninj1 (red) double-immunostaining in lower extremity muscle tissue from nondiabetic patients. Nuclei were labeled with DAPI (blue); scale bar = 100 μm. (G) The Ninj1 positive and PECAM-1 positive area were quantified by ImageJ; ***P<0.001. (H) Ninj1 protein expression in lower extremity muscle tissue from Type 2 diabetic and nondiabetic patients was measured by Western blot. (I) Quantitative analysis of Ninj1 protein based on normalized band intensity values by ImageJ; *P<0.05.

Figure 2
Proof of high level Ninj1 expression in endothelial cells of clinical specimens obtained from Type 2 diabetic patients

(AC) Representative PECAM-1 (green) and Ninj1 (red) double-immunostaining in lower extremity muscle tissue from Type 2 diabetic patients. Nuclei were labeled with DAPI (blue); scale bar = 100 μm. (DF) Representative PECAM-1 (green) and Ninj1 (red) double-immunostaining in lower extremity muscle tissue from nondiabetic patients. Nuclei were labeled with DAPI (blue); scale bar = 100 μm. (G) The Ninj1 positive and PECAM-1 positive area were quantified by ImageJ; ***P<0.001. (H) Ninj1 protein expression in lower extremity muscle tissue from Type 2 diabetic and nondiabetic patients was measured by Western blot. (I) Quantitative analysis of Ninj1 protein based on normalized band intensity values by ImageJ; *P<0.05.

Blocking Ninj1 restores HUVEC tube formation function and induces eNOS phosphorylation

We next sought to detect the effects of Ninj1 blocking on tube formation of HUVECs treated with HG. Compared with NG-cultured HUVECs and osmotic controls, HG-treated HUVECs showed decreased mesh numbers and tube length on Matrigel (Figure 3A–C). Treatment with low-dose (3 μM) or high-dose (10 μM) hNinj123–37 significantly reversed HG-induced impairment of HUVEC tube forming ability almost to a level comparable to that in the control groups. We further detected the activation and expression of eNOS, a known enzyme that has a protective function in the cardiovascular system. Compared with NG-cultured HUVECs and osmotic controls, the activation and expression of eNOS in HUVECs was not significantly changed with HG treatment with or without Ninj1ctr. As the administration of low-dose or high-dose hNinj123–37 significantly increased the phosphorylation of eNOS in HUVECs by nearly 1.79 ± 0.21 and 1.76 ± 0.1-fold respectively (Figure 3D and E), we concluded that hNinj123–37 exerts positive effects on tube formation of endothelial cells.

Functional blocking of Ninj1 restores HUVEC tube formation function and promotes eNOS phosphorylation

Figure 3
Functional blocking of Ninj1 restores HUVEC tube formation function and promotes eNOS phosphorylation

(A) HUVECs were precultured HG conditions for 24 h and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 for the times indicated. After treatment, HUVECs were plated on Matrigel, and the tube formation assay was performed. NG-cultured and L-Glu-treated HUVECs served as positive and osmotic controls respectively; scale bar: 200 μm. (B and C) Quantification of mesh numbers (number of meshes per area) (B) and relative tube length (compared with NG group) (C) analyzed by an Angiogenesis Analyzer for ImageJ (n=8); ****P<0.0001, ***P<0.001, **P<0.01, and *P<0.05. (D) eNOS expression and phosphorylation in HUVECs were measured by Western blot. (E) Relative ratio of phos-eNOS and eNOS based on normalized band intensity values processed by ImageJ (n=4); **P<0.01.

Figure 3
Functional blocking of Ninj1 restores HUVEC tube formation function and promotes eNOS phosphorylation

(A) HUVECs were precultured HG conditions for 24 h and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 for the times indicated. After treatment, HUVECs were plated on Matrigel, and the tube formation assay was performed. NG-cultured and L-Glu-treated HUVECs served as positive and osmotic controls respectively; scale bar: 200 μm. (B and C) Quantification of mesh numbers (number of meshes per area) (B) and relative tube length (compared with NG group) (C) analyzed by an Angiogenesis Analyzer for ImageJ (n=8); ****P<0.0001, ***P<0.001, **P<0.01, and *P<0.05. (D) eNOS expression and phosphorylation in HUVECs were measured by Western blot. (E) Relative ratio of phos-eNOS and eNOS based on normalized band intensity values processed by ImageJ (n=4); **P<0.01.

Functional blocking of Ninj1 promotes endothelial cell survival rather than proliferation under HG condition

We further studied the effect of blocking Ninj1 on cell proliferation. The proliferation of HUVECs exposed to HG for 48 h was slower than that of NG-cultured endothelial cells. The number of HG-cultured HUVECs was only 69.78 ± 8.1% of NG on day 7 (Figure 4A). We found that Ninj1ctr had no effect on the proliferation of either NG- or HG-cultured HUVECs (data not shown). However, Ninj1 blocking antibody maintained the total number and the EdU+ percentage of endothelial cells in the HG environment, though they were still lower than NG-cultured HUVECs (Figure 4B). Interestingly, Ninj1 blocking had no effect on the proliferation of NG-cultured HUVECs, probably due to the low expression of Ninj1 in the NG environment. The expression of Ang1, which was reported to mediate endothelial cell proliferation after blocking Ninj1 [29], was not regulated by culture condition or the addition of Ninj1 blocking antibody (Figure 4C) or Ninj1ctr (data not shown). Collectively, we hypothesized that functional blocking of Ninj1 could promote cell survival instead of proliferation in an HG environment.

Functional blocking of Ninj1 promotes HUVEC survival and ameliorates apoptosis

Figure 4
Functional blocking of Ninj1 promotes HUVEC survival and ameliorates apoptosis

(AC) HUVEC proliferation. (A) HUVECs were cultured with NG (5.6 mmol/l) or HG (33.6 mmol/l) and treated with or without 3 or 10 μM for 7 days. The cells are continuously passaged with numbers counted (n=4); ***P<0.001 and **P<0.01. (B) Quantification of EdU-positive (represents reproductive cells) HUVECs after culture with NG or HG and treatment with or without 3 or 10 μM for 24–48 h (n=12). The results from 48 h are shown; ****P<0.0001, ***P<0.001, and **P<0.01. (C) Ang1 protein expression in HUVECs measured by Western blotting. Cells were cultured with NG or HG and treated with or without 3 or 10 μM for 24–48 h (n=4). The results from 48 h are shown. No significance was found among groups. (D and E) HUVEC apoptosis. (D) Cells were precultured in HG for 24–48 h and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 for the times indicated. NG-cultured and L-Glu-treated HUVECs served as negative and osmotic controls respectively. After treatment, HUVECs were collected and stained with PE-Annexin V and PI and measured by flow cytometry. The results from 48 h are shown. (E) Quantification of apoptotic HUVEC percentage (n=6); **P<0.01 and *P<0.05. (F) Protein expression of apoptosis-related proteins cleaved caspase-3, Bcl-2, and Bax in HUVECs were measured by Western blot (n=4). (G and H) Quantitative analysis of cleaved caspase-3 protein (G) and relative ratio of Bcl-2 and Bax (H) based on normalized band intensity values processed by ImageJ (n=4); ****P<0.0001, ***P<0.001, and **P<0.01.

Figure 4
Functional blocking of Ninj1 promotes HUVEC survival and ameliorates apoptosis

(AC) HUVEC proliferation. (A) HUVECs were cultured with NG (5.6 mmol/l) or HG (33.6 mmol/l) and treated with or without 3 or 10 μM for 7 days. The cells are continuously passaged with numbers counted (n=4); ***P<0.001 and **P<0.01. (B) Quantification of EdU-positive (represents reproductive cells) HUVECs after culture with NG or HG and treatment with or without 3 or 10 μM for 24–48 h (n=12). The results from 48 h are shown; ****P<0.0001, ***P<0.001, and **P<0.01. (C) Ang1 protein expression in HUVECs measured by Western blotting. Cells were cultured with NG or HG and treated with or without 3 or 10 μM for 24–48 h (n=4). The results from 48 h are shown. No significance was found among groups. (D and E) HUVEC apoptosis. (D) Cells were precultured in HG for 24–48 h and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 for the times indicated. NG-cultured and L-Glu-treated HUVECs served as negative and osmotic controls respectively. After treatment, HUVECs were collected and stained with PE-Annexin V and PI and measured by flow cytometry. The results from 48 h are shown. (E) Quantification of apoptotic HUVEC percentage (n=6); **P<0.01 and *P<0.05. (F) Protein expression of apoptosis-related proteins cleaved caspase-3, Bcl-2, and Bax in HUVECs were measured by Western blot (n=4). (G and H) Quantitative analysis of cleaved caspase-3 protein (G) and relative ratio of Bcl-2 and Bax (H) based on normalized band intensity values processed by ImageJ (n=4); ****P<0.0001, ***P<0.001, and **P<0.01.

Functional blocking of Ninj1 ameliorates endothelial cells apoptosis via suppressing caspase-3 activation and promoting Bcl-2 expression

To discover reasons for the decreased HUVEC survival and to explore the biological relevance of Ninj1 blocking antibody, flow cytometry was performed to detect apoptosis of HG-treated HUVECs. We found that HG treatment with or without Ninj1ctr for 48 h triggered significantly higher apoptosis compared with control cells (Figure 4D and E). Of note, treatment with both low-dose and high-dose hNinj123–37 remarkably prevented HG-induced HUVEC apoptosis, especially high-dose hNinj123–37 (Figure 4D and E).

We further detected the expression of cleaved caspase-3, Bcl-2, and Bax using Western blot. The results showed that the cleaved caspase-3 was significantly higher in the HG-treated HUVECs than in NG-cultured controls, and both low-dose and high-dose hNinj123–37 significantly suppressed HG-induced caspase-3 activation (Figure 4F and G). What’s more, the expression of antiapoptotic protein Bcl-2 was remarkably up-regulated in both low-dose and high-dose hNinj123–37 treated groups, and the Bcl-2/Bax ratio was significantly increased by nearly 3.45 ± 0.74 and 3.85±0.62-fold respectively, compared with the NG group (Figure 4F and H).

Ninj1 blockade suppresses HG-induced inflammation by modulating ROS, p38 MAPK, and NF-κB

Endothelial dysfunction was associated with increased oxidative stress and proinflammatory mediators, especially in Type 2 DM. We found that HG significantly increased the intracellular generation of ROS in HUVECs by nearly 3-fold compared with NG controls (Figure 5A and C). Both low-dose and high-dose hNinj123–37 treatment significantly suppressed the production of ROS in HG-treated HUVECs (Figure 5A and C). Compared with NG-cultured HUVECs, HG treatment with or without Ninj1ctr was associated with a significant up-regulation of gene expression of ICAM-1, VCAM-1, IL-6, and MCP-1 (Figure 5B). These changes were significantly suppressed when the hNinj123-37 antibody, especially at a high-dose (10 μM), was administered.

Functional blocking of Ninj1 suppresses ROS production and inflammatory mediator expression in the HG condition

Figure 5
Functional blocking of Ninj1 suppresses ROS production and inflammatory mediator expression in the HG condition

(A) Representative microscopy images of DCFH-DA fluorescence in HG-cultured HUVECs after treatment for 24 h with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37. NG-cultured and L-Glu-treated HUVECs served as negative and osmotic controls respectively (n=6); scale bar: 100 μm. (B) qPCR for VCAM-1, ICAM-1, IL-6, and MCP-1 in HUVECs cultured in HG and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 for 24 h (n=6). NG-cultured and L-Glu-treated HUVECs served as negative and osmotic controls respectively; ****P<0.0001, ***P<0.001, **P<0.01, and *P<0.05. (C) Relative ROS fluorescence intensity (compared with the NG group) analyzed by ImageJ (n=6); ****P<0.0001 and ***P<0.001. (D) Protein expression of phos-NF-κB p65, NF-κB p65, phos-p38 MAPK, and p38 MAPK in HUVECs were measured by Western blot (n=4). (E and F) Relative ratio of phos-NF-κB p65 and NF-κB p65 (E) and relative ratio of phos-p38 MAPK and p38 MAPK (F) based on normalized band intensity values processed by ImageJ (n=4); ****P<0.0001, **P<0.01, and *P<0.05.

Figure 5
Functional blocking of Ninj1 suppresses ROS production and inflammatory mediator expression in the HG condition

(A) Representative microscopy images of DCFH-DA fluorescence in HG-cultured HUVECs after treatment for 24 h with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37. NG-cultured and L-Glu-treated HUVECs served as negative and osmotic controls respectively (n=6); scale bar: 100 μm. (B) qPCR for VCAM-1, ICAM-1, IL-6, and MCP-1 in HUVECs cultured in HG and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 for 24 h (n=6). NG-cultured and L-Glu-treated HUVECs served as negative and osmotic controls respectively; ****P<0.0001, ***P<0.001, **P<0.01, and *P<0.05. (C) Relative ROS fluorescence intensity (compared with the NG group) analyzed by ImageJ (n=6); ****P<0.0001 and ***P<0.001. (D) Protein expression of phos-NF-κB p65, NF-κB p65, phos-p38 MAPK, and p38 MAPK in HUVECs were measured by Western blot (n=4). (E and F) Relative ratio of phos-NF-κB p65 and NF-κB p65 (E) and relative ratio of phos-p38 MAPK and p38 MAPK (F) based on normalized band intensity values processed by ImageJ (n=4); ****P<0.0001, **P<0.01, and *P<0.05.

We then assessed the effects of hNinj123–37 on those signaling pathways related to ROS production and the expression of aforementioned inflammatory markers. Western blot analysis showed that p38 MAPK and NF-κB were activated 2.68 ± 0.17 and 2.06 ± 0.42-fold respectively, in HG-treated HUVECs compared with the NG group (Figure 5D–F). The increased activation of p38 MAPK and NF-κB was significantly suppressed by both low-dose and high-dose hNinj123–37 (Figure 5D–F), indicating that HG-induced p38 MAPK and NF-κB activation was mediated by Ninj1. Thus, blocking the Ninj1 attenuation of HG-induced endothelial inflammation is most likely due to negatively modulating p38 MAPK and NF-κB pathways.

Functional blocking of Ninj1 protects endothelial function and survival via activating the PI3K/Akt signaling pathway

As mentioned above, blocking Ninj1 increased the activation of eNOS and promoted the expression of Bcl-2. We then explored the upstream pathway that modulates these proteins, especially PI3K/Akt signaling. Western blot analysis showed no significant difference in total PI3K or total Akt expression among all groups. HG treatment with or without Ninj1ctr did not induce the phosphorylation of PI3K or Akt when compared with either the NG control or the osmotic control groups (Figure 6A–C). However, both low-dose and high-dose Ninj1 blocking antibody hNinj123–37 remarkably induced the phosphorylation of PI3K and Akt in the HG-treated HUVECs (Figure 6A–C). To verify whether the protective effect of blocking Ninj1 on endothelial cells was through the PI3K/Akt pathway, HUVECs were pretreated with the PI3K/Akt pathway inhibitor LY294002. We found LY294002 treatment significantly blocked the up-regulation of p-eNOS and Bcl-2 in hNinj123–37 treated groups (Figure 6D–F). Moreover, the restored HUVEC tube formation ability in the HG condition through blocking Ninj1 was further impaired after the PI3K/Akt pathway was inhibited (Figure 6G–I). This finding suggests that blocking Ninj1 positively modulated the phosphorylation of eNOS and expression of Bcl-2 in HG-cultured HUVECs at least partially through the PI3K/Akt signaling pathway.

Functional blocking of Ninj1 suppresses HUVEC apoptosis and promotes cell function via PI3K/Akt signaling pathway

Figure 6
Functional blocking of Ninj1 suppresses HUVEC apoptosis and promotes cell function via PI3K/Akt signaling pathway

(A) Protein expression of phos-PI3K, PI3K, phos-Akt, and Akt in HUVECs cultured in HG for 24 h and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 for the times indicated (n=4). NG-cultured and L-Glu-treated HUVECs served as negative and osmotic controls respectively. (B and C) Ratio of phos-PI3K and PI3K (B) and relative ratio of phos-Akt and Akt (C) based on normalized band intensity values processed by ImageJ (n=4); ****P<0.0001. (D) Protein expression of Bcl-2, Bax, phos-eNOS, and eNOS in HUVECs cultured in HG for 24 h and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 or PI3K/AKT signaling inhibitor LY294002 for the times indicated (n=4). (E and F) Relative ratio of Bcl-2 and Bax (E) and phos-eNOS and eNOS (F) based on normalized band intensity values processed by ImageJ (n=4); ****P<0.0001, ***P<0.001, and *P<0.05. (G) HUVECs were precultured in HG for 24 h and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 or PI3K/AKT signaling inhibitor LY294002 for the times indicated. After treatment, HUVECs were plated on Matrigel, and the tube formation assay was performed. NG-cultured HUVECs served as positive controls; scale bar: 200 μm). (H and I) Quantification of mesh numbers (number of meshes per area) (H) and relative tube length (compared with NG group) (I) were analyzed by the Angiogenesis Analyzer for ImageJ (n=8); ****P<0.0001, ***P<0.001, **P<0.01, and *P<0.05.

Figure 6
Functional blocking of Ninj1 suppresses HUVEC apoptosis and promotes cell function via PI3K/Akt signaling pathway

(A) Protein expression of phos-PI3K, PI3K, phos-Akt, and Akt in HUVECs cultured in HG for 24 h and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 for the times indicated (n=4). NG-cultured and L-Glu-treated HUVECs served as negative and osmotic controls respectively. (B and C) Ratio of phos-PI3K and PI3K (B) and relative ratio of phos-Akt and Akt (C) based on normalized band intensity values processed by ImageJ (n=4); ****P<0.0001. (D) Protein expression of Bcl-2, Bax, phos-eNOS, and eNOS in HUVECs cultured in HG for 24 h and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 or PI3K/AKT signaling inhibitor LY294002 for the times indicated (n=4). (E and F) Relative ratio of Bcl-2 and Bax (E) and phos-eNOS and eNOS (F) based on normalized band intensity values processed by ImageJ (n=4); ****P<0.0001, ***P<0.001, and *P<0.05. (G) HUVECs were precultured in HG for 24 h and treated with 10 μM Ninj1ctr or 3 or 10 μM blocking antibody hNinj123–37 or PI3K/AKT signaling inhibitor LY294002 for the times indicated. After treatment, HUVECs were plated on Matrigel, and the tube formation assay was performed. NG-cultured HUVECs served as positive controls; scale bar: 200 μm). (H and I) Quantification of mesh numbers (number of meshes per area) (H) and relative tube length (compared with NG group) (I) were analyzed by the Angiogenesis Analyzer for ImageJ (n=8); ****P<0.0001, ***P<0.001, **P<0.01, and *P<0.05.

Hindlimb blood flow in db/db mice was effectively maintained after Ninj1 blockade

DM was associated with reduced peripheral blood perfusion due to micro- and macrovascular endothelial cell injury. Having demonstrated that the Ninj1 blocking antibody promotes HUVEC survival and functions under HG conditions in vitro, we then further sought to validate the protective effect of Ninj1 blocking antibody on endothelial cells and blood flow perfusion. We intramuscularly injected the right hindlimbs of db/db mice with control Ninj1 antibody, Ninj1ctr, or mouse functional Ninj1 blocking antibody, mNinj126-37, weekly. In Ninj1ctr-treated or nontreated hindlimbs of db/db mice, the blood perfusion was only 43 ± 17% of that of age-matched db/+ mice, but in hindlimbs of db/db mice treated with mNinj126–37, blood perfusion was significantly maintained, and it was 78 ± 24% of the level of age-matched db/+ mice (Figure 7A and B). This finding indicates that functional Ninj1 blocking can at least partially prevent in vivo hyperglycemia-induced endothelial injury and vascular occlusion.

Blocking Ninj1 protects endothelial cells and maintains blood flow in the hindlimbs of db/db mice

Figure 7
Blocking Ninj1 protects endothelial cells and maintains blood flow in the hindlimbs of db/db mice

(A) Representative images of blood perfusion measured by laser Doppler at week 8 after different treatment weekly (left: db/+ control, middle: Ninj1ctr-treated db/db, right: mNinj126-37-treated db/db; red: highest perfusion, blue: lowest perfusion). (B) Relative ratio of blood perfusion compared with db/+ control based on normalized Doppler intensity values; ***P<0.001 and **P<0.01. (C) Representative PECAM-1 (green) and α-SMA (red) double-immunostaining in hindlimb muscle tissue from db/+ control (upper) and db/db mice 8 weeks after treatment with i.m. injection of Ninj1ctr (middle) or mNinj126–37 (lower) weekly respectively. Nuclei were labeled with DAPI (blue); scale bars: 100 μm. (D) PECAM-1 positive area in hindlimb muscle tissue quantified by ImageJ (n=8); ****P<0.0001 and **P<0.01. (E) α-SMA and PECAM-1 double positive vascular numbers (n=8); ****P<0.0001, **P<0.01, and *P<0.05.

Figure 7
Blocking Ninj1 protects endothelial cells and maintains blood flow in the hindlimbs of db/db mice

(A) Representative images of blood perfusion measured by laser Doppler at week 8 after different treatment weekly (left: db/+ control, middle: Ninj1ctr-treated db/db, right: mNinj126-37-treated db/db; red: highest perfusion, blue: lowest perfusion). (B) Relative ratio of blood perfusion compared with db/+ control based on normalized Doppler intensity values; ***P<0.001 and **P<0.01. (C) Representative PECAM-1 (green) and α-SMA (red) double-immunostaining in hindlimb muscle tissue from db/+ control (upper) and db/db mice 8 weeks after treatment with i.m. injection of Ninj1ctr (middle) or mNinj126–37 (lower) weekly respectively. Nuclei were labeled with DAPI (blue); scale bars: 100 μm. (D) PECAM-1 positive area in hindlimb muscle tissue quantified by ImageJ (n=8); ****P<0.0001 and **P<0.01. (E) α-SMA and PECAM-1 double positive vascular numbers (n=8); ****P<0.0001, **P<0.01, and *P<0.05.

Functional blocking of Ninj1 maintains α-SMA+ PECAM-1+ double positive vascular numbers in db/db mice

We further detected the in vivo vascular numbers and endothelial contents of mNinj126–37-treated hindlimbs through dual-label immunofluorescence analysis. Compared with the age-matched db/+ mice, the endothelial contents and the α-SMA+ PECAM-1+ double-positive normal vascular lumens were remarkably decreased in the hindlimbs of db/db mice with or without Ninj1ctr treatment (Figure 7C and E). Negative control of hindlimb muscle tissues stained with secondary antibodies only showed no positive signal (Supplementary Figure S3). Although some of the α-SMA+ vascular structures of db/db mice were preserved, the inner endothelium was severely damaged and almost undetectable, and as such, α-SMA+ PECAM-1 vessels were hardly able to maintain normal blood flow. After the application of Ninj1 blocking antibody mNinj126–37 rather than control protein, the vascular endothelial cells were effectively protected, and the α-SMA+ PECAM-1+ vascular lumens reached approximately 63% of age-matched db/+ mice, while Ninj1ctr treated mice were only approximately 27% of that of age-matched db/+ mice, which is consistent with Doppler perfusion results. These results indicate that the functional blocking antibody of Ninj1 at least partially rescues the endothelial cells in hyperglycemic conditions in vivo, exhibiting great potential in the prevention of endothelial dysfunction in diabetes.

Discussion

Inflammation-related endothelial dysfunction is a key therapeutic target of DM-related cardiovascular and peripheral vascular complications [35]. The role of Ninj1 in mediating the pathogenesis of endothelial dysfunction in HG conditions remains poorly understood. In the present study, we provided evidence that Ninj1 was up-regulated in endothelial cells in Type 2 diabetic patients and db/db mice and in primary-cultured HUVECs exposed to HG conditions. As the mechanisms shown in Figure 8, functional blocking of Ninj1 in vitro successfully reduced HG-induced endothelial inflammation and apoptosis and restored HUVEC tube formation abilities. Intramuscular injection of Ninj1 blocking antibody significantly inhibited HG-induced endothelial cell loss and maintained blood perfusion of the lower extremity in db/db mice. The functionality of the used antibody has been verified by previous studies [21,23]. On the basis of our study and previous work, we considered Ninj1 a potent target molecule for preventing endothelial dysfunction in DM.

Proposed mechanism of Ninj1-mediated endothelial dysfunction under high glucose condition

Figure 8
Proposed mechanism of Ninj1-mediated endothelial dysfunction under high glucose condition

Ninj1 mediates HG-induced activation of p38-MAPK and NF/κB and formation of ROS, thus promotes the expression of proinflammatory genes and may also enhance inflammation-related caspase-3 activation, leading to endothelial inflammation and apoptosis. Ninj1 may also mediate the inhibition of PI3K/Akt pathway under HG condition, thus aggravates apoptosis and endothelial dysfunction by attenuating Bcl-2 expression and eNOS activation.

Figure 8
Proposed mechanism of Ninj1-mediated endothelial dysfunction under high glucose condition

Ninj1 mediates HG-induced activation of p38-MAPK and NF/κB and formation of ROS, thus promotes the expression of proinflammatory genes and may also enhance inflammation-related caspase-3 activation, leading to endothelial inflammation and apoptosis. Ninj1 may also mediate the inhibition of PI3K/Akt pathway under HG condition, thus aggravates apoptosis and endothelial dysfunction by attenuating Bcl-2 expression and eNOS activation.

Previous studies have demonstrated that Ninj1 is involved in a variety of pathophysiological processes. In addition to its role in nerve regeneration [19], its involvement in inflammatory processes, such as MS, EAE [21,22,36] and sepsis [23], has also been verified. These studies confirmed that targeted blocking of Ninj1 can attenuate inflammation. DM is considered a low-grade inflammation disease and a risk factor for endothelial dysfunction [14]. Our experiments found a significant up-regulation of Ninj1 in endothelial cells both in vitro and in vivo in HG conditions, indicating that Ninj1 has a high likelihood of being mechanistically involved in diabetic endothelial dysfunction.

In the present study, we identified decreased proliferation and impairment in tube formation in endothelial cells when exposed to HG conditions, whereas the addition of the Ninj1 blocking antibody effectively protected HUVEC tube formation abilities. We found that blocking Ninj1 promoted HUVEC survival, instead of proliferation, via decreasing apoptosis in HG conditions. Most likely, due to discrepancies in cell subtypes, Ang1, which is involved in endothelial cell proliferation and may be activated after Ninj1 blocking [29], was not regulated in the present study. Another protein, eNOS, which plays important roles in regulating the survival and function of endothelial cells [37], is usually decreased by HG conditions, contributing to endothelial dysfunction and cardiovascular and peripheral vascular events [38,39]. In the present study, the phosphorylation of eNOS was not regulated by HG conditions. Interestingly, p-eNOS was significantly up-regulated after Ninj1 was blocked, and this finding is at least partially the reason for preservation of cell function.

Since we found no eNOS dysregulation in HG-treated HUVECs compared with NG controls, this explanation does not clarify the cell dysfunction in our experiments. We continued to look for other possibilities leading to endothelial dysfunction. We found that the level of ROS was significantly higher in HG-treated HUVECs than in NG controls. ROS is a sort of natural product in normal physiological conditions and is essential in cell signaling and homeostasis [40]. However, overproduction of ROS in DM causes endothelial dysfunction through triggering multiple damage effects, including but not limited to inducing apoptosis or cell death and overexpression of proinflammatory factors [41]. Importantly, overproduction of ROS is thought to be a unifying process in endothelial dysfunction caused by HG. In line with the study of Yin et al. [29], we found that functional blocking of Ninj1 significantly reduced the generation of ROS in HG-cultured endothelial cells.

Increased ROS production may stimulate the apoptosis of endothelial cells [42]. In the present study, we found that HG significantly stimulated the activation of caspase-3, which was significantly inhibited after Ninj1 was blocked. In addition, we also investigated other apoptosis-related proteins of the Bcl-2 family, but we found that HG treatment did not affect the expression of Bcl-2, Bax, or the Bcl-2/Bax ratio. Interestingly, the application of Ninj1 blocking antibody significantly stimulated the expression of Bcl-2, thus increasing the Bcl-2/Bax ratio. This outcome was in line with the eNOS phosphorylation results and revealed that functional blocking of Ninj1 exerted a significant antiapoptotic function. Further study uncovered that phosphorylation of PI3K and its downstream effector, Akt, were remarkably elevated when endothelial cells were treated by Ninj1 blocking antibody in HG conditions. In previous studies, the PI3K/Akt pathway was able to enhance cell survival by inducing eNOS activity [4244] and expression of antiapoptotic proteins such as Bcl-2 [45]. Moreover, activation of the PI3K/Akt pathway is important in preventing ROS-mediated endothelial cell injury [42,46]. At present, we provide evidence that blocking of Ninj1 could participate in antiapoptotic effects by up-regulating the Bcl-2/Bax ratio, possibly a new mechanism by which the Ninj1 blockade inhibits apoptosis.

In accordance with the above results, blocking Ninj1 also attenuated the expression of proinflammatory genes and decreased the activation of the transcription factor, NF-κB, in endothelial cells in HG conditions. The modulation of NF-κB is most likely due to HG-mediated Ninj1 expression and p38 MAPK because the Ninj1 blocking antibody significantly decreased p38 MAPK activity, and no aberrant regulation of Erk1/2 or JNK was observed (data not shown). Recently, a study found that inhibiting Ninj1 with either 5 or 20 μM of blocking antibody attenuated LPS-triggered systemic inflammation in macrophages and endothelial cells by modulating the activity of p38 and another transcription factor, AP-1 [23]. More importantly, Ninj1 was identified to directly bind to LPS and regulate macrophage-mediated inflammation [47]. Thus, Ninj1 is involved in HG-induced endothelial inflammation, and it could be partially reversed after Ninj1 is blocked.

In a Type 2 DM mouse model, we found that although the vascular structure (α-SMA+ lumens) was maintained, the endothelium and blood perfusion were severely damaged. In vivo blocking of Ninj1 significantly preserved endothelial contents and endothelium integrity (α-SMA+PECAM-1+ lumens), and blood flow in hindlimbs was remarkably rescued.

In short, we found that functional blocking of Ninj1 is an effective way to protect in vitro endothelial function through suppressing inflammation, ameliorating apoptosis, and enhancing survival and to preserve in vivo endothelial cells and blood perfusion in hindlimbs. Our study provides new evidence for the prevention of diabetic peripheral vascular complications by functional blocking of Ninj1. However, our research also has some limitations. First, it is reported that there are phenotypic, genetic, and protein differences among endothelial cells of arteries, veins, or microvasculars [48]. In the present study, HUVEC was used because it is still an important cell type for studying endothelial dysfunction in diabetes as recent studies has demonstrated [49,50]. Second, since this study focused on the direct effect of Ninj1 on endothelial cell function and its targeted blocking effect, we did not investigate the role of Ninj1 in the regulation of the function of inflammatory cells, which also contributes to endothelial dysfunction in diabetes. Further studies are needed to understand the exact mechanism.

Clinical perspectives

  • Endothelial dysfunction in Type 2 diabetic patients is associated with increased endothelial expression of Ninjurin1 (Ninj1).

  • Blocking Ninj1 ameliorates high glucose-induced endothelial inflammation and apoptosis and maintains endothelial cell function in vitro.

  • Blocking Ninj1 protects endothelial cells and restores blood perfusion of hindlimb in diabetic mice.

  • Ninj1 is a potential therapeutic target to prevent endothelial dysfunction in diabetes.

Funding

This study was supported by the National Natural Science Foundation of China [grant numbers 81370423, 81570432, 81500371, and 81601621].

Competing Interests

The authors declare that there are no competing interests associated with the manuscript.

Author Contribution

X.Wang, J.Q., X.Z., and Z.Z. performed most of the experiments and edited the manuscript. Z.P., K.Y., X. Wu, and X.Y. analyzed data and performed statistical analysis. H.S., X. Wang, X.G., X. Liu, and M.Y. conducted the experiments. X. Liu, M.Y., and X. Lu contributed to discussion and editing of the manuscript. All authors read and approved the final manuscript.

Abbreviations

     
  • α-SMA

    α-smooth muscle actin

  •  
  • Akt

    protein kinase B

  •  
  • ANOVA

    analysis of variance

  •  
  • Bax

    Bcl-2 associated X protein

  •  
  • Bcl-2

    B-cell lymphoma-2

  •  
  • cPC

    capillary pericyte

  •  
  • DCFH-DA

    2′,7′-dichlorofluorescin diacetate

  •  
  • DM

    diabetes mellitus

  •  
  • DMEM

    Dulbecco’s Modified Eagle’s Medium

  •  
  • EdU

    5-ethynyl-2′-deoxyuridine

  •  
  • eNOS

    endothelial nitric oxide synthase

  •  
  • HG

    high glucose

  •  
  • HUVEC

    human umbilical vein endothelial cells

  •  
  • ICAM-1

    intercellular cell adhesion molecule-1

  •  
  • IL-6

    interleukin-6

  •  
  • L-Glu

    L-glucose

  •  
  • MCP-1

    monocyte chemotactic protein-1

  •  
  • NF-κB

    nuclear factor kappa-light-chain-enhancer of activated B cells

  •  
  • NG

    normal glucose

  •  
  • Ninj1

    nerve injury-induced protein 1

  •  
  • p38 MAPK

    p38 mitogen-activated protein kinases

  •  
  • PCR

    polymerase chain reaction

  •  
  • PECAM-1

    platelet endothelial cell adhesion molecule-1

  •  
  • PI3K

    phosphatidylinositol 3 kinase

  •  
  • ROS

    reactive oxygen species

  •  
  • VCAM-1

    vascular cell adhesion molecule 1

References

References
1
Paneni
F.
,
Beckman
J.A.
,
Creager
M.A.
and
Cosentino
F.
(
2013
)
Diabetes and vascular disease: pathophysiology, clinical consequences, and medical therapy: part I
.
Eur. Heart J.
34
,
2436
2443
[PubMed]
2
Nanditha
A.
,
Ma
R.C.
,
Ramachandran
A.
,
Snehalatha
C.
,
Chan
J.C.
,
Chia
K.S.
et al. 
(
2016
)
Diabetes in Asia and the Pacific: implications for the global epidemic
.
Diabetes Care
39
,
472
485
[PubMed]
3
Beckman
J.A.
,
Paneni
F.
,
Cosentino
F.
and
Creager
M.A.
(
2013
)
Diabetes and vascular disease: pathophysiology, clinical consequences, and medical therapy: part II
.
Eur. Heart J.
34
,
2444
2452
[PubMed]
4
Wild
S.
,
Roglic
G.
,
Green
A.
,
Sicree
R.
and
King
H.
(
2004
)
Global prevalence of diabetes: estimates for the year 2000 and projections for 2030
Diabetes Care
27
,
2030
1053
5
Ghosh
P.
,
Sahoo
R.
,
Vaidya
A.
,
Chorev
M.
and
Halperin
J.A.
(
2015
)
Role of complement and complement regulatory proteins in the complications of diabetes
.
Endocr. Rev.
36
,
272
288
[PubMed]
6
Jia
G.
and
Sowers
J.R.
(
2014
)
Endothelial dysfunction potentially interacts with impaired glucose metabolism to increase cardiovascular risk
.
Hypertension
64
,
1192
1193
[PubMed]
7
Tousoulis
D.
,
Papageorgiou
N.
,
Androulakis
E.
,
Siasos
G.
,
Latsios
G.
,
Tentolouris
K.
et al. 
(
2013
)
Diabetes mellitus-associated vascular impairment: novel circulating biomarkers and therapeutic approaches
.
J. Am. Coll. Cardiol.
62
,
,
667
676
[PubMed]
8
Ho
F.M.
,
Liu
S.H.
,
Liau
C.S.
,
Huang
P.J.
and
Lin-Shiau
S.Y.
(
2000
)
High glucose-induced apoptosis in human endothelial cells is mediated by sequential activations of c-Jun NH-terminal kinase and caspase-3
.
Circulation
101
,
2618
2624
[PubMed]
9
Bhatt
M.P.
,
Lim
Y.C.
,
Kim
Y.M.
and
Ha
K.S.
(
2013
)
C-peptide activates AMPKα and prevents ROS-mediated mitochondrial fission and endothelial apoptosis in diabetes
.
Diabetes
62
,
3851
3862
[PubMed]
10
Bhatt
M.P.
,
Lim
Y.C.
,
Hwang
J.
,
Na
S.
,
Kim
Y.M.
and
Ha
K.S.
(
2013
)
C-peptide prevents hyperglycemia-induced endothelial apoptosis through inhibition of reactive oxygen species-mediated transglutaminase 2 activation
.
Diabetes
62
,
243
253
[PubMed]
11
Wang
D.
,
Luo
P.
,
Wang
Y.
,
Li
W.
,
Wang
C.
,
Sun
D.
et al. 
(
2013
)
Glucagon-like peptide-1 protects against cardiac microvascular injury in diabetes via a cAMP/PKA/Rho-dependent mechanism
.
Diabetes
62
,
1697
1708
[PubMed]
12
Kohner
E.M.
and
Henkind
P.
(
1970
)
Correlation of fluorescein angiogram and retinal digest in diabetic retinopathy
.
Am. J. Ophthalmol.
69
,
403
414
[PubMed]
13
Mazzone
T.
,
Chait
A.
and
Plutzky
J.
(
2008
)
Cardiovascular disease risk in type 2 diabetes mellitus: insights from mechanistic studies
.
Lancet
371
,
1800
1809
[PubMed]
14
Zhang
L.
,
Zalewski
A.
,
Liu
Y.
,
Mazurek
T.
,
Cowan
S.
,
Martin
J.L.
et al. 
(
2003
)
Diabetes-induced oxidative stress and low-grade inflammation in porcine coronary arteries
.
Circulation
108
,
472
478
[PubMed]
15
Halvorsen
B.
,
Santilli
F.
,
Scholz
H.
,
Sahraoui
A.
,
Gulseth
H.L.
,
Wium
C.
et al. 
(
2016
)
LIGHT/TNFSF14 is increased in patients with type 2 diabetes mellitus and promotes islet cell dysfunction and endothelial cell inflammation in vitro
.
Diabetologia
59
,
2134
2144
[PubMed]
16
Donath
M.Y.
(
2014
)
Targeting inflammation in the treatment of type 2 diabetes: time to start
.
Nat. Rev. Drug Discov.
13
,
465
476
[PubMed]
17
Giacco
F.
and
Brownlee
M.
(
2010
)
Oxidative stress and diabetic complications
.
Circ. Res.
107
,
1058
1070
[PubMed]
18
Tabas
I.
and
Glass
C.K.
(
2013
)
Anti-inflammatory therapy in chronic disease: challenges and opportunities
.
Science
339
,
166
172
[PubMed]
19
Araki
T.
and
Milbrandt
J.
(
1996
)
Ninjurin, a novel adhesion molecule, is induced by nerve injury and promotes axonal growth
.
Neuron
17
,
353
361
[PubMed]
20
Araki
T.
,
Zimonjic
D.B.
,
Popescu
N.C.
and
Milbrandt
J.
(
1997
)
Mechanism of homophilic binding mediated by ninjurin, a novel widely expressed adhesion molecule
.
J. Biol. Chem.
272
,
21373
21380
[PubMed]
21
Ahn B
J.
,
Le
H.
,
Shin
M.W.
,
Bae
S.J.
,
Lee
E.J.
,
Wee
H.J.
et al. 
(
2014
)
Ninjurin1 deficiency attenuates susceptibility of experimental autoimmune encephalomyelitis in mice
.
J. Biol. Chem.
289
,
3328
3338
[PubMed]
22
Ifergan
I.
,
Kebir
H.
,
Terouz
S.
,
Alvarez
J.I.
,
Lecuyer
M.A.
,
Gendron
S.
et al. 
(
2011
)
Role of Ninjurin-1 in the migration of myeloid cells to central nervous system inflammatory lesions
.
Ann. Neurol.
70
,
751
763
[PubMed]
23
Jennewein
C.
,
Sowa
R.
,
Faber
A.C.
,
Dildey
M.
,
von Knethen
A.
,
Meybohm
P.
et al. 
(
2015
)
Contribution of Ninjurin1 to toll-like receptor 4 signaling and systemic inflammation
.
Am. J. Respir. Cell Mol. Biol.
53
,
656
663
[PubMed]
24
Chen
J.S.
,
Coustan-Smith
E.
,
Suzuki
T.
,
Neale
G.A.
,
Mihara
K.
,
Pui
CH.
et al. 
(
2001
)
Identification of novel markers for monitoring minimal residual disease in acute lymphoblastic leukemia
.
Blood
97
,
2115
2120
[PubMed]
25
Woo
J.K.
,
Jang
Y.S.
,
Kang
J.H.
,
Hwang
J.I.
,
Seong
J.K.
,
Lee
S.J.
et al. 
(
2016
)
Ninjurin1 inhibits colitis-mediated colon cancer development and growth by suppression of macrophage infiltration through repression of FAK signaling
.
Oncotarget
7
,
29592
29604
[PubMed]
26
Jang
Y.S.
,
Kang
J.H.
,
Woo
J.K.
,
Kim
H.M.
,
Hwang
J.I.
,
Lee
S.J.
et al. 
(
2016
)
Ninjurin1 suppresses metastatic property of lung cancer cells through inhibition of interleukin 6 signaling pathway
.
Int. J. Cancer
139
,
383
395
[PubMed]
27
Chen
J.S.
,
Coustan-Smith
E.
,
Suzuki
T.
,
Neale
G.A.
,
Mihara
K.
,
Pui
C.H.
et al. 
(
2001
)
Identification of novel markers for monitoring minimal residual disease in acute lymphoblastic leukemia
.
Blood
97
,
2115
2120
[PubMed]
28
Yin
G.N.
,
Kim
W.J.
,
Jin
H.R.
,
Kwon
M.H.
,
Song
K.M.
,
Choi
M.J.
et al. 
(
2013
)
Nerve injury-induced protein 1 (Ninjurin-1) is a novel therapeutic target for cavernous nerve injury-induced erectile dysfunction in mice
.
J. Sex Med.
10
,
1488
1501
[PubMed]
29
Yin
G.N.
,
Choi
M.J.
,
Kim
W.J.
,
Kwon
M.H.
,
Song
K.M.
,
Park
J.M.
et al. 
(
2014
)
Inhibition of Ninjurin 1 restores erectile function through dual angiogenic and neurotrophic effects in the diabetic mouse
.
Proc. Natl. Acad. Sci. U.S.A.
111
,
E2731
E2740
30
Lee
H.J.
,
Ahn
B.J.
,
Shin
M.W.
,
Jeong
J.W.
,
Kim
J.H.
and
Kim
K.W.
(
2009
)
Ninjurin1 mediates macrophage-induced programmed cell death during early ocular development
.
Cell Death Differ.
16
,
1395
1407
[PubMed]
31
Fujita
Y.
and
Kawamoto
A.
(
2015
)
Ninjurin1 - a novel regulator of angiogenesis mediated by pericytes
.
Circ. J.
79
,
1218
1219
[PubMed]
32
Matsuki
M.
,
Kabara
M.
,
Saito
Y.
,
Shimamura
K.
,
Minoshima
A.
,
Nishimura
M.
et al. 
(
2015
)
Ninjurin1 is a novel factor to regulate angiogenesis through the function of pericytes
.
Circ. J.
79
,
1363
1371
[PubMed]
33
Liang
X.
,
Zhang
L.
,
Wang
S.
,
Han
Q.
and
Zhao
R.C.
(
2016
)
Exosomes secreted by mesenchymal stem cells promote endothelial cell angiogenesis by transferring miR-125a
.
J. Cell Sci.
29
,
2182
2189
34
Peng
Z.
,
Yang
X.
,
Qin
J.
,
Ye
K.
,
Wang
X.
,
Shi.
H.
et al. 
(
2017
)
Glyoxalase-1 overexpression reverses defective proangiogenic function of diabetic adipose-derived stem cells in streptozotocin-induced diabetic mice model of critical limb ischemia
.
Stem Cells Transl. Med.
,
6
,
261
271
,
35
Prattichizzo
F.
,
Giuliani
A.
,
De Nigris
V.
,
Pujadas
G.
,
Ceka
A.
,
La Sala
L.
et al. 
(
2016
)
Extracellular microRNAs and endothelial hyperglycaemic memory: a therapeutic opportunity
.
Diabetes Obes. Metab.
18
,
855
867
[PubMed]
36
Ahn
B.J.
,
Lee
H.J.
,
Shin
M.W.
,
Choi
J.H.
,
Jeong
J.W.
and
Kim
K.W.
(
2009
)
Ninjurin1 is expressed in myeloid cells and mediates endothelium adhesion in the brains of EAE rats
.
Biochem. Biophys. Res. Commun.
387
,
321
325
[PubMed]
37
Yu
J.
,
deMuinck
E.D.
,
Zhuang
Z.
,
Drinane
M.
,
Kauser
K.
,
Rubanyi
G.M.
et al. 
(
2005
)
Endothelial nitric oxide synthase is critical for ischemic remodeling, mural cell recruitment, and blood flow reserve
.
Proc. Natl. Acad. Sci. U.S.A.
102
,
10999
11004
38
Wang
X.
,
Lockhart
S.M.
,
Rathjen
T.
,
Albadawi
H.
,
Sorensen
D.
,
O'Neill
B.T.
et al. 
(
2016
)
Insulin downregulates the transcriptional coregulator CITED2, an inhibitor of proangiogenic function in endothelial cells
.
Diabetes
65
,
3680
3690
[PubMed]
39
Sowers
J.R.
(
2013
)
Diabetes mellitus and vascular disease
.
Hypertension
61
,
943
947
[PubMed]
40
Stone
J.R.
and
Collins
T.
(
2002
)
The role of hydrogen peroxide in endothelial proliferative responses
.
Endothelium
9
,
231
238
[PubMed]
41
Chen
Z.
,
Wen
L.
,
Martin
M.
,
Hsu
C.Y.
,
Fang
L.
,
Lin
F.M.
et al. 
(
2015
)
Oxidative stress activates endothelial innate immunity via sterol regulatory element binding protein 2 (SREBP2) transactivation of microRNA-92a
.
Circulation
131
,
805
814
[PubMed]
42
Ho
F.M.
,
Lin
W.W.
,
Chen
B.C.
,
Chao
C.M.
,
Yang
C.R.
,
Lin
C.R.
et al. 
(
2006
)
High glucose-induced apoptosis in human vascular endothelial cells is mediated through NF-kappaB and c-Jun NH2-terminal kinase pathway and prevented by PI3K/Akt/eNOS pathway
.
Cell. Signal.
18
,
391
399
[PubMed]
43
Song
C.L.
,
Liu
B.
,
Shi
Y.F.
,
Liu
N.
,
Yan
Y.Y.
,
Zhang
J.C.
et al. 
(
2016
)
MicroRNA-130a alleviates human coronary artery endothelial cell injury and inflammatory responses by targeting PTEN via activating PI3K/Akt/eNOS signaling pathway
.
Oncotarget
7
,
71922
71936
[PubMed]
44
Satake
A.
,
Takaoka
M.
,
Nishikawa
M.
,
Yuba
M.
,
Shibata
Y
,
Okumura
K.
et al. 
(
2008
)
Protective effect of 17beta-estradiol on ischemic acute renal failure through the PI3K/Akt/eNOS pathway
.
Kidney Int.
73
,
308
317
[PubMed]
45
Matsunaga
T.
,
Takemoto
N.
,
Sato
T.
,
Takimoto
R.
,
Tanaka
I.
,
Fujimi
A.
et al. 
(
2003
)
Interaction between leukemic-cell VLA-4 and stromal fibronectin is a decisive factor for minimal residual disease of acute myelogenous leukemia
.
Nat. Med.
9
,
1158
1165
[PubMed]
46
Chavakis
E.
,
Dernbach
E.
,
Hermann
C.
,
Mondorf
U.F.
,
Zeiher
A.M.
and
Dimmeler
S.
(
2001
)
Oxidized LDL inhibits vascular endothelial growth factor-induced endothelial cell migration by an inhibitory effect on the Akt/endothelial nitric oxide synthase pathway
.
Circulation
103
,
2102
2107
[PubMed]
47
Shin
M.W.
,
Bae
S.J.
,
Wee
H.J.
,
Lee
H.J.
,
Ahn
B.J.
,
Le
H.
et al. 
(
2016
)
Ninjurin1 regulates lipopolysaccharide-induced inflammation through direct binding
.
Int. J. Oncol.
48
,
821
828
[PubMed]
48
Chi
J.T.
,
Chang
H.Y.
,
Haraldsen
G.
,
Jahnsen
F.L.
,
Troyanskaya
O.G.
,
Chang
D.S.
et al. 
(
2003
)
Endothelial Cell Diversity Revealed by Global Expression Profiling
.
Proc. Natl. Acad. Sci. U.S.A.
100
,
10623
10628
49
Fetterman
J.L.
,
Holbrook
M.
,
Flint
N.
,
Feng
B.
,
Breton-Romero
R.
,
Linder
E.A.
et al. 
(
2016
)
Restoration of autophagy in endothelial cells from patients with diabetes mellitus improves nitric oxide signaling
.
Atherosclerosis
247
,
207
217
[PubMed]
50
Kumar
S.
,
Kim
Y.R.
,
Vikram
A.
,
Naqvi
A.
,
Li
Q.
,
Kassan
M.
et al. 
(
2017
)
Sirtuin1-regulated lysine acetylation of p66Shc governs diabetes-induced vascular oxidative stress and endothelial dysfunction[J]
.
Proc. Natl. Acad. Sci. U.S.A.
114
,
1714
1719

Author notes

*

These authors contributed equally to this work.