MicroRNA-200a/200b Modulate High Glucose-Induced Endothelial Inflammation by Targeting O-linked N-Acetylglucosamine Transferase Expression

Background and Aims: Increased O-linked N-acetylglucosamine (O-GlcNAc) modification of proteins by O-GlcNAc transferase (OGT) is associated with diabetic complications. Furthermore, oxidative stress promotes endothelial inflammation during diabetes. A previous study reported that microRNA-200 (miR-200) family members are sensitive to oxidative stress. In this study, we examined whether miR-200a and miR-200b regulate high-glucose (HG)-induced OGT expression in human aortic endothelial cells (HAECs) and whether miRNA-200a/200b downregulate OGT expression to control HG-induced endothelial inflammation. Methods: HAECs were stimulated with high glucose (25 mM) for 12 and 24 h. Real-time polymerase chain reaction (PCR), western blotting, THP-1 adhesion assay, bioinformatics predication, transfection of miR-200a/200b mimic or inhibitor, luciferase reporter assay, and transfection of siRNA OGT were performed. The aortic endothelium of db/db diabetic mice was evaluated by immunohistochemistry staining. Results: HG upregulated OGT mRNA and protein expression and protein O-GlcNAcylation levels (RL2 antibody) in HAECs, and showed increased intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule 1 (VCAM-1), and E-selectin gene expression; ICAM-1 expression; and THP-1 adhesion. Bioinformatics analysis revealed homologous sequences between members of the miR-200 family and the 3′-untranslated region (3′-UTR) of OGT mRNA, and real-time PCR analysis confirmed that members of miR-200 family were significantly decreased in HG-stimulated HAECs. This suggests the presence of an impaired feedback restraint on HG-induced endothelial protein O-GlcNAcylation levels because of OGT upregulation. A luciferase reporter assay demonstrated that miR-200a/200b mimics bind to the 3′-UTR of OGT mRNA. Transfection with miR-200a/200b mimics significantly inhibited HG-induced OGT mRNA expression, OGT protein expression; protein O-GlcNAcylation levels; ICAM-1, VCAM-1, and E-selectin gene expression; ICAM-1 expression; and THP-1 adhesion. Additionally, siRNA-mediated OGT depletion reduced HG-induced protein O-GlcNAcylation; ICAM-1, VCAM-1, and E-selectin gene expression; ICAM-1 expression; and THP-1 adhesion, confirming that HG-induced endothelial inflammation is partially mediated via OGT-induced protein O-GlcNAcylation. These results were validated in vivo: tail-vein injection of miR-200a/200b mimics downregulated endothelial OGT and ICAM-1 expression in db/db mice. Conclusion: miR-200a/200b are involved in modulating HG-induced endothelial inflammation by regulating OGT-mediated protein O-GlcNAcylation, suggesting the therapeutic role of miR-200a/200b on vascular complications in diabetes.


INTRODUCTION
Diabetic vascular disease-a low-grade, chronic inflammatory disease-is a major cause of death in developed countries. Increased levels of reactive oxygen species (ROS) in endothelial cells are involved in the pathophysiology of diabetic vascular complications by activating downstream pathways, leading to increases in polyol pathway flux, advanced glycation end product formation, advanced glycation end product receptor formation, protein kinase C isoform activation, and hexosamine biosynthetic pathway (HBP) flux (Giacco and Brownlee, 2010). Hyperglycemia is the primary abnormality causing diabetic endothelial dysfunction (Nakagami et al., 2005). Normally, glucose enters cells and is metabolized into glucose-6-phosphate and fructose-6-phosphate, the latter of the two being typically destined for the glycolytic pathway. Only a small amount of fructose-6-phosphate (2-5%) is converted to glucosamine-6phosphate by glutamine:fructose 6-phosphate amidotransferase. This conversion commits fructose-6-phosphate to entering the HBP, which produces UDP-N-acetylglucosamine (UDP-GlcNAc) as an important sensor of nutrition flux by integrating the metabolic processes of glucose, fatty acids, amino acids, and nucleotides (Bond and Hanover, 2015).
O-GlcNAc glycosylation (O-GlcNAcylation) is an important post-translational modification of nuclear, cytoplasmic, and mitochondrial proteins which modulates the stability, subcellular localization, and activity of proteins (Issad et al., 2010). O-GlcNAcylation modification is controlled by two enzymes: β-N-acetylglucosaminidase (OGA), which removes Olinked N-acetylglucosamine (O-GlcNAc) from O-GlcNAcylated proteins, and O-GlcNAc transferase (OGT), which adds O-GlcNAc from UDP-GlcNAc to the threonine and serine residues of targeted proteins (Bond and Hanover, 2015). More than 4000 O-GlcNAcylated protein targets have been identified; these proteins play a key role in multiple biological processes, including transcription, translation, proteasomal degradation, homeostasis, epigenetic regulation, and stress responses (Ma and Hart, 2014). Therefore, aberrant O-GlcNAcylation has been linked to the pathophysiology of several chronic diseases, including cancer, neurodegeneration, and type 2 diabetes (Bond and Hanover, 2013). Increased oxidative stress is closely linked to augmented O-GlcNAcylation in diabetes (Giacco and Brownlee, 2010). In contrast, increased activation of HBP with high-glucose (HG) and glucosamine has been reported to induce mesangial ROS production (Singh et al., 2007). Furthermore, O-GlcNAcylation depletion by shRNA-mediated downregulation of OGT has been shown to prevent HG-induced mesangial ROS production (Goldberg et al., 2011). These findings indicate significant interplay between redox signaling and O-GlcNAcylation modification in diabetes (Lima et al., 2012).
MicroRNAs (miRs) play an essential role in mediating the post-transcriptional regulation of the endothelial oxidative response (Marin et al., 2013). An increasing number of studies has suggested that multiple miRs are involved in the pathogenesis of diabetes (Shantikumar et al., 2012). Particularly, members of the miR-200 family are highly sensitive to ROS. Magenta and colleagues reported that miR-200 family members were increased in H 2 O 2 -stimulated human umbilical endothelial cells, which play a critical role in ROS-mediated senescence and apoptosis by downregulating the zinc-finger E-box binding homeobox 1 (ZEB-1) (Magenta et al., 2011). The five known miR-200 members are located on two chromosomes (chromosome 12: miR-141 and miR-200c; chromosome 1, miR-200a, miR-200b, and miR-429) and are highly conserved in higher vertebrates (Gheldof et al., 2012). Moreover, they can be separated into two functional groups that differ in seed sequences by only a single nucleotide (underlined). The seed sequence 5 ′ -AACACUG-3 ′ defines functional group 1 (miR-141/200a), while 5 ′ -AAUACUG-3 ′ defines functional group 2 (miR-200b/200c/429) (Senfter et al., 2016). The miR-200 family is now known to play an inhibitory role in cancer progression because of the strong suppressive effects of its members on cell transformation, migration, proliferation, invasion, and metastasis (Humphries and Yang, 2015). However, reports of miR-200 family members in diabetes-associated studies are conflicting. Some studies showed that miR-200 family members have harmful effects on diabetes progression (Filios et al., 2014;Belgardt et al., 2015), diabetes-induced inflammation (Reddy et al., 2012), and diabetes-induced endothelial dysfunction (Zhang et al., 2016), while others showed that members of the miR-200 family have protective effects because of their antiinflammatory (McArthur et al., 2011) and anti-fibrotic properties (Wei et al., 2014).
Although ROS is involved in modulating the miR-200 family and protein O-GlcNAcylation, the interplay between these factors is not clear. In this study, we investigated the role of miR-200 members in regulating endothelial inflammation mediated by OGT-induced O-GlcNAcylation in HG-treated human aortic endothelial cells (HAECs) and examined the aortic endothelial tissues in the miR-200a/200b mimics-treated db/db type 2 diabetic mice used as an in vivo model.

Endothelial Cell Culture
HAECs were purchased from Cell Applications, Inc. (San Diego, CA, USA) and cultured in endothelial cell growth medium (Cell Applications, Inc.) as previously described . Most of the chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA) unless otherwise specified. Highglucose (25 mM) stimulation for 12 and 24 h of HAECs was performed in different experiments. L-Glucose solutions (LG, 25 mM) were used as the osmotic control. THP-1 cells (i.e., a human monocytic cell line) were obtained from the American Type Culture Collection (Manassas, VA, USA) and maintained as previously described .

Luciferase Reporter Assay
A segment of the OGT 3 ′ -untranslated region (UTR) that includes both miR-200a and miR-200b binding sites was constructed into the pmirGLO vector (Promega, Madison, WI, USA). The sequence is listed in Supplemental Data 2. HAECs were cotransfected with 1 µg of constructed plasmids and 100 nM of miR-200a mimic, miR-200b mimic, or NC using Lipofectamine TM 2000 (Invitrogen). After 24 h of transfection, the cells were harvested to determine luciferase activity using the Luciferase Assay System Kit (Promega, E1500), as previously described .

OGT Gene Silencing
For OGT gene knockdown, the HAECs were transfected with 100 nM human OGT siRNA (GeneDirex, Keelung, Taiwan), as previously described Pai et al., 2017). Scrambled NC siRNA was included as a negative control. After OGT siRNA transfection, the medium was changed to fresh endothelial cell growth medium, and the HAECs were stimulated with HG for 24 h. After HG treatment, the expression levels of OGT mRNA, OGT, protein O-GlcNAcylation, ICAM-1 mRNA, VCAM-1 mRNA, E-selectin mRNA, and ICAM-1 were determined. THP-1 monocyte adhesion assays were performed after OGT siRNA transfection.

Type 2 db/db Diabetic Mouse Model Experiments
Animal studies were approved by the Hungkuang University Institutional Animal Care and Use Committee with a permit Number: HK 10607. Male db/db diabetic mice were obtained from the National Laboratory Animal Center (Nangang, Taipei, Taiwan). Fourteen-week-old db/db mice were injected with 100 µL miR-200a mimic, miR-200b mimic, or NC (13 µg per week, three times) by tail-vein injection using equal volume mixtures of Lipofectamine TM 2000 and miR-200a mimic, miR-200b mimic, or NC. The non-diabetic control db/m group was administered with equal volume mixtures of vehicle and PBS. Mice were sacrificed by CO 2 narcosis after 3 weeks. Aortic tissue was cautiously excised and fixed with formalin solution. Paraffin sections of the aorta were prepared for immunohistochemistry (IHC) staining.

Statistical Analysis
Statistical analyses were performed using the SPSS 12.0 statistical software (SPSS Inc., Chicago, IL, USA). All data are presented as the mean ± standard error of the mean (SEM). Pair-wise comparisons were carried out using Student's t-test. Three or more groups were compared by one-way analysis of variance with Tukey post-hoc tests. Significant differences were defined as p < 0.05.

HG Increases OGT Expression, Protein O-GlcNAcylation, and Inflammation in the HAECs
First, we examined the effects of HG stimulation on endothelial OGT expression. After 12 and 24 h treatment, HG stimulation caused significant (1.27-and 1.17-fold, respectively) increases in OGT gene expression in the HEACs compared with that in the unstimulated control ( Figure 1A). Similar to its stimulatory effect on OGT gene expression, HG treatment for 12 and 24 h also caused significant increases (1.19-and 1.28-fold, respectively) in OGA gene expression compared with that in the unstimulated control (Supplemental Data 3). In contrast, the osmotic control LG did not modulate the gene expression levels of OGT ( Figure 1B). The expression of OGT protein ( Figure 1C) were also significantly increased after 24 h of HG stimulation ( Figure 1C) and was associated with increased protein O-GlcNAcylation ( Figure 1D). Monocyte adhesion to the endothelium using cellular adhesion molecules (ICAM-1, VCAM-1, E-selectin) initiates endothelial inflammation. HG stimulation for 12 and 24 h caused 1.15-and 1.40-fold increases in ICAM-1 gene expression, respectively ( Figure 1E). Additionally, 12-and 24-h HG stimulation caused 1.11-and 1.27-fold increases in VCAM-1 gene expression and 1.36-and 1.46-fold increases in E-selectin gene expression, respectively ( Figure 1E). The expression of ICAM-1 was also significantly upregulated after 24 h HG stimulation ( Figure 1F) and was associated with a 1.95-fold increase in the adhesion of the THP-1 cells to the HAECs ( Figure 1G).

OGT Is a Direct Target of miR-200a and miR-200b
To identify possible binding partners of the 3 ′ -UTR of human OGT mRNA, in silico analyses using miRanda-mirSVR (www. microrna.org) and miRDB (www.mirdb.org) database with the help of manual sequence alignment were performed. All members of the miR-200 family were showed homology with the 3 ′ -UTR of human OGT mRNA in their seed sequences (Figure 2A), indicating potential regulation of OGT by miR-200 members. Positions 2-8 of the seed sequence of the functional group 1 (5 ′ -AACACUG-3 ′ ), and the positions 2-7 of the functional group 2 (5 ′ -AAUACU-3 ′ ) were predicated to bind to the 3 ′ -UTR of OGT mRNA. As shown in Figure 2B, real-time PCR analyses showed that HG stimulation for 12 h resulted in a significant decrease in miR-200a, miR-141, miR-200b, miR-200c, and miR-429 expression levels to 53, 38, 11, 42, and 33%, respectively, of the levels in the unstimulated control. Similarly, HG stimulation for 24 h caused a significant decrease in the miR-200a, miR-141, miR-200b, miR-200c, and miR-429 expression levels to 27, 18, 39, 71, and 39%, respectively, of the levels in the unstimulated control. Real-time PCR data revealed that the most highly expressed miRs among functional groups 1 and 2 were miR-200a and miR-200b, respectively. Therefore, miR-200a and miR-200b were selected for further analysis. Although HG caused a significant decrease in the miR-200a and miR-200b   -200a, miR-141, miR-200b, miR-200c, miR-429, and the 3 ′ -UTR of human OGT mRNA, indicating potential regulation of OGT by these miRs. | AU or CG.: GU wobble. (B) Real-time PCR showed that HG stimulation for 12 h decreased miR-200a, miR-141, miR-200b, miR-200c, and miR-429 expression levels to 53, 38, 11, 42, and 33% of the control level, respectively. Similarly, 24 h HG stimulation decreased miR-200a, miR-141, miR-200b, miR-200c, and miR-429 expression levels to 27, 18, 39, 71, and 39% of the control level, respectively N = 5. **p < 0.01 and ***p < 0.001 compared with control. miR expression levels were normalized to U6 expression. expression levels, the osmotic control LG did not modulate the expression levels of miR-200a and miR-200b ( Figure 2C). The transcriptional inhibitor ZEB-1 is a well-known target of the miR-200 family members through a reciprocal repression mechanism to promote epithelial-mesenchymal transition and cancer invasion (Burk et al., 2008). To determine whether HGinduced downregulation of miR-200 members also modulates ZEB-1 gene expression, we examined the effects of HG on ZEB-1 expression. After 12-and 24-h stimulation, HG caused significant (1.54-and 1.23-fold, respectively) increases in ZEB-1 gene expression in the HEACs compared with that in the in unstimulated control cells ( Figure 2D). Moreover, miR-200a/200b mimics significantly decreased the ZEB-1 gene expression levels to 83 and 80%, respectively, of the expression in the negative control group after 24 h HG-stimulation in the HAECs (Figure 2E). These results confirm that ZEB-1 is a target of miR-200a/200b. Furthermore, downregulation of miR-200 family members in the HG-stimulated HAECs can modulate the expression levels of ZEB-1. To investigate whether miR-200a/200b can interact with the 3 ′ -UTR of OGT mRNA, we performed a luciferase reporter assay. As shown in Figure 2F, cotransfection of pmirGLO-OGT-3 ′ -UTR and either the miR-200a or miR-200b mimic resulted in a decrease in the luciferase signal to 83 and 74%, respectively, of that in the negative control, confirming direct binding of miR-200a/200b to the 3 ′ -UTR of OGT mRNA.

miR-200a/200b Mimics Inhibit HG-Induced Endothelial OGT Expression, Protein O-GlcNAcylation, and Inflammatory Phenotypes
To examine the potential regulation of OGT expression by the miR-200 family, transfection assays with 200a/200b mimics or inhibitors were performed to determine whether miR-200a/200b mimics modulates OGT gene expression. As shown in Figure 3A, transfection of miR-200a/200b mimics in HGstimulated HAECs caused a significant decrease in the OGT gene expression level to 83 and 82%, respectively, of that in the negative control ( Figure 3A). Interestingly, transfection of miR-200a mimics in the HG-stimulated HAECs caused a nonsignificant decrease in OGA expression, to 95% (p = 0.089) of that in the negative control. Transfection of 200b mimics in the HG-stimulated HAECs caused a significant decrease in OGA expression, to 91% (p = 0.007) of that in the negative control (Supplemental Data 4). In contrast, transfection of miR-200a/200b inhibitors in the HG-stimulated HAECs caused a significant increase in OGT gene expression level, by 1.20-and 1.23-fold, respectively, of that in the negative control ( Figure 3B). Transfection of miR-200a/200b mimics also significantly downregulated OGT protein expression ( Figure 3C) and was associated with decreased protein O-GlcNAcylation in the HG-stimulated HAECs as compared with the negative control group (Figure 3D). To determine whether reduced OGT and protein O-GlcNAcylation expression are associated with reduced endothelial inflammation in the miR-200a/200b  mimics-transfected HG-stimulated HAECs, we measured ICAM-1, VCAM-1, and E-selectin gene expression levels, ICAM-1 expression, and THP-1 monocyte adhesion. As shown in Figure 3E, transfection of miR-200a/200b mimics respectively inhibited HG-induced endothelial ICAM-1 mRNA expression to 83 and 81%, VCAM-1 mRNA expression to 86 and 87%, and E-selectin mRNA expression to 74 and 70%, of the negative control expression levels. In contrast, transfection of miR-200a/200b inhibitors respectively increased HG-induced endothelial ICAM-1 mRNA expression to 1.25-and 1.32-fold, VCAM-1 mRNA expression to 1.28-and 1.35-fold, and E-selectin mRNA expression to 1.15-fold and 1.20-fold, of the negative control expression levels ( Figure 3F). ICAM-1 expression was also significantly decreased by transfection with miR-200a/200b mimics in the HG-stimulated HAECs ( Figure 3G) and was associated with decreased adhesion of the THP-1 cells to the HAECs, to 33 and 34%, respectively, of the adhesion levels in the negative control group (Figure 3H). These results indicate that miR-200a/200b mimics possess anti-inflammatory properties in HG-induced endothelial inflammation, possibly because of their ability to modulate protein O-GlcNAcylation levels via OGT downregulation.

OGT Depletion by siRNA Inhibits HG-Induced Endothelial Protein O-GlcNAcylation and Inflammation
To determine whether increased OGT expression was associated with HG-induced protein O-GlcNAcylation and inflammation, an OGT-specific siRNA transfection experiment was performed. As shown in Figure 4A, OGT-specific siRNA transfection caused a significant decrease in OGT mRNA levels-to 72% of the levels in the cells transfected with the scrambled negative control. Transfection of OGT siRNA also significantly attenuated OGT protein expression ( Figure 4B) and was associated with decreased protein O-GlcNAcylation (Figure 4C), confirming that HG-induced endothelial O-GlcNAcylation was mediated partially through OGT. To determine whether reduced OGT and protein O-GlcNAcylation expression were associated with reduced endothelial inflammation in OGT-depleted HGstimulated HAECs, we measured the ICAM-1, VCAM-1, and Eselectin gene expression levels, ICAM-1 expression, and THP-1 monocyte adhesion. Similar to the effects of the miR-200a/200b mimics, OGT siRNA transfection caused a significant decrease in ICAM-1, VCAM-1, and E-selectin mRNA levels-to 58, 71, and 67%, respectively, of that in the scrambled negative control ( Figure 4D). ICAM-1 expression also significantly decreased following transfection of OGT siRNA in the HG-stimulated HAECs ( Figure 4E) and was associated with decreased THP-1 cell adhesion to the HAECs-to 24% of the adhesion levels in the scrambled negative control group ( Figure 4F)-indicating that HG-induced endothelial inflammation is mediated in part through OGT.

miR-200a/200b Mimics Decreased Endothelial OGT and ICAM-1 Expression in Type 2 Diabetic Mice
To examine the effect of miR-200a/200b mimics on the expression of endothelial OGT and ICAM-1 in db/db type 2 diabetic mice, we performed IHC staining of the aortic tissues. As shown in Figures 5A,B, the immunoreactivities of OGT and ICAM-1 in the aortic endothelial layers were substantially downregulated in the miR-200a/200b mimic-treated db/db mice than in the negative control-treated db/db mice. In the endothelial layers, quantification of immunoreactivity signals showed that OGT expression was decreased significantly-to 0.37-and 0.43-fold in the db/db (miR-200a mimic) and db/db (miR-200b mimic) groups compared with that in the db/db (negative control) group. ICAM-1 expression in the endothelial layers was also decreased significantly-to 0.34-and 0.33fold in the db/db (miR-200a mimic) and db/db (miR-200b mimic) groups compared with that in the db/db (negative control) group ( Figure 5C). This suggests that miR-200a/200b mimics may play a therapeutic role by decreasing diabetesinduced endothelial inflammation through the downregulation of both OGT and ICAM-1. The potential mechanisms of miR-200a/200b in regulating HG-induced endothelial OGT expression, protein O-GlcNAcylation, and inflammation is illustrated in Figure 6.  and stress duration. Chen et al. reported that glucosamine and PUGNAc (an OGA inhibitor) inhibited ICAM-1 expression, cell adhesion, and nuclear factor-κB activity in tumor necrosis factor-α-stimulated human pigment epithelial cells . Hilgers et al. reported that glucosamine and Thiamet-G (an OGA inhibitor) suppressed tumor necrosis factor-α-induced hypocontractility and endothelial dysfunction by inhibiting tumor necrosis factor-α-induced inducible nitric  oxidase synthase expression and peroxynitrite production in the rat aortic ring (Hilgers et al., 2012). Xin et al. reported that glucosamine treatment for 14 days inhibited neointimal formation in the injured carotid artery in a rat acute balloon injury model (Xing et al., 2008). These studies support that O-GlcNAcylation has a protective role in acute stress.

DISCUSSION
In contrast, O-GlcNAcylation has been implicated in the development of cardiovascular complications in diabetes-related studies (Issad et al., 2010;Lima et al., 2012). James et al. reported that nuclear factor-κB-dependent promoter activation occurred because of increased flux in HBP in rat mesangial cells (James et al., 2002). Du et al. reported that HG-induced activation of HBP increased O-GlcNAcylation of Sp-1, Sp-1 transactivation, and Sp-1-dependent plasminogen activator inhibitor-1 and transforming growth factor-β gene expression in bovine aortic endothelial cells (Du et al., 2000). Federici et al. reported that HG and glucosamine-induced O-GlcNAcylation impaired insulin signaling by downregulating endothelial nitric oxide synthase phosphorylation and activity in human coronary endothelial cells (Federici et al., 2002). Makino et al. reported that OGA overexpression significantly decreased endothelial protein O-GlcNAcylation, and simultaneously increased left ventricle capillary density and restored the endotheliumdependent relaxation in streptozotocin-induced type 1 diabetic mice (Makino et al., 2015). This study demonstrates that HG induced the expression of the essential O-GlcNAcylation enzyme OGT. Importantly, we observed a crucial role for increased OGT expression and protein O-GlcNAcylation in the regulation of endothelial inflammation, as HG-induced ICAM-1, VCAM-1, and E-selectin mRNA expression; ICAM-1 expression; and THP-1 monocyte adhesion were reduced after OGT depletion by siRNA. Interestingly, our data showed that HG increased OGA mRNA. In addition, the miR-200b mimic not only decreased the expression levels of protein O-GlcNAcylation, but also decreased OCA mRNA. Previously, Zhang et al. reported that OGA mRNA and protein expression is sensitive to cellular protein O-GlcNAcylation levels . Our data also suggest that OGA plays a significant role in maintaining protein O-GlcNAcylation homeostasis by responding to the cellular protein O-GlcNAcylation levels (i.e., increase protein O-GlcNAcylation causing increase OGA expression and vice versa).
In this study, web-based bioinformatics analysis using miRanda-mirSVR and miRDB database revealed that the members of the miR-200 functional group 1, but not of group 2, may bind to the 3 ′ -UTR of human OGT mRNA. However, the miRanda-mirSVR database identified all five miR-200 family members as potential binding partners of the mouse Ogt 3 ′ -UTR. Based on manual sequence alignment, both human OGT and mouse Ogt 3 ′ -UTRs share an identical nucleotide sequence (5 ′ -AGUAUU-3 ′ ) that is complementary to the seed sequence of miR-200 functional group 2, suggesting that human OGT 3 ′ -UTR also binds to miR-200b/200c/429, similar to mouse Ogt. Alternatively, G:U wobble pairing may explain the binding of miR-200b (seed sequence: 3 ′ -GUCAUAA-5 ′ ) to the miR-200a binding site on OGT 3 ′ -UTR (5 ′ -CAGUGUU-3 ′ ), as Doench et al. reported that thermodynamically favorable G:U pairing in the seed region can retain significant gene repressive effects (Doench and Sharp, 2004). In this study, miR-200a/200b mimics regulated OGT gene and protein expression and protein O-GlcNAcylation. In addition, a luciferase reporter assay confirmed that miR-200a/200b mimics directly interact with the 3 ′ -UTR of OGT mRNA. These findings support that the miR-200a::OGT and miR-200b::OGT interactions are functional. While we did not investigate the role of miR-141 and miR-200c/429 in regulating OGT expression, it is possible that all members of the miR-200 family can regulate OGT expression, as they have the same seed sequence as either miR-200a or miR-200b.
Previously, Liu et al. reported that ICAM-1 is a direct target of miR-141 in the human umbilical endothelial cells (Liu et al., 2015). They further examined the role of an intravenous miR-141 mimic in a mouse model of myocardial ischemia-reperfusion injury and found that the miR-141 mimic attenuated myocardial ICAM-1 levels and infarction size (Liu et al., 2015). Although we did not examine the repressive effect of miR-141 on ICAM-1 expression, we found that ICAM-1 gene and protein expression and monocyte adhesion were significantly downregulated by the miR-200a mimic. Given that miR-200a and miR-141 have the same seed sequence, miR-200a may also regulate ICAM-1 expression in the HG-stimulated HAECs. In addition, similar to OGT 3 ′ -UTR, in silico analysis (www.microrna.org) did not identify functional group 2 members as targeting ICAM-1; however, we found that the miR-200b mimic significantly decreased ICAM-1 expression in HAECs. This suggests that either miR-200b can bind the miR-200a binding site of the ICAM-1 3 ′ -UTR through G:U wobble binding, or ICAM-1 was indirectly repressed by miR-200b. Interestingly, the miR-200 family members were not predicted to bind to the mouse ICAM-1 3 ′ -UTR; however, our results show that miR-200a/200b mimics significantly attenuated endothelial ICAM-1 expression, suggesting that ICAM-1 expression was indirectly regulated by miR-200a/200b mimics in the db/db diabetic mice. Based on OGT siRNA data, the ability of miR-200a/200b mimics to downregulate OGT expression and O-GlcNAcylation levels may act as a possible mechanism to attenuate hyperglycemiainduced endothelial inflammation by regulating nuclear factor-kB activity (James et al., 2002;Yang et al., 2008;Allison et al., 2012).
There is conflicting evidence regarding the regulatory role of miR-200 family members in diabetes. The miR-200 family has been reported to have a detrimental effect. Reddy et al. reported that vascular smooth muscle cells and aortic tissue from db/db mice showed increased miR-200b/429 expression, and an miR-200b mimic increased monocyte binding to db/+ vascular smooth muscle cells, whereas an miR-200b inhibitor decreased monocyte binding in the db/db vascular smooth muscle cells (Reddy et al., 2012). Belgardt et al. reported that β-cell specific overexpression of miR-200 family induced β-cell apoptosis and type 2 diabetes, whereas miR-200 family knockout reduced β-cell apoptosis and improved type 2 diabetes symptoms (Belgardt et al., 2015). Zhang et al. reported that miR-200c expression was increased in the diabetic mouse aorta and HG-stimulated mouse aortic endothelial cells. Moreover, they noted that a miR-200c mimic impaired endothelial-dependent relaxation in the non-diabetic mouse aorta, whereas an miR-200c inhibitor enhanced endothelial-dependent relaxation in the diabetic db/db mice (Zhang et al., 2016).
In contrast, the miR-200 family has also been reported to have protective effects. Wei et al. reported that tailvein injection of anti-miR-200a in streptozotocin-induced diabetic mice exacerbated cortical and glomerular fibrosis and increased urinary albumin excretion (Wei et al., 2014). Vascular endothelial growth factor plays an important role in the pathogenesis of diabetic retinopathy (Chiefari et al., 2016). McArthur et al. observed miR-200b downregulation and vascular endothelial growth factor (a target of miR-200b) upregulation in streptozotocin-induced diabetic rat retinas. Furthermore, intraocular injection of an miR-200b mimic prevented an increase in vascular endothelial growth factor expression and vascular permeability. Importantly, human retinal tissue from diabetic enucleated eyes showed lower miR-200b expression than non-diabetic retina (McArthur et al., 2011). Mitra et al. also reported the application of miR-200b nanoparticles for treating diabetic retinopathy in mice (Mitra et al., 2016). These findings suggest that the miR-200 family has cytoprotective roles, consistent with its role as a tumor suppressor. Previous studies reported that O-GlcNAcylation may be either beneficial or harmful, depending on the experimental design (e.g., type of species or tissue, acute vs. chronic stress, diabetic vs. non-diabetic samples), suggesting that the miR-200 family has various roles in inflammation-associated diseases, including diabetes. However, this study confirmed that the miR-200 family is dysregulated in HG-stimulated HAECs, and both in vitro and in vivo experiments revealed that miR-200a/200b have therapeutic potential in treating diabetic vascular disease.