Epigenetic down‐regulation of microRNA‐126 in scleroderma endothelial cells is associated with impaired responses to VEGF and defective angiogenesis

Abstract Impaired angiogenesis in scleroderma (SSc) is a critical component of SSc pathology. MicroRNA‐126 (miR‐126) is expressed in endothelial cells (MVECs) where it regulates VEGF responses by repressing the negative regulators of VEGF, including the sprouty‐related protein‐1 (SPRED1), and phosphoinositide‐3 kinase regulatory subunit 2 (PIK3R2). MVECs were isolated from SSc skin and matched subjects (n = 6). MiR‐126 expression was measured by qPCR and in situ hybridization. Matrigel‐based tube assembly was used to test angiogenesis. MiR‐126 expression was inhibited by hsa‐miR‐126 inhibitor and enhanced by hsa‐miR‐126 Mimic. Epigenetic regulation of miR‐126 expression was examined by the addition of epigenetic inhibitors (Aza and TSA) to MVECs and by bisulphite genomic sequencing of DNA methylation of the miR‐126 promoter region. MiR‐126 expression, as well as EGFL7 (miR‐126 host gene), in SSc‐MVECs and skin, was significantly down‐regulated in association with increased expression of SPRED1 and PIK3R2 and diminished response to VEGF. Inhibition of miR‐126 in NL‐MVECs resulted in reduced angiogenic capacity, whereas overexpression of miR‐126 in SSc‐MVECs resulted in enhanced tube assembly. Addition of Aza and TSA normalized miR‐126 and EGFL7 expression levels in SSc‐MVECs. Heavy methylation in miR‐126/EGFL7 gene was noted. In conclusion, these results demonstrate that the down‐regulation of miR‐126 results in impaired VEGF responses.

FLT-1) and VEGFR2 (FLK1/KDR) on endothelial cells, which result in phosphorylation of ERK1/2 MAPK, Akt and p38 MAPK, leading to endothelial cell proliferation and migration. [6][7][8][9] Several studies have shown that the expression of VEGFA, VEGFR1 and VEGFR2 is markedly up-regulated in SSc. [10][11][12][13][14] However, adaptive angiogenesis is absent despite the progressive loss of capillaries. 4 MiR-126, encoded by an intron of the EGF-like domain (EGFL7) gene, is abundantly expressed in the endothelium. MiR-126 regulates angiogenic signalling by regulating responses to VEGF in MVECs in part by direct repression of negative regulators of the VEGF signalling pathway, including the sprouty-related EVH1 domain containing 1 (SPRED1) and phosphoinositide-3 kinase regulator subunit 2 (PIK3R2), which negatively regulate VEGF signalling via the RAF1-MAP kinase and PI3 kinase pathways, respectively. SPRED1 and PIK3R2 are validated direct targets of miR-126, and the defective expression of miR-126 results in diminished responses to VEGF signalling and impaired angiogenesis. [15][16][17] Decreased expression levels of EGFL7, miR-126 host gene, were reported in SSc-MVECs. 18 In this study, we examined the expression levels of miR-126 in normal and SSc skin and MVECs. We also investigated the effects of miR-126 on the gene expression levels of SPRED1 and PIK3R2 and VEGF-dependent tube formation and migration in normal and SSc-MVECs. Moreover, we also inspected the effects of epigenetic regulators on miR-126 gene expression and the promoter DNA methylation status of the miR-126 gene in SSc-MVECs.

| MATERIAL S AND ME THODS
Cell and Cell culture. This study was approved by the University of Toledo Institutional Review Board. A five-mm skin biopsy was obtained from the forearm of healthy volunteers and patients with diffuse cutaneous SSc (n = 6) after obtaining a signed written consent form. MVECs were purified by CD31 magnetic beads as previously described. 19,20 The purity of isolated cells was >98% as determined by flow cytometry analysis using PE anti-human CD31. Cells were used at the 3-5th passage in the experiments. Fluorescent in situ hybridization (FISH) and immunofluorescent (IF) labelling. FISH and IF were performed as described previously. 21,22 Expression levels of hsa-miR-126 in MVECs identified by positive CD31 in skin biopsies were detected by miRCURY LNA miRNA ISH Optimization Kit 5 (FFPE) (miR-126) (Qiagen, Germantown, MD, USA).
In situ hybridization reaction was used with the DIG-labelled LNA miR-126 probe, a locked nucleic acid oligonucleotide probe labelled at both 5′ and 3′ ends with digoxigenin complementary to human miR-126. Immunologic detection was done with sheep anti-DIG-POD, Fab fragments (sigma) for miR-126 and primary antibody rabbit Anti-CD31 (Abcam) for labelling endothelial cells by incubating slides overnight at 4•C. Then, the slides were incubated at room temperature for 1 hour with Anti-rabbit-Alexa Fluor 594 (Invitrogen), and LNA miR-126 probes were labelled with FITC using a tyramide signal amplification (TSA) system (Perkin Elmer). Finally, the slides were mounted with ProLong  Matrigel tube assembly assay. Matrigel tube assembly assays were performed as previously described. 5 The per cent wound closure was measured by ImageJ. The cell migration was expressed as percentage wound closure (total area-area not occupied by the cells/total area ×100).
Endothelial Cell migration assay. The migration assay was per-

| Abundant miR-126 expression in control MVECs and reduced expression levels in SSc-MVECs and skin
The expression levels of miR-126 were assessed in control MVECs, control HDFBs (human dermal fibroblasts), control HDSMCs (human dermal smooth muscle cells) and SSc-MVECs by qPCR. The expression levels of miR-126 in NL-MVECs were approximately 500 times higher than in HDFBs and 5000 times higher than in HDSMCs ( Figure 1A). The data confirmed that miR-126 is expressed mainly in MVECs. MiR-126 expression was significantly down-regulated in SSc-MVECs by 0.16 ± 0.03 folds, compared to control ( Figure 1B; P < 0.01). MiR-126 expression levels were also examined in freshly isolated RNA obtained from skin biopsies of SSc and control subjects by qPCR (n = 3). Significant reduction in miR-126 expression was noted in SSc skin by 0.27 ± 0.06 folds compared to control samples ( Figure 1C; P < 0.01). Moreover, the expression of miR-126 was examined in paraffin sections of skin biopsies by in situ hybridization followed by quantitative densitometry analysis using ImageJ ( Figure 1D-1E). Co-localization of miR-126 and endothelial-specific marker CD31 was observed in skin biopsies ( Figure 1D). Similarly, the miR-126 expression levels in SSc-MVECs were also significantly down-regulated by 2.32 folds in SSc skin paraffin sections when compared to control ( Figure 1E; P < 0.01).

| Diminished SSc-MVECs VEGF-dependent angiogenesis responses
To examine the responses of MVECs to VEGF-induced angiogenesis, NL-MVECs and SSc-MVECs were plated onto Matrigel to investigate capillary tube assembly and plated into fibronectin-coated 6well plates and Corning FluoroBlok™ 96-well insert to test MVEC migration.
For Matrigel assay, MVECs were labelled with calcein AM and observed under a microscope. The capillary morphogenesis was quantified by measuring the total tube segment length. Figure 1F  Moreover, the tube length in SSc-MVECs was also significantly lower than in control MVECs at baseline levels ( Figure 1G; P < 0.05), which suggests that SSc-MVECs also have an impaired angiogenesis response to the low amount of growth factors in the Matrigel or to Matrigel itself.
In the scratch test, the addition of VEGF enhanced control MVECs migration ( Figure 1I) and resulted in 51.34 ± 8.64% wound closure in 24 hours versus 3.86 ± 0.62% in unstimulated control ( Figure 1J;  3). F and G, VEGF induced tube formation (F). Tube length expressed in micrometres (G). VEGF induced significant increase in tube length in normal MVECs, but failed to do so in SSc-MVECs (n = 6). Bar = 1000 μm. I-J, Scratch-wound assay analysis of NL and SSc-MVECs response to VEGF expressed as %wound closure. VEGF induced significant wound closure in normal MVECs compared to its untreated control (n = 6), but was unable to induce in SSc-MVECs. Bar = 200 μm. H, Cell migrations were quantitated using calcein AM labelling. SSc-MVECs did not exhibit cell migration under VEGF stimulation (n = 6). All values are expressed as mean ±SD. *P < 0.01, labelled group versus the control group. **P < 0.05, labelled group versus NL-MVECs/ctr. P < 0.05 was considered significant ( Figure 1H; P < 0.01), whereas SSc-MVECs with VEGF stimulation exhibited similar migration value to baseline values 1.35 ± 0.2 folds versus 1 ± 0.14 folds in control-SSc-MVECs ( Figure 1H; P >.05).

| Reduced miR-126 expression in SSc-MVECs is associated with the up-regulation of SPRED1 and PIK3R2
To explore the role of miR-126 in the defective VEGF-dependent angiogenesis in SSc-MVECs, we searched for potential direct mRNA  Figure 2B). These data suggested that the reduced miR-126 expression in SSc-MVECs is associated with up-regulation of SPRED1 and PIK3R2 expression ( Figure 2B). Western blot analysis also confirmed that conclusion on the protein levels ( Figure 2C).

| Down-regulation of miR-126 expression enhances SPRED1 and PIK3R2 and impairs angiogenesis response to VEGF in normal endothelial cells
To further validate that the up-regulation of SPRED1 and PIK3R2 was  Figure 3F). Similarly, NL-MVECs with knockdown miR-126 showed impaired migration potential after addition of VEGF with the migration rate at 1.26 ± 0.19 folds versus 1 ± 0.11 folds in control (P >.05; Figure 3G).

| Reduced expression of miR-126 in SSc-MVECs impaired phosphorylation response of ERK and AKT to VEGF
To examine the responses of MVECs to VEGF-induced activation of MAPK and AKT angiogenesis signalling pathways, NL-MVECs and SSc-MVECs were subjected to serum and growth factor withdrawal overnight and then subjected to VEGF stimulation at 50ng/ml for 15 minutes. Cell lysates were immunoblotted with the antibodies to determine the level of phosphorylated and total ERK, and phosphorylated and total AKT. Results showed that the VEGF stimulated To further test if miR-126 is an essential regulator for phosphorylation of AKT and ERK in response to VEGF, we transfected NL-MVECs with miR-126 inhibitor and examined the activation of MAP kinase and PI3 kinase by VEGF stimulation. As shown in Figure 5C

| The reduction of miR-126 in SSc-MVECs is linked with the hypermethylation of miR-126 promoter region
To explore the mechanism by which miR-126 is down-regulated in SSc-MVECs, we analysed the expression and the promoter methylation status of EGFL7 gene. MicroRNA-126 is an intronic micro-RNA, located within the intron of the EGFL7 locus, and mature miRNA-126 is produced from the processing of EGFL7 pre-mRNA transcript rather than from its promoter. 25 Western blot and realtime PCR analysis revealed that EGFL7 expression was significantly down-regulated in SSc-MVECs compared to healthy controls both at mRNA level (P < 0.01)) and at protein level ( Figure 6A), which were consistent with the down-regulated miR-126 expression in SSc-MVECs ( Figure 1C). These results support that miR-126 and EGFL7 share the same promoter, and their expression levels were controlled by the EGFL7 promoter in MVECs. EGFL7 expression levels may be considered as a biomarker for the miR-126 expression in tissue.
To investigate whether an epigenetic mechanism mediates underexpression of miR-126 in SSc-MVECs, NL-MVECs and SSc-MVECs were treated with the DNA methyltransferase inhibitor Aza the other is a CG-containing Ets1-binding site. The third one is a CG-containing EGR-binding site. All these sites are methylated in all three SSc samples, and they correspond to the 1st, 6th and 12th CGs ( Figure 6D). Previous studies reported that transcription factors binding to these sites are essential for the transcriptional regulation of EGFL7, [26][27][28][29] and therefore, the methylation of these sites of the promoter region can hinder transcriptional factor binding and repress the transcription of EGFL gene and miR-126 gene in SSc-MVECs. 30

| D ISCUSS I ON
The results of this study show that down-regulation of miR-126 is associated with impaired SSc endothelial cell responses to VEGF. We angiogenesis in cell-type and strand-specific manner. 15,16,31,32 Computational algorithms predicted that the specific genes, SPRED1 and PIK3R2, are potential targets of miR-126. It was also reported that miR-126 regulated and controlled the expressions relevant to other genes including PTPN9, PTEN, SDF-1, VCAM-1, F I G U R E 5 Impaired phosphorylation of ERK and AKT to VEGF in SSc-MVEC is associated with the reduced expression of miR-126 The phosphorylated and total ERK1/2 and AKT were assessed by Western blotting. The protein levels were quantitated by NIH ImageJ. The total ERK and AKT were used as the protein loading control for p-ERK and p-AKT, respectively. Values are fold change compared to the control without VEGF treatment. ECs were starvation overnight and then were stimulated with and without VEGF 50 ng/ ml for 15 minutes. A, B, Diminished activation of ERK and AKT in response to VEGF were seen in SSc-MVECs, while VEGF significantly increased the phosphorylation of ERK and AKT in NL- Phosphorylation of ERK1/2 MAPK, AKT and p38 MAPK by VEGFR2 activation appears to be necessary for stimulation of angiogenesis in endothelial. [7][8][9] MiR-126 has been shown to promote MAP kinase and PI3K signalling in response to VEGF and FGF by targeting negative regulators of these signalling pathways, including SPRED1 and PIK3R2. 15,16 SPRED1 negatively regulates the activation of the MAP kinase pathway by binding and inactivating RAF1, an upstream kinase of the pathway. [40][41][42] PIK3R2 acts as a suppressor of the PI3K/AKT signalling pathway activation, 33,35 and knockdown of PIK3R2 rescued the defect in VEGF-dependent phosphorylation of AKT in miR-126 knockdown HUVECs. 15 Consistent with these data, we found that the phosphorylation of ERK1/2 and AKT in response to VEGF is reduced in SSc-MVECs.
Similar responses were also noticed in miR-126 knockdown in NL-MVECs, while phosphorylation of ERK1/2 and AKT in response to VEGF was restored to normal levels in SSc-MVECs after transfection with miR-126 mimic. These data indicated that down-regulated miR-126 in SSc-MVEC impaired the VEGF-induced ERK and AKT

activation.
MicroRNA-126 is largely an endothelial-specific miRNA that is located within intron 7 of EGFL7. 15,16 MiR-126 and its host transcript, EGFL7, are highly expressed in endothelial cells. MiR-126 originates from the EGFL7 pre-mRNA. Usually, an intronic miRNA tends to be co-expressed with its host gene. 16,25,43 A previous study showed that the EGFL7 expression levels were significantly decreased in SSc-MVECs. EGFL7 transcript knockout mice were reported to display vascular abnormalities remarkably similar to those of miR-126 null mice, 44  It is known that an epigenetic mechanism is associated with the repression of the gene. 45,46 Methylation of promoter CpGs is thought to contribute to repression through two mechanisms: (1) direct inhibition of transcription factor binding which is necessary for recruitment of the transcription machinery (represented by RNA polymerase II), and (2) attraction of MeCPs (methyl CpG binding proteins) which associate with co-repressors such as histone deacetylases. 47 5-aza-2'-deoxycytidine, an inhibitor of all DNA methyltransferases that inhibits DNA remethylation after DNA replication. 48 Gene expression is also regulated by histone acetylation and de-acetylation. Moreover, our data show some methylated CpG islands overlap with the transcription factor binding sites, including SP1, ERG1 and Ets1 which regulate the expression of EGFL7 and miR-126 in endothelial cells. [26][27][28][29] Hypermethylation of miR-126 promoter in SSc-MVECs leads to the transcription repression of EGFL7 /miR-126 gene through obstructing TFs' binding and the recruitment of MeCP transcription repressing complex.
Some of the limitations of our study include a lack of testing of other angiogenic factors. We also did not evaluate the role of the antiangiogenic VEGF165b splice variant that was reported to be the limiting factor in SSc-MVECs angiogenic response in SSc. 49 Also, we did not investigate the relative role of histone alteration in the observed down-regulation of EGFL7 /miR-126 gene expression.
In conclusion, our findings show that SSc-MVECs express reduced levels of miR-126 and that miR-126 is required for angiogenesis in SSc-MVECs. MiR-126 enhances the VEGF-induced angiogenesis in SSc-MVECs by targeting SPRED1 and PIK3R2, which are the negative regulator of Raf-ERK and PI3K-AKT signalling separately. Extensive CpG sites' methylation was found in miR-126 promoter region in SSc-MVECs. These results may provide new insights into the pathogenesis of defective angiogenesis and vascular repair in SSc. Administration of proangiogenic miR-126 or molecular regulation of miRNA-126 expression might represent potential therapeutic approaches to promote effective angiogenesis and capillary regeneration in SSc.

CO N FLI C T O F I NTE R E S T
There are no financial support or other benefits from commercial sources for the work reported on in the manuscript, or any other financial interests that any of the authors may have, which could create a potential conflict of interest or the appearance of a conflict of interest regarding the work.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.