MicroRNA let‐7a regulates angiogenesis by targeting TGFBR3 mRNA

Abstract Angiogenesis has a great impact on human health, owing to its participation in development, wound healing and the pathogenesis of several diseases. It has been reported that let‐7a is a tumour suppressor, but whether it plays a role in angiogenesis is unclear. Here we showed that let‐7a, a microRNA conserved in vertebrates, regulated angiogenesis by concomitantly down‐regulating TGFBR3. Overexpression of let‐7a or knockdown of TGFBR3 in cell culture inhibited the tube formation and reduced migration rate. Moreover, xenograft experiments showed that overexpression of let‐7a or knockdown of TGFBR3 had smaller tumour size. Downstream genes, such as VEGFC and MMP9, were also down‐regulated in let‐7a overexpression or TGFBR3 knockdown groups. Therefore, our results revealed a novel mechanism that let‐7a regulate angiogenesis through post‐transcriptional regulation of TGFBR3.

namely TGFBR1 and TGFBR2, and therefore activate signalling transduction through phosphorylation of SMAD proteins. [6][7][8] To our interest, Tgfbr3-null mice died at embryonic day 14.5 due to defective coronary vasculogenesis, of which few evident vessels and persistent blood islands were found throughout the epicardium. 9 This finding suggests that TGFBR3 may play an important role in regulation of angiogenesis.
MicroRNAs (miRNAs) are a class of small non-coding RNAs,~22 nucleotides in length, which regulate gene expression by binding to the 3′-untranslated regions (3′UTR) of target mRNAs and causing subsequent degradation of the target mRNA and/or inhibition of mRNA translation. 10,11 The lethal-7 (let-7) gene family was initially discovered as an essential developmental module in Caenorhabditis elegans and later determined to act as miRNAs in other species. 12 Follow-up studies have comprehensively described the diverse biological functions of the 12 members of the let-7 miRNA family, including cell proliferation, 13 cell differentiation, 14 tumorigenesis 15,16 and angiogenesis. 17 It has been recently shown that let-7a, along with other miRNAs may favour the "Off" switch of tumour angiogenesis in a mouse model of breast tumour undergoing hormone therapy. 18 Intriguingly, TGFBR1, a member of TGFβ family, identified as potential target of let-7, was critical for endothelium inflammation and cell differentiation. [19][20][21][22][23] In this study, we focused on identification of let-7a-regulated genes involved in angiogenesis, particularly the TGFβs signalling pathway, 24 and found TGFBR3 to be a novel target for let-7a. Also, let-7a may impair angiogenesis via post-transcriptional regulation of TGFBR3. Our data likely identify a mechanistic explanation for the anti-angiogenesis function of let-7a and provide a potential therapeutic target for anti-or pro-angiogenesis strategies in cancers and other diseases.

| Animals, cell culture and transfection
Nude mice (7-8 weeks of age) were purchased from the Laboratory

| Plasmid construction
To construct the luciferase reporter vectors, a wild-type 3′UTR (3′ UTR-WT) fragment of human and mouse TGFBR3 containing putative binding sites for let-7a was amplified from genomic DNA (amplification primers are listed in Table 1). The amplified fragment (1229 bp and 349 bp in length respectively) was first inserted into pGEMT vector for site-directed mutagenesis. The mutant 3′UTR (3′ UTR-MUT) of TGFBR3, carrying a mutated sequence in the seeding region of let-7a, was mutated (Primers are listed in Table 1

| Luciferase reporter assay
For the luciferase reporter gene assay, HEK293T cells plated in a 24-well plate were co-transfected with 50 nmol/L of either Con mimics or let-7a mimics with 500 ng reporter comprising 3′UTR-WT or 3′UTR-MUT. Two days after the transfection, cells were lysed in  primary antibodies were used.

| Tube formation assay
The HUVECs tube formation assay was performed with Matrigel (BD Bioscience, Franklin Lakes, NJ, USA) according to the manufacturer's instructions. The sterile 96-well plates were coated with 50 μL Matrigel and incubated at 37°C for 1 hour to form gels.
Let-7a mimics-transfected (or siRNA against TGFBR3) and control HUVECs (1 × 10 5 ) were seeded into each well and incubated with cell-free culture supernatants from HUVECs at 37°C in a 5% CO 2 incubator for 6 hours. The degree of tube formation was evaluated using an inverted microscope (Olympus, Tokyo, Japan) and the number of tubes was calculated using Image J software (Media Cybernetics Inc.).

| Wound-healing assay
HUVECs were treated with miRNA/siRNA as described above.
Wounds were introduced by scrapping with a pipette tip after a 48 hours incubation when a monolayer culture was formed, and microscope photography captured images immediately (0 hour) and 6 hours after. The wound gaps at both time points were measured to be represented as a migration index, for example the distance migrated by let-7a mimics/TGFBR3 siRNA-treated cells relative to the distance migrated by NC-treated cells. 25

| Murine tumorigenesis
To generate hemangioendothelioma, the injection of MS1 cells in mice was performed as previously described with minor modifications. 26

| Statistical analysis
All experiments were independently repeated at least three times.
Quantitative RT-PCR (qRT-PCR), luciferase reporter assays and tube formation assays were performed in triplicate. All data were presented as mean ± SEM. The statistical analysis was performed using IBM SPSS Statistics 20 (IBM, Armonk, NY, USA) and GraphPad Prism 6 (Graphpad, La Jolla, CA, USA). The differences were considered statistically significant at P < 0.05 based on Student's t tests.

| RESULTS
3.1 | Identification of TGFBR3 as a novel target of let-7a We chose a multitude of candidate genes with high miRNA:mRNA alignment scores predicted by TargetScan and miRanda Web Servers. TGFBR3 emerged as a potential target gene of let-7a after F I G U R E 1 Identification of TGFBR3 as a novel target of let-7a. A, qRT-PCR analysis of the TGFBR3 mRNA level was significantly reduced (*P < 0.05). B, Western blot analysis showed reduction in TGFBR3 protein. C, The results of Western blotting were calculated using Image J and Vinculin was used as an internal control. Negative control expression was set as 100% (*P < 0.05). D, A schematic of the computational predicted seed region in the 3′UTR of TGFBR3 was shown, as well as the mutated sequences used in this study. E, HEK293T cells were cotransfected with either let-7a mimics or Con mimics and pmirGLO Vector comprising TGFBR3 3′UTR-WT or -MUT. The relative firefly luciferase activity normalized with Renilla luciferase was measured 48 h after transfection (***P < 0.001, ns means no significance). F, Schematic of let-7a potential binding site in Tgfbr3 3′UTR region. G, The relative luciferase activity mediated by let-7a and TGFBR3 3′UTR/ mutated 3′UTR (*P < 0.05, ns means no significance) screening candidate gene expression with qRT-PCR of let-7a mimics-transfected HEK293T cells. We observed approximately 60% reduction in TGFBR3 mRNA levels after overexpression of let-7a ( Figure 1A). Reduced mRNA also resulted in decreased TGFBR3 protein levels ( Figure 1B and C).
In order to verify that TGFBR3 is a direct target gene of let-7a, we cloned the 3′ untranslated region (3′UTR-WT) of TGFBR3 mRNA into the pmirGLO vector and then mutated the putative let-7a binding site by site-directed mutagenesis (3′UTR-MUT), as indicated in Figure 1D. As predicted, overexpression of let-7a significantly inhibited the luciferase activity mediated by 3′UTR-WT, while mutation of the let-7a binding site abolished this inhibitory effect ( Figure 1E), suggesting that TGFBR3 was indeed a target gene of let-7a. Targetscan identified that mouse TGFBR3 3′UTR region had a potential binding site for let-7a ( Figure 1F). Compared to control mimics group, let-7a could restrain Tgfbr3 WT 3′UTR fused luciferase activity, but not Tgfbr3 mut 3′UTR ( Figure 1G). This result indicates that mouse Tgfbr3 is a target of let-7a.
3.2 | Regulation of cell migration by let-7a and

TGFBR3 in HUVECs
Expression of TGFBR3 was examined in HUVEC culture using a simi- Vinculin was used as an internal control. Negative control expression was set as 100% (*P < 0.05, **P < 0.01). D, Wound healing assay was performed 48 h after transfection, and the distance between the wound edges was observed and photographed. E, The distance between the wound edges was evaluated using Image J from three independent experiments and is expressed as the percentage of the initial wound distance, n = 3 (**P < 0.01, *P < 0.05) TGFBR3 decreased HUVEC migration ( Figure 3C and D), phenocopying defective migration by let-7a overexpression in these cells ( Figure 2C and D). These results indicate that TGFBR3 might be a key mediator of let-7a-mediated migration in HUVECs.

| Regulation of angiogenesis by let-7a and TGFBR3 in HUVECs
Previous reports have found that let-7a plays a role in regulating angiogenesis. 17,29 We carried out a tube formation assay to determine if HUVEC angiogenic ability is affected by let-7a overexpression. We found that let-7a mimics the significantly lowered angiogenic activity of HUVECs ( Figure 4A and B). siRNA-mediated knockdown of the let-7a target gene TGFBR3 in HUVECs resulted in a similar phenotype to that of overexpression of let-7a ( Figure 4C and D). Furthermore, we found that overexpression of TGFBR3 could rescue the angiogenic activity mediated by let-7a mimics (Figure 4E and F). Together with data from the wound-healing assay, we confirmed that let-7a and TGFBR3 regulated angiogenesis in a linear pathway. Identifying the let-7a/TGFBR3 axis may explain the antiangiogenic function of let-7a.

| Let-7a/TGFBR3 signalling is required for hemangioendothelioma growth
To further validate the let-7a/TGFBR3 axis in vivo, we utilized the MS1 cell line. Overexpression of let-7a suppressed TGFBR protein levels by~70% and siRNA-mediated knockdown of TGFBR3 gave 40% reduction of TGFBR3 protein levels ( Figure 5A-D). Overexpression of let-7a mimics resulted in smaller hemangioendothelioma, which is similar to knockdown of Tgfbr3 ( Figure 5E and F).
These results suggest that let-7a targets Tgfbr3 in vitro and in vivo.  were calculated using Image J and Vinculin was used as an internal control. Negative control expression was set as 100% (**P < 0.01). C, Wound healing assay was performed 48 h after transfection, and the distance between the wound edges was observed and photographed. D, The distance between the wound edges was evaluated using Image J from three independent experiments and expressed as the percentage of the initial wound distance, n = 4 (***P < 0.001)

| DISCUSSION
Dysregulated angiogenesis is considered to be one of the fundamental mechanisms regulating initiation and development of many diseases, such as cancer, 20,32,33 heart disease 34,35 and inflammation. 36,37 In this study we found that TGFBR3 is a novel target of let-7a. Let-7a suppresses tube formation and migration by directly targeting TGFBR3 in HUVECs, resulting in defective TGFβ signalling in vascular ECs.
Therefore, we conclude that let-7a likely suppresses angiogenesis via down-regulation of TGFBR3 expression, and the relative TGFβ signalling pathway may be implicated in its mechanism.
Involvement of miRNAs in angiogenesis was first studied in 2006. In past decade, researchers found early regulation of miRNAs might underpin the rescue of diabetes-impaired angiogenesis. 38 Poliseno et al. used miRNA microarray as a discovery tool and identified 27 highly expressed miRNAs in HUVECs; 15 of which were F I G U R E 4 Modulation of angiogenesis by let-7a and TGFBR3 in HUVECs. A, HUVECs transfected with let-7a were subjected to tube-like network formation in matrigel. After 6 h, tube-like network formation was observed and photographed. B, The tube number was evaluated using Image J and let-7a mimics group was normalized by the Con mimics (negative control) group, n = 3 (**P < 0.01). C, HUVECs transfected with siRNA were subjected to tube-like network formation in matrigel. After 6 h, tube-like network formation was observed and photographed. D, The tube number was evaluated using Image J and TGFBR3 siRNA group was normalized to the Con siRNA group, n = 3 (****P < 0.0001). E, HUVECs transfected with let-7a mimics and co-injection with let-7a and TGFBR3 clone were subjected to tube-like network formation in matrigel. After 6 h, tube-like network formation was observed and photographed. F, The tube number was evaluated using Image J, let-7a mimics group and let-7a mimics+TGFBR3 clone was normalized to the Con mimics group, n = 3 (**P < 0.01, ***P < 0.001, ****P < 0.0001) predicted to regulate expression of cell-surface receptors of angiogenic factors. 17 In pathological condition, the combination tamoxifen and/or letrozole with exercise training up-regulated let-7a expression, leading to reduced angiogenesis and tumour growth (mi-croRNA-206, let-7a and microRNA-21) pathways involved in the anti-angiogenesis effects of the interval exercise training and hormone therapy in breast cancer. 19 Let-7 also repressed pathological ocular angiogenesis by targeting HMGA2. Overexpression of let-7 by adenovirus suppressed endothelial cell migration and networking, blockade of let-7a with anti-miR promote migration. 39 Let-7a, one of the highly expressed miRNAs in HUVECs, has been identified to be an important inhibits role in multiple types. Let-7a was shown to control Etv2 (Ets-variant protein 2) expression and EC differentiation through a post-transcriptional mechanism in zebrafish. 40 Our in vivo and in vitro data were consistent with the previous study's conclusions that let-7a imposed anti-angiogenesis activity in a mouse model of breast cancer. 19,41 In particular, multiple evidence suggest let-7a suppress cancer cell migration, invasion and proliferation. For instance, overexpression of let-7a suppressed PTC cell migration and tumour growth by targeting AKT2. 42 Exogenous IGF-2 F I G U R E 5 Let-7a/TGFBR3 is required for hemangioendothelioma growth. Evaluation of TGFBR3 protein level by Western blot after transfection of let-7a mimics (A) or TGFBR3 siRNA (B) in MS1 cell line. C and D, Quantification of TGFBR3 protein levels from three independent experiments by Image J and Vinculin was used as an internal control. Negative control expression was set as 100% (***P < 0.001, *P < 0.05). E, Representative images showing xenograft hemangioendothelioma in response to the knockdown of TGFBR3 or overexpression of let-7a (F) the tumour weights in response to the knockdown of TGFBR3 or overexpression of let-7a, n = 6 (**P < 0.01, *P < 0.05) F I G U R E 6 Screening of downstream genes of let-7a/TGFBR3 axis. A, qRT-PCR analysis of relative genes showed that mRNA levels were significantly reduced in HUVECs transfected with let-7a mimics (*P < 0.05, ***P < 0.001). B, qRT-PCR analysis of the VEGFC and MMP9 mRNA level was significantly reduced in HUVECs transfected with TGFBR3 siRNA, but changes of the ID1 and PROX1 mRNA levels were of no significance (*P < 0.05, ns means no significance). C, Western blot analysis of relative genes showed that VEGFC and MMP9 protein level were significantly reduced in HUVECs transfected with let-7a mimics. D, Western blot analysis of the VEGFC and MMP9 protein level was significantly reduced in HUVECs transfected with TGFBR3 siRNA. E and F, Quantification of TGFBR3, VEGFC, and MMP9 protein levels from three independent experiments by Image J and Vinculin was used as an internal control. Negative control expression was set as 100% (***P < 0.001, *P < 0.05) expression, a potent stimulus facilitated cancer progression, could reversed NEAT1-knockdown-induced growth inhibition, might be a direct target of let-7a. 43 Thus, overexpression of let-7a suppressed cell migration and angiogenesis, which also supported our study that shown in Figure 2.
To explore the molecular mechanism of let-7a in HUVECs, we used TargetScan and miRanda Web Servers, and reviewed considerable literature to screen its target genes in relation to angiogenesis.
Several candidate genes were chosen for additional validation, which not only yielded high miRNA:mRNA alignment scores, but also belong to the TGFβ signalling pathway. After qRT-PCR and Western blotting analyses, TGFBR3 appeared to be a target gene of let-7a owing to the suppression of its mRNA and protein levels by let-7a overexpression.
Furthermore, we identified the let-7a-binding site on the 3′UTR of TGFBR3 and Tgfbr3 mRNA and confirmed that it is indeed a direct target gene of let-7a using reporter gene assay systems ( Figure 1).  46 Depletion of Tgfbr3 has led to mice embryonic lethal due to failed coronary vasculogenesis. 9 Additional phenotypes associated with decreased proliferation and invasion of Tgfbr3-knockout (KO) epicardial cells in vitro have been observed. 47,48 In zebrafish, knockdown of Tgfbr3 by injection of antisense morpholinos lead to impaired angiogenesis in morphant embryos. 49 Similarly, we knocked down the TGFBR3 gene in HUVECs via transfecting siRNAs, resulting in decreased angiogenesis and migration (Figures 3 and 4). Particularly, we observed that let-7a/TGFBR3 was required for hemangioendothelioma growth in the animal study, which could be an approach for in vivo strong evidence ( Figure 5). Other studies utilizing gene Based on the existing data, we speculated that let-7a down-regulates these genes by targeting TGFBR3 mRNA via the TGFβ signalling pathway. Through qRT-PCR analysis, VEGFC and MMP9 were expectedly down-regulated after transfection with let-7a mimics or TGFBR3-siRNA. However, reduction of ID1 and PROX1 mRNA levels were not significant. Furthermore, The Western blot showed the decreased expression of VEGFC and MMP9 (Figure 6), revealing the importance of MMP9 and VEGFC in angiogenesis. MMP9 and VEGF were downregulated by S100A4 silencing, resulting in suppression of cell proliferation, angiogenesis and invasion in thyroid cancer cells. 58 Interestingly, preclinical data have reported the presence of a positive feedback regulation between VEGF and MMP9. MMP inhibitors were capable of inhibiting VEGF-induced angiogenesis in vivo, indicating that MMP9 acts downstream of VEGF. Additionally, MMP9s in vivo angiogenic response was inhibited with a neutralizing VEGF antibody, suggesting that MMP9 also acts upstream of VEGF. 59 In conclusion, our present study highlights that TGFBR3 is a novel and direct target gene of let-7a and let-7a suppresses endothelial angiogenesis and migration through the TGFβs signalling pathway. The TGFBR3/MMP9/VEGFC signalling pathway may be a potential therapeutic target for anti-or pro-angiogenesis strategies in the survival and growth of solid cancers and other diseases. However, because we studied the mechanism of let-7a on angiogenesis in vitro, further investigation is required in the future to determine whether it has a similar effect in vivo.

ACKNOWLEDG EMENTS
This work was supported by grants from the National Natural

CONFLI CT OF INTEREST
The authors indicate no potential conflict of interest.