Investigation of the Regulation of Roundabout4 by Hypoxia-Inducible Factor-1α in Microvascular Endothelial Cells

Abstract

Purpose.: We determined if hypoxia-inducible factor-1α (HIF-1α) and Roundabout4 (Robo4) colocalized in fibrovascular membranes (FVM) from patients with proliferative diabetic retinopathy (PDR), and investigated the regulation of HIF-1α on Robo4 in microvascular endothelial cells under normoxic and hypoxic conditions in vitro.

Methods.: Immunofluorescence and confocal laser scanning microscopy were done to analyze the colocalization of HIF-1α and Robo4 in the FVM. Expression of HIF-1α was knocked down by small interfering RNA (siRNA) technology to study its effects on Robo4 expression of human retinal endothelial cells (HREC) and human dermal microvascular endothelial cells (HDMEC) under normoxic and/or hypoxic conditions. Full-length human HIF-1α gene was transfected into HREC and HDMEC using GFP lentivirus vectors to overexpress HIF-1α under normoxic conditions. The HIF-1α and Robo4 mRNA and protein expressions were quantified by real-time PCR and Western blot. A cell proliferation, migration assay, and flow cytometry were used to analyze the effect of HIF-1α regulation on Robo4 in HREC under hypoxic conditions.

Results.: Colocalization of HIF-1α and Robo4 in vessels of FVM was confirmed by immunofluorescence staining. Knockdown of HIF-1α expression by siRNA in the HREC and HDMEC inhibited Robo4 expression in mRNA and protein level, while overexpressed HIF-1α increased Robo4 mRNA and protein expression. Silencing HIF-1α in endothelial cells under hypoxic conditions inhibited cell invasion and proliferation, which showed that HIF-1α and Robo4 overexpression due to hypoxic conditions correlated with HREC migration and proliferation.

Conclusions.: Both HIF-1α and Robo4 may have a vital role during the formation of FVM. The increased or decreased expression of Robo4 by stimulation or knockdown of HIF-1α suggesting that Robo4 is positively regulated by HIF-1α under normoxic and hypoxic conditions in microvascular endothelial cells in vitro. The HIF-1α gene promotes HREC invasion and proliferation by transcriptionally upregulating Robo4 under hypoxic conditions.

Diabetic retinopathy is a severe microvascular complication associated with diabetes mellitus, and a hallmark of this complication is the formation of a fibrovascular membrane (FVM) that can greatly threaten patients' visual function. An increasing number of reports demonstrate that chronic retinal hypoxia and ischemia have important roles in FVM development. Due to its essential role in systemic responses to hypoxia, hypoxia-inducible factor 1α (HIF-1α), an oxygen sensitive transcription factor, has been associated with angiogenesis and FVM development.^1–3 Particularly, an increased expression of HIF-1α is found in the vitreous humor and in the fibrovascular tissues of eyes of proliferative diabetic retinopathy (PDR) patients. Experimentally and clinically, HIF-1α has a mediating or contributing role in PDR, and, therefore, it can serve as a potential target for therapeutic intervention of retinal neovascularization.^4,5 

In the conditions of low oxygen, hundreds of proteins related to angiogenesis, cell proliferation, cell survival and metabolism were activated through HIF-1 pathway.⁶ To stimulate the expression of hypoxia induced-protein, for example VEGF, erythropoietin, and angiopoietins, HIF-1α has to translocate to the nucleus, dimerizes with HIF-1β, and binds to hypoxia response elements within the promoters of several genes.^7–10 

Roundabout4 (Robo4), the fourth member of the Robo gene family, receives the most attention among its gene family, because it is expressed specifically in the vasculature and is upregulated at sites of angiogenesis.^11–13 It is reported that either knockdown or overexpression of Robo4 on zebrafish impairs vessel formation, suggesting Robo4 has a key role in embryonic angiogenesis.¹⁴ Moreover, in Robo4-knockout mice vascular leakage was found more severe than normal mice, indicating Robo4 is essential in stabilizing the blood vessels.¹⁵ Furthermore, the levels of Robo4 mRNA and the presence, distribution, and role in retinal cells of Robo4 were studied in the FVM from PDR patients. These studies suggest that Robo4 may have a role in the formation of FVM and have physiological functions in the cells of the retina.¹⁶ More recently, Keijiro et al.¹⁷ demonstrated that Robo4 gene downregulation is associated with retinal hyperoxia in an oxygen-induced retinopathy model, which may correspond to the marked regression of the superficial network of vessels that in the central retina and delayed development of the deep plexus. 

The Robo4 gene was found overexpressed only in endothelial cells when exposed to hypoxia in vitro,¹⁸ consistent with the theory that Robo4 is a gene regulated by hypoxia. However, the regulatory mechanism has not been fully explained to date. Based on the positive expression of HIF-1α and Robo4 in FVM in vivo,^3,4,16 we hypothesized that HIF-1α, a transcriptional regulatory factor, may have a regulatory role on Robo4 expression. To confirm this hypothesis, colocalization of HIF-1α and Robo4 in the FVM were analyzed by immunofluorescence and confocal laser scanning microscopy, and Robo4 expression was quantified during HIF-1α downregulation or overexpression in human retinal endothelial cells (HREC) and human dermal microvascular endothelial cells (HDMEC) under normoxic and/or hypoxic conditions in vitro. Our results reveal for the first time to our knowledge that HIF-1α and Robo4 colocalized in the vessels of FVM and that Robo4 was positively regulated by HIF-1α under normoxic and hypoxic conditions in microvascular endothelial cells in vitro. 

Materials and Methods

Tissue Samples

The study protocol was approved by the Ethics Committee of Jilin University, and informed consent was obtained from all patients according to the World Medical Association Declaration of Helsinki. A total of 12 type II diabetes mellitus patients with PDR got involved in this research, four of whom were males and eight were females. All patients were aged 47 to 72 years. They accepted pars plana vitrectomy with membrane peeling. The FVM specimens surgically obtained were fixed in 4% paraformaldehyde (PFA), paraffin embedded, and sections cut at 4 μm. 

Confocal Immunofluorescence

Sections were dewaxed and rehydrated through an alcohol to water gradient, rinsed in 0.01 M PBS for 5 minutes and blocked with 10% normal goat serum (Sigma-Aldrich Corp., St. Louis, MO, USA) for 30 minutes at 37°C. Then, the tissue sections were applied by 1:50 anti-mouse HIF-1α monoclonal antibody (Cat No.ab1; Abcam, Cambridge, UK) with 1:50 anti-rabbit Robo4 polyclonal antibody (Cat No.ab10547; Abcam), 1:50 anti-rabbit HIF-1α polyclonal antibody (Boster; Boster Co., Wuhan, China) with 1:50 anti-mouse CD34 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), 1:50 anti-rabbit Robo4 polyclonal antibody with 1:50 anti-mouse CD34, respectively, at 4°C overnight. After washing with PBS, the sections were incubated for 1 hour at 37°C with 1:100 FITC and Cy3-conjugated goat anti-mouse and goat anti-rabbit secondary antibodies (Invitrogen Corp., Camarillo, CA, USA), respectively. After incubation, the slides were washed with PBS and cell nuclei were stained with Hoechst33342 (Sigma-Aldrich Corp.). Negative controls, including omission of the primary antibody and use of an irrelevant polyclonal or isotype-matched monoclonal primary antibody, were adopted in each immunostaining procedure. In all cases, negative controls showed only weak staining after incubation at 4°C overnight. Each antibody specimen was immunolabeled twice at least, and appropriate Ig controls were included for each experiment. Slides were mounted in 20% glycerol/PBS, coverslipped, and sealed with nail varnish. The observation was carried by an Olympus FV-1000 laser confocal microscope (Olympus, Center Valley, PA, USA) and image software (FV10-ASW1.7). 

Cell Culture

The HREC and HDMEC cells were kindly provided by Prof Liu Xiaoqing of Tongji University and Prof Pei of the Second Hospital of Jilin University, respectively. The HREC cells were cultured in Endothelial Cell Medium (Invitrogen Corp.) in an incubator maintained at 37°C in an atmosphere containing 5% CO₂ and air, supplemented with 5% fetal bovine serum (FBS; Gibco, Invitrogen Corp.), 100 units/mL penicillin, 100 μg/mL streptomycin (Sigma-Aldrich Corp.), 1% Endothelial Cell Growth Supplement (Beijing Maichen, Bio Co., China). The HDMEC cells were cultured in Dulbecco's modified eagle medium (DMEM; Invitrogen Corp.) with 10% FBS, 100 units/mL penicillin, 100 μg/mL streptomycin at 37°C in a humidified atmosphere containing 5% CO₂ and air. 

Hypoxic conditions were gained by culturing cells in a sealed, anaerobic workstation (Concept 400; Ruskin Technologies, Pencoed, Wales, UK), in which the hypoxic environment (1% O₂, 94% N2, and 5% CO₂), temperature (37°C), and humidity (90%) were kept constant. 

Small Interfering RNA (siRNA) and Transfection Assays

Using a previously described method,¹⁹ the HIF-1α–specific siRNAs were synthesized chemically. The sequence of targeted HIF-1α was CTGGACACAGTGTGTTTGA. Human nonsilencing siRNA acted as a negative control (NC) and was used to control for any effects of the siRNA and transfection reagent. The control sequence was AGUCUCCACGUGUACGUTT. Cells were transfected with siRNAs using Lipofectamine 2000 reagent (Invitrogen Corp.) according to the manufacturer's instructions. The HIF-1α siRNA was used to transfect the cells at different concentrations (such as 10, 20, 50, and 100 nM) for 24, 48, and 72 hours. It was determined that transfection with 100 nM HIF-1α siRNA for 48 hours was the most effective and was selected for the next experiments (data not shown). Under normoxic conditions, HREC and HDMEC were randomly divided into control hypoxia group (N), vector group (NC), and HIF-1α siRNA group (HIF-1α si). The HREC cells were randomly divided into a normoxia group and hypoxia group. The hypoxia group was randomly divided into control hypoxia group (N), vector group (NC), and HIF-1α siRNA group (HIF-1α si). No transfection was done in the normoxia group and control hypoxia group. After 6 hours of incubation with the liposome-DNA complex, the cells were washed and cultured for 24 hours and then cultured under normoxia (21% O₂) or hypoxia (1% O₂) for an additional 24 or 48 hours. 

Recombinant HIF-1α Lentiviral Vector and Production

Lentiviral plasmid pLenti-V5-D-TOPO-HIF-1α encoding human HIF-1α full length sequence was donated by Prof Niles (Marshall University, Huntington, WV, USA).²⁰ To construct recombinant lentiviral vector encoding HIF-1α with green fluorescence reporter, pLVX-hHIF1α-IRES-ZSGreen was designed and constructed. Briefly, human HIF-1α full-length sequence (2.5 kb) was amplified using the template pLenti-V5-D-TOPO-HIF-1α and the following primers: pLVX-hHIF1α-XbaI-F: 5′-CTAGTCTAGACACCATGGAGGGCGCCGGCG GCGCGA-3′, pLVX-hHIF1α-BamHI-R: 5′-CGCGGATCCTCAGTTAACTTGATCCAAAGCTCT G-3′. The pLVX-hHIF1α-IRES-ZSGreen was produced by double digestion at the restriction enzyme cutting sites XbaI and BamHI. The identity of the gene was confirmed by sequencing (Sangon, Shanghai, China). Lentivirus was produced from cotransfection with four plasmids pLVX-hHIF1α-IRES-ZSGreen or pLVX-IRES-ZsGreen, PLP1, PLP2, and PLP-VSVG (pLVX-IRES-ZsGreen and the additional three plasmids were provided by Prof Liu Xiao-qing of Tongji University). The PLVX-IRES-ZsGreen served as negative control. Recombinant lentivirus was harvested 72 hours following cotransfection of the pLVX-hHIF1α-IRES-ZSGreen (10 μg) or pLVX-IRES-ZsGreen (10 μg), PLP1 (5 μg), PLP2 (5 μg), and PLP-VSVG (5 μg) into 293T cells cultured in DMEM with 10% FBS. Transfections were performed using a high efficiency calcium phosphate transfection kit (Beijing Maichen) with the manufacturer's recommendations. Then, the virus supernatant was purified, and the viral titer was detected. 

In Vitro Transduction and Analysis of GFP Expression

The HREC and HDMEC cells were seeded into 96-well plates at 2 × 10³ cells/well, respectively, and incubated in 0.5 mL growth medium for 24 hours before infection. Viral particles were added to the wells at a multiplicity of infection (MOI) of 2, 5, 10, 20, 30, and 50. To improve lentiviral vector transduction, 10 μg/mL hexadimethrine bromide (Polybrene; Sigma-Aldrich Corp.) was added simultaneously per well. After 24 hours incubation at 37°C in 5% CO₂, the virus-containing medium was removed and replaced with 0.5 mL fresh culture medium per well. An inverted fluorescence microscope (IX71; Olympus) was applied to observe the transduction efficiency every day. Transduction efficiency was quantified by measuring the percentage of GFP-expressing cells in total visible cells. The following studies were done at the MOI of 20 of the lentivirus. Then, HREC and HDMEC were seeded into 12-well plates at 1 × 10⁴ cells/well, respectively, and incubated in 1 mL growth medium for 24 hours before infection. Then, we added viral particles (4 μL) into each well. The infection was repeated three times every 48 hours. The onset and time course of GFP fluorescence and morphology of GFP-positive cells were observed at days 3, 5, 7, 10, and 14 after infection with the same settings. The HREC and HDMEC cells were harvested at 7 and 14 days, respectively, post infection for quantitative real-time PCR or for Western blot analysis. Three parallel wells were performed in identical procedure and repeated at least 3 times for each viral vector infection, then the data from each infection were summarized and averaged. 

Real-Time PCR

The RNAiso kit (TakaRa; TakaRa Biotechnolgical Co., Dalian, China) was used to extract total RNA from cells according to the manufacturer's protocol. All the RNA preparations were measured to have an OD260:OD280 ratio of 1.9:2.0. Total RNA (2 μg) was reversibly transcribed by aid of the Revertaid first-strand cDNA Synthesis Kit (lot K1622; Fermentas China, ThermoFisher Scientific, Waltham, MA, USA). No template control was included in each set of samples. The real-time PCR assays were performed in Light Cycler 480II (Hoffmann-La Roche, Basel, Switzerland) according to the manufacturer's instructions. Both HIF-1α and Robo4 were normalized to GAPDH expression and calculated by the Light Cycler 480II software. All primers used were synthesized by Shanghai Sangon Biological Engineering Technology & Services Corporation (Shanghai, China). See the Table for sequences. 

Table

 

Gene Subtype Oligonucleotide Primers

 

Western Blot Analysis

Protein extracts were electrophoresed on 8% SDS polyacrylamide gels, transferred onto a 0.22 mm polyvinylidene difluoride membranes (Invitrogen) and then probed with specific antibodies. Primary antibodies were used at the following dilutions: anti-Robo4 (1:1000), anti-HIF-1α (1:1000; Abcam) and GAPDH (1:1000; Hangzhou Goodhere Bio Co., China). Immunoreactive bands were visualized with EasySee Western blot kit (Beijing TransGen Biotech Co., China) according to the manufacturer's instructions. Band densities of HIF-1α and Robo4 proteins were normalized to GAPDH internal control. Western blots were repeated three times and qualitatively similar results were obtained. 

HREC Migration Assay

A migration study was performed using a Transwell system (Corning Life Sciences, Costar, Tewksbury, MA, USA). Briefly, HREC were grown for 48 hours after transfection with 100 nM HIF-1α siRNA in ECM containing 5% FBS with 1% Endothelial Cell Growth Supplement. Media were collected and loaded onto the lower chamber. Then, 1 × 10⁵ HREC in 100 uL of ECM (containing 5% FBS with 1% Endothelial Cell Growth Supplement) were loaded into the top chamber. Then, the cells were cultured under normoxia (21% O₂) or hypoxia (1% O₂) for an additional 18 hours at 37°C. Migrating cells were fixed and stained with crystal violet and counted in five high power fields (×20). 

Bromodeoxyuridine (BrdU) Incorporation Assay

Cell proliferation was analyzed by a 5-bromo-2′-deoxyuridine (BrdU) incorporation assay. The HREC cells were grown 24 hours after transfection with 100 nM HIF-1α siRNA in complete medium under hypoxic conditions and then pulsed with 10 μM BrdU (Sigma-Aldrich Corp.) for 24 hours. Harvested cells by trypsinization were washed with PBS and fixed in 4% cold ethanol for 30 minutes. After washing with PBS, the fixed cells were permeabilized by incubation in 2 M HCl for 10 minutes at room temperature and then washed with PBS three times. The cells were incubated with PBS containing 5% BSA for 30 minutes at room temperature. Next, the cells were washed with PBS and incubated with a monoclonal anti-BrdU antibody (1:200 dilution, Sigma-Aldrich Corp.) or PBS as a negative control in dark for 1 hour at room temperature. Then, the cells were washed twice with PBS and resuspended and mixed with FITC-conjugated secondary antibody (1:500 dilution; ThermoFisher Scientific). Fluorescence signals of FITC-BrdU were measured by flow cytometry using a Flow Cytometer (Beckman Coulter EPICS XL-MCL; Beckman Bioscience, Brea, CA, USA), and the data were analyzed with EXPO32 ADC Analysis software (Beckman Bioscience). 

Statistical Analysis

All experiments were performed at least three times. Results were presented as the mean ± SD of three independent experiments. Student's t-test was used to do statistical evaluation. A value of P < 0.05 was considered to indicate statistical significance. 

Results

Colocalization of HIF-1α and Robo4 in FVM Sections

To determine whether HIF-1α colocalizes with Robo4 in the FVM, human FVM sections were double stained with mouse anti-HIF-1α and rabbit anti-Robo4 antibodies. In all FVM specimens, intense staining of HIF-1α and Robo4 was detected and their expression was colocalized (Fig. 1A). To further investigate the distribution of HIF-1α and Robo4 in FVM specifically, the sections were double stained with an endothelial cell marker, CD34 antibody, and either an anti-Robo4 (Fig. 1B) or anti-HIF-1α (Fig. 1C) antibody. Consistent with previous results on the staining patterns of tumors and FVM,^12,16 Robo4 was expressed in the vessels of the FVM. Furthermore, HIF-1α also coexpressed with CD34. 

Figure 1

 

Colocalization of HIF-lα and Robo4 in FVM from a PDR patient ([A–C], original magnification: ×40). (A) Staining of HIF-lα and Robo4 by mouse anti-HIF-1α antibody (green) and rabbit anti-Robo4 antibody (red). Double immunofluorescence staining shows colocalized expression of HIF-lα and Robo4 in the FVM (Robo4/HIF-lα, yellow). (B) Staining of Robo4 and CD34 by rabbit anti-Robo4 (red) and CD34 (green) antibody, respectively. Based on the double immunofluorescence staining, Robo4 expresses in the vessels of the FVM (Robo4/ CD34, yellow). (C) Staining of HIF-lα (red) and CD34 (green) by rabbit anti-HIF-lα and CD34 antibody, respectively. The HIF-lα expresses in the vessels of the FVM (HIF-lα/CD34, yellow). The colocalization of HIF-1α and Robo4 on blood vessel was indicated by white arrow. (D) Quantitative analysis of colocalization of HIF-lα and Robo4 in FVM. Pearson's correlation coefficient (PCC) and overlap coefficient according to Manders (MOC) indicated a high degree of colocalization of HIF-lα and Robo4 proteins, and a moderate degree of colocalization of HIF-lα or Robo4 and CD34. Image-Pro Plus Software was used to calculate colocalization coefficients. An average of MOC and PCC of three examined samples for each time point is shown. P < 0.05. Error bars indicate standard deviation.

Quantitative colocalization analysis was performed using Image-Pro Plus Software. Pearson's correlation coefficient (PCC) and Mander's overlap coefficient (MOC) were examined (Fig. 1D). Colocalization of HIF-1α and Robo4 was supported by the results of coefficients calculations: PCC was 0.74, while MOC was 0.757. It is important to note that the two coefficients, while revealing different aspects of the colocalization process, showed a similar pattern of change among the study groups, proving the applicability of the calculations to investigate the degree of colocalization of HIF-1α and Robo4. Colocalization of HIF-1α or Robo4 and CD34 was indicated by the results of coefficients calculations: PCC was 0.52 or 0.54, while MOC was 0.661 or 0.634, respectively. The fluorescence of CD34 was stronger than HIF-1α and Robo4 in FVM, so MOC, which represents the true degree of colocalization, was more meaningful. 

Effect of HIF-1α Silencing on Robo4 mRNA and Protein Expression In Vitro

To further determine the relationship between HIF-1α and Robo4, HIF-1α RNA in both HREC and HDMEC was knocked down by HIF-1α siRNA under normoxic conditions, and Robo4 mRNA and protein expressions were evaluated by real-time PCR and Western blot assays, respectively (Fig. 2). 

Figure 2

 

Inhibition of HIF-1α with siRNA decreases Robo4 expression in HREC and HDMEC under normoxic conditions. Mean and standard deviation of three independent real-time PCR experiments are presented for HIF-1α and Robo4 mRNA (A, B). The HIF-1α siRNA (100 nM) abolished HIF-1α mRNA expression and downregulated Robo4 mRNA expression in HREC (A) and HDMEC (B). Representative Western blot analysis and the mean and standard deviation of densitometric analysis from three independent Western blots in HREC (C) and HDMEC (D) are shown. Robo4 protein levels also decreased after HIF-1α siRNA transfection in both cell types after 48 hours (C, D). Control cells were transfected with scrambled siRNA. *Denote values significantly different between HIF-1α siRNA-treated group and control groups (P < 0.01).

At the mRNA level, real-time PCR demonstrated that HIF-1α siRNA depleted HIF-1α mRNA levels by 67 ± 8% in HREC (P < 0.01; Fig. 2A) and 48 ± 4% in HDMEC (P < 0.01; Fig. 2B). On the contrary, there was no significant difference between the cells transfected with control siRNA (negative control, NC) and nontransfected cells (normal group, N; P > 0.05). Accordingly, Robo4 mRNA and protein expressions in HREC and HDMEC were inhibited. The levels of Robo4 mRNA were downregulated by 33 ± 3% in HREC and 61 ± 4% in HDMEC in the HIF-1α siRNA group when compared to the normal group (P < 0.01; Figs. 2A, 2B), which is consistent with the decreased protein expression in both cell types (Figs. 2C, 2D). 

Based on the results during normoxia conditions, we transfected HREC with HIF-1α siRNA to knock down HIF-1α and cultured cells under hypoxic conditions for 24 and 48 hours. Then, Robo4 mRNA and protein expressions by real-time PCR and Western blot assays were analyzed (Figs. 3, 4). 

Figure 3

 

Inhibition of HIF-1α with siRNA decreases Robo4 mRNA expression in HREC during hypoxic conditions. Mean and standard deviation of three independent real-time PCR experiments are presented for HIF-1α and Robo4 mRNA. The HIF-1α siRNA (100 nM) abolished HIF-1α mRNA expression (A) and down regulated Robo4 mRNA expression (B) in HREC in culture under hypoxic conditions for 24 and 48 hours. Control cells were transfected with scrambled siRNA. *Denote values significantly different between HIF-1α siRNA-treated group and control groups (P < 0.01).

Figure 4

 

Inhibition of HIF-1α with siRNA decreases Robo4 protein expression in HREC during hypoxic conditions. Representative Western blot analysis and the mean and standard deviation of densitometric analysis from three independent Western blots in HREC during hypoxic conditions are shown (A). The HIF-1α protein expression increased under hypoxic conditions in culture for 24 hours, and more significantly for 48 hours, when compared to normal cells during normoxic conditions. However, HIF-1α siRNA (100 nM) abolished HIF-1α protein expression (B) and downregulated Robo4 protein expression (C) in HREC under culture in hypoxic conditions for 24 and 48 hours. Control cells were transfected with scrambled siRNA. *Denote values significantly different between HIF-1α siRNA-treated group and control groups (P < 0.01).

At the mRNA level, real-time PCR demonstrated that HIF-1α mRNA in HREC under hypoxic culture for 24 and 48 hours increased by 1.52 ± 0.12-fold and 2.11 ± 0.21-fold respectively, when compared to the normal group under normoxic conditions (P < 0.01, Fig. 3A). The HIF-1α siRNA depleted HIF-1α mRNA levels by 57 ± 6% and 81 ± 9% in HREC under hypoxic culture for 24 and 48 hours, respectively (P < 0.01, Fig. 3A). On the contrary, there was no significant difference between the cells transfected with control siRNA (negative control, NC) and nontransfected cells (normal group, N) under hypoxic conditions (P > 0.05). At the protein level, HIF-1α expression was increased under hypoxic culture for 24 and 48 hours when compared to the normal group under normoxic conditions, and it decreased in the HIF-1α siRNA group (P < 0.01; Figs. 4A, 4B). Accordingly, Robo4 mRNA and protein expressions were inhibited in the HIF-1α siRNA group under hypoxic conditions. Robo4 mRNA was upregulated by 1.76 ± 0.14-fold and 2.78 ± 0.17-fold in HREC under hypoxic culture for 24 and 48 hours, respectively, when compared to the normal group under normoxic conditions (P < 0.01; Fig. 3B). Robo4 mRNA expressions were decreased by 61 ± 7% and 79 ± 3% in the HIF-1α siRNA group when compared to the normal group under hypoxic culture for 24 and 48 hours (P < 0.01, Fig. 3B), which is consistent with the decreased protein expression (Figs. 4A, 4C). 

Efficient Transduction by Lentiviral Vector

Lentiviral vector-mediated HIF-1α transduction efficiency in HREC and HDMEC by GFP detection was first visualized by fluorescence microscopy three days after transduction. Fluorescence peak in HREC and HDMEC were detected at 7 and 14 days, respectively (data not shown). 

Overexpression of HIF-1α Upregulates Robo4 Expression in HREC and HDMEC Cells

Based on real-time PCR results, relative levels of HIF-1α and Robo4 mRNA in the study group were markedly increased in HREC (1.76 ± 0.08-fold and 1.61 ± 0.11-fold) and HDMEC (3.03 ± 0.36-fold and 3.32 ± 0.48-fold) when compared to the levels in the GFP-negative control groups (cells transfected with GFP vectors alone, pLVX-ZsGreen) and levels in the nontransfected control cells (mock; P < 0.01; Figs. 5A, 5B). The HIF-1α-GFP lentiviral vector transfection also was confirmed by examination of HIF-1α protein expression via Western blot. Both HIF-1α and Robo4 protein expression levels in the HREC and HDMEC were found significantly higher in the study groups than in the control groups (Figs. 5C, 5D). 

Figure 5

 

Overexpression of HIF-1α by lentiviral vector transduction upregulates Robo4 expression under normoxic conditions. Real-time PCR analysis (A, B) and Western blot analysis (C, D) of HIF-1α overexpression and the changes in Robo4 levels in HREC and HDMEC after transfection by lentiviral vectors. At 7 (HREC) and 14 (HDMEC) days after transfection, mRNA and protein were extracted from cells of the study groups (transfected with HIF-1α–GFP lentiviral vector pLVX-hHIF1α), cells of the negative GFP control groups (transfected with GFP lentiviral vector pLVX -ZSGreen), and cells of the negative control groups without transfection (mock). Real-time PCR was performed to measure the relative HIF-1α and Robo4 mRNA levels, and readings were normalized to human GAPDH mRNA levels. Each value represented the mean ± SD of three replicates. The HIF-1α mRNA levels were significantly higher in the study group than the negative GFP control and mock groups (*P < 0.01, [A, B]). Accordingly, Robo4 mRNA levels were significantly higher in the study group than negative GFP control and mock groups (*P < 0.01, [A, B]). Protein expression levels of HIF-1α and Robo4 were assessed by Western blot analysis in HREC (C) and HDMEC (D). The HIF-1α and the Robo4 protein were significantly higher in the study groups (pLVX-hHIF1α) than in both control groups (C, D).

Hypoxia/HIF-1α Induces HREC Proliferation and Invasion

To determine the importance of HIF-1α regulation on Robo4 in HREC under hypoxic conditions, we examined the effects of HIF-1α siRNA treatments on HREC invasion and cell proliferation by an invasion assay using Transwell Matrigel-coated chambers and a BrdU assay, respectively. 

The cell migration assay demonstrated that HIF-1α RNA intervention (HIF-1α siRNA group, 193 ± 9.77) markedly reduced the number of cells that migrated through the chamber compared to the control siRNA-treated cells (NC, 444 ± 36.36, P < 0.01) and nontransfected cells (N, 517 ± 12.16; P < 0.01) under hypoxic conditions, while those two groups' migration ability was significantly higher than the normoxic groups (131 ± 11.99; P < 0.01; Figs. 6A, 6B). In contrast, there was no significant difference between the N and NC groups (P > 0.05, Figs. 6A, 6B). 

Figure 6

 

Inhibition of HIF-1α with siRNA decreases migration and proliferation ability in HREC during hypoxic conditions. (A) After transfection with 100 nM HIF-1α siRNA for 48 hours, 10⁵ cells were seeded onto Matrigel coated Transwell inserts and allowed to migrate through the filter in normoxia or hypoxia conditions for 18 hours. Migrated cells were fixed, stained, and captured at ×200 magnification using the camera on an inverted microscope. (B) The HIF-1α siRNA significantly decreased the migration ability of HREC. The HIF-1α siRNA group showed a markedly decreased migration ability compared to the control siRNA-treated cells (NC, negative control; *P < 0.01) and nontransfected cells (N, normal; *P < 0.01) under hypoxic conditions, which migrated significantly more than the normoxic groups (#P < 0.01). (C) The BrdU staining assays revealed that HIF-1α siRNA group cell growth was significantly lower than in the control siRNA-treated cells and nontransfected cells under hypoxic conditions (*P < 0.01). Data are presented as the mean ± SD. Each group experiment was repeated 3 times (n = 3).

By using a BrdU assay, we found that the growth of cells transfected with HIF-1α siRNA (HIF-1α siRNA group, 13.11 ± 1.59%) was markedly inhibited compared to the untreated group (N, 27.64 ± 0.79%) and the control siRNA-treated group (NC, 24.84 ± 1.09%, P < 0.01, Fig. 6C). There was no significant difference between the N and NC groups (P > 0.05). 

Discussion

Retinal neovascularization (RNV), a major cause of blindness in humans, is an abnormal proliferation and migration of new blood vessels from pre-existing vessels in the retina. This kind of neovascularization is deficient in tight junctions and, hence, causes plasma to leak into surrounding tissue including the vitreous resulting in vitreous hemorrhage. It attributes to the formation of FVM in PDR. The condition of RNV is stimulated by one or more angiogenic factors released by the retina under hypoxic or ischemic conditions.²¹ The HIF-1 signal pathway regulates the adaptive responses to O₂ tensions at cellular levels, and it controls the expression of many genes involved in angiogenesis, cell survival,²² tumor growth,²³ and genetic instability.²⁴ 

The identification of hypoxia-inducible factor-1 (HIF-1) as a key transcription factor that mediates increased expression of hypoxia-regulated genes, such as VEGF, significantly aided the understanding of RNV pathogenesis. The HIF-1 has two subunits: HIF-1α expression is induced in hypoxic tissue, while HIF-1β is constitutively expressed. Expression of HIF-1α is detected in FVM of diabetic retinopathy (DR) patients, indicating a potential target for therapeutic intervention of RNV.^2,3,5 

Recent studies have suggested that Robo4, an endothelial-specific member of the roundabout family, is required to maintain blood vessel integrity by counteracting VEGF and acts as a negative regulator of angiogenesis in model system.¹⁵ It also has been reported that Robo4-UNC5B signaling maintains vascular integrity by counteracting VEGF signaling in endothelial cells.²⁵ However, there is no study on how Robo4 is regulated. 

In our study of PDR membranes, we demonstrated for the first time to our knowledge that HIF-1α and Robo4 were positively expressed and colocalized. Further, HIF-1α and/or Robo4 and CD34, an endothelial cell marker, were observed to colocalize, consistent with previous studies^16,21 where HIF-1α and Robo4 were closely related with angiogenesis. Robo4 was specially expressed in the vascular system, particularly in the vascular endothelial cells.^11–13 Several studies have suggested that HIF-1α and Robo4 may have a role in the formation of FVM and participate in physiological functions within the cells of the retina.^2–4,16 Therefore, we studied the role of HIF-1α and Robo4 in PDR and notably found that HIF-1α and Robo4 colocalized in the FVM of PDR patients (Figs. 1A, 1D). Therefore, we hypothesized that HIF-1α, as a transcriptional regulatory factor, may have a part in Robo4 synthesis during the formation of FVM. 

To further explore the regulation of HIF-1α on Robo4, HIF-1α expression was controlled by siRNA and stable transfection in two kinds of microvascular endothelial cells (HREC and HDMEC) under normoxic conditions. The HREC is a type of primary cell line, and HDMEC is an immortalized cell line. To investigate the relationships between Robo4 and HIF-1α accurately, we tested both cells. We demonstrated that HIF-1α siRNA specifically knocks down HIF-1α RNA in HREC and HDMEC, and Robo4 mRNA and protein levels decreased accordingly under normoxic conditions (Fig. 2). 

In this study, we exposed HREC to hypoxic conditions in vitro to mimic the hypoxia experienced by endothelial cells in ischemic retinal diseases in vivo. We found that Robo4 mRNA and protein expressions were inhibited in the HIF-1α siRNA group under hypoxic conditions (Figs. 3B, 4A, 4C). This is consistent with a previous study that indicated Robo4 expression is increased in endothelial cells under hypoxic conditions.¹⁸ In summary, we found HIF-1α positively controls the expression of Robo4 in hypoxic conditions just as it does in normal conditions. 

According to the research that proves that gene transfer of a constitutively active form of HIF-1 induced RNV in the absence of retinal ischemia,²⁶ we observed that at peak HIF-1α lentiviral transfection in HREC and HDMEC under normoxic conditions, HIF-1α mRNA and protein levels along with Robo4 mRNA and protein levels were markedly overexpressed (Fig. 5). Therefore, overexpressed HIF-1α is responsible for the upregulation of Robo4 expression. 

These results suggest that Robo4 is positively regulated by HIF-1α under normoxic and hypoxic conditions in vitro. In vivo relationships between HIF-1α and Robo4 will be conducted in further investigations. We predicted the possible transcription factors of Robo4 by AliBaba 2.1 software, but HIF-1α was not found, suggesting that there is no direct binding between HIF-1α and Robo4 proteins. The HIF-1α effects on Robo4 expression may act through another pathway. Okada et al.^27–29 firstly reported the cloning and characterization of the human Robo4 promoter, and confirmed that SP1 bond to the human Robo4 promoter and induced the promoter activity. Several findings suggest that hypoxia activates some genes and proteins, such as Redd1³⁰ and phenylethanolamine N-methyltransferase (PNMT),^31,32 indirectly via HIF-1α stimulation of Sp1. So, the regulation of HIF-1α on Robo4 expression may be mediated by activation of Sp1, but detail mechanism requires further exploration. 

To study the importance of HIF-1α regulation on Robo4 in endothelial cells under hypoxic conditions, we examined the effects of HIF-1α siRNA transfection on HREC invasion and proliferation and found that silencing HIF-1α inhibited cell invasion and proliferation. Our results show that HIF-1α and Robo4 overexpression due to hypoxic conditions correlates with HREC migration and proliferation (Figs. 3, 4, 6). As illustrated in Figures 3 and 4, the increased expression of Robo4 induced by hypoxia was attenuated in HIF-1α knockdown cells. These results demonstrated that hypoxia/HIF-1α promotes HREC invasion and proliferation by transcriptionally upregulating Robo4. This is consistent with a previous study that found Robo4-specific siRNA was an effective and specific inhibitor of migration and proliferation of choroid–retina endothelial (RF/6A) and human RPE cells.¹⁶ Furthermore, these results indicated that the RNA intervention method used in our study to knock down hypoxia-induced HIF-1α expression also is efficient at inhibiting HREC migration and proliferation. Targeted therapy toward HIF-1α and Robo4 may have possible therapeutic implications in proliferative diabetic retinopathy. 

In conclusion, our study indicated that HIF-1α and Robo4 may cooperate in the formation of FVM. Silencing HIF-1α expression in HREC and HDMEC cells inhibited Robo4 expression, while lentiviral vector-mediated overexpression of HIF-1α in HREC and HDMEC cells was associated with increased Robo4 expression, suggesting Robo4 is positively regulated by HIF-1α under normoxic and hypoxic conditions in vitro. The HIF-1α promotes HREC invasion and proliferation by transcriptionally upregulating Robo4 under hypoxic conditions. The regulatory mechanism and function change remains unclear and requires further exploration. 

Acknowledgments

The authors thank Richard M. Niles for generously providing the lentiviral plasmid pLenti-V5-D-TOPO-HIF-1α used in this study. 

Supported by Key Projects of Science and Technology Development Plan of Jilin province (Grant No. 20090453) and Research Fund for Key Specialty-Ophthalmology Construction Program of Jilin Department of Health. 

Disclosure: R. Tian, None; Z. Liu, None; H. Zhang, None; X. Fang, None; C. Wang, None; S. Qi, None; Y. Cheng, None; G. Su, None