MMP-2 Alters VEGF Expression via αVβ3 Integrin-Mediated PI3K/AKT Signaling in A549 Lung Cancer Cells

Abstract

Vascular endothelial growth factor (VEGF) is one of the most important angiogenic growth factors for tumor angiogenesis. Here, we sought to explore whether RNA interference (RNAi) targeting Matrix metalloproteinase-2 (MMP-2) could disrupt VEGF mediated angiogenesis in lung cancer. MMP-2 siRNA inhibited lung cancer cell-induced tube formation of endothelial cells in vitro; addition of recombinant human-MMP-2 restored angiogenesis. MMP-2 transcriptional suppression decreased VEGF, PI3K protein levels, and AKT phosphorylation in lung cancer cells. In addition, MMP-2 suppression decreased Hypoxia inducible factor-1α (HIF-1α), a transcription factor for VEGF, as determined by Electrophoretic mobility shift assay (EMSA). We also show that MMP-2 suppression disrupted phosphatidylinositol 3-kinase (PI3K) dependent VEGF expression; ectopic expression of myr-AKT restored VEGF inhibition. Further, MMP-2 suppression decreased the interaction of integrin-αVβ3 and MMP-2 as confirmed by immunoprecipitation analyses. Studies with either function blocking integrin-αVβ3 antibody or MMP-2 specific inhibitor (ARP-100) indicate that suppression of MMP-2 decreased integrin-αVβ3-mediated induction of PI3K/AKT leading to decreased VEGF expression. Moreover, A549 xenograft tissue sections from mice that treated with MMP-2 siRNA showed reduced expression of VEGF, and the angiogenic marker, Factor-VIII. The inhibition of tumor angiogenesis in MMP-2 suppressed tumor sections was associated with decreased co-localization of integrin-αVβ3 and MMP-2. In summary, these data provide new insights into the mechanisms underlying MMP-2-mediated VEGF expression in lung tumor angiogenesis.

INTRODUCTION

Tumor growth and metastasis require angiogenesis when the tumor reaches 1-2mm in diameter.¹ Angiogenesis, a physiological process involving the growth of new blood vessels from pre-existing vessels, occurs due to a series of events, which begins with the disruption of the endothelial cell basement membrane and extracellular matrix (ECM) by proteolytic enzymes.² Degradation of ECM releases basement membrane-sequestered angiogenic factors, such as Vascular endothelial growth factor (VEGF), basic Fibroblast Growth Factor (bFGF), and Transforming Growth Factor-β (TGF-β).³ VEGF and it's receptors (VEGFRs) are key modulators of angiogenesis.⁴ Recent studies show that the functions of VEGF may not be limited to endothelial cells and that VEGF may play an important role in survival, proliferation, and migration of tumor cells.^5,6 It has also been demonstrated that VEGF produced by tumor cells is essential for tumor growth^2,7 and is thought to increase proliferation of endothelial cells from neighboring blood vessels through interactions with VEGFR-2. However, the lung cancer cells themselves increased expression levels of VEGF and VEGFR-2 on tumor epithelial and stromal cells as compared to their normal tissue counterparts. These studies suggest that tumor-produced VEGF has additional biological functions, perhaps promoting the proliferation and survival of tumor cells.^8-10 Anti-angiogenic therapies based on inhibition of VEGF/VEGFR signaling were reported to be powerful clinical strategies in oncology.¹¹

Phosphoinositide 3-kinase (PI3K) has been shown to play a role in cell proliferation and anti-apoptosis, both of which are crucial to cancer development.¹² AKT is one of the most important downstream targets of PI3K. AKT transmits oncogenic signals and mediates a variety of cellular responses including cell growth, transformation, differentiation, motility, and cell survival.¹³ Recent data clearly suggest that hypoxia-inducible factor-1 (HIF-1) is a downstream gene in the PI3K/AKT pathway.^14-16 VEGF expression is regulated at transcriptional level by HIF-1 in response to hypoxia and growth factor stimulation. HIF-1 is a heterodimer of the HIF-1α and HIF-1β subunits. HIF-1α stabilization can be induced by hypoxia, growth factors, and oncogenes.¹⁷ Further, activation of PI3K has been shown to increase levels of HIF-1α, and these levels correlated with tumorigenicity and angiogenesis in nude mice.^18,19

Matrix metalloproteinases (MMPs) comprise a highly regulated family of structurally-related enzymes capable of degrading most, if not all, of the components of the extracellular matrix.^18,19 Several MMPs are believed to be important in the process of angiogenesis, especially MMP-2 and MMP-9. These two MMPs preferentially degrade basement membrane components such as type-IV collagen, unmask cryptic biologically relevant sites in ECM components, modulate angiogenic factors, and are involved in the production of endogenous angiogenic inhibitors. Several investigators have described a possible role of MMP-2 in tumor angiogenesis using mouse models^20-22 and synthetic inhibitors (e.g., BHPA, BMS-275291).^23,24 MMP-2 and MMP-9 have been associated with increased tumor spread and poor prognosis in lung cancer.^25,26 In recent years, MMP inhibitors have been the focus of anti-cancer research and have been used in clinical trials to block angiogenesis in tumors as well as tumor metastasis.^18,19 However, the exact mechanisms of how MMP-2 contributes to angiogenesis still remain obscure and may involve multiple pathways. Clinical trials of the broad-spectrum MMP inhibitor marimastat and its analogue batimastat have been disappointing, mainly because of unexpected side effects such as musculoskeletal pain and inflammation. Overall, the results from clinical trials of broad spectrum MMP inhibitors point to the need to determine the role of specific MMPs in specific stages of disease progression and the need to design more selective inhibitors that are devoid of adverse reactions induced by broad-spectrum inhibitors.

Here, we have demonstrated the role of MMP-2 in lung cancer cell-induced angiogenesis using siRNA-based approach in in vitro and in vivo models. Our findings indicate that MMP-2 transcriptional inactivation significantly reduced integrin-αVβ3-mediated, PI3K/AKT-induced VEGF expression, which ultimately decreased tumor cell-induced angiogenesis.

MATERIALS AND METHODS

Cells and Reagents

A549 and H1299 lung adenocarcinoma cells were cultured in RPMI 1640 (ATCC Manassas, VA) supplemented with 10% fetal bovine serum (Invitrogen, Carlsbad, CA), 50 units/mL penicillin, and 50 μg/mL streptomycin (Life Technologies, Inc., Frederick, MD). For human microvascular dermal endothelial cells (HMEC-1), glutamine, EGF and hydrocortisone (Stem Cell Technologies, British Columbia, Canada) were added to Advanced MEM medium (Invitrogen, Carlsbad, CA). Cells were incubated at 37 °C in a humidified 5 % CO₂ atmosphere.

We used antibodies specific for MMP-2, VEGF (VEGF-A), VEGFR-2, HIF-1α, GAPDH, integrin-αVβ3 (Clone 23C6), MT1MMP, (Santa Cruz Biotechnology, Santa Cruz, CA), PI3K, AKT, phospho-AKT (Ser-473) (Cell Signal Technology, Boston, MA), functional blocking integrin-αVβ3 (Clone 23C6; Cat# CBL544; Millipore Corporation, Temecula, CA) and anti-Human Von-Willebrand Factor (Factor-VIII; Dako North America, Inc, CA), and HRP/Alexa Fluor^® conjugated secondary antibodies (Santa Cruz Biotechnology, Santa Cruz, CA). We also used constitutively active-AKT (myr-AKT) plasmid (Addgene, Plasmid 10841), human recombinant-MMP-2, human recombinant-VEGF₁₆₅ (Millipore Corporation, Temecula, CA), human VEGF quantikine ELISA Kit, human MMP-2 quantikine ELISA Kit (R&D Systems, Minneapolis, MN) and ARP-100, a MMP-2 specific inhibitor (TOCRIS Bioscience, Ellisville, MO) in this study.

Adenoviral siRNA constructs and infection

Adenoviral siRNA constructs for MMP-2 (Ad-MMP-2-Si) and scrambled vector (Ad-SV) were constructed and amplified as described by us previously.²⁷ Viral titers were quantified as pfu/mL following infection of 293 cells. Titers were obtained for Ad-SV (7.6 × 10¹¹ pfu/mL) and for Ad-MMP-2-Si (5.0 × 10¹¹ pfu/mL). The amount of infective adenoviral vector per cell (pfu/cell) in culture media was expressed as multiplicity of infection (MOI). Virus constructs were diluted in serum-free culture media to the desired concentration, added to cells, and incubated at 37 °C for 1 h. The necessary amount of complete medium was then added and cells were incubated for the desired time periods.

Gelatin zymography

Tumor conditioned medium was prepared as follows: after 36 h of mock, 100 MOI of either Ad-SV or Ad-MMP-2-Si infection, medium was removed from A549 cells, washed with PBS, 3 mL of serum free medium was added, and cells were incubated for another 12 h. MMP-2 secretion into conditioned medium was determined by gelatin zymography as described previously.²⁷ Briefly, conditioned medium containing equal amounts of protein was resolved over gelatin-SDS-polyacrylamide gels. For immunoprecipitated samples, immunocomplexes were incubated with 1X sample buffer for 30 min at RT and were resolved over gelatin-SDS-polyacrylamide gels. Gels were washed in 2.5% Triton X-100 to remove SDS and followed by overnight incubation at 37 °C in Tris-CaCl₂ buffer (pH 7.6). Gels were stained with Coomassie brilliant blue and subsequently de-stained for 1h. Gelatinolytic activities were identified as clear zones of lyses against a dark blue background.

Immunoprecipitation and western blotting

A549 cells were infected with 100 MOI of either Ad-SV or Ad-MMP-2-Si for 48 h. Whole cell lysates were prepared by lysing cells in radioimmunoprecipitation assay (RIPA) lysis buffer with proteinase inhibitors. For the hypoxic experiments, cells were infected for 36 h followed by incubation at 1 % O₂ for 12 h in serum-free medium. Equal amounts of protein fractions or immunoprecipitates of lysates with indicated antibodies were resolved over SDS-PAGE and transferred to PVDF membrane. Proteins were detected with primary antibodies followed by HRP-conjugated secondary antibodies. Comparable loading of proteins on the gel was verified by re-probing the blots with an antibody specific for the housekeeping gene, GAPDH.

Reverse transcription PCR

Total RNA was extracted from cells using TRIzol reagent as per manufacturer's instructions. The ImProm-II Reverse Transcription (Promega Corporation, Madison, WI) with 2 μg of RNA and poly-dT primers was used for synthesis of cDNA. To determine the mRNA transcript level form cDNA, PCR was carried out using the following primers: MMP-2, forward 5′-GTGCTGAAGGACACACTAAAGAAGA-3′ and reverse 5′-CCTACAACTTTGAGAAGGATGGCAA-3′; VEGF-A, forward 5′-CTGTTCTCGCTTCGGAGGAG-3′ and reverse 5′-AAACCATGAACTTTCTGCTGTCTTGGG-3′; VEGFR-2, forward 5′-CTGGCATGGTCTTCTGTGAAGCA-3′ and reverse 5′-TGGCTTCCACCAGAGATTCCAT-3′; HIF-1α, forward 5′-AGTCTGCAACATGGAAGG-3′ and reverse 5′-CACGACTTGATTTTCTCCC-3′; and GAPDH, forward 5′-GGAGTCAACGGATTTGGTCGTAT-3′ and reverse 5′-GTCTTCACCACCATGGAGAAGGCT-3′. PCR products were resolved on 1-2 % agarose gels, visualized under UV light.

ELISA

Conditioned medium was prepared as mentioned above and VEGF-A concentrations in conditioned medium were determined using a commercial Human VEGF quantikine ELISA Kit as per manufacturer's instructions. For determining the MMP-2 concentrations in cell lysates, commercial Human MMP-2 quantikine ELISA Kit was used. Absorbance at 450 nm was measured and corrected using the 540 nm reading on a Benchmark microplate reader (Bio Rad, Model 680, Hercules, CA). Data analysis was done using Microplate Manager III software (Bio-Rad Laboratories).

In vitro angiogenic assay

Angiogenesis was performed as described earlier.²⁸ Briefly, conditioned medium collected from mock, Ad-SV-infected or Ad-MMP-2-Si-infected cells was added to HMEC-1 (2×10⁴ cells/well) in 96-well plates coated with Matrigel. HMEC-1 were allowed to grow overnight and the formation of capillary-like structures was captured using a microscope attached to a CCD camera. The degree of angiogenesis was measured as the total tube length per field (average of 8-10 fields) and mock treatment condition normalized as 100; the product indicates the degree of angiogenesis.

Electrophoretic mobility shift assay (EMSA)

Nuclear extracts were prepared from Ad-MMP-2-Si-infected A549 cells after 48 h as described previously.²⁸ The protein was quantified and 5μg nuclear extracts were incubated in 20 μL of buffer (20 mM HEPES pH 7.9, 0.4 mM EDTA pH 8.4, 0.4 mM DTT, 10 % glycerol, 100 μg/mL poly-dI/dC, 1 % NP-40) with ³²P-labeled, double-stranded oligonucleotides containing the HIF-1α binding site with 75,000 cpm for 15 min at RT. For gel supershift, HIF-1α antibodies were used. The reaction mixture was electrophoresed through a 5 % polyacrylamide gel, and the gel was dried and subjected to autoradiography using phosphor screens at −70 °C.

Xenograft animal models

Severe combined immunodeficiency (SCID) mice were maintained and experiments were performed as described previously.²⁷ Briefly, A549 cells (5×10⁶/0.1 mL) were grafted into the flanks of 6-week-old female SCID mice. When the mean tumor diameter reached 3-4 mm (~12–14 days), the mice were separated into three groups (5/group). Each group was injected intravenously with PBS or 5×10⁸ pfu of either Ad-SV or Ad-MMP-2-Si virus thrice with 3-day intervals. After treatment, subcutaneous tumor growth was measured every 3^rd day and tumor volume was calculated as 0.5 (R_max × R_min²). Tumor growth was followed for up to 5 weeks. Mice were euthanized when the tumor diameter in control mice reached 1.1-1.2 cm. Subcutaneous tumors were either frozen immediately at −80 °C or fixed in 10 % phosphate buffered formaldehyde.

Immunocytochemical and immunohistochemical analyses

Immunocytochemical and immunohistochemical analyses were performed as described previously.²⁹ Briefly, A549 cells were cultured in chamber slides and infected with mock or 100 MOI of either Ad-SV or Ad-MMP-2-Si for 48 h. Cells were washed with PBS and fixed in 4 % Para-formaldehyde and permeabilized in 0.1 % Triton-X100. Non-specific binding was blocked by BSA in PBS, followed by incubation with anti-integrin-αVβ3 and anti-MMP-2 antibodies. For immunohistochemical analysis, tissue sections (4-5 μm) were de-paraffinized in xylene, rehydrated in graded ethanol solutions, permeabilized in 0.1 % Triton X-100 and 0.1 % sodium citrate, and incubated overnight with primary antibodies for MMP-2, Factor-VIII, VEGF or integrin-αVβ3. Mouse IgG was used as a negative control. Expression was detected with either Alexa Fluor^®-conjugated antibody or HRP-conjugated secondary antibody followed by 3,3-diaminobenzidine solution. For nucleus counterstaining, DAPI or hematoxylin was used. Slides were photographed with microscope attached with CCD camera.

Statistical analysis

Results were analyzed using a two-tailed Student's t-test to assess statistical significance. Statistical differences are presented at probability levels of p<0.05, p<0.01 and p<0.001.

RESULTS

Ad-MMP-2-Si inhibits MMP-2 expression in A549 cells

We have previously shown that transcriptional suppression of MMP-2 using adenoviral-mediated siRNA against MMP-2 (Ad-MMP-2-Si) reduced invasion and angiogenesis in A549 cells and inhibited tumor growth and lung metastasis in a mouse model.²⁷ In this study, we sought to determine the mechanisms underlying Ad-MMP-2-Si-mediated inhibition of the angiogenic factor VEGF (VEGF-A). We previously confirmed the specificity of MMP-2 transcriptional suppression with Ad-MMP-2-Si, which inhibits MMP-2 but no other MMPs, such as MMP-1, MMP-7 and MMP-9.²⁹ Initially, to test the efficacy of the siRNA sequence, we infected A549 cells with 100 MOI of Ad-MMP-2-Si or scrambled vector (Ad-SV). The MMP-2 protein secreted into the conditioned medium was examined using gelatin zymography. MMP-2 gelatinolytic activity was higher in cells infected with PBS (mock) and Ad-SV. However, the gelatinolytic activity of MMP-2 was reduced by 88.98 ± 2.29 % (p<0.01) in Ad-MMP-2-Si-infected cells as compared to mock and Ad-SV-infected controls (Fig. 1A). To support these results, we performed western blot analysis using an anti-MMP-2 antibody. Ad-MMP-2-Si infection decreased the expression of MMP-2 by 87.87 ± 3.59 % (p<0.01) when compared to mock and Ad-SV-infected cells as determined by densitometric analysis of western blot bands (Fig. 1B). To determine whether decreased levels of MMP-2 were caused by gene transcription, we examined the transcript levels of MMP-2 using semi-quantitative RT-PCR. As shown in figure 1B, MMP-2 mRNA levels were decreased by 87.09 ± 2.50 %; (p<0.01) in Ad-MMP-2-Si-infected cells when compared to the controls. Further, we also show that Ad-MMP-2-Si inhibits MMP-2 protein and mRNA levels in H1299, another adenocarcinoma cell line (Supplementary Fig. 1).

Figure 1. Ad-MMP-2-Si infection inhibits MMP-2 expression in A549 cells.

A549 cells were infected as described in Materials and Methods. Briefly, A549 cells were infected with 100 MOI of either Ad-SV or Ad-MMP-2-Si, the medium was aspirated after 36 h of incubation, 3 mL of serum-free medium was added, and cells were incubated overnight. (A) Gelatin zymographic analysis for secreted MMP-2 activity in to the tumor conditioned medium. The gelatinolytic band intensities of MMP-2 were quantified by densitometric analysis using ImageJ software (National Institutes of Health) and normalized with the intensity of the gelatinolytic band in mock-conditioned medium. Columns: mean of triplicate experiments; bars: SE (Standard Error); *p<0.01, significant difference from Ad-SV infected control. (B) Top: Western blot analysis of MMP-2 expression in cell lysates, which were prepared 48 h after Ad-MMP-2-Si infection. The experiments were repeated three times and a representative blot is shown. The blot was stripped and re-probed with GAPDH antibody to detect total amount of the respective proteins. Protein band intensities were quantified by densitometric analysis using ImageJ software (National Institutes of Health). The levels of MMP-2 protein were normalized to protein level in mock infected cells. Columns: mean of triplicate experiments; bars: SE; *p<0.01, significant difference from Ad-SV control. Bottom: Semi-quantitative RT-PCR analysis showing reduced MMP-2 mRNA transcription in Ad-MMP-2-Si-infected cells. Total RNA was extracted, as mentioned in Materials and Methods, from cells infected with mock or 100 MOI of either Ad-SV or Ad-MMP-2-Si, and cDNA was synthesized as described in Materials and Methods. The PCR reaction was set up using the first-stand cDNA as the template for MMP-2. PCR products were resolved on agarose gels and quantified by densitometry using ImageJ software (National Institutes of Health). The level of MMP-2 mRNA transcripts was normalized to the mock infected control product. GAPDH served as a control. Columns: mean of triplicate experiments; bars: SE; *p<0.01, significant difference from Ad-SV control.

Ad-MMP-2-Si infection inhibits tumor cell-induced microtubule network formation in endothelial cells

Since MMP-2 is know to play a role in angiogenesis,^20-22 we examined the effect of Ad-MMP-2-Si on A549 cell-induced tube formation by HMEC-1. Extensive microtubule network formation was visualized in endothelial cells cultured in the presence of tumor conditioned medium (TCM) from A549 cells infected with mock and Ad-SV. In contrast, the microtubule network formation was significantly decreased by 89.05 ± 2.51 % (p<0.01) in endothelial cells cultured in the presence of TCM from Ad-MMP-2-Si-infected A549 cells (Fig. 2A).

Figure 2. Ad-MMP-2-Si inhibits tumor cell conditioned medium-induced angiogenesis in HMEC-1 cells and inhibits VEGF and VEGFR-2 expression in A549 cells.

(A) In vitro angiogenesis: A549 cells were infected with mock, Ad-SV or Ad-MMP-2-Si for 36 h, medium was removed, minimum amounts of serum-free medium were added to cover the cells, and cells were incubated for 12 h. Tumor conditioned medium (TCM) was collected and added into 96-well plates, which were coated with Matrigel and seeded with human dermal microvascular endothelial cells-1 (HMEC-1; 2×10⁴ cells/well). After overnight incubation at 37°C, cells were observed under the bright field microscope for formation of capillary-like structures. The degree of angiogenic induction by mock-TCM, Ad-SV-TCM and Ad-MMP-2-Si-TCM was quantified for the numerical value of the product of the relative capillary length per microscopic field. Columns: mean of triplicate experiments; bars: SE; *p<0.01, significant difference from Ad-SV-TCM. (B) Tumor conditioned medium was prepared as described above and ELISA was performed for detection of VEGF-A levels according to the manufacturer's instructions. Quantification of VEGF concentration is shown. Columns: mean of triplicate experiments; bars: SE; *p<0.01, significant difference from Ad-SV control. (C) Western blot analysis of tumor conditioned medium (CM) for VEGF (VEGF-A) protein levels was carried out using a VEGF-specific antibody. For preparation of cell lysates (CL) infected cells were incubated for 48 h as described in Materials and Methods. Western blot analysis of cell lysates for VEGF and VEGFR-2 protein levels was carried out using a VEGF and VEGFR-2 specific antibody. The blots were stripped and re-probed with GAPDH antibody to detect total amounts of the respective proteins. Protein band intensities were quantified by densitometric analysis using ImageJ software (National Institutes of Health). The levels of VEGF and VEGFR-2 proteins were normalized to respective protein level in mock infected A549 cells. Columns: mean of triplicate experiments; bars: SE; *p<0.01 and *p<0.05, significant difference from Ad-SV control. (D) Semi-quantitative RT-PCR analysis showing reduced VEGF and VEGFR-2 mRNA expression in Ad-MMP-2-Si-infected cells. Total RNA was extracted, as described in Materials and Methods, from A549 cells infected with mock or 100 MOI of either Ad-SV or Ad-MMP-2-Si, and cDNA was synthesized as described in Materials and Methods. The PCR reaction was set up using the first-stand cDNA as the template for VEGF and VEGFR-2. PCR products were resolved on agarose gels and quantified by densitometry using ImageJ software (National Institutes of Health). The levels of VEGF and VEGFR-2 mRNA transcripts were normalized to the mock infected control product. GAPDH served as a control. Columns: mean of triplicate experiments; bars: SE; *p<0.01 and **p<0.05, significant difference from Ad-SV control.

Ad-MMP-2-Si inhibits VEGF-mediated angiogenesis in vitro

Since lung cancer cells often express very high levels of VEGF (VEGF-A), a key mediator of blood vessel growth,^18,19 we determined the effect of Ad-MMP-2-Si infection on VEGF expression using a Human VEGF Quantikine ELISA Kit. MMP-2 inhibition decreased VEGF shedding into TCM as compared to mock and Ad-SV-infected cells. The extent of VEGF secretion into the TCM in Ad-MMP-2-Si-infected A549 cells was 2.01 ± 0.15 pg/μg (p<0.01) as compared to 7.22 ± 0.84 pg/μg in Ad-SV-infected cells (Fig. 2B). Western blot analysis indicated that VEGF and VEGFR-2 protein expression was decreased in Ad-MMP-2-Si-infected cells as compared to mock and Ad-SV-infected cells (Fig. 2C). Ad-MMP-2-Si infection inhibited VEGF and VEGFR-2 expression by >85 % (p<0.01) and 65.00 ± 4.45 % (p<0.05), respectively. To test whether Ad-MMP-2-Si infection inhibited VEGF and VEGFR-2 mRNA transcription, we performed semi-quantitative RT-PCR using total RNA. As shown in figure 2D, Ad-MMP-2-Si infection inhibited VEGF and VEGFR-2 mRNA transcription by 89.65 ± 2.01 % (p<0.01) and 68.65 ± 2.44% (p<0.05), respectively. Moreover, addition of recombinant human-VEGF (rhVEGF; 25 nM) restored Ad-MMP-2-Si-mediated inhibition of angiogenesis in vitro (Fig. 4C). To further confirm the role of MMP-2 in VEGF expression, we infected A549 cells with Ad-MMP-2-Si for 24 h and incubated cells with recombinant human-MMP-2 (rhMMP-2; 25 ng/mL). As shown in Figure 5, rhMMP-2 restored Ad-MMP-2-Si-mediated inhibition of VEGF expression from 15.55 ± 6.21 % (p<0.01) to 67.96 ± 4.01 % (p<0.01). Similarly, addition of rhMMP-2 to Ad-MMP-2-Si-infected cells reversed Ad-MMP-2-Si-mediated inhibition of capillary network formation by HMECs (Fig. 4C). These results suggest that MMP-2 downregulation inhibits expression of VEGF at the mRNA transcription level in A549 cells.

Figure 4. Ad-MMP-2-Si inhibits PI3K/AKT-mediated VEGF expression in A549 cells.

(A) A549 cells were infected with mock or 100 MOI of either Ad-SV or Ad-MMP-2-Si for 48 h and cell lysates western blotted for PI3K, AKT and phosphorylated-AKT (Ser-473) using specific antibodies. The experiments were carried out three times and representative western blots are shown. The blots were stripped and re-probed with GAPDH antibody to detect total amounts of the respective proteins. Protein band intensities were quantified by densitometric analysis using ImageJ software (National Institutes of Health). The levels of PI3K and phospho-AKT (Ser-473) proteins were normalized to respective protein level in mock infected A549 cells. Columns: mean of triplicate experiments; bars: SE; *p<0.01, significant difference from Ad-SV control. (B) A549 cells were transfected with a plasmid expressing constitutively active AKT (myr-AKT) 24 h before Ad-MMP-2-Si infection. We then performed western blot analysis for VEGF levels in tumor conditioned medium (CM) and VEGF, AKT and phosphorylated-AKT (Ser-473) levels in cell lysates (CL) using specific antibodies. Protein band intensities were quantified by densitometric analysis using ImageJ software (National Institutes of Health). The levels of VEGF and phospho-AKT (Ser-473) proteins were normalized to respective protein level in mock infected A549 cells. Columns: mean of triplicate experiments; bars: SE; *p<0.01, significant difference from Ad-SV control; **p<0.01, significant difference from Ad-MMP-2-Si infection alone. (C) A549 cells were transfected with a plasmid expressing constitutively active AKT (myr-AKT) 24 h before Ad-MMP-2-Si infection. After 24 h of Ad-MMP-2-Si infection, recombinant human MMP-2 (25 ng/mL) or recombinant human VEGF (25 nM) was added to A549 cells. Cells were incubated for 12 h, the medium was aspirated, cells were washed with PBS three times, a minimum amount of serum-free medium was added, and cells were incubated for another 12 h. An in vitro angiogenic assay was performed under the same conditions as described in Figure 2. The experiments were carried out three times, and representative pictures are shown. The degree of angiogenic induction was quantified for the numerical value of the product of the relative capillary length per microscopic field. Columns: mean of quadruplicate experiments; bars: SE; **p<0.01, significant difference from respective Ad-SV infected control; *p<0.01, significant difference from respective Ad-MMP-2-Si infection alone.

Figure 5. Recombinant human-MMP-2 inhibits Ad-MMP-2-Si-inhibited PI3K, phosphorylation of AKT and VEGF expression in lung adenocarcinoma cells.

(A & B) After 24 h of Ad-MMP-2-Si infection, A549/H1299 cells were incubated with rhMMP-2 (25 ng/mL) for 12 h and the medium was aspirated, cells were washed with PBS three times, a minimum amount of serum-free medium was added, and cells were incubated for another 12 h. We then performed western blot analysis for VEGF in conditioned medium (CM) and expression of PI3K, phosphorylated-AKT (Ser-473), VEGF and VEGFR-2 in cell lysates (CL). Protein band intensities were quantified by densitometric analysis using ImageJ software (National Institutes of Health). The levels of PI3K, phospho-AKT (Ser-473) and VEGF proteins were normalized to respective protein level in mock infected cells. Columns: mean of triplicate experiments; **p<0.01, significant difference from Ad-SV control; *p<0.05 significant difference from Ad-MMP-2-Si infection.

Ad-MMP-2-Si inhibits HIF-1α expression and activity in A549 cells

A key stimulus for increased VEGF expression in lung cancer is hypoxia. Hypoxia is prevalent in lung tumors^18,19 and leads to stabilization and increased expression of the α subunit of the transcription factor HIF-1. Since the HIF-1α is an upstream molecule and a transcription factor for VEGF, we quantified the expression and activity of HIF-1α. The effect of Ad-MMP-2-Si infection on HIF-1α stabilization under normoxic and hypoxic conditions was determined in cell lysates from Ad-MMP-2-Si-infected A549 cells. We could not detect HIF-1α expression under normoxic conditions; however, HIF-1α was detected in A549 cells exposed to hypoxia (1% O₂) for 12-16h. As Figure 3A shows, Ad-MMP-2-Si infection inhibited HIF-1α expression by 96.13 ± 1.98 % (p<0.01) when compared to the controls. We next determined whether Ad-MMP-2-Si infection inhibited HIF-1α expression at the mRNA transcription level using semi-quantitative RT-PCR with total RNA. As shown in Figure 3B, Ad-MMP-2-Si infection inhibited HIF-1α mRNA transcription by 81.21 ± 3.47 % (p<0.01) as compared to the controls. Since, HIF-1α is a transcription factor for VEGF; we determined HIF-1α binding activity by EMSA using nuclear extracts from Ad-MMP-2-Si-infected A549 cells. As shown Figure 3C, HIF-1α activity substantially decreased in Ad-MMP-2-Si-infected A549 cells as compared to mock and Ad-SV infected cells. Pre-incubation of nuclear extracts with HIF-1α antibody indicated a shift, thereby confirming the specificity of the band.

Figure 3. Ad-MMP-2-Si inhibits HIF-1α expression in A549 cells.

A549 cells were infected with mock, Ad-SV or Ad-MMP-2-Si for 36 h, medium was removed, minimum amounts of serum-free medium were added to cover the cells, and cells were incubated at 1.0% O₂ for 12 h. Finally, we collected the conditioned medium and the cells. (A) Cell lysates were used for western blot analysis for HIF-1α expression using a HIF-1α-specific antibody. The experiments were carried out three times and a representative western blot is shown. The blots were stripped and re-probed with GAPDH antibody to detect total amounts of the respective proteins. Protein band intensities were quantified by densitometric analysis using ImageJ software (National Institutes of Health). The levels of HIF-1α protein were normalized to protein level in mock infected A549 cells. Columns: mean of triplicate experiments; bars: SE; *p<0.01, significant difference from Ad-SV control. (B) Semi-quantitative reverse transcription-PCR analysis showing reduced HIF-1α mRNA expression in Ad-MMP-2-Si-infected cells. Total RNA was extracted, as described in Materials and Methods, from A549 cells infected with mock or 100 MOI of either Ad-SV or Ad-MMP-2-Si, and cDNA was synthesized as described in Materials and Methods. The PCR reaction was set up using the first-stand cDNA as the template for HIF-1α. GAPDH served as a control. PCR products were resolved on agarose gels and quantified by densitometry using ImageJ software (National Institutes of Health). The level of HIF-1α mRNA transcripts was normalized to the mock infected control product. GAPDH served as a control. Columns: mean of triplicate experiments; bars: SE; *p<0.01 significant difference from Ad-SV controls. (C) A549 cells were infected as described above and nuclear extracts prepared as described in Materials and Methods. Nuclear protein (5 μg) was subjected to EMSA for HIF-1α binding to its consensus sequence.

Ad-MMP-2-Si reduces PI3K/AKT-mediated VEGF expression and angiogenesis in vitro

Several studies have confirmed a link between PI3K/AKT pathway activation and increased VEGF and HIF-1α expression.^14,15 We therefore determined the effect of MMP-2 transcriptional suppression on PI3K/AKT. As shown in Figure 4A, Ad-MMP-2-Si infection inhibited the expression of PI3K and phosphorylation of AKT (Ser-473) by 94.7 ± 1.95 % (p<0.01) and 98.9 ± 1.02 % (p<0.01), respectively. To determine the contribution of the PI3K pathway, which involves PI3K and its downstream effector, AKT on VEGF expression, A549 cells were transiently transfected with constitutively active-AKT (myr-AKT) 24 h before Ad-MMP-2-Si infection. As shown in Figure 4B, ectopic expression of constitutively active-AKT restored AKT phosphorylation and VEGF expression in Ad-MMP-2-Si-infected A549 cells. To test the functional consequence of ectopic expression of myr-AKT in Ad-MMP-2-Si-infected cells, we performed tube formation assays with supernatant from cultured A549 cells. As shown in Figure 4C, the ectopic expression of myr-AKT restored Ad-MMP-2-Si-inhibited tube length from 11.09 ± 2.12 % (p<0.01) to 75.11 ± 5.57 % (p<0.01).

To further validate the role of MMP-2 in regulating PI3/AKT and VEGF expression, we infected A549 lung adenocarcinoma cells with Ad-MMP-2-Si for 24 h and incubated cells with rhMMP-2 (25 ng/mL) for another 24 h. The addition of rhMMP-2 to Ad-MMP-2-Si-infected A549 cells restored PI3K expression (from 13.25 ± 2.36 %; p<0.01 to 40.11 ± 2.32 %; p<0.01), phospho-AKT (from 9.45 ± 4.21 % p<0.01 to 60.22 ± 3.33 %; p<0.01) and VEGF (from 15.55 ± 6.21 %; p<0.01 to 67.96 ± 4.01 %; p<0.01) in A549 cell line (Fig. 5A). Similar results were observed in H1299, another adenocarcinoma cell line (Fig. 5B). Likewise, addition of rhMMP-2 restored Ad-MMP-2-Si-mediated inhibition of capillary formation (Fig. 4C). Quantification of tube length indicated that rhMMP-2 addition to Ad-MMP-2-Si-infected cells restored tube length from 11.02 ± 2.12 % (p<0.01) to 57.65 ± 2.55 % (p<0.01) as compared to cells treated with only Ad-MMP-2-Si. These data show that MMP-2 transcriptional suppression decreases the angiogenic activity of VEGF via the PI3K/AKT pathway.

Transcriptional suppression of MMP-2 inhibits VEGF expression via integrin-αVβ3 in A549 cells

Angiogenesis depends on input from various growth factors and vascular cell adhesion signals. Direct interaction between MMP-2 and the integrin-αVβ3 has been demonstrated,^18,19 and coordinated interplay between both these molecules during blood vessel formation has been reported.^21,30,31 These observations prompted us to determine the interaction of MMP-2 and integrin-αVβ3 in MMP-2-regulated VEGF expression in A549 cells. Immunocytochemical analysis revealed that the co-localization of MMP-2 and integrin-αVβ3 decreased in Ad-MMP-2-Si-infected cells as compared to mock and Ad-SV-infected cells (Fig. 6A). Immunoprecipitation studies also confirmed decreased interaction between MMP-2 and integrin-αVβ3 by 87.15 ± 3.46 % (p<0.01) in Ad-MMP-2-Si-infected cells as compared to Ad-SV-infected cells (Fig. 6B). Similarly, inhibition of catalytic activity of MMP-2 using a pharmocologic inhibitor, ARP-100 (TOCRIS Bioscience, Ellisville, MO) also decreased MMP-2 and integrin interactions. Cells were treated with MMP-2 inhibitor (ARP-100) and cell lysates were immunoprecipitated with integrin-αVβ3 and MMP-2 activity and protein levels were determined by western blotting and gelatin zymography. Figure 6C indicates that MMP-2 inhibitor at 50 and 100 μM decreased MMP-2 interaction with integrin-αVβ3 by ~65 and ~90 %. To further characterize the effect of the MMP-2 and integrin-αVβ3 interaction on VEGF expression, we added rhMMP-2 to Ad-MMP-2-Si infected cells and performed co-immunoprecipitation using an anti-integrin-αVβ3 antibody followed by gelatin zymography, and western blot analysis with an anti-MMP-2 antibody. As anticipated, the interaction of MMP-2 and integrin-αVβ3 increased with the exogenous addition of rhMMP-2 in control and Ad-MMP-2-Si-infected cells (Fig. 6D). Addition of rhMMP-2 markedly increased the phosphorylation of AKT and expression of PI3K and VEGF in mock and Ad-MMP-2-Si-infected cells (Fig. 6E) as compared to cells treated with Ad-MMP-2-Si only. Likewise, incubation of A549 cells with functional blocking anti-integrin-αVβ3 antibodies prevented the rhMMP-2-induced increase in p-AKT, PI3K and VEGF (Fig. 6E).

Figure 6. Inhibition of MMP-2 inhibits integrin-αVβ3-mediated VEGF expression in A549 cells.

(A) A549 cells were plated in 8-well chamber slides and infected with mock or 100 MOI of either Ad-SV or Ad-MMP-2-Si for 48 h. Immunocytochemistry was performed for co-localization (yellow) of MMP-2 (green) and integrin-αVβ3 (Red) using specific antibodies. (B) A549 cells were infected with mock or 100 MOI of either Ad-SV or Ad-MMP-2-Si for 48 h, and proteins were immunoprecipitated with an anti-integrin-αVβ3 antibody. Western blot analysis was performed for MMP-2 using immunoprecipitated proteins. Protein band intensities were quantified by densitometric analysis using ImageJ software (National Institutes of Health). The levels of MMP-2 protein were normalized to protein level in mock infected A549 cells. Columns: mean of triplicate experiments; *p<0.01, significant difference from Ad-SV control. (C) After 36 h of Ad-MMP-2-Si infection, A549 cells were incubated with a MMP-2 specific inhibitor (ARP-100; at a concentration of 50 and 100 μM) for 12 h. Proteins were immunoprecipitated with anti-integrin-αVβ3 antibody and performed gelatin zymography and western blot for MMP-2. Columns: mean of triplicate experiments; *p<0.01, significant difference from Ad-SV/solvent control. (D) After 24 h of Ad-MMP-2-Si infection, A549 cells were incubated with rhMMP-2 (25 ng/mL) for 24 h. Gelatin zymography and western blot analysis was performed for MMP-2 from immunoprecipitated protein samples with an anti-integrin-αVβ3 antibody. Protein band intensities were quantified by densitometric analysis using ImageJ software (National Institutes of Health). The levels of MMP-2 protein were normalized to protein level in mock infected A549 cells. Columns: mean of triplicate experiments; *p<0.01, significant difference from Ad-SV control; **p<0.01 significant difference from Ad-MMP-2-Si infection. (E) After 24 h of Ad-MMP-2-Si infection, A549 cells were incubated with rhMMP-2 (25 ng/mL), non-specific IgG and/or anti-integrin-αVβ3 antibodies for another 24 h. We then performed western blot analyses for PI3K, phosphorylation of AKT and VEGF levels using total lysates. GAPDH served as a loading control. Protein band intensities were quantified by densitometric analysis using ImageJ software (National Institutes of Health). The levels of PI3K, phospho-AKT (Ser-473) and VEGF proteins were normalized to respective protein level in mock infected A549 cells. Columns: mean of triplicate experiments; *p<0.01, significant difference from Ad-SV infected control; **p<0.05 significant difference from Ad-SV plus integrin-αVβ3 blocking antibody treatment.

Ad-MMP-2-Si treatment inhibits VEGF expression and reduces blood vessel density in vivo

We next sought to determine the effect of Ad-MMP-2-Si treatment on VEGF expression and angiogenesis in vivo. Our previous studies demonstrate that Ad-MMP-2-Si treatment decreased tumor growth and completely abolished lung nodules in a spontaneous lung metastasis model.²⁷ Evaluation of the tumor tissue sections indicated decreased expression of VEGF, co-localization of MMP-2 and integrin-αVβ3 in tumors from mice treated with Ad-MMP-2-Si as compared to tumor tissue sections from mice that received Ad-SV treatment (Fig. 7). We used anti-Human Von-Willebrand Factor (Factor-VIII) staining to compare the density of micro-vessels in the tumors treated with Ad-MMP-2-Si and in the control tumors. We observed a drastic reduction in the expression of Factor-VIII (an angiogenic endothelial cell marker) in xenograft tumor tissue sections from Ad-MMP-2-Si-treated mice when compared to tumor tissue sections from mock and Ad-SV-treated controls (Fig. 7).

Figure 7. MMP-2 inhibition using Ad-MMP-2-Si decreased angiogenesis, VEGF expression, and co-localization of MMP-2 and integrin-αVβ3 in A549 lung tumors in vivo.

H&E staining was performed as per standard protocol and representative pictures (400 X magnifications) of tumor sections from mock, Ad-SV and Ad-MMP-2-Si-treated mice were shown. Immunohistochemistry was performed for expression and co-localization (yellow) of MMP-2 (green) and integrin-αVβ3 (Red) using specific antibodies. Also showed immunohistochemical analysis for anti-Human Von-Willebrand Factor (Factor-VIII) and VEGF expression using specific antibodies. Data shown are representative fields (400 X magnifications). Also shown is the negative control where the primary antibody was replaced by non-immune serum (inserts).

DISCUSSION

Angiogenesis is essential for embryo development, female reproduction, tissue repair, inflammatory diseases, and tumor growth and metastasis. The angiogenesis process includes the breakdown of the basement membrane, migration and proliferation of endothelial cells, formation of new vessels by the recruitment of pericytes.³² Most human tumors overexpress VEGF, which stimulates angiogenesis and tumor growth.³³ Inhibition of angiogenesis via VEGF targeting is becoming an important approach for cancer treatment. In this study, we demonstrate that transcriptional inactivation of matrix metalloproteinase-2 decreased integrin-αVβ3-mediated inhibition of the PI3K/AKT pathway. Transcriptional suppression of MMP-2 resulted in decreased VEGF expression, which led to decreased angiogenesis.

MMP inhibition studies support the notion that angiogenesis is dependent, at least in part, on the actions of MMPs since both TIMPs and synthetic MMP inhibitors [e.g., BB94 (Batimastat) and BMS-275291] display anti-angiogenic properties in vitro and in vivo.^24,34,35 In a Swarm rat chondrosarcoma model, in which MMP-2 activation was associated with the angiogenic phenotype, suppression of MMP-2 activity by antisense oligonucleotides inhibited angiogenic potential and tumor growth.³⁶ Further, Itoh and co-workers demonstrated that MMP-2-deficient mice displayed reduced tumor-induced angiogenesis.²¹ The overexpression of MT1-MMP, the activator of pro-MMP-2, exerted similar effects and increased the tumorigeneity of glioma xenografts.³⁷ Similarly, we have previously demonstrated that siRNA-mediated transcriptional suppression of MMP-2 decreased VEGF protein levels in glioblastoma cell lines and efficiently blocked angiogenesis both in vitro and in vivo.³⁸ In the present study, siRNA-mediated knockdown of MMP-2 at transcriptional level inhibited VEGF expression and tube formation in an in vitro angiogenic assay. Moreover, addition of recombinant human-MMP-2 (rhMMP-2; 25 ng/mL) reversed Ad-MMP-2-Si inhibition of VEGF expression and angiogenesis in vitro.

MMP-mediated alterations in VEGF expression have been previously documented. Recently, MMP-2 inhibition with pharmacological inhibitors blocked radiation-induced VEGF upregulation and inhibited melanoma tumor growth and angiogenesis in vivo.³⁹ Similarly, MMP-9 and MMP-2 directly induced VEGF levels in ovarian carcinoma cell lines.⁴⁰ Silletti and coworkers demonstrated that the interaction of MMP-2 and integrin-αVβ3 on the cell surface is important for the induction of downstream signaling during angiogenesis. They used an organic molecule (TSRI265), which blocked the interaction MMP-2 and integrin-αVβ3, but does not have any effect on integrin-αVβ3 binding to its extracellular matrix ligand, vitronectin or on MMP-2 activation or catalytic activity.^18,19 In our study, western blot and zymography analysis for MMP-2 protein and gelatinolytic activity in anti-integrin-αVβ3 immunoprecipitated proteins of Ad-MMP-2-Si-infected A549 cells revealed that MMP-2 downregulation led to decreased interaction of MMP-2 and integrin-αVβ3 in A549 cells. In addition, our studies also indicate that MMP-2 suppression inhibits MMP-2 proteolytic interaction with integrin-αVβ3 as pharmacologic inhibition of MMP-2 activity decreased MMP-2 interaction with integrin-αVβ3 in immunoprecipitation studies. Further, addition of rhMMP-2 increased interaction of MMP-2 and integrin-αVβ3 on the cell surface and induced the downstream PI3K/AKT signaling pathway. These results demonstrate that the role of MMP-2 in regulating VEGF expression occurs through interference with cell signaling through interaction with receptors and the integrin-αVβ3.

Our studies also indicate that transcriptional suppression of MMP-2 decreased hypoxia-induced transcription factor, HIF-1α, which regulates VEGF expression. One of the major stimuli of cancer angiogenesis is hypoxia, which activates hypoxia-inducible transcription factors (HIFs), which in turn induce the expression of VEGF and other angiogenesis factors.⁴¹ In normal cells, HIF-1α is transiently expressed as a result of the action of the HIF-prolyl hydroxylase that targets HIF-1α to an ubiquitin ligase complex containing von-Hippel-Lindau, which marks it for destruction by the proteasome. In tumor cells, a number of factors stabilize HIF-1α and translocate HIF-1α into the nucleus.¹⁹ Enatsu and co-workers observed in lung adenocarcinoma that expression of HIF-1α was significantly higher in cases with vascular invasion, lymph node involvement, and VEGF-A expression.⁴² In our studies, under normoxic conditions, HIF-1α expression was not observed either in Ad-MMP-2-Si-treated cells or the controls. This is in accordance with the known data, which has suggested that under normal conditions, the tumor suppressor gene VHL binds to HIF-1α and targets it for degradation. Nevertheless, under hypoxic conditions, strong expression of HIF-1α was observed in mock and Ad-SV-treated controls. The stabilization of HIF-1α expression and activity were inhibited in MMP-2-transcriptionally inactivated A549 cells.

PI3K and AKT have emerged as important components of critical signaling pathways that are activated by growth factor receptors. PI3K and AKT are crucial for the signaling of angiogenic growth factors that promote endothelial cell migration, proliferation, microtubule formation, and survival.⁴³ The cellular proliferation, survival and migration required for vascular sprouting, and endothelial cell differentiation leading to microtubule formation is primarily driven by VEGF/VEGFR activation that can in turn trigger the PI3K/AKT pathway. In this study, we show that siRNA-mediated MMP-2 knockdown inhibits PI3K, AKT and the downstream survival signal transduction pathway. MMP-2 and MT1-MMP were identified as important mediators of tumor cell induced expression of endothelial genes that are involved in the formation of vasculogenic like networks.⁴⁴ Previous studies indicate that PI3K regulates the protein levels of MT1-MMP.⁴⁵ Consistent with these observations we show that, MT1-MMP levels were decreased in Ad-MMP-2-Si infected cells, possibly due to the inhibition of PI3K signaling in these cells (Supplementary Fig.2). We also show that adding rhMMP-2 to Ad-MMP-2-Si-infected cells restores Ad-MMP-2-Si inhibited PI3K/AKT and VEGF expression. Further, blocking of integrin-αVβ3 with functional blocking antibody in these cells also inhibited PI3K/AKT and VEGF expression, thereby implying that interaction of MMP-2 and integrin-αVβ3 is required for PI3K/AKT-mediated VEGF expression. Taken together, these results demonstrate that transcriptional suppression of MMP-2 inhibits VEGF expression via inhibition of the integrin-αVβ3-mediated PI3K/AKT signaling.

We also observed that capillary tube formation, which was significantly inhibited when HMECs were cultured in tumor conditioned medium from MMP-2-transcriptionally inactivated A549 cells, and was reversed when HMECs were cultured in tumor conditioned medium from A549 cells transfected with constitutively active AKT prior to Ad-MMP-2-Si infection. Consistent with the in vitro results, the immunoreactivity of anti-factor-VIII antibody, which allows for the visualization of blood vessels, the immunoreactivity of the VEGF, and co-localization of integrin-αVβ3 and MMP-2 were all decreased in tumor tissue sections from mice that received Ad-MMP-2-Si treatment. These results clearly show that the transcriptional inactivation of MMP-2 leads to angiogenic inhibition in vivo.

In conclusion, our study demonstrates that transcriptional suppression of MMP-2 may attenuate tumor angiogenesis by reducing VEGF-A, a primary pro-angiogenic protein. This relationship was also determined to be mediated by the integrin-αVβ3 and the PI3K/AKT signaling pathway (Scheme 1). Taken together, this study shows that siRNA-mediated transcriptional knockdown of MMP-2 exhibits significant anti-angiogenic potential and may be a novel therapeutic agent for lung cancer treatment targeting VEGF expression.

Scheme 1. MMP-2 interaction with integrin-αVβ3 induces PI3K/AKT-mediated VEGF expression resulting in angiogenesis.

MMP-2 siRNA inhibits PI3K/AKT-mediated VEGF expression resulting in decreased angiogenesis.

Abbreviations used

MMP

matrix metalloproteinase

siRNA

short interfering RNA

RT-PCR

reverse transcription polymerase chain reaction

VEGF

Vascular endothelial growth factor

ECM

extracellular matrix

VEGFR-2

Vascular endothelial growth factor receptor-2

HMEC-1

microvascular dermal endothelial cells

HIF

Hypoxia inducible factor

EMSA

Electrophoretic mobility shift assay

PI3K

phosphatidylinositol 3-kinase

rhVEGF

recombinant human-VEGF

rhMMP-2

recombinant human-MMP-2