The role of vascular endothelial growth inhibitor in wound healing

Abstract

Vascular endothelial growth inhibitor (VEGI) is an endothelial cell‐specific cytokine and a potent inhibitor of endothelial cell proliferation and angiogenesis. The role of VEGI in angiogenesis related to tissue repair has previously not been investigated. Biopsies from different wound types were analysed by immunohistochemistry and quantitative polymerase chain reaction for the presence of VEGI protein and transcript, respectively. Human vascular endothelial cell line was transfected with VEGI expression plasmid and tested for their in vitro angiogenesis properties. Immunohistochemical staining for VEGI showed reduced expression in the dermal layer of the acute wounds compared with the chronic wound or normal skin. The ability of VEGI to prevent angiogenesis by in vitro assays showed that VEGI acts as a suppressor to the proliferation and microtubule formation of endothelial cells, and the addition of Hepatocyte Growth Factor had little effect on the ability of cell lines expressing the VEGI gene to increase microtubule formation. The aberrant expression of VEGI in different wound types appears to be linked to the outcome of the healing in these wounds. The altered expression of VEGI in chronic wounds constitutes an important target of future therapies.

Introduction

Vascular endothelial growth inhibitor (VEGI) is an endothelial cell‐specific gene and a potent inhibitor of endothelial cell proliferation, angiogenesis and tumour growth (1). This protein is a type II transmembrane protein, with a molecular mass of 13 kDa, belonging to the tumour necrosis factor (TNF) superfamily of cytokines (2). Unlike other members of the TNF family, VEGI is expressed specifically in endothelial cells. Total RNA obtained from many adult organs and tissues contains VEGI messenger RNA (mRNA), suggesting its expression by endothelial cells of the quiescent vasculature 3, 4. Therefore, suggesting that the VEGI gene may be involved in suppressing the proliferation and differentiation of endothelial cells in a normally quiescent vasculature in adults (4). The role of VEGI in angiogenesis related to tissue repair has previously not been investigated.

VEGI inhibits angiogenesis in part by directly inhibiting endothelial cell proliferation (1). These results support the notion that VEGI is most likely not competing for receptors of the different growth factors but exerting its activity by binding to a specific cell surface receptor to initiate a unique signalling pathway that would lead to the termination of angiogenesis (1). The effects of VEGI on Hepatocyte Growth Factor‐induced (HGF‐induced) endothelial tubule formation and proliferation of endothelial cells have not been investigated previously.

Aims

The aim of this study was to (1) assess the expression of VEGI in various wounds in vivo and (2) assess the role of VEGI in HGF‐induced in vitro angiogenesis.

Materials and methods

Materials

Anti‐human VEGI was purchased from Santa‐Cruz Biotechnologies Inc, Santa‐Cruz, CA. A universal immunohistochemical kit, Elite ABC Kit, was purchased from Vector Laboratories^®, Peterborough, UK. RNA extraction kit and reverse transcription kit were obtained from AbGene Ltd (Surrey, Epsom, UK). Polymerase chain reaction (PCR) primers were designed using Beacon Designer (Palo Alto, CA) and synthesised by Invitrogen Ltd. (Paisley, UK). Molecular‐biology grade agarose and DNA ladder were from Invitrogen. Master Mix for routine PCR and quantitative PCR was from AbGene. HGF was a generous gift from Dr T. Nakamura (Osaka, Japan). Matrigel (reconstituted basement membrane) was purchased from Collaborative Research Products (Bedford, MA).

HECV cell line

The human endothelial cell line, HECV, was obtained from Interlab Cell line Collection (ICLC), Naples, Italy.

Skin biopsies

Skin biopsies were retrieved from the departmental tissue bank (Table 1). The following biopsies were included.

Table 1.

Demographics of patients who underwent wound biopsy or biopsy of healthy skin

Patient number	Age	Ulcer type	Sex
1	32	Acute pilonidal	M
2	21	Acute pilonidal	M
3	17	Acute pilonidal	M
4	19	Acute pilonidal	F
5	29	Acute pilonidal	F
6	26	Acute pilonidal	M
7	22	Acute pilonidal	M
8	17	Acute pilonidal	M
9	23	Acute pilonidal	F
10	22	Acute pilonidal	M
11	70	Chronic venous	M
12	81	Chronic venous	F
13	60	Chronic venous	F
14	82	Chronic venous	M
15	73	Chronic venous	F
16	73	Chronic venous	M
17	68	Chronic venous	F
18	64	Chronic venous	M
19	63	Chronic venous	M
20	82	Chronic venous	F
21	71	Chronic arterial	M
22	59	Chronic arterial	M
23	65	Chronic arterial	M
24	63	Chronic arterial	F
25	72	Chronic arterial	M
26	69	Chronic arterial	F
27	64	Chronic arterial	M
28	22	Normal skin	F
29	35	Normal skin	M
30	21	Normal skin	F
31	23	Normal skin	F
32	31	Normal skin	F
33	45	Normal skin	M
34	32	Normal skin	F
35	23	Normal skin	F
36	22	Normal skin	F
37	26	Normal skin	M

F, female; M, male.

Chronic wound tissue

Biopsies from 17 patients with chronic venous leg ulcers were used during the study. Venous disease was diagnosed by Duplex ultrasonography, and all wounds had been present for a minimum of 6 months, with no evidence of healing occurring 6 weeks before biopsy. The wounds had a minimum area of 4 cm² before biopsy and were clinically free of infection. Using an aseptic technique, 6‐mm punch biopsies were taken following the application of local anaesthetic (1% Lidocaine) from the wound margin, incorporating epidermis and dermis at the wound edge (WE) with adjacent granulation tissue (Figure 1).

Figure 1.

Sub‐division of wound biopsy: (1) wound edge (WE), i.e. skin immediately adjacent to the wound bed, (2) approximately 2 mm from the WE abbreviated to MFE (Medium Far from the wound Edge) and (3) approximately 4 mm from the WE abbreviated to FFE (Far From the wound Edge). Sub‐dividing the wound in this way enabled comparison of the dermis at regular intervals from the wound bed. (Adapted with permission of Prof. KG Harding.)

Acute wound tissue

Single wedge biopsies were obtained from 10 patients with acute surgical wounds after undergoing excision of pilonidal disease. These wound were judged to be clinically non infected by a wound healing expert. The biopsies were obtained from the edge of the healing wound within 6 weeks from the surgical excision.

Normal skin

To compare with wound tissue, normal, unwounded skin was also examined. Under local anaesthetics, 3‐mm punch biopsies were taken from the inner aspect of the upper arm of 10 healthy volunteers working within the Wound Healing Research Unit.

Immunohistochemical staining

The method was modified from VEGI staining protocol recently reported (5). The wound and skin tissues were frozen sectioned using a Leica cryostat. Frozen sections were first air‐dried, fixed in acetone for 15 minutes and air‐dried again. Frozen sections of the biopsies were allowed to reach room temperature, and after the foil was removed, the sections were fixed in acetone (Fischer Scientific Ltd, Loughborough, UK) for 15 minutes. Excess acetone was removed by air‐drying the sections for 10 minutes before being washed three times in Tris‐buffered saline (TBS) for 5 minutes each time. Sections were then incubated at room temperature in a humid box with normal blocking solution (Dako Ltd^®, High Wycombe, UK). Excess blocking serum was removed, and the working dilution of primary antibody (made in 1% TBS/BSA) was applied, and the sections were incubated for 30 minutes. Individual antibodies were tested at various dilutions to obtain the optimal working concentrations for use in the immunohistochemical procedure.

Antibody localisation was then identified by a standard streptavidin–biotin peroxidase technique using Vector Elite ABC Kit (Vector Laboratories^®). This involved incubating the sections with a relevant biotinylated secondary antibody for 30 minutes, followed by incubation with the avidin–biotin complex reagent provided in the kit for an additional 30 minutes. The final reaction product was developed for 10 minutes with 3,3′‐diaminobenzidine (DAB) substrate (0·005%) serum (Dako Ltd). The sections were then rinsed in TBS, followed by tap water and then counterstained with Ehrlich’s haematoxylin solution (BDH‐Merck, Poole, UK) for 30 seconds and then washed again in tap water for 5 minutes.

Finally, sections were dehydrated through a graded series of alcohol solutions (BDH‐Merck) for 5 minutes in each and mounted in DPX medium (BDH‐Merck) before mounting under a cover slip.

Positive staining was seen as a brown/black deposit, and non stained cells as blue counterstained nucleated cells, with no associated brown DAB stain. Images were obtained from a digital camera. Staining intensity was semi‐quantified using a method established in our laboratory (6), modified based on a previous report (7). Briefly, grey‐scale digitised images were imported into the Optimas™ software (Optimas 6.0™; Optimus Corp, Bothell, WA). The intensity of the cell staining in the dermis was measured at a magnification (×40) in 10 locations from randomly chosen cells. Control staining (without primary antibody) was used for extraction of the background staining. Intensity data were exported to a Microsoft™ Excel spreadsheet for statistical analysis and are presented as the mean intensity of staining.

Generation of complementary DNA from cell lines and reverse transcription–PCR

Individual biopsies were placed in separate Eppendorfs tubes and rapidly thawed. The biopsy was then homogenised using the Ultra‐Turrax T8 (IKA Labortechnik, Staufen, Germany) in an RNA extraction buffer (AbGene). Total cellular RNA thus extracted from the homogenised biopsies was quantified using a spectrophotometer (WPA UV 1101; Biotech Photometer, Cambridge, UK).

Reverse transcription was performed from 0·5 μg of the RNA sample using oligo‐dT primer according to the manufacturer’s instructions and a reverse transcription kit (Sigma, Poole, Dorset, UK). Complementary DNA (cDNA) was prepared by heating the samples at 47°C for 50–60 minutes, followed by incubation at 75°C for 10 minutes to inactivate any reverse transcriptase.

Conventional PCR primers were designed using Beacon Designer software to allow amplification of regions that have no overlap with other known genes and span at least one intron. Primers were synthesised by Life Technologies (Paisley, Scotland) (Table 2). Conventional PCR was performed using the cDNA from the wound biopsies together with the PCR master mix using the respective primers. The PCR was performed in a GeneAmp PCR system 2400 thermocycler (Perkin‐Elmer, Norwalk, CT) using Reverse‐iT™ First Strand Synthesis Kit (Abgene). The reaction conditions were 94°C for 20 seconds, 60°C for 30 seconds (30 cycles), 72°C for 40 seconds and a final phase of 7 minutes. The PCR products were separated on a 2% agarose gel and stained with 10‐ml ethidium bromide before examination under UV light, and a photograph was taken.

Table 2.

Sequence of the VEGI primers sets, primer sets for detecting the VEGI plasmid and primers sets to detect the VEGI ribozyme knockouts. These primers sets were incorporated into the human endothelial cell line (HECV). The primers were constructed and validated in the host laboratory [Parr et al. (5)]. The sequences of the primer sets used to detect VEGI in quantitative reverse transcription–polymerase chain reaction are shown in the bottom row. The quality of DNA was verified using β‐actin

Primer	Direction	Sequence (5′–3′)
VEGI expression	Forward	VEGIEXF1: ATGAGACGCTTTTTAAGCAA
	Reverse (z‐sequence)	VEGIEXR1: CTATAGTAAGAAGGCTCCAAAGA
		VEGIEXR2: CTATAGTAAGAAGGCTCCAAAG
VEGI ribozyme 1	Forward (z‐sequence)	VEGIRIB1F: CTGCAGTCATTGGGAAACTGTACTG ATGAGTCCGTGAGGA
	Reverse	VEGIRIB1R: ACTAGTGAGACGCTTTTTAAGCAA AGTTTCGTCCTCACGGACT
VEGI ribozyme 2	Forward (z‐sequence)	VEGIRIB2F: CTGCAGTCTCACAACTGGAAACTGATGAGTCCGTGAGGA
	Reverse	VEGIRIB2R: ACTAGTTAATCCTCTTTCTTGTTTCGTCCTCACGGACT
Primers for detecting plasmid		T7F: TAATACGACTCACTATAGGG and BGHR: TAGAAGGCACAGTCGAGG
Primers to detect ribozymes		RBTPF: CTGATGAGTCCGTGAGGACGAA and RBBMR: TTCGTCCTCACGGACTCATCAG
β‐actin	Forward	ATGATATCGCCGCGCTCGTC
	Reverse	CGCTCGGTGAGGATCTTCA
VEGI quantitative PCR	Forward	CAACGTCTAC AGTTTCCCAAT
	Reverse (z‐sequence)	ACTGAAGCTGACCGTACATGATTTTTAAAGTGC TGTGTG

VEGI, vascular endothelial growth inhibitor.

Real‐time quantitative reverse transcription–PCR

Real‐time quantitative PCR was carried out using the iCycler iQ™ system (Bio Rad, Hemel Hemstead, UK) to determine the level of expression of the VEGI and β‐actin transcript in acute and chronic wound tissue and in normal skin, using a method recently described (8). Briefly, the iCycler iQ™ system incorporates a gradient thermocycler and a 96‐channel optical unit. Amplifluor™ detection system was used in the current study (Intergen New York, NY, USA), which included the use of specific sense primer, a universal FAM‐labelled probe and a specific anti‐sense primer that incorporate a z‐sequence that is complementary to the probe (primer sequences, Table 1). In each of the test plates, a serially diluted internal standard was included for calculation purpose (9). Quantitative PCR was carried out in 96‐well plate with 10 pmol FAM‐probe, using a custom hot‐start Q‐PCR master mix, with the following conditions: 95°C for 15 minutes, followed by 50 cycles at 95°C for 15 seconds, 55°C for 40 seconds and 72°C for 15 seconds. The copy number of each transcript was calculated from the internal standards and shown here as copies/50 ng RNA.

Preparation of the expression cassette for human VEGI

Sets of primers designed to amplify the entire coding region of human VEGI gene were used (Table 2). The primers amplified the coding region and were cloned into the respective mammalian expression vector within open reading frame. cDNA used for amplification was from normal human mammary tissues. The successful PCR product was then TA cloned into a pEF6‐V6/His cloning vector (Invitrogen). Following TA cloning, the product was used to transform chemically competent Escherichia coli (OneShot™; Invitrogen). Transformed E. coli was then plated over an ampicillin‐containing LB agar plate and incubated at 37°C overnight. Discrete E. coli colonies were then marked, and tested using a PCR reaction, which allows identification of positive colonies that had the VEGI correctly inserted, both in its ligation direction and in the presence of the product. The expression plasmid and control plasmid were amplified, purified and used to transfect HECV cells by electroporation (Easyjet, Flowgen, Surry, England, UK). Following selection with a blasticidin, stably transfected HECV cell was verified for its over‐expression of VEGI and used for subsequent analysis.

Construction of ribozyme transgenes that specifically target human VEGI

This was based on a method recently described in our laboratories 10, 11. Briefly, suitable sites for hammerhead ribozyme were selected based on the secondary structure of VEGI, which was generated using the Zucker’s mFold programme. Two respective ribozymes were generated using touch‐down PCR with the respective primers (Table 2). Ribozyme transgenes targeting human VEGI were constructed using a mammalian expression vector using a similar approach as given earlier. Three different VEGI ribozyme transgenes (VEGI ribozymes 1, 2 and 3) were thus generated and used to transfect HECV cells as described above. VEGI ribozyme 2 was deemed to be expressed more effectively by the transfected HECV cells and was used in the subsequent assays.

VEGI – expression and cell growth

An expression and cell growth assay was performed to determine changes in cell growth of the various transfected and non transfected cell lines. Four 96‐well culture plates were used. Cells were counted with a haemocytometer counting chamber, and a specific number of cells (5000) were seeded to each well with culture medium (DMEM). The culture plates were then incubated at 37°C for 1, 3 and 5 days. The wells were then treated with 4% formaldehyde, before being stained with 5% (w/v) crystal violet solution. The crystal violet staining extracted with 10% acetic acid, and absorbance was read on a multi‐channel plate reader.

VEGI – expression and tubule formation

The endothelial tubule formation assay was based on that reported previously and modified in our laboratory (12). Aliquots (100 μl) of 200 μg/ml Matrigel (reconstituted basement membrane; Becton Dickinson, Bedford, MA, USA) were added to wells in a 96‐well plate and gelled at 37°C. HECV (in 100 μl DMEM or 100 μl of DMEM containing 50 μl of HGF) were seeded onto re‐hydrated Matrigel at 10⁵ cell/well and incubated at 37°C for 2 hours in order to attach to Matrigel. The medium was then carefully aspirated, and a second layer of Matrigel was added to sandwich the cells. After gellation of the second layer of Matrigel, 100 μl of DMEM (or 100 μl of DMEM containing 50 μl of HGF) was added to each well. The cells were incubated at 37°C for 24 hours. Endothelial microtubule formation was visualised microscopically. Lengths of these tubules are given in units (number of pixels was used) using Optimas 6.0™ software package. The mean length was obtained from two independent experiments (8–10 frames per experiment). Tubule‐forming ability increased with the addition of HGF, although none of these differences reached statistical significance.

Results

Immunohistochemical identification of VEGI in human chronic and acute wound tissues and in normal skin

Standard immunohistochemical techniques were used to morphologically identify the distribution of VEGI in different types of skin tissues, namely normal skin and acute and chronic wound tissues. Staining of the respective protein was first assessed qualitatively using standard light microscopy and then quantitatively analysed for staining intensity using the Optimas 6.0™ software system.

The skin adjacent to the wound bed in acute and chronic wound biopsies was sub‐divided into three at a magnification of ×40: (1) WE, i.e. skin immediately adjacent to the wound bed, (2) approximately 2 mm from the WE abbreviated to MFE (Medium Far from the wound Edge) and (3) approximately 4 mm from the WE abbreviated to FFE (Far From the wound Edge). Sub‐dividing the wound in this way enabled comparison of the dermis at regular intervals from the wound bed (Figure 1).

When assessed qualitatively, VEGI staining was found to be greatest in the dermal layer of the skin and the endothelial cells in the skin tissues (Figure 2). Quantitative assessment of VEGI staining using image analysis software (Optimas 6.0™) revealed that the staining in the dermal layer of the acute wound tissue was similar to that of the chronic wound tissue and normal skin (Figure 3). However, staining for VEGI was reduced at the WE in the acute wound tissue and increased moving away from the WE towards normal skin. This is in contrast to the chronic wound tissue where the staining for VEGI was greatest at the WE and decreased moving towards normal skin.

Figure 2.

Immunohistochemical staining for vascular endothelial growth inhibitor within the wound tissue dermis. Where positive staining is seen as a brown/black deposit and non stained cells as blue. Positive dermal staining was assessed semi‐quantitatively at a magnification of ×40 by randomly selecting 10 area (marked as 1–10). The staining intensity of these random areas minus the mean negative control staining intensity was analysed by Optimas 6.0™ software. The chronic and acute wound tissues were compared at the wound edge (WE), 2 mm from the WE (MFE, Medium Far from the wound Edge) and 4 mm from the WE (FFE, Far From the wound Edge). The wound tissue was compared with normal skin in a similar manner.

Figure 3.

Comparison of vascular endothelial growth inhibitor staining between acute and chronic wounds and normal skin. Error bars represent 1 standard deviation from the mean.

Analysis of VEGI mRNA, qualitative and quantitative approaches

Both conventional reverse transcription–polymerase chain reaction (RT‐PCR) (Figure 4) and quantitative real‐time RT‐PCR (Figure 5) were used to analyse the presence and quantity of messages of VEGI. Comparison was made between acute wound tissue, chronic wound tissue and normal skin. Expression of VEGI transcripts in acute wounds was less than that in chronic wounds and normal skin, although these differences did not reach statistical significance (Figure 6).

Figure 4.

Reverse transcription–polymerase chain reaction of vascular endothelial growth inhibitor (VEGI) messenger RNA (mRNA) from acute wound tissue, chronic wound tissue and normal skin separated by agarose gel electrophoresis, stained with ethidium bromide, viewed under UV light and compared against a molecular weight markers (marked in KDa on the left of the figure). There is a lower concentrations of VEGI mRNA transcripts in acute wound tissue than in the chronic wound tissue and normal skin. VEGI is an anti‐angiogenic cytokine that is believed to maintain the vasculature of the adult in a quiescent state. Acute wounds require a pro‐angiogenic environment, whereas one of the reasons why chronic wounds are slow to heal is poor angiogenesis. The higher expression of VEGI in chronic wound tissue may in part explain this.

Figure 5.

Quantitative real‐time reverse transcription–polymerase chain reaction of VEGI transcripts in acute and chronic wound tissues and in normal skin. Typically, amplification of a given complementary DNA over time follows a curve, with an initial flat phase, followed by an exponential phase. Finally, as the experiment reagents are used up, DNA synthesis slows down and the exponential curve flattens into a plateau. Initial template levels are determining when the signal rises above the set threshold value (horizontal orange line). Data analysis including standard curve (reference standard for extrapolating quantitative information for mRNA targets of unknown concentrations) is generated automatically. Each curve represents a separate sample (acute wound tissue, n = 10; chronic wound tissue, n = 17 and normal skin, n = 10).

Figure 6.

Mean number of transcript copies/5 μl complementary DNA of VEGI transcripts in acute and chronic wound tissue and in normal skin. Error bars represent 1 standard deviation from the mean.

Genetically manipulating the expression of VEGI and the impact on the growth and tubule formation from endothelial cells

We generated a panel of endothelial cell lines that had different profile of VEGI expression: VEGI‐over‐expressing cells (transfected with VEGI expression vector), VEGI non expression cells (transfected with VEGI ribozyme transgene) and plasmid control cells (transfected with empty control plasmid), together with the wild‐type HECV cells. These cells were used in the tubule‐forming assay and cell growth assays (Figure 7). HGF increased the formation of tubules in all three cell types. VEGI expression cells showed a shorter tubule length compared with wild type. However, ribozyme transgene had only limited impact on the tubules. The mean cell growth after 5 days for VEGI expression plasmid was 0·131 (0·061–0·148) and 0·284 (0·212–0·497) for VEGI ribozyme knockout. The wild‐type HECV had a mean growth after 5 days of 0·436 (0·342–0·599), while the PEFα plasmid control had a mean of 0·295 (0·230–0·432) (Figure 8). These differences did not reach statistical significance (mixed analysis of variance, P > 0·05 at 5 days).

Figure 7.

HECV microtubule formation with or without 50 μg of HGF. ⋄, plasmid control. Error bars represent 1 standard deviation from the mean. WT, wild type.

Figure 8.

The cell growth assay was performed to quantify changes in cell growth of the various transfected and non transfected cell lines at 1, 3 and 5 days. Cell growth of HECV line expressing the VEGI was less than the cell lines expressing the VEGI ribozyme, plasmid control or the wild‐type (WT) HECV. These differences did not reach statistical significance (mixed analysis of variance, p > 0·05 at 5 days).

Discussion

The role of VEGI in angiogenesis related to tissue repair has previously not been investigated. This study shows that the staining in the dermal layer of the acute wound tissue was similar to that of the chronic wound tissue and normal skin. However, staining for VEGI was reduced at the WE in the acute wound tissue and increased moving away from the WE to a maximum at 4 mm. This is in contrast to the chronic wound tissue where the staining for VEGI was greatest at the WE and decreased moving away. Expression of VEGI transcripts is reduced in acute wound tissue compared with chronic wound tissue or normal skin. The expression of VEGI transcripts in the chronic wound tissue was similar to that in the normal skin. These trends did not reach statistical significance.

Overall, the distribution of VEGI in various wound types is similar to the other inhibitors of HGF (HGF activation inhibitors, i.e. HAI‐1 and HAI‐2). These results corroborate the assertion that VEGI is involved in suppressing the proliferation and differentiation of endothelial cells in a normally quiescent vasculature in adults. Normal wound healing requires the formation of new blood vessels from existing vessels at the WE in order to achieve successful healing. An increased expression of VEGI, as seen at the WE of chronic wound tissue, would imply that formation of new blood vessels is inhibited in these wounds, which ultimately are slow to heal or remain non healed. The increased expression of VEGI in normal skin also corroborates this assertion, as it is important to maintain a quiescent vasculature in healthy tissue.

There is accumulating evidence indicating that angiogenesis is regulated by both positive and negative regulatory factors, the regulatory mechanisms balancing the angiogenic versus angiostatic drive remain poorly understood (13). The effects of VEGI on HGF‐induced endothelial microtubule formation and proliferation of endothelial cells has not been investigated previously.

This study shows endothelial proliferation after 5 days for the VEGI expression cassette to be markedly reduced compared with the VEGI ribozyme knockouts, wild‐type HECV and the pEFα plasmid control.

The sum microtubule length for HECV expressing the VEGI gene was less than that for HECV cells expressing the VEGI ribozyme or the pEFα plasmid control. The wild‐type HECV and the pEFα plasmid control HECV cell lines had a greater ability to form microtubule than the cell lines expressing the VEGI gene. The VEGI ribozyme blocked the anti‐angiogenic effects of the VEGI gene, and overall microtubule formation was greater than the HECV cell lines expressing the VEGI gene.

Addition of HGF to the endothelial microtubule formation assay increased the ability of the wild‐type HECV and the cell lines expressing the VEGI ribozyme to form microtubules. HGF had little effect on the ability of cell lines expressing the VEGI gene to increase microtubule formation. Although none of these differences reached statistical significance.

The results of this study are in keeping with the results from previous studies, which reported VEGI functions, in part, by directly inhibiting endothelial cell proliferation (1). The results of this study also support the notion that VEGI is most likely not competing for growth factor receptors but exerting its activity by binding to a specific cell surface receptor to initiate a unique signalling pathway that leads to a termination of angiogenesis.