Aims Vascular endothelial growth factors (VEGFs) and their receptors (VEGF-Rs) are among the most powerful factors regulating vascular growth. However, it has remained unknown whether stimulation of VEGF-R1, VEGF-R2 or both of the receptors produces the best angiogenic responses in myocardium. The aim of this study was to compare the VEGF-R1-specific ligand VEGF-B186, VEGF-R2-specific ligand VEGF-E and VEGF-A165, which stimulates both receptors, regarding their effects on angiogenesis and left ventricular function in mice.
Methods and results High-resolution echocardiography was used to guide the closed-chest injections of adenoviral (Ad) vectors expressing VEGF-B186, VEGF-E, and VEGF-A165 into the anterior wall of the left ventricle in C57Bl/6J mice. Angiogenic and functional effects were analysed using histology, ultrasound and perfusion analyses 6 (D6) and 14 (D14) days after the Ad injection. AdVEGF-A165 induced a strong angiogenic response seen as an enlargement of myocardial capillaries whereas angiogenesis induced by AdVEGF-B186 and AdVEGF-E seemed more physiological. The increase in the capillary area was accompanied with an increase in myocardial perfusion at D6 after the gene injection. AdVEGF-A165 and AdVEGF-E induced endothelial-specific proliferation whereas AdVEGF-B186 mostly induced proliferation of cardiomyocytes. AdVEGF-A165 induced more pronounced tissue damage than AdVEGF-B186 and AdVEGF-E. Left ventricular function measured as ejection fraction did not change during the follow-up. AdVEGF-A165 increased both VEGF-R1 and VEGF-R2 protein expression whereas AdVEGF-B186 and AdVEGF-E did not affect endogenous receptor expression levels.
Conclusion AdVEGF-B186 and AdVEGF-E are equally potent in inducing therapeutic angiogenesis in mouse myocardium and produce less side effects than AdVEGF-A165.
Left ventricular function
Coronary artery disease (CAD) remains the leading cause of death in the western world in spite of improved control of the risk factors, availability of medical therapies and highly effective invasive treatments.1 This makes the therapeutic vascular growth, which involves stimulation of angiogenesis and arteriogenesis, an intriguing concept for the treatment of CAD.1–4
Members of the vascular endothelial growth factor (VEGF) family are among the most powerful factors to modulate vascular growth.5 The VEGF family consists of several members: VEGF-A, VEGF-B, VEGF-C, VEGF-D, VEGF-E, VEGF-F, and placental growth factor (PlGF). These share similar structures but differ in their physiological and biological properties largely due to their different interaction with three specific tyrosine kinase receptors: VEGF receptor (VEGF-R)1, VEGF-R2, and VEGF-R3.6
VEGF-A, also called VEGF, is considered as the master regulator of angiogenesis and vascular permeability.5,7,8 VEGF-A is a specific mitogen for vascular endothelial cells (ECs) and a ligand for both VEGF-R1 and VEGF-R2. VEGF-A has four major isoforms: VEGF-A121, VEGF-A165, VEGF-A189, and VEGF-A206, of which VEGF-A165 is the most used isoform in therapeutic angiogenesis having heparin-binding properties.9–11 However, recent findings suggest that VEGF-A has a narrow therapeutic window and severe side effects have been reported from even two- to three-fold over expression.12
VEGF-B is a ligand for VEGFR-1 and neuropilin (Nrp)-1 and is expressed in the heart, skeletal muscle, adipose tissue, and smooth muscle cells in adults.5 It is produced in two different isoforms: heparin-binding VEGF-B167 and diffusible VEGF-B186.13,14 VEGF-B has shown angiogenic potency in non-ischaemic tissues, likely through indirect mechanisms.15 Recent results also suggest some tissue-specificity of VEGF-B induced angiogenesis.16
VEGF-E is a novel, Orf-virus derived protein which is a specific ligand for VEGF-R2 and Nrp-1. Humanized, diffusible VEGF-E with PlGF-derived amino acid residues at both the amino and carboxyl terminal ends has shown angiogenic potency without oedema in the skin of VEGF-E/PlGF transgenic mice12 and rat hindlimb ischaemia17 but its effects have not been studied in myocardium.
VEGF-R1 and VEGF-R2 are expressed in the adult murine heart.18 VEGF-R1 and VEGF-R2 play different roles in physiological and pathological angiogenesis. VEGF-R1 has a dual role in angiogenesis, a negative control during embryogenesis and a positive role in adulthood. In addition to activating EC proliferation, VEGF-R1 promotes tumour growth, metastasis, and inflammation whereas VEGF-R2 is considered the major inducer of angiogenesis.19 By activating both VEGF-R1 and VEGF-R2, VEGF-A stimulates EC proliferation, migration, and survival, tubular formation, and vascular permeability in vivo. Activation of only VEGF-R1 by its specific ligand VEGF-B stimulates only some of the above-mentioned phenomena. Signalling mediated through VEGF-R2 is activated by its specific ligand VEGF-E.20
The aim of this study was to analyse the effects of VEGF-R1-specific ligand VEGF-B186 and VEGF-R2-specific ligand VEGF-E on their angiogenic response and left ventricle function in mouse and to compare them to VEGF-A165, which stimulates both VEGF-R1 and VEGF-R2, to find the most suitable factor for therapeutic angiogenesis in the heart.
2.1 Experimental animals
A total of 201 C57Bl/6J male mice from 9 to 13 weeks of age were used to study the effects of different VEGFs in myocardium. The mice were kept in standard housing conditions in The National Laboratory Animal Center of Kuopio University. Diet and water were provided ad libitum. All animal procedures were carried out in accordance with the guidelines of the Experimental Animal Committee of Kuopio University. The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996).
2.2 Viral constructs
Human clinical grade first-generation serotype 5 replication-deficient (E1, partially E3 deleted) adenoviruses were produced under GMP conditions in 293 cells and analysed to be free from endotoxin and microbiological contaminants.21,22 All viral constructs were tested in vitro for their protein production by western blotting.
2.3 ELISA and western blotting
The amount of human VEGF-A165, VEGF-B186, total mouse VEGF-A, VEGF-R1, or VEGF-R2 in the left ventricle was determined using ELISA (human VEGF-A ELISA kit, mouse VEGF-A ELISA kit, mouse VEGF-R1 and VEGF-R2 ELISA kits, R&D Systems, Minneapolis, MN, USA and human VEGF-B ELISA designed in Ulf Eriksson's laboratory). The amount of VEGF-E (antibody C-20, Sánta Cruz Biotechnology, Inc., Santa Cruz, CA, USA) in the left ventricle was determined and quantified by western blotting. Functionality of viral constructs was determined by in vitro transduction of C2C12 cell line (ATCC, Manassas, VA, USA) cultured as instructed by ATCC in complete growth media. Conditioned media were analysed with western blot as stated above using the following antibodies: VEGF-A165 (C-1, Sánta Cruz Biotechnology, Inc.), VEGF-B186 (Af751, R&D Systems), and VEGF-E (C-20 Sánta Cruz Biotechnology, Inc.).
2.4 Echocardiography-guided myocardial injections
Adenoviral (Ad) constructs of VEGF-A165, VEGF-B186, VEGF-E, LacZ, or 0.9% saline (B. Braun Melsungen AG, Germany) were injected into the anterior wall of the left ventricle under ultrasound guidance. Viral constructs were diluted in 0.9% saline to a final concentration of 1 × 1012 viral particles (vp) in millilitre and a total of 1 × 1010 vp in 10 µL was injected into the mouse myocardium. The injections were carried out mainly as described by Springer et al.23 Briefly, a 30-gauge disposable needle in a 50 µL Hamilton syringe was connected to a micromanipulator (VisualSonics Inc., Toronto, ON, Canada). The mouse and the needle were aligned with the probe so that the anterior wall of the left ventricle was approached with the needle in an angle of approximately 20°. The needle was penetrated through the chest substernally until the needle reached the target position in the left ventricle wall. Viral injection was performed and visualized by ultrasound image and documented as video clips. After the operation, analgesic (carprofen 50 mg/mL, Rimadyl, Pfizer Inc., NY, USA) was given to the mice.
2.5 Echocardiography, ECG, and myocardial perfusion
A high-resolution imaging system specially developed for small animal research (Vevo 770, VisualSonics Inc., Toronto, ON, Canada) was used to gain echocardiographic data from the mice on the injection days (D0), 6 (D6), and 14 (D14) after the injection. A high-frequency ultrasound probe (RMV-707B) operating at 30 MHz, with a focal depth of 12.7 mm was used. The animals were anaesthetized with isoflurane (induction: 4.5% isoflurane, 450 mL air, maintenance: 2.0% isoflurane, 200 mL air, Baxter International, Inc., Deerfield, IL, USA). After careful removal of hair from the chest with a depilatory (Veet, Reckitt Benckiser, UK), mice were placed in supine position on a heated platform (THM100, Indus Instruments, Houston, TX, USA) to maintain the body temperature at 36–37°C. The temperature of the animals was monitored throughout the study protocol via a rectal probe. The ECG signal was obtained from the electrode pads on the mouse platform. Warm ultrasound gel (Aquasonic 100, Parker Laboratories, Inc., Orange, Fairfield, NJ, USA) was applied to the shaved chest before the placement of the ultrasound probe. Ejection fraction and left ventricle diastolic wall thickness were determined from parasternal short-axis M-mode measurements. Ejection fraction was calculated by Vevo770 software by using the Teicholz formula:
where LVID;d is the left ventricular internal diameter (diastole) andLVID;s the left ventricular internal diameter (systole).
ECG was monitored during the echocardiography. An ECG sample of 30 s of each mouse was analysed. Time intervals (PQ time, QRS-time, QTc-time) were measured from the mean curve of 30 s with a specially made MatLab analysis program (Kubios HRV analysis program version 2.0 beta 4, Department of Physics, University of Kuopio, Finland). QT time was normalized to QTc time with the following formula:
Cardiac perfusion was measured at D0, D6, and D14 with Cadence™ contrast pulse sequence (CPS) ultrasound by Acuson Sequoia C256 using 15L8 probe (Siemens Medical Solutions, Mountain View, CA, USA) and a second generation microbubble contrast agent (SonoVue®, Bracco Diagnostics, Inc., Princeton, NJ, USA) diluted 1:3 in 0.9% saline.24 The contrast agent was administered as a 50 µL bolus via tail vein and the CPS was performed using the following parametres at 14 MHz: frame rate, 15 Hz; dynamic range, 80 dB; power, 10 dB; MI, 0.25; CPS gain, 10; S1/0/2/4 Δ4; and depth, 20 mm. CPS signal intensities (dB) of video clips were quantified with Datapro (v2.13, Noesis SA, Courtaboeuf, France) and intensity of the contrast agent in anterior wall was normalized to that of the left ventricle. PowerShowCase (v4.95, Trillium Technology, Inc., Ann Arbor, MI, USA) and WinAVI Video Converter (v7.7, ZJMedia Digital Technology Ltd.) were used for the processing of ultrasound video files for the Windows Media Video 9 format (Microsoft, Redmond, WA, USA).
Avidin–biotin–HRP and avidin–biotin–alkaline phosphatase systems were used for immunohistochemical analyses of 5 µm thick paraffin-embedded sections fixed with 4% paraformaldehyde in 7.5% sucrose for 4 h. The injection site was determined and general histology studied using haematoxylin/eosin stainings. The immunohistological stainings were performed on sections right next to the section with the largest needle tract. Endothelium was immunostained using lectin antibody (Biotonylated Griffonia (Bandeiraea) Simplicifolia Lectin I, dilution 1:100, Vector Laboratorios, CA, USA) and proliferating cells were immunostained using PCNA (Proliferating cell nuclear antigen, dilution 1:100, Zymed Laboratorios, Inc., CA, USA). Cardiomyocytes were immunostained using connexin43/GJA1 antibody (ab11370, Abcam, Cambridge, MA, USA). VEGF-R1 was immunostained using Flt-1 antibody (C-17, Sánta Cruz Biotechnology, Inc.) and VEGF-R2 using Flk-1 antibody (Affinity Purified anti-mouse Flk-1, 14-5821, eBioscience, Inc., San Diego, CA, USA). Photographs of the sections were taken with Olympus AX70 microscope (Olympus, Tokyo, Japan), AnalySIS software (Soft Imagining System GmbH, Germany) and were processed with Adobe Photoshop (v7.0, Adobe Systems, Inc., San Jose, CA, USA).
2.7 Capillary enlargement, cell proliferation, and tissue damage
Mean capillary areas and capillary per square millilitre ratios were measured from five microscopic fields of lectin immunostained sections taken from the surrounding area of the needle tract at ×200 magnification by using AnalySIS software. Cell proliferation was measured as the mean value from five microscopic fields of PCNA/lectin double-stained sections at ×400 magnification by using AnalySIS software. The tissue damage, which means infiltrated inflammatory cells and necrosis, was analysed in a blinded fashion by three observers from haematoxylin/eosin stained sections in ×12.5 magnification using the following grading criteria: +, for mild damage (only the needle mark visible); ++, moderate damage; and +++, severe damage.
2.8 Statistical analyses
Results are presented as means ± SEM, statistical significance was evaluated using t-test, one-way ANOVA or repeated measures two-way ANOVA with Dunnett's multiple comparison test or Bonferroni was used as post-tests. The used statistical analyses are specified in the figure legends. P < 0.05 was considered statistically significant. The following symbols are used in the figures: *P < 0.05 (Figures 3A, 6A and C), **P < 0.01 (Figures 3A and B, 5B, and 6C) and ***P < 0.001 (Figures 3A, 5A, and 6B).
3.1 ELISA and western blotting
After Ad transductions-specific protein products encoded by the transgenes were detected from cell culture media using western blotting (data not shown). Protein concentrations in the AdVEGF-A165, AdVEGF-B186, and AdVEGF-E transduced hearts were the highest at D6 after the injection decreasing towards D14 (Figure 1, n = 5–6). AdVEGF-B186 or AdVEGF-E injections did not increase the amount of endogenous VEGF-A in mouse myocardium compared with the AdLacZ injected group (data not shown).
Protein expression in AdVEGF-A165 (A), AdVEGF-B186 (B) and AdVEGF-E (C) transduced left ventricle at D6 and D14. Protein expression of VEGFs was clearly increased at D6 compared with the LacZ control where levels of the above-mentioned growth factors were below the detection limit. Expression decreased towards D14 time point as expected after adenoviral transduction. nd, not detectable; a.u., arbitrary units; n, 5–6.
3.2 Angiogenic response, proliferation, and inflammation
AdVEGF-A165 induced strong dilatation of capillaries in mouse myocardium. The effect was highest at D6 after the injection, the capillary area being more than 20-fold compared with the AdLacZ control group (Figures 2 and 3A, n = 6–7). Capillaries formed vascular lacunae that were surrounded by ECs. The effect diminished rapidly towards the later time points (Figures 2I, M, J, and N and 3A, n = 6). In AdVEGF-B186- and AdVEGF-E-treated groups the increase in the capillary area was milder, eight- and nine-fold at D6 time point, respectively, compared with the AdLacZ control group. Also, morphology of the capillaries was more physiological and uniform in shape (Figures 2E, G and H and 3A n = 6–7). This effect diminished more slowly than in AdVEGF-A165-treated group: the enlargement of the capillary area was three- and two-fold in AdVEGF-B186-treated group and four- and two-fold in AdVEGF-E-treated group at D14 and D28, respectively, compared with the AdLacZ group (Figures 2I, K, L, M, O, and P and 3A, n = 6–7). In AdVEGF-A165 group the capillary area decreased rapidly at D14, whereas in the AdVEGF-B186 and AdVEGF-E groups the mean area continued to enlarge until D14 (Figure 3A). The effects of adenovirus and β-galactosidase on the morphology of the capillaries were controlled by saline injections. No differences were seen within the morphology of the capillaries between AdLacZ and saline injected myocardium (data not shown). The effects of ADVEGF-A165 with one logarithm lower dose (1 × 1011 vp/mL) were also tested to see if that dose induced more physiological angiogenesis. After the injection of this lower dose, the increase in the capillary area was three-fold compared with the control mice at D6 and returned to the normal level at D14 (data not shown). Gene transfers did not influence capillaries per square millilitre ratio (data not shown).
Angiogenic responses in mouse myocardium at D6, D14, and D28 after AdLacZ, AdVEGF-A165, AdVEGF-B186 and AdVEGF-E gene transfers. In the top row haematoxylin/eosin (HE) stainings of D6 samples (A–D). Lectin endothelial stainings in different time points (E–P). Six days after AdVEGF-A165 gene transfer capillaries were massively enlarged and formed vascular lacunae (B, F) that recovered to uniform but still enlarged capillaries towards the later time points (J, N). AdVEGF-B186 (C, G, K, O) and AdVEGF-E (D, H, L, P) gene transfers induced a more physiological and prolonged angiogenesis, and the enlarged capillaries were uniform in shape in all time points. Scale bars 100 µm, n = 6–7, asterisks indicate the higher power insert area.
The effect of different VEGFs on mean capillary area (A) and cell proliferation in mouse myocardium (B–J). Capillary enlargement was the most prominent in AdVEGF-A165-treated animals but the mean area decreased rapidly towards the later time points. In AdVEGF-B186 and AdVEGF-E-treated groups the angiogenic effect was more physiological, yet significant, and the mean capillary area remained increased at D14 time point (A). The total number of proliferating cells was significantly increased in AdVEGF-A165 (B, D) group compared with the AdLacZ (B, C). In AdVEGF-A165 (B, D, H) and AdVEGF-E (B, F, J) groups the majority of proliferation was in endothelial cells whereas in AdVEGF-B186 (B, E, I) group proliferating cells were mainly in other cell types such as cardiomyocytes. Double stainings (C–F) with lectin (blue) and PCNA (brown), in (G–J) with connexin43 (brown) and PCNA (blue). Scale bars 10 µm, n = 11–13, asterisks indicate the higher power insert area. Statistical analyses were done by one-way ANOVA with Dunnett's multiple comparison as a post-test.
The profile of proliferating cells was different between the groups (Figure 3B–J, n = 11–13). The total number of proliferating cells was significantly higher in AdVEGF-A165 group compared with the AdLacZ group. In AdVEGF-B186 and AdVEGF-E groups a trend towards a higher total cell proliferation was seen. In AdVEGF-A165 group over 90% of the proliferating cells were ECs but in AdVEGF-B186 the percentage of proliferating ECs was only 25, with most of the proliferation being in other cell types, such as cardiomyocytes. In AdVEGF-E group the percentage of proliferating ECs was 61, and the profile resembled that of AdVEGF-A165.
Tissue damage was moderate at D6 in all adenovirus injected groups (Figure 4A, n = 11–13). At D14 the area of inflammatory cells and necrosis was moderate in AdLacZ, AdVEGF-B186, and AdVEGF-E groups and severe in AdVEGF-A165 group. In the saline-injected group only the needle track could be detected at both time points (Figure 4A–G, n = 11–13). Gene injection itself or Ad vector did not have any effects on ECG (Figure 4H–K, n = 6). Also, gene transfer of growth factors did not affect the ECG compared with the saline or AdLacZ control groups (data not shown). The ECG time intervals for the control groups' were the following: PQ (44.0 ± 2.6, 40.9 ± 1.5 s), QRS (16.7 ± 3.1, 13.7 ± 0.9 s), or QTc (40.4 ± 2.3, 42.2 ± 2.3 s) in saline and AdLacZ group, respectively, at D0 and PQ (40.7 ± 1.4, 39.0 ± 2.0 s), QRS (15.9 ± 2.1, 15.8 ± 3.0 s), and QTc (39.1 ± 3.0, 40.2 ± 4.2 s) in saline and AdLacZ group, respectively, at D6. No differences were seen in heart rate between the groups (data not shown).
Tissue damage based on necrosis and infiltration of inflammatory cells and general histology after gene transfers (A–G) and the effect of injections on ECG (H–K). Tissue damage grading was done in a blinded fashion: + refers to mild (needle mark), ++ moderate, and +++ severe damage. Tissue damage was mild in saline group at D6 and D14, moderate in all adenovirus injected groups at D6 as well as in AdLacZ, AdVEGF-B186 and AdVEGF-E groups at D14 and severe in AdVEGF-A165 group at D14. In representative histology sections of saline injected myocardium D6 (B) and D14 (E) only the needle mark was detected (arrow). In AdLacZ group at D6 (C), D14 (F) and in AdVEGF-A165 group at D6 (D) moderate and in AdVEGF-A165 group D14 (G) severe overall tissue damage was seen. The effect of the injection itself and adenovirus on the ECG was controlled by saline and LacZ injections. Representative ECG curves of saline and AdLacZ groups are shown at D0 (H and J, respectively) and D6 (I and K, respectively). No difference was seen in PQ, QRS, or QTc intervals at D0 or D6 between the groups compared with the saline or AdLacZ control groups (one-way Anova). Scale bars 500 µm, n = 11–13 in tissue damage grading, n = 6 in ECG measurements.
3.3 VEGF-R1 and VEGF-R2 expression
VEGF-R1 protein expression was significantly increased in AdVEGF-A165 group compared with the AdLacZ group and to intact mouse heart (Figure 5A, n = 5–6). AdVEGF-B186 and AdVEGF-E did not have any effects on VEGF-R1 protein levels. VEGF-R2 protein level was up-regulated in AdVEGF-A165 group but remained unchanged in other treatment groups compared with the AdLacZ group. (Figure 5B, n = 5–6). VEGF-R1 expression was mostly located in cells in the interstitial space as well as in ECs of larger vessels whereas VEGF-R2 expression was located exclusively in ECs (data not shown).
VEGF-R1 (A) and VEGF-R2 (B) protein concentration D6 after injection with different growth factors. Both VEGF-R1 and VEGF-R2 protein levels were significantly increased in AdVEGF-A165 group but stayed at the level of AdLacZ control group in AdVEGF-B186 and AdVEGF-E groups (A and B). One-way ANOVA with Dunnett's multiple comparison test, n = 5–6.
3.4 Left ventricular function, wall thickness, and myocardial perfusion
Ejection fraction did not change significantly in any of the groups during the 2 week follow-up time (Figure 6A, n = 5–6); however, in the AdVEGF-A165 group the ejection fraction was increased at D6. Left anterior wall thickness was significantly increased at D6 and D14 after AdVEGF-A165 gene transfer indicating myocardial oedema (Figure 6B, n = 5–6). No significant wall thickening was seen in other groups. The increase in the capillary areas led to significant increases in perfusion in the anterior wall in the AdVEGF-A165, AdVEGF-B186, and AdVEGF-E groups D6 after the gene transfer compared with the AdLacZ control group (Figure 6C, n = 5–7).
Ejection fraction (A), left ventricle anterior wall thickness in diastole (B) and myocardial perfusion (C) after injection of saline, AdLacZ, and adenoviruses expressing different growth factors. Viral vectors did not have a decreasing effect on ejection fraction compared with either AdLacZ or saline injected groups (A). Ejection fraction stayed at the normal level in AdVEGF-B186 and AdVEGF-E groups and was increased in AdVEGF-A165 group at D6 compared with the AdLacZ group (A). A significant increase was seen in the left ventricle anterior wall thickness at D6 and D14 after AdVEGF-A165 injection indicating ventricular wall oedema (B). No such oedema was seen after treatment with saline, AdLacZ, AdVEGF-B186 or AdVEGF-E. Increased capillary area was reflected as a significant increase in myocardial perfusion in all growth factor-treated groups at D6 post-injection (C). n = 5–7, statistical analyses with repeated measures two-way ANOVA (A and B) and one-way ANOVA (C) with Dunnett's multiple comparison test.
We analysed the angiogenic properties of ligands for different VEGF receptors: VEGFR-1-specific ligand VEGF-B186, VEGFR-2-specific ligand VEGF-E and VEGF-A165 which binds to both receptors. Adenoviruses expressing the above-mentioned growth factors were injected using ultrasound guidance through the chest directly into the anterior wall of the left ventricle. The effect of the gene transfer was assessed by using histology 6, 14, and 28 days as well as echocardiography and myocardial perfusion analysis D6 and D14 after the gene injection. Ultrasound-guided Ad injection directly into the mouse myocardium was found a feasible method that avoids traumatic open-chest surgery. The injection itself did not have any measurable effects on the left ventricular function or ECG. In this study we used a sophisticated new technique to assess the myocardial perfusion.24 This CPS technique allows repeated bolus injections and follow-up the development of myocardial perfusion in the treated area in the same animal.
The results showed a strong acute angiogenic response, which was however short term, after AdVEGF-A165 injection. Angiogenesis was seen as an enlargement of capillaries. That also caused oedema in the anterior wall of the left ventricle at D6. Vascular lacunae were also detected. Similar structures were also seen in rabbit hind limb after AdVEGF-A165 transduction.7 One logarithm lower dose of AdVEGF-A165 induced more physiological vascular structures with a major reduction in the degree of angiogenesis. Simultaneous activation of VEGF-R1 and VEGF-R2 by prolonged and repeated administration of recombinant VEGF-A protein has also been shown to enhance the hind limb perfusion and to stimulate arteriogenesis.25 The angiogenic effects of AdVEGF-A165 diminished much faster than those of AdVEGF-B186 and AdVEGF-E. In AdVEGF-B186- and AdVEGF-E-treated groups angiogenesis was more physiological and did not cause tissue oedema. The increase in the capillary area was accompanied with an increase in myocardial perfusion at D6 after the gene transfer in all groups compared with the AdLacZ group. The effect of Ad vector on necrosis and infiltration of inflammatory cells was studied by comparing Ad groups to saline-injected control group. Only a needle track with mild tissue damage was detected in the saline injected animals. The tissue damage was moderate in all adenovirus injected groups at D6 and in AdLacZ, AdVEGF-B186, and AdVEGF-E groups at D14 but severe in AdVEGF-A165 group at D14. VEGF-A induces macrophage migration that is mainly mediated via VEGF-R1,20 which could be the mechanism leading to the more severe tissue damage seen after AdVEGF-A165 transduction. The strong angiogenic response and tissue oedema cause macrophage infiltration, which could also explain the more severe damage in the AdVEGF-A165 group, since there was no such inflammatory response in the AdVEGF-B186-transduced group after VEGF-R1 activation.
AdVEGF-B186 gave promising results as a myocardial angiogenic factor, as angiogenic effects seen after the transduction with AdVEGF-B186 were more physiological and no oedema was detected. These results agree with recent findings observed with AdVEGF-B186 in pig myocardium.16 Because of the strong oedema seen after the AdVEGF-A165 injection, AdVEGF-B186 as a VEGF-R specific ligand is a more potential candidate for therapeutic angiogenesis, although VEGF-R2 is considered as the main VEGF-R mediating angiogenesis via EC proliferation.19,20 The effects of AdVEGF-B186 could potentially be mediated by up-regulation of endogenous VEGF-A. However, we did not see any significant up-regulation of endogenous VEGF-A compared with the AdLacZ injected group which is in agreement with the recent findings of Lahteenvuo et al.16 Also, it has been suggested, that excess VEGF-R1 specific ligands, such as VEGF-B186, could occupy available binding sites on VEGF-R1 and therefore more VEGF-A could be available to bind to VEGF-R2 to induce angiogenesis.26,27 However, deficiency of VEGF-B does not decrease the activity of VEGF-A via VEGF-R2 even though more VEGF-R1-binding sites are available because of the absence of endogenous VEGF-B.28 It has been shown that angiogenic effects of VEGF-B186 are dependent on Nrp-1, which is abundantly expressed in cardiomyocytes in normal and ischaemic AdVEGF-B186-transduced pig myocardium, whereas in AdVEGF-A-transduced normal and ischaemic pig myocardium it is predominantly expressed in capillary endothelium.16
It has been shown in vitro that VEGF-E has a similar affinity for VEGF-R2 and that it activates similar intracellular signalling as VEGF-A165. VEGF-E has also been reported to have a similar mitotic activity on ECs as VEGF-A165, suggesting a similar angiogenic activity in vivo.17,29 However, we saw more modest but yet more physiological angiogenesis after the VEGF-E transduction as compared with that after VEGF-A165 injection. Angiogenesis induced by activation of VEGF-R2 by VEGF-E was comparable to that induced by activation of VEGF-R1 by VEGF-B186, which might suggest that these effects are due to VEGF-E's signalling through Nrp-1. This question needs to be addressed in future studies.
Cell proliferation induced by AdVEGF-A165 and AdVEGF-E was mostly seen in ECs, as expected due to the signalling via VEGF-R2, whereas AdVEGF-B186-induced proliferation was mostly seen in other cell types, such as cardiomyocytes. This difference in the proliferation profiles suggests that in the AdVEGF-B186 group the angiogenic effects were due to other factors, such as increased cardiomyocyte metabolism accompanied with increased need for perfusion in the proliferating myocardium. This suggests that AdVEGF-B186 activity may lead to capillary enlargement and increased blood flow at least partly via an increased metabolism of myocardial cells, a finding similar to that of Lahteenvuo et al.16 This finding may offer a new approach for the regeneration of damaged myocardium and warrants further investigation.
AdVEGF-B186 or AdVEGF-E transduction did not change the expression of VEGF-R1 compared with the intact heart or AdLacZ transduced heart. Interestingly, we saw a nine-fold increase in the VEGF-R1 protein expression after AdVEGF-A165 injection. VEGF-A has been reported to up-regulate the VEGF-R1 and soluble VEGF-R1 expression in vitro.30 Part of the VEGF-R1 up-regulation could be soluble VEGF-R1, which binds excess VEGF-A to reduce the potential negative effects. We did not see increased amount of VEGF-R1 in immunohistochemical stainings in the AdVEGF-A165 injected group compared with the AdLacZ group, which supports the theory of soluble VEGF-R1 up-regulation. VEGF-R1 staining was mostly seen in cardiomyocytes, whereas VEGF-R2 staining was seen especially in ECs. In the AdVEGF-A-treated group, VEGF-R2 expression was upregulated, whereas AdVEGF-B186 or AdVEGF-E did not induce any changes in the expression of VEGF-R1 or VEGF-R2.
We conclude that AdVEGF-B186 as a VEGF-R1 specific ligand and AdVEGF-E as a VEGF-R2 specific ligand are promising factors for the treatment of myocardial ischaemia in vivo. Simultaneous activation of both VEGF-R1 and VEGF-R2 by AdVEGF-A165, on the other hand, leads to aberrant vascular growth involving vascular lacunae and significant myocardial oedema. We found that AdVEGF-B186 and AdVEGF-E induced formation of enlarged capillaries that appear more physiological than those formed by AdVEGF-A. The functional benefits of these vessels formed by AdVEGF-B186 and AdVEGF-E have to be examined in more detail in myocardial ischaemia models. Interestingly, AdVEGF-B186 may also have a role in the regeneration of damaged myocardium.
This work was supported by Finnish Academy, Finnish Foundation for Cardiovascular Research, Sigrid Juselius Foundation, Aarne and Aili Turunen Foundation, and Clinigene EU Network of Excellence (LSHB-CT-2006-018933).
The authors wish to thank Sari Järveläinen, Tiina Koponen, and Seija Sahrio for technical assistance, and the staff at the National Laboratory Animal Center of Kuopio University for maintenance of the animals. The authors also wish to thank Med. Cand. Heidi Oksman for immunohistochemical assistance.
. Adenoviral catheter-mediated intramyocardial gene transfer using the mature form of vascular endothelial growth factor-D induces transmural angiogenesis in porcine heart. Circulation 2004;109:1029-1035.
. Increased vascularity detected by digital subtraction angiography after VEGF gene transfer to human lower limb artery: a randomized, placebo-controlled, double-blinded phase II study. Mol Ther 2002;6:127-133.
. Safety and feasibility of catheter-based local intracoronary vascular endothelial growth factor gene transfer in the prevention of postangioplasty and in-stent restenosis and in the treatment of chronic myocardial ischemia: phase II results of the Kuopio Angiogenesis Trial (KAT). Circulation 2003;107:2677-2683.
. A novel type of vascular endothelial growth factor, VEGF-E (NZ-7 VEGF), preferentially utilizes KDR/Flk-1 receptor and carries a potent mitotic activity without heparin-binding domain. J Biol Chem 1998;273:31273-31282.