Up-regulation of soluble vascular endothelial growth factor receptor-1 prevents angiogenesis in hypertrophied myocardium

Abstract

Aims

Inadequate capillary growth in pressure-overload hypertrophy impairs myocardial perfusion and substrate delivery, contributing to progression to failure. Capillary growth is tightly regulated by angiogenesis growth factors like vascular endothelial growth factor (VEGF) and endogenous inhibitors such as the splice variant of VEGF receptor-1, sVEGFR-1. We hypothesized that inadequate expression of VEGF and up-regulation of VEGFR-1 and its soluble splice variant, sVEGFR-1, restrict capillary growth in pressure-overload hypertrophy.

Methods and results

Neonatal New Zealand White rabbits underwent aortic banding. mRNA (qRT–PCR) and protein levels (immunoblotting) were determined in hypertrophied and control myocardium (7/group) for total VEGF, VEGFR-1, sVEGFR-1, VEGFR-2, and phospho-VEGFR-1 and -R-2. Free VEGF was determined by enzyme-linked immunoassay (ELISA) in hypertrophied myocardium, controls, and hypertrophied hearts following inhibition of sVEGFR-1 with placental growth factor (PlGF). VEGFR-1 and sVEGFR-1 mRNA (seven-fold up-regulation, P = 0.001) and protein levels were significantly up-regulated in hypertrophied hearts vs. controls (VEGFR-1: 44 ± 8 vs. 23 ± 1, P = 0.031; sVEGFR-1: 71 ± 13 vs. 31 ± 3, P = 0.016). There was no change in VEGF and VEGFR-2 mRNA or protein levels in hypertrophied compared with controls hearts. A significant decline in free, unbound VEGF was found in hypertrophied myocardium which was reversed following inhibition of sVEGFR-1 with PlGF, which was accompanied by phosphorylation of VEGFR-1 and VEGFR-2.

Conclusion

Up-regulation of the soluble VEGFR-1 in pressure-loaded myocardium prevents capillary growth by trapping VEGF. Inhibition of sVEGFR-1 released sufficient VEGF to induce angiogenesis and preserved contractile function. These data suggest sVEGFR-1 as possible therapeutic targets to prevent heart failure.

1. Introduction

In response to persistent elevation in wall stress, due to sustained increase in workload on the ventricle, the heart attempts to compensate by remodelling. The process of remodelling relates to the progressive changes that occur in myocardial structure and function. During the initial phase of this dynamic process the ventricle compensates for the increased workload and maintains adequate contractile function, normal geometric shape and supernormal ventricular mass-to-volume ratio. If elevated afterload is unrelieved, decompensation follows with impaired contractile function and ventricular dilation. In response to mechanical stress, cardiomyocytes enlarge through addition of sarcomeres and progressively increase in size as hypertrophy develops. However, capillary density does not concomitantly increase as hypertrophy progresses and therefore the volume of tissue supplied by one capillary increases.¹ We have previously shown that the ability of endothelial and non-myocyte cells to respond to exogenous angiogenesis inducing growth factors such as vascular endothelial growth factor (VEGF) is preserved.¹ However, unopposed VEGF production has also been shown to result in continued new vessel formation, which as long as the inductive VEGF stimulus was present, resulted in an exaggerated angiogenic response to the extent that new vessels destroyed the normal organ architecture. This finding demonstrates the necessity of an effective negative feedback to limit VEGF-driven neovascularization.² Thus, angiogenesis is the result of a dynamic balance between angiogenesis stimulators and angiogenesis inhibitors.³

VEGF is a key regulator of angiogenesis in the myocardium.⁴ The biological activity of VEGF is mediated by two receptor tyrosine kinases, VEGFR-2 and VEGFR-1.^4,5 VEGFR-2 is the main angiogenesis inducing receptor, whereas the precise role of VEGFR-1 is less understood.⁵ The affinity of VEGFR-1 for VEGF is very high, with a K_d of about 2–10 pM, which is at least one order of magnitude higher than that of VEGFR-2.⁶ On the other hand, the tyrosine kinase activity of VEGFR-1 is relatively weak, and VEGF does not stimulate the proliferation of cells over-expressing VEGFR-1.⁷ VEGFR-1 functions as a ligand-binding receptor (decoy function) rather than a signal-transducing receptor⁸ but may participate in the regulation of the proliferative response of the endothelium to VEGF by forming heterodimers with VEGFR-2.^9,10 Another important feature of the VEGFR-1 gene is its ability to encode not only the mRNA for a full-length receptor but also a short mRNA for a soluble form of the VEGFR-1 protein which carries only the extracellular domain.^9,11,12 The soluble form of VEGFR-1 is likely to be a negative regulator of VEGF availability by binding and sequestering VEGF^11,13,14 or by forming non-signalling complexes with VEGFR-2.¹⁵ sVEGFR-1 is the main component to maintain the cornea avascular despite the ongoing production and release of VEGF.¹⁶

Placental growth factor (PlGF) and VEGF belong to the same gene family but PlGF is not expressed in healthy tissue, except the placenta. PlGF, which binds with high affinity to VEGFR-1/sVEGFR-1 but not to VEGFR-2, lacks direct mitogenic properties or the ability to stimulate tyrosine phosphorylation in endothelial cells.^17,18 It has previously been shown when PlGF saturated VEGFR-1/sVEGFR-1-binding sites, the activity of VEGF was augmented in vivo and in vitro, which makes it an ideal inhibitor for sVEGFR-1.¹⁷ On the basis of these observations, we hypothesized that impaired capillary growth in hypertrophied left ventricular (LV) myocardium is either the result of an inadequate compensatory pro-angiogenic stimulus due to lack of up-regulation of VEGF, or a consequence of increased expression of sVEGFR-1 resulting in reduced VEGF availability for binding to the angiogenesis inducing receptor, VEGFR-2. Furthermore, we sought to determine whether inhibition of sVEGFR-1 with PlGF releases VEGF, induces capillary growth, and thereby preserves contractile function.

2. Methods

2.1. Animal model of pressure-overload hypertrophy

LV pressure-overload hypertrophy was achieved by banding the descending aorta in 10-day-old New Zealand White rabbits (Millbrook Farms, Amherst, MA, USA) as we have previously described.^19–21 Animals are followed by weekly transthoracic echocardiography. LV mass/volume (M/V) ratio was used as an indicator for progression of hypertrophy and shortening fraction as measure of contractile function. Data collection and analysis were performed as previously described for this model by our group.^19–23 Tissue harvest was performed at 4 weeks of age (maximal hypertrophy) and at 7 weeks of age (onset of ventricular dilatation).²⁰ At the latter time point, hypertrophied hearts had already developed significant changes on a cellular level,^19,21,22 accompanied by extracellular remodelling²³ as we have previously shown. Age-matched controls (n = 7/group), untreated hypertrophied (n = 7/group), and treated hypertrophied animals (n = 7/group) were euthanized and LV myocardial tissue was harvested, frozen in liquid nitrogen and stored at −80°C for further analysis.

2.2. Real-time quantitative RT–PCR

mRNA expression levels of total VEGF, and VEGF receptors (VEGFR-2, VEGFR-1, and sVEGFR-1) were quantified by qRT–PCR. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as reference gene for normalization. In order to verify that GAPDH levels were not altered in hypertrophied myocardium, we also measured 18S as an additional internal control (Ambion, Austin, TX, USA). Total RNA was isolated from frozen LV myocardial tissue samples (150–200 mg) using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Purity of RNA was assessed by 260/280 ratio, and RNA integrity was determined by agarose gel electrophoresis. SYBR^® Green RT–PCR using a PTC-200 Peltier thermal cycler and Chromo 4 continuous fluorescence detector (MJ Research, Ramsey, MN, USA) were used. Sequence-specific oligonucleotide primers were purchased from Operon (Operon, Huntsville, AL, USA). Primer design was as follows: VEGF forward CCTTGCCTTGCTGCTCTACC and reverse AGGTTTGATCCGCATGATCTG; VEGFR-1 forward CTCAACGAGGCAGCATTACA and reverse CTGCTTGTGGAACTCATCCA; VEGFR-2 forward CAAGTGCATCCACAGAGACCTG and reverse GGAAAATATCTCCCAGAGCAACAC; and sVEGFR-1 forward GAACCTGCTCCTCAAGAACG and reverse CCTTTTTGTTGCAGTGCTCA. After initial transcription to cDNA and a denaturation step at 94°C for 10 min, a two-step cycle procedure was performed (denaturation at 94°C for 15 s, annealing and extension at 53.4°C for 30 s) for 40 cycles. Each sample was run in triplicates. A gradient including melting curve determined primer optimization. Fluorescence representing each gene was normalized to GAPDH mRNA and 18S.

2.3. Immunoblotting

Protein levels were determined by immunoblotting using standard protocols as we have previously reported.^19–23 Equal protein loading was confirmed by staining of gels with Coomassie brilliant blue (Bio-Rad, Hercules, CA, USA). For specific protein detection, the following primary antibodies were used: VEGF (Santa Cruz Biotechnology, Santa Cruz, CA, USA), VEGFR-2 (Millipore, Temecula, MA, USA), VEGFR-1 (Sigma-Aldrich, St. Louis, MO, USA), sVEGFR-1 (Zymed Laboratories, San Francisco, CA, USA), PlGF (R&D Systems, Minneapolis, MN, USA), phospho-VEGFR-1 (Millipore), and phospho-VEGFR-2 (Cell Signaling Technology Inc., Danvers, MA, USA). Horseradish peroxidase-conjugated antibodies were used as secondary antibodies (GE Healthcare, Piscataway, NJ, USA) followed by detection with enhanced chemiluminescence according to the manufacturer's instruction (GE Healthcare). After exposure on films, quantitative protein analysis was conducted using laser densitometry. Immunoblotting data are expressed as arbitrary densitometry units.

2.4. Determination of free, unbound VEGF by ELISA

In order to quantify whether the amount of free, soluble VEGF is decreased in hypertrophied hearts due to binding/trapping by sVEGFR-1, VEGF levels were measured by a sandwich enzyme-linked immunoassay (ELISA) (R&D Systems). The ELISA relies on the fact that it does not detect VEGF when it is bound to sVEGFR-1, which is supported by findings from other investigators.²⁴

Protein isolation was performed as described above. Samples of equal protein were loaded onto a microplate, pre-coated with a monoclonal antibody specific for VEGF. Any free, unbound VEGF present in the lysate was bound by the immobilized antibody and detected by incubation with an enzyme-linked polyclonal antibody specific for VEGF after addition of a substrate solution. The colour development was stopped and colour intensity was measured at 450 and 535 nm, respectively. In order to account for optical imperfections in the plate the readings at 535 nm were subtracted from readings at 450 nm. These results were plotted against a VEGF standard curve to determine the concentration of free VEGF.

2.5. Administration of PlGF to displace VEGF from sVEGFR-1

We used the concept known as ligand shifting.¹⁷ We treated hypertrophied hearts with PlGF, a ligand specifically binding to VEGFR-1 but not to VEGFR-2, following the same protocol of intrapericardial administration (2 µg/kg of rhPlGF protein from R&D Systems Inc.) as we had previously reported for this model.^1,22,23 rhPlGF protein was administered at 4 weeks of age (compensated hypertrophy) and again at 6 weeks of age (early failure). Tissue was harvested for further analysis within 24 h following administration (4 weeks old) and 1 week after the second administration of PlGF (7 weeks old). The second time point was determined based on previous observations that by this age more than one-third of untreated, hypertrophied animals had succumbed to the natural progression of the disease. We determined free, unbound VEGF by ELISA and release of VEGF from binding to sVEGFR-1 by co-immunoprecipitation. Briefly, immunoprecipitation was performed as we have previously reported with anti-sVEGFR-1 followed by immunoblotting using an anti-VEGF antibody.²¹

To exclude PlGF's additional effects on angiogenesis, we measured PlGF protein levels by immunoblotting in control, untreated hypertrophied and PlGF-treated hypertrophied hearts and an ELISA was used to determine VEGF/PlGF heterodimers (R&D Systems). VEGFR-1 phosphorylation was determined using an antibody directed at tyrosine residue 1213 and VEGFR-2 phosphorylation by immunoblotting.

2.6. Determination of capillary density

LV cross-sections were obtained from controls, hypertrophied and PlGF-treated hypertrophied hearts and fixed in 4% paraformaldehyde/phosphate buffered saline followed by paraffin embedding. The sections were de-paraffinized, cardiomyocytes were stained with desmin (Epitomics Inc., Burlingame, CA, USA) and a red fluorescent secondary antibody (Alexa-594 fluorophore; Invitrogen), capillaries with CD-31 (Dako Corporation, Carpinteria, CA, USA) and a green fluorescent secondary antibody (Alexa-488 fluorophore; Invitrogen), and nuclei with blue fluorescent 4′,6-diamidino-2-phenylindole (DAPI) nucleic acid stain (Invitrogen). Cover-slips were applied with fluorescent mounting medium (Dako Corporation) and slides were visualized using an Axiovert 35 Microscope (Carl Zeiss, Jena, Germany) with a Nikon 20x objective (NA 20x/0.75). Microvascular density was determined on 15 different, randomly selected fields of cross-sectioned fibres from each slide by a blinded microscopist. Within a calibrated graticule, all vessels identified by the computer software image analyser were counted using the MetaMorph Imaging System software (Universal Imaging Corporation, West Chester, PA, USA).

2.7. Animal care

All animals received humane care from the Animal Resources of Children's Hospital Boston and the investigation conforms to the ‘Guide for the Care and Use of Laboratory Animals’ prepared by the National Academy of Sciences and published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). The protocol was reviewed and approved by the Institutional Animal Care and Use Committee at Children's Hospital Boston.

2.8. Statistical analysis

Data were analysed using SPSS software package (version 16.0, SPSS Inc., Chicago, IL, USA) and are reported as mean ± standard error of the mean (SEM). A two-tailed unpaired student's t-test or analysis of variance with Bonferroni post hoc analysis where applicable were used for comparison between groups if normality was passed. A value of P ≤ 0.05 was considered statistically significant.

3. Results

3.1. mRNA and protein levels for VEGF

VEGF mRNA levels (relative expression ratio: 0.934) and protein levels by immunoblotting (control: 184 ± 1 vs. hypertrophy: 174 ± 4 arbitrary densitometry units; P = 0.063) were unchanged for hypertrophied hearts compared with control hearts. A representative immunoblot and data summary are shown in Figure 1.

Figure 1.

(A) A representative immunoblot for total VEGF and (B) a summary of densitometry data are depicted. Corresponding to protein levels, total VEGF mRNA is not different between control and hypertrophied myocardium (0.934-fold up-regulation; data not shown).

3.2. mRNA and protein levels for VEGFR-1, sVEGFR-1, and VEGFR-2

As indicated in Figure 2A, mRNA levels of VEGFR-1 and its alternative splice variant sVEGFR-1 were significantly up-regulated (expression ratio: VEGFR-1: 6.653 and sVEGFR-1: 6.984; P = 0.001) in hypertrophied myocardium. Corresponding to the mRNA results, immunoblotting analysis revealed that VEGFR-1 and sVEGFR-1 protein levels in hypertrophied rabbits were significantly increased in comparison with the control group (VEGFR-1: control 23 ± 1 vs. hypertrophy 44 ± 8 arbitrary densitometry units, P = 0.031 and sVEGFR-1: control 31 ± 1 vs. hypertrophy 71 ± 3 arbitrary densitometry units; P = 0.016). Figure 2B and C show representative immunoblots on top, and a summary of densitometry data below. The angiogenesis-promoting receptor VEGFR-2, however, was unchanged in hypertrophied vs. control myocardium on mRNA (expression ratio: −0.683) as indicated in Figure 2A, and protein level (control: 18 ± 2 vs. hypertrophy: 20 ± 4 arbitrary densitometry units; P = 0.662; Figure 2D).

Figure 2.

(A) The bar graph depicts the relative expression of mRNA of VEGFR-2, VEGFR-1, and sVEGFR-1 comparing age-matched controls with hypertrophied hearts. VEGFR-1 and sVEGFR-1 mRNA levels were significantly up-regulated (seven-fold) in hypertrophied hearts compared with age-matched control hearts (*P = 0.001) but VEGFR-2 remained unchanged. (B) A representative immunoblot on top and a summary of all densitometry data show that the full-length VEGFR-1 is significantly up-regulated in hypertrophied myocardium. *P = 0.031. (C) A representative immunblot of sVEGFR-1 and a bar graph depicting cumulated data show that sVEGFR-1 protein levels corresponding to the mRNA levels are significantly up-regulated in hypertrophied myocardium. *P = 0.016. (D) In contrast to VEGFR-1, VEGFR-2 was unchanged in hypertrophied vs. control hearts on mRNA and protein level.

3.3. Effects of PlGF on angiogenesis

In order to determine whether sVEGFR-1 was sequestering VEGF, we measured free, unbound VEGF protein levels in untreated hypertrophied, PlGF-treated hypertrophied hearts and age-matched controls. As indicated in Figure 3, acute as well as long-term application of PlGF resulted in higher levels of free, unbound VEGF. There is no endogenous VEGF production immediately following PlGF administration which was determined by qRT–PCR. This finding indicates that free, unbound VEGF protein detected in the PlGF-treated hypertrophied hearts is a consequence of release of VEGF from sVEGFR-1. We validated these findings by measuring VEGF not complexed to sVEGFR-1 by immunoprecipitation followed by immunoblotting. Figure 4 shows results following acute administration of PlGF. These data indicate that sVEGFR-1 was sequestering VEGF, rendering it unavailable for binding to VEGFR-2 to induce angiogenesis and PlGF reversed this inhibitory effect of sVEGFR-1 on VEGF. In addition, we found a significant up-regulation of VEGFR-2 phosphorylation in PlGF-treated hearts compared with almost undetectable levels in untreated hypertrophy (Figure 4B). This was a direct result of VEGF binding since we did not detect any PlGF protein in the hypertrophied myocardium within 24 h following treatment. We also ruled out that VEGF/PlGF heterodimers activated VEGFR-2 by measuring VEGF/PlGF heterodimers with an ELISA. There were no VEGF/PlGF heterodimers detectable in any of our samples (data not shown). Since PlGF can potentially stimulate angiogenesis through VEGFR-1, we also measured VEGFR-1 phosphorylation (Figure 4C). Phosphorylation of VEGFR-1 tyrosine 1213 was significantly up-regulated in PlGF-treated hypertrophied hearts compared with untreated hypertrophied hearts, which is a direct effect of VEGF binding to this receptor.

Figure 3.

(A) and (B) Free, soluble VEGF protein was measured by ELISA in tissue lysates obtained from untreated hypertrophied, PlGF-treated hypertrophied, and age-matched control hearts. VEGF protein levels were determined within 24 h following PlGF treatment (4 weeks of age) and at 7 weeks of age. At both time points, hypertrophied hearts showed decreased levels of free, soluble VEGF whereas inhibition of sVEGFR-1 by PlGF released VEGF (*P < 0.001 vs. hypertrophy). On mRNA level, VEGF was unchanged following PlGF treatment at 4 weeks but was significantly up-regulated at 7 weeks (6.48-fold) indicating a long-term effect on endogenous VEGF production (P = 0.008; data not shown).

Figure 4.

(A) Immunoprecipitation followed by immunoblotting confirmed that PlGF application releases VEGF from sVEGFR-1. Free, unbound VEGF is significantly higher in PlGF-treated hypertrophied hearts compared with untreated hypertrophied hearts (*P = 0.05). (B) PlGF released sufficient VEGF for binding to VEGFR-2 with subsequent activation whereas in untreated hypertrophy, VEGFR-2 phosphorylation was almost undetectable (*P = 0.001 vs. untreated hypertrophy). (C) Phosphorylation of VEGFR-1 on Tyr 1213 residue was detected in control hearts, hypertrophied and PlGF-treated hypertrophied hearts which is a direct consequence of VEGF binding. This is not PlGF mediated since PlGF induces phosphorylation of a different tyrosine residue. There was a significant increase in VEGFR-1 phosphorylation following PlGF treatment (*P = 0.03 vs. untreated hypertrophy).

3.4. Capillary density following PlGF

We have previously shown that increasing VEGF levels results in capillary growth and preserves contractile function despite the presence of increased pressure loading on the ventricle.^1,22,23 In the presence of PlGF, an inhibitor of sVEGFR-1, more unbound VEGF is available resulting in VEGFR-2 phosphorylation. Therefore, we determined capillary density in hypertrophied myocardium following PlGF application (Figure 5). There was a significant increase in the number of capillaries expressed per number of nuclei following inhibition of sVEGFR-1 with PlGF.

Figure 5.

(A) Representative immunohistochemistry sections are shown where capillaries are stained with CD-31 (green), cardiomyocytes with desmin (red), and nuclei with DAPI (blue). (B) Capillary density is expressed per total number of nuclei. Blocking of sVEGFR-1 resulted in capillary growth through release of VEGF. *P < 0.001 vs. control, ^#P < 0.001 vs. PlGF treated.

3.5. Echocardiography

LV M/V ratio was used as a measure for progression of hypertrophy. Initially, the LV responds by increasing muscle mass to maintain wall stress normal. After a peak at 4 weeks of age, the ventricle starts to dilate, indicated by a fall in M/V ratio in untreated hypertrophied hearts. In contrast, PlGF-treated hypertrophied hearts maintained hypertrophic growth consistently and continuously during the entire observation period (Figure 6A). Corresponding with the dilatation of the ventricle, contractile function (shortening fraction) deteriorates, which is prevented by PlGF administration to the hypertrophied myocardium (Figure 6B).

Figure 6.

(A) Animals were followed by weekly echocardiography. LV M/V ratio shows a significant rise at 4 weeks of age indicative of compensatory muscle growth followed by a steady decline, indicative of ventricular dilatation in untreated hypertrophied hearts. PlGF application prevented ventricular dilatation by maintaining the increase in LV muscle mass (*P < 0.05 vs. hypertrophy and ^#P < 0.05 vs. control). (B) Shortening fraction was used as indicator of contractile function. There was a significant decline in contractile function over time as hypertrophy progressed to failure in untreated hypertrophied hearts, whereas PlGF treatment preserved contractile function (*P < 0.05 vs. hypertrophy).

4. Discussion

The most important finding of this study is that VEGFR-1 and its soluble splice variant, sVEGFR-1, were significantly up-regulated in hypertrophied LV myocardium at the onset of failure whereas VEGFR-2 levels were unchanged. At the same time, total VEGF remained unchanged in hypertrophied myocardium and significantly lower levels of free, unbound VEGF were present in hypertrophied hearts, indicative of VEGF sequestration by sVEGFR-1 which resulted in lack of VEGFR-2 binding and downstream activation of pro-angiogenic signals. Blocking of sVEGFR-1 with PlGF increased myocardial levels of free VEGF resulting in capillary growth which maintained hypertrophic growth and preserved contractile function. Thus, we concluded that increased levels of sVEGFR-1 sequester VEGF and thereby restrict the amount available for VEGFR-2 to induce angiogenesis in pressure-overload hypertrophied myocardium.

Angiogenesis is a tightly regulated process that leads to the formation of new blood vessels by sprouting from pre-existing microvasculature, and is essential for normal mammalian development. VEGF is the most important inducer of angiogenesis in the heart. All major cell types in cardiac tissue including cardiac myocytes, fibroblasts, and endothelial cells, have been shown to produce VEGF. VEGF induces the proliferation and movement of endothelial cells, remodelling of the extracellular matrix and the formation of capillary tubules. VEGF exerts its cellular effects by binding to two high-affinity transmembrane tyrosine kinase receptors: VEGFR-1 and VEGFR-2.^4,25,26 VEGFR-2 is the main receptor involved in VEGF-stimulated angiogenesis.^5,9 During normal cardiac growth between childhood and young adulthood (physiologic hypertrophy), coronary microvascular growth parallels the degree of cardiomyocyte growth.²⁷ It can occur under physiologic conditions such as endurance training but is more likely in young rather than older animals.^28,29 In pathologic hypertrophy, this tight relationship appears to be lost. There is considerable controversy on age of onset of hypertrophy and impaired capillary density. It is generally accepted that adults with hypertrophy suffer from capillary rarefaction.³⁰ Studies in young animals, however, show discrepant results due to differences in age of onset of hypertrophy, method of pressure-overload hypertrophy, and gradual vs. acute increase in pressure loading.^27,31–34 Our model is unique in the sense that pressure loading is imposed almost immediately following birth and that aortic constriction is gradually achieved by growth of the animals, tightening the banding site which mimics the disease in humans with congenital heart defects affecting the LV outflow tract. In our model, cardiomyocytes enlarge with hypertrophy and a mismatch develops between the number of capillaries and cardiomyocytes per unit area, suggesting that the volume of tissue supplied by one capillary increases. Increased pressure loading on the ventricle produces a sustained abnormal increase in myocardial wall stress, which leads to progressive ventricular remodelling. The stress imposed on the cardiomyoyctes is a major stimulator for VEGF production and release.¹² One would therefore expect increased production and release of VEGF. As we have previously reported, hypertrophying myocardium suffers from early onset of cardiomyocyte loss which also potentially restricts the amount of VEGF released from the remaining hypertrophied cardiomyocytes.²² It is therefore reasonable to assume that mRNA levels of total VEGF are not up-regulated in the hypertrophied myocardium because of insufficient production of the VEGF isoforms by the remaining hypertrophied cardiomyocytes. At the same time, protein levels of VEGF were also unchanged, which indicates that neither newly produced VEGF nor matrix bound, thus stored VEGF, were available in sufficient amounts for binding to VEGF receptors to induce angiogenesis. As indicated by our data, VEGF levels in hypertrophied hearts were not sufficient to induce VEGFR-2 phosphorylation to the degree seen in age-matched control hearts.

VEGFR-1 null mutant mice die at embryonic stage due to an overgrowth and disorganization of blood vessels and not due to poor vascularization which indicates that VEGFR-1 is essential for the organization of embryonic vasculature but not for endothelial cell differentiation.³⁵ Events mediated through VEGFR-1 are mainly responsible for early stages of VEGF-induced angiogenesis, such as modulation of cell motility and adhesion of endothelial cells, whereas VEGFR-2 is more important in regulating endothelial cell proliferation.⁹ The affinity of VEGFR-1 for VEGF is very high, with a K_d of about 2–10 pM, which is at least one order of magnitude higher than that of VEGFR-2.⁶ On the other hand, the tyrosine kinase activity of VEGFR-1 is relatively weak, and VEGF does not stimulate the proliferation of cells over-expressing VEGFR-1.^7,9 Since VEGFR-1 and VEGFR-2 can form heterodimers, VEGFR-1 may participate in the regulation of the proliferative response of endothelium to VEGF.³⁶ This suggests VEGFR-1 to be both a negative regulator by sequestering (trapping) VEGF via its ligand-binding domain, and on the other hand, a positive regulator by its decoy function attracting VEGF or due to its mild tyrosine kinase signalling activity. VEGFR-1 is a 180 kDa transmembrane protein that is composed of seven extracellular immunoglobulin-like (Ig-like) domains, which include the binding site for VEGF in the second Ig-like domain, a single transmembrane domain and an intracellular tyrosine kinase region.^8,9 Another important aspect of VEGFR-1 is that the gene encodes not only the mRNA for a full-length receptor but also a short mRNA for a soluble form of the VEGFR-1 protein which carries only the extracellular domain, and has a unique 31 amino-acid C-terminus.^9,11 It was first reported in patients with pre-eclampsia who showed abnormally high levels of sVEGFR-1 associated with inadequate vascularization of the placenta.¹⁴ The soluble form of VEGFR-1 is likely to be a negative regulator of VEGF availability because it forms heterodimers with membrane-bound VEGFR-1 and VEGFR-2 thus acts as a receptor blocker of both VEGFR-1 and VEGFR-2¹⁵ and also because it sequesters VEGF, limiting the amount of free, soluble VEGF for induction of angiogenesis.

It has been shown that VEGFR-2 is the main mediator of VEGF-induced angiogenesis and releasing VEGF by blocking sVEGFR-1 may displace VEGF from sVEGFR-1 which makes it available for binding to VEGFR-2, a concept known as ligand shifting.¹⁷ Besides VEGF, there are other growth factors binding to VEGF receptors. PlGF is a disulphide-linked dimeric protein which belongs to the same gene family as VEGF.^37,38 VEGF exerts its biologic activities by interacting with both receptors, VEGFR-1 and VEGFR-2. PlGF on the other hand does not bind to VEGFR-2 but has a high affinity to VEGFR-1/sVEGFR-1, and therefore competes with VEGF for binding to those receptors. Our data are in support of the notion that PlGF's competitive binding to sVEGFR-1 releases VEGF trapped by this receptor. More free, unbound VEGF available in the hypertrophied myocardium leads to increased binding to VEGFR-2, indicated by phosphorylation of this receptor, the initial step in angiogenesis. Other potential mechanisms for PlGF to promote angiogenesis include that PlGF can transmit intracellular signals through VEGFR-1,^39,40 or in VEGFR-1 mediated transactivation of VEGFR-2,⁴⁰ or VEGF/PlGF heterodimers have also been described that can act through VEGFR-2.^40–42 We ruled out these effects of PlGF by measuring PlGF protein levels at the time point when we detected VEGFR-2 phosphorylation in PlGF-treated hypertrophied hearts. PlGF remained undetectable in the hypertrophied and non-hypertrophied myocardium, making it more convincing that VEGF released from sVEGFR-1 mediated this effect on angiogenesis. Also, VEGFR-2 phosphorylation was significantly lower in untreated hypertrophied hearts compared with age-matched controls directly corresponding to the lack of VEGF available for binding. Secondly, we determined the level of VEGF/PlGF heterodimer formation which remained undetectable. Thirdly, we excluded a direct effect of PlGF on VEGFR-1 by measuring VEGFR-1 phosphorylation on tyrosine 1213 residue. It has been shown that PlGF and VEGF have different effects on VEGFR-1 by stimulating autophosphorylation of different tyrosine residues in VEGFR-1.⁴⁰ Thus, we excluded all other possibilities of PlGF direct effects on angiogenesis and our results support the conclusion that PlGF displaces VEGF from sVEGFR-1, thereby increasing the fraction of VEGF available to activate VEGFR-2. As a long-term beneficial effect, PlGF application also resulted in endogenous production of VEGF in hypertrophied myocardium most likely as positive feedback to increased VEGF protein levels. PlGF is not an ideal inhibitor since it potentially can have direct effects on other cells in the myocardium. It requires further studies to determine these effects of PlGF which we are focusing on at the present time.

In conclusion, inhibition of capillary growth in hypertrophied myocardium is a result of inadequate VEGF protein levels due to lack of up-regulation of endogenous VEGF production and at the same time trapping of VEGF by the angiogenesis inhibitor sVEGFR-1. Inhibition of sVEGFR-1 released sufficient VEGF to induce angiogenesis and preserved contractile function. Therapeutic strategies aimed at preserving VEGF by inhibition of the soluble VEGFR-1 may be useful in maintaining capillary density and preventing heart failure.

Conflict of interest: none declared.

Funding

This work was supported by grants from National Heart, Lung, and Blood Institute HL-075430 (to I.F.) and HL-063095 (to P.J.d.N.). E.K. was supported by the Max Kade Foundation. I.F. work was supported in part by the Eleanor and Miles Shore Fellowship Program for Scholars in Medicine, Harvard Medical School and Children's Hospital Boston.