Differential Effects of Vascular Endothelial Growth Factor-C and Placental Growth Factor-1 on the Hydraulic Conductivity of Frog Mesenteric CapillariesHillman N.J.a · Whittles C.E.b · Pocock T.M.b · Williams B.a · Bates D.O.b
aCardiovascular Research Institute, University of Leicester, Leicester, bDepartment of Physiology, University of Bristol, Bristol, UK Corresponding Author
Vascular endothelial growth factors (VEGFs) are known to increase vascular permeability. VEGF-A acts on two receptor tyrosine kinases, VEGF receptor-1 (VEGF-R1 or flt-1) and VEGF receptor-2 (VEGF-R2, flk-1 or KDR). VEGF-C acts only on VEGF-R2 on vascular endothelial cells, whereas placental growth factor-1 (PlGF-1) acts only on VEGF-R1. The effects of perfusion of these receptor-specific proteins on hydraulic conductivity (Lp) was measured in frog mesenteric capillaries. The effect of PlGF on Lp was not conclusive, and overall fluid flux did not increase during that time. VEGF-C acutely and transiently increased Lp (4.5 ± 0.9-fold), which was more obvious in a subset of vessels, in a similar manner to that reported for VEGF-A. In the subset of vessels in which VEGF-C significantly increased Lp acutely, there was a sustained 12-fold increase in Lp 20 min after perfusion, but this was not seen in those vessels which did not respond acutely to VEGF-C, or in vessels exposed to PlGF-1. Lp was also increased 24 h after perfusion with VEGF-C, but not with PlGF-1. Western blot analysis showed that VEGF-R1 and VEGF-R2 are both present in frog tissue. These data show that the VEGFs that stimulate VEGF-R2 chronically increase Lp, but not those that stimulate VEGF-R1 only. This supports the hypothesis that chronic increases in microvascular permeability induced by VEGF are mediated via activation of VEGF-R2 rather than VEGF-R1.
Copyright © 2001 S. Karger AG, Basel
Vascular endothelial growth factors (VEGFs) are a family of endothelial cell-specific peptides that act to increase the delivery of nutrients to tissue by increasing the rate of angiogenesis, stimulating vasodilatation and increasing microvascular permeability . VEGF production is associated with a number of normal physiological processes including wound healing and the female reproductive cycle , and VEGFs have been implicated in the pathogenesis of a variety of vascular diseases including tumour growth , myocardial infarction , hypertension  and diabetes . VEGFs are currently being investigated as treatment for coronary and peripheral ischaemic disease . VEGF-A is the most common secreted form of VEGF and is currently being used in clinical trials for revascularisation . Previous studies have shown that the perfusion of microvessels in vivo with VEGF-A causes an acute, transient increase in microvascular hydraulic conductivity (Lp)  and diffusive permeability to albumin . This acute increase is followed by a chronic, sustained increase in Lp after 24 h, but no increase in mean pore size, since there is not a reduction in the oncotic reflection coefficient to albumin . Despite the undoubted importance of VEGF in a variety of pathologies, and the likely use of VEGF in clinical medicine in the future, very little is known about the mechanisms which underlie either the acute or chronic increase in permeability mediated by VEGF in microvessels in vivo.
VEGF-A is known to act principally via two receptor tyrosine kinases in vascular endothelial cells, i.e. VEGF receptor-1 (VEGF-R1 or flt-1) and VEGF receptor-2 (VEGF-R2, flk-1 or KDR) . The aim of this study was to determine the separate roles and importance of the two receptors in the initiation of both the acute and chronic permeability increases caused by VEGF. We have therefore used receptor-specific endogenous agonists. Placental growth factor-1 (PlGF-1) acts only on VEGF-R1 , and VEGF-C acts only on VEGF-R2 in vascular endothelial cells . We have studied the separate effects of these reagents on microvascular permeability. This work has been presented in abstract form to the Physiological Society [14, 15].
materials and methods
All experiments were carried out on adult male frogs (Rana temporaria) supplied by Blades, UK. All chemicals were purchased from Sigma unless otherwise specified. The frogs were anaesthetised by immersion in 0.1% 3-aminobenzoic acid ethyl ester (MS222) in tap water. The animal was laid supine and the limbs were lightly secured. A small incision (8–10 mm) was made in the right lateral skin and body wall, and the distal ileum was exteriorised and draped over a transparent quartz pillar 1 cm in diameter. The upper surface of the mesentery was kept continuously superfused with frog Ringer’s solution, pH 7.4 [in mM: 111 NaCl, 2.4 KCl, 1 MgSO4, 1.1 CaCl2, 0.2 NaHCO3, 2.63 N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic (HEPES) acid and 2.37 HEPES sodium salt], containing 0.02% MS222 at room temperature (20–22°C). The capillaries in the mesentery were visualised using a microscope and videoed through an electronic timer for off-line analysis.
The Lp of isolated mesenteric capillaries was measured using the Landis micro-occlusion method  modified as described previously . Baseline Lp was defined as the mean Lp measured during 3–5 micro-occlusions while perfusing with 1% bovine serum albumin (BSA; Sigma, A4378) in Ringer’s solution (adjusted to pH 7.4 with 0.115 mM NaOH) prior to perfusion with reagent. Microvessels were chosen that had divergent flow at the start of the vessel and convergent flow at the end (defined as capillaries). They were suitable if they had brisk flow with no white cells adhering or rolling along the wall, were at least 800 μm long and 15–25 μm in diameter with no side branches, and had a baseline Lp of <10 × 10–7 cm·s–1·cm H2O–1 . Vessels with a baseline Lp greater than 10 × 10–7 cm·s–1· cm H2O–1 were excluded. Each vessel was cannulated with a glass micropipette made from a pulled and bevelled glass capillary tube. The micropipette was filled with 1% BSA in frog Ringer’s solution with rat red blood cells as flow markers. Rat red cells were collected by direct cardiac puncture of 5% halothane-anaesthetised rats. The rats were then killed by cervical dislocation. The rat red blood cells were washed three times in frog Ringer’s solution before use. The filled micropipette was clamped in a holder that was connected to a water manometer to control the perfusion pressure. The capillary was perfused with the solution and occluded with a glass rod downstream of the cannulation site. The occlusion was maintained for 3–7 s, then the glass rod was removed and free flow was allowed for a few seconds before occluding again. The Lp of the capillary was determined from the filtration rate per unit area, calculated as described below, and the hydrostatic and oncotic pressures across the capillary wall, according to Starling’s hypothesis .
The rate of filtration per unit area of capillary wall (Jv/S) was calculated from the initial velocity of the flow marker red blood cells (dl/dt) after occlusion of the microvessel and the surface area of the vessel available for filtration (S). S was calculated from the internal radius of the capillary (r) and the length between the marker cell and the point of occlusion (l). Thus:
Jv/S = (dl/dt)(r/2l)
All of these measurements were made off-line from a videotape recording of the whole experiment. The Lp was calculated from the Starling equation:
Lp = (Jv/S)/ΔP
where ΔP is the effective hydrostatic and oncotic pressure difference between the capillary and the surrounding tissue. As 1% BSA has an effective oncotic pressure of 3.6 cm H2O, the pressure used to calculate Lp was 3.6 cm H2O lower than the pressure set in the manometer, assuming that tissue hydrostatic pressure was negligible and tissue oncotic pressure was equivalent to that in the superfusate (zero).
Following baseline Lp measurement in a single microvessel, the micropipette was refilled with Ringer’s solution containing 1% BSA, marker red cells and reagent. A refilling system based on that described by Neal  was used. The refilling system consisted of two syringes on an electric push-pull pump (SP120, WPI, Stevenage). The pushing syringe (5 ml) was connected to flexible polyethylene tube (PE50, Portex) that was connected to hand-drawn PE50 tubing glued into a male-male Luer connector. The drawn part was fed through the micropipette holder and up to the shoulder of the micropipette. The pulling syringe (20 ml) was connected to a similar tube that led through the micropipette holder to the back of the micropipette. Both syringes were attached to the tubes via taps that allowed the syringes to be closed off from the tubes and therefore from the micropipette (fig. 1). Before use, the tubes were filled with 1% BSA in frog Ringer’s solution, expelling all air bubbles, and were closed off from the syringes. After cannulation, the pipette was refilled with 1% BSA containing a low haematocrit (1–10%) of red blood cells. Following baseline Lp measurement, the pushing syringe was filled with a solution of 1% BSA in frog Ringer’s solution, with rat red blood cells and either 10 nM VEGF-C (a kind gift from Prof. K. Alitalo) or 10 nM PlGF-1 (a kind gift from Dr. M.G. Persico, or from R&D Systems). Both syringes were opened to the tubes and micropipette, and the pump was activated. The solution containing reagent therefore filled the front of the micropipette and the baseline solution was removed from the back of the micropipette. Switching of the solution was complete within 7 s (as determined both by filling with dye or by the change in haematocrit, pumping at maximum rate and switching 0.75 ml). As soon as the refilling was complete, the syringes were isolated from the micropipette by closing the taps to allow pressure equilibration, and the vessel was occluded as soon as possible to allow Lp measurement. In order to determine whether the system refilled without raising the pressure in the pipette, control experiments were carried out by refilling with 1% BSA. The filtration rate did not increase when measured at intervals during refilling. Lp was measured approximately every 15 s over the next 5 min.
Fig. 1. Diagram of the refilling system. The new agonist to be perfused is placed in a 5-ml infusion syringe (a). This is connected to a 20-ml withdrawal syringe (b), on a push-pull electronic pump (c). When the pipette is to be refilled, the pump drives 0.75 ml of solution through the filling line (d) into the tip of the pipette (e), and provides suction to the back of the pipette through the suction line (f). Once the solution has filled the pipette, the taps to the syringes (g, h) are closed, and the pressure equilibrates with that of the water manometer connected to the back of the pipette (i).
The effect on microvascular permeability 20 min after perfusion with reagent was determined in 13 vessels perfused with 10 nM VEGF-C and 8 vessels perfused with 10 nM PlGF-1. Following the Lp measurements described above, the micropipette was refilled with 1% BSA in frog Ringer’s solution with no rat blood cells and the vessel was perfused at 30 cm H2O for 20 min. The micropipette was then refilled with 1% BSA in frog Ringer’s solution with rat red blood cells and further Lp measurements were made.
The effect on microvascular permeability 24 h after perfusion with reagent was determined in 8 vessels perfused with 10 nM VEGF-C and 8 vessels perfused with 10 nM PlGF-1. Following the Lp measurements described above, a map of the mesentery was drawn in order that the same vessel could be located 24 h later. The frog gut mesentery consists of panels of connective tissue bordered by arteries and veins. The characteristic spacing of the arteries and veins allows an individual panel to be easily located from the map. All the visible microvessels in the connective tissue panel and the location of the arteries and veins crossing the mesentery were drawn. The location of the relevant panel of connective tissue in relation to the ileo-caecal junction was also noted and the cannulated vessel was marked on the map. The gut was then replaced in the body cavity and the body wall and skin were sutured. The animal was untied, allowed to recover whilst partially immersed in water, and then kept at room temperature. After 24 h, the frog was anaesthetised and the mesentery exposed as before. The same panel of connective tissue was found from the vascular architecture and its position relative to the ileo-caecal junction. The experimental capillary was located in the panel from the map drawn on day 1. The vessel was cannulated downstream of the cannulation site from the previous day, and blocked upstream of the block site from the previous day. Permeability was therefore measured on an undamaged section of the same vessel, which also formed the majority of the vessel investigated the previous day. Baseline Lp (perfusion with 1% BSA in frog Ringer’s solution with rat red blood cells) was measured as before. The criteria for acceptable Lp measurement on day 2 were similar to that of day 1, in that the capillary must have contained mobile red blood cells and no white blood cells adhering to the vessel wall. However, on day 2, there was no maximum Lp over which the vessel would be discarded, and the minimum available length of the capillary was only 400 μm.
To control for possible artefacts of anaesthetic, surgery or cannulation, Lp was measured in a group of vessels in exactly the same way as described above (15 vessels studied acutely, 5 after 20 min and 8 after 24 h). However, in each case, the pipette was always refilled with 1% BSA in frog Ringer’s solution with rat red blood cells, and never with either PlGF-1 or VEGF-C.
For analyses of the time course of the acute response to growth factors, the data were combined by grouping each measurement of Lp, or Lp relative to baseline, into 15-second intervals for each microvessel. The mean and standard error of the mean (SEM) were then calculated for each 15-second group. The peak increase in Lp reported in the results section is the mean of the individual maximum measured, not the peak value of the mean responses, and will therefore be greater than the time-averaged response seen in the figures. Relative changes (fold increases) in Lp were not normally distributed, and therefore non-parametric statistics (including median ± interquartile ranges) were used to analyse these changes. A value of p < 0.05 was considered significant.
In order to ensure that VEGF-R1 and VEGF-R2 were present in frog vessels, various frog tissues were processed for Western blot analysis with antibodies to human VEGF-R1 and VEGF-R2. Frog tissues were collected from animals pithed by cervical dislocation and destruction of the brain. Parasites were removed from the lung if present, and the tissue was then gently homogenised in 100 μl of phosphate-buffered saline (PBS) containing protease inhibitors: 0.2 μg·ml–1 leupeptin, 0.2 μg·ml–1 pepstatin A, 80 μg·ml–1 4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride and 0.1 mg/ml ethyldiaminotetraacetic acid, sodium salt. Protein was then extracted by resuspending the homogenate in an equal volume of extraction buffer (300 mM NaCl, 20 mM Tris, 10 mM EDTA) containing 2% Triton X-100, 0.2% sodium dodecyl sulphate (SDS), 2 mM sodium orthovanadate (Na3VO4) and 1 mM phenyl methyl sulphonyl fluoride. Homogenates were spun in a benchtop centrifuge for 10 min at 10,000 g at 4°C to separate the protein and cell debris, and 10-μl aliquots were removed to determine the protein concentration by the Bradford method (BioRad). For each sample, 105 μg (except the heart tissue for the VEGF-R2 blot, where 20 μg was used) of protein was used for SDS polyacrylamide gel electrophoresis. The gel was electroblotted onto a polyvinylidine difluoride membrane at 250 mA for 3 h. The membrane was washed in 1 × PBS-0.1% Tween (PBS-T) then blocked in 10% marvel in PBS-T for 2 h, and probed with a mouse anti-VEGF-R2 antibody (Santa Cruz, SC-6251) or a goat anti-VEGF-R1 antibody (Santa Cruz, SC316-G) at 1:1,000 (VEGF-R2) or 1:300 (VEGF-R1) in 5% marvel. The membrane was washed in PBS-T, incubated with HRP-conjugated anti-goat (Santa Cruz, 1:3,000) or anti-mouse (Dako, 1:1,000) secondary antibody and then washed 5 times for 5 min in PBS-T (anti-VEGF-R2) or 3 times for 1 h in PBS-T (anti-VEGF-R1). Protein was detected with enhanced chemiluminescence Western blotting detection reagent (Roche, UK) and exposed to photo-sensitive film for 1 min and developed. Bands present were measured against coloured markers (BioRad) also loaded onto the gel. Figures are representative of three gels run for each antibody.
Results from Western blot analysis showed that both VEGF-R1 and VEGF-R2 are present in frog tissues using antibodies to the human receptors. Bands were seen in the lung, kidney, liver and heart, which have previously been shown to be sites for expression in mammalian systems, and importantly, both were also present in the mesentery (fig. 2).
Fig. 2. Western blot of protein extracted from various frog tissues. Protein samples were equally loaded (105 μg), except for heart on the VEGF-R2 blot (20 μg). a Blot probed using antibody to VEGF-R1. Bands were seen at approximately 200 and 150 kD, consistent with glycosylated and pre-glycosylated mammalian forms in frog liver, kidney, heart, lung and mesentery. b Blot probed using antibody to VEGF-R2. Bands were seen at 150 kD, consistent with mammalian VEGF-R2 in frog heart, lung, kidney and mesentery, but not in liver.
Baseline Lp was defined as the Lp during perfusion with 1% BSA in frog Ringer’s solution prior to perfusion with reagent. For each vessel, 5 or 6 baseline readings were taken and averaged. This value was the baseline Lp for that vessel and was taken as time point zero. Vessels were then perfused with reagent and Lp measurements were taken over a 5-min period. Perfusion of 34 vessels with 10 nM VEGF-C caused an acute and transient median 2.5 ± 0.8-fold increase in Lp [from a baseline Lp (mean ± SEM) of 3.5 ± 0.5 to a peak of 13.0 ± 3.2 × 10–7 cm·s–1· cm H2O–1; p < 0.001, Wilcoxon paired test (WPT)]. Lp returned to baseline values during the 5-min perfusion with VEGF-C (to 5.1 ± 1.2 × 10–7 cm·s–1·cm H2O–1). For comparison of the time course of the response, Lp measurements during 5 min (300 s) of perfusion with VEGF-C for each vessel were grouped into 15-second bins and the mean Lp for each time point was calculated (fig. 3). Lp was significantly increased compared to the baseline value (point zero) during 12 of the first 13 15-second time points (up to 200 s), after which, the Lp was not significantly greater than the baseline value. If one chooses even stricter criteria for a response, such as those described in a previous report (i.e. the peak reading during perfusion with reagent is greater than 2 times the standard deviation of the mean baseline readings of that entire group of vessels ), which yields only 44% of the vessels being responders, these vessels behaved like the whole group. Since a heterogeneous population of vessels was being investigated, we separated the results from the vessels that showed a response (responders), from those that did not (non-responders). The time-averaged data for the two groups is shown in figure 3b. Again, during the 5-min perfusion period, there was a significant increase in Lp compared to baseline for the first 200 s for responding vessels, after which Lp returned to baseline values.
Fig. 3. The acute effect of VEGF-C on Lp (actual values). a Time-averaged Lp measurements (15-second bins over 5 min) from 34 vessels perfused with 10 nM VEGF-C (○) or 1% BSA (■). b Time-averaged Lp measurements (15-second bins over 5 min) of vessels categorised as to whether the maximal Lp reading during perfusion with reagent was more than 2 standard deviations greater than the mean baseline Lp [n = 15, responders (○)] or not [n = 19, non-responders (▲)].
The maximal value of Lp during perfusion of 19 vessels with 10 nM PlGF-1 was significantly greater than the baseline value (10.9 ± 3.0 compared to 3.4 ± 0.6 × 10–7 cm·s–1·cm H2O–1; p < 0.05, WPT). However, there were no consistent time points at which the maximum Lp measurement occurred, and the mean Lp at any particular time point after the start of perfusion was not significantly greater than the average baseline value. Neither was Lp significantly different from baseline values after 5 min (6.5 ± 2.3 × 10–7 cm·s–1·cm H2O–1). This can be seen statistically, since none of the time-averaged data (fig. 4a) showed a significant increase in Lp. To determine whether there was a significant increase in water flux during VEGF perfusion, the mean area under the permeability-time curve was calculated. (For each vessel, the mean baseline Lp for that vessel prior to perfusion with PlGF-1 was subtracted from each Lp reading during PlGF-1 perfusion and the results were plotted. The area under this curve was calculated and the mean of the group of vessels was taken.) Comparison of the mean area under the curve for PlGF-1 perfusion (all vessels: 742 ± 512 cm·cm H2O–1) with the mean area under the curve for 1% BSA perfusion (–102 ± 63 cm·cm H2O–1) over the same time interval showed that the increase in Lp during PlGF-1 perfusion was not statistically significant (p = 0.45, Mann-Whitney U test). Therefore, in these experiments, the effects of PlGF were inconclusive, in contrast to those of VEGF-C and VEGF-A.
Fig. 4. The acute effect of PlGF-1 on Lp (actual values). a Time-averaged Lp measurements (15-second bins over 5 min) from 19 vessels perfused with 10 nM PlGF-1 (○) or 1% BSA (■). b Time-averaged Lp measurements (15-second bins over 5 min) of vessels categorised as to whether the peak Lp reading during perfusion with reagent was more than 2 standard deviations greater than the mean baseline Lp [n = 5, responders (●)] or not [n = 14, non-responders (△)].
Perfusion with 1% BSA in frog Ringer’s solution with rat red blood cells and no reagent did not significantly alter microvascular permeability (15 vessels; increase from a baseline value of 2.3 ± 0.4 to a peak of 3.2 ± 0.5 × 10–7 cm·s–1·cm H2O–1; p = 0.19, WPT).
Lp was measured in 13 vessels 20 min after perfusion with VEGF-C and in 8 vessels 20 min after perfusion with PlGF-1. Lp was significantly increased in vessels perfused with VEGF-C 7.4 ± 2.9-fold after 20 min (from 3.4 ± 0.6 to 48.2 ± 20 × 10–7 cm·s–1·cm H2O–1; p < 0.05, WPT). In contrast, there was no significant increase in Lp in vessels perfused with PlGF-1 after 20 min (4.6 ± 0.7 compared to 5.8 ± 1.2 × 10–7 cm·s–1·cm H2O–1; p = 0.38, WPT) (fig. 5).
Fig. 5. The acute and chronic effect of growth factors (relative to baseline on day 1). VEGF-C (10 nM; n = 8, solid bars), PlGF-1 (10 nM; n = 8, grey bars) or 1% BSA alone (n = 8, open bars) were perfused into vessels and Lp was measured acutely, 20 min after perfusion and 24 h after perfusion. Values are medians ± interquartile ranges.
In order to determine whether the raised Lp 20 min after perfusion with 10 nM VEGF-C was linked to the initial response, responsive and unresponsive vessels were analysed separately. In responsive vessels (n = 7), the baseline value was 4.1 ± 1.0 × 10–7 cm·s–1·cm H2O–1, which increased 12.2 ± 5.9-fold after 20 min to 84.8 ± 31.4 × 10–7 cm·s–1·cm H2O–1 (p < 0.05, WPT), but in the unresponsive vessels (n = 6), there was a small and insignificant change from a baseline value of 2.6 ± 0.6 to a value of 5.4 ± 4.0 × 10–7 cm·s–1·cm H2O–1 (p = 0.51, WPT) after 20 min. A significant increase in Lp 20 min after perfusion with VEGF-C was therefore only seen in vessels that had responded acutely to VEGF-C (fig. 6). There was also no effect of 1% BSA on Lp after 20 min (n = 5; change from baseline value of 2.3 ± 0.8 to 10.5 ± 4.3 × 10–7 cm·s–1·cm H2O–1 after 20 min; p = 0.13, WPT). These results suggest that the increase in Lp after 20 min only occurs in vessels that show an acute Lp increase in response to VEGF-C, and that PlGF-1 does not increase Lp over a longer time period.
Fig. 6. The acute and 20-min increases in Lp are linked (values shown are relative to baseline). The median peak Lp is shown for vessels perfused with 10 nM VEGF-C (n = 13) or PlGF-1 (n = 8), measured during (grey bars) or 20 min after (solid bars) growth factor perfusion. Vessels in which Lp was acutely increased during VEGF-C perfusion had a significantly increased Lp 20 min later. This was not true for vessels that did not show acutely increased permeability, nor was it true for any vessels perfused with PlGF-1. Values are medians ± interquartile ranges. * p < 0.05.
Lp was significantly increased 7.6 ± 3.0-fold in 8 vessels 24 h after perfusion with VEGF-C, from 2.7 ± 1.1 to 19.6 ± 8.0 × 10–7 cm·s–1·cm H2O–1 (p < 0.05, WPT). In most of the vessels, no overt angiogenesis was seen, although in one vessel (data not included), a new side branch was clearly seen, which made Lp measurement impossible. In contrast, there was no increase in Lp in 8 vessels 24 h after perfusion with PlGF-1 (from a baseline value of 2.7 ± 1.0 to 5.2 ± 1.3 × 10–7 cm·s–1·cm H2O–1 24 h later; p = 0.11, WPT; fig. 5). Further, 1% BSA perfusion did not result in an increase in permeability the following day either (n = 8; 2.5 ± 0.6 to 3.7 ± 1.3 × 10–7 cm·s–1·cm H2O–1). This was consistent with previously published data on the chronic effect of perfusion of single vessels with 1% BSA . There was no relationship between the increase in Lp on the first day and the size of the Lp on the second day in any of the vessels perfused (with VEGF-C, PlGF-1 or 1% BSA; fig. 7). However, there was a significant relation between the acute and the 20-min Lp for VEGF-C, but not for PlGF-1 or 1% BSA, confirming that the acute increase in Lp is somehow linked to the more sustained increase.
Fig. 7. The relationship between the acute and chronic changes in permeability. a There was no relationship between the acute change in permeability and the Lp 24 h later for any of the groups studied. b There was a significant correlation (r = 0.769, p < 0.02) between the magnitude of the acute response and the response 20 min later in vessels perfused with VEGF-C (solid regression line). There was no relationship between the Lp during PlGF-1 (r = 0.181; dotted line) or BSA (r = –0.186; dashed line) perfusion and the Lp 20 min later.
The diameters of the vessels perfused with VEGF-C or PlGF-1 were measured on both days. Although there was no change in diameter in vessels perfused with PlGF-1, VEGF-C perfusion, in contrast, resulted in a significant increase in vessel diameter, from 20.5 ± 1.4 to 25.2 ± 2.0 μm (n = 8; p < 0.05, paired t test). This increase is similar to that described for VEGF-A  (fig. 8).
Fig. 8. VEGF-C but not PlGF-1 increased capillary diameter. The diameter of capillaries was measured before (grey bars) and 24 h after (solid bars) perfusion with 10 nM VEGF-C or PlGF-1. The diameter was significantly greater 24 h after VEGF-C perfusion (* p < 0.02, paired t test), but no different in PlGF-1-perfused vessels. There was no relationship between the diameter of the vessels and any of the permeability measurements.
VEGFs were originally isolated according to their ability to increase the extravasation of macromolecules , over 10 years ago. However, it was a further 6 years before the permeability-enhancing effects of VEGF were described in single perfused microvessels . Despite the importance of this family of growth factors in the generation of leaky vessels, pathological oedema and tumour-related effusions, the mechanisms through which VEGFs act have still not been clearly elucidated . What has been shown is that VEGF-A, the most common isoform of VEGF, can cause an acute, transient increase in Lp  and albumin permeability in vivo  that is followed by a chronic, sustained increase in Lp after 24 h . VEGF has been shown in frogs in vivo to acutely increase permeability through a pathway that is dependent on calcium influx  and independent of calcium store release . This pathway has been shown in pig coronary vessels to be dependent on nitric oxide formation . There is evidence that this may be through phospholipase C (PLC) activation [21, 22], and activation of protein kinase C. Although there is now some evidence that PLC is specifically activated by VEGF-R2 and not VEGF-R1 , it is not known which of the two VEGF receptors is responsible for one of the functional end points of VEGF stimulation – increased permeability. The results described here show that VEGF-C can cause an acute, transient increase in microvascular permeability, and can also cause the sustained, more chronic increase in Lp. VEGF-C has been known to stimulate only VEGF-R2 in previously studied tissues, whereas PlGF-1 can only stimulate VEGF-R1 . This suggests that the chronic increase was mediated via VEGF-R2 and not VEGF-R1. Although it is possible that the PlGF was not given at a high enough concentration to stimulate VEGF-R1, this is unlikely, since the affinity of VEGF-R1 is higher for PlGF (230 pM)  than that of VEGF-R2 for VEGF-C (410 pM) . In addition, we were using doses nearly two orders of magnitude higher than the affinities of the receptors for their ligands. However, further interpretation should be cautious in the absence of a dose response curve for both VEGF-C and PlGF. Since it has previously been shown that VEGF-R1 stimulation does not result in PLC=γ activation , but that VEGF-R2 activation does , then these data are consistent with the hypothesis that VEGF acts to increase permeability through VEGF-R2-mediated activation of PLC=γ . Some vessels studied failed to show a change in Lp in response to VEGF-C perfusion. There was no visible difference between these vessels and responsive vessels, and it was impossible to predict in advance whether vessels would show a response or not. It may be that unresponsive vessels do not express VEGF-R2, or that the receptor is present but inactive.
VEGF-A acts either as a homodimer or as a heterodimer with other members of the VEGF family. VEGF-A has two separate receptor-binding domains specific to VEGF-R1 or VEGF-R2. Due to the arrangement of these binding sites, the VEGF dimer can bind two receptors, either as homo- or heterodimers . PlGF has been shown to down-regulate the angiogenic effect of VEGF-A , and this may be due to stimulated VEGF-R1 attenuating the effect of stimulation of VEGF-R2. We have not studied the effects of heterodimerisation; we perfused capillaries with only one receptor-specific reagent at a time. However, our results with VEGF-C showed a chronic increase in Lp after 20 min, which is not seen with VEGF-A. There is also a more marked increase in Lp after 24 h with VEGF-C compared to previous results with VEGF-A . This suggests that perfusion of microvessels with VEGF-A may result in the stimulation of both receptors, and VEGF-R1 may attenuate the effect of VEGF-R2, resulting in a lesser effect on permeability than stimulation of VEGF-R2 alone.
It has previously been shown that perfusion of microvessels with VEGF-A increased the diameters of vessels after 24 h . In this study, this effect was also seen in vessels perfused with VEGF-C, but not in vessels perfused with PlGF-1. An increase in diameter has been shown to be dependent on mitogen-activated protein kinase (MAPK) activation , and both PlGF-1 and VEGF-C have been shown to activate MAPK [26, 30]. One hypothesis is that there was no VEGF-R1 expression on the mesenteric capillaries (unlikely since there was VEGF-R1 expression in the mesentery, which consists of connective tissue, mesothelial cells and blood vessels). Alternatively, it may be that the rest of the signalling pathway necessary to stimulate the increase in diameter is not stimulated by VEGF-R1.
In summary, our data show that VEGF-C, but not PlGF-1, increases the Lp of perfused mesenteric capillaries after 20 min and 24 h. Although VEGF-C clearly and consistently results in an acute increase in Lp lasting a few minutes, the acute effects of PlGF-1 were inconclusive. These data are consistent with the hypothesis that the previously described acute and chronic increases in Lp in frog mesenteric microvessels are stimulated by VEGF activation of VEGF-R2. In addition, the increase in diameter brought about by VEGF is also seen when VEGF-C is used to stimulate vessels, but not when PlGF-1 is used, and therefore these data are consistent with the hypothesis that the diameter increase is also brought about by VEGF-R2 activation.
This study was supported by the Wellcome Trust (No. 050742 to N.J.H.), the British Heart Foundation (PG97198 to T.M.P., FS98023 to D.O.B.) and the National Kidney Research Foundation (R35/1/98 to C.E.W.).
Dr. David Bates
Department of Physiology, The University of Bristol, The Vet School
Southwell Street, Bristol BS2 8EJ (UK)
Tel. +44 117 928 7823, Fax +44 117 925 4794
Received: Received: April 14, 2000
Accepted after revision: November 9, 2000
Number of Print Pages : 11
Number of Figures : 8, Number of Tables : 0, Number of References : 32
Journal of Vascular Research
Founded 1964 as Angiologica by M. Comèl and L. Laszt (1964–1973) continued as Blood Vessels by J.A. Bevan (1974–1991)
Vol. 38, No. 2, Year 2001 (Cover Date: March-April 2001)
Journal Editor: M.J. Mulvany, Aarhus
ISSN: 1018–1172 (print), 1423–0135 (Online)
For additional information: http://www.karger.com/journals/jvr