Main

Vascular endothelial growth factor (VEGF) is a potent endothelial mitogen with angiogenic, vascular permeability, and vasoactive properties (1–3). VEGF exists as at least five isoforms (121–, 145–, 165–, 189–, and 206–amino acid isoforms) generated by alternate splicing of a single gene (1–3). VEGF165, a 46-kD protein, is the most potent of the isoforms and is abundantly expressed in the lung vascular smooth muscle cells, macrophages, and epithelial cells (4). It is critical to early vascular development because gene ablation of a single allele of VEGF in the mouse prevents normal vascular development and causes early embryonic lethality (5,6).

In the vascular system, two types of receptors are encountered: VEGF receptor 1 (VEGFR1; also called Flt-1) and VEGFR2 (also called Flk-1/KDR) both are tyrosine kinase receptors located on endothelial cells. VEGFR2 is initially synthesized as a 150-kD protein, rapidly glycosylated to a 200-kD intermediate form, and then further glycosylated, at a slower rate, to a 230-kD mature form expressed at cell surface (7). The overall processing steps are essentially the same for VEGFR1. Both receptors bind VEGF with high affinity, but VEGFR1 has the highest affinity for VEGF165, with a dissociation constant of ≈10–20 pM. VEGFR2 has a lower affinity with a dissociation constant of ≈75–125 pM (2,8) and is responsible for the dilator response to VEGF (9). In fact, at least part of the dilator response to VEGF has been ascribed to induction of nitric oxide (NO) release (10). Of interest, a soluble form of VEGFR1 also exists; it is secreted by endothelial cells and was shown to inhibit VEGF activity in vitro by sequestering the ligand (1,11–13).

At birth, with the initiation of ventilation, pulmonary vascular resistance decreases markedly (14). Persistent pulmonary hypertension of the newborn (PPHN) is a clinical syndrome characterized by abnormal pulmonary vascular tone, reactivity, and structure, leading to persistent elevation of pulmonary vascular resistance and severe hypoxemia at birth (15,16). In human newborns with PPHN, circulating levels of VEGF are reduced (17). In adult models of hypoxia-induced pulmonary hypertension using rats, it was shown that lung VEGF overexpression protects against pulmonary hypertension (9), whereas treatment with a VEGFR2 inhibitor (SU-5416) causes pulmonary hypertension and vascular remodeling (19). Recently, controversy has been raised by two studies done in fetal lambs using a model of increased pulmonary blood flow to induce pulmonary hypertension. In the whole lung, one study reported that VEGF expression is down-regulated (20), whereas the other noted that VEGF, VEGFR1, and VEGFR2 expression is up-regulated (21). Although, the exact role of VEGF and its receptors in the development of neonatal pulmonary hypertension remains obscured, Le Cras et al. (22) showed that inhibition of VEGF function by SU-5416 early in life causes pulmonary hypertension.

Knowing that maintenance of fetal levels of hypoxia postnatally results in intense vasoconstriction and vascular remodeling, preventing the decrease in vascular resistance that normally occurs at birth, we wanted to determine the contribution of the VEGF system in the development of neonatal hypoxia-induced pulmonary hypertension. We hypothesized that in the newborn undergoing hypoxic exposure, the vasoprotective effect of VEGF is blunted. Using the neonatal piglet model of hypoxia-induced pulmonary hypertension, we determined the effect of hypoxia on the ontogeny of pulmonary vascular expression/concentration of VEGF and its receptors, VEGFR1 and VEGFR2, and assessed the effect of hypoxia on the dilator response to VEGF. Furthermore, we assessed the expression of endothelial NO synthase (eNOS) and inducible NO synthase (iNOS) to determine whether we could establish a temporal link with the VEGF system. We report that pulmonary vascular levels of VEGF are increased by exposure to hypoxia, and these changes coincide with a decrease in VEGF-induced vasodilation. Furthermore, VEGFR1 and VEGFR2 expressions are transiently altered by hypoxia, possibly affecting the efficacy of VEGF.

METHODS

Animals.

Experiments were performed with 1-d-old piglets with the approval of the Animal Care Committee of McGill University. As previously described (23–26), piglets were maintained in fraction of inspired oxygen (Fio2) 0.10 ± 0.005 (hypoxia) or room air (control) for a period of 1, 3, 7, or 14 d in a 440-L Plexiglas chamber. Hypoxia was achieved by a continuous mixture of separate sources of air and nitrogen (Floxal; VitalAire, Montreal, PQ, Canada). Fio2 and fraction of inspired CO2 were measured at least three times per day with an electrochemical cell O2 analyzer (model S3-A/1; Ametek, Pittsburgh, PA) and sensor (model N-22M; Ametek) and an infrared CO2 analyzer (model CD-3A; Ametek) and sensor (model P-61B; Ametek). Fio2 was also monitored continuously with an Oxychek oximeter (Critikon, Tampa, FL). Fraction of inspired CO2 was kept at <0.005 via adjustment of total gas flow. Piglets were maintained in a thermoneutral environment, and ambient temperature was adjusted (26–36°C) according to age using a heater (model PCW-4; Noma, Scarborough, ON, Canada) with a temperature controller (model 689-0000; Barmant, Barrington, IL). Humidity level was maintained at <70% using a condensing coil as a dehumidifier. A 12-h/12-h light/dark cycle was established, and animals were fed ad libitum with balanced artificial milk (Wet Nurser; Jefo Import Export, St-Hyacinthe, PQ, Canada). Daily care of the animals was done without interruption of hypoxia. Control piglets were raised under identical conditions, except that Fio2 was kept at 0.21. Some normoxic animals were raised on the farm, as preliminary experiments did not reveal differences in measurements. All piglets received an intramuscular injection of iron in the immediate postnatal period. Animals from eight study groups were used either for isolation of arteries for peptide expression/concentration studies or for lung perfusion, in addition to cardiac weight measurements: group 1, 1-d normoxia (n = 10); group 2, 1-d hypoxia (n = 12); group 3, 3-d normoxia (n = 10); group 4, 3-d hypoxia (n = 11); group 5, 7-d normoxia (n = 6); group 6, 7-d hypoxia (n = 5); group 7, 14-d normoxia (n = 9); and group 8, 14-d hypoxia (n = 9). In animals that were not used for lung perfusion, the heart was removed and dissected into right ventricle (RV) and left ventricle plus septum (LV+S). The ratio of RV to LV+S, as an index of right ventricular hypertrophy, then was calculated. In 3- and 14-d-old animals, blood was also withdrawn from the left ventricle to measure circulating VEGF.

Drugs.

The Krebs solution used for the isolated lung experiments had the following composition (in mM): 119 NaCl, 4.0 KCl, 1.2 MgSO4, 1.2 KH2PO4, 2.5 CaCl2, 5.5 dextrose, and 25 NaHCO3. The solution was supplemented with 50 g/L albumin (BSA; Sigma Chemical Co. Chemical, St. Louis, MO). VEGF was provided by Dina Washington (Genentech, San Francisco, CA). The stable endoperoxide analog 9,11-dideoxy-9α,11α-methano-epoxy prostaglandin F2α (U-46619; Biomol Research Laboratories, Plymouth Meeting, PA), was used to raise the tone.

Tissue preparation.

As described previously (23–25), lungs were excised and immersed in cold (4°C) HEPES-buffered salt solution, which had the following composition (in mM): 20 HEPES, 135 NaCl, 2.68 KCl, 1.8 CaCl2, and 2.05 MgCl2. After exposure of animals to normoxia or hypoxia for 1, 3, 7, and 14 d, pulmonary arteries were dissected from the hilum down to 100-μm diameter. Arteries then were snap-frozen in liquid nitrogen and stored at −80°C to be processed as whole-artery homogenate. A portion of the artery homogenate was used as is, and the other portion was used for subcellular fractioning. Thus, measurements on cellular fractions correspond to measurements that were obtained in homogenate as they originate from the same pool of vessels.

For each preparation, lungs from several animals (n = 1–4) were used. Tissues were ground and homogenized at 0–4°C in sample buffer that contained 50 mM Tris-HCL, 250 mM sucrose, 5 mM EDTA, 10 mM EGTA, 10 mM benzamidine, 50 μg/mL phenylmethylsulfonyl fluoride, 10 μg/mL leupeptin, and 5 mM DTT, with a Polytron PT-3000 homogenizer (Brinkmann Instruments Canada, Rexdale, ON, Canada) at a speed of 13,500 rpm with six bursts of 20 s separated by 10-s cooling intervals. The homogenate was centrifuged at 16,000 × g for 30 min at 4°C. The supernatant then was filtered through a 100-μm mesh filter (Nytex, Zurich, Switzerland). This filtered fraction constituted the whole-artery homogenate.

For subcellular fractionation, the whole-artery homogenate was further centrifuged at 100,000 × g for 1 h at 4°C; the resulting supernatant constituted the cytosolic fraction, and the pellet represented the membrane fraction of cellular proteins. The microsomal pellet was resuspended in the sample buffer. Protein concentration was measured by the Bradford method. Preliminary experiments were conducted to determine whether specific proteins were better expressed in whole homogenate or in cytosolic or microsomal fraction. Therefore, protein contents of VEGFR2, eNOS, and iNOS were measured in whole-artery homogenate. VEGFR1 signal was markedly stronger in the cytosolic fraction compared with the whole homogenate. Therefore, VEGFR1 was measured in cytosolic and microsomal fractions. For comparison purposes, the endothelial cell marker platelet endothelial cell adhesion molecule (PECAM) and the smooth muscle cell marker α-actin were assessed. PECAM was measured in the microsomal fraction, and α-actin was measured in the whole-artery homogenate.

Western blot analysis.

On each gel, one sample of each condition and duration of exposure was loaded. Equal amounts of samples (15 μg of protein) were resolved by the appropriate reduced SDS-PAGE and electrotransferred to polyvinylidene fluoride membranes. Prestained molecular weight protein markers were also loaded onto each gel. Posttransfer staining of the gel by Coomassie blue confirmed the efficacy of the transfer. Nonspecific binding sites of the membrane were blocked by 2-h incubation at room temperature with 5% wt/vol skim milk powder in PBS (pH 7.4) with 0.1% Tween 20. Immunoblots were performed on the membrane-bound proteins at room temperature with incubation with rabbit polyclonal antibodies specific to VEGFR1 (1:500), VEGFR2 (1:200), eNOS (1:1000), iNOS (1:500), and PECAM (1:200; Santa Cruz Biotechnology, Santa Cruz, CA). A mouse MAb (1:25000, smooth muscle specific) was used to detect α-actin (Calbiochem, San Diego, CA). Goat anti-rabbit and a goat anti-mouse antibody (BioRad, Hercules, CA) were used as second antibody (1:7500). Signal detection was done using enhanced chemiluminescence (ECL Plus; Amersham Life Sciences) on Hyperfilm ECL (Amersham Biosciences, Inc., Quebec, Canada). Their corresponding scans (HP ScanJet 5P, Mississauga, ON, Canada) were quantified with Fujifilm Science Lab 97 Image Gauge Version 3.0 software (Tokyo, Japan). The following size bands were obtained: 130, 200, and 230 kD for VEGFR1; 150, 200, and 230 kD for VEGFR2; 135 kD for eNOS; 170 kD for iNOS; 130 kD for PECAM; and 43 kD for α-actin. Gels were also run in the presence of blocking peptides when available, i.e. VEGFR1, eNOS, and iNOS.

Measurement of VEGF concentration in plasma and cytosol.

After thoracotomy and before lung excision, blood was sampled from the left ventricle and placed in chilled tubes with EDTA. Blood was immediately centrifuged at 3000 × g, at 4°C for 20 min, and the plasma was stored at −70°C until analyzed. Plasma VEGF concentration was measured by quantitative sandwich enzyme immunoassay technique, using QuantiGlo human VEGF immunoassay kits from R&D Systems (Minneapolis, MN). For the measurement of VEGF in the cytosolic fraction of pulmonary arteries, the subcellular fraction was prepared as described above. Samples were analyzed in duplicate. Two antibodies, directed at different epitopes of the VEGF molecule, were used. An enhanced luminol/peroxide substrate solution was added, and the intensity of the light emitted was measured with a microplate luminometer (MicroLumat Plus LB96V; EG (Beid, Germany & Berthold, Wildbad, Germany). The intra-assay coefficient of variation for the kit is 3–7%. The sensitivity of the assay for VEGF is 0.78 pg/mL. The cross-reactivity with human VEGF165/PlGF heterodimer is 27%.

Isolated perfused lung.

Animals were anesthetized with sodium pentobarbital (65 mg/kg, i.p.), tracheostomized, and ventilated with 25% O2/5% CO2 in N2 with a tidal volume of 20 mL/kg. The isolated-perfused lung preparation was set up as previously described (23–26). Briefly, a cannula was placed in the pulmonary artery, and lungs were perfused with Krebs solution supplemented with 5% albumin at a constant flow of 30 mL · min−1 · kg body weight−1 using a peristaltic pump (Masterflex, model 7553-30; Cole-Palmer, Chicago, IL). Perfusate drained by gravity through a cannula placed in the left ventricle. Throughout the experiments, perfusion pressure, which reflects pulmonary artery inflow pressure, was monitored with a transducer (model P2310; Gould Instruments, Cleveland, OH) and a chart recorder (model 7E; Grass Instruments, Quincy, MA). Baseline perfusion pressure has been shown previously to be stable for ≥2.5 h. Integrity of the preparation was assessed by the stability of the baseline perfusion pressure and/or by the presence of edema as revealed by macroscopic examination.

Lung preparations were allowed to stabilize for 30 min after being mounted and perfused inside the chamber. After the stabilization period, for achieving comparable baseline pressure among all groups, vascular tone was raised as needed by adding to the perfusate the endoperoxide analog U-46619. The target pressure was ≈12 mm Hg. In hypoxic preparations, U-46619 was almost never required as tone was already at target level. VEGF then was added to the perfusion bath and was recirculated continuously through the preparation. An interval of at least 20 min was allowed between each concentration of VEGF tested in a cumulative manner.

Data analysis.

Data are presented as means ± SD. Statistical analysis of RV-to-LV+S ratio was performed using ANOVA, with p < 0.05 being considered significant. In normoxia, for standardizing the results between the different blots, protein expression at different ages was normalized against protein expression measured at day 1. In hypoxia, protein expression for each duration of exposure was expressed as a ratio of hypoxia over normoxia. For accounting for underrepresentation of endothelial cell proteins in the context of increased smooth muscle cell mass as a result of remodeling in hypoxia, expression in hypoxia over normoxia of VEGFR1, VEGFR2, eNOS, and iNOS was normalized against expression in hypoxia over normoxia of PECAM. Ratios were calculated for each pool of arteries. Expressions of VEGFR1, VEGFR2, eNOS, iNOS, PECAM, and α-actin as well as VEGF concentration were analyzed using two-tailed or one-tailed t test as appropriate, with p < 0.05 being considered significant.

The dilator response to VEGF was evaluated by measuring the lowest perfusion pressure reached for each concentration tested. The significance of the effect of VEGF on perfusion pressure at different concentrations was assessed statistically using ANOVA for repeated measures, with p < 0.05 being considered significant.

RESULTS

Physiologic measurements.

As expected, the RV/LV+S weight ratio decreased from day 3 to day 14 in normoxic piglets (p < 0.0001), whereas the ratio increased in hypoxic animals (p < 0.0001; Fig. 1). In fact, the RV/LV+S ratio was significantly increased after 3 (p < 0.005), 7 (p < 0.0001), and 14 d (p < 0.0001) of exposure to hypoxia compared with normoxia.

Figure 1
figure 1

Comparisons of RV/LV+S in normoxic (n = 25) and hypoxic piglets (n = 32). Values are means ± SD; *p < 0.005 vs normoxic group of corresponding age; †p < 0.0001 vs normoxic group of corresponding age; ‡p < 0.0001 vs 3-d group of corresponding condition.

Full size image

VEGF concentration.

Circulating VEGF was not affected by age or by exposure to hypoxia (Table 1). The cytosolic concentration of VEGF was not affected by age but was augmented by hypoxic exposure (Table 1).

Table 1 Concentrations of VEGF
Full size table

Endothelial cell and smooth muscle cell markers.

To assess whether our findings in hypoxic vessels were simply the result of changes in the proportion of endothelial cells versus smooth muscle cells induced by hypoxia, we evaluated the expression of the endothelial cell marker PECAM and the smooth muscle cell marker α-actin. PECAM expression was measured in the microsomal fraction, whereas α-actin expression was measured in the homogenate of piglet pulmonary arteries. PECAM and α-actin expression remained stable during the first 2 wk of life (Figs. 2 and 3). In hypoxia, PECAM expression was decreased as exposure was prolonged, reaching significance at 3 (p < 0.05), 7 (p < 0.05), and 14 d (p < 0.0005; Fig. 3). Further supporting that hypoxia affects specifically certain proteins is that the expression of α-actin was not affected by hypoxia (Fig. 3).

Figure 2
figure 2

PECAM expression in pulmonary artery homogenate of piglets that were exposed to normoxia or hypoxia for 1, 3, 7, or 14 d. Values are means ± SD; n = 4. (A) Percentage of day 1 for different ages in normoxia. No difference was observed across ages. (B) Ratio of hypoxia (H) over normoxia (N) for each duration of exposure. A value less than one represents a decrease in hypoxia in comparison with normoxia. *Expression significantly reduced in hypoxia (p < 0.05); †expression significantly reduced in hypoxia (p < 0.0005).

Full size image
Figure 3
figure 3

α-Actin expression in pulmonary artery homogenate of piglets that were exposed to normoxia or hypoxia for 1, 3, 7, or 14 d. Values are means ± SD; n = 4. (A) Percentage of day 1 for different ages in normoxia. No difference was observed across ages. (B) Ratio of hypoxia (H) over normoxia (N) for each duration of exposure. A value less than one represents a decrease in hypoxia in comparison with normoxia. Expression was not affected by hypoxia.

Full size image

VEGFR1 expression.

Three bands were identified at ≈130, 200, and 230 kD. The signal was remarkably stronger in cytosolic than in microsomal fraction. In fact, VEGFR1 expression in the microsomes was only 5% of the one measured in the cytosol, and for this reason, results of the microsomal fraction are not shown. In normoxia, the expression of the 230- and 200-kD bands increased with age (Fig. 4). However, the expression of the 130-kD band was decreased. In hypoxia, the normalized expression against PECAM of all bands was dramatically increased after most durations of exposure to hypoxia (Fig. 4).

Figure 4
figure 4

VEGFR1 expression in pulmonary artery cytosolic fraction of piglets that were exposed to normoxia or hypoxia for 1, 3, 7, and 14 d. Three different bands were identified at 230, 200, and 130 kD; values are means ± SD; n = 3. (A–C) Percentage of day 1 for different ages in normoxia. *Expression significantly altered compared with day 1 (p < 0.05); †expression significantly changed compared with day 1 (p < 0.005). (D–F) Ratio of the expression of hypoxia (H) over normoxia (N) over the expression of hypoxia over normoxia of PECAM for each duration of exposure. A value less than one represents a decrease in hypoxia in comparison with normoxia. ‡Expression significantly increased in hypoxia (p < 0.005); §expression significantly different in hypoxia (p < 0.05).

Full size image

VEGFR2 expression.

The expression of VEGFR2 was measured in piglet pulmonary artery homogenates. Three bands were also identified at ≈150, 200, and 230 kD (Fig. 5). In normoxia, the expression of these three bands was not developmentally changed. However, when animals were exposed to hypoxia, the normalized expression of the 230-kD band against PECAM was dramatically reduced at 14 d (p < 0.05). The normalized expression of the 200-kD band was not affected by hypoxia, but the expression of the 150-kD band was augmented at all ages (p < 0.05).

Figure 5
figure 5

VEGFR2 expression in pulmonary artery homogenate of piglets that were exposed to normoxia or hypoxia for 1, 3, 7, and 14 d. Three different bands were identified at 230, 200, and 150 kD; values are means ± SD; n = 4. (A–C) Percentage of day 1 for different ages in normoxia. (D–F) Ratio of the expression of hypoxia (H) over normoxia (N) over the expression of hypoxia over normoxia of PECAM for each duration of exposure. A value less than one represents a decrease in hypoxia in comparison with normoxia. *Expression significantly altered in hypoxia (p < 0.05); †expression significantly increased in hypoxia (p < 0.005).

Full size image

eNOS expression.

The expression of eNOS was measured in piglet pulmonary artery homogenates. In normoxia, eNOS expression was slightly reduced in 3- and 7-d animals (p < 0.05 and p < 0.01, respectively; Fig. 6). Hypoxic exposure did not affect the normalized eNOS expression.

Figure 6
figure 6

eNOS expression in pulmonary artery homogenate of piglets that were exposed to normoxia or hypoxia for 1, 3, 7, or 14 d. Values are means ± SD; n = 4. (A) Percentage of day 1 for different ages in normoxia. *Expression significantly reduced at day 3 (p < 0.05); †expression significantly decreased at day 7 (p < 0.005). (B) Ratio of the expression of hypoxia (H) over normoxia (N) over the expression of hypoxia over normoxia of PECAM for each duration of exposure. A value less than one represents a decrease in hypoxia in comparison with normoxia.

Full size image

iNOS expression.

The expression of iNOS was measured in piglet pulmonary artery homogenates. In normoxia, iNOS expression was not developmentally affected (Fig. 5). However, hypoxic exposure diminished normalized iNOS expression at 7 d (p < 0.05; Fig. 7).

Figure 7
figure 7

iNOS expression in pulmonary artery homogenate of piglets that were exposed to normoxia or hypoxia for 1, 3, 7, or 14 d. Values are means ± SD; n = 4. (A) Percentage of day 1 for different ages in normoxia. No difference was observed across ages. (B) Ratio of the expression of hypoxia (H) over normoxia (N) over the expression of hypoxia over normoxia of PECAM for each duration of exposure. A value less than one represents a decrease in hypoxia in comparison with normoxia. *Expression significantly reduced in hypoxia (p < 0.05).

Full size image

Pulmonary vascular response to VEGF.

Baseline perfusion pressure was increased in animals that were raised in hypoxia compared with those in normoxia (11.5 ± 1.4 versus 6.4 ± 1.4 mm Hg; p < 0.05). VEGF (10−11–10−9 M) caused a dose-dependent relaxation in normoxic piglets (Fig. 8). The dilator response started after 1 to 2 min of infusion, peaked at ≈5 min, and was partially or completely reversed at 20 min. When animals were exposed to hypoxia for 14 d, vasodilation to VEGF was completely abolished (p < 0.05; Fig. 8).

Figure 8
figure 8

Vasodilator response to VEGF in isolated perfused lungs of newborn piglets that were exposed to normoxia or hypoxia for 14 d. Responses are the lowest perfusion pressure reached for each concentration tested. Values are mean ± SD; n = 5. B, baseline perfusion pressure; RB, baseline perfusion pressure raised by the addition of the endoperoxide analog U-46619. *Baseline perfusion pressure is higher in hypoxic animals compared with normoxic animals (p < 0.05); †significantly reduced vasodilator response in hypoxia compared with normoxia (p < 0.05).

Full size image

DISCUSSION

This is the first report to suggest that in the neonatal piglet pulmonary vasculature, chronic hypoxia is associated with multiple disruptions in the integrity of the VEGF pathway. Whereas the cytosolic concentration of VEGF is increased in hypoxic pulmonary arteries, there is a reduction in VEGFR2 expression in the whole vascular bed, an increase in VEGFR1 expression, and an abolishment of the vasodilator response to VEGF.

As we have shown previously (23–26), the exposure to chronic hypoxia soon after birth prevents the decrease in pulmonary vascular resistance and induces significant pulmonary hypertension. The presence of an increased ratio of RV/LV+S weight after only 3 d of hypoxic exposure further demonstrates that these changes occur rapidly.

In the present study, we have found that chronic hypoxia is associated with unchanged circulating levels of VEGF in the neonatal piglet, whereas in the pulmonary vasculature, VEGF concentration is increased. Unlike us, Lassus et al. (17) reported decreased circulating plasma levels of VEGF in human infants with PPHN. Others have reported, as in our cytosolic measurements, that VEGF is augmented. For example, Mata-Greenwood et al. (21) noted that whole-lung VEGF expression was increased in a fetal lamb model of pulmonary hypertension with increased pulmonary blood flow. In addition, lung VEGF expression was found to be increased in the adult rat with hypoxia-induced pulmonary hypertension (27,28). Furthermore, VEGF was augmented in lungs of newborn infants with congenital diaphragmatic hernia and pulmonary hypertension (29). However, Grover et al. (20) recently described a marked reduction in whole-lung VEGF expression in a fetal lamb model of neonatal pulmonary hypertension with increased pulmonary blood flow. The conflicting results noted between this study and our experiment may represent a species-, a model-, a site-, or an antibody type–related difference.

We have shown that in normoxia, VEGFR1 expression increases with age. The reason for this increase in the postnatal period remains to be determined but may represent another way of regulating angiogenesis. Because of its higher binding affinity with VEGF, it has been suggested that VEGFR1 sequesters VEGF, thus preventing its binding to VEGFR2 (2,8). VEGFR1 transcription can be stimulated directly in endothelial cells in response to hypoxia via HIF-1α (1). In our model, VEGFR1 expression is rapidly induced by hypoxia, and this increase is real as it persisted after normalization against PECAM. Thus, the dilutional effect on endothelial cell proteins by the increase in smooth muscle cell proteins is accounted for by normalizing against PECAM. However, Mata-Greenwood et al. (21) reported a later rise in VEGFR1 in a different model of pulmonary hypertension. Whether this increased VEGFR1 expression is due to increased transcription or translation in our animal model remains to be determined. In our experiments, VEGFR1 expression was remarkably strong in cytosol and very weak in microsomes. However, the majority of VEGFR1 expression should lie within the membrane fraction (1). Contamination of the cytosolic fraction with microsomes could not explain our results, as the signal was weak in the microsomes. A soluble form of VEGFR1, lacking the seventh Ig-like domain, transmembrane sequence, and cytoplasmic domain, has been identified in human umbilical vein endothelial cells (12,13). It is possible that the higher molecular band of VEGFR1 found in the cytosolic fraction corresponds in fact to a complex formed with two soluble VEGFR1 and a VEGF dimer, as described by others (13,30). Similar complexes of 220 kD have been reported in human umbilical vein endothelial cells (12). Furthermore, Hornig et al. (31) reported that only 5–10% of the soluble receptor is detected as uncomplexed receptor protein. In pulmonary artery cytosolic fraction, three bands were identified for VEGFR1. As these bands all were specific, they most likely correspond to the immature, intermediate, and mature forms of the receptor (7).

Overall, most functional endothelial VEGF cell signaling described to date is mediated via VEGFR2 or strongly suspected to involve VEGFR2 (3). In hypoxia, the expression of the mature form of VEGFR2 in neonatal piglet pulmonary vasculature is reduced after prolonged exposure, contrasting with the report of Mata-Greenwood et al. (21). As with VEGFR1, the decrease in VEGFR2 was detected even after normalization for endothelial cell mass and protein loading with PECAM expression in the same lung sample. Worsening of pulmonary hypertension after VEGFR2 blockade in chronically hypoxic adult rats (19) or resulting pulmonary hypertension in rats that were treated as newborns with VEGFR blocker (22) supports our hypothesis that VEGFR2 may limit the action of VEGF.

As with VEGFR1, three specific bands were noted for VEGFR2, and they most likely correspond to the immature, intermediate, and mature forms of the receptor (7). The reduction in VEGFR2 expression, namely the 230-kD band, could be due to down-regulation induced by increased VEGF expression. In fact, VEGFR2 expression was shown to be regulated by VEGF (32–34). However, that the immature form of VEGFR2 was increased by hypoxia whereas the mature form was decreased suggests that a defect in receptor maturation may occur in hypoxia. As mentioned above, VEGFR2 is initially synthesized as a 150-kD protein, rapidly glycosylated to a 200-kD intermediate form, and then further glycosylated, at a slower rate, to a 230-kD mature form expressed at cell surface (7). Thus, as hypoxia can affect the expression of glucose transporters (35), it could potentially affect the glycosylation of VEGFR2.

In this study, we have shown that eNOS expression is not affected by hypoxia in the neonatal piglet. This differs from our previous report on eNOS activity (23). This suggests that enzymatic activity can be altered even if expression is unchanged.

We have shown that VEGF can induce significant vasodilation in the neonatal piglet pulmonary vasculature. Recently, we demonstrated that this vasodilator response in the neonatal piglet pulmonary vasculature is mainly due to the release of NO (36). This is in line with the report of Grover et al. (37) in which VEGF caused fetal pulmonary vasodilation by releasing NO through activation of phosphatidylinositol-3-kinase. The vasodilator response to VEGF was abolished in animals that were exposed to hypoxia. The lack of vasodilator response to VEGF is not due to reduced eNOS expression. However, in a previous report (25), eNOS activity was reduced as well as the response to NO, released by SIN-1, and to 8-bromo-cGMP. This suggests that the NO signaling cascade is also affected in the context of hypoxia. Thus, the lack of vasodilator response to VEGF may be due to multiple changes occurring at the receptor level down to the signaling cascade. However, vascular remodeling induced by hypoxia could limit the response to all vasodilators. The expression of iNOS was decreased by exposure to hypoxia. This is in line with our previous report on iNOS activity (23). In endothelial cells from porcine aorta, it was shown that VEGF can induce iNOS expression via VEGFR2 (38). Whether there is link between iNOS and VEGFR2 remains to be shown in the neonatal piglet pulmonary vasculature.

In our experiments, PECAM expression was decreased in hypoxia. This can be due to underrepresentation of endothelial cell proteins in the context of an increased smooth muscle cell mass. This is why normalization of protein expression against PECAM provides a way to ensure that results really reflect changes in expression and not in concentration. We cannot exclude that in our model of hypoxia-induced pulmonary hypertension, PECAM expression may also be affected by hypoxia in the context of cell proliferation (39). In hypoxia-induced pulmonary hypertension, smooth muscle proliferation occurs (26). However, α-actin expression was not affected by hypoxia. Others have reported that α-actin expression decreases in proliferating neonatal lung vascular cells after hypoxia (40) and that the actin phenotype changes from α- to β-actin during proliferation (41,42). This may explain the lack of change in α-actin expression in our model.

In conclusion, in neonatal hypoxia-induced pulmonary hypertension, VEGF is increased, whereas vasodilation to VEGF is abolished. This reduced vasodilator response may be due to decreased VEGFR2 expression. We speculate that sequestration by VEGFR1, to some extent, may limit the vascular protecting effect of VEGF, thus contributing to the pathophysiologic changes seen in neonatal hypoxia-induced pulmonary hypertension.