Increased Vascular Endothelial Growth Factor and Receptors: Relationship to Angiogenesis in Asthma

Asthma is a chronic inflammatory lung disease that is associated with airway wall remodeling. Remodeling of the microvasculature is probably an important contributor to increased inner airway wall thickness, lumenal narrowing, and especially the airway hyperresponsiveness in asthma (1-4). Vascular endothelial growth factor (VEGF) has been postulated to act as an angiogenic agent in altering subepithelial vascularity. Blood vessels respond to VEGF by sprouting angiogenesis (5) and increased levels of VEGF have been suggested to exist within asthmatic airways (6). Indeed, a central role for VEGF in asthma pathogenesis and pathophysiology has been proposed on the basis of sophisticated animal studies (7).

Several fibrogenic growth factors that are up-regulated in asthma also have angiogenic potential including transforming growth factor-ß, tumor necrosis factor-a, and basic fibroblast growth factor (8). VEGF is more specific and a powerful angiogenic factor that can also induce capillary hyperpermeability in vivo (9). Vascular dilatation can be induced by inflammatory mediators that are released in asthma, such as histamine (10), bradykinin (11), leukotrienes (12), platelet-activating factor (13), and substances released by autonomic nerves (14). Together, increased bronchial vascularization, vasodilatation, and increased permeability and leakiness can result in mucosal engorgement and edema (15, 16), thus contributing to airway inner wall thickening and subsequent airflow limitation, perhaps especially relevant in acute exacerbation.

VEGFR1 and VEGFR2 are the two main receptors for VEGF signaling in human airways (17). Both receptors are located on the vascular endothelium, but have divergent functions in vivo. VEGFR2 has been demonstrated to be the active receptor involved in the mediation of major growth and permeability actions of VEGF (17), whereas VEGFR1 has been postulated to act as a modulating decoy to VEGFR2, thereby inhibiting VEGFR2-VEGF binding (17). Interestingly, it is the decoy receptor VEGFR1 that has the higher affinity for the VEGF ligand (18), emphasizing its potential strategic importance.

After vessel maturation, angiopoietins begin to play important roles in the microvascular remodeling process. Angiopoietin-1 (Ang1) is important in maintaining the quiescence and stability of the mature vasculature. In the adult, disruption of Ang1 stabilization corresponds to the reinitiation of vascular remodeling under the influence of coincident VEGF (i.e., increased sprouting), as occurs in the estrous adult female reproductive system or in tumors (19). Although the importance of both the angiopoietins and VEGF synergistically affecting angiogenesis has been established (20), until more recently angiopoietins have remained largely unexplored in asthma (21).

Angiogenic sprouting is perhaps the predominant mechanism by which a vascular bed such as the airway vasculature is thought to grow (9, 22, 23). Immunohistochemical evidence from the present and previous studies (24) suggests that this may be true in the airways. The classic sprouting process involves endothelial cell migration, proliferation, and tube formation (18). VEGF contributes to this process by stimulating vascular splitting and sprouting. Embryos lacking a single VEGF allele have fewer vascular sprouts (25, 26).

Our objective was to assess VEGF, the VEGF receptors (VEGFR1 and VEGFR2), and Ang1 in airway biopsies and bronchoalveolar lavage (BAL) fluid from subjects with asthma and normal control subjects to determine whether VEGF is likely to be contributing to microvascular remodeling in asthmatic airways. Furthermore, we have undertaken some novel preliminary investigations into budlike sprouting structures that we have observed within the airway vessel walls and that may well be early indicators of angiogenesis.

METHODS

The study was approved by the institutional ethics committee, and all subjects gave written, informed consent.

Thirty-five patients with mild to moderate yet symptomatic atopic asthma were recruited. All fulfilled American Thoracic Society criteria for asthma, including proven reactivity to skin allergen prick tests, a positive PD^sub 20^ (provocative methacholine dose causing a 20% fall in FEV^sub 1^) of less than 2 mg of methacholine at screening, baseline FEV^sub 1^ at screening of at least 60% predicted FEV^sub 1^ and at least 1.5 L, and current symptom levels for active asthma, such that there was a need for ß-agonist therapy for at least 10 of 14 d. Thus, all patients had well-documented diagnosed asthma for at least 12 mo. Their mean (SD) age was 39 (13) yr (range, 20-70 yr). None were current smokers; 16 were ex-smokers, having given up a median of 16 yr (interquartile range, 6-22 yr) before recruitment. There was no evidence of smoking-related disease in any participant in the study. Subjects were using only inhaled albuterol for relief of symptoms and no "preventer" medication: no subject had received treatment with regular inhaled corticosteroids or inhaled long-acting ß^sub 2^-agonists for at least 3 mo before the study, 34 subjects had not received oral corticosteroid treatment for at least 12 mo before the study, and 1 subject had received a course of oral corticosteroid treatment 3 mo before the study. Patients who were receiving ß-receptor antagonists (ß-blockers), bronchodilator therapy other than inhaled ß-agonists, or who had received oral corticosteroids within 4 wk of the screening visit or maintenance therapy were also excluded, as were patients with clinical or laboratory evidence of serious uncontrolled systemic disease and those who had been hospitalized for any aspect of their asthma in the 4 wk before the screening visit.

Twenty-two normal healthy volunteers were recruited as control subjects and also underwent bronchoscopic examination and physiologic evaluation. None of the normal individuals had any history of respiratory illness; six were atopic. Their mean (SD) age was 37 (13) yr (range, 19-61 yr). All were nonsmokers; six were ex-smokers, having given up a median of 13 yr (interquartile range, 10-18 yr) before recruitment with no evidence of smoking-related disease. Subject demographics are summarized in Table 1.

After recruitment to the study there was a 2-wk run-in period during which the subjects with asthma were continued on albuterol alone as needed for symptomatic relief. After this, bronchoscopic examination was performed, with up to six airway biopsy specimens taken. All airway pathologic endpoints were read blind to any clinical information, with material randomized in the laboratory between normal subjects and subjects with asthma.

For bronchoscopic examination, subjects were premedicated with 0.4 mg of atropine (administered intravenously) and 5 to 15 mg of midazolam. Nebulized salbutamol (5 mg) was also given 15 to 30 min before bronchoscopic examination to both the subjects with asthma and normal subjects. Lignocaine (4%) was applied topically to the nose, pharynx, and larynx, and 2% lignocaine was administered below the cords in 2-ml aliquots as required up to a maximum dose of 6 ml. Subjects were monitored by pulse oximetry and administered oxygen during the procedure. BAL fluid was obtained from the right middle lobe of the lung, with the bronchoscope wedged into the subsegmental bronchus. Three aliquots (60 ml each) of phosphate-buffered saline were immediately aspirated under low pressure (100 cm H2O or less). Endobronchial biopsy specimens were taken from the subsegmental carinae of the right lower pulmonary lobe of each patient, using alligator forceps (FB-15C; Olympus, Tokyo, Japan). At each bronchoscopy, no more than six endobronchial biopsies were obtained.

For vessel staining and quantification (of vessels and sprouts), biopsy specimens were embedded into the hydrophilic polymer glycol methacrylate (GMA). Coded blocks were cut on a semithin resin microtome (Leica RM 2310; Leica Microsystems, Nussloch, Germany), using glass triangular knives. The 2-

For growth factor detection, biopsies were fixed in 10% formalin for 2 h and then embedded in paraffin wax. Blocks were cut on a Leica RM 2155 microtome (Leica Microsystems) at a thickness of 3 µm and again adhered to silanized glass slides. Extensive preliminary studies had informed us that GMA was suboptimal for these endpoints. Staining was performed with monoclonal antibodies for VEGF (R&D Systems, Minneapolis, MN), VEGFR1 and VEGFR2 (Santa Cruz Biotechnology, Santa Cruz, CA), and Ang1 (Jomar Diagnostics, Stepney, Australia), along with an ABC avidin-HRP kit (VECTASTAIN Elite ABC kit; Vector Laboratories, Burlingame, CA) for secondary antibody binding and color resolution. Total growth factor staining per unit area was determined, as well as area occupied by the growth factor (expressed as a percentage). The concentration of VEGF in BAL fluid was measured by QuantiGlo human VEGF immunoassay (R&D Systems).

What we have termed "angiogenic sprouts" were assessed by multicolor immunofluorescence, performed with monoclonal and polyclonal antibodies, respectively, to colocalize staining for collagen type IV and the endothelial marker von Willebrand factor. Sprouts were located in blood vessel walls as apparent cystic enclaves within the vessel walls. They were quantified as the number of sprouts per vessel and per square millimeter of subepithelial lamina propria.

In all cases, image analysis was performed with a Leica DM LB microscope (Leica Microsystems), Dage-MTI DC200 one-chip video camera (SciTech, Preston, Australia), and Image-Pro version 4.1 software (Media Cybernetics, Silver Spring, MD). Differences between normal subjects and subjects with asthma were analyzed by Student t test, performed either on the raw data (if parametric) or on log^sub 10^-transformed data (if nonparametric). All data were expressed as means ± SEM. Differences were considered significant at p < 0.05.

RESULTS

Airway Vascularity

The density of vessels was significantly greater in subjects with asthma than in control subjects (p = 0.05). The mean percentage of area occupied by vessels (percentage vascularity) was greater in subjects with asthma compared with control subjects; however, this was not significant (p = 0.15). Average vessel size (as described by vascular area [square micrometers]) was similar between groups, although there was a skew toward smaller vessels in asthma, with a significant increase (p < 0.001) in vessels less than 25 µm^sup 2^ in diameter (Figure 1).

Growth factors and receptors were analyzed both as absolute level of staining per area of lamina propria, and as staining of growth factor as a proportion of total stained vessels. This distinction is especially important in our examination of the relative amounts of growth factor staining, as absolute vessel numbers between subjects with asthma and control subjects were different and thus any changes in growth factor staining per unit area may be a reflection of this.

Quantitation of VEGF and Ang1

Subjects with asthma had significantly more VEGF-positive vessels compared with control subjects (Figure 2). VEGF-positive vessels per total vessels were not significantly different. VEGF in the BAL fluid samples of subjects with asthma was significantly elevated compared with control subjects (p < 0.05; Figure 3 and Table 2).

We found several significant correlations in the asthmatic group: VEGF staining significantly correlated with the total number of vessels per square millimeter (r = 0.69, p = 0.001) and percentage of vessels that were VEGF positive also correlated with vessels per square millimeter (r = 0.49, p < 0.02); that is, more vessels resulted in a greater proportion of VEGF staining. Percentage of vascularity (area of lamina propria occupied by vessels) also correlated both with VEGF-positive vessels per square millimeter (r = 0.58, p < 0.01; Figure 4A) and with percentage of vessels that were VEGF positive (r = 0.53, p < 0.05). No significant correlations were observed in normal subjects.

VEGF concentrations in BAL fluid in the asthmatic group were also significantly correlated with the total number of vessels per square millimeter (r = 0.59, p = 0.005; Figure 4B) and subepithelial vascular area (r = 0.39, p < 0.05).

Like VEGF, a greater number of Ang1-positive vessels per square millimeter was observed in subjects with asthma compared with control subjects (Figure 5). Ang1-positive vessels per stained vessels were not statistically significantly different between the subject groups (Table 2).

Quantification of VEGF Receptors

VEGFR1 expression (i.e., total vessels stained) was increased in subjects with asthma compared with control subjects; however, this difference was of borderline significance (p = 0.06). The number of vessels positively stained for VEGFR1 was not different between normal subjects and subjects with asthma. For VEGFR2, no significant differences were observed in total vessel staining or in VEGFR2 staining per vessel in asthma (p = 0.21). However, there was a suggestion that subjects with asthma could have had a lower relative quantity of VEGFR2-stained vessels than did normal subjects (Table 2). Indeed, the ratio of VEGFR1 to VEGFR2 staining was significantly higher in subjects with asthma (2:3) than in control subjects (1:3, ?^sup 2^ = 0.01).

In the asthmatic group, VEGF immunostaining significantly correlated with VEGFR2 immunostaining (r = 0.52, p < 0.01) but not with VEGFR1. In control subjects, there was no correlation between numbers of VEGF- and VEGFR2-positive vessels, but a significant correlation did exist between VEGF and VEGFR1 (r = 0.65, p < 0.01). There was a significant correlation between VEGFR1 and VEGFR2 immunostaining only in the asthmatic group (r = 0.72, p = 0.02; Figure 6), with a ratio of VEGFR1 to VEGFR2 of approximately 1:1.5.

Examination of Angiogenic "Sprouts"

Structures referred to as sprouts were observed within the endothelial basement membrane of preexisting vessels located within the lamina propria of bronchial biopsies from both normal subjects and subjects with asthma (Figure 7A). Immunofluorescence staining and confocal imaging (Figure 7B) were unable to show that these cystic structures were lined with endothelial cells, positively expressing von Willebrand factor (27).

The number of sprouts per square millimeter of lamina propria was significantly higher in subjects with asthma (log^sub 10^ no. sprouts/mm^sup 2^; 3.02 ± 0.04), compared with control subjects (2.77 ± 0.06; p < 0.005). Further, the number of sprouts per vessel was higher, but not quite significantly so, in subjects with asthma (4.19 ± 0.44), compared with control subjects (3.07 ± 0.39; p < 0.07).

In the asthma group only, there were significant correlations both between the number of sprouts per square millimeter and total vessels per square millimeter (r = 0.55, p < 0.005; Figure 8A), and also the number of sprouts per square millimeter and VEGF-positive vessels per square millimeter (r = 0.451, p < 0.05; Figure 8B).

DISCUSSION

Although it is known that asthmatic airways have increased vessel numbers in their subepithelial lamina propria, we have now shown that these increased numbers are associated with increases in both VEGF and Ang1 staining of vessels; furthermore, we have shown increased VEGF concentrations in the BAL fluid of subjects with asthma. That the increased levels of VEGF in the airway lumen are related to the extent of subepithelial vascularity supports a concept that there is a global increase in VEGF in asthmatic airways, and that VEGF may indeed be an active participant in the microvascular remodeling process.

The possibility that VEGF plays an active role in the microvascular remodeling process was recognized by Hoshino and coworkers (24), who also found that VEGF was up-regulated in asthma. Our observations are consistent with theirs, despite large variations in the actual data for VEGF-positive immunostaining between the two studies; this variation is most likely due to the method of quantification and analysis techniques. The changes in percentage vascularity and vessels per square millimeter in our study were consistent with the findings of some other groups who used GMA-embedding medium to quantify vessels (28, 29), although findings in paraffin-embedded tissue have been mixed (4, 30). A comparison between GMA- and paraffin-embedding techniques in canine arteries demonstrated that paraffin embedding caused considerable tissue shrinkage, and that GMA minimized volume changes of the tissue (31). We have also found that paraffin preparations do distort absolute vessel counts (data not shown). Although using GMA resin across the study would have been preferable, the increased rigidity of the resin is paid for in the loss of antigenicity; we found that growth factor staining was not possible in this medium and we were limited to performing growth factor staining in formalin-fixed, paraffin-embedded tissues.

Our observed change in distribution of vessel size, with subjects with asthma having a greater predominance of smaller vessels, is consistent with active angiogenesis. Although no significant differences were observed between normal subjects and subjects with asthma for any vessel size category, the distribution is significantly skewed to smaller vessels in asthma. In contrast, Li and Wilson (3) found significantly more vessels that were larger than 300 µm^sup 2^ in asthmatic airways, which may be suggestive of vascular dilatation via inflammatory mediators such as histamine and bradykinin, rather than angiogenesis.

Functional effects of increased VEGF within the asthmatic airway are likely to include the formation of abnormal, leaky vessels, resulting in pronounced tissue swelling (32-34). Increased vascular permeability would allow plasma proteins to leak into the extravascular space, causing potentially profound alterations in the extracellular matrix (35, 36). Further examination of vessel leakiness and vascular permeability (using markers such as a^sub 1^-macroglobulin) in asthmatic airways is warranted. VEGF has also been linked to modulation of the immune system and found to underlie many of the chronic changes described in asthma (7).

Although we did not observe any significant differences between subjects with asthma and control subjects in the direct comparison of staining of VEGF receptors, we believed that the near significance of this result (p = 0.06) was likely to represent a type 2 statistical error, and warranted some mechanistic discussion. Our data indicate that relative to the total number of vessels, VEGFR1 immunostaining was increased in asthma, whereas VEGFR2 tended to be decreased in asthmatic airways. The ratio between receptor staining was significantly different between subjects with asthma and normal subjects. This might suggest an abnormal balance of VEGF signaling activity in asthma. However, VEGF immunostaining significantly correlated with VEGFR2 in the airways of subjects with asthma, but with VEGFR1 in the airways of normal subjects. It is plausible, although somewhat speculative, that the VEGFR2 receptor is actively engaged in enhanced VEGF signaling and VEGF activity in asthma, possibly contributing to an increase in VEGF-induced microvascular remodeling. Moreover, subjects with asthma did possess relatively more VEGFR2-positive than VEGFR1-positive vessels (Figure 6). The increase in absolute amount of VEGFR2, along with an increase in the VEGFR1:VEGFR2 ratio, in subjects with asthma might suggest that VEGFR1 is serving as a braking mechanism to enhance VEGF-VEGFR2 activity.

In this study, we evaluated all positively stained blood vessels within the subepithelial lamina propria (excluding smooth muscle and submucosal glands). Others (37) have evaluated all "positively stained cells" located within the same area, and found much lower numbers, although as mentioned earlier, with relatively increased numbers in asthma consistent with our findings. This seems to be due to quantitation differences between studies, and it is not quite certain from their description what was being counted. In our study, we found that nonendothelial cellular staining was limited and inconsistent between immunohistochemical staining procedures, and thus we chose to emphasize VEGF-positive vessels and their potential contribution to the microvascular remodeling process in asthma.

To the best of our knowledge, this is the first study to examine Ang1 in the airways of subjects with asthma. Although Ang1-positive vessels per square millimeter were increased in asthma, Ang1 expression was not higher than control levels when expressed relative to the total number of vessels (as a percentage). We could not find a simple relationship between Ang1 and VEGF staining, but both were increased in asthma, which again suggests a complex process in which increases in VEGF stimulation of angiogenesis could be at least partly balanced by appropriate suppressive feedback systems. Further study of this should be done.

Vascular "sprouts" were present in the airways of both normal subjects and subjects with asthma, and thus were not specific to asthmatic disease processes. They were, however, increased in asthma. As far as we could tell, these structures were not endothelialized. A difficulty in interpreting these structures is that von Willebrand factor may not be present on developing endothelium, and CD31 or CD34 markers related to endothelial growth and implicated in tumor formation may be more applicable. However, at present, we are unable to stain for these in the GMA resin, whereas the sprouts themselves, although visible in paraffin-embedded tissue, are then more indistinct and difficult to define. An investigation of these structures in another rigid medium is warranted.

If these sprouting processes do indeed represent angiogenic mechanisms, our data would suggest that angiogenesis is likely to have an ongoing role in normal airways, contributing to the dynamic balance between growth and regression of the vascular bed. Elevated numbers of sprouts in subjects with asthma, compared with normal control subjects, even per vessel, may therefore be an index of the angiogenic modulation of microvascular remodeling within asthmatic airways. In addition, the increased number of sprouts per unit area of subepithelial lamina propria within the airways of subjects with asthma was demonstrated to be positively associated with the level of VEGF staining. This supports the notion that VEGF is stimulating the microvascular remodeling of existing vasculature in asthma.

There is the possibility that what we have termed vascular sprouts are indicative of the degradation or destabilization of the existing basement membrane of subepithelial blood vessels. This degradation of the basement membrane could occur via the action of matrix metalloproteinases, which are known to be required for the subsequent development of true angiogenic sprouts (38). It is interesting to note that VEGF also stimulates microvascular leakage (39). This allows hexose transport to meet increasing energy demands of tissue involved in growth and development (40) and tissue infiltration of plasma proteins (9, 41). These "sprouts" could, therefore, be more to do with transport mechanisms for fluid and nutrients than angiogenesis.

We conclude that increased numbers of blood vessels in the lamina propria of the airway wall in asthma are accompanied by increases in vessel staining of the angiogenic growth factors, VEGF and Ang1, as well as VEGF receptors and of VEGF in the BAL fluid. Relationships observed between VEGF and its receptors and vascularity suggest a complex, coordinated control feedback system, even within the remodeling process. Cystlike structures in the vessel walls may be angiogenic sprouts or at least structures directly relating to VEGF activity in the airways.

Conflict of Interest Statement: B.N.F. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. D.W. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. L.Z. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. C.W. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. D.R. received $5,000 of unrestricted research funding from Novartis Pharmaceuticals in 2001 and in 2002. R.H. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. E.H.W. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript.