INTRODUCTION

TOP ABSTRACT INTRODUCTION BACKGROUND OF ANGIOGENESIS AND... MYCOPLASMA INFECTION MODEL OF... LEAKINESS OF REMODELED... THERAPEUTIC APPROACHES TO... CONTRASTING EFFECTS OF VEGF... REFERENCES

Keywords: angiogenesis; endothelial cells; microvasculature; Mycoplasma pulmonis; plasma leakage; vascular permeability; vascular remodeling

Angiogenesis and microvascular remodeling are elements of the tissue remodeling in chronic inflammatory diseases and tumors. Both types of change in the microvasculature result from endothelial cell proliferation and often occur together, but they represent different phenomena and responses to different stimuli. Angiogenesis is the growth of new blood vessels from existing ones, whereas microvascular remodeling involves structural alterationsusually enlargementof arterioles, capillaries or venules, without the formation of new vessels. As inflammatory or neoplastic diseases evolve, the microvasculature undergoes progressive changes in structure and function. Blood vessels enlarge or proliferate to supply nutrients to accumulations of inflammatory cells in chronically inflamed tissues or dividing cancer cells in enlarging tumors.

Changes in the microvasculature in chronic disease may be out of proportion to the increased metabolic needs of tissues because of the overproduction of growth factors that stimulate vessel growth and remodeling. Also, blood vessels in diseased tissues usually have multiple abnormalities, ranging from the expression of molecules not found on normal vessels to alterations in endothelial barrier function and leakiness. The strategy of using the abnormal vasculature of a diseased organ as a therapeutic target is now being used in promising efforts toward inhibiting angiogenesis in cancer, arthritis, and diabetic retinopathy. In addition, vessel leakiness is being exploited to enable tissue access of liposome delivery systems, viral vectors, or other therapeutic agents that do not readily cross the normal endothelium. Despite this progress, research on the pathophysiologic and therapeutic implications of angiogenesis and microvascular remodeling in airway disease is still at an early stage.

	BACKGROUND OF ANGIOGENESIS AND MICROVASCULAR REMODELING IN AIRWAY DISEASE

TOP ABSTRACT INTRODUCTION BACKGROUND OF ANGIOGENESIS AND... MYCOPLASMA INFECTION MODEL OF... LEAKINESS OF REMODELED... THERAPEUTIC APPROACHES TO... CONTRASTING EFFECTS OF VEGF... REFERENCES

The literature on angiogenesis and microvascular remodeling in human airway disease is relatively sparse, but there are clues that changes in the airway microvasculature have long been recognized as a feature of asthma. For example, it was recognized many years ago that the airway mucosa in fatal asthma is edematous and contains dilated, congested blood vessels (1). Early studies also showed that the airway wall of subjects with asthma is abnormally thick (4, 5), a feature that has been confirmed more recently by morphometric studies (6) and computed tomography (9). The increased wall thickness would necessitate an expansion of the microvasculature to accommodate the extra tissue mass, and the abundant, enlarged, and congested mucosal blood vessels contribute to the wall thickness. This vascular contribution is functionally important, because even modest increases in wall thickness can amplify decreases in airway conductance produced by bronchoconstriction (6, 12).

Some early studies of the pathology of asthma probably missed changes in the airway vasculature because the tiny mucosal blood vessels are inconspicuous in conventional histological sections. However, immunohistochemical methods using antibodies to vascular markers have made it much easier to visualize these vessels in bronchial biopsies and autopsy specimens, and changes in the airway microvasculature in human inflammatory respiratory disease are now better documented. The presence of angiogenesis in asthma and other airway diseases is being documented by an increasing number of studies (15). Furthermore, blood vessels previously described as enlarged, congested capillaries are now known to be a manifestation of microvascular remodeling instead of simple vasodilatation (22, 27). Nonetheless, the mechanism and therapeutic implications of alterations in airway blood vessels are just beginning to be elucidated, and changes in the microvasculature still represent an important gap in the understanding of the pathophysiology of asthma and other chronic inflammatory airway diseases.

We have sought to develop a better understanding of angiogenesis and vascular remodeling in chronic airway inflammation through the use of animal models. One objective has been to characterize changes in vascular architecture and endothelial cell phenotype in chronic airway inflammation. Another objective has been to learn the mechanism of leakiness of the new or remodeled blood vessels. A third objective has been to contrast the roles in angiogenesis and microvascular remodeling of two endothelial cell-specific growth factors, vascular endothelial growth factor (VEGF) and angiopoietin 1 (Ang1).

	MYCOPLASMA INFECTION MODEL OF CHRONIC AIRWAY INFLAMMATION

TOP ABSTRACT INTRODUCTION BACKGROUND OF ANGIOGENESIS AND... MYCOPLASMA INFECTION MODEL OF... LEAKINESS OF REMODELED... THERAPEUTIC APPROACHES TO... CONTRASTING EFFECTS OF VEGF... REFERENCES

Mycoplasma pulmonis infection in mice and rats (28, 29) has proven to be a useful model for studying angiogenesis and microvascular remodeling in chronic inflammatory airway disease (30). This infection in rodents has certain features in common with both asthma and chronic bronchitis in humans. Pathogen-free mice or rats are inoculated intranasally with M. pulmonis and then are maintained in a barrier facility to avoid other infections for days, weeks, or months as a chronic respiratory disease develops (28, 29). The microvasculature of the airway mucosa begins to change soon after infection, and angiogenesis and microvascular remodeling become long-lasting features of the disease (27, 31). In this model, growth factors and other cytokines produced by resident airway cells and inflammatory cells drive extensive remodeling of the airway wall, providing an opportunity to examine, step by step, the changes occurring during the development of chronic inflammation. Remodeling of the airway wall reflects changes in the microvasculature, inflammatory cell influx, epithelial thickening, mucous gland hypertrophy, and fibrosis of the airway wall (27, 28, 32, 33). However, the roles of specific types of cells, growth factors, and inflammatory mediators are still being elucidated.

The peak in endothelial cell proliferation in the airway mucosa occurs only 5 d after the onset of M. pulmonis infection (31), even though changes in the microvasculature continue to evolve and persist throughout the life of the animal. Angiogenesis is the dominant change in the microvasculature of the airway mucosa of rats (Figure 1A and 1B) (33, 34). Both microvascular remodeling and angiogenesis occur in mice, with the relative proportions of each being genetically controlled, in part because of strain-related differences in the immune response to M. pulmonis infection. Both remodeling and angiogenesis occur in the airways of C57BL/6 mice (Figure 1C and 1D), but microvascular remodeling is the main change in C3H mice (Figure 1D and 1E) (27).

View larger version (123K): [in this window] [in a new window]   Figure 1.   Comparison of the microvasculature in the tracheal mucosa of pathogen-free and M. pulmonis-infected rats (A-B, F-I ) and mice (C-E ). Tracheal whole mounts showing microvasculature stained with Lycopersicon esculentum lectin (A-G). (A) Pathogen-free rat: simple pattern of normal vasculature, with relatively straight capillaries (arrows) spanning cartilaginous ring (33). (B ) M. pulmonis-infected rat at 4 wk: abundant, tortuous, capillary-size angiogenic vessels (arrows), some of which are located in focal regions of lymphoid tissue (33). (C ) M. pulmonis- infected C57BL/6 mouse (8 wk) showing vascular remodeling consisting of new capillary-like vessels (angiogenesis, arrow) in addition to enlargement of existing vessels (27). (D) Pathogen-free mouse showing normal tracheal capillaries (arrow) (27). (E ) M. pulmonis-infected C3H mouse (8 wk) showing prominent microvascular enlargement (arrow) without new capillary-like vessels (27). (F, G) Differences in the microvasculature of tracheas in pathogen-free (F ) and M. pulmonis-infected (G) Wistar rats (42). Ricinus communis agglutinin I lectin stains the tracheal microvasculature uniformly in the pathogen-free rat (F ), but sites of lectin leakage, appearing as diffuse extravascular staining (arrows), are prominent in the infected rat (G). (H-I ) Distribution of leaky, Monastral blue- labeled blood vessels in rat tracheal mucosa after substance P (33). (H) Pathogen-free rat: Monastral blue labeling of normal postcapillary venules that became leaky after substance P (arrows). (I ) M. pulmonis- infected rat at 4 wk: Monastral blue labeling of numerous, tortuous angiogenic blood vessels (arrows) that are smaller and more heavily labeled than most normal postcapillary venules. Tissue stained by esterase histochemistry to demonstrate regions of abundant mucosal lymphoid tissue (brown, I ). Scale bar in I applies to all figures, A-B and H-I, 100 µm; C-E, 90 µm; F-G, 70 µm.

Functional changes in the microvasculature accompany the conspicuous morphological changes in blood vessel number, size, and architecture. Endothelial cells of the remodeled vessels upregulate P-selectin and support leukocyte adherence and migration (31, 35). The remodeled endothelial cells also avidly take up cationic liposomes and upregulate neurokinin 1 (NK1) receptors, leading to an unusual sensitivity to substance P not found in pathogen-free mice (36).

	LEAKINESS OF REMODELED MICROVASCULATURE IN CHRONIC AIRWAY INFLAMMATION

TOP ABSTRACT INTRODUCTION BACKGROUND OF ANGIOGENESIS AND... MYCOPLASMA INFECTION MODEL OF... LEAKINESS OF REMODELED... THERAPEUTIC APPROACHES TO... CONTRASTING EFFECTS OF VEGF... REFERENCES

Blood vessels in the airway mucosa after M. pulmonis infection might be expected to leak, because newly formed and remodeled blood vessels typically have abnormalities in endothelial barrier function (39). When we initially examined this issue in infected rats, no significant increase in baseline leakage was detected (30). By using a more sensitive approach, we found that the microvasculature of the infected tracheal mucosa is indeed leakier than normal under baseline conditions in the absence of other stimuli (42). Compared with the airways of pathogen-free rats, where the small amount of baseline leakage and the baseline clearance are in equilibrium, the airways of rats infected for 4 wk accumulate two to five times as much Evans blue tracer over a period of 30 min. The increased baseline leakage comes from newly formed and remodeled blood vessels, as shown by binding of extravasated Ricinus communis agglutinin I (RCA-I) lectin (Figure 1F and 1G), focal extravasation of the tracer Monastral blue, staining of endothelial cell borders with silver nitrate, and direct observation by scanning electron microscopy (42).

Several factors are likely to participate in the leakiness of remodeled blood vessels. An increase in endothelial permeability resulting from focal separations ~ 400 nm in diameter between endothelial cells is likely to be involved (42). The increase in the lumenal surface area created by angiogenesis and microvascular enlargement would add to the leakage. In addition, the enlargement of arterioles may lower upstream resistance and increase the transmural driving force for leakage. Impaired clearance of extravasated proteins via lymphatics could be another factor, although this has not been documented experimentally.

In normal vessels, histamine, bradykinin, substance P, and 5-hydroxytryptamine (5-HT) cause plasma leakage through the formation of focal gaps between endothelial cells (43). However, endothelial gaps may not be the only route for extravasation in remodeled vessels. Endothelial fenestrae, transcytotic vesicles, vesiculo-vacuolar organelles (VVOs), and monolayer defects all may contribute to increased plasma extravasation in pathological conditions accompanied by angiogenesis and microvascular remodeling (47). The size and other biophysical properties of each route across the endothelium would determine the size of molecules or particles affected, but the relative contributions of these factors have been difficult to quantify. The largest route for leakage, a pathway that can accommodate particles < 2 µm in diameter, results from defects in the endothelial monolayer of tumor vessels (41, 49), but similar defects have not been described in airway inflammation.

Angiogenic and remodeled vessels that form after M. pulmonis infection are abnormally sensitive to certain stimuli that evoke plasma leakage. In particular, substance P and the sensory nerve irritant capsaicin trigger an abnormally large amount of plasma leakage in the airways of infected rats (Figure 1H and 1I) (32, 42, 50). This vascular hyperreactivity is due to an increase in the expression of NK1 receptors on endothelial cells of the remodeled vessels (37). The hyperreactivity appears to be specific to substance P and other NK1 receptor agonists, as leakage produced by platelet-activating factor (PAF) and 5-HT is not exaggerated in the infected rats (42). Therefore, the remodeled microvasculature overresponds to substance P but apparently does not have a generalized hypersensitivity to inflammatory mediators.

	THERAPEUTIC APPROACHES TO DECREASING MICROVASCULAR REMODELING AND PLASMA LEAKAGE

TOP ABSTRACT INTRODUCTION BACKGROUND OF ANGIOGENESIS AND... MYCOPLASMA INFECTION MODEL OF... LEAKINESS OF REMODELED... THERAPEUTIC APPROACHES TO... CONTRASTING EFFECTS OF VEGF... REFERENCES

Many substances are known to cause plasma leakage by increasing vascular permeability but only a few have the opposite effect. ₂-Adrenergic receptor agonists constitute one class of agents that decrease plasma leakage. These agents are commonly used in the treatment of chronic airway disease and are known to inhibit plasma leakage evoked by a variety of stimuli including antigen, substance P, bradykinin, and PAF (51).

In the airways of rats with M. pulmonis infection, baseline leak is reduced by inhalation of a single nebulized dose of the ₂-agonist salmeterol over 10 min (42, 54). Dose-response studies have shown that this treatment significantly reduces baseline leakage in infected rats. The reduction occurs at the same dosage of salmeterol that abolishes ovalbumin-induced late-phase leakage and leukocyte adhesion in pathogen-free rats (54). A 5-h delay after treatment is necessary to achieve maximal inhibition because of the relatively slow onset but prolonged action of salmeterol (54). The highest dose used (5 mg/ml in the nebulizer), which completely eliminates the late-phase leakage after allergen, reduces baseline leakage in airways of infected rats by 60%. The inhibitory effect of salmeterol on plasma leakage is thought to be mediated by ₂-receptors because it is blocked by prior administration of the ₂-receptor antagonist ICI-118,551 (54).

The antileak action of ₂ agonists is probably related to the inhibition of endothelial gap formation (53). However, effects on airway smooth muscle cells, mast cells, epithelial cells, and nerves may contribute, considering the multiplicity of actions of these agents (55). For example, ₂ agonists can reduce the release of leak-producing inflammatory mediators from sensory nerves (56) and mast cells (57). The ₂-agonist salmeterol also may reduce angiogenesis and vascular remodeling in the airways of individuals with asthma (23).

Glucocorticoids can reverse the remodeling and the leakiness of the airway vasculature produced by M. pulmonis infection. Rats that develop chronic disease during the 6 wk after M. pulmonis infection and then are treated with dexamethasone for 4 wk have a normal airway microvasculature and a normal response to substance P (34). Treatment with oxytetracycline to reduce the number of infecting organisms has the same effect. In the absence of treatment, infected rats have extensive microvascular remodeling and as much as a 3-fold increase in substance P-induced plasma leakage (34). Similar results have been obtained in C3H mice infected with M. pulmonis for 4 d and then treated with dexamethasone for 10 d (35). These findings indicate that even severe structural remodeling and leakiness of the airway microvasculature are reversible in these animal models (34, 35). Related findings in subjects with asthma suggest that some reversal is also possible in human airway disease (23).

	CONTRASTING EFFECTS OF VEGF AND Ang1 AS ENDOTHELIAL CELL-SPECIFIC GROWTH FACTORS

TOP ABSTRACT INTRODUCTION BACKGROUND OF ANGIOGENESIS AND... MYCOPLASMA INFECTION MODEL OF... LEAKINESS OF REMODELED... THERAPEUTIC APPROACHES TO... CONTRASTING EFFECTS OF VEGF... REFERENCES

The endothelial cell-specific growth factors, VEGF and angiopoietin 1 (Ang1), have potent and clinically relevant actions on the microvasculature (58). Both of these growth factors play essential but separate roles in vascular development in the embryo. VEGF is key to the formation of the initial vascular plexus in early development. Indeed, in the absence of VEGF, the primitive vasculature does not develop normally and the embryo dies (59). VEGF expression increases in the airways of subjects with asthma and correlates with mucosal vascularity (24, 25). Ang1 is also essential for development of the vasculature but in a different way than VEGF. Mice lacking Ang1, or its tyrosine kinase receptor Tie2, die because primitive endothelial cell tubes do not evolve into mature vessels (60, 61). Ang1 appears to be essential for maturation of the vasculature from primitive tubes into a hierarchical network of vessels composed of endothelial cells and pericytes or smooth muscle cells. A second angiopoietin, angiopoietin 2 (Ang2), antagonizes the effects of Ang1 on Tie2 receptors and in some contexts acts as a natural inhibitor of Ang1 (62). Whereas Ang1 is widely expressed in normal adult tissues, Ang2 is expressed mainly at sites of vascular remodeling such as the ovary, placenta, uterus, and tumors (62). Ang2 expression in the presence of VEGF is accompanied by angiogenesis, but Ang2 expression in the absence of VEGF is associated with vascular regression (63).

In collaboration with G. Yancopoulos (Regeneron Pharmaceuticals, Tarrytown, NY), who discovered the angiopoietins, we examined some of the distinctive differences in the actions of VEGF and Ang1 on the adult microvasculature by using transgenic mice and adenovirus-transfected mice that overexpress these growth factors. The model of transgenic overexpression in the skin, using the keratin 14 (K14) promoter, made it possible to deliver VEGF, Ang1, or both in a sustained, tissue-specific manner. In the skin of transgenic mice that overexpress VEGF in basal keratinocytes (K14-VEGF mice), capillary-like blood vessels are unusually abundant and tortuous (Figure 2A and 2B) (65). There also is exaggerated leukocyte rolling and adhesion (66). By contrast, the skin of transgenic mice that overexpress Ang1 in the epidermis (K14-Ang1 mice) appears reddened because of enlarged dermal blood vessels, yet the number of vessels is about normal (Figure 2A and 2C) (65, 67). Although the enlarged vessels in K14-Ang1 mouse skin have the location of capillaries, the vessels resemble venules in size and expression of P-selectin and von Willebrand factor (vWF), which are not normally expressed by dermal capillaries in mice (27, 65). The skin of double transgenic mice, produced by breeding K14-Ang1 mice with K14-VEGF mice, overexpresses both Ang1 and VEGF (K14-Ang1/VEGF mice) and have enlarged venule-like vessels, similar to those in K14-Ang1 mice, as well as abundant capillary-like vessels, similar to those in K14-VEGF mice (Figure 2D) (65).

View larger version (100K): [in this window] [in a new window]   Figure 2.   (A-D) Comparison of ear skin vasculature, stained by perfusion of biotinylated Lycopersicon esculentum lectin, in wild-type mouse (A), K14-VEGF transgenic mouse (B), K14-Ang1 transgenic mouse (C ), and K14-Ang1/VEGF double transgenic mouse (D) (65). (E ). Comparison of Evans blue leakage, with or without mustard oil, in ear skin of same four types of mice as shown in (A-D) (65) (*p < 0.05 compared with baseline; p < 0.05 compared with wild type). (F ) Comparison of Evans blue leakage, with or without mustard oil, in controls and 6 d after systemic injection of Ad-Ang1 or Ad-Ang2 (70). (G-I ) Ricinus communis agglutinin I lectin-stained control venule in wild-type mouse (G, no mustard oil and no leaky sites) compared with leaky venule after mustard oil in wild-type mouse (H, abundant leaky sites, arrows) and with nonleaky venule after mustard oil in K14-Ang1 mouse (I, one leaky site, arrow) (65). ( J-L) Blood vessels (arrows) in ear skin are the same size in untreated wild-type mouse ( J ) and in wild-type mouse 5 d after Ad-Ang1 (K ). By comparison, skin blood vessels are markedly enlarged (arrow) in K14-Ang1 transgenic mouse (L) (65, 70). Scale bar in (L) applies to all figures: (A-D) and ( J-L), 40 µm; (G-I ), 15 µm.

Despite their relatively normal appearance, the unusually abundant dermal blood vessels in K14-VEGF mice are leaky under baseline conditions (65, 66), as might be expected because of the leak-producing action of VEGF (40). The leaky vessels are also abnormally sensitive to inflammatory stimuli as exemplified by their exaggerated leakage response to mustard oil (Figure 2E). Because of their venule-like features, the enlarged dermal vessels in K14-Ang1 mice seemed likely to be leaky and especially sensitive to inflammatory mediators. However, these vessels turned out to have normal baseline leakage and unusual resistance to leakage induced by mustard oil (Figure 2E) as well as by 5-HT, platelet-activating factor, and VEGF (65). Because of the contrasting actions of VEGF and Ang1 in these models, we determined whether the proleakage effect of VEGF or the antileakage effect of Ang1 dominates in the skin of K14-Ang1/VEGF double transgenic mice (65). We found that the antileakage phenotype of Ang1 dominates in these mice. Baseline plasma leakage in K14-Ang1/VEGF mice is in the normal range, and mustard oil-induced leakage is significantly less than that in K14-VEGF mice (Figure 2E) (65).

Studies of the mechanism of the antileakage effect of Ang1 are just beginning. One of the first questions to be answered is whether the decrease in leakage results from a reduction in the number or size of leaky sites in the endothelium or from a decrease in the hydrostatic driving force across the endothelium due to hemodynamic changes. We addressed this question by taking advantage of the observation that leaky sites expose focal regions of the endothelial basement membrane to the vessel lumen (45). We visualized these focal sites by perfusing through the vasculature a biotinylated lectin (RCA-I) that avidly binds to regions of exposed basement membrane (68). In untreated vessels, RCA-I faintly stains the lumenal surface of the endothelium but does not extravasate and does not stain the basement membrane (Figure 2G). However, after treatment of wild-type mice with mustard oil, RCA-I strongly stains focal regions of exposed basement membrane in venules in addition to the faint staining of the lumenal surface (Figure 2H). This intense focal staining is restricted to venules, as is the leakage in other models of acute inflammation (45). In contrast, in K14-Ang1 mice treated with mustard oil, RCA-I faintly stains the lumenal surface of vessels but rarely the basement membrane because of the lack of leaky sites (Figure 2I) (65). The presence of occasional focal staining of exposed endothelial basement membrane of venules in K14-Ang1 mice (Figure 2I) indicates that Ang1 reduces leakage and does not abolish the RCA-I-binding properties of the basement membrane. Arterioles and capillaries do not have indications of leakage in K14-Ang1 mice, as in wild-type mice (65). These results suggesting that the anti-leakage action of Ang1 is not due to hemodynamic changes are consistent with the leak-inhibiting effect of Ang1 on monolayer cultures of endothelial cells in vitro (69).

Collectively, these findings led to two unexpected conclusions: (1) Ang1 can block vessel leakiness produced by VEGF as well as by inflammatory mediators; and (2) new capillary-like vessels formed by VEGF overexpression need not be leaky, as they do not leak in a setting where Ang1 is coexpressed transgenically.

Because of these intriguing findings in transgenic mice, we asked whether Ang1 can reduce leakage in the adult without prolonged expression inherent in K14-Ang1 transgenic mice. For this purpose we used a model in which VEGF or Ang1 is produced systemically after intravenous injection of an adenoviral vector-growth factor gene construct that turns the liver into a bioreactor (70). This approach made it possible to deliver sustained high concentrations of Ang1 or VEGF throughout the body and to examine their effects on blood vessel leakiness. We used adenoviral vectors that expressed Ang1 (Ad-Ang1), VEGF (Ad-VEGF), or a control gene (Ad-green fluorescent protein).

Published data indicate that localized administration of Ad-VEGF is followed by the formation of bizarre blood vessels at the injection site (71). By comparison, our experiments revealed that systemic delivery of a large dose (> 2.5 × 10⁸ PFU) of Ad-VEGF causes lethal plasma leakage (70). Mice die with severe multiorgan vessel leakiness and edema within 2 d after injection of Ad-VEGF in the dose we used. The injection of the same dose of Ad-Ang1 has the opposite effect. Ad-Ang1-treated mice are healthy and the amount of plasma leakage in ear skin is in the normal range, but the leakage after mustard oil is significantly reduced (Figure 2F) (70). Administration of Ad-Ang1 also diminishes the leakage produced by local injection of VEGF. This protection against leakage is evident within 1 d after injection of Ad-Ang1 and lasts more than 10 d. The antileak effect of Ad-Ang1 is not accompanied by discernible enlargement of the skin microvasculature (Figure 2J and 2K), unlike the vascular enlargement found in the skin of K14-Ang1 transgenic mice (Figure 2L).

These adenoviral transfection experiments suggest that the antileakage effect of Ang1 occurs rapidly in mice with a wild-type genetic background and is independent of the vessel enlargement found in K14-Ang1 transgenic mice. Studies using Ad-Ang1 showed that the antileakage effect in the skin is not accompanied by enlargement of the skin microvasculature, thus separating the antileakage action from the remodeling action (70). However, enlarged dermal blood vessels are evident in the skin of K14-Ang1 transgenic mice (65), and our preliminary studies have shown signs of venular enlargement in the airway mucosa of mice treated with Ad-Ang1. As for the mechanism of Ang1-induced microvascular enlargement, endothelial cell proliferation presumably contributes, but the cell survival action of Ang1 may also be a factor (72, 73). The presence of prominent enlargement of venules in airway inflammation, as described above (Figure 1E) (27), adds to the interest in understanding the mechanism of Ang1-induced microvascular remodeling.

In conclusion, alterations in the microvasculature participate in the pathophysiology of chronic inflammatory airway disease. Vascular changes that are well documented in animal models have also been found in human asthma. Angiogenesis and microvascular remodeling not only result in more or larger blood vessels in the airway mucosa but also result in functionally abnormal blood vessels. Sustained leakage, leukocyte adhesion, and selective upregulation of NK1 receptors are examples of the abnormalities. However, these abnormalities and the vascular remodeling are potentially reversible by therapeutic intervention. VEGF and Ang1 may play complementary and coordinated roles in vascular remodeling, but they have opposite effects on blood vessel leakiness. The discovery of the antileakage action of Ang1 offers a possible new strategy for reducing airway edema in chronic inflammatory airway diseases such as bronchitis and asthma.