SEARCH

SEARCH BY CITATION

Keywords:

  • respiratory tract;
  • blood vessels;
  • angiogenesis;
  • vascular regression;
  • hypoxia;
  • VEGF

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Recent advances have documented the development of lung vasculature before and after birth, but less is known of the growth and maturation of airway vasculature. We sought to determine whether airway vasculature changes during the perinatal period and when the typical adult pattern develops. On embryonic day 16.5 mouse tracheas had a primitive vascular plexus unlike the adult airway vasculature, but instead resembling the yolk sac vasculature. Soon after birth (P0), the primitive vascular plexus underwent abrupt and extensive remodeling. Blood vessels overlying tracheal cartilage rings regressed from P1 to P3 but regrew from P4 to P7 to form the hierarchical, segmented, ladder-like adult pattern. Hypoxia and HIF-1α were present in tracheal epithelium over vessels that survived but not where they regressed. These findings reveal the plasticity of airway vasculature after birth and show that these vessels can be used to elucidate factors that promote postnatal vascular remodeling and maturation. Developmental Dynamics 239:2354–2366, 2010. © 2010 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Early in embryonic development, mesodermal cells differentiate into endothelial cells that form the primitive plexus of small, anastomotic, poorly differentiated blood vessels through the process of vasculogenesis (Coffin and Poole,1988; Risau and Flamme,1995; Drake and Fleming,2000; Yancopoulos et al.,2000). Later in development, the vessels undergo pruning and remodeling to form a hierarchical network of arteries, capillaries, and veins, which expands through angiogenesis (Risau,1997; Gerber et al.,1999; Coultas et al.,2005; Thurston et al.,2005). The formation of the primitive vascular plexus has been the focus of multiple studies (Sato et al.,1995; Suri et al.,1996; Thurston et al.,2005), but less is known of how the plexus is remodeled into the organ-specific vasculature of the adult.

Understanding of the embryonic development of the lung vasculature has advanced with recent studies (Metzger and Krasnow,1999; Chinoy,2003; Cardoso and Lu,2006; White et al.,2007; Metzger et al.,2008). The pulmonary and bronchial vascular networks are thought to develop through varying contributions of vasculogenesis and angiogenesis (Burri,1984; Sparrow and Weichselbaum,2003; Stenmark and Gebb,2003; Anderson-Berry et al.,2005). The pulmonary circulation appears during branching morphogenesis in the pseudoglandular stage of lung development. As the part of the systemic circulation providing nutrients to the trachea and conducting airways, the bronchial vasculature appears during the cannalicular stage of lung formation, when branching of the bronchial tree is almost complete (Sparrow and Weichselbaum,2003).

The vasculature of the trachea and main bronchi of adult mice and rats has been used as a convenient model of the airway vasculature because of its accessibility, monolayer nature, and simple, segmented pattern, (McDonald,1994,2008; Baluk et al.,2005), but little is known about how this stereotypical architecture develops and whether it results from remodeling of a more primitive embryonic vascular plexus. A distinctive feature of the adult tracheal vasculature is the ladder-like pattern of capillaries that cross cartilage rings and the location of most arterioles and venules in tissue between cartilage rings (McDonald,1994; Thurston et al.,1998). The adult pattern of the tracheal vasculature is known to be present at late postnatal stages (Baffert et al.,2004; Thurston et al.,2005), but the vasculature continues to mature after birth, as indicated by age-dependent changes in sensitivity to angiogenesis inhibitors and factors that promote vascular remodeling (Baffert et al.,2004; Thurston et al.,2005).

We sought to determine how the vasculature of the airways develops and matures, by focusing on the transition from the embryonic period to the postnatal period. In particular, we asked whether a primitive vascular plexus develops initially and then is replaced by or remodeled into the distinctive adult vascular pattern. We also examined the mechanism of the regression and regrowth and whether hypoxia and vascular endothelial growth factor (VEGF) were involved.

By examining the architecture and phenotype of blood vessels in tracheas of mice from embryonic day (E) 17.5 through postnatal day (P) 70, we found that a primitive vascular plexus—unlike that of the adult—was present at E17.5 through P0. The primitive plexus regressed between P0 and P3 and then grew back to form the adult pattern by P7. VEGF was found to be a survival factor for airway blood vessels during this period.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Stereotypical Pattern of Tracheal Vasculature

The tracheal vasculature of adult mice was segmentally organized in relation to the cartilage rings (Fig. 1A,B). Arterioles and venules were located in the mucosa between cartilage rings, and most capillaries were located in the mucosa overlying the rings, where they had a ladder-like pattern with a rung spacing of approximately 85 μm (Fig. 1A). Capillaries had longitudinally oriented pericytes (Fig. 1A,D). Arterioles were relatively straight, narrow vessels, with elongated, spindle-shaped endothelial cells and circumferential smooth muscle cells (Fig. 1C). Venules were larger, had more branching, were more variable in caliber, and had irregularly shaped pericytes (Fig. 1D). Lymphatic vessels, with weak PECAM-1 immunoreactivity, were located between cartilage rings (Fig. 1A, asterisks).

thumbnail image

Figure 1. Vasculature of adult mouse trachea. Whole-mount preparation showing the vasculature of the adult trachea (endothelial cells, PECAM-1, green) with pericytes and smooth muscle cells (desmin, red) at postnatal day (P) 70. Long axis of the trachea oriented horizontally. A: Confocal micrograph showing ladder-like arrangement of capillaries (arrows) that span three cartilage rings. Most arterioles, venules, and weakly stained lymphatic vessels (*) are confined to regions between cartilages. Boxed regions are enlarged in C,D. B: Cartilages (arrows) of entire panel A are shown here in a brightfield image. C: Arteriole with densely packed circumferentially oriented smooth muscle cells and a weakly stained lymphatic (*). D: Venule and capillaries with obliquely or longitudinally oriented pericytes stained for desmin. Scale bar in D = 50 μm in A, 150 μm in B, 20 μm in C,D.

Download figure to PowerPoint

Development of Tracheal Vascular Pattern

The microvasculature of embryonic tracheas prepared at E16.5 (Fig. 2A,E,I) and E17.5 (Fig. 2B,F,J) consisted of a dense anastomotic, honeycomb-like plexus of primitive undifferentiated blood vessels (Fig. 2A,D,G) having a pattern substantially different from the segmented, hierarchical vasculature of the adult trachea. The embryonic plexus was denser rostrally than caudally, and lacked the hierarchical arrangement of arterioles and venules between cartilages and ladder like arrangement of capillaries over the cartilages typical of the adult. At E16.5 and E17.5, the regions of the vascular plexus over cartilage rings resembled the pattern between the rings much more than after P5 or in the adult, where the composition and pattern was entirely different (Fig. 1).

thumbnail image

Figure 2. Stages of pre- and postnatal development of tracheal vasculature. A–L: Confocal micrographs of tracheal whole-mounts with blood vessels stained for PECAM-1 at E16.5 (A,E,I), at E17.5 (B,F,J), P2 (C,G,K) and P5 (D,H,L). A,B: The primitive vascular plexus of undifferentiated blood vessels is highly anastomotic at embryonic day (E) 16.5 (A) and E17.5 (B). C: Vascular pruning stage: Segmentation of vasculature at postnatal day (P) 2 by regression of the primitive plexus over cartilage rings (*). D: Vascular regrowth and adult stage: Adult vascular pattern starts to become evident at P5, after growth of horizontal capillaries over cartilage rings, where almost no vessels are present at P2. Vertically oriented arterioles and venules are present between the rings. E–H: Panels A–D shown in inverted polarity grayscale to emphasize the conspicuous differences in the vasculature patterns. I–L: Higher magnification views of the vascular architecture at E16.5, E17.5, P2, and P5. Scale bar = 80 μm in A–C,E–G, 100 μm in D,H, 20 μm in I–L.

Download figure to PowerPoint

To determine the sequence of changes that transformed the embryonic plexus into the adult pattern, we examined tracheas of 10–20 mice on each day of development from E16.5 to P7 and then on 2- or 3-day intervals to P14, based on timed pregnancies. We found that the transformation occurred in three stages.

In the first stage, from E16.5 to P0, the tracheal vasculature consisted of a dense anastomotic plexus of blood vessels (Fig. 2A,B,E,F,I,J). The second stage, from P1 to P3, was characterized by pruning of certain blood vessels. In particular, blood vessels over cartilage rings underwent regression. At P2, few blood vessels were located in the mucosa over the rings (Fig. 2C,G,K). Staining with antibodies to different endothelial cell markers, including PECAM-1 (Fig. 2), VEGFR-2 (Fig. 3C,E,G), MECA-32, VE-cadherin, and endoglin/CD105 (data not shown), gave essentially the same segmented pattern, with few or no vessels over cartilage rings. Vascular pruning in the trachea was more extensive caudally than rostrally and was greater at the ends of the rings than at their midpoint in the ventral midline (Fig. 2C).

thumbnail image

Figure 3. Magnitude of changes in tracheal vasculature from embryonic day (E) 17.5 to postnatal day (P) 70. Confocal microscopic images. A: White box marking the region over a cartilage ring where area densities of PECAM-1–immunoreactive blood vessels were measured. B: Area densities of PECAM-1 immunoreactivity expressed as percentage of total area. Values are highest from E17.5 to P0, decrease significantly from P0 to P1, remain low through P3, increase significantly at P4, and plateau to an intermediate value at P5 typical of the adult. C: Examples of capillaries (arrows) crossing over a cartilage ring. D: Capillaries, expressed per millimeter of ring length, counted at the midline of each cartilage ring. Age-related changes in crossing capillaries follow a pattern similar to those of area densities of PECAM-1–positive blood vessels. E: Examples of capillary loops (asterisks) over a cartilage ring. F: Capillary loops, expressed per mm2 of cartilage ring, are numerous from E17.5 through P0 and decrease significantly between P0 and P1. G: Examples of vascular sprouts or projections (arrows). H: Endothelial projections, expressed per mm2 of cartilage ring, are abundant from E17.5 through P5 and then gradually decrease to the low level found in the adult. No abrupt changes occurred between E17.5 and P5. PECAM-1 (green); VEGFR-2 (red). Scale bar = 50 μm in A,C; 30 μm in E,G. Asterisks in graphs mark values significantly different from the adult (P < 0.05. Cross (†) in 3D marks values that were significantly different from the preceding values.

Download figure to PowerPoint

The third stage, beginning at P4, was distinguished by the presence of abundant vascular sprouts and new blood vessels in regions of mucosa over cartilage rings. The new vessels had the organized, ladder-like pattern typical of the adult (Fig. 2D,H,L). By P7, most of the vasculature over rings had the adult pattern. This stage began in the rostral trachea (P4–P5), then progressed to the caudal trachea (P5–P7), and was not complete in main stem bronchi until approximately P14. Remodeling and reorganization of blood vessels in regions of mucosa between cartilage rings, from P0 to P5, contributed to the transformation into the adult pattern (Fig. 2D,H,L).

Time Course of Tracheal Vascular Remodeling From E17.5 to Adult

Measurements of blood vessels over cartilage rings revealed that tracheas had 68% greater overall vascularity at E17.5 (area density 37%) than in the adult (area density 22%) (Fig. 3A,B). Tracheal vascularity abruptly decreased after birth, reaching the lowest value at P2 (area density 9%). The reduction in the first 2 days after birth reflected the regression of 76% of the vascular plexus in regions over the cartilage rings (Fig. 3B). From P2 to P5 tracheal vascularity more than doubled to an area density of 20%, which was approximately the same as the adult (Fig. 3B). The first significant increase in vascularity occurred between P3 and P4 (Fig. 3B).

Measurements of blood vessels crossing cartilage rings showed changes with a time course similar to the area density values. More than twice as many crossing vessels were present at E17.5 (21/mm) as in the adult (10/mm; Fig. 3C,D). Crossing vessels decreased from 21/mm at E17.5 to 17/mm at P0 (Fig. 3D) but fell to only 6/mm at P2 (71% decrease from E17.5). However, by P4 the number of crossing vessels had increased to the adult value (Fig. 3D). In the later postnatal stages and in adult tracheas, almost all ladder-like capillaries crossing the cartilage rings were conspicuously oriented along the longitudinal axis of the trachea.

The embryonic vascular plexus was highly branched and had abundant anastomoses that created capillary loops in the mucosa over the cartilage rings (Fig. 3E) rarely seen in the adult. Measurements revealed 316 loops per mm2 of cartilage at E17.5 compared with only 9 loops per mm2 in the adult (Fig. 3F). The number of capillary loops decreased slightly from E17.5 to E18.5 and from E18.5 to P0, but these changes were not statistically significant. The first significant reduction in loops occurred between P0 and P1 (Fig. 3F). The postnatal decrease in loops was progressive. Values at P2 were 85% less than at E17.5 and at P21 were 97% less than at E17.5.

Sprout-like projections from tracheal blood vessels were abundant from E17.5 through P9 but were rare in the adult (Fig. 3G,H). On average, vascular projections, identified by PECAM-1 and/or VEGFR-2 immunoreactivity, were nearly 40 times as numerous at E17.5 (198/mm2) as in the adult (5/mm2) (Fig. 3H). The number of projections was constant from E17.5 through P0, tended to increase during stages of regression and regrowth from P1 to P5, and decreased steadily thereafter (Fig. 3H). Many sprouts had strong VEGFR-2 immunoreactivity, but interconnecting blood vessels had weaker VEGFR-2 immunoreactivity.

Postnatal Development of Vascular Hierarchy and Phenotype

To determine whether the phenotype of tracheal blood vessels changed as they underwent remodeling after birth, we examined α-SMA–immunoreactive mural (smooth muscle) cells typical of arterioles, P-selectin–immunoreactive endothelial cells typical of venules, vascular basement membrane, and pericytes typical of capillaries and postcapillary venules.

Blood vessels in the primitive vascular plexus were not accompanied by α-SMA–immunoreactive cells at E17.5 (Fig. 4A). The first vessels with α-SMA–immunoreactive cells were found in regions overlying the trachealis muscle as early as at P2, but at P7 were found in most inter-cartilaginous regions (Fig. 4B). At P14, blood vessels typical of arterioles with circumferential α-SMA–positive cells along their entire length had the adult abundance and pattern (Fig. 4C).

thumbnail image

Figure 4. Development of vascular hierarchy and phenotype in mouse trachea. Differentiation of primitive vessels into arterioles or venules from embryonic day (E) 17.5 to adult. A–C: α-SMA immunoreactivity as a marker of arterial smooth muscle cells not found at E17.5, but present on scattered vessels at postnatal day (P) 7 and with adult abundance and pattern by P14. D,E: P-selectin immunoreactivity, a venular marker, faint and sparse at E17.5, but conspicuous in the adult (P70). F,G: Ephrin B2 immunoreactivity, an arteriole marker, not found in vascular plexus at P0 but weakly expressed on arteriole (arrow) at P7. H,I: EphB4 immunoreactivity, a venular marker, present on some blood vessels (arrows) and lymphatic vessels (arrowheads) at P0 and at P7. J,K: Type IV collagen immunoreactivity (vascular basement membrane marker) faint and diffuse on blood vessels at E17.5, but with adult intensity and pattern at P21. Scale bar in K = 120 μm in A–C, 40 μm in D,E, 20 μm in F–I, 40 μm in J,K.

Download figure to PowerPoint

At E17.5, faint P-selectin immunoreactivity was found on scattered endothelial cells but not on entire blood vessels (Fig. 4D). P-selectin immunoreactivity was stronger in intercartilaginous regions at P2, increased further at P7, and by P14 had the pattern typical of venules in adult (P70) mice (Fig. 4E).

We further examined the development of the tracheal vascular plexus for Ephrin B2 and EphB4 immunoreactivity as markers of arterial and venular phenotype, respectively. We choose regions of the mucosa between the cartilages at P0 and P7, because this is the location of most arterioles and venules. Ephrin B2 immunoreactivity was not detected in the vascular plexus at P0 (Fig. 4F) and was weak at P7 (Fig. 4G). EphB4 immunoreactivity was weak on some blood vessels at P0 and at P7 (Fig. 4H,I) and was also weak on lymphatic vessels, which were located in the intercartilaginous regions (Fig. 4H,I).

Basement membrane marked by type IV collagen or entactin/nidogen immunoreactivity was found on blood vessels at E17.5 but was faint and barely distinguishable from the remainder of the extracellular matrix (Fig. 4J). Staining for type IV collagen was still faint from P2 through P7 but had the adult pattern at P21 (Fig. 4K).

Pericytes identified by desmin immunoreactivity were sparse and loosely attached to the primitive vascular plexus at E17.5 (Fig. 5A) but became numerous after birth. Pericyte coverage was more than twice as extensive at P3 and P7 (Fig. 5B,C) as at E17.5. From P28 onward, pericyte coverage reflected by the proportion of tracheal vessels with pericytes was similar to that of the adult (P70; Fig. 5D) and reached maturity, as assessed by number of pericyte cell bodies per length of vessel, at P56 (Fig. 5E).

thumbnail image

Figure 5. Pericyte coverage of tracheal vasculature. A–E: Confocal micrographs of tracheal whole-mounts stained for PECAM-1 (green) and desmin (red) comparing pericyte coverage of vasculature in mice at embryonic day (E) 17.5, postnatal day (P) 3, and P7. A: At E17.5 pericytes are sparse. B,C: During capillary regrowth, new vessels over cartilage rings have more complete pericyte coverage at P3 and P7 than at E17.5. Scale bar = 20 μm. D,E: Measurements of pericyte maturity expressed as (D) percentage of tracheal capillary length accompanied by pericytes and (E) number of pericyte cell bodies per millimeter of capillary. Asterisks indicate significantly fewer pericytes (P < 0.05, Dunnett test) than in adult mice (P70).

Download figure to PowerPoint

These observations indicate that tracheal blood vessels of the primitive vascular plexus in late-stage embryos lack segment-specific markers of arterioles, capillaries, and venules, but then regress after birth, and regrow into the hierarchical network of mature vessels typical of the adult.

Vascular Remodeling in Other Organs

To address the question of whether the postnatal changes found in the tracheal vasculature occurred in all organs, we examined vessels of the kidney, brain, and liver from E17.5 to P7. The distinctive, specialized vasculature of kidney glomeruli was evident at E17.5 and changed little after birth. The vasculature of brain and liver at E17.5 was also largely similar to that of the adult. No primitive vascular plexus was present in these organs at E17.5, and extensive remodeling was not evident immediately after birth (data not shown).

Endothelial Cell Apoptosis, Regression, and Proliferation During Vascular Remodeling

To investigate the mechanisms of the postnatal vascular remodeling, we examined tracheal whole-mounts during the vascular regression phase (P0–P3) and during vascular regrowth (P5–P7). During the regression phase, apoptotic cells marked by activated caspase-3 immunoreactivity, were sparse (Fig. 6A), but most were associated with blood vessels (Fig. 6B,C). In contrast, during the vascular regrowth phase, dividing nuclei marked by phosphohistone H3 were abundant (Fig. 6D–F), but few were associated with blood vessels (Fig. 6E,F).

thumbnail image

Figure 6. Distribution of activated caspase-3 and phosphohistone H3 at postnatal day (P) 3/P7. A–C: Whole-mount of P3 trachea stained for activated caspase-3 immunoreactivity to mark apoptotic cells and PECAM-1 to mark endothelial cells. Some activated caspase-3–positive cells (arrows) are associated with blood vessels. B and C are enlargements of boxes in A and B. D–F: Whole-mount of P7 trachea stained for phosphohistone H3 immunoreactivity to mark dividing cells and for PECAM-1. Many cells stain for phosphohistone H3 (arrows). Although most (*) are not associated with blood vessels, some were in the wall or lumen of blood vessels. E and F are enlargements of boxes in D and E. Scale bar in D = 140 μm in A,D, 50 μm in B,E, 20 μm in C,F.

Download figure to PowerPoint

Hypoxia in Newborn Trachea

Next, we examined the amount and distribution of hypoxia in perinatal mouse tracheas using antibodies to pimonidazole adducts formed in hypoxic tissues. In P0 tracheas, pimonidazole staining of the epithelium was strong and uniformly distributed over and between cartilages, as was the vascular plexus (Fig. 7A–C,G). On P3, pimonidazole staining was markedly reduced, and was absent in most of the epithelium over cartilages (Fig. 7D–F,H). At this stage, few blood vessels were present over the cartilages, and most of those that remained were located in regions where the epithelium had pimonidazole staining (Fig. 7D–F,H). In cryostat sections of trachea at P0, the strong pimonidazole staining in the epithelium colocalized with HIF-1α immunoreactivity (Fig. 7I,J). Cartilage and other cells in the tracheal mucosa were not stained. No staining of the epithelium was present when pimonidazole and the HIF-1α primary antibody were omitted (Fig. 7K).

thumbnail image

Figure 7. Hypoxia in newborn mouse trachea. A–H: Whole-mounts of mouse tracheas at postnatal day (P) 0 or P3 stained for hypoxic cells (pimonidazole, green) and blood vessels (PECAM-1, red). A–C: At P0, pimonidazole staining of tracheal epithelium is strong and, like the vascular plexus, relatively uniformly distributed. D–F: However, at P3 pimonidazole staining is weaker in regions between cartilage rings and is absent in regions over cartilages where blood vessels are also absent except in scattered hypoxic regions (*). G,H: Enlargements of regions in boxes in C and F. I,J: Cross-section of P0 trachea showing the epithelium is stained for pimonidazole (I) and HIF-1α (J), but cartilages (*) are not. K: Negative control shows little staining by secondary antibodies when pimonidazole and HIF-1α primary antibody were omitted. Scale bar in K = 140 μm in A–F,I–K, 50 μm in G,H.

Download figure to PowerPoint

Destruction of Most of Tracheal Vasculature by VEGFR-2 Blockade From P0–P7

We asked whether blood vessels in the trachea at P0 were dependent on VEGF as a survival factor by blocking VEGFR-2 with DC101 during the first week of life. In untreated mice at P7, tracheal capillaries had the typical ladder-like pattern over the cartilages, and larger vessels were present between the cartilages (Fig. 8A,B). In sharp contrast, in mice treated with DC101, almost no capillaries were present over the cartilages and few other blood vessels were present elsewhere in the trachea (Fig. 8C,D). However, tracheal lymphatic vessels were still present after treatment with DC101, but LYVE-1 immunoreactivity was reduced (Fig. 8D).

thumbnail image

Figure 8. Destruction of most of tracheal vasculature by VEGFR-2 blockade from postnatal day (P) 0–P7. A–D: Blood vessels (PECAM-1) in whole-mounts of P7 tracheas untreated (A,B) or treated from P0 to P7 with VEGFR-2 blocking antibody DC101 (C,D). Normal ladder-like capillaries over cartilages in untreated trachea, but few capillaries in trachea treated with DC101. B,D: Enlargements of boxes in A and C showing many blood vessels between cartilages without treatment (C) but only one after DC101. Lymphatic vessels (LYVE-1) are still present. Scale bar in D = 140 μm in A,C, 50 μm in B,D.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

This study revealed that the vasculature of the mouse trachea underwent abrupt and rapid reorganization and change in pattern just after birth and then the arterioles, capillaries, and venules gradually acquired their adult phenotype. Before birth, the tracheal mucosa had a primitive vascular plexus similar to that found in the yolk sac. The primitive plexus was present at E17.5 but underwent conspicuous remodeling after birth. Blood vessels over cartilage rings regressed from P0 to P3, and new capillaries grew back from P3 to P7 to form the ladder-like pattern typical of the adult.

Studies of the embryonic yolk sac have contributed important elements of the understanding of the development of the primitive vascular plexus by vasculogenesis and subsequent growth and remodeling of the vasculature (Risau and Flamme,1995; Auerbach et al.,1997; Drake et al.,2000; Lucitti et al.,2007). The vasculature also undergoes remodeling after birth as animals grow but in most organs does not exhibit major changes in architecture. The vasculature of the eye and ovary are well-characterized exceptions, and these are exploited as powerful models for studying vascular growth, maturation, and regression (Augustin,2000; Gerhardt et al.,2003; Fraser and Duncan,2005; Fruttiger,2007). In the eye, the embryonic hyaloid vasculature begins to regress around birth, as the retinal vasculature grows radially from the optic disc across the retinal surface (Connolly et al.,1988; Ito and Yoshioka,1999). The retinal vasculature begins as a primitive anastomotic plexus, which expands by angiogenic sprouting and undergoes remodeling into a hierarchical, multilayered network of arterioles, capillaries, and venules (Connolly et al.,1988; Fruttiger,2007). While cyclic growth of ovarian follicles and corpus luteum formation and regression have many specialized features, they also demonstrate successive waves of angiogenesis and regression (Augustin,2000; Fraser and Duncan,2005). Previous studies of vascular regression in the papillary membrane in the newborn mouse eye suggest the involvement of macrophages triggering apoptosis in isolated endothelial cells, followed by cessation of blood flow and deprivation of VEGF as a survival factor (Lang and Bishop,1993; Lang et al.,1994; Meeson et al.,1999). Our finding of scattered endothelial cells at P0–P3 staining for activated caspase-3, a marker of apoptosis, is entirely consistent with these models.

We are not aware of another example of a system in which a primitive vascular plexus in a mammal regresses rapidly after birth and is then replaced by the adult pattern. The primitive vascular plexus of the trachea resembles the one in the yolk sac and retina, where an array of vascular polygons remodels into a hierarchical network of arterioles, capillaries, and venules. Most vessels of the tracheal vascular plexus at P0 had the typical endothelial markers PECAM-1, VEGFR-2, VE-cadherin, endoglin/CD105, and MECA-32. However, the markers of venular identity EphB4 and P-selectin were absent over cartilages and the few vessels that had these markers were located between the cartilages were venules are located. Markers of arterial identity (Ephrin B2 and circumferential smooth muscle cells) were not evident until P7. Blood vessels that grew from P2 to P5 to form the ladder-like capillaries over cartilages probably sprouted from surviving components of the primitive vascular plexus.

In the trachea, the architecture of the new vasculature is organized in a unique ladder-like pattern across cartilage rings. As in the other organs, the primitive vascular plexus of the trachea has more abundant vessels than the adult. Some 76% of the vessels over cartilage rings of the trachea were pruned as the vascular network with the adult pattern formed. Examination of the vasculature of the kidney, brain, and liver from E17.5 to P7 did not reveal vascular remodeling similar to that found in the trachea, but the question of whether vascular changes occur in other organs during this period deserves further study.

Our finding of abundant sprout-like projections and mitotic nuclei during the period of greatest vascular remodeling is consistent with angiogenesis. The stronger VEGFR-2 immunoreactivity of many angiogenic sprouts is consistent with endothelial tip cells (Gerhardt et al.,2003), while the weaker VEGFR-2 immunoreactivity of some interconnecting vessels may indicate regression. The prevalence of both types of vascular segments during the most dynamic period from P0 to P5 fits with the overlap of vessel regression and rapid regrowth by angiogenic sprouting.

Empty basement membrane sleeves, identified by type IV collagen or entactin/nidogen immunoreactivity (Inai et al.,2004), were not found in the trachea during the periods of rapid vascular regression or regrowth. Sleeves of basement membrane are present at sites of blood vessel regression in tumors and normal adult organs, in which endothelial cell loss follows inhibition of VEGF signaling (Inai et al.,2004; Baffert et al.,2006; Kamba et al.,2006; Mancuso et al.,2006). The lack of detection of empty basement membrane sleeves in the neonatal trachea may reflect the sparse basement membrane of the primitive vascular plexus or their susceptibility to degradation.

Changes in oxygen tension after birth may initiate or contribute to vascular regression and regrowth. Exposure of the retina to hyperoxic conditions after birth results in vascular pruning, and restoration of normoxia leads to new vessel growth (Campochiaro and Hackett,2003). In the trachea, the transition from the relatively hypoxic intrauterine environment to atmospheric oxygen may contribute the vessel pruning that occurs between P0 and P2. A more hypoxic microenvironment resulting from vascular regression would favor angiogenesis. However, such a mechanism does not explain why the vascular remodeling was restricted to the airways, where tissues are exposed to oxygen both from the airway lumen and from the blood supply.

Numerous studies have established the important roles of hypoxia, HIF-1α, and VEGF in the development of the lung and pulmonary vasculature (Tuder and Yun,2008), but to our knowledge, this is the first report of their contribution to postnatal vascular regression and regrowth in airways. Many reports have shown that hypoxia induces HIF family members and VEGF and promotes blood vessel growth and survival (Ferrara,2009; Fraisl et al.,2009). Hypoxia promotes and hyperoxia inhibits proper lung development (Compernolle et al.,2002; van Tuyl et al.,2005; Voelkel et al.,2006). Mice are born with immature lungs still at a saccular stage of development; alveolarization starts at P4 and is complete around P14 (Amy et al.,1977). Hypoxic regions of the embryo express HIF-1α and VEGF (Lee et al.,2001). Mice deficient in HIF-1α die within hours of birth from defects in lung development and symptoms similar to respiratory distress syndrome. Blockade of VEGF signaling impairs fetal lung maturation, and treatment with VEGF prevents fatal respiratory distress in premature mice (Compernolle et al.,2002; Zhao et al.,2005). A novel finding of our study is the persistence of hypoxia in the airway epithelium of newborn mice. Hypoxia then decreases, and regions with the least hypoxia match the regions with the greatest blood vessels regression.

The strong VEGF-dependence of blood vessels in the airways of newborn mice is consistent with a previous study of slightly older mouse pups, and with evidence that sensitivity to VEGF inhibitors of age-dependent (Gerber et al.,1999; Baffert et al.,2004; Thurston et al.,2005; Baffert et al.,2006). Indeed, we were surprised that pups treated with DC101 from P0 to P7 survived, because almost all tracheal blood vessels were eliminated. Blood vessels undergoing regression included not only those overlying the cartilages but also most that were located between the cartilages. Most of the vessels that remained were larger trunk vessels or lymphatics. However, these findings contrast with the relatively small reduction in the microvasculature of adult mice after treatment with inhibitors of VEGF signaling (Gerber et al.,1999; Baffert et al.,2004,2006; Inai et al.,2004; Thurston et al.,2005; Kamba et al.,2006).

Further studies are needed to determine the basis of the differing susceptibility of the vasculature of newborns and adults to VEGF inhibitors. More information is also needed about factors such as fibroblast growth factor 9 (FGF9) and sonic hedgehog that influence the spatial and temporal expression of VEGF (White et al.,2007; Lavine et al.,2008). Blood vessel remodeling in development is orchestrated by a complex interplay of VEGFs, angiopoietins, PDGFs, HGF, ephrin/Eph family members, and other growth factors (Carmeliet,2005; Shibuya,2008; Augustin et al.,2009). Naturally occurring angiogenesis inhibitors may also contribute (Cooke and Kalluri,2008). For example, cartilage is avascular and has been reported to contain angiogenesis inhibitors (Moses et al.,1999; Liu et al.,2001; Shukunami et al.,2008).

Other factors may be mechanical stresses imposed on conducting airways at the onset of breathing. Mechanical forces can have pronounced effects on blood vessel development (Lucitti et al.,2007), and shear stress on endothelial cells from flowing blood can activate genes involved in angiogenesis (Risau and Flamme,1995). Cyclical stretch applied to coronary microvessels increases VEGFR-2 protein, tube formation, and microvessel branching in vitro (Zheng et al.,2008). The rostral–caudal gradient in the development of the tracheal microcirculation supports the contribution of mechanical forces to the postnatal vascular remodeling. The rostral end of the trachea is anchored more securely and may experience larger mechanical forces with the onset of breathing than the caudal end, which has more freedom to move. The kidney, brain, and liver, which did not have conspicuous vascular remodeling after birth, are not exposed to the same mechanical forces as the airways. Mechanical ventilation of newborn mouse lungs can induce changes in VEGFR-2 mRNA, lung remodeling, and angiogenesis that is not driven by hypoxia (Mokres et al.,2010). A conspicuous feature of the newly formed capillaries overlying the cartilage rings is that they were aligned with the direction of cyclical stretch that tracheas experience after the onset of breathing at birth. The primitive embryonic plexus is not exposed to such forces.

In conclusion, the vasculature of the airway mucosa in late-stage embryos is completely different than in the adult. The primitive vascular plexus undergoes rapid remodeling after birth, whereby segments of the plexus regress in the first few days after birth and then capillaries rapidly regrow to form the typical adult vascular pattern. Hypoxia, HIF-1α, and VEGF are likely to play a role in this remodeling. In examining the mechanisms underlying the rapid onset of vascular remodeling after birth, changes in tissue oxygen tension and mechanical forces associated with breathing are worthy topics for future investigation. Whatever the mechanism, the dynamic nature of the airway vasculature during postnatal development is consistent with its extraordinary plasticity in disease and provides a robust model for investigating factors that promote or inhibit vascular remodeling after birth.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Mice

C57BL/6 mice (Charles River, Hollister, CA) of either sex were housed under barrier conditions. Mice were examined at E17.5, E18.5, P0 (equivalent to E19.5), at daily intervals from P1 through P7, at 2- or 3-day intervals from P7 to P14, and at P21 and P70. We determined embryonic age by counting from discovery of the vaginal plug. Although the gestation period of C57BL/6 mice is reported as 18.5 days (http://jaxmice.jax.org/manual/index.html#breeding), in fact, we observed pups born from E18.5 to E20.5. The Institutional Animal Care and Use Committee of the University of California at San Francisco approved all experimental procedures.

Fixation by Immersion or Vascular Perfusion

Pregnant females were anesthetized (intramuscular ketamine, 83 mg/kg, and xylazine, 13 mg/kg), the uterus was removed, and the embryos were removed and euthanized by immersion in ice-cold phosphate-buffered saline (PBS) for 30 min without exposure to room air. Embryos were then immersed in 1% paraformaldehyde (PFA) in PBS (pH 7.4), and tracheas were removed under a dissecting microscope and fixed for 1 hr at room temperature. Mice at P0 and older were anesthetized with ketamine and xylazine, and tracheas were fixed by vascular perfusion of 1% PFA in PBS (pH 7.4) for 2 min at a pressure of 120 mmHg through a cannula (20-gauge for ages P0 to P7; 18-gauge for older mice) inserted into the ascending aorta by means of the left ventricle. The atria were punctured for drainage of blood and fixative. After perfusion, tracheas were removed and fixed further by immersion in 1% PFA at room temperature for 1 hr.

Immunohistochemistry

For ease of making flat whole-mounts, the cylindrical tracheas from E16.5 to P7 were cut along the dorsal longitudinal axis, and older tracheas were cut along the ventral longitudinal axis. The whole-mounts were then permeabilized in PBS containing 0.3% Triton X-100 (PBS-T), and then incubated in PBS-T containing 5% normal goat or donkey serum (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA), 0.2% bovine serum albumin, and 0.1% sodium azide for 1 hour at room temperature to block nonspecific antibody binding.

Tracheas were stained for immunohistochemistry by incubation in a combination of two or three primary antibodies overnight at room temperature. Endothelial cells were stained with anti-PECAM-1 (1:500, hamster anti-mouse, clone 2H8, Thermo Scientific, Waltham, MA or rat anti-mouse, clone MEC 13.3, BD Biosciences, San Diego, CA), anti-VEGFR-2 (1:2,000, rabbit polyclonal T014; Rolf Brekken, University of Texas Southwestern Medical Center, Dallas, TX), MECA-32 (1:500, rat anti-mouse clone MECA-32, BD Biosciences), or anti-endoglin/CD105 (1:500, rat anti-mouse CD105, clone MJ7/18, eBioscience, San Diego, CA). Pericytes were identified with anti-desmin (1:2,000, rabbit polyclonal 8592, Abcam, Cambridge, MA or 1:500, rabbit monoclonal, clone Y66, Millipore, Billerica, MA). Smooth muscle cells on arterioles and venules were stained with anti-α-smooth muscle actin (1:500, Cy3-conjugated mouse monoclonal anti-α-SMA, clone 1A4, Sigma-Aldrich, Saint Louis, MO or desmin). Arterial segments were identified by staining for Ephrin B2 (1:250, goat polyclonal AF496, R&D Systems, Minneapolis, MN). Lymphatic vessels were identified by staining for LYVE-1 (1:500, rabbit polyclonal 11-034, AngioBio, Del Mar, CA). Basement membrane was stained with anti-type IV collagen (1:10,000, rabbit polyclonal LSL-LB -1403, Cosmo Bio, USA, Carlsbad, CA) or anti-entactin/nidogen (1:1,000, rat monoclonal ELM1, Millipore). Apoptotic cells were stained with activated caspase-3 (1:500, rabbit polyclonal AF835, R&D Systems) and mitotic nuclei were identified by phosphohistone H3, (1:500, rabbit polyclonal 06-570, Millipore). Endothelial cells of venules were identified by anti-P-selectin (1:400, rat anti-CD62p, clone RB40.34, BD Biosciences or by EphB4, 1:250, goat polyclonal AF446, R&D Systems). Hypoxia was detected immunohistochemically (Samoszuk et al.,2004) using a Hypoxyprobe-1 kit (Hpi, Inc, Burlington, MA). In brief, pimonidazole hydrochloride was injected subcutaneously at a dose of 60 mg/kg body weight and mice were perfused 60 min later. In most cases, pimonidazole adducts bound to hypoxic tissues were detected by a fluorescein isothiocyanate (FITC) -labeled mouse antibody (1:500, clone 4.3.11.3), but we confirmed the results using a rabbit polyclonal antibody 2627 (1:500, Hpi, Inc). Hypoxia-Inducible Factor 1α (HIF-1α) was detected using rabbit polyclonal antibody NB100-479 (1:500, Novus Biologicals, Littleton, CO).

After incubation with primary antibodies, tracheas were rinsed with PBS-T and incubated overnight at room temperature with species-specific secondary antibodies (IgG labeled with FITC, Cy3, or Cy5, 1:500, Jackson ImmunoResearch Laboratories Inc.) diluted in 5% normal serum in PBS-T. Negative controls included omitting the primary antibody or substituting a nonspecific IgG of the same species as the primary antibody, or for detection of hypoxia, omitting the injection of pimonidazole. Tracheas were then rinsed with PBS-T, fixed briefly in 4% PFA, rinsed again with PBS, and mounted on glass slides in Vectashield (Vector Laboratories, Burlingame CA). Fluorescence images were acquired with a conventional fluorescence microscope (Zeiss Axiophot) or confocal microscope (Zeiss LSM510). For clarity, some fluorescence images were converted into grayscale images with inverted polarity by using Photoshop software. Cartilage rings were visualized in tracheal whole-mounts by transmitted light and images were superimposed on fluorescence images.

Measurements of Vascular Density

The vascularity of regions over cartilage rings of tracheal whole-mounts was assessed by measuring the fractional area density occupied by PECAM-1–immunoreactive blood vessels (Thurston et al.,1998; Baffert et al.,2004). Briefly, the live video image was overlaid with a computer-generated square lattice (20-μm squares) and then the number of points of the lattice that intersected vessels was scored. Area density was expressed as the percentage of lattice points that intersected blood vessels. Measurements were made on a region 1.7 mm2 in area of each of 5 to 10 rostral-most cartilage rings (n = 3–5 mice per group).

Measurements of Vascular Organization and Complexity

The number of crossing capillaries, vascular loops, and vascular projections (sprouts) was measured using Image-J software (http://rsb.info.nih.gov/ij) on digital images measuring 480 × 640 μm (×10 objective, ×2 Optovar zoom) or 960 × 1,280 μm (×10 objective, ×1 Optovar zoom) obtained from the same regions used for area density measurements (n = 3–5 mice per group).

The number of PECAM-1–immunoreactive capillaries crossing the midline of a cartilage ring and connecting to vessels in an adjacent intercartilaginous region was counted (Thurston et al.,1998; Baffert et al.,2004). The values were expressed as number of vessels per millimeter length of cartilage ring.

Vascular loops were counted within regions of interest, averaging 0.08 mm2 in area of cartilage ring. Vascular loops were defined as vessels, identified by PECAM-1- or VEGFR-2–immunoreactivity, which joined other vessels to form loops over cartilage rings. Values were expressed as number of loops per square millimeter.

Vascular projections (sprouts) were counted as thin, tapered, PECAM-1- or VEGFR-2–immunoreactive extensions from blood vessels in regions over cartilage rings used for vascular loop measurements. Values were expressed as number of projections per square millimeter of cartilage ring.

Measurement of Pericyte Coverage

Pericyte coverage of capillaries overlying cartilage rings was expressed as the proportion of capillary length accompanied by desmin-positive pericytes in tracheal whole-mounts stained for PECAM-1 and desmin (4). Pericyte cell bodies, identified as desmin-positive cells closely associated with capillaries, were expressed as the number of cell bodies per millimeter length of PECAM-1–positive capillary. Measurements were made with a digitizing tablet on real-time fluorescent video images of capillaries overlying five cartilage rings of each trachea (×20 objective; n = 3–5 mice per group).

Inhibition of VEGFR-2

VEGF signaling was blocked by a neutralizing antibody to mouse VEGFR-2 (rat monoclonal antibody DC101, UCSF Cell Culture Facility; original clone from ImClone, New York, NY) injected subcutaneously into pups at a dose of 100 μg/g body weight on P0 and 40 μg/g on P3 (Witte et al.,1998; Baluk et al.,2005). Control pups received no treatment. Mice were fixed by vascular perfusion at P7.

Statistics

Values were expressed as means ± SEM (n = 3–7 mice per group). The significance of differences among groups was assessed using two-way ANOVA followed by Dunn-Bonferroni or Dunnett post hoc tests for multiple comparisons. P values < 0.05 were considered significant.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We thank Rolf Brekken at the University of Texas Southwestern Medical Center (Dallas, TX) for the kind gift of the rabbit VEGFR-2 antibody used for immunohistochemistry, and Dr. Tatsuma Okazaki for many helpful discussions during this work.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES