Introduction

Abstract Introduction Materials and Methods Results Discussion References

Neonatal respiratory function depends on the prenatal development of a pulmonary capillary bed that is adequate to allow efficient diffusion of oxygen from the air-liquid interface of the alveolus to the pulmonary vasculature. Although much is known about the histologic changes that accompany the development of the air-blood barrier in the lung, the factors regulating this process are unknown (1).

Vascular endothelial growth factor (VEGF) is a potent inducer of endothelial cell growth in vitro and angiogenesis in vivo (2, 3). VEGF is a secreted peptide that acts specifically on vascular endothelial cells through two different high-affinity receptors, designated KDR/flk-1 (4) and flt-1 (5). The messenger RNA (mRNA) for these receptors, as well as the mRNA for VEGF and VEGF protein, have been identified in the lung of the rat (6). High levels of VEGF mRNA (7) and protein (8) have also been localized in airway epithelial cells of midtrimester human fetal lung, which suggests that capillary development in the lung is modulated through secretion of VEGF by alveolar epithelial cells.

Human VEGF mRNA exists as four transcripts, derived by alternative splicing of a single precursor mRNA, which encode polypeptides of 121, 165, 189, and 206 amino acids (9). The VEGF isoforms VEGF₁₂₁ and VEGF₁₆₅ are secreted as homodimeric glycoproteins, in contrast to the VEGF₁₈₉ and VEGF₂₀₆ isoforms, which are bound to the extracellular matrix (10, 11). VEGF gene expression is known to be modulated by a variety of factors, including oxygen (13), cyclic adenosine monophosphate (cAMP) (14), glucocorticoids (14), transforming growth factor- (15), platelet-derived growth factor (PDGF), and platelet-activating factor (PAF) (16).

A frequently used model for studying the effects of regulatory factors on lung development is human fetal lung explants maintained in vitro. The differentiation of midtrimester human fetal lung tissue in explant culture has been characterized both biochemically and morphologically (17, 18, 19). After 4 d in culture in serum-free defined medium and 95% air/5% CO₂ (20% O₂) at 37°C, the undifferentiated epithelial cells of lung tissue derived from 15- to 20-wk abortuses spontaneously differentiate into typical alveolar type II cells. This process involves the transformation from a tissue containing abundant connective tissue surrounding small prealveolar ducts lined with an undifferentiated columnar epithelium to tissue with thin connective tissue septa separating large luminal spaces lined almost entirely by lamellar body-containing, cuboidal type II pneumonocytes (18). The spontaneous differentiation of human fetal lung in vitro is influenced by glucocorticoids and factors that modulate cAMP levels within the tissue (20).

In the present study we used human fetal lung explants maintained in vitro to study the regulation of VEGF gene expression and to gain information about the possible role of VEGF in the developing human lung. The study had two objectives: (1) to determine the effects of incubation time, low oxygen, cAMP, and glucocorticoids on VEGF mRNA expression in human fetal lung maintained in vitro; and (2) to localize VEGF protein and VEGF mRNA in human fetal lung tissues after several days in culture and after exposure to regulatory factors.

	Materials and Methods

Abstract Introduction Materials and Methods Results Discussion References

Lung Tissues and Organ Culture

Lung tissues were obtained from human abortuses of 16- to 20-wk gestation from a local clinic or from Advanced Biosciences Resources (Alameda, CA), as approved by the University of Iowa Human Subjects Review Committee. As previously described (21), fetal lung tissue was dissected free of major blood vessels and airways, minced into 1 to 2-mm³ pieces with a sterile scalpel blade, and placed in organ culture. Lung tissues from one or two fetuses were used for each experiment. Cultures were maintained in serum-free Waymouth's MB752/1 medium (GIBCO, Grand Island, NY) in the absence or presence of the cAMP analog Bt₂cAMP (1 mM) or the synthetic glucocorticoid dexamethasone (1 to 100 nM). Tissues were maintained in standard incubator conditions of 37°C and 5% CO₂/95% air (20% O₂), or in modular incubators (Billups-Rothenberg, Del Mar, CA) flushed with a gas mixture of 2% O₂, 5% CO₂, and 93% N₂ for a total of 4 to 6 d of incubation. The modular incubators were flushed for 3 min at a flow rate of 10 liter/min (lpm), equilibrated for 30 min at 37°C, flushed again for 3 min at 10 lpm, and then sealed for 24 h of incubation. The gaseous atmospheres and the media were changed every 24 h. Incubated tissues were harvested after 2, 4, or 6 d of culture, frozen in liquid N₂, and stored at 70°C.

Immunoblot Analysis

Midtrimester human fetal lung tissues that had been maintained in explant culture for 4 d were homogenized in ice-cold water containing phenylmethysulfonyl fluoride (1 mM). The homogenates were centrifuged at 600 × g for 5 min. Supernatant fractions were collected and assayed for total protein (22). Fifty micrograms of protein were separated by electrophoresis on a 10% polyacrylamide gel using a Tris-glycine-sodium dodecyl sulfate (SDS) buffer. The separated proteins were transferred to Immobilon membranes (Millipore Corp., Bedford, MA) by electrophoresis. The membranes were blocked with 5% nonfat dry milk (wt/vol) in Tris-buffered saline (TBS; 0.1 M, pH 7.5) that contained 0.05% (vol/vol) Tween 20, and were then incubated overnight at 4°C with rabbit antihuman VEGF antibody (1:5,000) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). The membranes were rinsed three times for 10 min each in TNT buffer (Tris-buffered saline with 0.05% [vol/vol] Tween 201) and then incubated for 45 min at room temperature with goat antirabbit IgG conjugated to alkaline phosphatase (1:30,000) (Sigma, St. Louis, MO). Following three 10-min rinses with TBS, the membranes were incubated for 5 min in 5-bromo-4-chloro-3- indolylphosphate/nitroblue tetrazolium substrate (Sigma). The membranes were then rinsed and air-dried.

Immunostaining

The frozen fetal lung tissue was mounted in ornithyl carbamyltransferase compound, and 7-µm sections were prepared with a cryostat and thaw-mounted on glass slides. Sections were fixed for 10 min at room temperature in freshly prepared 10% formalin in phosphate-buffered saline (PBS). The sections were then rinsed twice for 10 min each in PBS. The sections were stained with a Vectastain Elite kit (Vector Laboratories, Burlingame, CA). Nonspecific binding sites were blocked by incubating the sections with 2% normal goat serum at room temperature. The sections were quickly rinsed in PBS and then incubated for 1 h in a humidified chamber at room temperature with a rabbit antihuman VEGF antibody (Santa Cruz Biotechnology) at a dilution of 1:50 in PBS. The tissue sections were washed twice in PBS for 5 min per washing, and then incubated for 30 min in biotinylated secondary antibody before being rinsed twice in PBS for 5 min per rinse. The sections were next incubated for 45 min in avidin-peroxidase reagent. After rinsing twice in PBS for 5 min per rinse, the sections were incubated in diaminobenzidine (700 µg/ml) for 1 to 3 min. Sections were rinsed in PBS for 5 min, rinsed quickly in distilled water, and then dehydrated and mounted with glass coverslips. In some experiments the sections were counterstained with hematoxylin for 30 s. Negative controls were incubated with secondary antibody alone for all experimental conditions. Sections were viewed and photographed with a Nikon FX (Melville, NY) photomicroscope.

RNA Isolation and Northern Blot Analysis

Frozen tissues that had been incubated in 2% or 20% oxygen atmospheres for the final 2 d of culture were thawed in 4.0 M guanidinium thiocyanate and homogenized. Total RNA was isolated (23) and quantitated by determining the absorbance at 260 nm. Ten micrograms of total RNA for each sample were separated on a 1.2% agarose/formaldehyde gel and then transferred to a Nytran membrane (Schleicher & Schuell, Keene, NH) by vacuum blotting (Bio-Rad Laboratories, Hercules, CA) for Northern blot analysis for VEGF mRNA. Absence of ethidium bromide staining of the gels after transfer confirmed that all transfers were complete. RNA was crosslinked to membranes by ultraviolet irradiation for 5 min followed by baking of the blots for 90 min at 80°C. Membranes were prehybridized at 42°C for 6 to 12 h in hybridization buffer (bovine serum albumin [BSA; 0.2% wt/vol], polyvinylpyrrolidone [0.2% wt/vol], ficoll [0.2% wt/vol], Tris-HCl [50 mM, pH 7.4], Na pyrophosphate [0.1% wt/vol], SDS [1% wt/vol], formamide [50% vol/vol], dextran sulfate [10% wt/vol], NaCl [1 M], and denatured salmon sperm DNA [100 µg/ ml]). The membranes were hybridized overnight at 42°C in hybridization buffer containing 1 × 10⁶ cpm/ml of ³²P- labeled complementary DNA (cDNA) specific for human VEGF (VEGF cDNA was a kind gift of Dr. Donald S. Torry, University of Tennessee, Knoxville, TN) (24). After hybridization, the blots were washed twice for 5 min at room temperature in 2× saline sodium citrate (SSC; NaCl [300 mM], sodium citrate [30 mM], pH 7.0), twice for 30 min each at 42°C in 1× SSC plus SDS (1% wt/vol), and once for 15 min at room temperature in 0.1× SSC. Blots were exposed to X-ray film with an intensifier screen at 70°C for 24 to 72 h. After the blots were hybridized to VEGF cDNA, they were stripped and reprobed with a radiolabeled cDNA for human 18S ribosomal RNA (rRNA) (American Type Culture Collection, Rockville, MD). Autoradiograms of the hybridized blots were quantified by scanning densitometry. Values were corrected for RNA loading and transfer errors by adjustment to the relative amounts of 18S rRNA.

In Situ Hybridization

Frozen sections (7 µm thick) were cut at 20°C and mounted on glass slides (Superfrost Plus; Fisher, Chicago, IL). Sections from undifferentiated distal fetal lung tissue and cultured human fetal lung explants were mounted on each slide to facilitate comparisons. The methods used for in situ hybridization were modifications of those originally described by Angerer and Angerer (25). The sections were allowed to come to room temperature and then fixed in freshly prepared paraformaldehyde (4% wt/vol, pH 7.5). After several rinses in PBS (pH 7.4), the tissues were incubated in Pronase solution (0.25 mg/ml in Tris HCl [50 mM, pH 7.5]) and ethylenediamine tetraacetic acid (EDTA; 5 mM). After incubation for 10 min in PBS that contained glycine (2 mg/ml), the sections were rinsed in triethanolamine buffer (TEA; 0.1 M, pH 8.0) and then treated with acetic anhydride (0.25%) in TEA buffer for 10 min. Sections were rinsed in 2× SSC (1× = NaCl [150 mM], sodium citrate [15 mM], pH 7.0) and were then dehydrated through an ethanol series.

Fifty microliters of hybridization buffer (NaCl [300 mM], Tris HCl [10 mM, pH 8.0], EDTA [1 mM], formamide [50% vol/vol], 1× Denhardt's solution [Sigma], dextran sulfate [10% wt/vol], and yeast transfer RNA [tRNA; 0.28 mg/ml]) that contained the antisense VEGF [³H]cRNA probe was applied to each slide. A coverslip (HybriWell Chambers; Lab Vision Corp., Fremont, CA) was then placed on top of the sections and sealed. The slides were incubated overnight at 60°C in a humid chamber. After removal of the coverslip, the slides were rinsed four times in 4× SSC for 10 min each time. The sections were then incubated for 30 min at 37°C in a solution of ribonuclease A (RNAse A; 20.0 µg/ml) and RNAse T1 (3.0 U/ml) in buffer (Tris HCl [10 mM, pH 8.0], EDTA [1 mM], NaCl [0.5 M]). The sections were then rinsed for 30 min at 37°C in Tris-HCl (10 mM, pH 8.0), EDTA (2.5 mM), and NaCl (0.5 M), and then in 2× SSC for 30 min at room temperature, in 0.1× SSC at 56°C for 30 min, and in 0.1× SSC at room temperature for 30 min. The sections were dehydrated through an ethanol series, air-dried, and coated with NTB-2 emulsion (Kodak, Rochester, NY), after which they were dried and exposed in light-proof boxes at 4°C. The photographs of VEGF mRNA hybridization presented in Figures 1, 2, and 3 are from slides that were exposed for approximately 3 wk. The autoradiograms were developed in D-19 developer, rinsed in distilled water, fixed in rapid fixer, rinsed in distilled water, stained in hematoxylin, dehydrated, and mounted with glass coverslips.

View larger version (180K): [in this window] [in a new window]   Figure 1.   Low-magnification (A, B) and high-magnification (C, D) brightfield and darkfield photomicrographs of in situ hybridization for VEGF mRNA in midtrimester human fetal lung. The connective tissue (CT and arrows), the lumina of distal airways (L), and distal airway epithelial cells (arrowheads) are indicated. Bars = 100 µm.

View larger version (191K): [in this window] [in a new window]   Figure 2.   Brightfield and darkfield photomicrographs of in situ hybridization for VEGF mRNA in midtrimester human fetal lung maintained in vitro. (A, B) Tissue incubated in 20% O2 and serum-free medium for 4 d. (C, D) Tissues incubated in 2% O2 and some serum-free medium for 4 d. The connective tissue (CT and arrows), lumina of distal airways (L), and distal airway epithelial cells (arrowheads) are indicated. Bar = 100 µm.

View larger version (121K): [in this window] [in a new window]   Figure 3.   Brightfield and darkfield photomicrographs of in situ hybridization for VEGF mRNA in midtrimester human fetal lung maintained in vitro. Tissue was incubated in 20% O2 and serum-free medium for 4 d and then for an additional 2 d in the absence (A, B) (control), or presence (C, D) of Bt2cAMP. (E, F) Tissue incubated in 20% O2 and hybridized with the sense probe for VEGF mRNA, demonstrating specificity of the antisense probe. Connective tissue (CT), lumina of distal airways (L), and distal airway epithelial cells (arrowheads) are shown. Bar = 100 µm.

Statistical Analysis

Statistical comparisons of the densitometric mRNA data were made with a two-tailed Dunnett's test (26). Some comparisons were made with two-way repeated-measures analysis of variance (ANOVA), as stated in the text. Differences were considered significant when P < 0.05. Each experiment was done with lung tissues obtained from one or two fetuses, and each experiment was repeated three to four times with tissues from different fetuses.

	Results

Abstract Introduction Materials and Methods Results Discussion References

Effect of Incubation Time and Oxygen on VEGF mRNA and Protein in Midtrimester Human Fetal Lung

Midtrimester human fetal lung explants were incubated at 37°C under standard incubator conditions of 5% CO₂/95% room air (20% O₂) or 5% CO₂/2% O₂, and were harvested after 2 and 4 d of incubation. VEGF mRNA was detectable in tissues before culture and after incubation of the explants in either 2% O₂ or 20% O₂ for 2 or 4 d (Figure 4). VEGF mRNA levels were significantly increased in tissues incubated for 2 d (4.9 ± 0.6-fold) and 4 d (9.8 ± 2.5-fold) in 2% O₂, and for 2 d (3.7 ± 0.6-fold) and 4 d (6.3 ± 1.2-fold) in 20% O₂ compared with VEGF mRNA levels in midtrimester human fetal lung (starting tissue). Two-way repeated-measures ANOVA of the densitometric data presented in Figure 4 revealed a significant effect of 2% O₂ versus 20% O₂ (P = 0.05) and a nearly significant effect of 2 d versus 4 d in culture (P = 0.07) when these variables were controlled independently. Northern blot analysis demonstrated the major VEGF mRNA transcript to be 3.7 kb long (Figure 4). Minor transcripts were consistently observed at 1.4 kb and 1.8 kb. The levels of minor transcripts did not change independently with duration of incubation of the tissues or with oxygen concentration. Autoradiograms of Northern blots for VEGF mRNA that were overexposed showed a band at 4.4 kb as well, but this was not observed consistently (data not shown).

View larger version (17K): [in this window] [in a new window]   Figure 4.   Graph and representative autoradiogram of VEGF mRNA levels from Northern blot analysis of RNA from midtrimester human fetal lung maintained for 2 or 4 d in 2% O2 or 20% O2. Values represent the means ± SEM (n = 4) of densitometric determinations of the 3.7-kb band, with normalization to the "start" (not incubated) condition. *P < 0.05 versus start condition; +P < 0.05 versus Day 2, 20% O2 condition; oP = 0.10 versus Day 2, 2% O2 condition.

VEGF protein was detected by immunoblot analysis of protein from midtrimester human fetal lung before and after incubation in 2% O₂ or 20% O₂ (Figure 5). In tissues that had not been maintained in vitro, two bands of approximately 23 kD each were observed; however, a single band of 23 kD, consistent with VEGF₁₆₅, was observed in incubated tissues (Figure 5). VEGF protein levels were increased in tissues incubated in 2% O₂ compared with both 20% O₂-treated tissues and tissues that had not been incubated (Figure 5).

View larger version (42K): [in this window] [in a new window]   Figure 5.   Representative immunoblot for VEGF protein of midtrimester human fetal lung tissues cultured in 2% and 20% O2 for 2 and 4 d of incubation in serum-free medium. SDS-polyacrylamide gel electrophoresis was used to separate 50 µg of protein homogenate from tissue that was not incubated (Start) or from tissues incubated for 2 to 4 d in either 2% or 20% O2. Human recombinant VEGF (50 ng) was run as a standard (Stand). The blot is representative of three separate experiments.

To determine the cellular distribution of VEGF protein, immunostaining was done on human fetal lung tissues that had not been incubated, and on explants that had been incubated in 2% O₂ or 20% O₂ atmospheres for 4 d. Undifferentiated midtrimester human fetal lung tissue exhibited weak immunoreactivity for VEGF protein, with staining observed primarily in the epithelial cells of the distal airways (Figure 6A). We did not observe an effect of gestational age of the midtrimester human fetal lung on the amount or location of VEGF immunoreactivity. Immunoreactivity was markedly increased in tissues incubated in 2% O₂ or 20% O₂ for 4 d compared with that in tissues before culture, with the greatest immunoreactivity observed in tissues incubated in 2% O₂ (Figures 6C and 6D). In accord with previous observations of lack of morphologic change in human fetal lung tissues maintained in low-oxygen environments (21), explants incubated in 2% O₂ failed to develop the dilated lumina observed in 20% O₂-treated tissues (Figures 6C and 6D). Examination at high power of tissues incubated in 2% and 20% O₂ revealed differences in the localization of VEGF immunoreactivity for the different conditions (Figure 6E versus 6F). Tissues maintained in 20% O₂ showed immunoreactivity localized primarily to the basement membrane region subjacent to distal airway epithelial cells (Figure 6E). In contrast, immunoreactivity in tissues incubated in 2% O₂ was observed in both airway epithelial cells and the basement membrane region subjacent to the airway epithelial cells (Figure 6F). The pattern and density of immunoreactivity was generally observed to be uniform throughout the tissues maintained under each particular condition. However, in a few instances, increased immunostaining was observed in cells in the interior of the explant maintained in 20% O₂ compared with cells in the periphery of the explants (data not shown). Under no condition was VEGF immunoreactivity confined to cells in the interior of the explants.

View larger version (160K): [in this window] [in a new window]   Figure 6.   Photomicrographs of midtrimester human fetal lung tissues immunostained for VEGF. (A) Tissue that was not incubated. (B) Control tissue that was stained with PBS instead of the primary antibody against VEGF, demonstrating the specificity of the antibody staining. (C) Immunoreactivity for VEGF in tissue incubated in 20% O2 and serum-free medium for 4 d. (D) Immunoreactivity for VEGF in tissue incubated in 2% O2 and serum-free medium for 4 d. Higher-magnification photomicrograph of VEGF immunostaining of midtrimester human fetal lung tissue maintained for 4 d in (E) 20% O2 or (F) 2% O2. The connective tissue (CT), lumina of distal airways (L), distal airway epithelial cells (arrowheads), and basement membrane (arrows) are indicated. Bar in lower right corner of each panel represents 100 µm.

In Situ Hybridization for VEGF mRNA in Midtrimester Human Fetal Lung

In situ hybridization studies were performed to determine the cells of origin for VEGF mRNA in midtrimester human fetal lung before and after incubation. VEGF mRNA was detectable primarily in airway epithelial cells and to a lesser extent in connective tissue cells of undifferentiated midtrimester human fetal lung (Figure 1). Cultured explants maintained in 2% O₂ and 20% O₂ also exhibited hybridization to VEGF mRNA in both distal airway epithelial cells and connective tissue cells (Figure 2). Both the 2% O₂- and the 20% O₂-treated explants showed increased hybridization signal compared with human fetal lung tissue that had not been maintained in culture, with the greatest amount of signal present in the 2% O₂-treated tissues (Figure 2). VEGF mRNA hybridization was observed to be uniform throughout the explants, with the exception of an occasional 20% O₂-treated tissue explant in which a greater VEGF mRNA signal was observed in the interior cells of the explant than in more exterior cells. However, in no tissues was VEGF mRNA hybridization observed only in the interior cells of the explants. These data show that VEGF mRNA production increases in human fetal lung tissue maintained in explant culture, and that VEGF mRNA production is induced in both airway epithelial cells and connective tissue cells by low-oxygen environments.

Effect of Regulatory Factors on VEGF mRNA and Protein Expression in Human Fetal Lung In Vitro

To determine whether regulatory factors other than oxygen modulate VEGF mRNA levels in developing human lung, we examined the effects of the cAMP analog Bt₂cAMP and the synthetic glucocorticoid dexamethasone. Midtrimester human fetal lung tissues were maintained in 20% O₂ atmospheres and serum-free media for 4 d to induce differentiation of type II cells, and were then further incubated for 48 h in either 2% O₂ or 20% O₂ in the absence or presence of Bt₂cAMP (1 mM) or dexamethasone (100 nM). Total RNA was isolated from the tissues, and Northern blot analysis for VEGF mRNA was performed. Although there was a marked increase in levels of VEGF mRNA in all 2% O₂-treated tissues compared with tissues incubated in 20% O₂, no effect of either Bt₂cAMP or dexamethasone treatment was observed in these explants (Figure 7). In contrast to those in the 2% O₂-treated tissues, VEGF mRNA levels in tissues maintained in 20% O₂ were increased nearly 3-fold in tissues incubated with Bt₂cAMP compared with the 20% O₂ control (Figure 7). The levels of VEGF mRNA in fetal lung tissues that were first incubated for 4 d in 20% O₂ to induce epithelial cell differentiation were not affected by dexamethasone (Figure 7). However, when explants were incubated from the outset in the presence of various concentrations of dexamethasone in 20% O₂ for 4 d, VEGF mRNA levels increased with the dose of dexamethasone (Figure 8). Under these conditions, VEGF mRNA levels were increased 2.2-fold (range: 1.7- to 2.7-fold, n = 3) in dexamethasone (10⁷ M)-treated tissues compared with controls.

View larger version (20K): [in this window] [in a new window]   Figure 7.   Graph and representative autoradiogram of VEGF mRNA levels from Northern blot analysis of RNA from midtrimester human fetal lung maintained for 4 d in 20% O2 and serum-free medium followed by an additional 2 d of incubation in either 2% O2 or 20% O2 in the absence (C) or presence of dexamethasone (D) or Bt2cAMP (B). Values represent the means ± SEM (n = 4) of densitometric determinations of the major (3.7-kb) band, corrected for loading and transfer errors by 18s rRNA values and then normalized to the 20% O2 control values. *P < 0.05 versus 20% O2 control (C) condition.

View larger version (83K): [in this window] [in a new window]   Figure 8.   Autoradiograph of a representative Northern blot of total RNA probed for VEGF mRNA. Midtrimester human fetal lung was placed in explant culture in 20% O2 and 5% CO2 in the absence (C) and presence of dexamethasone (109, 108, 107 M) for 4 d. The blot was stripped and reprobed for 18s rRNA as demonstrated.

In situ hybridization of VEGF mRNA in explants incubated in the absence or presence of Bt₂cAMP demonstrated a marked increase in signal in the Bt₂cAMP-treated tissues over that in the untreated tissues (Figure 3). VEGF mRNA signal was increased in both the airway epithelial cells and the connective tissue cells of Bt₂cAMP-treated issues.

Immunostaining for VEGF protein in tissues incubated for 2 d in the presence of dexamethasone after induction of differentiation of the epithelium by incubation in 20% O₂ for 4 d demonstrated no differences in the intensity of immunoreactivity or location of staining as opposed to those in 2% O₂- or 20% O₂-treated tissues incubated without dexamethasone (data not shown). In contrast, Bt₂cAMP treatment resulted in increased immunostaining for VEGF in the basement membrane region of the prealveolar ducts in the 20% O₂-treated tissues compared with control tissues (Figure 9). Immunostaining for VEGF protein was not observed to be different in 2% O₂-treated tissues incubated with Bt₂cAMP than in controls (Figure 9).

View larger version (129K): [in this window] [in a new window]   Figure 9.   High-power photomicrographs of VEGF immunostaining of midtrimester human fetal lung tissues that were maintained for 4 d in 20% O2 to induce type II-cell differentiation, followed by an additional 2 d of incubation in (A) 20% O2, (B) 20% O2 with Bt2cAMP, (C) 2% O2, or (D) 2% O2 with Bt2cAMP. The connective tissue (CT), lumina of distal airways (L), epithelial cells (arrowheads), and basement membrane (arrows) are indicated. Bar = 50 µm.

These data show that in addition to oxygen, cAMP regulates VEGF mRNA levels in human fetal lung in vitro. Dexamethasone treatment results in increased VEGF mRNA levels only when lung tissue explants are treated from the outset of incubation. Dexamethasone treatment does not induce VEGF mRNA levels in tissues that have been preincubated in 20% O₂ for 4 d to induce type II cell differentiation.

	Discussion

Abstract Introduction Materials and Methods Results Discussion References

The present study shows increases in VEGF gene expression in distal airway epithelial cells in association with the spontaneous differentiation of human fetal lung in vitro. Maintenance of human fetal lung explants in vitro in serum-free medium is characterized by differentiation of distal airway epithelial cells into type II alveolar pneumocytes (18). Thus, our results suggest that VEGF mRNA and protein levels increase in distal airway epithelial cells as the lung develops. Interestingly, a localization of VEGF protein to the basement membrane region of distal airway epithelial cells was also associated with differentiation of the human fetal lung in vitro and with treatment of the tissue with a cAMP analog. These data suggest that VEGF production and elaboration by differentiating alveolar epithelial cells play a role in supporting and/or directing the development of the alveolar capillary bed.

In the human fetus, VEGF mRNA is found in abundance in tissues that become highly vascularized, such as the kidney, spleen, and lung, and to a lesser extent in heart, adrenal gland, pancreas, liver, and brain (8). Vascular development in some organs, such as the brain and kidney, occurs through the process of angiogenesis, which involves the sprouting of capillaries from existing vessels with subsequent invasion into adjacent tissue (27). In contrast, capillary development in the lung occurs through the process of vasculogenesis (28), which is defined by the de novo proliferation and organization of endothelial cell precursors, which assemble into capillaries and then associate with the systemic circulation (29). VEGF is a potent and specific endothelial cell mitogen that induces angiogenesis in a variety of in vivo models, including the chicken chorioallantoic membrane (30), the rabbit cornea (31), the ischemic hindlimb of the rabbit (32), and the ischemic myocardium of the dog (33). The absolute requirement of VEGF for vasculogenesis has been demonstrated by inactivation studies of VEGF alleles (34, 35) and of the VEGF receptors flt-1 (36) and KDR/flk-1 (37). In each of these studies, inactivation of the target genes resulted in lethal phenotypes characterized by deficient organization of endothelial cells and endothelial cell precursors. Additionally, because inactivation of VEGF alleles and VEGF receptor genes resulted in early embryonic death, the studies provided no information specific to development of the pulmonary capillary bed, since the fetal lung had not begun to develop at the time of embryonic death. Given the effect of VEGF in directing the assembly of blood vessels and stimulating the proliferation of capillary endothelial cells, our data showing localization of VEGF protein to the basement membrane region of distal airway epithelial cells suggests that the production of VEGF by type II cells and type II cell precursors, along with translocation of the protein to the basal lamina, may be important, if not necessary, for the establishment of a normal air-blood barrier in the developing alveolus.

The best-characterized regulator of VEGF gene expression is oxygen. VEGF mRNA levels increase in low-oxygen environments and decrease in a dose-dependent manner with increasing oxygen exposure in a number of cultured cell models (38). This "hypoxic" regulation of VEGF is similar to that of erythropoietin (12) in that both factors require new protein synthesis for the inhibitory effect of oxygen to occur, and in that each factor is stimulated by cobalt and inhibited by carbon monoxide, suggesting the involvement of a heme-binding protein in the modulation of these proteins by oxygen (12). To determine whether VEGF protein and mRNA levels are regulated by oxygen in the developing human lung, we incubated midtrimester human fetal lung tissues in atmospheres of 2% O₂ and 20% O₂. In accord with cell-culture data, we observed a significant increase in VEGF mRNA levels in tissues incubated in 2% O₂ versus tissues incubated in 20% O₂. VEGF protein levels were similarly shown to be increased by a low oxygen concentration in immunoblots and by immunohistochemistry. These data indicate that VEGF is regulated by oxygen in the developing human lung, but they raise the question of why VEGF mRNA levels are greater in tissues incubated in 20% O₂ than in midtrimester human fetal lung, which has not been exposed to this relatively high oxygen concentration. The most plausible explanation is that VEGF expression is increased in human fetal lung explants as a function of type II cell differentiation in tissues maintained in vitro. Thus, the increased expression of VEGF mRNA that we observed in 2% O₂-treated tissues suggests that the induction of VEGF by a low oxygen concentration may be a separate effect from that of induction of VEGF by tissue differentiation. Consistent with our observation of increased VEGF gene expression with in vitro maturation of the human fetal lung was the observation by Amin and colleagues (39) of an increase in VEGF mRNA expression with maturation of the fetal and neonatal mouse lung. The relationship between low-oxygen environments, VEGF gene expression, and capillary development in the fetal lung is unclear. Our investigation does not shed further light on the role of a low oxygen environment in normal lung development. If a low-oxygen environment does play a role in regulating VEGF expression during lung development, this may come about as the lung grows and tissue mass increases to the point at which oxygen delivery to the tissues is inadequate. However, we speculate that in normal lung development, it is the differentiation of the epithelial cells that is primarily responsible for the increase in VEGF expression.

To determine whether VEGF mRNA levels are regulated in human fetal lung by factors known to modulate the differentiation of human fetal lung in vitro, we evaluated the effects of a cAMP analog and the synthetic glucocorticoid dexamethasone on VEGF mRNA levels and immunolocalization of VEGF protein in human fetal lung maintained in organ culture. The spontaneous differentiation of human fetal lung in vitro is associated with increased levels of endogenous prostaglandins and can be inhibited by inhibition of prostaglandin synthesis (40). This prostaglandin induction of type II cell differentiation in midtrimester human fetal lung in vitro is mediated by cAMP and is prevented by the incubation of explants in the presence of the cAMP-dependent protein kinase inhibitor H-8 (41). Furthermore, exogenous cAMP stimulation results in the accelerated differentiation of type II cells in human fetal lung in vitro (42). Interestingly, the spontaneous differentiation of human fetal lung in vitro, in either the absence or presence of exogenous cAMP, is prevented by incubation of tissues in low-oxygen environments (21). Thus, in the present study, we observed significant increases in VEGF mRNA levels in tissues that had been preincubated in 20% O₂ for 4 d to induce type II cell differentiation, and then incubated in the presence of the cAMP analog Bt₂cAMP and 20% O₂, relative to control explants. Decreased VEGF mRNA levels were present in these tissues compared with those incubated in 2% O₂ alone, indicating the relative potency of a low oxygen concentration as a regulator of VEGF mRNA levels. Treatment with cAMP was observed to overcome some of the inhibitory effect of incubation at the relatively high oxygen concentration of 20% on VEGF mRNA levels. Of perhaps greater significance was our observation of increased immunostaining for VEGF protein in tissues incubated in 20% O₂ and Bt₂cAMP. This immunoreactivity was localized to distal airway epithelial cells (type II cells in incubated tissues), and was especially dense in their subjacent basement membranes. These data strongly suggest that type II cells produce VEGF, which translocates to the subjacent basement membrane in order to direct the assembly of the pulmonary capillary bed. Others have demonstrated increases in VEGF mRNA and protein levels induced by cAMP and by factors that increase cAMP levels in cardiac myocytes (13) and osteoblasts (14). The production of VEGF mRNA by adult type II alveolar cells during recovery from hyperoxic injury has also been demonstrated (43), as has VEGF mRNA expression by a subpopulation of type II cells from neonatal rabbit lung (44), further supporting the possible importance of alveolar type II cells as a source of VEGF in the lung.

Because the synthetic glucocorticoid dexamethasone also influences the differentiation of midtrimester human fetal lung in vitro (45), we examined the effect of dexamethasone on VEGF mRNA levels and VEGF immunostaining in human fetal lung in vitro. We found that VEGF mRNA levels were increased in lung tissues maintained in dexamethasone for 4 d from the beginning of incubation in 20% O₂. In contrast, when the epithelium of the explants was first induced to differentiate by incubation in 20% O₂ for 4 d and was then treated with dexamethasone, VEGF mRNA levels were not increased. The differentiation of distal airway epithelial cells to type II cells with 4 d of incubation in serum-free medium and 20% O₂ has been well documented (18, 42). Dexamethasone treatment induces type II cell differentiation in human fetal lung explants (45). Therefore, it is possible that the increases in VEGF mRNA levels that we observed in tissues incubated at an early point with dexamethasone were due to the effects of dexamethasone on type II cell differentiation, and were not a direct effect of this glucocorticoid on VEGF gene expression. The lack of effect of dexamethasone after tissues were preincubated to induce epithelial cell differentiation supports this possibility. As with the cAMP-treated tissues, there was also no detectable effect of dexamethasone treatment on VEGF mRNA levels in tissues incubated in 2% O₂. It also is possible that any effect of either cAMP or dexamethasone on VEGF mRNA levels was undetectable because it was masked by the low-oxygen-induced increase caused by 2% O₂. It is interesting to note, however, that dexamethasone has been observed to decrease VEGF mRNA levels in an osteoblastic cell line (14) and to inhibit the inducing effects of PDGF and PAF on VEGF mRNA levels in vascular smooth-muscle cells (16). The discrepancy in the effect of dexamethasone on VEGF mRNA levels may be due to the differences in the systems studied (whole-tissue explant versus cell culture) or to the possibility that a dominant effect of dexamethasone on the differentiation of type II cells in human fetal lung explants resulted in a relative increase in VEGF mRNA. Furthermore, in the present study we found that dexamethasone treatment had no effect on the staining intensity or localization of immunoreactivity for VEGF protein in cultured human fetal lung.

In summary, we have demonstrated the in vitro modulation of VEGF mRNA and protein in midtrimester human fetal lung by incubation of explants in 2% O₂ versus 20% O₂ and by cAMP. The observed localization of VEGF protein to distal airway epithelial cells and to the basement membrane subjacent to type II epithelial cells supports the possibility that capillary development is regulated by airway epithelial cells and VEGF in the human lung. Further investigation is underway to determine the precise role of VEGF in regulating both the assembly and growth of the nascent capillary bed in the human lung. Such studies may lead to strategies for inducing development of the capillary bed in the immature lung.