INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

POSTNATAL LUNG DEVELOPMENT involves the formation of alveoli from existing but immature respiratory saccules by means of septation, maturation of epithelial and mesenchymal units, and development of the alveolar capillary system (4, 21). In rats, alveolar septation begins at postnatal days (P) 3-4 and continues throughout the second week of postnatal development. Maturation of newly formed units occurs in parallel and continues throughout the fourth week of postnatal lung development. These processes are tightly regulated by direct interactions between epithelial and mesenchymal cells (5).

Several intrinsic and extrinsic factors have been identified as modifiers of postnatal lung growth. For example, hypoxia over prolonged periods can affect lung growth (27). There has been some controversy, however, regarding the overall effects of hypoxia on the postnatal development of the lung (3, 22). Whether such issues are related to age or stage of development at the time of exposure is not well defined, and possible mediators of the effects of hypoxia on lung morphology remain essentially unexplored.

Whereas the molecular mechanisms underlying the effects of hypoxia on lung growth remain to be delineated, significant advances have been made in our understanding of the mediators controlling prenatal lung morphogenesis and branching. Members of the transforming growth factor (TGF)- family have emerged as crucial factors controlling the formation, patterning, and maturation of lung tissue (13, 24). TGF- belongs to a superfamily of polypeptides including 1) TGF- isoforms themselves, 2) activins, and 3) a complex third subfamily consisting of morphogenic proteins (bone morphogenic proteins, nodal, Xenopus Vg-1, Drosophila dpp, and screw). On most cell types studied, three classes of receptors for TGF- (TR) have been identified: the type I (TRI), type II (TRII), and type III (TRIII) receptors. Biological responses to TGF- are induced on binding of activated TGF- ligand to TRII, which induces the formation of a heterooligomeric complex of TRI and TRII. TRI is then phosphorylated by the constitutively active kinase domain of TRII, and downstream signaling is initiated (20, 25). TGF- is an important regulator of cellular differentiation and proliferation in the lung (e.g., by inhibition of epithelial cell proliferation, or induction of mesenchymal cell proliferation) and thus represents an important candidate in the control of lung development and growth. Furthermore, early lung branching is inhibited by TGF- (31, 32), providing functional evidence for a key role of TGF- as a negative regulator of lung branching and prenatal lung development. Little is known, however, about the regulation of TGF- activity and the expression patterns of TRs in the lung and whether they are affected by low-O₂ conditions. In this study, we, therefore, investigated the effects of chronic hypoxia on postnatal lung development, with a focus on characterization of the TGF- system under normal and low-O₂ conditions.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Study animals. Sprague-Dawley rats (Charles River Laboratories, Wilmington, MA) were placed into 9.5% oxygen at P3 and reared in hypoxia until P14. For each set of experiments, litters were trimmed to eight pups to minimize confounding variables related to litter size and lung growth. Weights of P14 hypoxic animals were reduced by ~54% compared with controls (14.7 ± 1.2 vs. 34.1 ± 1.12 g), whereas there was no weight difference in the adult rats. Other observations, such as enlargement of the right ventricle, were present in both hypoxic P14 and adult animals (data not shown). In rare instances, pups from the hypoxic P14 group died, whereas all animals in the remaining groups survived. No data were generated from dead or moribund animals. Similarly, adult rats raised in room air were placed into 9.5% oxygen for 2 wk before death with their normoxic controls. The adult rats examined were separate animals and not the mothers of the neonatal litters exposed. At the end of each experimental period, hypoxic and control rats were anesthetized and killed, as approved by Animal Services. Lung tissue was harvested, rinsed, and immediately frozen in dry ice. All experimental groups were housed in pathogen-free chambers and monitored daily for any signs of infection by the Animal Services Facility of Yale University School of Medicine.

Lung processing and histological analysis. In selected animals from each experimental group, lungs were inflated through tracheotomies to 20 cmH₂O pressure with 4% paraformaldehyde (pH 7.4), and tracheas were ligated. Lungs and hearts were excised en bloc, submersed in 4% paraformaldehyde overnight, and processed by Research Histology (Critical Technologies Program, Dept. Pathology, Yale University) for paraffin embedding and sectioning. Lung tissues were cut into 3-µm sections and stained with Trichrome. The degree of alveolar septation was analyzed under light microscopy with a 1 × 1 mm grid at ×200 magnification. The grid was divided into 11 equally spaced horizontal lines; each septum, which crossed the horizontal lines, was counted. A total of 10 separate grids from three randomly selected animals was analyzed for each slide, and the average number of septae per grid was calculated. Similarly, septal thickness was analyzed with the Image Pro software. In brief, slides were examined at ×500 magnification under a grid of five equally spaced horizontal lines. Each grid examined consisted solely of alveolar tissue; bronchioles were intentionally omitted from the field of view. The thickness of each septum crossing a given horizontal line was measured perpendicular to its course at that crossing point. A total of 10 separate grids from three randomly selected animals was analyzed, and the average septal thickness was then calculated for each grid.

Collection of bronchoalveolar lavage fluid. Tracheotomies were performed in selected animals from each experimental group, and tracheas were cannulated with an appropriately sized catheter. Lungs were lavaged with either 1 ml (P14 animals) or 3 ml (adult animals) of PBS (pH 7.4) and immediately frozen in 70°C. Protein concentration was equalized by using the Bradford assay according to the manufacturer's instructions.

Luciferase assay for TGF- activity. Active TGF- levels in bronchoalveolar lavage fluid (BALF) were measured by using the p3TP-Lux reporter construct and MvLu1 cells, as described previously (1). Briefly, BALF was obtained from three different animals of each experimental group, as indicated. MvLu1 cells transfected with the TGF- responsive reporter p3TP-Lux were then incubated in DMEM containing 1% fetal bovine serum, 10 mM HEPES, and 20% of each respective BALF for 24 h. Supernatants were removed, and Luciferase activity was measured in total cell lysates after 24 h. Luciferase activities were normalized by transfection controls, as described (11). To investigate the contribution of each of the three TGF- isoforms (TGF-₁, TGF-₂, and TGF-₃) to BALF-induced p3TP-Lux statement, each sample was preincubated with the respective isoform-specific neutralizing anti-TGF- antibody in a separate set of experiments, as recommended by the manufacturer (R&D Systems), before and during incubation of the BALF sample on MvLu1 cells.

Western blot analysis. Frozen lung tissue was homogenized by using a tissue grinder in lysis buffer consisting of 20 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, and 1 mM -glycerophosphate. The proteinase inhibitor Complete (Roche Molecular Biochemicals, Indianapolis, IN) and Na₃VO₄ (2 mM) were added to the lysis buffer immediately before homogenization. Homogenates were sheared by multiple aspiration through a 24-gauge needle, agitated at 4°C for 30 min, and centrifuged at 15,000 rpm (4°C) for 10 min. Supernatants were taken as whole tissue lysates, and protein concentration was measured by using the Bradford assay according to the manufacturer's instructions. Equal amounts of protein (4 µg) were separated on 7.5% SDS-PAGE gels and transferred to nitrocellulose, as described earlier (10). Western blots were performed with antibodies against constitutive heat shock cognate 70 (HSC70; 1:5,000; Stressgen, Victoria, BC), TRI (1:1,000; Santa Cruz Biotechnologies, Santa Cruz, CA), TRII (1:1,000; Santa Cruz Biotechnologies), and TRIII (betaglycan; 1:500; R&D Systems). Specific bands were visualized after incubation with the respective secondary antibodies by autoradiography by using enhanced chemiluminescence, according to the manufacturer (SuperSignal, Pierce Chemicals, Rockford, IL). Specificity of primary antibodies was assessed by incubation with their respective blocking peptides.

Densitometry measurements. Densitometry measurements of Western blots from each experimental group were obtained (n  6 for each group), and absolute values were equalized with constitutive HSC70.

Immunohistochemical analysis. Lung tissue sections were deparaffinized in xylene for 3 × 5 min and rehydrated in 100% ethanol for 2 × 10 min, 95% ethanol for 2 × 10 min, and distilled water for 1 × 5 min. Antigen retrireview was performed in a pressure cooker using 6.5 mM sodium citrate (pH 6.0). Immunohistochemical analysis was performed by using the indicated primary antibodies at 1:50 dilutions. Immunoreactivity was detected by avidin-biotin complex staining according to the manufacturer's instructions (Santa Cruz Biotechnologies). Endogenous peroxidase activity was quenched by incubating the sections in 1.5% hydrogen peroxidase-methanol for 10 min. Specificity was assessed through staining of control sections with primary antibodies in the presence of their respective blocking peptides. Because blocking peptides for TRIII were not available, negative control slides for TRIII were generated by incubation in the presence of a species-corresponding, unspecific primary antibody. Peroxidase-labeled sections were counterstained with hematoxylin-eosin for 10 s.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Morphological and histological changes in rat lungs exposed to hypoxia. Exposure of rats to chronic hypoxia dramatically affected normal postnatal development of the lung (Fig. 1). Histological examination of lungs obtained from hypoxic rats at P14 (Fig. 1D) revealed a remarkable arrest of alveolarization compared with their normoxic controls (Fig. 1B). Furthermore, lungs from hypoxic animals at P14 showed little progression in development compared with newborn animals, with a predominance of terminal saccules in peripheral tissue rather than the well-formed alveoli seen in normoxic controls. In comparison, lung tissue from adult rats exposed to hypoxia (Fig. 1E) showed no obvious differences compared with their normoxic controls (Fig. 1C). Quantitative analysis of septal number and thickness demonstrated a significant reduction in average septal number per grid in hypoxic animals at P14 compared with controls (71 ± 8.34 vs. 87.4 ± 6.6 septae/grid, P < 0.005). In adult animals, consistent with histological data, we found no statistically significant differences (92 ± 7.4 vs. 80.8 ± 7.84 septae/grid for hypoxia vs. normoxia). Similar results were obtained when we quantified the average septal thickness. In lungs from hypoxic animals at P14, septal thickness was reduced by >45% compared with control lungs (5.28 ± 0.55 vs. 9.56 ± 2.17 µm, P < 0.001). In adult animals, average septal thickness was not significantly altered when the animals were exposed to chronic hypoxia (4.42 ± 0.35 vs. 4.56 ± 0.19 µm).

View larger version (118K): [in this window] [in a new window]   Fig. 1.   Histological changes in lung tissue exposed to 9.5% O2. Lung sections from postnatal day 3 rats (A) before any hypoxic exposure show large terminal saccules with very little alveolar septation. Compared with normoxic animals at postnatal day 14 (P14; B), lungs from hypoxic animals at P14 (D) fail to show the normal septation that occurs during the first 2 wk of postnatal life, and their septae appeared thinner and less cellular than their normoxic counterparts. There were no observable differences in adult lungs exposed to hypoxia (C and E). Each photograph is representative of 6 separate animals, taken from at least 3 different litters. All photographs were taken under ×400 magnification.

Measurement of bioactive TGF- in hypoxic rat lungs. Hypoxia affected lung morphology specifically during the period of alveolar development, where it led to an arrest of alveolarization. Because the TGF- system functions as a potent inhibitor of prenatal lung development, we sought to analyze bioactive TGF- levels in rat lungs using a TGF--sensitive assay (1). As depicted in Fig. 2A, active TGF- levels were significantly increased in BALF obtained from hypoxic animals at P14, compared with their normoxic controls (1,120 ± 153 vs. 510 ± 130 relative light units, P < 0.02). Interestingly, we found no statistically significant differences in TGF- bioactivity in BALF when comparing hypoxic and normoxic adult rats (Fig. 2A). The Luciferase construct used to measure TGF- bioactivity, p3TP-Lux, is similarly responsive to all three TGF- isoforms. Therefore, we performed additional reporter gene experiments after incubation of BALF samples with isoform-specific, neutralizing antibodies directed against TGF-₁, TGF-₂, or TGF-₃. Figure 2B demonstrates that the greater than twofold increase in bioactive TGF- levels observed in BALF from P14 hypoxic rats is abrogated by coincubation with anti-TGF-₁ and anti-TGF-₂ antibodies (P < 0.05), whereas incubation with anti-TGF-₃ antibody did not result in a significant decrease. The greatest part of TGF- bioactivity in hypoxic BALF samples can, therefore, be attributed to the presence of TGF-₁ and TGF-₂ isoforms.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Measurement of bioactive transforming growth factor (TGF)- levels in rat bronchoalveolar lavage fluid (BALF). A: in BALF from P14 animals, chronic hypoxia caused a significant increase in TGF- activity compared with normoxia (* P < 0.02). No statistically significant difference was observed between normoxic and hypoxic adults. B: when BALF samples were incubated with neutralizing, isoform-specific antibodies against TGF-1 or TGF-2, the increase in bioactive TGF- levels with hypoxia was partially abrogated (* P < 0.05). Incubation with anti-TGF-3 antibody did not result in a significant difference compared with controls. All experiments were done in triplicate with 3 separate BALF samples. Values are means ± SE.

Statement of TR in hypoxic rat lungs. Because chronic hypoxia resulted in increased TGF- activity in BALF samples at P14 but not in adults, we further analyzed the expression patterns of TRs in hypoxic and control lungs to determine whether hypoxia exerted complementary changes in receptor expression. Western blot analysis revealed the presence of all three receptors, TRI, TRII, and TRIII, at the expected sizes in whole lung extracts (Fig. 3A). Chronic hypoxia led to a dramatic upregulation of TRI in the P14 group (Fig. 3A, lanes 1-4). Further analysis of this effect by densitometry demonstrated that this upregulation was fourfold above the expression levels seen in normoxic animals (Fig. 3B). Interestingly, this upregulation of TRI expression was not observed in adult animals (Fig. 3A, lanes 5-8). Here, densitometry measurements revealed no significant changes in TRI expression (Fig. 3B). Similar to TRI, TRII expression was upregulated at P14 with hypoxia (Fig. 3A, lanes 1-4). Densitometry revealed that the increase in expression was ~30-35% above control levels (Fig. 3B). In adult animals, hypoxia caused no significant difference in TRII expression (Fig. 3A, lanes 5-8). Different observations were made, however, for TRIII. TRIII expression was dramatically downregulated in response to hypoxia, both at P14 (Fig. 3A, lanes 1-4) and in adult animals (Fig. 3A, lanes 5-8), which was confirmed by densitometry (Fig. 3B).

View larger version (43K): [in this window] [in a new window]   Fig. 3.   TGF- receptor (TR) isotype expression in rat lungs exposed to hypoxia. A: total lung proteins were extracted, and aliquots (4 µg) were analyzed by Western blot analyses using antibodies specific for TRI [activin receptor-like kinase (ALK)-5], TRII, TRIII, and heat shock cognate 70 (HSC70), as indicated. Note the dramatic increase in TRI and TRII expression at P14, whereas TRIII showed a marked downregulation at P14 (lanes 1-4). In comparison, adult animals exposed to hypoxia demonstrated no change in TRI and TRII levels and a decrease in TRIII statement (lanes 5-8). Equal protein loading was assessed with an antibody against constitutive HSC70. The 110-kDa band of TRIII represents the core protein, whereas the >200-kDa band represents the mature protein, which has been subjected to posttranslational modifications, including glycosylation (9, 30). Of note, densitometry measurements for TRIII reflect integration of the >200- and 110-kDa bands, because both are specific for TRIII. Specificity of primary antibodies was assessed by incubation with their respective blocking peptides (data not shown). Molecular sizes are indicated on the left. N, normoxic rats; H, hypoxic rats. B: densitometry was performed to quantify the data obtained from Western blot analysis. All changes in receptor isotype expression noticed on Western blot analyses of P14 animals were statistically significant (* P < 0.01). In adult hypoxic animals, only the decrease in TRIII expression reached statistical significance (n  6 rats for all experimental groups). Values are means ± SE.

Localization of TR in rat lungs. Because we observed major changes in the expression levels of TRs in response to hypoxia, specifically in P14 animals, we further studied the localization patterns of these receptors in hypoxic and normoxic lungs from animals at this age. Immunohistochemical analysis of lung tissue obtained from P14 rats demonstrated that all three receptors were predominantly localized to bronchiolar epithelium of larger airways and alveolar tissue (Fig. 4, A-C, arrows). Less intense signals were also noted in endothelial cells. Consistent with Western blot data, the staining for TRI and TRII appeared to be more intense in hypoxic lungs (Fig. 4, D and E), particularly in alveolar tissue. Similarly, the staining for TRIII was less intense in hypoxic tissues (Fig. 4F) compared with normoxic tissues. Of note, smooth muscle cells stained only weakly for all three receptors. To assess antibody specificity for TRI and TRII, hypoxic slides were also stained after incubation of the primary anti-TRI and anti-TRIII antibodies with their corresponding blocking peptides (Fig. 4, G and H, respectively). A blocking peptide for the TRIII antibody was not available, as such; control stainings for this receptor were performed by using species-corresponding, unspecific primary antibody. These experiments clearly demonstrated the specificity of our stainings (Fig. 4, G and H).

View larger version (121K): [in this window] [in a new window]   Fig. 4.   Localization of TR isotypes in P14 rat lung. Lungs from normoxic and hypoxic rats were fixed in 4% paraformaldehyde, processed for immunohistochemistry, and hybridized with antibodies against TRI (ALK-5), TRII, and TRIII. Staining for TRI (A), TRII (B), and TRIII (C) was observed primarily in bronchiolar (arrows) and alveolar epithelium of normoxic P14 animals, with less intense immunoreactivity seen in endothelium. Of note, smooth muscle cells stained only weakly for all 3 receptors. With hypoxia, immunoreactivity for TRI (D) and TRII (E) became particularly intense in alveolar tissues, whereas signal for TRIII (F) was barely detectable. Antibody specificity for TRI (G) and TRII (H) was assessed by preincubating each primary antibody with its respective blocking peptide before staining. For TRIII (I), control slides were stained with species-compatible, unspecific primary antibody. All photos were taken under ×200 magnification, with insets of higher magnification.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In the present study, we characterized the effects of chronic hypoxia on TGF- and TR isotype expression in rat lungs obtained from P14 and adult animals. We found that bioactive TGF- levels, as well as the expression levels of the receptor isotypes TRI and TRII, were increased in hypoxia. These effects of hypoxia on TGF- activity and TR expression were age dependent; they were only observed early in postnatal development but not in adult animals. Interestingly, the overall increase in TGF- activity was observed at a developmental stage when hypoxia led to an arrest in alveolar development (Fig. 1 and Refs. 3, 22). These different responses to hypoxia at two distinct stages of development imply that specific and possibly different regulatory mechanisms control lung development and morphology at these specified ages. The molecular mechanisms underlying such differences, however, have thus far remained obscure. Importantly, we were able to perform the histological and molecular analyses simultaneously in identical animals. Hence, we had an ideal opportunity to concurrently investigate both the morphological changes induced by chronic hypoxia, as well as the potential mechanisms underlying these alterations in structure.

With respect to morphology, hypoxia clearly affected normal postnatal lung development. Because we could demonstrate major changes in the TGF- system in response to hypoxia, we believe that these changes can explain, at least in part, the observed lung phenotype for three major reasons. First, an increase in bioactive TGF- ligand was observed only in P14 animals exposed to hypoxia, and only P14 animals exhibited the described morphological changes with hypoxia. Indeed, it has been shown that prenatal lung branching in vitro is inhibited by TGF- signaling (31, 32). Although our present study focuses on the effects of hypoxia during postnatal life, these past investigations, nevertheless, provide functional evidence that TGF- is a negative regulator of lung morphogenesis. Interestingly, adult animals exposed to hypoxia did not show any differences in TGF- activity, coinciding with the fact that alveolar morphology was not altered with hypoxia at this age. It is important to note, however, that our study did not rely on measurements of total TGF- in BALF, because there is a large amount of inactive TGF- in biological samples, e.g., pools deposited within the extracellular matrix (2, 20, 26).

Second, we found a major upregulation in TRI and, to a lesser degree, TRII expression in hypoxic P14 animals. This, in combination with the increased levels of bioactive TGF-, would lead to an overall upregulation of the signal transducing components of this system. In addition, although there is significant controversy surrounding the exact role of TRIII (or betaglycan), there is evidence that, in certain cell types, especially epithelial cells, this receptor may function as an inhibitor of TGF- signaling (9). In this respect, it has become increasingly clear that the cell surface expression of TRs, especially TRI and TRII, is able to regulate not only the cellular response to TGF-, but also the activation of specific downstream signaling molecules (6-8). Accordingly, the observed upregulation of TRI and, to a lesser extent, TRII expression specifically in hypoxic P14 animals would result in an increased TGF- response. The dramatic downregulation of TRIII (betaglycan) expression could also lead to enhanced downstream signaling. The resulting overall response would lead to an arrest in early lung development, either by premature differentiation, inhibition of proliferation, or a combination thereof (20, 25).

Finally, our observation that all three TRs mostly localized to airway and alveolar epithelial cells, which are ultimately responsible for the generation of new lung, also supports our belief that the observed changes in the TGF- system have an impact on the phenotypic alterations seen early in development. Indeed, TGF- is a well-known inhibitor of epithelial cell proliferation (19, 25). Importantly, the localization of TRI and TRII did not change with hypoxia, but rather became more intense in alveolar cells. Similarly, the localization of TRIII did not change, but rather became less intense in both bronchiolar and alveolar cells.

Although we present data supporting the fact that chronic hypoxia induced the observed morphological changes by directly affecting target cells such as alveolar epithelial cells, we realize that, in vivo, hypoxia also produces a variety of hormonal, nutritional, cardiovascular, and biochemical alterations that may lead indirectly to such morphological changes. Indeed, the decreases in whole body weights of hypoxic P14 animals compared with controls may suggest additional synergisms with nutritional issues. Past investigators, however, have reported that hypoxia induces morphological alterations in the lung that are independent of, although accentuated by, any nutritional differences (27). Therefore, it is possible that, in our model, hypoxia induces its effects directly or indirectly.

Hypoxia is a potent stimulus that causes complex changes in multiple coexisting signal transduction pathways, which may confound its overall effect on the lung as a single organ. As such, it will be important to examine specific regulatory mechanisms of postnatal lung development and then assess how hypoxia specifically affects these processes. For example, it is well known that epithelial-mesenchymal interactions are of crucial importance in controlling lung development (18, 21). During alveologenesis, close interactions of proliferating and differentiating epithelial cells with their surrounding endothelial cells are required for the subsequent proper formation of alveoli. It follows that, in the case of deficient or improper angiogenesis, the formation of alveoli is also inhibited, a mechanism that has been suggested to contribute to diseases such as bronchopulmonary dysplasia (BPD) or emphysema (14, 28). In this respect, it has recently been shown that TRI receptor expression is a major regulator of vascular development. Whereas the activin receptor-like kinase (ALK)-1 isoform of TRI induces endothelial cell migration and proliferation, expression of the ALK-5 TRI isoform does the reverse (12). In our studies, we have found dramatic upregulation of the ALK-5 isoform of TRI in rat lungs subjected to hypoxia (Fig. 3), which coincided with an arrest of alveolarization with hypoxia. Therefore, it is clearly possible that such molecular processes governed by increased TGF-1 and ALK-5 activities are a major cause for the arrest of alveolarization observed in our model of hypoxic rat lungs.

Although such speculations are tempting, the exact molecular mechanisms underlying the hypoxia-induced arrest in alveolarization cannot entirely be deduced from these studies. Other methods are needed to answer these questions directly. Ideally, age-specific TGF-₁ overexpression, as well as selective modulation of lung epithelial cell TR expression by transgenic technologies, will help to address these questions in the future. At present, only the selective overexpression of soluble secreted proteins in the lung is a feasible approach to analyze the morphological effects of growth factors or cytokines (35). The presented findings, nevertheless, suggest an essential role for the TGF- system in the response of lung epithelial cells to chronic hypoxia at different developmental stages.

In this respect, manipulation of TGF- signaling and/or receptor expression may become a promising pharmacological target in diseases involving an arrest or interruption of alveolar development, such as BPD. Although hyperoxia rather than hypoxia has been implicated in the pathophysiology of BPD, similar morphologic characteristics (i.e., inhibition of alveolarization) have been reported on pathological examination of lung sections from BPD patients, and it remains to be elucidated whether early hypoxia indeed represents an initial trigger for disease onset (14, 17, 23). Whereas TGF- has been implicated in the pathogenesis of BPD before, a paucity of work exists in this field (28). Further investigations addressing specific regulatory mechanisms, as well as elucidation of the various downstream mediators of TGF- signaling, may help to address the different phenotypes associated with increased TGF- activity, both developmentally and with exposure to hypoxia.