Chronic lung disease in preterm lambs: effect of daily vitamin A treatment on alveolarization

Kurt H. Albertine, Mar Janna Dahl, Linda W. Gonzales, Zheng-ming Wang, Drew Metcalfe, Dallas M. Hyde, Charles G. Plopper, Barry C. Starcher, David P. Carlton, Richard D. Bland

Abstract

Neonatal chronic lung disease is characterized by failed formation of alveoli and capillaries, and excessive deposition of matrix elastin, which are linked to lengthy mechanical ventilation (MV) with O2-rich gas. Vitamin A supplementation has improved respiratory outcome of premature infants, but there is little information about the structural and molecular manifestations in the lung that occur with vitamin A treatment. We hypothesized that vitamin A supplementation during prolonged MV, without confounding by antenatal steroid treatment, would improve alveolar secondary septation, decrease thickness of the mesenchymal tissue cores between distal air space walls, and increase alveolar capillary growth. We further hypothesized that these structural advancements would be associated with modulated expression of tropoelastin and deposition of matrix elastin, phosphorylated Smad2 (pSmad2), cleaved caspase 3, proliferating cell nuclear antigen (PCNA), VEGF, VEGF-R2, and midkine in the parenchyma of the immature lung. Eight preterm lambs (125 days' gestation, term ∼150 days) were managed by MV for 3 wk: four were treated with daily intramuscular Aquasol A (vitamin A), 5,000 IU/kg, starting at birth; four received vehicle alone. Postmortem lung assays included quantitative RT-PCR and in situ hybridization, immunoblot and immunohistochemistry, and morphometry and stereology. Daily vitamin A supplementation increased alveolar secondary septation, decreased thickness of the mesenchymal tissue cores between the distal air space walls, and increased alveolar capillary growth. Associated molecular changes were less tropoelastin mRNA expression, matrix elastin deposition, pSmad2, and PCNA protein localization in the mesenchymal tissue core of the distal air space walls. On the other hand, mRNA expression and protein abundance of VEGF, VEGF-R2, midkine, and cleaved caspase 3 were increased. We conclude that vitamin A treatment partially improves lung development in chronically ventilated preterm neonates by modulating expression of tropoelastin, deposition of elastin, and expression of vascular growth factors.

  • bronchopulmonary dysplasia
  • VEGF
  • VEGF-R2
  • midkine
  • tropoelastin
  • elastin
  • TGF-β
  • Smad
  • lung growth and development
  • alveolar formation
  • alveolar capillary growth

neonatal chronic lung disease (CLD), a modified form of bronchopulmonary dysplasia, is histopathologically characterized by excessive, aberrant matrix elastin deposition and alveolar simplification in preterm neonates that are chronically managed by mechanical ventilation (MV) (4, 23). Excessive, aberrant matrix elastin deposition is due to excessive synthesis of tropoelastin expression (19, 47, 59). Alveolar simplification is manifest as reduced formation of alveolar secondary septa and capillaries (3, 24). Despite surfactant treatment and gentler ventilation modes, rates of neonatal CLD remain relatively constant (29, 32) or have increased (66, 67), and neonatal CLD is not restricted to the most severely ill preterm neonates (66, 67). Consequently, neonatal CLD continues to be a major cause of mortality and long-term morbidity in premature infants (75).

Several treatment approaches to potentially limit neonatal CLD show promise, including vitamin A, inhaled nitric oxide, and caffeine. Among these promising approaches, vitamin A supplementation reduces the need for oxygen (O2)-rich gas at 36 wk postmenstrual age in subsets of preterm neonates at risk of neonatal CLD (9, 72). The clinical rationale for vitamin A supplementation is that plasma and tissue concentrations of vitamin A are low in preterm infants who are at risk of developing neonatal CLD (17, 65). The biological rationale for vitamin A supplementation is that vitamin A and its downstream metabolites, such as all-trans retinoic acid, are morphogens that participate in lung development and injury repair (21, 48, 49). However, whether vitamin A supplementation improves structural formation of alveoli, with linkage to expression patterns of molecules that participate in that process in the lung of preterm infants with neonatal CLD, is not known.

Vitamin A supplementation may be anticipated to improve formation of alveolar secondary septa through its effects on expression and deposition of matrix elastin (20). During normal lung development, vitamin A modulates tropoelastin expression and matrix elastin deposition among cells of mesenchymal origin (51, 52). Following injury, vitamin A supplementation may be anticipated to reduce tropoelastin expression and deposition of matrix elastin in the lung of chronically ventilated preterm neonates because all-trans retinoic acid downregulates tropoelastin promoter activity following injury (39). Also, retinoic acid reduces transforming growth factor-β (TGF-β) signaling because retinoic acid decreases hyperoxia-induced TGF-β receptor expression and Smad abundance in the lung (57). However, effects of vitamin A supplementation on tropoelastin expression and matrix elastin deposition, as well as indices of TGF-β signaling as reflected by Smad expression, in the lung of preterm infants with neonatal CLD are incompletely understood.

Thinning of the mesenchymal tissue cores that separate adjacent distal air spaces in the immature lung is necessary for efficient gas exchange across the lung's parenchyma. During prolonged MV of preterm lambs, however, the mesenchymal cores remain cellular and thick, in association with the opposing consequences of reduced apoptosis and increased cell proliferation among mesenchymal cells (62). Because all-trans retinoic acid inhibits lung fibroblast proliferation (69), an anticipated outcome may be that vitamin A supplementation shifts the balance in favor of apoptosis of mesenchymal cells in the distal air space walls, thereby contributing to thinning the mesenchymal cores. Whether vitamin A supplementation shifts balance between apoptosis vs. proliferation among mesenchymal cells and reduces thickness in the distal air space walls of chronically ventilated preterm neonates is not known.

Vitamin A supplementation also may be anticipated to improve alveolar formation through its effects on expression of growth factors in the lung that participate in alveolar capillary growth. Expression of vascular endothelial growth factor (VEGF) and its functional receptor in the lung, VEGF receptor 2 (VEGF-R2; also called fetal liver kinase 1 or Flk-1), is reduced in the lung of chronically ventilated preterm infants (10), baboons (45), and lambs (15). These findings, along with the finding that blocking VEGF-R2 decreases pulmonary capillary growth and alveolar formation (35), have contributed to the “vascular hypothesis” of neonatal CLD (1). On the other hand, a rationale for supplemental vitamin A or all-trans retinoic acid is these retinoids increase VEGF and VEGF-R2 expression (22, 61, 64), as well as midkine expression (81), in the lung. However, the effects of vitamin A supplementation on alveolar capillary growth and expression of vascular growth factors in the lungs of preterm infants with neonatal CLD also are incompletely understood.

Our laboratory has reported that alveolar simplification occurs in the lung of preterm lambs managed by MV for 3 wk (3). At the time of delivery (∼125 days' gestation), the lungs are at the saccular stage of lung development, which is equivalent to ∼28-wk gestation in humans. At the end of 3 wk of MV, the lungs of the preterm lambs also have more expression of tropoelastin mRNA and aberrant deposition of matrix elastin, more expression of TGF-β, more cellular and thicker mesenchyme in the distal air space walls, and less expression of VEGF and VEGF-R2 and fewer alveolar capillaries (12, 15, 59, 62). As a next step, we gave supplemental vitamin A daily to preterm lambs during 3 wk of MV. We hypothesized that vitamin A supplementation during prolonged MV, without confounding by antenatal steroid treatment, would improve alveolar secondary septation, decrease cellularity and thickness of the mesenchymal tissue cores between distal air space walls, and increase alveolar capillary growth. We further hypothesized that these structural advancements would be associated with modulated expression of tropoelastin and deposition of matrix elastin, phosphorylated Smad2, cleaved caspase 3, proliferating cell nuclear antigen (PCNA), and VEGF, VEGF-R2, and midkine in the parenchyma of the immature lung.

METHODS

Study Groups

The protocols adhered to American Physiological Society/National Institutes of Health guidelines for humane use of animals for research, and were approved by the IACUC at the University of Utah, Health Sciences Center.

Two groups of four fetal lambs each were delivered prematurely at ∼125 days of gestation (term is ∼150 days) and mechanically ventilated for 3 wk. Before delivery, one group of four preterm lambs was prospectively assigned to receive vitamin A daily {Aquasol A palmitate, 5,000 IU·kg−1·day−1 (1.5 mg·kg−1·day−1) diluted in saline, intramuscularly [intramuscularly (im)]; Mayne Pharma USA, Paramus, NJ} beginning immediately after birth. The remaining four preterm lambs received vehicle (saline, im; identified as vehicle control). The timing of delivery was at the transition from the saccular to the alveolar stage of lung development (7, 27).

Two groups of lambs were included in this study to provide developmental perspective for the preterm lambs so that we could address the question of whether vitamin A treatment restored alveolar formation relative to normal gestation. One group of fetal lambs was delivered surgically at ∼125 days of gestation (designated as f125) to collect their unventilated lungs. This fetal lamb group served as the gestation-age reference for when the preterm lambs were delivered. The other group was born vaginally after spontaneous labor at term gestation and breathed spontaneously for 4–8 h (designated as term) before their lungs were collected. That group served as the gestation-age reference for the preterm lambs, had they not been delivered prematurely 3 wk earlier.

The vehicle control and fetal and term gestation-age reference lambs were included in a recent report (15). However, different tissue samples were analyzed and different analytical methods were used for the present study. All of the lamb studies were done within a 15-mo period at the University of Utah.

Preparation of Chronically Ventilated Preterm Lambs

The experimental animal preparation has been reported previously (3, 1215, 43, 59). None of the ewes were treated with antenatal glucocorticoids. All of the preterm lambs were treated with surfactant (Infasurf, 10 ml of 35 mg/ml; generous gift of ONY, Amherst, NY) just before operative delivery. After an initial period of stabilization, usually ∼24 h after birth, the preterm lambs were managed by MV, using a time-cycled, pressure-limited infant respirator (model IV-100B; Sechrist, Anaheim, CA) set at a rate of 20 breaths/min, inspiratory time 0.75 s, and end-expiratory pressure ∼5 cmH2O. Peak inflation pressure was adjusted to keep PaCO2 between 35 and 45 mmHg, and the fraction of inspired oxygen was adjusted to keep PaO2 between 50 and 90 mmHg. The use of a slow breath rate, long inspiratory time, and relatively low target level of PaCO2, was to induce prolonged cyclic lung stretch with O2-rich gas and thereby create neonatal CLD. All preterm lambs received buprenorphine, 0.01 mg/kg iv, every 4 h to prevent agitation and either pentobarbital sodium (Vet Lab, Lenexa, KS), 3–5 mg/kg, or phenobarbital sodium (Wyeth Laboratories, Philadelphia, PA), 10 mg/kg, as needed for sedation during MV. The ductus arteriosus was surgically ligated within 2 days of birth. Daily management and monitoring protocols have been described (3, 1215, 43, 59). Plasma retinol concentration was determined by high-pressure liquid chromatography (11). This measurement was made at birth and at weekly intervals thereafter. After 3 wk of MV, the preterm lambs received an intravenous injection of pentobarbital sodium (35 mg/kg), after which their chest was opened to remove the lungs for postmortem analyses.

Postmortem Analyses

Lung tissue preparation.

After 3 wk of MV of the preterm lambs, and at the time of birth of the fetal gestation-age reference lambs, an intravenous injection of pentobarbital sodium (35 mg/kg) was given, after which the chest was opened to remove the lungs for postmortem analyses. We prepared the lungs of all lambs for 1) molecular and biochemical analyses and 2) morphometric and stereological analyses, according to methods that we have described (3, 5, 6, 12, 15, 43, 62).

For the molecular and biochemical analyses, peripheral tissue (devoid of visceral pleura and central airways/arteries) from the right caudal and middle lobes was snap-frozen in liquid nitrogen and stored at −80°C. RT-PCR, immunoblot, and radioimmunoassay were performed on homogenates of the frozen peripheral lung tissue. For the morphological analyses, the following lobes were sampled by design-based methods: 1) right middle lobe, 2) left lingula, 3) left caudal lobe, and 4) cardiac lobe. The lobes and left lingula were double-clamped with peripheral vascular clamps at end inspiration, while the lungs were in the chest and ventilated at the same pressures that were used to manage each preterm lamb. The clamped, excised right middle lobe was immersed in Carnoy's fixative for 24 h (4°C). The right middle lobe was used to quantify alveolar secondary septation. The excised left lingula had its lobar artery cannulated for perfusion of 10% buffered neutral formalin (DNase- and RNase-free). The bronchus also was cannulated to insufflate the same fixative (25 cmH2O static pressure). The hilum of the left lingula was clamped to retain fixative in the vasculature and airways before immersion in formalin overnight (4°C). The left lingula was used to quantify capillary surface density and small vessel number, by stereological analysis of immunostained tissue sections. The left caudal lobe was insufflated with 10% buffered neutral formalin (25 cmH2O static pressure) via its lobar bronchus and immersed in the same fixative overnight (4°C). That lobe was used to localize VEGF-A, VEGF-R2, and tropoelastin mRNAs by in situ hybridization, and midkine, cleaved caspase 3, PCNA, pSmad2 proteins, and elastin. The cardiac lobe was injected, using a 27-gauge needle, with 2.5% glutaraldehyde/1% paraformaldehyde in cacodylate buffer before immersion in the same fixative (4°C). The cardiac lobe was used for transmission electron microscopy.

Quantitative real-time RT-PCR.

Peripheral, whole-lung tissue homogenates were used for analysis of VEGF, VEGF-R2, and tropoelastin mRNA expression, using ovine-specific primers and probes (Supplemental Table 1) (32–34).

In situ hybridization.

Formalin-fixed, paraffin-embedded tissue sections were used for in situ hybridization. Ovine-specific riboprobes were made for in situ hybridization of VEGF-A, VEGF-R2, and tropoelastin, using standard methods (38). Supplemental data for this article is available online at the AJP-Lung web site.

Immunoblot.

Homogenates of peripheral, whole-lung frozen samples were used to quantify the relative abundance of VEGF-A165/122 (hereafter designated VEGF) and VEGF-R2 proteins by standard methods (14, 15, 43, 58, 62).

Immunohistochemistry.

We used standard methods to localize VEGF and VEGF-R2, using the same antibodies that we used for immunoblot, colocalized VEGF and surfactant apoprotein B, PECAM-1, midkine, cleaved caspase 3, PCNA, phosphorylated Smad2 (pSmad2, Ser465/467), and elastin (5, 14, 15, 31, 43, 58, 62).

Radioimmunoassy.

Analysis of desmosine content, an indicator of mature, cross-linked elastin, was performed on hydrolyzed samples, using a competitive radioimmunoassay that has been described previously (37, 59).

Morphometry and stereology.

We used design-based sampling (16, 33) to quantify lung parenchymal structure, using morphometric and stereological methods previously reported by our group (3, 6, 12, 13, 43, 62). We followed the principles of systematic, uniform, random sampling to evaluate non-overlapping calibrated fields. Tissue sections were analyzed without knowledge of the lamb group from which the tissue was taken. We quantified alveolar secondary septation of distal air space walls by three independent methods: 1) radial alveolar count (RAC), 2) volume density of secondary septa (Vv2′ septa), and 3) mean face length (2, 3, 28, 34). We used histochemistry (Hart's elastic fiber stain) to determine the volume density of matrix elastin in the distal air space walls. We also used immunohistochemistry (PECAM) combined with stereology to determine the surface density of capillaries (Svcap) and epithelium (Svepi) in the distal air space walls, and the number of pre- and postcapillary microvessels.

Data Analysis

Data are reported as means ± 1 SD. We used analysis of variance followed by Fisher's PLSD nonparametric test for statistical analyses, using a commercial computer application (StatView5; SAS Institute, Cary, NC). Statistical significance was accepted as P < 0.05 (80).

RESULTS

Characteristics of the Two Groups of Preterm Lambs Managed by MV for 3 Wk

Gestational age (vehicle controls, 124 ± 4 days; vitamin A treated, 124 ± 4 days), birth weight (2.77 ± 0.46 kg vs. 2.49 ± 0.44, respectively), and weight at death (3.59 ± 0.82 kg vs. 3.11 ± 0.87, respectively) were not significantly different between the two groups of preterm lambs. Plasma retinol concentrations at the time of birth and at death are shown in Fig. 1. At the time of birth, plasma retinol concentration was the same for both groups of preterm lambs. Daily vitamin A treatment for 3 wk significantly increased plasma retinol concentration compared with vehicle controls (P < 0.05). Plasma retinol concentration in the vitamin A-treated preterm lambs was also greater than the reference lambs that were born at term gestation (P < 0.05) and equal to that in adult ewes (24 ± 5 mg/dl; n = 3). During the course of the 3-wk study, mean arterial blood pressure was >40 mmHg and was not different between the groups.

Fig. 1.

Plasma retinol concentration. Daily vitamin A treatment (+ Vitamin A) of preterm lambs during 3 wk of mechanical ventilation (MV) increased plasma retinol concentration compared with vehicle controls (*P < 0.05; means ± SD; n = 4/group). Plasma retinol concentration in the vitamin A-treated preterm lambs was equivalent to that in lambs that were born at term gestation (Term). Term lambs served as the gestation age-matched reference for the 3-wk ventilation studies. Plasma retinol concentration in the vehicle controls was low at birth (d0) and remained low at the conclusion of the 3-wk study. That group's plasma retinol concentration is equal to that of fetal lambs that were delivered at 125 days of gestation (f125). Those fetal lambs served as the gestation age-matched reference at the time the preterm lambs were delivered. †Significant difference compared with the paired plasma concentration for the same group at d0 of life, P < 0.05. ‡Significant difference compared with the f125 reference group, P < 0.05.

Alveolar Secondary Septation

Histologically, terminal respiratory units (TRU) of the vitamin A-treated preterm lambs appeared to have more alveoli compared with the vehicle controls, which uniformly had alveolar simplification (Fig. 2, A and B). However, TRU architecture in the vitamin A-treated preterm lambs was heterogeneous, with a mixture of alveolar simplification and appropriate alveolar formation. In the lung of the vitamin A-treated preterm lambs, proximal parts of TRUs, that is, respiratory bronchioles and initial part of alveolar ducts, were distended, secondary septa were stunted when present, and distal air space walls were thick and cellular, much like the architecture of TRUs in the vehicle controls. By comparison, the distal parts of the same TRUs in the vitamin A-treated preterm lambs had more and longer alveolar secondary septa, and thinner, less cellular distal air space walls.

Fig. 2.

Terminal respiratory unit (TRU) architecture. A preterm lamb treated daily with vitamin A (+ Vitamin A; B) during 3 wk of MV has TRU architecture that is heterogeneous. The proximal parts (respiratory bronchioles and initial segments of alveolar ducts) have alveolar simplification (*). That is, the air spaces are distended, alveolar secondary septa are infrequent, and those that are present are stunted, and the air space walls are thick and cellular. The distal parts of the same TRU have air spaces that are small, alveolar secondary septa that are numerous, long, and thin (arrows), and air space walls that are thin and less cellular (arrowhead). The TRU in the vehicle control (− Vitamin A; A) has uniformly distended air spaces, few alveolar secondary septa, and those that are present are short and thick (arrows), and distal air space walls are thick and cellular (arrowhead). A and B are the same magnification. Quantitative histology (C–E) showed that radial alveolar count (C) and volume density of alveolar secondary septa (Vv2′ septa; D) are greater in vitamin A-supplemented preterm lambs compared with vehicle controls (*P < 0.05; means ± SD; n = 4/group). However, radial alveolar count and Vv2′ septa in the vitamin A-treated preterm lambs is less than in lambs born at term gestation (Term; ◊P < 0.05). Term lambs served as the gestation age-matched reference for the 3-wk ventilation studies. Mean face length (E), on the other hand, is shorter in the vitamin A-treated preterm lambs than in the vehicle controls (*P < 0.05; means ± SD; n = 4/group); however, mean face length is longer than in lambs born at term gestation (Term; ◊P < 0.05). All 3 structural indices of alveolar formation in the vehicle controls are the same as in f125 reference group. ‡Significant difference compared with the f125 reference group, P < 0.05. ◊Significant difference compared with both preterm groups, P < 0.05.

We used three independent morphometric/stereological approaches to quantitatively compare alveolar secondary septation between the vitamin A-treated and vehicle controls, and between both preterm groups and term lambs. The three approaches were RAC, Vv2′ septa, and mean face length. RAC and Vv2′ septa should become larger, whereas mean face length should become shorter, as the saccules are subdivided into anatomic alveoli by formation of new alveolar secondary septa. These analyses showed that RAC and Vv2′ septa were significantly greater in the lung of vitamin A-treated preterm lambs compared with vehicle controls (Fig. 2, C and D; P < 0.05). By comparison, mean face length was significantly shorter in vitamin A-treated preterm lambs compared with vehicle controls (Fig. 2E; P < 0.05). These analyses also showed that both groups of preterm lambs had less alveolar secondary septation (Fig. 2, C–E) than term gestation-age reference lambs had (P < 0.05).

Tropoelastin mRNA Expression, Elastin Abundance, and Matrix Elastin Localization

We used fluorescence in situ hybridization and histochemical staining methods to topographically assess differences in lung tropoelastin mRNA expression and matrix elastin deposition, respectively, between the two groups of preterm lambs. In situ hybridization revealed that the tips of developing alveolar secondary septa had less topographic expression of tropoelastin mRNA in the lung of vitamin A-treated preterm lambs compared with vehicle controls (Fig. 3, A–C). Histochemical staining, using Hart's elastic fiber stain, also revealed that deposition of matrix elastic fibers was less in the lung vitamin A-treated preterm lambs compared with vehicle controls (Fig. 3, D and E). Based on these topographic observations, we quantified tropoelastin expression and elastin abundance. These analyses showed that lung tropoelastin mRNA expression (Fig. 3F), desmosine content (Fig. 3G), and parenchymal elastic fiber volume density (Fig. 3H) were significantly less in vitamin A-treated preterm lambs than vehicle controls (P < 0.05). These analyses also showed that both groups of preterm lambs had significantly more tropoelastin mRNA expression (Fig. 3F), desmosine content (Fig. 3G), and parenchymal elastic fiber volume density (Fig. 3H) in their lungs than the term gestation-age reference lambs (P < 0.05).

Fig. 3.

Elastin expression and deposition. A preterm lamb treated daily with vitamin A (+ Vitamin A; B) during 3 wk of MV has less tropoelastin mRNA localization in the lung parenchyma by in situ hybridization (green in B) and matrix elastin deposition revealed by Hart's stain (arrows in E) than a vehicle control (− Vitamin A; A and D, respectively). C is the sense control for in situ hybridization (blue fluorescence is DAPI-stained nuclei). No green immunofluorescence is visible, indicating specificity of the antisense probe used for tropoelastin. A–C are the same magnification, as are D and E. Daily vitamin A treatment quantitatively stabilized tropoelastin mRNA expression (F), desmosine content (G), and parenchymal elastic fiber volume density (H) compared with vehicle controls (− Vitamin A). However, expression and protein abundance in the vitamin A-treated preterm lambs is greater than in lambs born at term gestation (Term; ◊P < 0.05). Term lambs served as the gestation age-matched reference for the 3-wk ventilation studies. ‡Significant difference compared with the f125 reference group, P < 0.05. ◊Significant difference compared with both preterm groups, P < 0.05.

VEGF and VEGF-R2 mRNA Expression and Protein Abundance and Immunolocalization

We semiquantitatively compared VEGF and VEGF-R2 mRNA expression and protein abundance between the vitamin A-treated and vehicle control preterm lambs and between both groups of preterm lambs and term gestation-age reference lambs. VEGF-A and VEGF-R2 mRNA expression (Fig. 4, A and B) and protein abundance (Fig. 4, C and D) were significantly greater in the lungs of vitamin A-treated preterm lambs than vehicle controls (P < 0.05). For the immunoblot analysis, the anti-VEGF165/122 antibody revealed two protein bands (46-kDa [VEGF165] and 28-kDa [VEGF122]; Fig. 4C). The immunoblot analysis for VEGF-R2 protein showed one protein band (200 kDa; Fig. 4D). These semiquantitative densitometric analyses also showed that both groups of preterm lambs had significantly less VEGF and VEGF-R2 mRNA expression (Fig. 4, A and B) and protein abundance (Fig. 4, C and D) than term gestation-age reference lambs had (P < 0.05).

Fig. 4.

VEGF and VEGF-R2 expression. Daily vitamin A treatment (+ Vitamin A) of preterm lambs during 3 wk of MV increased VEGF-A and VEGF-R2 mRNA expression (A and B, respectively) and protein abundance (C and D, respectively) compared with vehicle (− Vitamin A) preterm controls (*P < 0.05; means ± SD; n = 4/group). However, mRNA expression and protein abundance in the vitamin A-treated preterm lambs is less than in lambs born at term gestation (Term; ◊P < 0.05). Term lambs served as the gestation age-matched reference for the 3-wk ventilation studies. VEGF and VEGF-R2 mRNA expression and protein abundance in the vehicle controls are the lowest among the groups. VEGF-A and VEGF-R2 mRNA expression is normalized for GAPDH mRNA expression, whereas abundance of the respective proteins is normalized for endogenous protein revealed by MemCode reversible protein stain. ‡Significant difference compared with the f125 reference group, P < 0.05. ◊Significant difference compared with both preterm groups, P < 0.05.

We also used fluorescence in situ hybridization and immunohistochemistry to localize VEGF and VEGF-R2 mRNAs and proteins, respectively, in lung (Fig. 5). VEGF-A mRNA was expressed in cuboidal epithelial cells (Fig. 5, A–C). VEGF-A mRNA expression was also present in distal airway epithelial cells (not shown). VEGF-R2 mRNA was expressed in mesenchymal (interstitial) cells located in the distal air space walls (Fig. 5, D–F). Endothelium of extra-alveolar blood vessels were labeled, too (not shown). Specificity of the localization was provided by the sense control results, which had no immunofluorescence. The topographic distribution of both mRNAs was recapitulated by localization of the corresponding proteins, using immunohistochemistry. VEGF protein was localized in cuboidal epithelial cells that were dispersed along the distal air space walls, especially at their corners (Fig. 5, G and H). VEGF protein also was immunolocalized in distal airway epithelial cells (not shown). Double immunofluorescence showed that the cuboidal cells in which VEGF protein was immunolocalized were alveolar type II epithelial cells because SP-B protein colocalized with VEGF protein (Fig. 5, I–L). VEGF-R2 protein was immunolocalized in capillary endothelial cells and other cells of mesenchymal origin such as vascular smooth muscle cells (Fig. 5, M and N). Some alveolar type II epithelial cells and distal airway epithelial cells also were immunostained for VEGF-R2 protein.

Fig. 5.

Localization of VEGF and VEGF-R2. A preterm lamb treated daily with vitamin A (+ Vitamin A) during 3 wk of MV has more VEGF-A and VEGF-R2 mRNA expression visible (red fluorescence in B and E, respectively) compared with a vehicle control (− Vitamin A; A and D, respectively). C and F are the sense controls for in situ hybridization (blue fluorescence is DAPI-stained nuclei). No red immunofluorescence is visible, indicating specificity of the antisense probe used for VEGF-A and VEGF-R2, respectively. A–F are the same magnification. In lung tissue from the same preterm lamb treated daily with vitamin A, more VEGF and VEGF-R2 protein localization is visible (brown immunostain in H and N, respectively) compared with the vehicle control (G and M, respectively). G, H, M, and N are the same magnification. VEGF protein that is immunolocalized in domed epithelial cells in G and H (arrows) is specifically localized in alveolar type II epithelial cells (yellow fluorescence in I and J) because VEGF protein (red fluorescence in K) colocalized with SP-B protein (green fluorescence in L). I and J are the same magnification. VEGF-R2 protein is immunolocalized in capillary endothelial cells (arrowhead in N) and other cells such as vascular smooth muscle cells and epithelial cells (arrows of M and N).

Capillary and Extra-Alveolar Microvessel Formation

The observations that mRNA expression and protein abundance of VEGF and VEGF-R2 were greater in the lungs of vitamin A-treated preterm lambs compared with vehicle controls raised the possibility that blood vessel growth might be enhanced in the vitamin A-treated preterm lambs. We used two approaches to assess that possibility. One approach was transmission electron microscopy, which showed that capillaries traversed the length of many, but not all, alveolar secondary septa in the vitamin A-treated preterm lambs (Fig. 6B). By comparison, capillaries were absent from the stunted alveolar secondary septa in the vehicle controls (Fig. 6A). The other approach used morphometry to determine if vitamin A treatment quantitatively increased lung capillary and microvessel formation in preterm lambs. We used immunohistochemistry combined with stereology to estimate two blood vessel parameters: 1) surface density of capillaries per surface density of air space epithelium (Svcap/Svepi) and 2) the number of pre- and postcapillary microvessels (>20 but <100 μm in diameter) per 100 parenchymal tissue points. Immunohistochemistry was performed with an antibody directed against PECAM-1 protein to mark vascular endothelial cells (Fig. 6, C and D). Stereology showed that the lungs of vitamin A-treated preterm lambs had significantly greater Svcap (Fig. 6E) and more extra-alveolar microvessels (Fig. 6G) compared with vehicle controls (P < 0.05) for equivalent epithelial-cell surface density. We measured the same amount of epithelial-cell surface density, which we used as the reference space, because alveolar secondary septation was not equal between the two groups of preterm lambs. By keeping the reference space the same, the stereological analysis was optimized to identify if a difference in capillary surface density occurred between the two groups of preterm lambs. In addition, these analyses showed that both groups of preterm lambs had significantly lower Svcap (Fig. 6, E and F) and fewer extra-alveolar microvessels (Fig. 6G) than the term gestation-age reference lambs (P < 0.05).

Fig. 6.

Alveolar secondary septal ultrastructure and capillary localization. A preterm lamb treated daily with vitamin A (+ Vitamin A; B) during 3 wk of MV has an alveolar secondary septum (Sec Septum) that is longer, thinner, and has capillaries (C) along its length compared with the vehicle control (A; shown at the same magnification as B). Alv Wall, alveolar wall; E, elastin; M, mesenchymal cell. Immunohistochemical localization of PECAM-1 protein highlights endothelial cells with brown immunostain (arrows in C and D; shown at the same magnification). Daily vitamin A treatment quantitatively increased capillary surface density (E), referenced to equal epithelial cells surface density (F), and extra-alveolar microvessel number/100 parenchymal points (>20 but <100 μm in diameter; G) compared with vehicle controls (− Vitamin A; *P < 0.05; means ± SD; n = 4/group). However, capillary surface density and microvessel number in the vitamin A-treated preterm lambs is less than in lambs born at term gestation (Term; ◊P < 0.05). Term lambs served as the gestation age-matched reference for the 3-wk ventilation studies. ‡Significant difference compared with the f125 reference group, P < 0.05. ◊Significant difference compared with both preterm groups, P < 0.05.

Midkine, Cleaved Caspase 3, PCNA, and pSmad2 Immunolocalization

The differences in lung parenchymal architecture between the two groups of preterm lambs (Figs. 26) are likely influenced by a number of molecules, in addition to VEGF, VEGF-R2, tropoelastin, and matrix elastin. Therefore, we also assessed topographic localization of midkine, cleaved caspase 3, PCNA, and pSmad2 (Fig. 7). Midkine and cleaved caspase 3 proteins were immunolocalized in the cytoplasm and nuclei, respectively, of mesenchymal cells in the walls of distal air spaces. Brown immunostain was most evident in the vitamin A-treated group of preterm lambs and least evident in the vehicle-treated controls. Conversely, PCNA and pSmad2 proteins were immunostained in the nuclei and cytoplasm, respectively, of mesenchymal cells in the walls of distal air spaces. For PCNA and pSmad2, brown immunostain was least evident in the vitamin A-treated group of preterm lambs and most evident in the vehicle-treated controls.

Fig. 7.

Localization of midkine, cleaved caspase 3, PCNA, and pSmad2. A preterm lamb treated daily with vitamin A (+ Vitamin A) during 3 wk of MV has more midkine and cleaved caspase 3 protein localization visible among mesenchymal cells (brown stain in B and E, respectively) compared with a vehicle control (− Vitamin A; A and D, respectively). Conversely, lung tissue sections from the vitamin A-treated preterm lamb have less immunostain for PCNA and pSmad2 among mesenchymal cells (brown stain in H and K, respectively) compared with the vehicle control (G and J, respectively). Immunostained cells in the mesenchyme of the distal air space (DAS) walls are labeled with small, thin arrows. The bold arrow in the panels with an inset image identifies the region illustrated at greater magnification in the corresponding inset image. All panels are the same magnification. All inset images are the same magnification (twice the magnification of the panels). The column of images on the right shows immunohistochemical staining results for a term newborn lamb, as a gestation age-matched reference for the preterm lambs. Immunolocalization patterns for midkine (C), cleaved caspase 3 (F), PCNA (I), and pSmad2 (L) are most similar to the preterm lamb treated daily with vitamin A during 3 wk of MV.

Respiratory Gas Exchange For the Two Groups of Preterm Lambs Managed By MV For 3 Wk

Respiratory gas exchange variables for the two groups of preterm lambs that were ventilated for 3 wk are shown in Table 1. The fraction of inspired oxygen, PaO2, PaCO2, arterial pH, peak inspiratory pressure, mean airway pressure, and tidal volume were not significantly different between the two preterm groups at the beginning or end of the 3-wk ventilation period.

View this table:
Table 1.

Summary of physiological data for preterm lambs treated with vehicle or vitamin A

DISCUSSION

Vitamin A supplementation reduces the need for O2-rich gas at 36-wk postmenstrual age in subsets of preterm neonates at risk of neonatal CLD (9, 72). While the effect of supplemental vitamin A may be anticipated to improve alveolar formation that is associated with modulated expression of molecules that regulate alveolar formation, structural and molecular effects have not been well characterized in the lungs of preterm infants who are given supplemental vitamin A during prolonged MV. Our results demonstrate that daily vitamin A supplementation increases alveolar secondary septation, decreases thickness of the mesenchymal tissue cores between the distal air space walls, and increases alveolar capillary growth. These structural advancements are associated with less tropoelastin mRNA expression, less matrix elastin deposition, and less pSmad2 and PCNA localization in the mesenchymal tissue core of the distal air space walls. On the other hand, these structural advancements are associated with more topographic presence of cleaved caspase 3 among mesenchymal cells in the distal air space walls, as well as more mRNA expression and protein abundance of VEGF, VEGF-R2, and midkine. Therefore, our study provides structural and molecular insights into some of the diverse biological effects of supplemental vitamin A on alveolar formation in the setting of prolonged MV.

During normal alveolar formation, expression of tropoelastin and the subsequent deposition of matrix elastin are necessary for alveolar secondary septation (18, 41, 51, 52, 76). Vitamin A modulates tropoelastin expression and matrix elastin deposition among cells of mesenchymal origin during normal lung development (51, 52). On the other hand, the histopathology of neonatal CLD is characterized by upregulation of tropoelastin expression, as well as aberrant and excessive deposition of matrix elastin (3, 12, 15, 19, 24, 45, 47, 50, 59, 71). Our study shows that vitamin A supplementation is associated with decreased expression of tropoelastin and deposition of matrix elastin, as well as increased secondary septation of distal air spaces. These beneficial outcomes are different from the outcomes following vitamin A supplementation of chronically ventilated preterm baboons (60). For both studies, the preterm neonates were treated with the same vitamin A compound and dosage (Aquasol; 5,000 IU·kg−1·d−1 im). Reasons for the differing results is unclear. However, one reason may be that our preterm lambs were not exposed to antenatal corticosteroids, whereas the preterm baboons were exposed to antenatal betamethasone. Because we did not treat with antenatal corticosteroids, our study was not confounded by actions of antenatal corticosteroids, such as increased expression of tropoelastin (42). Another reason may be the duration of experiments: our preterm lamb studies were 3 wk, whereas the preterm baboon studies were 2 wk. The longer study period that we used may be necessary to detect molecular and structural manifestations of vitamin A supplementation during MV.

A molecular mechanism by which tropoelastin expression and matrix elastin deposition increase in the lung is upregulation of TGF-β signaling (8, 53, 63). Increased tropoelastin expression and matrix elastin deposition inhibit alveolar secondary septation in the lung (30, 74), a histological characteristic of neonatal CLD. We recently showed that TGF-β signaling is increased in the lung of preterm lambs that are managed by MV for 3 wk (15). The results of the present study suggest that vitamin A supplementation may decrease TGF-β signaling in the lung, based on less immunohistochemical localization of pSmad2 protein among mesenchymal cells in the distal air space walls. In this regard, pSmad2 has been used to assess TGF-β activation in the lung of neonatal rats exposed to high tidal volume ventilation (77) and neonatal mice managed by MV for 24 h (55). Because all-trans retinoic acid downregulates Smad signaling in alveolar type II cells (57), it is reasonable to suggest that an effect of supplemental vitamin A in the preterm lambs may be diminished TGF-β/Smad signaling, resulting in reduced expression of tropoelastin and deposition of matrix elastin in the distal air space walls.

Our observation that daily vitamin A supplementation during prolonged MV is associated with less pSmad2 protein localization in mesenchymal cells may have implications about the pathogenic mechanisms leading to cellular, thick mesenchyme in the distal air space walls, which is another histological characteristic of neonatal CLD. We recently showed that this phenotypic characteristic is due to imbalance between apoptosis and cell proliferation among mesenchymal cells in the distal air space walls when preterm lambs are managed by MV (62). During prolonged MV, mesenchymal cell apoptosis is reduced, while cell proliferation is enhanced. Coincidently, the distal air space walls remained thick. In the present study, supplemental vitamin A modulated that imbalance, such that mesenchymal cells in the distal air space walls had more immunohistochemical evidence of apoptosis and less immunohistochemical evidence of cell proliferation. Less immunohistochemical evidence of mesenchymal cell proliferation in the vitamin A-treated preterm lambs is consistent with the observation that all-trans retinoic acid inhibits cell proliferation of lung fibroblasts (69). We propose that a mechanism by which vitamin A leads to less cellular and thinner distal air space wall structure may be downregulation of TGF-β/Smad signaling.

The vascular hypothesis of neonatal CLD proposes that inhibition of pulmonary capillary growth directly impairs alveolar formation (1). This hypothesis is supported by the observation that alveolar capillary growth is reduced in neonatal rats by blocking vascular growth factor signaling, using anti-angiogenic drugs such as VEGF-R2 blocker (SU-5146) or thalidomide (35). Additionally, alveolar capillary growth is decreased by blocking PECAM-1-mediated endothelial cell migration or in PECAM-1-null mice (26). Reduced capillary growth in each of these studies is associated with decreased alveolar formation. These results are consistent with the lung phenotype of neonatal CLD, in which capillary growth is reduced along with diminished expression of vascular growth factors (10). Downregulation of VEGF-R2 expression also occurs when TGF-β signaling is increased (44, 54). The results of the present study show that vitamin A supplementation increases expression of VEGF, VEGF-R2, and midkine among epithelial and mesenchymal cells in the lung parenchyma. Because all-trans retinoic acid upregulates endothelial nitric oxide synthase expression (73), it is possible that nitric oxide contributed to some of the molecular and structural improvements in capillary growth that occurred in the vitamin A-treated preterm lambs in the present study. We immunolocalized midkine protein because it has retinoic acid response regions in its promoter region (56, 70), resulting in direct upregulation of midkine (36, 40, 81). Our results also show that the increased expression of vascular growth factors in the vitamin A-treated preterm lambs is associated with more capillary growth. We suggest, therefore, that another beneficial effect of supplemental vitamin A is increased vascular growth in the lung.

While our study shows that vitamin A supplementation to chronically ventilated preterm lambs modulates expression of genes involved in structural development of the lung and improves formation of alveolar secondary septa and capillaries, the molecular and structural effects are not uniform across the lung or complete. Because these effects are heterogeneous and incomplete, it may not be unexpected that we did not find improved respiratory gas exchange in the vitamin A-treated preterm lambs compared with the vehicle controls. In this regard, limited lung functional improvement also is reported in studies of compensatory lung growth in dogs following pneumonectomy and daily retinoic acid treatment (25, 78, 79). The remaining lung had persistent distal air space wall thickening. Instability of the very premature lambs over the prolonged period in the intensive care setting contributed to considerably variable respiratory and cardiovascular measurements at the end of 3 wk of management in the vehicle control group. This could have contributed to the absence of statistical differences in respiratory gas exchange between groups. Another contributing factor may be that the preterm lambs received a relatively large tidal volume (∼12 ml/kg body wt) to reach the ventilation target. The large tidal volume likely contributed to the heterogeneous architecture of the terminal respiratory units in the vitamin A-treated preterm lambs. Because respiratory gas exchange reflects the sum of gas diffusion across terminal respiratory units (68), we speculate that the heterogeneous architecture may have precluded detecting improvement in respiratory gas exchange in the vitamin A-treated group compared with the vehicle control group. While we did not detect statistical differences, trends are evident toward better oxygenation and ventilation for the vitamin A group at 3 wk of treatment compared with the vehicle controls at the same endpoint (Table 1). It is possible that smaller applied tidal volume may result in more homogeneous architecture and thus association with respiratory gas exchange. New studies in our laboratory are testing this possibility.

In summary, the contribution of our study is it shows that daily vitamin A treatment during MV for 3 wk modulates expression of VEGF, VEGF-R2, midkine, tropoelastin, cleaved caspase 3, PCNA, and pSmad2 in the lung of chronically ventilated preterm lambs compared with vehicle controls. These molecular influences are associated with improved alveolar formation and capillary growth, although the improvements are heterogeneous in their distribution within terminal respiratory units. We conclude that vitamin A treatment partially improves lung development in chronically ventilated preterm neonates by modulating expression of tropoelastin, deposition of elastin, and expression of vascular growth factors. We speculate that treatment approaches that further enhance these positive effects may lead to more complete alveolar formation and capillary growth such that respiratory gas exchange will be improved.

GRANTS

This work was supported by National Institutes of Health Grants R01-HL-062875 (K. H. Albertine), R01-HL-62512 (R. D. Bland), P01-HL-056401 (Specialized Center of Research, Project V; R. D. Bland), R01-HL-086631 (R. D. Bland), March of Dimes Birth Defects Foundation Grant 6-FY97-0138 (R. D. Bland), T35-HL-07744 (K. H. Albertine), P01-ES-000628 and NCRR RR-00169 (D. M. Hyde).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

ACKNOWLEDGMENTS

We thank Phil Clair, Melinda Williams, Nancy Chandler, Jiancheng Sun, and Li Dong for technical assistance. In addition, we acknowledge the help of numerous people who are not listed as authors, including the many medical students who worked tirelessly to accomplish the 21-day studies. Three of the medical students (Jeffrey Gardner, Seth Spanos, Niloufar Tabatabaei) were supported by Research Training Grant T35-HL-07744.

Present address of D. P. Carlton: Dept. of Pediatrics, Emory University, Atlanta, GA 30322; present address of R. D. Bland: Dept. of Pediatrics, Stanford University, Palo Alto, CA 94305.

REFERENCES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
View Abstract