Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

There is growing evidence that retinoids, and in particular all-trans retinoic acid (tRA), are involved in the development of both the conducting airway and alveolar portions of the lungs (1). Retinoids are a family of molecules whose biologically active members are collectively called "vitamin A." Most cells acquire retinol from plasma, although a few organs like the liver, kidney, and lungs contain retinyl ester hydrolases and can presumably acquire retinol from the hydrolysis of endogenous retinyl esters (2). Before interacting with retinoid responsive genes, retinol is usually converted to tRA. Massaro and Massaro (3, 4) have shown that tRA increases the number of alveoli in neonatal rats and restores alveolar number and alveolar surface area in an animal model of emphysema. These data suggest that tRA enhances alveolar formation, but the mechanisms responsible for this enhancement have not been elucidated.

Alveolar formation involves a coordinated increase in epithelial and mesenchymal cells, the protrusion of primary and secondary alveolar septa into the air spaces, and the formation of an extracellular connective tissue framework for the newly formed septa (5). Elastic fibers are a critical element of this supporting framework and are essential to the alveolar septation process (6). More recently, Boström and associates (7) have shown that the platelet-derived growth factor-A null deletion in mice results in a marked reduction in the number of -smooth muscle actin containing alveolar septal cells that produce elastin at Postnatal Day 10 and older. The marked diminution of elastin-producing interstital cells was accompanied by a failure of the formation of secondary septa. This resulted in markedly dilated distal air sacs at Postnatal Day 19. Elastin is a hydrophobic protein that confers elastic properties to the lung and reduces the mechanical energy required for respiration (8). Elastin is synthesized as a soluble monomer, tropoelastin (TE), and is secreted into the extracellular space where it is covalently cross-linked to other TE molecules, in the presence of various fibrillar glycoproteins, to form insoluble elastic fibers (8).

Endogenous retinoids increase lung elastin gene expression during Gestational Days 19 through 22 in the rat lung when retinyl esters are most abundant and may be required for concurrent saccular development (9). We have previously shown that exogenous tRA increases elastin gene expression by a subpopulation of lipid-laden, myofibroblast-like lung fibroblasts in vitro (12, 13). These fibroblasts are elastin-producing lung fibroblasts that accumulate lipid and are abundant during the first two postnatal weeks in mice (14). Because the development of the elastic fiber network in the pulmonary interstitium occurs hand-in-hand with alveolar morphogenesis, we hypothesized that the stimulation of pulmonary alveolar formation by tRA is accompanied by an increase in elastin gene expression. To test this hypothesis, we examined elastin gene expression, elastic fiber formation, and alveolar number and surface area in mice whose endogenous retinoid signaling pathway had been disrupted by deletions of specific retinoic acid receptor (RAR) genes.

Retinoids alter gene expression by binding to RARs, which bind tRA or its isomer 9-cis-RA, and to retinoid X receptors (RXR), which bind only 9-cis-RA (15). RARs and RXRs heterodimerize in solution, thereby forming a complex that binds with high affinity to DNA response elements, termed retinoic acid response elements (RARE), upstream from the promoters of target genes. RAR is expressed in the murine lung during embryogenesis primarily in late gestation (after 15 d of gestation in the mouse) and appears in the mesenchymal cells of the distal lung (16).

Specific RAR and RXR gene deletions have been created in mice using homologous recombination (17). In general, deletion of one family member, such as RAR or RAR, does not result in an abnormal phenotype in the lung (18). In contrast, the deletion of RAR is associated with defects in tracheal cartilages, and as a group, the animals have a shortened life span (19). RXR and RXR gene deletions have also been described (20, 21). RXR is expressed in the lungs of adult rats and mice (22). RXR / mice demonstrate a delayed embryonic maturation of their lungs and liver, but the most pronounced abnormality is a hypoplastic thin cardiac ventricular wall, which precludes development beyond Gestational Day 17 (20). Mice with compound deletions of RAR and RXR have been reported but their lungs have not been studied in detail (23).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Mouse RAR and RXR Gene Targeting and Breeding

Female RAR +/ and male RXR +/ mice were provided by Dr. Pierre Chambon and bred at the Animal Research Facility at the Iowa City Veterans Affairs Medical Center, following an approved protocol. The targeted disruption of these genes has been described previously (19, 20). The mice were fed Harlan Teklab-7001 mouse and rat chow and watered ad libitum. Male and female mice that were RAR +/,RXR +/ (heterozygotic for both the RAR and RXR null mutations) were obtained and bred. Tail biopsies were performed at Postnatal Day 5 and genomic DNA was isolated. DNA bearing RAR and RXR deletions was identified using the polymerase chain reaction (PCR) and Southern blot analysis. Genomic DNA was also subjected to Southern blot analysis to confirm the genotypes that were determined using PCR (20). Probe 1 was used to detect the RAR gene and is a complementary DNA (cDNA) that is derived from intron 7, immediately 5' to exon 8 (19). Probe B was used to detect the RXR gene and is a cDNA that is derived from intron 2, exon 3, and a portion of intron 3 (20). Based on the results from analysis by PCR, homozygotes for the RAR deletion were killed between Postnatal Days 10 and 12, and the lung was removed. The body and lung weights were recorded and a portion of the lung was immediately frozen in liquid nitrogen and stored at 75°C. The remainder of the lung was used for the isolation of lipid-laden interstitial fibroblasts (LIFs). The isolation procedure was similar to that for rats except that the discontinuous Percoll gradients were formed in 15-ml rather than 50-ml tubes and contained 2.5 ml per layer (13). Greater than 80% of the cells that sedimented at a density of 1.04 contained neutral lipid droplets and vimentin intermediate filaments, characteristic of lipid-laden mesenchymal cells (13).

Administration of tRA to Neonatal Rats

Because others have shown that postnatal administration of tRA increases alveolar number, we determined whether this treatment resulted in an increase in TE messenger RNA (mRNA) (3). Rats were chosen because the effects of tRA on the early postnatal alveolar development have been more completely described in rats than in mice (24). Two groups of four rats were given daily intraperitoneal injections from Postnatal Days 2 through 11 of either 0.5 mg/kg of tRA (1 mg/ml) in safflower oil or safflower oil alone. On Postnatal Day 12, the rats were killed and lung tissue was removed. Some of the lung tissue was used directly for RNA isolation while the remainder was used to isolate LIFs, using the procedure that has been described (25). Northern blot analysis was performed to compare the steady-state levels of TE mRNA in control and tRA-treated animals (25). Three experiments were performed using different litters of rats.

Analysis of TE mRNA

Total RNA was isolated from either whole lung tissue or isolated LIFs obtained from mice or rats, using guanidine isocyanate, subjected to denaturing electrophoresis on 1.5% agarose, and transferred to cationic nylon membranes (25). The membranes were successively probed with the partial cDNAs for rat TE and ribosomal phosphoprotein P-0 (RP-0) (10). RP-0 is constitutively expressed and served as a control for loading differences. Autoradiograms were prepared and the exposure times were limited so the densities were within the linear range of the film. Autoradiograms were scanned using a densitometer, and the densities of the bands containing TE mRNA were expressed relative to the densities of the corresponding bands containing RP-0.

Quantification of Elastin

The lungs from control and transgenic mice at Postnatal Day 28 were perfused with phosphate-buffered saline and removed; the lobes were separated, blotted dry, and weighed, and frozen in liquid nitrogen. A portion of the tissue was used to isolate elastin by extracting with hot alkali (12). The washed, alkali-resistant, insoluble elastin residue was hydrolyzed for 20 h in 6 N HCl under vacuum, and the HCl was removed by evaporation under a stream of nitrogen. The amino-acid composition of the hydrolysate was analyzed using reverse-phase high performance liquid chromatography following a procedure that has been described previously (12).

Histologic Preparation and Analysis

Mice were weighed and killed at Postnatal Day 28, and the anterior chest wall was removed. The lungs were perfused with 2% paraformaldehyde via the right ventricle, buffered in 0.1 M Na-phosphate, pH 7.0. The trachea was cannulated and the cannula was tied firmly in place. The trachea and lungs were infused with 2% paraformaldehyde at 20 cm H₂O pressure and maintained at this pressure for 18 h at 4°C. The lungs were monitored carefully for the initial 10 min, and only lungs that did not leak during this time were used for fixation and embedding. After fixation, the lungs were removed from the chest cavity, the lobes were separated, and the heart and mediastinal tissues were removed. The lungs were dehydrated through a graded series of ethanol. Finally, the lobes were placed into individual cassettes and embedded in paraffin. The central portions of the blocks were sectioned at 3.5-micron intervals, and the sections were mounted on glass slides, deparaffinized, hydrated, and stained with hematoxylin and eosin. Elastic fibers were also identified histologically. Lung tissue sections were obtained from mice with each genotype and were stained concurrently using a modified Van Gieson stain (Sigma Chemical Co., St. Louis, MO), so the intensity of elastic tissue staining could be compared among the various genotypes (26).

Morphometry

Sections were chosen at random, and randomly selected microscopic fields from peripheral (subpleural) and central regions of the lungs were photographed. The photographs were enlarged uniformly, overlaid with transparent grids, and used for morphometric analysis (27). The volume densities of air and tissue were determined using point counting; mean cord lengths (L_m) were determined by counting intersections with an array of lines; and the numbers of alveoli in rectangles of defined areas were recorded. All morphometric measurements were made by two independent observers who were unaware of the genotypes of the animals being analyzed. Three to five animals were analyzed per condition and six photographs were analyzed per animal. The mean linear intercept and alveolar surface areas were calculated as described previously and the values determined for control, RAR /,RXR +/+, and RAR /,RXR +/ animals. Alveolar number was normalized per unit volume (mm³) of the inflated, paraffin-embedded lung. Means and standard deviations were calculated and statistical comparisons were performed using unpaired Student's t test.

Statistical Analyses

The data are expressed as means ± standard error of the mean (SEM). Comparisons of TE mRNA and elastin in wild-type and animals bearing gene deletions and in the neonatal rats that received tRA or safflower oil vehicle were made using a two-way analysis of variance. Morphometric parameters from wild-type and the respective RAR /,RXR +/+ or RAR /, RXR +/ groups were made using Student's t test for unpaired variables. Differences were considered significant when P < 0.05 (28).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

TE mRNA Is Decreased in Pulmonary LIF at Postnatal Day 12

Lung tissue was removed, and LIFs were isolated from control mice and mice bearing the RAR gene deletion at Postnatal Day 12 when the level of TE mRNA is near its maximum (25). The mean body weight and whole lung wet weights of wild-type mice at Postnatal Day 12 were 4.98 ± 0.54 and 0.105 ± 0.006 g (mean ± SEM, n = 5), respectively. The corresponding weights for RAR /,RXR +/+ mice were 5.00 ± 0.73 and 0.103 ± 0.01 g (n = 4), respectively, and for RAR /,RXR +/ mice, 3.36 ± 0.34 and 0.082 ± 0.01 g (n = 4, P < 0.05 compared with wild-type mice). The number of LIFs that were isolated from wild-type mice (7.60 ± 0.85 × 10⁶, n = 13) was higher when compared with the RAR /,RXR +/ mice, 3.07 ± 0.39 × 10⁶ (mean ± SEM, n = 3, P < 0.03), whereas wild-type mice were similar to RAR /,RXR +/+ mice (7.83 ± 1.66 × 10⁶, n = 5). A representative composite autoradiogram demonstrating the abundance of TE (3.5 kb) and RP-0 (1.1 kb) mRNA in wild-type mice and mice bearing a RAR gene deletion is shown in Figure 1A. LIFs isolated from mice that bore the RAR / deletion demonstrated an approximately 2-fold decrease in their TE mRNA that was statistically significant (P < 0.01) when the mice were also RXR +/ (Figure 1B). However, the steady-state levels of TE mRNA in whole lung tissues from RAR /,RXR +/ or RAR /, RXR +/+ mice were similar to levels in wild-type mice (Figure 1C).

View larger version (14K): [in this window] [in a new window]   View larger version (13K): [in this window] [in a new window]   View larger version (59K): [in this window] [in a new window]   Figure 1.   TE mRNA is decreased in isolated LIFs but not in whole lung tissue from RAR null mice. (A) LIFs were isolated from control, RAR /, and RAR /, RXR +/ mice at Postnatal Day 12, and total RNA was isolated from the LIF. Northern blot analysis was performed using 4 µg per lane, and the filters were successively probed with 32P rat TE cDNA (to identify a 3.5-kb mRNA) followed by 32P rat RP-0 cDNA (to identify a 1.1-kb mRNA). RNA from wild-type mice is shown in lanes 1 to 3 and RNA from RAR /,RXR +/+ or RAR /, RXR +/ mice is shown in lanes 4 to 6. Autoradiograms were exposed for varying times to ensure that the signal was within the linear range of the film. RP-0 served as a control to normalize for inequalities in the amounts of RNA loaded in the various lanes. (B) Autoradiograms were subjected to densitometry and the density of the bands representing the various mRNAs were normalized to the density of the corresponding band for RP-0 mRNA for each sample. The error bars are 1 SEM. * P < 0.01. For control mice, n = 9; for RAR /, RXR +/+ mice, n = 3, and RAR /,RXR +/ mice, n = 4. (C) Lungs were removed from control, RAR /, and RAR /,RXR +/ mice at Postnatal Day 12 and total RNA was isolated. Northern blot analysis was performed using 8 µg per lane, and the filters were successively probed with 32P rat TE cDNA followed by 32P rat RP-0 cDNA.

Effects of tRA on Neonatal Rat Lung Elastin Gene Expression

Administering tRA to neonatal rats from Postnatal Days 2 through 11 increases alveolar number (3). Therefore, studies were performed to determine whether tRA administration also produced an increase in TE mRNA (24). The results of these studies are summarized in Figure 2 and indicate that TE mRNA is increased in LIFs but not in whole lung tissue obtained from tRA-treated animals when compared with levels in vehicle-treated control animals.

View larger version (11K): [in this window] [in a new window]   Figure 2.   Administration of retinoic acid (RA) increases TE mRNA in rat LIF at Postnatal Day 12. Total RNA was isolated from LIFs and whole lung tissue that were obtained from control rats (n = 6) that received only safflower oil or rats that received tRA (n = 6) in safflower oil from Postnatal Days 2 through 11. Northern blot analysis was performed as in Figure 1. Bars are 1 SEM. * P < 0.05.

Lung Elastin Content Is Reduced in RAR /,RXR +/ Mice

Lungs were removed from wild-type mice and mice bearing RAR and RXR gene deletions at 28 d of age, and elastin was isolated and subjected to amino-acid analysis. Elastin content was significantly decreased in the RAR /, RXR +/ mice (P < 0.05) and was also decreased, but not significantly, in RAR / mice when compared with levels in wild-type mice (Figure 3). Lung tissue stained with van Gieson elastic stain revealed a reduction in alveolar wall elastic fibers in RAR /,RXR +/ mice when compared with control mice (Figure 4). In contrast, the elastic fiber density in airway and vascular walls was not altered in the RAR /,RXR +/ mice.

View larger version (10K): [in this window] [in a new window]   Figure 3.   Lung elastin is decreased in RAR /,RXR +/ mice. Insoluble elastin was isolated from whole lung tissue at Day 28 and quantified by amino-acid analysis. Error bars represent 1 SEM, n = 4 for each of the three groups. * P < 0.05, wild-type versus RAR /,RXR +/ mice.

View larger version (61K): [in this window] [in a new window]   Figure 4.   Van Gieson elastic tissue stain of lung tissue at 28 d. Lungs from RAR /, RXR +/ and wild-type mice were fixed at an inflation pressure of 20 cm and paraffin sections were prepared. These were stained using the van Gieson elastin stain and photographed. Lung tissues from wild-type mice were stained simultaneously so that the incubation times were equivalent for the two sources of tissue. Panels A and C are from a representative wild-type mouse and panels B and D are from a representative RAR /, RXR +/ mouse. Arrows point to elastic fibers in alveolar septa. airway, A; blood vessel, BV.

Alveolar Number Is Reduced in Mice Bearing the RAR / Gene Deletion

A morphometric analysis was performed to compare parenchymal lung tissues from control mice and mice bearing the RAR and RXR gene deletions. At Postnatal Day 28, wild-type mice had a mean body weight of 15.65 ± 0.75 g (mean ± SEM, n = 8), whereas mice bearing the RAR deletion alone had a mean weight of 11.29 ± 1.26 g (mean ± SEM, n = 5, P < 0.02, Student's t test for unpaired variables). The weight of RAR /,RXR +/ mice did not differ from that of RAR /,RXR +/+ mice. Representative photographs in Figure 5 show that the alveolar spaces were enlarged in both RAR /, RXR +/ (Figure 5B) and RAR /,RXR +/ mice (Figure 5C). This observation was confirmed by morphometric measurements (Table 1). The alveolar wall volume density, L_m, alveolar surface area, and alveolar number were significantly lower (P < 0.01) in the RAR /, RXR +/ group compared with the corresponding wild-type control group. RAR /,RXR +/+ mice also had fewer alveoli and a larger L_m than the corresponding control group, but the decreases were not as marked as in the RAR /,RXR +/ mice. Alveolar surface area was not significantly lower in the RAR /,RXR +/+ mice when compared with the wild-type control group.

View larger version (70K): [in this window] [in a new window]   Figure 5.   Alveolar size is increased in RAR /,RXR +/ mice. Lungs obtained from control mice and mice bearing the RAR and RXR gene deletions were fixed inflated under 20 cm pressure at Postnatal Day 28, embedded in paraffin, sectioned at 3.5 µm, and stained with hematoxylin and eosin. Microscopic fields were selected at random and photographed. Bar represents 50 µm. (A) Wild-type; (B) RAR /,RXR +/; (C) RAR /,RXR +/+.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Massaro and Massaro (24) have shown that tRA increases alveolar number and the volume density of gas exchange tissue in rats that received supplemental tRA during Postnatal Days 3 through 13. They also demonstrated that tRA ameliorates the decreases in alveolar number and the volume density of gas exchange tissue that are induced by the postnatal administration of dexamethasone (3, 24). Massaro and Massaro (29) had previously shown that postnatal administration of dexamethasone is associated with thinning of the alveolar walls, which results from an attenuation of the normal developmental increase in the absolute volume of interstitial cells that occurs between Postnatal Days 6 and 14. The decrease in the volume of interstitial cells was accompanied by a decrease in the volume of lipid-laden fibroblasts and a decrease in the volume density of their lipid droplets. More recent studies by the same investigators indicate that tRA increases alveolar number and surface area in adult rats that developed pulmonary emphysema after receiving a single intratracheal administration of elastase (4). These important studies indicated for the first time that alveolar septal formation could be increased by a pharmacologic maneuver, but they did not identify which RARs and which types of cells are involved. Our previous findings indicate that the lipid-laden interstitial cell is centrally involved in the tRA-mediated effects on elastin gene expression (12, 25). The present study shows that elastin gene expression in neonatal LIF is increased by exogenous tRA and is influenced by RAR-mediated signaling during alveolar formation in vivo but is only one component of retinoic acid-mediated signaling during alveolar formation.

We also desired to minimize the likelihood that the observed pulmonary abnormalities were secondary to defects that occurred outside the lung. RAR is primarily expressed in skin and lung in adult mice, and the level of RAR expression is highest in lung fibroblasts immediately after birth and before the time when elastin gene expression reaches its maximum (25, 30). By studying mice with a RAR gene deletion, we focused on a nuclear retinoid receptor that predominates in pulmonary mesenchymal cells under the influence of endogenous ligands (16). Previous studies using mice with RAR1/RAR or RXR/ RAR compound gene deletions demonstrated that RARs were critically involved in embryonic lung development, as the animals demonstrated lung hypoplasia or agenesis (18, 23). Because alveolar development occurs primarily postnatally in rodents, it was necessary to study a phenotype that was not uniformly lethal during the fetal life or immediately postpartum.

The cooperative interactions between the deletion of the RAR gene and the RXR gene that we observed in alveolar formation have been shown to occur in other contexts (23). A detailed analysis of cardiac defects showed that RAR /,RXR +/ mice developed ventricular thinning, whereas RAR /,RXR +/+ mice did not (23). Similarly, RAR /,RXR +/ mice demonstrated a higher frequency of tracheal ring malformations and ocular defects than did mice with either RXR / or RAR / deletions alone (23). It has been hypothesized that this phenomenon reflects the requirement that the concentration of RXR and RAR both exceed critical thresholds to form RAR/RXR heterodimers (23). The gradation of the effects that we observed in the severity of the elastin deficiency and alveolar enlargement progressing from RAR / to RAR /,RXR +/ is consistent with this hypothesis.

The early (by 4 wk) onset of alveolar air-space enlargement that we observed in RAR / mice is most consistent with a developmental defect in alveolar formation rather than excessive elastic fiber destruction, which requires longer to produce a morphometrically detectable defect. This is illustrated in two genetic models of airspace enlargement in mice. Airspace enlargement that results from developmental abnormalities (either defects in the mottled locus that reduce TE and lysyl oxidase, or a defective fibrillin-1 gene, which results in aberrant elastic fiber formation) is evident at 1 mo of age (31, 32). This contrasts with mice bearing a defect at the pallid locus in which airspace enlargement is thought to primarily represent the destruction of elastic fibers that is not evident until at least 8 mo of age (31, 33).

Alterations in TE mRNA associated with the RAR / genotype or retinoic acid administration were limited to LIF and were not observed in whole lung tissue. One explanation for why LIFs may be unique in this regard follows from our previous observations of RAR gene expression in isolated LIFs and whole lung tissue. Only LIFs demonstrated a temporal correlation between the quantity of endogenous retinoic acid and the level of RAR mRNA (25). This indicates that RAR signaling in LIF may be more influenced by endogenous retinoids than in other types of pulmonary cells. Alternatively, the regulation of elastin gene expression in the LIF may be more dependent on both endogenous and exogenous retinoids than is elastin gene expression in other pulmonary mesenchymal cells where alternative regulatory factors may dominate. Whatever explanation applies, one can conclude that LIF are important contributors to alveolar elastin synthesis because elastic fibers were reduced in RAR /, RXR +/ mice that had diminished TE mRNA at Postnatal Day 12. Our studies do not exclude the possibility that the role of the LIFs in the alveolar development of RAR / mice differs from their role in alveolar development in wild-type mice.

Our observation that TE mRNA levels are similar in whole lung tissue from control and RAR / mice at Postnatal Day 12 is discrepant with our observation that the content of elastin protein is less in RAR / mice at Postnatal Day 28. Several factors may contribute to this apparent disparity. First, a larger fraction of TE mRNA is found in association with blood vessels at Postnatal Day 11 than at Postnatal Days 15 and 21 in the rat (34). Thus, the level of TE mRNA in whole lung tissue is more heavily dependent on expression by blood vessels at Day 12 than at older ages. Therefore, the level of TE mRNA whole lung tissue at Day 12 is less representative of the expression in alveolar wall fibroblasts than is the level of TE mRNA in isolated LIFs. Because TE mRNA persists at higher levels in the alveolar wall longer than in airways and vessels after Postnatal Day 12, its translation would disproportionately contribute to the synthesis of elastin in the alveolar wall, which is the most abundant location of elastin in the lung at Day 28. Second, retinoid signaling may regulate other processes involved in elastin synthesis and elastic fiber assembly that do not impact the steady-state levels of TE mRNA, which only reflect pretranslational events. Elastic fiber assembly is a complex process that is dependent on an orderly progression of cellular growth, migration, differentiation, and the post-translational processing of proteins (8). Other proteins provide examples of how retinoids can alter one or more of these steps (35). Likewise, retinoids have been shown to influence multiple cellular functions that are involved in alveolar septal formation, such as proliferation, migration, and terminal differentiation of cells (1). Further investigation will be required to identify which cellular functions are involved in alveolar septal formation and how retinoids influence them.