Mild Vitamin A Deficiency Delays Fetal Lung Maturation in the Rat

Bernadette Chailley-Heu, Nadia Chelly, Martine Lelivre-Pgorier, Anne-Marie Barlier-Mur, Claudie Merlet-Bnichou, and Jacques R. Bourbon
INSERM U319, Dveloppement Normal et Pathologique des Fonctions Epithliales, Universit Paris 7-Denis Diderot, Paris, France

During late pregnancy, the fetal lung stores surfactant in preparation for extrauterine life. Surfactant deficiency, most often due to prematurity, precipitates respiratory distress syndrome (RDS) of the neonate. Although vitamin A (retinol) and retinoic acid have been shown to enhance the synthesis of phospholipid surfactant components, their effect on surfactant-specific proteins is unclear. No attempt has been made to evaluate the consequences of vitamin A restriction on surfactant phospholipid storage or on the expression of the life-essential surfactant protein-B (SP-B). We induced in rats a partial vitamin A deficiency leading to a 3060% reduction in blood retinol, a status compatible with maintenance of gestation and absence of gross abnormalities in offspring. At term, lung surfactant phospholipids were reduced by 21%, and the major surfactant phospholipid, disaturated phosphatidylcholine (DSPC), was reduced by 27% in vitamin Adeficient (VAD) fetuses. The decrease in surfactant phospholipids and DSPC correlated linearly with plasma retinol, and reached about 50% in fetuses with the lowest retinol concentrations; it was accompanied by reduced expression of the gene for fatty acid synthase, a key enzyme in the synthetic pathway for surfactant-phospholipid lipid precursors. The amounts of SP-A, SP-B, and SP-C messenger RNAs were decreased by 46%, 32%, and 28%, respectively, in VAD fetuses. Consistently, amounts of SP-A and SP-B proteins were diminished as assessed by Western blotting. The proportion of type II cells determined after SP-B labeling was unchanged in VAD as compared with control lungs. Vitamin A deficiency is therefore a cause of lung maturational delay. In view of its rather large incidence in human populations, it may represent an increased risk for RDS and an aggravating factor for prematurity. Chailley-Heu, B., N. Chelly, M. Lelivre-Pgorier, A.-M. Barlier-Mur, C. Merlet-Bnichou, and J. R. Bourbon. 1999. Mild vitamin A deficiency delays fetal lung maturation in the rat. Am. J. Respir. Cell Mol. Biol. 21:8996.

Vitamin A (or retinol) and its active derivative retinoic acid (RA) play multiple roles in lung development and function (1, 2). Thus, vitamin A deprivation during pregnancy (3) or double knockouts of RA nuclear receptors in the mouse (4) both profoundly and similarly disturb lung organogenesis. In the isolated lung anlage, RA was found to perturb lung branching morphogenesis (5) and to alter the expression of pattern-related genes (6). Postnatally, RA was shown to stimulate the process of alveolarization and to prevent the glucocorticoid-induced inhibition of septation (7). In adult animals deficient in retinol, airways
(Received in original form September 23, 1998 and in revised form January 29, 1999 ) Address correspondence to: Jacques R. Bourbon, INSERM U319, Universit Paris 7-Denis Diderot, Case courrier 7126, 2 Place Jussieu, 75251 Paris Cedex 05, France. E-mail: bourbon@paris7. jussieu.fr Abbreviations: choline-phosphate cytidylyltransferase, CPCT; disaturated phosphatidylcholine, DSPC; epidermal growth factor, EGF; high-pressure liquid chromatography, HPLC; messenger RNA, mRNA; retinoic acid, RA; total phospholipids, TPL; surfactant protein-B, SP-B; vitamin A-deficient, VAD.
Am. J. Respir. Cell Mol. Biol. Vol. 21, pp. 8996, 1999 Internet address: www.atsjournals.org

undergo squamous metaplasia, a transformation of the mucociliary epithelium into squamous cells (8). Quite similar changes are observed in bronchopulmonary dysplasia, a chronic lung disease often encountered in infants ventilated for neonatal respiratory distress, and it has been suggested that vitamin A deficiency is part of the pathogenesis of this condition (2). However, the effects of retinoids on alveolar type II cells, the epithelial cells that secrete lung surfactant, have been a matter of debate. Surfactant is a phospholipidprotein material that lines alveoli and prevents their collapse. Although additional retinol and RA were found to enhance surfactant phospholipid synthesis in vivo (9) and in vitro (10), they have been reported to decrease the amount of surfactant protein (SP)-A in the rat fetus in vivo (9) and to decrease SP-A and SP-C messenger RNAs (mRNAs), but to increase SP-B mRNA in human fetal lung explants (11). In embryonic rat lung organ cultures, RA was reported either to inhibit gene expression for all three of these surfactant proteins at a pretranslational level in a dose-dependent way (5), or conversely, to enhance their expression at the same concentrations (12). RA has also been reported to upregulate SP-B gene expression, but not

90

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 21 1999

that of SP-A, in the H441 cell line (13). Changes in lung retinol during development nonetheless argue for a major role of stored retinoids in the control of surfactant storage in preparation for birth. Vitamin A is effectively stored as retinyl esters in fetal rat lung from gestational Days 14 to 18 (the lung primordium appears on Day 11, and term is 22 d), and is mobilized between gestational Day 19 and the first postnatal days (14). This period corresponds to prenatal maturation of the alveolar epithelium with appearance of surfactant-containing lamellar bodies (15) at a time when bronchial branching morphogenesis is completed. We thought that a model of vitamin A deficiency in vivo would help in evaluating the actual role of retinoids in the process of surfactant storage. Since total vitamin A deficiency induces severe malformations and does not allow normal pregnancy to proceed to completion, pregnant rats were exposed to partial vitamin A deficiency. This model allows overall normal fetal development until term. Fetal lung maturity was assessed through determination of the various surfactant components, including disaturated phosphatidylcholine (DSPC), the major surfactant component both quantitatively and functionally, and SP-B, the only surfactant protein thus far demonstrated to be essential to life (16). Possible implication of vitamin A in lung maturation is especially important from a clinical perspective, since surfactant deficiency is the precipitating factor in neonatal respiratory distress syndrome (RDS), and since vitamin A deficiency is a massive public health problem (17). Moreover, women of childbearing age are particularly at risk for such deficiency because of the extra requirements for vitamin A nutrient during pregnancy and lactation (18). The prevalence of vitamin A deficiency is particularly high in developing countries, where it is often part of complex nutritional deficiencies (19), but even in developed countries inadequate intakes of vitamin A may be a common feature as a result of dietary habits, and consequently, the resulting low vitamin A stores may not be sufficient to meet increased demands during pregnancy (18, 2022). Herein we report that all surfactant components in fetal rat lung are effectively reduced by vitamin A deprivation, and that alterations in surfactant parallel the decrease in circulating retinol. Vitamin A deficiency in humans may therefore have delaying effects on fetal lung surfactant accumulation, and consequently may influence the incidence of RDS.

major growth retardation in the offspring. Females having a sperm-positive vaginal plug on the morning after mating were housed separately; this day was considered to be Day 0 of gestation. The same two diets were pursued until gestational Day 22, when fetal blood and lung tissue were collected. Body weight was recorded three times a week throughout gestation. Sample Collection Fetuses were retrieved by cesarean section from pentobarbital-anesthetized pregnant females in both the control and VAD groups. Fetal blood was collected from axillary vessels for plasma retinol determination. Fetal lungs were rapidly removed, immediately frozen in liquid nitrogen, and kept at 80C until use. A blood sample for plasma retinol determination was also taken from each pregnant female before killing. Plasma was immediately separated from blood cells by centrifugation and kept at 20C until analysis. Plasma Retinol and Lung Retinyl Palmitate Determinations Plasma retinol and lung retinyl palmitate concentrations were determined by reverse-phase high-pressure liquid chromatography (HPLC) as previously described (23). Retinyl palmitate is the major storage form of retinol in fetal lung (14). In brief, vitamin A was extracted from plasma samples with ethanol/n-hexane/butylated hydroxytoluene, retrieved in the n-hexane phase, dried, and dissolved in methanol/dichloromethane, 65:35 (vol/vol). Fetal lung lipids were extracted from tissue homogenates in phosphate-buffered saline (PBS) with chloroform/methanol/water, 1:2:0.8 (vol/vol), the chloroform phase was dried up and redissolved in methanol/dichloromethane, 65:35 (vol/vol). In both instances, samples in methanol/dichloromethane were injected into an HPLC pump (Waters, Saint Quentin en Yvelines, France) linked to a multiwavelength detector. Detection was done at 325 nm at a flow rate of 2 ml/min on a Nucleosil C-18 column (Life Science International, Cergy Pontoise, France). The external standard was retinol or retinyl palmitate, and the internal standard was retinol laurate. All reagents were of ultrapure grade (Sigma, Saint Quentin Fallavier, France). Protein and DNA Assays Lung homogenate protein concentrations for Western blots were determined with the Bradford assay kit (BioRad, Ivry sur Seine, France) and bovine serum albumin (BSA) (Sigma) as a standard. Lung DNA was measured with the colorimetric diphenylamine method (24). Lung proteins and DNA were measured in an aliquot of homogenate and in the centrifugation pellets after surfactant fractionation (see next paragraph), respectively. Isolation of Surfactant Material and Phospholipid Analysis Surfactant was isolated by sucrose gradient ultracentrifugation through a method that was previously shown to collect essentially lamellar bodies, to yield a product with all the characteristic biochemical and biophysical features of surfactant, and to be quantitative (25, 26). In brief, tissue

Materials and Methods

Animals Partial vitamin A deficiency was established in rats as previously described (23). In brief, 30-d-old female Sprague Dawley rats weighing around 100 g were assigned to two diet groups. The control group received the standard chow for pregnant and lactating rats (13,000 IU/kg vitamin A). The vitamin A-deficient (VAD) group was fed the same diet devoid of vitamin A. Both diets were purchased from UAR, Villemoison s/Orge, France. The females were mated overnight after having been maintained on their respective diets for around 5 wk. This duration is adequate to obtain a vitamin A deficiency minor enough to allow pregnancy to reach term without gross abnormality and

Chailley-Heu, Chelly, Lelivre-Pgorier, et al.: Vitamin A Deficiency Delays Fetal Lung Maturation in Rat

91

homogenates in isotonic Trisethylenediaminetetraacetic acid (EDTA)NaCl buffer were deposited on 0.75M sucrose and centrifuged for 1 h at 48,000 g at 4C. Material at the interface was collected, diluted in the same buffer, deposited on a discontinuous sucrose gradient of 0.25M and 0.68M, and centrifuged for 1 h at 64,000 g. The final surfactant fraction, which banded over 0.68M sucrose, was collected. A trace amount of [ 14C]dipalmitoylphosphatidylcholine (113 mCi/mmol [4.2 GBq/mmol], Amersham France, Les Ulis, France) was added to determine the yield of the procedure, and lipids were extracted with chloroform/methanol/water, 1:2:0.8 (vol/vol). An aliquot fraction was taken for total phospholipid analysis, and the remainder of the recovered material was treated with osmium tetroxide and chromatographed on silica gel 60 chromatography plates (Merck, Darmstadt, Germany) in chloroform/methanol/water, 65:25:4 (vol/vol) for separation of DSPC (27). Radioactivity was determined in an Optiscint scintillation cocktail (EEG Instrument, Evry, France) with the aid of a Packard Tri-Carb 2100 TR counter (Meriden, CT). Total phospholipid and DSPC contents were determined through phosphate determination in mineralized samples (25). Choline-Phosphate Cytidylyltransferase Assay Choline-phosphate cytidylyltransferase (CPCT) (EC 2.7.7.15), the enzyme that catalyzes the principal rate-limiting step in the biosynthesis of phosphatidylcholine (28), was assayed in fetal lung homogenates by measuring the rate of incorporation of phosphoryl[14CH3]choline (50 mCi/mmol [1.85 GBq/ mmol]; DuPont NEN, Les Ulis, France) into cytidine triphosphate to form cytidine diphosphate (CDP)-choline according to the method of Weinhold and colleagues (29). Isolation of RNA and Northern Blot Analysis Total RNA was isolated from lung tissue with the RNA plus kit of Bioprobe Systems Laboratories (Montreuil sous Bois, France). The precipitated RNA was dissolved in sterile H2O and quantified by absorbance at 260 nm. Thirty micrograms of RNA were fractionated by electrophoresis through 1% agarose, 2.2 M formaldehyde gels, and were blotted onto nylon membranes (Gene Screen; DuPont NEN). The plasmid used to make probe for fatty acid synthase (FAS) was clone FASg57pB1.8, encoding rat spleen FAS (a gift from Dr. S. D. Clarke, Colorado State Univ., Fort Collins, CO). SP-A, SP-B, and SP-C complementary DNA (cDNA) probes were amplified by reverse transcriptionpolymerase chain reaction (RTPCR) from RNA extracted from adult rat lung. The primers used were: for SP-A, 5-GGAAGCCCTGGGATCCCTGGA-3 and 5-TGGTGTGGTTGACCATGGGTC-3, amplifying a 552-bp fragment between positions 88 and 639; for SP-B, 5-GCTACTGCTCCTTCCTAC-3 and 5-AGAGGTGTGGGGTTTGGA-3, amplifying an 1,107-bp fragment between positions 30 and 1136; and for SP-C, 5-ATGGGTAGCAAAGAGGTA-3 and 5-GAGTATGGACAGGAGCAG-3, amplifying a 673-bp fragment between positions 19 and 691. PCR products were separated by agarose gel electrophoresis and eluted from the gel with GenElute Spin Columns (Supelco, Bellefonte, PA). Probes were labeled with [ -32P]deoxycytosine triphos-

phate ([-32P]dCTP) (ICN, Irvine, CA) using the Rediprime DNA labeling system from Amersham France and purified on NucTrap probe purification columns (Stratagene, Cambridge, UK). Conditions for prehybridization, hybridization, and stripping of blots were as previously described (30). To allow correction for variations in the amount of RNA loaded, the blots were also hybridized with a 18S rRNA probe. Autoradiographs were made by exposing blots for at least 1 h at 80C to X-ray film (Reflection; DuPont NEN). Quantification of signal intensity was performed by densitometric analysis of autoradiograms using image analysis software (NIH Image; U.S. National Institutes of Health, Bethesda, MD). Western Blot Analysis of Surfactant Proteins Lung homogenate samples were solubilized under reducing conditions for SP-A analysis, and under nonreducing conditions for SP-B analysis. Proteins were separated by 12% (SP-A) or 15% (SP-B) sodium dodecylsulfatepolyacrylamide gel electrophoresis (SDSPAGE). Proteins were detected after blotting with antibodies raised in rabbits and directed against SP-A (a gift of Dr. M. Post, Hospital for Sick Children, Toronto, ON, Canada) or SP-B (antibody 28031, gift of Prof. J. A. Whitsett, Childrens Hospital Medical Center, Cincinnati, OH). Each blotting membrane was normalized by analysis with a monoclonal anti--actin antibody (Sigma) to avoid bias from possible variations in the amount of protein loaded. A secondary antirabbit or antimouse IgG antibody conjugated with horseradish peroxidase was applied and the resulting complex was visualized by chemiluminescence (ECL Western Blotting Analysis System, Amersham France). Band densitometry was done with image analysis software (NIH Image). Immunofluorescence Study To determine the percentage of SP-Bimmunoreactive cells in fetal lungs, a double labeling was performed as previously described (31). Briefly, frozen sections were incubated with the antiSP-B antibody for 2 h and were then labeled with a Texas red-conjugated anti-IgG antibody (Amersham France), followed by nuclear labeling with the bisbenzimide dye Hoechst 33258 (Sigma) at 0.25 g/ml for 10 min. For each sample, counts were made on photographs of different fields of lung parenchyma, excluding the large bronchial areas. Up to 2,000 cells were counted for each fetal lung. Statistical Analysis Data are presented as mean SEM. Mean comparisons between control and VAD fetuses were made with a twotailed t test for unpaired values, with P 0.05 considered the limit of statistical significance.

Results
Characterization of the Experimental Model The various determinations were made in fetuses from six control litters and 15 VAD litters. The number of control litters was smaller than that of VAD litters because controls were homogeneous with regard to plasma retinol

92

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 21 1999

TABLE 1

Body weight, lung weight, lung weight/body weight ratio, lung proteins, and lung DNA in control and VAD fetuses
Groups BW (g) LW (mg) LW/BW (mg/g) L Prot (mg/g) L DNA (mg/g)

Controls VAD rats

5.32 0.23 (20) 5.09 0.10 (32)

151 10 (20) 145 5 (32)

28.2 1.5 (20) 28.5 1.0 (32)

59.2 1.7 (10) 55.5 1.5 (24)

7.7 0.3 (10) 6.8 0.2* (24)

* P 0.01 versus control group. Definition of abbreviations: BW body weight; L DNA lung DNA content; L Prot lung protein content; LW lung weight; VAD vitamin A-deficient rats.

level, whereas VAD pregnant rats showed large variations in plasma retinol level. It was therefore necessary to explore a larger number of VAD litters in order to deter-

mine whether fetal lung changes correlated with the degree of vitamin A deficiency. Maternal and fetal plasma retinol, maternal and fetal body weight, and fetal lung weight were determined in each individual. Samples from several litters with different degrees of vitamin A deficiency were used for each type of biochemical analysis. In general, from the approximately 12 fetuses per litter, from one to three fetuses were used for a particular analysis. Body growth before mating, and fertility rate, were the same in control and VAD female rats (23). Maternal growth during pregnancy was regular in both groups and resulted in an identical weight gain (170 15 g and 172 10 g, respectively). The same average number of fetuses was retrieved in both groups (12.8 0.7 and 11.9 0.6, respectively). No obvious developmental defect was observed in any of the fetuses from either group. At killing, maternal plasma retinol was 31.5 1.8 g/dl in control rats (range: 2436 g/dl) and 16.1 1.1 g/dl in VAD rats (P 0.001, range: 12.528 g/dl). Fetal plasma retinol was 25.1 1.3 g/dl (n 29, range: 1829 g/dl) and 14.6 0.5 g/dl (n 45, P 0.001, range: 821 g/dl) in the control group and VAD group respectively, which corresponds to a 42% reduction on average in the latter. Lung retinyl palmitate was 0.62 0.05 and 0.02 0.002 pmol/ mg in control and VAD fetuses, respectively ( n 6 for both groups, P 0.001). Neither fetal body weight nor lung weight, nor the lung to body weight ratio were significantly changed in VAD fetuses (Table 1). Fetal Lung Protein and DNA Contents As shown in Table 1, fetal lung protein content was unchanged in VAD fetuses as compared with controls. By contrast, lung DNA was decreased by 12.5% (P 0.01) in VAD fetuses, which indicates a reduction of total cell number with an average slight cell hypertrophy. Surfactant Phospholipids Total phospholipids (TPL) and DSPC of surfactant material were significantly decreased on average, by 21% and 27%, respectively, in the VAD fetuses as compared with control fetuses (Figure 1, left panel ). It should be emphasized, however, that for both TPL and DSPC the decrease was 50% in fetuses with the lowest plasma retinol concentrations (Figure 1, right panel ). Furthermore, TPL and DSPC correlated linearly with fetal plasma retinol in the VAD group but not in the control group (Figure 1, right panel ).

Figure 1. Effects of vitamin A deficiency on fetal lung surfactant phospholipids. Surfactant material was isolated as stated in MATERIALS AND METHODS. TPL total phospholipids; DSPC disaturated phosphatidylcholine; LW lung weight. Left panel: Comparison between control and VAD fetuses for mean TPL and DSPC values. Mean SEM for 10 and 24 individuals from four and 10 different litters, respectively (significant difference from controls at **P 0.01; ***P 0.001). Right panel: Plot of individual TPL and DSPC values versus corresponding fetal plasma retinol concentrations in control and VAD fetuses. Linear regression: continuous line for control fetuses and dotted line for VAD fetuses. TPL and DSPC contents of surfactant correlated with plasma retinol in VAD fetuses (r 0.612 and r 0.661, respectively, P 0.01), but not in control fetuses ( r 0.390 and r 0.161, respectively, P 0.1).

Chailley-Heu, Chelly, Lelivre-Pgorier, et al.: Vitamin A Deficiency Delays Fetal Lung Maturation in Rat

93

rate-limiting step of the CDP-choline pathway of phosphatidylcholine synthesis, or whether it resulted from changes in the synthesis of lipid precursors of phospholipids, lung CPCT activity and FAS gene expression were determined. The activity of FAS is effectively entirely regulated by adjustment of its synthesis at the steady-state level of its mRNA (32). CPCT activity was found to be identical in VAD and control fetuses (158 11 dpm/g protein/h, n 7, versus 161 18 dpm/g protein/h, n 6, respectively). By contrast, the steady-state level of FAS mRNA was found to be reduced by 41% in VAD fetuses as compared with controls (Figure 2). Surfactant Protein Gene Expression As shown in Figures 3A and 3B, the mRNAs for SP-A, SP-B, and SP-C were significantly decreased, by 46%, 32%, and 28% on average, respectively, in VAD fetuses as compared with controls. The mRNA level correlated with the plasma retinol concentration (P 0.01 for SP-A and SP-B, P 0.05 for SP-C), and the reduction was considerable, reaching 6080% in those fetuses with the lowest plasma retinol concentrations (Figure 3C). Surfactant Proteins To determine whether the decreased amounts of surfactant protein mRNAs were reflected in changes in the amount of surfactant proteins themselves, we performed a semiquantitative Western blot analysis of SP-A and SP-B. Results are presented in Figure 4. The antibody directed against rat pulmonary SP-A strongly labeled the characteristic SP-A monomeric bands (unglycosylated 26 kD,
Figure 3. Effects of vitamin A deficiency on surfactant protein gene expression in fetal lung. Northern hybridization analysis of relative level of mRNAs coding for the surfactant proteins SP-A, SP-B, and SP-C, and of 18S rRNA, using specific cDNA probes, as described in MATERIALS AND METHODS. (A) Representative autoradiograms; lanes 1 and 2: control fetuses (plasma retinol: 26.1 g/dl and 25.7 g/dl, respectively), lanes 3 and 4: VAD fetuses (plasma retinol: 15.3 g/dl and 11.5 g/dl, respectively). Sizes of transcripts are 1 kb and 1.7 kb for SP-A, 1.4 kb for SP-B, and 0.9 kb for SP-C. (B) Densitometric analysis of autoradiograms; results are expressed as percentage of the control value after normalization for RNA loading on the basis of hybridization of 18S rRNA (mean SEM for nine and 15 individuals from four and 10 different litters in the control and the VAD groups, respectively). Significant difference from control fetuses at **P 0.01; ***P 0.001. (C) Plot of individual SP-A, SP-B, and SP-C mRNA levels (densitometric analysis) versus corresponding fetal plasma retinol concentration in control ( closed symbols) and VAD (open symbols) fetuses. All three protein gene expressions correlated linearly with plasma retinol (r 0.714 for SP-A, P 0.01; r 0.750 for SP-B, P 0.01; r 0.549 for SP-C, P 0.05).

Figure 2. Effect of vitamin A deficiency on FAS gene expression in fetal lung. Northern hybridization analysis of relative level of FAS mRNA and of 18S ribosomal RNA (rRNA) done with specific cDNA probes, as described in MATERIALS AND METHODS. (A) Representative autoradiograms. (B) Densitometric analysis of autoradiograms; results are expressed as percentage of the control value after normalization for RNA loading on the basis of hybridization of 18S rRNA (mean SEM for seven individuals from six different litters). Significant difference with control fetuses at **P 0.01.

CPCT Activity and Fatty Acid Synthase Gene Expression In order to determine whether the decreased DSPC content of the surfactant pool was due to a decrease in the

94

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 21 1999

Figure 4. Effect of vitamin A deficiency on fetal lung surfactant proteins. Western blot analysis of SP-A, SP-B, and -actin after protein separation on SDS-polyacrylamide gels (12% for SP-A, 15% for SP-B). Representative immunoblots of an experiment that was repeated on three independent series of individuals (three controls and nine VAD fetuses from 10 different litters). Lane 1: control fetus (plasma retinol: 26.6 g/dl), lanes 2 to 4: VAD fetuses (plasma retinol: 19.8 g/dl, 15.2 g/dl, and 11.7 g/ dl, respectively). A progressive reduction of SP-A was observed with decreasing plasma retinol in the lungs of VAD fetuses; a marked decrease was observed only in the last two lanes for SP-B. Right margin indicates Mr 103.

glycosylated 32 kD and 36 kD bands) and the dimeric band (60 kD). Labeling of SP-A progressively decreased with decreasing fetal blood retinol in samples from VAD fetuses. Band-densitometric analysis of the different membranes, normalized to -actin, was done to avoid bias due to variations in protein loading. It indicated a decrease in all SP-A forms, which reached about 25%, 65%, and 75% in fetuses with plasma retinol concentrations that were decreased by 25%, 40%, and 55%, respectively, as compared with the mean control value. The anti-SP-B antibody labeled an 18-kD band corresponding to the unreduced dimeric form of SP-B; densitometric analysis indicated that SP-B was decreased by 5075% in fetuses with plasma retinol values that were decreased by 40% or more. Surfactant-Producing Cells SP-B labeling was used as a feature for characterizing SPBproducing cells, (i.e., type II cells). The number of SP-B immunoreactive cells in each fetal lung was referred to the total number of parenchymal cells as determined through labeling of nuclei. Three VAD (plasma retinol: 16.0 2.5 g/dl) and three control (plasma retinol: 24.7 1.1 g/dl) fetuses were studied. The percentage of SP-Bpositive cells was found to be unchanged in VAD lungs (20.1% 2.5%) as compared with control lungs (20.8% 4.2%).

Discussion
The role of retinoids in alveolar cell maturation is not clearly established. Most previous studies have addressed

this question either through the use of severely depleted animals and knockout models, or through organ culture approaches using high dosages of retinoids. Since partial vitamin A deficiency may occur quite frequently, in industrialized countries as well as others, we developed a model of vitamin A restriction in pregnant rats leading to mild deficiency, and we studied the consequences on several key parameters of lung-functional maturation at the end of gestation. Our results clearly indicate that vitamin A is a limiting factor for normal type II cell maturation, and that vitamin A deficiency reduced the prenatal accumulation of surfactant components, including the functionally crucial components DSPC and SP-B. The reason why fetal lung stores of retinyl palmitate were almost exhausted in VAD fetuses, whereas plasma retinol was reduced by only 40% on average, does not appear clearly. It should be emphasized, however, that at the stage close to term at which we studied fetuses, fetal lung retinol stores have already been largely used (14). Most likely, the peak level reached on Day 18 in VAD fetuses was reduced as compared with that of control rats, and the decreased amount of disposable retinol was already almost completely mobilized on gestational Day 22. The present study is the first to investigate the consequences of vitamin A deficiency during pregnancy on surfactant phospholipids, especially on DSPC. Although DSPC is the major component of surfactant, it is not surfactant-specific, since only 30% of whole-lung DSPC originates from type II cells (33). We therefore extracted surfactant material from fetal lung before phospholipid quantification, in order to specifically evaluate the effects of vitamin A restriction on surfactant DSPC. Our finding of a decreased amount of surfactant DSPC is consistent with those in previous studies by our group, that both retinol and RA enhanced DSPC synthesis when given in vivo (9) or added in vitro to fetal type II cell cultures (10). Also in agreement with our finding of decreased amount of surfactant DSPC is the report of a reduced rate of DSPC synthesis in type II cells isolated from the lungs of vitamin A-deprived adult rats (34). Of particular interest is that surfactant phospholipid storage, including that of DSPC, appeared not to be correlated with plasma retinol levels in control fetuses, whereas it was proportional to plasma retinol concentrations in VAD fetuses. Results presented in Figure 1 suggest that below a certain concentration, plasma retinol becomes a rate-limiting factor in surfactant phospholipid synthesis and storage. This threshold appears to be around 18 g/dl. It can therefore be concluded that a reduction of circulating retinol by only 30% can significantly affect surfactant production. As also shown in Figure 1, a reduction of blood retinol to about half the mean control value led to a deficit in surfactant DSPC of 5060%. The amount of surfactant DSPC in these latter fetuses was not much higher than that in the normal rat lung on Day 20.5 of gestation (25) (i.e., at a still very immature, nonfunctional stage). This shows that vitamin A deficiency does not need to be severe to lead to a considerable lung-maturational delay. Considering the possible underlying mechanisms for the observed reduction in DSPC, CPCT, the enzyme that catalyzes the rate-limiting step of the CDP-choline path-

Chailley-Heu, Chelly, Lelivre-Pgorier, et al.: Vitamin A Deficiency Delays Fetal Lung Maturation in Rat

95

way of phosphatidylcholine synthesis, was unaffected by vitamin A deficiency. Since the developing type II cell has been shown to be dependent on proper lipogenesis for meeting the needs of surfactant synthesis and storage (35), we questioned whether FAS, a key enzyme in the lipogenesis pathway, was involved. The reduced FAS gene expression observed in our study suggested that decreased storage of surfactant phospholipids effectively resulted from a decreased supply of their lipid precursors. In accord with this conclusion is the previous finding that RA increased the incorporation of acetate, a precursor of fatty acids, in phospholipids of isolated fetal type II cells (10). The present study is the first to show that vitamin A deficiency affects gene expression and lung content of surfactant proteins. Western blot analysis of SP-A and SP-B proteins corroborated the findings obtained at the mRNA level and clearly indicated a protein deficit, which was all the more marked because the plasma retinol concentration was lowered. SP-A appeared to be more sensitive to vitamin A availability than were SP-B and SP-C. SP-A mRNA was effectively decreased to a greater extent on average, and SP-A protein was already less abundant in fetuses with small plasma retinol reductions, in which SP-B did not appear to have yet been affected. Although the consequences of vitamin A deprivation on SP-B gene expression had not been explored before, SP-A and SP-C gene expression had been previously found to be unaffected in VAD rat fetuses (36). Two differences may account for this discrepancy between the previous and the present investigations. First, although the control diets in the two studies contained the same amount of vitamin A, the vitamin A-deprived diet used formerly contained small amounts of vitamin A (0.06 mg vitamin A/kg or 200 IU/ kg), whereas the diet used in the present study was totally devoid of vitamin A. Thus, although fetuses in the previous study were effectively deprived of vitamin A, as shown by the diminished retinyl-ester content of fetal lungs, the depletion may have been insufficient to cause a change in surfactant protein mRNAs; however, the lack of blood retinol determination in the prior study prevents a direct comparison. Second and most likely, the discrepancy may result from the stage at which fetal lung was studied (i.e., in the previous study it was examined at 18 and 20 d of gestation, when fetal lung is still immature, and in the present study at term, when the lung is mature, with sufficient surfactant to achieve normal respiratory function). Taking into account the considerable increase in surfactant protein expression and storage between Day 20 and term (26, 37), changes could have become detectable at term in the previous experiments. Our study also provides insights into the effects of retinoids on surfactant protein gene expression. Several previous investigations, using in vitro models, suggested that retinoids would downregulate surfactant protein gene expression, mostly for SP-A and SP-C (5, 11). It should be emphasized that the retinoid concentrations used were high, and that in one of these studies (5), fetal rat lung was explanted at an early stage (13 d), far preceding the time of surfactant storage. The findings of Cardoso and coworkers (5, 6) suggest that a high local RA concentration would favor bronchial-type differentiation. In another

study, using lung explants taken at a later stage, RA was conversely reported to increase surfactant protein gene expression, and for SP-A and SP-B this was observed with low concentrations of RA (12). Our results clearly indicate that a normal vitamin A supply is necessary for normal development of surfactant proteins, both in terms of gene expression and protein synthesis. Taken together with the findings of Bogues and associates (12), they suggest that at lower concentration and/or at later developmental stages, retinoids are stimulating factors for surfactant protein synthesis. The decreased lung surfactant content in VAD fetuses in our study could have resulted from a decreased number of surfactant-producing cells or from a decreased rate of surfactant synthesis. SP-B labeling did not indicate any reduction of type II cell proportion in VAD fetuses. Therefore, the decreased lung cell number indicated by the 12.5% decrease in lung DNA cannot account for the 30 50% decrease in various surfactant parameters. Most likely, decreased surfactant content reflects a decreased capacity of surfactant synthesis on a per cell basis. We conclude that surfactant ontogeny (i.e., type II cell maturation) was delayed by vitamin A deficiency. The present in vivo approach does not allow us to determine whether the effects of retinoids are exerted directly on type II cells, and if so, whether these effects occur directly at the level of expression of the considered genes, through RA receptor-mediated transactivation. Our previous findings of increased DSPC synthesis in isolated type II cells exposed to RA (9, 10) argue for the possibility of direct effects for the phospholipid moiety of surfactant. With regard to SP-B control, a direct effect is demonstrated by enhanced SP-B synthesis in pulmonary adenocarcinoma cells exposed to RA or transfected with RA-receptor expression vectors and by identification of an RA-responsive element mediating RA stimulation of the human SP-B promoter (38). Several other reports conversely suggest the existence of indirect effects. For instance, it has been suggested that the RA-stimulated proliferation of cells isolated from developing mouse lung involved an epithelial mesenchymal interaction, and was in part produced by stimulation of epidermal growth factor (EGF) receptor expression (39). RA was subsequently shown to be a potent enhancer of EGF-receptor binding in fetal rat lung fibroblasts (40). Since EGF is itself a stimulating factor of fetal lung maturation, EGF and EGF receptors may also relay the maturational effect of RA. To explore these possible mechanisms, further studies conducted with preparations such as isolated type II cells or mixed cells will be necessary. As a whole, this study shows that a partial vitamin A deficiency during pregnancy leads to lung immaturity at term, with a decrease in both the phospholipid and protein moieties of surfactant. Until recently, only severe vitamin A deficiencies or excess of vitamin A had been considered as risk factors to the fetus. In previous studies, longer vitamin A deprivation, likely to have completely exhausted maternal vitamin A stores, resulted in multiple fetal malformations, growth retardation, or death (3, 41). Recent investigations in our laboratory (23), using the experimental design used in the present study, have shown that par-

96

AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 21 1999

tial vitamin A deficiency alters nephrogenesis and leads to a permanent nephron deficit. Thus, a moderate vitamin A deficiency that has no obvious consequences on overall development effectively leads to impaired development in at least two organs, the kidney and the lung. By contrast with a nephron deficit, which could have possible longterm consequences for renal function, the consequences of insufficient surfactant for respiratory function will be immediate at birth. As far as the present observations can be extended to human development, vitamin A deficiency may represent a factor for increased risk of RDS and an aggravating factor for prematurity. In view of the large incidence of relative vitamin A deficiency in human populations, the risk of delayed fetal lung maturation should not be neglected, and vitamin A status should be checked during pregnancy in groups at risk.
Acknowledgments: The authors are grateful to Dr. Jean Bastin for helpful critical reading of the manuscript.

References
1. Chytil, F. 1992. The lungs and vitamin A. Am. J. Physiol. 262 (Lung Cell. Mol. Physiol. 6):L517L527. 2. Zachman, R. D. 1995. Role of vitamin A in lung development. J. Nutr. 125:1634S1638S. 3. Wilson, J. G., C. B. Roth, and J. Warkany. 1953. An analysis of the syndrome of malformations induced by maternal vitamin A deficiency: effects of restoration of vitamin A at various times during gestation. Am. J. Anat. 92:189217. 4. Mendelsohn, C., D. Lohnes, D. Decimo, T. Lufkin, M. LeMeur, P. Chambon, and M. Mark. 1994. Function of RA receptors (RARs) during development: II. Multiple abnormalities at various stages of organogenesis in RAR double mutants. Development 120:27492771. 5. Cardoso, W. V., M. C. Williams, S. A. Mitsialis, M. Joyce-Brady, A. K. Rishi, and J. S. Brody. 1995. Retinoic acid induces changes in the pattern of airway branching and alters epithelial cell differentiation in the developing lung in vitro. Am. J. Respir. Cell Mol. Biol. 12:464476. 6. Cardoso, W. V., S. A. Mitsialis, J. S. Brody, and M. C. Williams. 1996. Retinoic acid alters the expression of pattern-related genes in the developing lung. Dev. Dynam. 207:4759. 7. Massaro, G. D., and D. Massaro. 1996. Postnatal treatment with retinoic acid increases the number of pulmonary alveoli in rats. Am. J. Physiol. 270 (Lung Cell. Mol. Physiol. 14):L305L310. 8. MacDowell, E. M., K. P. Keenan, and M. Huang. 1984. Effects of vitamin A deprivation on hamster tracheal epithelium: a quantitative morphologic study. Virchows Arch. B. Cell Pathol. 45:197219. 9. Fraslon, C., and J. R. Bourbon. 1994. Retinoids control surfactant phospholipid biosynthesis in fetal rat lung. Am. J. Physiol. 266 (Lung Cell. Mol. Physiol. 10):L705L712. 10. Fraslon, C., and J. R. Bourbon. 1994. Differential effects of retinoic acid, gastrin-releasing peptide and epidermal growth factor on fetal type-II cell maturation in vitro. Prog. Respir. Res. 27:106112. 11. Metzler, M. D., and J. M. Snyder. 1993. Retinoic acid differentially regulates expression of surfactant-associated proteins in human lung. Endocrinology 133:19901998. 12. Bogue, C. F., H. C. Jacobs, D. W. Dynia, C. M. Wilson, and I. Gross. 1996. Retinoic acid increases surfactant protein mRNA in fetal rat lung in culture. Am. J. Physiol. 271 (Lung Cell. Mol. Physiol. 15):L862L868. 13. George, T. N., O. L. Miakotina, K. L. Goss, and J. M. Snyder. 1998. Mechanism of all trans-retinoic acid and glucocorticoid regulation of surfactant protein mRNA. Am. J. Physiol. 274 (Lung Cell. Mol. Physiol. 18):L560 L566. 14. Shenai, J. P., and F. Chytil. 1990. Vitamin A storage in lungs during perinatal development in the rat. Biol. Neonate 57:126132. 15. Geevarghese, S. K., and F. Chytil. 1994. Depletion of retinyl esters in the lungs coincides with lung prenatal morphological maturation. Biochem. Biophys. Res. Commun. 200:529535. 16. Noggee, L. M., D. E. De Mello, L. P. Dehner, and H. R. Colten. 1993. Deficiency of pulmonary surfactant protein B in congenital alveolar proteinosis. N. Engl. J. Med. 328:406410.

17. West, K. P., Jr. 1994. Vitamin A deficiency: its epidemiology and relation to child mortality and morbidity. In Vitamin A in Health and Disease. R. Blomhoff, editor. Marcel Dekker, New York. 585614. 18. Ortega, R. M., P. Andres, R. M. Martinez, and A. M. Lopez-Sobaler. 1997. Vitamin A status during the third trimester of pregnancy in Spanish women: influence on concentrations of vitamin A in breast milk. Am. J. Clin. Nutr. 66:564568. 19. Filteau, S. M., and A. M. Tomkins. 1995. Vitamin A supplementation in developing countries. Arch. Dis. Child. 72:106107. 20. Hercberg, S., P. Preziosi, P. Galan, M. Devanlay, H. Keller, C. Bourgeois, G. Potier de Courcy, and F. Cherouvrier. 1994. Vitamin status of a healthy French population: dietary intakes and biochemical markers. Int. J. Nutr. Res. 64:220232. 21. Duitsman, P., L. Cook, S. Tanumihardjo, and J. Olson. 1995. Vitamin A inadequacy in socioeconomically disadvantaged pregnant Iowan women as assessed by the modified relative dose response (MRDR) test. Nutr. Res. 15:12631276. 22. Van den Berg, H. 1996. Vitamin A intake and status. Eur. J. Clin. Nutr. 50 (Suppl. 3):S7S12. 23. Lelivre-Pgorier, M., J. Vilar, M.-L. Ferrier, E. Moreau, N. Freund, T. Gilbert, and C. Merlet-Bnichou. 1998. Mild vitamin A deficiency leads to inborn nephron deficit in the rat. Kidney Int. 54:14551462. 24. Burton, K. 1956. A study of the conditions and mechanism of the diphenylamine reaction for the colorimetric estimation of deoxyribonucleic acid. Biochem. J. 62:315323. 25. Bourbon, J. R., P. M. Farrell, E. Doucet, D. J. Brown, and C. Valenza. 1987. Biochemical maturation of fetal rat lung: a comprehensive study including surfactant determination. Biol. Neonate 52:4860. 26. Farrell, P. M., J. R. Bourbon, R. H. Notter, L. Marin, L. M. Noggee, and J. A. Whitsett. 1990. Relationships among surfactant fraction lipids, proteins and biophysical properties in the developing rat lung. Biochim. Biophys. Acta 1044:8490. 27. Patterson, C. E., K. S. Davis, D. E. Beckman, and R. A. Rhoades. 1986. Fatty acid synthesis in the fetal lung: relationship to surfactant lipids. Biochim. Biophys. Acta 878:110126. 28. Post, M., J. J. Batenburg, L. M. G. Van Golde, and B. T. Smith. 1984. The rate-limiting reaction in phosphatidylcholine synthesis by alveolar type II cells isolated from fetal rat lung. Biochim. Biophys. Acta 795:558563. 29. Weinhold, P. A., M. E. Rounsifer, and D. A. Feldman. 1986. The purification and characterization of CTP:phosphorylcholine cytidylyltransferase from rat liver. J. Biol. Chem. 261:51045110. 30. Fraslon-Vanhulle, C., B. Chailley-Heu, J. J. Batenburg, R. Elfring, and J. R. Bourbon. 1994. Ontogeny of surfactant proteins and lipid-synthesizing enzymes in cultured fetal lung epithelial cells. Am. J. Physiol. 267 (Lung Cell. Mol. Physiol. 11):L375L383. 31. Benachi, A., B. Chailley-Heu, A. L. Delezoide, M. Dommergues, F. Brunelle, Y. Dumez, and J. R. Bourbon. 1998. Lung growth and maturation after tracheal occlusion in diaphragmatic hernia. Am. J. Respir. Crit. Care Med. 157:921927. 32. Wakil, S. J., J. K. Stoops, and V. C. Joshi. 1983. Fatty acid synthesis and its regulation. Annu. Rev. Biochem. 52:537579. 33. Young, S. L., S. A. Kremers, J. S. Apple, J. D. Crapo, and G. W. Brumley. 1981. Rat lung surfactant kinetics: biochemical and morphometric correlation. J. Appl. Physiol. 51:248253. 34. Zachman, R., X. Chen, A. K. Verma, and M. A. Grummer. 1992. Some effects of vitamin A deficiency on the isolated rat lung alveolar type II cell. Int. J. Vit. Nutr. Res. 62:113120. 35. Bourbon, J. R., and C. Fraslon. 1991. Developmental aspects of the alveolar epithelium and the surfactant system. In Pulmonary Surfactant: Biochemical, Functional, Regulatory and Clinical Concepts. J. R. Bourbon, editor. CRC Press, Boca Raton, FL. 257325. 36. Zachman, R. D., and M. A. Grummer. 1998. Effect of maternal/fetal vitamin A deficiency on fetal rat lung surfactant protein expression and the response to prenatal dexamethasone. Pediatr. Res. 43:178183. 37. Schellase, D. E., P. A. Emrie, J. H. Fisher, and J. M. Shannon. 1989. Ontogeny of surfactant apoproteins in the rat. Pediatr. Res. 26:167174. 38. Yan, C., M. Ghaffari, J. A. Whitsett, X. Zeng, Z. Sever, and S. Lin. 1998. RA-receptor activation of SP-B gene transcription in respiratory epithelial cells. Am. J. Physiol. 275 (Lung Cell. Mol. Physiol. 19):L239L246. 39. Schuger, L., J. Varani, R. Mitra, Jr., and K. Gilbridge. 1993. Retinoic acid stimulates mouse lung development by a mechanism involving epithelialmesenchymal interaction and regulation of epidermal growth factor receptors. Dev. Biol. 159:462473. 40. Dammann, C. E. L. 1998. Regulation of the epidermal growth factor receptor in fetal rat lung fibroblasts during late gestation. Endocrinology 139: 16711677. 41. Takahashi, Y. I., J. E. Smith, M. Winick, and D. W. S. Goodman. 1975. Vitamin A deficiency and fetal growth and development in the rat. J. Nutr. 105:12991310.