DEVELOPMENTAL DYNAMICS 20R47-59 (1996) Retinoic Acid Alters the Expression of Pattern-Related Genes in the Developing Rat Lung WELLINGTON V. CARDOSO, S. ALEX MITSIALIS, JEROME S. BRODY, AND MARY C. WILLIAMS The Pulmonary Center, Department of Medicine, Boston University School of Medicine, Boston, Massachusetts 02118 ABSTRACT Exogenous retinoids alter pattern formation and differentiation in many developing systems, such as limb, vertebrae, and central nervous system. Many of these effects are mediated by changes in expression of patterning genes such as Hox genes and Sonic hedgehog. We have previously shown that exogenous retinoic acid, administered to the embryonic rat lung in culture alters the structural pattern of the developing lung, suppressing formation of distal lung and favoring growth of proximal tubules. To determine whether these retinoic acid-induced changes in lung development were linked to alterations in pattern-related genes, we characterized the expression of Hoxa-2, Hoxh-6, and Sonic hedgehog mRNAs in vivo and in vitro, with or without 10-'M retinoic acid, by in situ hybridization and quantitative polymerase chain reaction. Each of these genes demonstrated unique timing and distribution of expression that was similar in vivo and in control cultured embryonic lungs. Hoxb-6 and Sonic hedgehog mRNAs both decreased during lung development in vivo or in vitro. From the patterns of mRNA expression we propose that Hoxb-6 is involved in distal airway branching while Homz-2 is involved in differentiation of proximal mesenchymal derivatives and vasculogenesis in the lung. RA upregulated all three genes, changing their developmental pattern of distribution and preventing the developmental decrease in Sonic hedgehog expression. We propose that RA acts to maintain high levels of expression of these and likely other pattern-related genes in a fashion that is characteristic of the immature lung, promoting continued formation of proximal lung structures and preventing formation of typical distal lung structures of the mature lung. 0 1996 Wiley-Liss, Inc. Key words: Retinoic acid, Hox genes, Sonic hedgehog,Lung development INTRODUCTION Retinoic acid (RA) is largely known for its effects on pattern formation during development and regeneration of structures such as vertebrae, limb, face, neural tube, and other organs (Hill et al., 1993; reviewed in Tabin, 1991, and Gudas, 1994). In the adult lung retinoids play an important role in maintenance of the difB 1996 WILEY-LISS,INC. ferentiated state of the respiratory epithelium, preventing squamous metaplasia (reviewed in Chytil, 1992). The developing lung is responsive to exogenous retinoids. Administration of retinoic acid (RA) to pregnant hamsters results in alterations in the lobation pattern of the fetal lung (Shenefelt, 1972). Retinoic acid interacts with a number of retinoic acid receptors (RARs) which have been shown to be essential for normal lung formation. Double knock-outs of RARs, such as FtARalp2, cause dramatic abnonnalities in the respiratory tract of transgenic mice. Such abnormalities are similar to those described in vitamin A deficient animals (Wilson et al., 19531, and include developmental arrest and growth retardation (Mendelsohn et al., 1994). Short exposures to exogenous FtA appear to stimulate branching in murine lung cultures previously growth arrested with serum (Schuger et al., 1993). Recently we have shown that longer exposure (5-9 days treatment) of the 13.5 day embryonic rat lung to RA (10-6-10-5M) alters its normal developmental pattern in vitro. RA suppresses formation of the distal lung and induces a proximal-like phenotype in the lung epithelial tubules (Cardoso et al., 1995). Many of the retinoid effects on pattern formation are associated with changes in expression of patterning genes, such as Hox (Kessel and Gruss, 1991; Papalopulu et al., 1991) or Sonic hedgehog (Shh)(Riddle et al., 19931,which are expressed in the developing lung (Riddle et al., 1993; Bitgood and McMahon, 1995; reviewed in Cardoso, 1995). Hox genes encode transcriptional regulators involved in specification of cell fate along the anterior-posterior axis of the body (Krumlauf, 1994). Shh is a mouse homolog of the Drosophila segment polarity gene hedgehog that encodes a secreted protein implicated in patterning of the central nervous system, somites, and limb (Echelard et al., 1993; Laufer et al., 1994; Johnson and Tabin, 1995). RA has been shown to upregulate Hox mRNA in fetal lung explants (Bogue et al., 1994). Although Shh is known to be induced by exogenous RA in the limb mesenchyme, no information is available on how Shh responds to RA in the lung. In the present work, we hypothesized that the proximalized phenotype induced by exogenous RA in culReceived December l, 1995;accepted March 18, 1996. Address reprint requestskorrespondence to Wellington V. Cardoso, MD, PhD, The Pufmonary Center, Boston University School of Medicine, 80 East Concord Street, R-304,Boston, MA 02118. 48 CARDOSO ET AL. p-actin (20 ngheaction). The oligonucleotide primer sequences used for PCR amplification were as follows: Hoxa-2: 5' oligo: AAA CCC AGT GCA AGG AGA A, 3' oligo: GCT GAA ATA TCT ACA GGA CTG (465 bp, Patel et al., 1992);Hoxb-6: 5' oligo: GAG AGC AAA CTG CTC AGT GC, 3' oligo: GGA GAC ATC AAG TGA GTC CG (354 bp, Shen et al., 1991); p-actinr 5' oligo: ATG CCA TCC TGC GTC TGG A, 3' oligo: CAC ATC TGC TGG AAG GTG G (552 bp, Nude1 et al., 1983). Samples were PCR-amplified for 15 cycles as follows: 1 min at 95°C (melting), 1 min a t 55°C (hybridization), and 1.5 min at 72°C (extension); this protocol was shown previously to give a n optimal signal on a Southern blot within the linear range of amplification for the genes studied (data not shown). Amplified DNA was blotted onto nitrocellulose membranes (Hybond'" -N, MATERIALS AND METHODS Amersham, Arlington Heights, IL), and hybridized Embryonic Lung Cultures with random primer labelled probes (Feinberg and VoEmbryonic lungs were obtained from timed-preg- gelstein, 1983) a t 68°C overnight and washed at high nant Sprague-Dawley specific pathogen-free rats stringency (0.3M NaCl). Each filter was hybridized (Zivic-Miller, Zelionople, PA) and were sacrificed a t with probes for Hoxa-2 and Hoxb-6 and subsequently gestational day 13.5 according to standard procedures reprobed for P-actin. Autoradiographic bands were an(Cardoso et al., 1995). Lung explants were cultured for alyzed by densitometry (ImageQuant, Molecular Dy12 hr to 9 days in either a control defined, serum-free namics, Eugene, OR). The values obtained from the medium (BGJb, GIBCO, Gaithersburg, MI) with 0.2 densitometric measurements for each Hox gene were mg of ascorbic acid and 50 U / I J .of ~ penicillidstrepto- normalized for loading efficiency of the PCR by the mycin (GIBCO) (Slavkin et al., 1989), or the same me- corresponding p-actin band in the same lane of the dia containing all-trans-retinaic acid (Sigma, St. Louis, same filter. Densitometric measurements were exMO), diluted in ethanol at lOV5M,a concentration that pressed as % of control values. A statistical analysis has been previously shown to cause a marked change was performed (ANOVA, Fisher PLSDTest, Statview); in lung patterning in vitro (Cardoso et al., 1995). Har- differences were considered significant a t P < 0.05. vested cultures were photographed, and quick-frozen in Each experimental condition was tested several times liquid nitrogen (mRNA extraction) or fixed in 4% para- with samples from different litters. The number of samples "n" used in this procedure is shown in the formaldehyde (in situ hybridization studies). legend of Figure 4. Each sample (n) represents a pool of Quantitation of Hox mRNA Abundance by three lungs from a same litter. Reproducibility of the Southern Blot Analysis of RT/PCR Products method has been previously tested (Cardoso et al., Poly-A RNA was extracted from freshly isolated rat 1995). lungs (gestational day 13.5, 15.5, 18.5, and 21.5) or cultured lungs (control, RA-treated) which had been In Situ Hybridization Analysis Paraffin sections (6 bm) were obtained from freshly quick-frozen in liquid nitrogen (Micro-Fast Track mRNA Isolation Kit, Invitrogen, La Jolla, CAI. To isolated (gestational days 13-14, 18.5, and 21.5) and avoid contamination with genomic DNA, samples were 5-day cultured (control, RA-treated) lungs fixed in 4% treated with DNase prior to RT/PCR analysis. Products paraformaldehyde. After PCR amplification, Hoxu-2 of the reverse transcriptase (RT) reaction were labelled (465 bp, nucleotides 784-1,249, Patel et al., 1992) and with r3H]thymidine 5'-triphosphate, tetrasodium salt Hoxb-6 (354 bp, nucleotides 1,880-2,234, Shen et al., (f3H]-dTT,10 p,Ci/pl, specific activity of 100 Ci/mmol) 1991) rat cDNA fragments were cloned, respectively, to allow loading of equal counts and to equalize the into a PCR- I1 vector (TA cloning kit, Invitrogen) and amount of cDNA template used for each reaction. The an EcoRI site of a pBluescript KS+ vector. The Shh final concentration of dATP, CTP, GTP, in the RT re- cDNA consisted of a 640 bp EcoR-I fragment cloned action was 0.5 mM, whereas dTTP was 0.1 mM. Unin- into a pBluescript I1 SK of the chicken Shh gene that is corporated [3H]-dTTwas separated from cDNA by eth- highly homologous to the mouse Shh. In addition to anol precipitation. Approximately 4 ng of DNA were these genes we also assessed expression of surfactantused as template for PCR. For PCR we used 50 mM associated protein C (SP-C), a marker of distal lung KC1, 10 mM Tris (pH 8.31, 1.5 mM MgCl,, 0.01%gel- differentiation. SP-C probe consisted of a 574 bp cDNA atin, 1.25 M each of dATP, dCTP, dTTP, and dGTP, 2.5 (nucleotides 22-596, Fisher et al., 1989) cloned into a U of recombinant Taq DNA polymerase (Perkin Elmer pBluescript KS+. Recombinant Hoxa-2, Hoxb-6, Shh, Cetus, Norwalk, CT), and specific upstream and down- and SP-C plasmids were linearized to achieve transtream primers of Hoxa-2, Hoxb-6 (75 ngjreaction), and scription of sense and antisense templates. The tran- tured lung explants is associated with changes in pattern-related genes. Thus, we assessed mRNA expression of Shh and representative anterior and relatively posterior Hox genes Hoxa-2 and Hoxb-6 in vivo, in vitro, and in lungs whose development was altered by exogenous RA. We found that these genes are developmentally regulated in the lung, and that exposure of lung explants to 10-6M RA consistently upregulates Hoxu-2, Shh, and to a lesser extent Hoxb-6 mRNAs. We propose that RA may act to maintain high levels of expression of these and other pattern-related genes in a fashion that is characteristic of the immature lung, thereby promoting continued formation of proximal lung structures and preventing formation of the distal lung structures of the mature lung. RETINOIC ACID AND PATTERNING OF THE LUNG 49 Fig. 1. Airway branching in the embryonic rat lung in culture for 5 days under control conditions (A), in the presence of RA lO-%l (B), and after RA washout (C) (RA for 3 days, and in control medium for additional 2 days). RA treatment prevents formation of typical distal bud-like structures seen in controls and favors growth of proximal-like tubules. Distal buds typical of controls (arrows) branch from the proxirnalized tubules after RA is withdrawn. Whole mounts magnification: A, B, and C, x 25. scription reaction consisted of 300 ng of linearized plasmid template, 2 pl of 10 x transcription buffer (BRL), 2 p1 of 100 mM DTT, 0.8 U of RNAase inhibitor (RNAsin, Promega, Madison, WI),80 pCi t35Sl-UTP (40 mCi/ml stock, Amersham), 1 pl each of 10 mM stocks of ATP, CTP, GTP, in a total volume of 19 p1. Reaction was started by adding 1 p1 (20 U) of the appropriate RNA polymerase (BRL)that allowed synthesis of sense and antisense riboprobes. Samples were incubated at 37°C for 2 hr. Plasmid DNA template was digested by adding 1 ~1 of RNAase-free DNAase (Promega) and incubated a t 37°C for 15 min. The labeled probe was submitted to alkaline hydrolysis for 60 min and purified on G-50 Sephadex columns. Sections were then prehybridized and hybridized with sense and antisense RNA probes, being subsequently washed and treated with RNAase (Sassoon et al., 1988; Rishi et al., 1995). The slides were finally dehydrated, dried, and coated with Kodak NTB2 emulsion. Slides were developed in Kodak D-19 developer a t 1 and 3 weeks and photographed in a Leitz Orthoplan microscope. Each experimental condition was tested 2-3 times. of the epithelial tubules; however higher RA concentrations to 10-'M) suppress formation of distal tubules and favor growth of proximal-like tubules (Fig. 1B and Cardoso et al., 1995). In the present work we studied the lungs treated with RA lO-'M, in which the effects on patterning were most evident. These lungs, although showing proximal features of differentiation such as ciliated cells in the epithelium, appeared less mature than equivalent controls (compare proximal epithelia of day 13.5, control, and RA at day 9 in Fig. 2). The 10-5M RA-treated lung phenotype promptly reverted to the control phenotype by culturing the RAtreated lungs in control medium (Fig. 1C). Newly formed distal buds arising from the proximalized tubules showed the same features as distal tubules of controls (Fig. 1C and Fig. 9). Hox mRNA Quantitation In Vivo Densitometric measurements were performed on RT/ PCR Southern blots of rat lungs freshly isolated a t gestational days 13.5, 15.5, 18.5, and 21.5 and probed for Hoxa-2 and Hoxb-6 (Fig. 3). Hoxu-2 mRNA levels remained unchanged throughout development. Hoxb -6 levels decreased significantly and were almost undetectable at late gestation (Figs. 3 and 41, as previously shown by Northern analysis (Bogue et al., 1994). RESULTS Morphology of Control and RA-Treated Lungs Control cultured lungs exhibited airway branching and differentiation of both epithelium and mesenchyme, reproducing the overall proximal-to-distal pattern seen in lungs in vivo (Fig. 1A; see also Cardoso et Hox mRNA Quantitation In Vitro and Effects al., 1995).Briefly, proximal tubules (airways) are lined of RA The relative levels of Hoxa-2 and Hoxb-6 mRNA in by a tall columnar epithelium with ciliated and secretory cells (Fig. 2B) and are surrounded by either a control lungs a t days 3 and 5 mirrored the changes in smooth muscle layer or cartilage primordia. Distal ep- mRNA levels observed a t equivalent time points in ithelial tubules (putative bronchioles and alveoli) are vivo (Fig. 4).At lO-'M RA Horn-2 mRNA levels are acinus-shaped, with cuboidal cells and lack an under- significantly higher than controls at day 5. We could lying smooth muscle layer (not shown). Treatment of not detect statistically significant differences in H o d 8 mRNA levels between control and RA-treated lung explants with low RA concentrations ( lO-'M) does not affect the overall pattern of branching lungs by quantitative PCR (Fig. 4). 50 CARDOSO ET AL. Fig. 2. Histological features of the epithelia of a day 13.5 embryonic rat lung (A), proximal tubules of a 9-day control cultured lung (B), and a proximalired tubule of a 9-day IO-%l RA-treated lung (C).The epithelium of the RA-treated lung, although showing proximal features of differentiation such as the presence of ciliated cells (arrowheads), resembles the immature epithelium of the 13.5 day lung. Toluidine blue-stained plastic sections; magnification: A and C, x 650; B, x 950. In Situ Hybridization:Hox and Shh mRNA Expression In Vivo and In Vitro and Effects of RA Hoxa-2. Hoxa-2 mRNA is expressed in mesenchyma1 cells of the day 13-14 rat lung along the main axis of the respiratory tract; trachea and lung exhibited a similar intensity of signal (Fig. 5A). At later gestational ages, such as days 18.5 (Fig. 5C,D) and 21.5 (Fig. 5E,F), while Hoxu-2 mRNA levels remain unaltered (Fig. 4), its distribution changes, localizing mostly to the mesenchyme associated with cartilage, smooth muscle layer of airways, and vascular smooth muscle (Fig. 5C-F). Control cultured lungs a t day 5, similarly to day 18.5 fetal lungs, show Hoxu-2 mRNA signals in cartilage primordia and smooth muscle of large proximal airways (Fig. 5G). Distal mesenchyme of the cultured lungs also expressed Hoxu-2 mRNA, however diffusely distributed and at much lower levels than in proximal mesenchyme. Vascular expression was not seen in vitro because blood vessels appear not to develop in vitro. We assessed Hoxa-2 mRNA expression in lungs treated with lO-'M RA for 12 hr (see Fig. and days (Fig*5H)*At both time points treated lungs showed much higher H o ~ a - 2mRNA levels than equivalent control lungs. At day 5, the lung proximalto-distal pattern was altered and, in contrastto controls, Hoxa-2 mRNA was homogeneously distributed in the mesenchyme, with no preferential localization to particular structures of the lung (Fig. 5H). 7ApB) Fig. 3. Hox gene expression in embryonic rat lungs: representative Southern blots of rat cDNA from rat lungs. After RT/PCR amplification gels were blotted and hybridized to a Hoxa-2 and Hoxb-6 32P labelled cDNA probe. Filters were reprobed with p-actin probe. Hoxa-2 mRNA levels do not change throughout gestation, however Hoxb-6 mRNA significantly decrease in late lungs. 51 RETINOIC ACID AND PATTERNING OF THE LUNG --T i n vivo- W T BJ HOXZI-2 Hoxb-6 100 h s v 9 0 50 0 13.5 15.5 18.5 21.5 gestational d a y Hoxb-6 Hoxa-2 200 -^__ 100 ** 1 T z Y n s Y 0 0 100 ti rT- 0 50 0 0 13.5 C 3 C5 13.5 C 3 C5 R5 R5 Fig. 4. Hox gene expression in embryonic rat lungs in vivo and in vitro, and effects of RA. Histograms show quantitation of RT/PCR-Southern blots from freshly isolated embryonic rat lungs (gestational day 13.521.5) and cultured lungs in control medium (C3, C5) and in the presence of RA 10-5M (R5) for 5 days. Filters were probed with Hoxa-2 and Hoxb-6;the densitometric values obtained from Southern blots were normalized by respective p-actin values as shown in blots in Figure 3. Values (bars and lines representing mean and standard deviation) were expressed as percentage from the gestation day 13.5 lung value in both in vivo and in vitro studies. Number of samples "n" equals 3 (gestation day 15.5,C3,C5, and R5),4 (gestation days 13.5and 18.5), and 5 (gestation day 21.5). Hoxa-2 mANA levels do not change throughout gestation in vivo or with time in culture. Hoxb-6 levels significantly decrease with gestation age and time in culture. Five-day RA-treated lungs show significantly higher levels of Hoxa-2 mRNA than respective controls and day 13.5 lungs. Hoxb-6 mRNA levels in RA-treated lungs at day 5 are not significantly higher than respective controls. 'Significantly different (P < 0.05) from day 13.5 lung but not different from respective control; ""significantly different (P i0.05) from day 13.5 lung and respective control. Hoxb-6. Lungs of gestation day 13-14 show a gradient of Hoxb-6 mRNA expression in the respiratory tract. Background levels of grains are seen in trachea and proximal lung; high levels of expression are found in mesenchymal cells associated with the distal epithelial tubules (Fig. 6A). At late stages of development, days 18.5 (Fig. 6C,D) and 21.5 (not shown) fetal lung expresses progressively lower levels of Hoxb-6 mRNA. Signals are almost undetectable in the 18.5-day lung (Fig. 6C,D). Similarly Hoxb-6 signals are barely detectable in the mesenchyme of 5-day control cultured lungs, and the proximal-distal mRNA gradient previously observed is no longer seen (Fig. 6E).Short exposures to 10m6MRA (12 hr) appears to expand slightly the domain of Hoxb-6 expression to a more proximal location (Fig. 7D).In contrast to the results obtained by quantitative PCR (Fig. 41, in situ hybridization of 5-day RA-treated lungs appear to display much stronger Hoxb-6 mRNA signals than 5-day controls (Fig. 6E,F). Signals are no longer localized to distal lung as in the 13.5 day lung, but are seen throughout the mesenchyme of the entire RA-treated explant. Shh. High levels of Shh mRNA expression are observed in the epithelia of the 13-14 day rat esophagus and the respiratory tract (arrows in Fig. 8A). Shh transcripts are not detected a t sites other than the epithelium, and signals are similar in trachea and lung; thus no proximal-distal gradient was identified (Fig. 8A). Days 18.5 (Fig. 8B,C) and 21.5 (Fig. 8D,E) fetal lungs and trachea had progressively less Shh expression in 52 CARDOSO ET AL. Fig. 5. RETINOIC ACID AND PATTERNING OF THE LUNG their epithelia. Low levels of Shh transcripts, comparable to day 18.5 lungs, are also found in 5-day control cultured lungs (Fig. 8F).Exposure to 10m5MRA for 12 hours did not seem to alter the preexisting high levels of expression of Shh mRNA of the early cultured lung (data not shown). In contrast, RA exposure for 5 days strongly upregulated Shh in the lung epithelium and prevented the normal decrease in Shh seen in controls (Fig. 8G). When RA is washed out, distal buds form (Fig. lC, and also indicated by arrowheads in Fig. 9) and appear similar to the distal lung buds of control lungs, expressing low levels of epithelial Shh mRNA (arrowheads in Fig. 9B). Areas that retain the morphologic features of the proximalized lung (large arrows in Fig. 9) continue to express high levels of Shh (large arrow in Fig. 9B). SP-C. We have previously shown by in situ hybridization that SP-C mRNA is localized to the distal tubules of control lungs and that its expression is suppressed by 10p5MRA treatment (Cardoso et al., 1995). In lungs where RA effects have been reversed by washing out RA, the newly formed buds express high levels of SP-C mRNA (arrowheads in Fig. 9A). DISCUSSION We have previously shown that RA affects the pattern of lung formation during embryonic development, tending to proximalize the developing lung and to prevent formation of distal lung tubules (Cardoso et al., 1995). The demonstration of RARs and retinoid binding proteins in the developing lung (Dolle et al., 1990) suggests that RA-responsive signalling pathways may participate in the structural and biochemical alterations we have observed. The potential role of RA in lung development has been demonstrated by double null mutations of RARs that result in the absence or hypoplasia of the lung (Mendelsohn et al., 1994).These observations lead us to propose that RA and associated molecules may play important roles in normal lung development. A number of studies of developing systems suggest that RA acts in part via its ability to influence the site and time of expression of Hox genes. Hox genes are present in the lung a t early stages of development (reviewed in Cardoso, 1995), although it has not been shown that lung development depends on any single ~~ Fig. 5. In situ hybridization of Hoxa-2 mRNA expression in the developing rat lung at gestational days 13.5 (A, anti-sense; B, sense strand negative control) and 18.5 (C, antisense; D, phase contrast), trachea at gestational day 21 (E, antisense; F, phase contrast), control 5-day cultured lung (0)and 5-day 10-5MRA-treated lungs (H).Hoxa-2 mRNA (arrows) is diffusely expressed in the mesenchyme of the 13.5 day lung. Later in gestation and in culture Hoxa-2 is abundant in the mesenchyme associated with smooth muscle layer of vessels (v), airways (a), and cartilage (ct) in proximal airways; low Hoxa-2 signal is also seen diffusely in distal mesenchyme of control cultured lungs. Under RA-treatment Hoxa-2signal is strongly upregulated in the mesenchyme along the entire proximal-distal axis of the lung. Magnification: A and 8, x 70; C and D, x95: E and F, x50; G, x55; H, x60. 53 homeobox gene, as in the case of the spleen (Roberts et al., 1994) or on a Hox code established by overlapping domains of Hox gene expression, as in the case of the limb (Hunt et al., 1991, Izpisua-Belmonte et al., 1990). The main purpose of our study was to see if changes in lung patterning were associated with changes in expression of pattern-related genes. Thus, we designed the study to investigate Hox and Shh gene expression a t the RA concentration that generated an easily recognizable patterning phenotype, without evidence of cellular toxicity (i0-5M). Although we expected to find an overall increase in Hox gene expression in our system upon RA administration, we chose to study the RA effects on the expression of one anterior (Hoxa-2) and a relatively posterior (Hoxb-6)Hox gene in the lung because they would be expected to respond differently to RA (Simeone et al., 1990) and because they have different patterns of expression along the respiratory tract; Hoxa-2 is present in the embryonic trachea and lung, while Hoxb-6 is confined to the lung itself (reviewed in Cardoso, 1995). The in vivo and in vitro patterns of Hox gene expression seen in our study suggest that these genes play distinct roles in lung development. Hoxa-2 appears to participate in the formation of proximal mesenchymal derivatives and vasculogenesis because its expression is progressively restricted to the mesenchyme associated with differentiating cartilage primordia, airway smooth muscle, and vessels. Hoxb-6 may be part of a complex network that regulates lung distal epithelial branching during development because its expression appears to be spatially and temporally associated with the process of formation of distal lung tubules. We speculate that such a network, like that described in other systems (Rancourt et al., 1995), may involve cooperativity with other homeobox genes, such as H o x b d , which share the same pattern of expression with Hoxb-6 (Holland and Hogan, 1988; Volpe and Nielsen, 1994), and the thyroid transcription factor-1 (TTF-1) (Lazzaro et al., 19911,as well as several growth factors. One of the problems of implicating Hox genes individually in lung development is that there is probably functional overlap of these genes. Perhaps because of such redundancy, targeted knockout of either Hoxa-2 or Hoxb-6 results in no abnormal lung phenotype (Rijli et al., 1993; Gendron-Maguire et al., 1993; Rancourt et al., 1995). Shh is a signalling molecule in limb bud morphogenesis, present in the limb mesenchyme. Shh mRNA is activated by RA and induces Hoxd gene expression (Riddle et al., 1993; reviewed in Johnson and Tabin, 1995). Shh mRNA and protein are expressed at many sites of epithelial-mesenchymal interactions, such as the branchial arches, gut, respiratory and urinary tract in the mouse embryo, where its role is still unclear (Echelard et al., 1993; Bitgood and McMahon, 1995; Marti et al., 1995). We found that Shh mRNA is expressed at high levels in the epithelium of the day 13-14 rat lung and trachea, with no proximal-distal 54 CARDOSO ET AL. Fig. 6. In situ hybridization of Hoxb-6 mRNA expression in gestational day 13.5 rat lung (A, antisense; 6,sense), gestational day 18.5 lung (C, antisense; D, phase contrast), control 5-day cultured lung (E) and 5day 10M6MRA-treated lungs (F).Hoxbd signals (large arrows) are restricted to the distal mesenchyme of the 13.5 day lung. Hoxb-6 mRNA levels are almost undetectable later in gestation and in culture. Hoxb-6 is upregu- lated in RA-treated lungs; no proximal-distal gradient of expression is observed. The coarse grains of variable size (small double arrows in C and E) are not true signal, and result from tissue birefringency. tr, trachea; a, airways in the lung in vivo and in vitro. Magnification: A-D, x 80; E and F, x 60. RETINOIC ACID AND PATTERNING OF THE LUNG 55 Fig. 7. Effect of 12-hr exposures of the lung to RA 10-5M on Hoxa-2 (A, 8 ) and Hoxb-6 (C, D) mRNA expression. Control (A, C). RA-treated (6,D) lungs. Short exposures to RA do not alter patterning of the lung; however, they markedly upregulate Hoxa-2 mRNA signals (arrows in A and B) along its entire proximal-distal axis. RA appears to extend ff~xb-6expression (arrows in C and D)to a slightly more proximal site. Magnification: A-D, x 50. mRNA gradient. Lung maturation in vivo and in vitro was associated with a decline in the Shh levels in the epithelium; however in the presence of exogenous RA, lung explants maintain Shh expression a t high levels with a distribution similar to that seen in day 13-14 lungs (Fig. 8A and G). RA therefore prevents the normal decline in Shh mRNA seen a t late time points in controls, and appears to hold the lung in a proximal immature pattern. This hypothesis is further supported by histologic features of immaturity seen in the proximalized epithelium (Fig. 2). Interestingly, at late times in culture both control and RA-treated lungs express Shh and SP-C mRNA in distal epithelium in an inverse fashion; when Shh is upregulated, SP-C is downregulated and vice versa. In contrast, distal buds of uncultured day 13.5 lungs simultaneously express Shh (Fig. 8A) and SP-C (data not shown, Wert et al., 1993)mRNAs, and likely both genes are present in the same cells in the early lung bud. When RA is removed from the culture media, newly formed epithelial buds 56 CARDOSO ET AL. Fig. 8. In situ hybridization of Shh mRNA expression in the developing rat lung at gestational days 13.5 (A, inset shows isolated lung) and 18.5 (8, antisense; C, phase contrast), gestational day 21.5 trachea (D, antisense; E, phase contrast), control 5-day cultured lung (F), and 5-day l 0 - W RA-treated lungs (G, antisense; H, sense strand negative control). Shh signals (arrows) are very strong in the epithelium of the day 13.5 esophagus (es), trachea (tr), and lung (lu). Shh mRNA is uniformly ex- pressed along the proximal-distal axis of the respiratory tract. Shh signals decrease in the airway epithelium of the lung and trachea later in gestation (8, D) and after 5 days in culture (F). RA markedly increases Shh mRNA signals in the lung epithelium (G), and does not alter the proximaldistal mRNA expression pattern seen in controls. Magnification: A, F. G, and H, x60; inset, x80; B and C, x70; D and E, x50. express high levels of SP-C and low levels of Shh mRNAs (Fig. 9) in a fashion that is comparable to controls of the same age. Thus it seems that RA withdrawal releases the lung from the proximalized pattern and allows newly formed distal buds to express Shh and SP-CmRNAs in a fashion that is typical of mature lungs. We have shown that RA rapidly alters Hox gene expression, prior to changes in lung morphology and SP-C mRNA suppression. We hypothesize that such early changes in Hox expression are a major determinant of subsequent alterations in developmental events. Longer exposures to RA showed more marked changes in both lung patterning and Hox gene expression. It is likely that in our model, RA acts by maintaining prolonged high levels of Hoxu-2, Hozb-6, and other RA-responsive patterning genes from the Ho3t family as well as Shh. The proximalizing phenomenon may involve retinoid signalling molecules such as RAR-p and RAR-y which are characteristically expressed in the proximal respiratory tract a t early stages of the development (Dolle et al., 1990). Difference in Hoxb-6 expression between day 5 control and RA-treated samples appeared greater by in situ hybridization than by PCR analysis. Such difference may be ascribed to several factors and includes: (1)variability in the response of the explants to RA in the PCR analysis, as shown by standard deviations in Figure 4; the mean value of Hoxb-6 expression in RA treated samples appears twice that of controls and, likely, an increase in the number of samples could demonstrate a statistically significant difference between RETINOIC ACID AND PATTERNING OF THE LUNG 57 Fig. 9. In situ hybridizationof Sf-C(A)and Sbh(B) mRNA expression in lungs treated with RA lO-’M for 3 days, washed out and cultured in control medium for additional 2 days. Newly formed distal buds (arrowheads in A, B, and also shown in Fig. 1C) express Sf-C, a marker of distal lung epithelium and Shh mRNA in the same fashion of control 5-day cultured lungs (see Shh expression in distal lung shown in Fig. 8F and Fig. 9s). The remaining proximalized tubules (large arrow in A and 6)do not express SP-C (Fig. 9A), but express high levels of Shh mRNA (Fig. 9B) characteristic of the RA-treated lung epithelium (shown in Fig. 8G). Magnification: A and B, x 65. these groups; (2) Apparent non-linearity of in situ hybridization method on detection of signals, so that increases in Hoxb-6 expression appear more noticeable when the baseline signal is low (as in controls); (3) changes in the proportion of epithelium vs. mesenchyme; there is less mesenchyme in controls than in RA-treated explants, thus, signals by in situ hybridization appear more prominent; (4) Lower sensitivity of the PCR analysis, as compared to in situ hybridization on detection of changes in Hoxb-6, which is less inducible by RA than other 3’ genes (such as Hoxu-2) (Simeone et al., 1990). We have not yet characterized the molecular sequence of events and roles that Hox and Shh genes play in lung patterning, but it is clear that exogenous RA profoundly affects this process. Good candidates for participating in this network are members of the fibroblast growth factor (FGF) family and their receptors. Limb bud formation appears to be regulated by a network involving endogenous RA, Shh, Hoxb, and Hoxd genes, as well as FGF-4 and FGF-8 (reviewed in Tabin, 1995). FGFs and FGF receptors are expressed in the developing lung (Stark et al., 1991; Peters et al., 1992; Han et al., 1992;Mason et al., 1994; Finch et al., 1995,) and appear to have a crucial role in lung epithelial branching (Peters et al., 1994; Nogawa and Ito, 1995). Our studies suggest that molecular paradigms developed for skeletal structures, limb bud, and brain development apply in general to the development of branch- ing visceral organs such as the lung. The proximalized lung is intended to be a model where pattern formation is altered; it does not reproduce a physiological situation, However, this model allows us to infer that high concentrations of RA at early stages of development are incompatible with distal lung formation, fostering a more immature proximal phenotype. ACKNOWLEDGMENTS We thank Dr. Andrew P. McMahon (Harvard Biolabs) for the gift of the Shh cDNA clone. We thank Barbara Miranda and Anne Hinds for technical assistance. This work was supported by grants from the National Institute of HealtWNHLBI (PO1 HL47049 and R01 HL34768) and the Parker B. Francis Foundation. REFERENCES Bitgood, M.J., and McMahon, A.P. (1995) Hedgehog and Bmp genes are coexpressed at many diverse sites of cell-cell interaction in the mouse embryo. Dev. Biol. 172:126-138. Gross,I., Vasavada, H., Dynia, D. W., Wilson, C. M., and Bogue, C. W., Jacobs, H. C. (1994) Identification of H a genes in newborn lung and effects of gestational age and retinoic acid on their expression. Am. J. Physiol. 266:448-454. Cardoso, W. V.(1995) Transcriptionfactors and pattern formation in the developing lung. Am. J. Physiol. 269:L429-L442. Cardoso, W. V.,Williams, M. C., Mitsialis, S. A., Joyce-Brady, M., Rishi, A. K., and Brody, J. S. (1995) Retinoic acid changes the pattern of airway branching and alters epithelial cell differentiation in 58 CARDOSO ET AL. the developing lung in vitro. Am. J. Respir. Cell Mol. Biol. 12464476. Chytil, F. (1992) The lungs and vitamin A. Am. J. Physiol. 262:L517L527, Dolle, P., Ruberte, E., Leroy, P., Morriss-Kay, G., and Chambon, P. (1990) Retinoic acid receptors and celular retinoid binding proteins. I. A systematic study of their differential pattern of transcription during mouse organogenesis. Development 1101133-1 151. Echelard, Y., Epstein, D. J., StJacques, B., Shen, L., Mohler, J., McMahon, J., and McMahon, A. P. (1993) Sonic hedgehog, a member of a family of putative signaling molecules, is implicated in the regulation of CNS polarity. Cell 75:1417-1430. Feinberg, A. P., and Vogelstein, B. (1983) A technique for radiolabelling DNA restriction endonucleasefragments to high specific activity. Anal. Biochem. 132:6-13. Finch, P. W., Cunha, 0.R., Rubin, J. S., Wong, J., and Ron, D. (1995) Pattern of keratinocyte growth factor and keratinocyte growth factor receptor expression during mouse fetal development suggests a role in mediating morphogenetic mesenchymalepithelial interactions. Dev. Dyn. 2033223-240. Fisher, J . H., Shannon, J. M., Hofman, T., and Mason, R. J. (1989) Nucleotide and deduced amino acid sequence of the hydrophobic surfactant protein SP-C from r a t Expression in alveolar type I1 cells and homology with SP-Cfrom other species. Biochim. Biophys. Acta 995:225-230. Gendron-Maguire, M., Mallo, M. Zhang, M., and Gridley, T. (1993) Horn-2 mutant mice exhibit homeotic transformation of skeletal elements derived from cranial neural crest. Cell 751317-1331. Gudas, L. J. (1994) Retinoids and vertebrate development. J. Biol. Chem. 269:15399-15402. Han, R. N., Liu, J., Tanswell, K., and Post, M. (1992) Expression of basic fibroblast growth factor and receptor: Immunolocalization studies in developing rat fetal lung. Pediatr. Res. 31:435-440. Hill, D. S:,Ragsdale, C. W., and Brockes, J. P. (1993) Isoform-specific immunological detection of newt retinoic acid receptor 6 in normal and regenerating limbs. Development 117:937-945. Holland, P. W. H., and Hogan, B. L. M. (1988) Spatially restricted patterns of expression of the homeobox-containing gene H m 2.1. during mouse embryogenesis. Development 102159-174. Hunt, P., Gulisano, M., Cook, M., Sham, M., Faiella, A., Wilkinson, D., Boncinelli, E., and Krumlauf, R. (1991) A distinct Hox code for the branchial region of the head. Nature 3535361-864. Izpisua-Belmonte,J.4,Tickle, C., Dolle, P., Wolpert, L., and Duboule, D.(1991) Expression of the homeobox Hox-4 genes and the specification of position in chick wing development. Nature 350:585-589. Johnson, R. L., and Tabin, C. (1995) The long and short of hedgehog signaling. Cell 81:313-316. Kessel, M., and Gruss, P. (1991) Homeotic transformations of murine vertebrae and concomitant alterations of Hox codes induced by retinoic acid. Cell 67:89-104. Krumlauf, R. (1994) €€ox genes in vertebrate development. Cell 78: 191-201. Laufer, E., Nelson, C. E., Johnson, R. L., Morgan, B., and Tabin, C. (1994) SORCChedgehog and Fgf-4 act through a signaling cascade and feedback loop to integrate growth and patterning of the developing limb bud. Cell 79:993-1003. Lazzaro, D., Price, M., de Felice, M., and dilauro, R. (1991) The transcription factor TTF-1 is expressed at the onset of thyroid and lung morphogenesis and in restricted regions of the fetal brain. Development 113:1093-1 104. Marti, E., Takada, R., Bumcrot, D.A., Sasaki, H., and McMahon, A.P. (1995) Distribution of Sonic hedgehog peptides in the developing chick and mouse embryo. Development 121:2537-2547. Mason I. J., Fuller-Pace, F., Smith, R., and Dickson, C . (1994) FGF-7 (keratinocyte growth factor) expression during mouse development suggests roles in myogenesis, forebrain regionalization and epithelial-mesenchymal interactions. Mech. Dev. 4515-30. Mendelsohn, C., Lohnes, D., Decimo, D., Lufiin, T., LeMeur, M., Chambon, P., and Mark, M. (1994) Function of retinoic acid receptors (RARs)during development. (11) Multiple abnormalities a t various stages of organogenesis in RAR double mutants. Development 1202749-2771. Nogawa, H., and Ito, T. (1995) Branching morphogenesisof embryonic mouse lung epithelium in mesenchyme-free culture. Development 121:lOlS-1022. Nudel, U., Zakut, R., Shani, M., Neuman, S., Levy, Z., and Yaffe, D. (1983) The nucleotide sequence of rat cytoplasmic p-actin gene. Nucleic Acids Res. 11:1759-1771. Papalopulu, N., Clarke, J. D. W., Bradley, L., Wilkinson, D., Krumlauf, R., and Holder, N. (1991) Retinoic acid causes abnormal development and segmental patterning of the anterior hindbrain in Xenopus embryos. Development 113:1145-1158. Patel, C. V., Gorski, D. H., LePage, D. F., Lincecum, J., and Walsh, K. (1992) Molecular cloning of a homeobox transcription factor from adult aortic smooth muscle. J. Biol. Chem. 67:26085-26090. Peters, K.G., Chen, W.G., and Williams, L.T. (1992) T w o FGF receptors are differentially expressed in epithelial and mesenchymal tissues during limb formation and organogenesis in the mouse. Development 114933-243. Peters, K., Werner, S., Liao, S., Wert, S., Whitsett, J., and Williams, L. (1994) Targeted expression of a dominant negative FGF receptor blocks branching morphogenesis and epithelial differentiation of the mouse lung. EMBO J. 133296-3301, Rancourt, D. E.,Tsuzuki, T., and Capecchi, M. (1995) Genetic interaction between Hoxb-5 and Hmb-6 is revealed by nonallelic noncomplementation. Genes Dev. 9108-122. Riddle, R. D., Johnson, R. L., Laufer, E., and Tabin, C. (1993) Sonic hedgehog mediates the polarizing activity of the ZPA. Cell 75:14011416. Rijli, F. M., Mark, M., Lakkaraju, S., Dierich, A., Dolle, P., and Chambon, P. (1993) A homeotic transformation is generated in the rostra1 branchial region of the head by disruption of Horn-2, which acts as a selector gene. Cell 751333-1349. Rishi, A. K., Joyce-Brady, M., Fisher, J., Dobbs, L. G., Floros, J., Vanderspek, J., Brody, J. S., and Williams, M. C. (1995) Cloning, characterization, and developmental expression of a rat lung alveolar type I cell gene in embryonic endodermal and neural derivatives. Dev. Biol. 167:294-306. Roberts, C. W. M., Shutter, J. R., and Korsemeyer, S. J. (1994) H o d 1 controls the genesis of the spleen. Nature 368:747-749. Sassoon, D. A., Garner I., and Buckingham, M. (1988) Transcripts of a-cardiac and a-skeletal actins as early markers of myogenesis in the mouse embryo. Development 104155-164. Schuger, L., Varani, J., Mitra R. Jr., and Gilbride, K. (1993) Retinoic acid stimulates mouse development by a mechanism involving epithelial-mesenchymal interaction and regulation of epidermal growth factor receptors. Dev. Biol. 159:462-473. Shen, W. F., Detmer, K., Simonitch-Eason,T. A., Lawrence, H. J., and Largman, C. (1991) Alternative splicing of the Hox 2.2 homeobox gene in human hematopoietic cells and murine embryonic and adult tissues. Nucleic Acids Res. 19539-645. Shenefelt, R. (1972) Morphogenesis of malformations in hamsters caused by retinoic acid: Relation to dosage and stage at treatment. Teratology 5:103-1 18. Simeone, A., Acampora, D., Arcioni, L, Andrews, P. W., Boncinelli, E., and Mavilio, F. (1990) Sequential activation of Hor2 homeobox gene by retinoic acid in human embryonal carcinoma cells. Nature 346:763-766. Slavkin, H. C., Johnson, R., Oliver, P., Bringas P. Jr., Don-Wheeler, G., Mayo, M., and Whitsett, J. A. (1989) Lamellar body formation precedes pulmonary surfactant apoprotein expression during embryonic mouse lung development in vivo and in vitro. Differentiation 41:223-236. Stark, K. L., McMahon, J. A,, and McMahon, A. P. (1991) FGFR-4, a new member of the fibroblast growth factor receptor familily, expressed in the definitive endodem and skeletal muscle lineages of the mouse. Development 113:641-651. Tabin, C. (1991) Retinoids, homeoboxes and growth factors: Toward molecular models for limb development, Cell 66:199-217. Tabin, C. (1995) The initiation of the limb bud: Growth factors, Hox genes and retinoids. Cell 80:671-674. Volpe, M. V., and Nielsen, H. C. (1994) The regulation of Hox-2.1 protein in mouse lung during development and by growth factors. Pediatr. Res. 35798, 1994. RETINOIC ACID AND PATTERNING OF THE LUNG 59 Wed, S., Glasser, S. W., Korfhagen, T. R., and Whitsett, J. A. (1993) Wilson, J. G., Roth, C. B., and Warkany, J. (1953)An analysis of the syndrome of malformations induced by maternal vitamin A defiTranscriptional elements from the human SP-Cgene direct expresciency. Effects of restoration of vitamin A at various times during sion in the primordial respiratory epithelium of transgenic mice: gestation. Am. J. Anat. 92:189-217. Dev. Biol. 156:426-443.