Repulsive axon guidance molecule Sema3A inhibits branching morphogenesis of fetal mouse lung

Mechanisms of Development 97 (2000) 35±45

www.elsevier.com/locate/modo

Repulsive axon guidance molecule Sema3A inhibits branching morphogenesis of fetal mouse lung Takaaki Ito a, Masako Kagoshima b, Yukio Sasaki b, Chanxia Li b, Naoko Udaka a, Takashi Kitsukawa c, Hajime Fujisawa c, Masahiko Taniguchi d, Takeshi Yagi e, Hitoshi Kitamura a, Yoshio Goshima b,f,* a Department of Pathology, Yokohama City University School of Medicine, 3-9 Fuku-Ura, Kanazawa-ku, Yokohama 236-0004, Japan Department of Pharmacology, Yokohama City University School of Medicine, 3-9 Fuku-Ura, Kanazawa-ku, Yokohama 236-0004, Japan c Group of Developmental Neurobiology, Division of Biological Science, Nagoya University Graduate School of Science, Chikusa-ku, Nagoya 464-8602, Japan d Department of Biochemistry and Molecular Biology, Graduate School of Medicine, University of Tokyo, Tokyo 110-0033, Japan e Department of Neurobiology and Behavioral Genetics, National Institute for Physiological Sciences, Myodaiji, Okazaki 444, Japan f CREST, Japan Science and Technology Corporation (JST), Kawaguchi 332-0012, Japan b

Received 25 February 2000; received in revised form 15 May 2000; accepted 26 June 2000

Abstract Semaphorin III/collapsin-1 (Sema3A) guides a speci®c subset of neuronal growth cones as a repulsive molecule. In this study, we have investigated a possible role of non-neuronal Sema3A in lung morphogenesis. Expression of mRNAs of Sema3A and neuropilin-1 (NP-1), a Sema3A receptor, was detected in fetal and adult lungs. Sema3A-immunoreactive cells were found in airway and alveolar epithelial cells of the fetal and adult lungs. Immunoreactivity for NP-1 was seen in fetal and adult alveolar epithelial cells as well as endothelial cells. Immunoreactivity of collapsin response mediator protein CRMP (CRMP-2), an intracellular protein mediating Sema3A signaling, was localized in alveolar epithelial cells, nerve tissue and airway neuroendocrine cells. The expression of CRMP-2 increased during the fetal, neonate and adult periods, and this pattern paralleled that of NP-1. In a two-day culture of lung explants from fetal mouse lung (E11.5), with exogenous Sema3A at a dose comparable to that which induces growth cone collapse of dorsal root ganglia neurons, the number of terminal buds was reduced in a dose-dependent manner when compared with control or untreated lung explants. This decrease was not accompanied with any alteration of the bromodeoxyuridine-positive DNA-synthesizing fraction. A soluble NP-1 lacking the transmembrane and intracellular region, neutralized the inhibitory effect of Sema3A. The fetal lung explants from neuropilin-1 homozygous null mice grew normally in vitro regardless of Sema3A treatment. These results provide evidence that Sema3A inhibits branching morphogenesis in lung bud organ cultures via NP-1 as a receptor or a component of a possible multimeric Sema3A receptor complex. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Branching morphogenesis; Lung; Sema3A; Neuropilin-1; CRMP-2

1. Introduction Lung development during the fetal period involves cell proliferation, migration, branching morphogenesis and cell differentiation, and the ontogenic sequence of these events in lung organogenesis needs to be well coordinated. Lung branching morphogenesis during development occurs in a stereotyped manner, providing a good model for analyzing the molecular and cellular mechanisms of morphogenesis (Serra et al., 1994; Nogawa and Ito, 1995; Shiratori et al., 1996; Hogan and Yingling, 1998). The development of the * Corresponding author. Fax: 181-45-785-3645. E-mail address: [email protected] (Y. Goshima).

lung appears to depend on precise signaling between epithelial cells and mesenchymal cells, and the signaling molecules, both known and unknown, should be in concert with each other in a complex but ordered manners. SemaphorinIII/collapsin-1 has been well documented as an axon guidance molecule mediating repulsive and inhibitory navigation of neurons (Kolodkin et al., 1993; Luo et al., 1993; Wright et al., 1995). SemphorinIII/collapsin-1 potently induces growth cone collapse of a speci®c subset of neurons and has been characterized most extensively at the molecular and cellular levels (Fan et al., 1993; Kolodkin et al., 1993; Luo et al., 1993; Wright et al., 1995; Goshima et al., 1995, 1997, 1999, 2000). SemaphorinIII/ collapsin-1 is assigned to the secreted type of semaphorin class 3 and is

0925-4773/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(00)00401-9

36

T. Ito et al. / Mechanisms of Development 97 (2000) 35±45

now termed Sema3A. The semaphorins share a Sema domain, a region of sequence conservation spanning about 500 residues, which is potentially important for the function of the proteins. Neuropilin-1 (NP-1) has been shown to be a receptor, or a component of a receptor, for Sema3A (He and Tessier-Lavigne, 1997; Kolodkin et al., 1997). NP-1 and NP-2, an NP-1- related protein, are expressed by developing neuronal and non-neuronal cells, and each of these NP subtypes is expressed in a pattern that is roughly non-overlapping (Chen et al., 1997; Giger et al., 1998). Both NPs and the semaphorin family members have been expressed in neuronal and non-neuronal tissues including the lung, heart and bone (Luo et al., 1993; Behar et al., 1996; Takahashi et al., 1997; He and Tessier-Lavigne, 1997; Kolodkin et al., 1997; Giger et al., 1998), and this suggests a possible role of NP and semaphorin interaction in the morphogenesis of non-neuronal as well as neuronal tissues. In an attempt to characterize the components of the collapsin-1/Sema3A pathway, we analyzed a Xenopus laevis oocyte expression system and isolated a collapsin response mediator protein (CRMP-62) from chick dorsal root ganglion. CRMP is an intracellular protein, and seems to be required for Sema3A signaling (Goshima et al., 1995). CRMP-62 shares homology with the C. elegans protein UNC-33 (Goshima et al., 1995; Li et al., 1992). In the rat, rCRMP-1, -2, -3 and -4 have been isolated, being differentially expressed almost solely in the nervous system (Wang and Strittmatter, 1996). CRMPs are also referred to as TOAD-64 (turned on after division, 64 kDa) DRP (DHP related protein), and Ulip (Unc-33-like phosphoprotein) (Minturn et al., 1995; Hamajima et al., 1996; Byk et al., 1996; Kamata et al., 1998). Of these, rCRMP-2 is most closely related to chick CRMP-62 and is the most widely expressed CRMP within the nervous system. The only exception of the neuronal speci®city of CRMP expression is a barely detectable level of lung rCRMP-2 with Northern blot analysis (Wang and Strittmatter, 1996). In addition, several small cell lung cancer cell lines exhibit homozygous deletions in chromosome region 3p21.3 where Sema3A-related secreted semaphorins, Sema3B and human Sema3F, are located (Roche et al., 1996; Sekido et al., 1996; Xiang et al., 1996). This raises the possibility that some members of the semaphorin family could function in regulation of cell motility, proliferation and differentiation during morphogenesis and/or tumorigenesis in the lung. All of the above ®ndings suggest a possible role of an interaction between Sema3A and NP-1 in the development of the fetal lung. In the present study, we have investigated fetal and adult mouse lungs with immunohistochemistry for Sema3A, NP-1 and CRMP-2, and examined the effect of Sema3A on branching morphogenesis in the explant culture of fetal mouse lung. We have further investigated whether there exist phenotypic changes in branching morphogenesis and responsiveness to Sema3A in lung tissues from neuropilin-1 (np1) mutant mice. A preliminary report of this work

was presented elsewhere (Kagoshima-Maezono et al., 1998). 2. Results 2.1. Immunohistological localization of Sema3A, NP-1 and CRMP-2 in embryonic and adult lungs To explore possible role of Sema3A and its receptor NP-1 in lung branching morphogenesis, we tried immunohistochemical detection for Sema3A, NP-1 and CRMP-2 in developing lung. In the embryonic lungs, Sema3A immunoreactivity was weak in the developing epithelium of embryonic day 12 (E12) mouse (data not shown). In E14 mouse, Sema3A immunoreactivity was strong in the terminal airway epithelial cells, but was weak in the large airway epithelial cells (Fig. 1A). Sema3A immunoreactivities were seen in the bronchiolar and alveolar epithelia in the lungs of late gestational and adult periods (Fig. 1C). A slight but signi®cant Sema3A immunoreactivity was detected in endothelial cells of bronchiolar arteries (Fig. 1C). NP-1 immunoreactivity of fetal lung epithelia was weak in E12 (data not shown), and was strongly positive in the developing alveolar epithelial cells in late fetal days (Fig. 1D). The NP-1 immunoreaction in the adult lungs was predominantly seen in the alveolar cells, but not in the bronchiolar cells (Fig. 1E,F). Endothelial cells showed immunohistochemical staining for NP-1 (data not shown). CRMP-2 immunoreactivity was seen in the terminal airway epithelium by E14 (Fig. 1G,H). In the fetal and adult lungs, nerve tissues and airway neuroendocrine cells were positively stained for CRMP-2. With serial sections stained for CRMP-2 and calcitonin gene-related peptide (CGRP), CRMP-2-positive airway epithelial cells gave positive CGRP-immunoreactivity (data not shown), and the CRMP-2-positive airway epithelial cells are considered to be neuroendocrine cells. Positive CRMP-2 immunoreactivity in pulmonary neuroendorine cells suggests a kind of neuronal phenotype expression of this cell type, as neuroendocrine cells share common neuronal phenotypes (Ito, 1999). In the negative control sections, no speci®c immunohistochemical reactions were obtained, and peptide competition control immunoreactions were non-speci®c for Sema3A (Fig. 1B) and CRMP-2 (data not shown). 2.2. RT-PCR for Sema3A and NP-1 and Western blot analysis for CRMP-2 Semiquantitative RT-PCR revealed the presence and developmental changes of mRNA for Sema3A and NP-1 in fetal and adult lung tissue. Each of the PCR products showed a single band, which corresponded to the size predicted for the corresponding transcript. No speci®c products for RT-PCR reaction without reverse transcriptase were detectable by gel electrophoresis (Fig. 2). The pro®le of the reactions appeared to decrease with age in Sema3A

T. Ito et al. / Mechanisms of Development 97 (2000) 35±45

37

Fig. 1. Immuno¯uorescence staining for Sema3A (A±C), NP-1 (D,E), CRMP-2 (G) and Clara cell 10 kDa protein (F,H). (A) Strong Sema3A immunostaining (yellow) is seen in the developing terminal epithelium (t) on E14. Lb (lobar bronchus). Counterstained with phalloidin (red) £300. (B) A serial section of (A) with peptide competition. The positive immunostaining for Sema3A seen in (A) is abolished by incubation of the Sema3A antibody with the corresponding peptide. Red blood cells are non-speci®cally stained (arrow). Counterstained with phalloidin (red) £300 (scale bar, 50 mm). (C) In the fetal mouse lung on E18, Sema3A immunoreactivity (yellow) is seen in the alveolar (a) and bronchiolar (b) epithelium. A slight but signi®cant immunoreactivity is seen in endothelial cells of a small artery (*). Counterstained with phalloidin (red) £300 (scale bar, 50 mm). (D) NP-1 immunoreactivity is detected in the alveolar cells (a) but not in the bronchiolar cells (b) of E18 mouse lung. Immunoreactivity in alveolar interstitial tissues is seen, but it is dif®cult to de®ne cell type which shows positive staining. Counterstained with phalloidin (red) £320 (scale bar, 50 mm). (E,F) NP-1 (E) and Clara cell protein (F) immunostainings of serial sections of adult lung. Positive NP-1 immunostaining is predominantly seen in the alveolar region. Terminal bronchiole (tb) counterstained with phalloidin (red) £480 (scale bar, 20 mm). (G,H) CRMP-2 (G) and Clara cell protein (H) immunostainings of serial sections of E18 mouse lung. CRMP-2 immunostaining (green; G) is predominantly seen in the alveolar region, but bronchiolar epithelium including Clara cell protein-positive cells (green to yellow; H) is only weakly stained for CRMP-2 (G). Counterstained with phalloidin (red) in (H) £300 (scale bar, 50 mm).

and increase with age in NP-1. The expression of lung Sema3A was highest at E12 and decreased by 95% in the adult. On the other hand, the expression of NP-1 appeared to be upregulated at E14, and thereafter became relatively constant, with the highest expression in the adult. Intensities of b-actin product were similar among the tissue samples tested. In Western blot analysis of CRMP-2, expression of CRMP-2 in the lungs increased with development, which roughly paralleled the level of the message of NP-1 (Figs. 2 and 3). We con®rmed that there was no staining in the presence of the CRMP peptide antigen (Goshima et al., 1995).

2.3. Dose-dependent reduction by Sema3A in the number of terminal buds of fetal mouse lung explant Using lung explant culture of the developing lung, we investigated the effects of Sema3A on branching morphogenesis. Before cultivation, the lungs of E11 mice have a few terminal buds originating from the lobar bronchi. In serum-free medium, the lung explants grew and terminal buds increased with culture (Fig. 4A). The number of terminal buds in the cultured lung explants was fewer with Sema3A treatment than that without Sema3A (Figs. 4 and 5). We also con®rmed the effect by use of puri®ed recom-

38

T. Ito et al. / Mechanisms of Development 97 (2000) 35±45

Fig. 2. Sema3A and NP-1 mRNA expression in the fetal (E12, 14, 18), neonatal, and adult lungs by RT-PCR method. PCR ampli®cation was performed under conditions of linearity for each individual primer set (20 cycles of ampli®cation for Sema3A and NP-1 and 15 cycles for b-actin). The RT-PCR products (RT1) are shown with negative controls (RT2), which are gained after the similar PCR processes without RT pretreatment.

binant Sema3A (data not shown). Bromodeoxyuridine (BrdU) immunohistochemistry, however, revealed that the BrdU-positive DNA-synthesizing fraction was similar among the explants regardless of Sema3A treatment (Fig. 5A,B). In order to determine a functional role for NP-1 in mediating Sema3A signaling, we have employed sNP-1, a soluble NP-1 lacking the transmembrane and intracellular region (Goshima et al., 1999). When combined with Sema3A, sNP-1 signi®cantly attenuated the reduction in the number of terminal buds induced by Sema3A (Fig. 6). Conditioned medium from mock-transfected cells alone produced no effect on lung branching morphogenesis, and did not affect Sema3A-induced change in the number of terminal bronchial buds (Fig. 6). These ®ndings suggest that sNP-1 competes for the same binding site on Sema3A as does the receptor mediating inhibition of branching morphogenesis. The simplest explanation is that the receptor for Sema3A is NP-1. Exogenous sNP-1 alone, however, produced no effect on the number of terminal branches (Fig. 6), which was an unexpected result. Since exogenous Sema3A inhibits branching morphogenesis of embryonic lung, exogenous sNP-1 alone should promote branching. Although the exact reasons were unknown, it might be possible that the concentrations of sNP-1 that are required to block the effect of endogenous Sema3A were higher than those required to perturb the effect of exogenous Sema3A. Indeed, sNP-1 blocks the effect of exogenous Sema3A to induce growth cone collapse of dorsal root ganglia (DRG) neurons in a competitive fashion (Goshima et al., 1999). 2.4. Morphology of the lungs from np1-mutant mice and failure of Sema3A to reduce the number of terminal bronchial buds in explant culture of fetal lung from np1 homozygous null mice To further con®rm whether NP-1 mediates the effect of Sema3A, we employed explant culture of lung from neuropilin-1 (np1) homozygous null animals. In the np1 homozygous null animals (np1 (2/2)), the lung was smaller and the number of branches in the left lungs was signi®cantly lower than that in the wild-type (np1 (1/1)) and heterozy-

gous np1 mutant (np1 (1/2)) animals (Fig. 7): mean ^ SD for the number of branches was 2.43 ^ 0.77 in np1 (1/1), 2.41 ^ 0.73 in np1 (1/2), and 1.86 ^ 0.66 in np1 (2/2) fetal lungs, respectively (P , 0:01, and P , 0:05, when compared to np1 (1/1) and np1 (1/2); n ˆ 14±35). After cultivation with Sema3A (10 units/ml), the increase in the number of terminal buds was not prominent in the lung explants from np1 (1/1) and (1/2) mice relative to that with vehicle control. In spite of the smaller number of branches before cultivation in the np1 homozygous null animals, the lung explants from the np1 (2/2) animals grew normally in vitro: the relative number of terminal buds of np1 (2/2) fetal lung after cultivation was similar to that seen in np1 (1/1) or (1/2) lung (Fig. 8, open column). Addition of Sema3A to the culture media did not reduce the number of terminal buds of the lung explants from np1 (2/2) animals (Fig. 8). 3. Discussion We present the ®rst evidence of a non-neuronal effect of Sema3A in developing fetal lungs. Sema3A negatively regulated lung morphogenesis with doses which were comparable to those to induce growth cone collapse in chick sensory neurons (Goshima et al., 1995). This inhibitory effect was mediated by NP-1, because blocking NP-1 binding sites on Sema3A with sNP-1 prevented Sema3Ainduced inhibition of branching morphogenesis. Furthermore, the present study clearly demonstrated that Sema3A failed to affect branching morphogenesis in culture of lung explant from np1 homozygous null mice. These results indicate that NP-1 is an essential cell surface component that mediates Sema3A-induced inhibition of branching morphogenesis of the fetal lung. Immunohistochemical and immunoblot analysis using anti-Sema3A, -NP-1 and -CRMP-2 antibodies and a semiquantitative RT-PCR method for Sema3A and NP-1 mRNAs revealed that the mouse lungs had at least three components which are considered to be essential for the Sema3A-signaling (Goshima et al., 1995; He and TessierLavigne, 1997; Kolodkin et al., 1997; Kitsukawa et al., 1997). Our ®ndings are consistent with the data on Northern blot analyses for Sema3A (Luo et al., 1993), NP-1 (Takagi et al., 1991; Kawakami et al., 1996) and CRMP-2 (Wang and Strittmatter, 1996) in lung tissue. Sema3A alkaline

Fig. 3. Expression of CRMP-2 in the fetal (E12, 14, 18), neonatal, and adult lungs by Western blot analysis. The crude homogenates (about 10 mg of protein/lane) were separated on SDS-polyacrylamide gels and immunoblotted.

T. Ito et al. / Mechanisms of Development 97 (2000) 35±45

39

Fig. 4. Effect of Sema3A treatment on branching morphogenesis of the fetal mouse lungs (E11.5). (A) Fetal lung explants (E11.5) before culture (a and c) and after 2 days cultivation (b,d) with vehicle control (a,b) and with 10 units/ml Sema3A (c,d). Note fewer terminal branches(d) in comparison with (b) £60 (scale bar, 100 mm). (B) The relative number of terminal buds was calculated as the number of buds at the end of the culture/the number of buds at the beginning of the culture. The graph shows that the number of terminal buds after 2 day cultivation is reduced by treatment of Sema3A with dose-dependent manner. Data are mean ^ SD from 15 experiments. aP , 0:05, bP , 0:005, cP , 0:001, when compared to control.

phosphatase fusion protein (Sema3A-AP) has been used as a probe to label NP-1 (He and Tessier-Lavigne, 1997; Kolodkin et al., 1997; Takahashi et al., 1997). High levels of staining with Sema3A-AP are present throughout the developing lung (Takahashi et al., 1997). In the present immunohistochemical study, Sema3A immunoreaction was detected in the terminal airway epithelial cells, and in the bronchiolar and alveolar epithelia of the adult lung. NP-1 levels rose dramatically during development, while Sema3A levels were high in the early stages and then fall to nearly undetectable levels in later stages. Sema3A immunoreactivities were seen in the bronchiolar and alveolar epithelia, while NP-1 immunoreactivities were mainly seen in alveolar epithelium, but not in the bronchiolar cells. These distribution patterns of Sema3A and its receptor NP-1 suggest that airway epithelial cells secreting Sema3A make a concentra-

tion gradient of Sema3A to the peripheral region, and could function in lung morphogenesis and maintenance of alveolar tissue. CRMP-2 immunoreactivity was also detected in developing and adult alveolar epithelium in parallel with NP-1 immunoreactivity, thereby further suggesting that alveolar epithelial cells are responsive to Sema3A. As discussed below, an alternative view is that NP-1 has two roles during lung development: an early role in branching morphogenesis as a receptor for Sema3A, and a later role in vascularization as a receptor for VEGF (Soker et al., 1998). Branching morphogenesis could be regulated by many factors, which can induce cell proliferation activity or cell motility. The observed effects of Sema3A in fetal lung development appeared not to be due to a direct action on epithelial growth, because Sema3A did not affect the BrdUpositive DNA-synthesizing fraction. In contrast, several

Fig. 5. Bromodeoxyuridine (BrdU) incorporation in cultured lung explant. (A) Immuno¯uorescence of bromodeoxyuridine of the fetal mouse lung after 2 days cultivation with (b) or without (a) Sema3A 10 units/ml. No remarkable differential staining was detected, £60 (scale bar, 100 mm). (B) Values of BrdU labeling indices were not different among experimental groups. Data shown are mean ^ SD from ten experiments.

40

T. Ito et al. / Mechanisms of Development 97 (2000) 35±45

Fig. 6. Effect of soluble NP-1 (sNP-1; at 1:20 dilution) on Sema3A-induced effect of the fetal mouse lung explants (E11.5). Note that sNP-1 treatment did not alter the relative number of terminal buds of cultured lung with or without Sema3A treatment (10 units/ml). Data are mean ^ SD from ten experiments. aP , 0:01, when compared to control. bP , 0:05, when compared to sNP-1 or Sema3A 1 sNP-1.

other growth factors affect proliferation and/or induce apoptosis (Serra et al., 1994; Shiratori et al., 1996). Transforming growth factor b (TGFb) has been demonstrated to inhibit expression of N-myc and negatively regulate lung development (Serra et al., 1994). The inhibition of branching morphogenesis by TGFb is accompanied by inhibition of DNA synthesis in epithelium and accumulation of extracellular matrix, both of which are considered to cause inhibition of lung bud formation. On the other hand, keratinocyte growth factor (KGF) and acidic ®broblast growth factor stimulate proliferation and differentiation of alveolar epithelial type II cells via KGF receptor, but inhibit branching morphogenesis of fetal lung (Shiratori et al., 1996). These ®ndings imply that differentiation and growth activity of lung cells affect branching morphogenesis. In addition, our present study suggests that factors that inhibit or facilitate chemotaxis or cell migration could exert negative or positive regulatory function in branching morphogenesis of developing lung: Sema3A potently inhibits growth cone

motility by causing actin-depolymerization (Fan et al., 1993), and it is possible that Sema3A modulated migration or motility of epithelial cells, and inhibited branching morphogenesis in the fetal lung. This idea is compatible with a recent ®nding that Sema3A affects on cell migration of endothelial and neural crest cells (Miao et al., 1999; Eickholt et al., 1999). Interestingly, hepatocyte growth factor (HGF), a scattering factor that enhances cell motility (Ohmichi et al., 1998) promotes lung branching morphogenesis. NP-1 and Sema3A immunohistochemical reactions were detected both in the alveolar epithelial and endothelial cells. This observation suggests that the Sema3A-NP-1 system perhaps play a roles in branching morphogenesis in early fetal days, and NP-1 may be additionally involved in other biological activities such as alveolar vascularization at the later stage. In fact, NP-1 also acts as an isoform-speci®c receptor for vascular endothelial growth factor165 (VEGF165), a major regulator of angiogenesis (Soker et al., 1998; Kawasaki et al., 1999). VEGF165 binds to both NP-1 and kinase domain containing receptor (KDR) with a similar high af®nity of a 100-pM range. KDR appears to be the major transducer of VEGF signals in endothelial cells that result in chemotaxis, mitogenicity, actin reorganization, and gross morphological changes in target cells. When coexpressed in cell with KDR, NP-1 enhances the binding of VEGF165 to KDR and VEGF165-mediated chemotaxis (Soker et al., 1998). Although there are no reports concerning the effect of VEGF on branching morphogenesis in fetal lung explant, lung acinar tubules are decreased and type I alveolar cell differentiation is inhibited when VEGF is overexpressed in developing alveolar epithelial cells under the control of the promoter from the human surfactant protein C gene (Zeng et al., 1998). Very recently Sema3A has been shown to inhibit the motility of vascular endothelial cells, probably by a competitive inhibition of Sema3A against VEGF165-induced cell motility (Miao et al., 1999). The molecular mechanisms and signaling pathway by which Sema3A inhibits branching morphogenesis are unknown. We have analyzed a Xenopus laevis oocyte expression system in an attempt to characterize components of the collapsin/Sema3A response pathway, and isolated CRMP-62 (chick CRMP-2) from chick DRG (Goshima et

Fig. 7. Morphological observation of the lungs from neuropilin-1 (np1) mutant mice. Size and number of branches were under-developed in np1 (2/2) mouse lungs than np1 (1/2) mouse lungs. (A,B) Phase contrast view of whole lungs of fetal mice (E11.5) with np1 (1/2) (A) and np1 (2/2) (B) £60 (scale bar, 100 mm). (C,D) Transverse section of the chest of the fetal mice (E12) with np1 (1/2) (C) and np1 (2/2) (D) £50 (scale bar, 100 mm).

T. Ito et al. / Mechanisms of Development 97 (2000) 35±45

al., 1995). Although its mechanisms of action remain illde®ned, CRMP-2 is required for Sema3A-induced growth cone collapse in the DRG (Goshima et al., 1995). Our present study provides the ®rst immuno-histochemical evidence that the lung CRMP-2 is localized in alveolar epithelial and neuroendocrine as well as neuronal cells. Although it remains to be determined that CRMP-2 also mediates Sema3A-induced inhibition of branching morphogenesis of the lung, there is strong evidence in the form of the expression of CRMP-2 and NP-1 in the developing and adult alveolar epithelia and the expression pro®le of these molecules during lung development to support the idea that the NP-1-CRMP system could also play a role in the lung alveolarization and in maintenance of the alveolus. Similar parallel and dynamic changes of NP-1 and CRMP-2 expression levels are seen with the regenerating primary olfactory nerve after axotomy (Pasterkamp et al., 1998). Despite the expression of CRMP in neuroendocrine cells, however, NP1 was not detected in this cell population. This may imply that CRMP-2 plays a role other than that of a protein mediating response initiated by binding of Sema3A to NP-1 (Goshima et al., 1995). Studies of the sema3A- and np1-targeted animals have revealed that these genes are critical for peripheral nerve formation (Taniguchi et al., 1997; Kitsukawa et al., 1997). We observed, however, no remarkable morphologic abnormalities in the lung of sema3A- de®cient mice (data not shown). Although our present ®nding on the Sema3A action may present a novel mode of action on lung branching morphogenesis, the phenotype in sema3A- de®cient mice argues against the idea that only Sema3A exerts a unique function in the fetal lung. The lung expression of other semaphorins such as Sema3C and Sema3F, together

41

with their receptor components NP-1 and/or NP-2 has been reported (Puschel et al., 1995; Eckhardt and Meyerhans, 1998; Chen et al., 1997; Giger et al., 1998; Encinas et al., 1999). Moreover, we detected mRNA expression of these molecules in the developing lung of normal mouse embryo (Kagoshima et al., unpublished observation). Because of the highly related molecular structure of class 3 semaphorins, these molecules may compensate for the function of Sema3A in sema3A-de®cient mice. On the other hand, in the np1 homozygous null mice, retarded lung morphogenesis was detected. It is possible that other NP-1 ligands are more important in lung morphogenesis, because np1 (2/2) lungs are mildly abnormal, while sema3A (2/2) lungs are normal. In fact, we have recently found that other type 3 semaphorins such as Sema3C, which binds to either NP-1 or NP-2, promote branching, while Sema3A inhibits it (data submitted). Regulation of branching morphogenesis by semaphorins in the developing lung may thus depend on the balance between these counteracting actions of multiple semaphorins. Alternatively, since the lung explants from the np1 homozygous null mice appeared to grow normally for 2 days in vitro, this abnormal development of the lung is perhaps not manifested by the lack of Sema3A-NP-1 signaling, but is rather secondary to systemic circulatory disturbance associated with cardiovascular abnormalities, which are probably caused by the absence of VEGF-NP-1 signaling (Kitsukawa et al., 1997; Soker et al., 1998). More recently, NP-1 and plexin 1 have been demonstrated to form a stable receptor complex for Sema3A (Takahashi et al., 1999; Tamagnone et al., 1999). It will thus be interesting to know whether plexins are also involved in the effects of semaphorins on lung branching morphogenesis. In conclusion, Sema3A negatively controls fetal lung branching via NP-1. This provides a novel regulatory system for branching morphogenesis in the developing lung. The expression of Sema3A, NP-1 and CRMP-2, and the inhibitory effect of Sema3A on branching morphogenesis suggests that developing lung tissue or the epithelial cells may provide a useful model to investigate mechanisms by which Sema3A and other semaphorins exert their biological actions during development of neuronal and nonneuronal tissues.

4. Experimental procedures 4.1. Animals Fig. 8. Effect of Sema3A (10 units/ml) on lung branching morphogenesis of fetal lung explants of mice with different np1 status (E11.5). Note that the number of branches is reduced signi®cantly in the lung explants from np1 (1/1) and np1 (1/2) mice by treatment of Sema3A, but no apparent effect of Sema3A on the lung explants from np1 (2/2) mice. Data are mean ^ SD from at least ®ve experiments. aP , 0:05, bP , 0:02 when compared to corresponding control.

ICR mice were purchased from Japan SLC (Shizuoka, Japan). To obtain fetal lungs, mice were mated during the night, and the day of the discovery of the vaginal plug was counted as fetal day (E) 0. For studies of expression of Sema3A, NP-1 and CRMP-2, fetal (E12, E14, and E18), neonatal, and adult (6 week old, male) lungs of ICR mice were used. For organ culture, lungs from E11.5 mice were

42

T. Ito et al. / Mechanisms of Development 97 (2000) 35±45

used. Mice de®cient for np1 were generated using homologous recombination techniques as described previously (Kitsukawa et al., 1997). 4.2. Immunohistochemistry of embryonic and adult mouse lungs Immunohistochemical detection in embryonic lung in vivo was performed as described (Ito et al., 1997). Brie¯y, at E12, E14, and E18, six pairs of fetal lungs at each gestational day, lungs from 6 neonatal mice, and lungs from 6 adult male mice were ®xed in a phosphate-buffered 4% paraformaldehyde solution (pH 7.3) for 3 h at room temperature. The ®xed tissues were washed in a phosphate buffered solution, embedded in OCT compound, and frozen with liquid nitrogen. Frozen sections were incubated overnight at 48C with a goat polyclonal antibody against Sema3A (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and rabbit polyclonal antibodies against NP-1 (Kawakami et al., 1996) and against chicken CRMP-2 (Goshima et al., 1995). After washing in the phosphate buffer, sections were incubated with biotinylated anti-goat (for Sema3A staining) and anti-rabbit (for NP-1 and CRMP-2 stainings) IgG (Chemicon, Temecula, CA, USA), and treated with ¯uorescein isothiocyanate-conjugated avidin (Sigma). Background staining with tetramethylrhodamine isothiocyanate-conjugated phalloidin (Sigma, St. Louis, MO) was also added. They were then mounted with a p-phenylenediamine and glycerol mixture and observed in an epi¯uorescent microscope and a confocal laser microscope equipped with argon ion and helium neon lasers (LSM GB200UV; Olympus, Tokyo, Japan). To clarify localization of the immunoreaction for Sema3A and the related antigens, serial frozen sections were made, and were stained immunohistochemically for these three antigens in combination with immunohistochemistry for Clara cell, non-ciliated bronchiolar cell, with a rabbit polyclonal antibody against Clara cell 10 kDa protein (generous gift from Dr G. Singh), and for CGRP with a rabbit polyclonal antibody (Amersham International, Little Chalfont, UK) to detect neuroendocrine cells. Negative controls were performed by applying nonimmune sera to all pertinent subclasses of immunoglobulin. In addition, for the Sema3A and CRMP-2 immunohistochemistry, the antibody preincubation with the corresponding peptides was substituted for the active antibodies against Sema3A and CRMP-2. 4.3. Semiquantitative reverse-transcriptase-polymerase chain reaction (RT-PCR) Total RNA was extracted from fetal, neonatal and adult lungs by the guanidinium thiocyanate method. Extracted total RNA was reverse transcribed using SuperScript TM Preampli®cation System (GIBCO-BRL, Rockville, MD, USA) with oligo(dT)12±18 primer. PCR ampli®cation was performed with primer sets of Sema3A, NP-1 and b-actin. The forward and reverse PCR primers for mouse Sema3A

and NP-1 were: 5 0 -GAG GCG CAC AAG ACG ACA AG3 0 (nucleotide #1760±1779 of mouse sema3A cDNA [Acc#: X85993]), and 5 0 -TGG CGC AAG CTT GAC ACT TCT GGG TGC CCG-3 0 (complimentary to #2409±2426 of the cDNA and extra-Hind III site); 5 0 -TGG AAT TCA CAT TGG GCG TTA TTG TGG GC-3 0 (nucleotide #1013± 1035 of mouse NP-1 cDNA [Acc#: D50086] and extraEco RI site), and 5 0 -TGC TCG AGC CTC TTA CTA TCT TCT CAT CTC CC-3 0 (complimentary to #1814± 1837 of the cDNA and extra-Xho I site), respectively. The resulting product sizes were 679 bp for Sema3A and 839 bp for mouse NP-1. The forward and reverse PCR primers for mouse b-actin were 5 0 -CGT GGG CCG CCC TAG GCA CCA-3 0 (nucleotide #182±202 of mouse cytoskeletal bactin cDNA [Acc#:X03672]), and 5 0 -TTG GCC TTA GGG TTC AGG GGG G-3 0 (complimentary to #403±424 of the cDNA), respectively. The resulting product size was 243 bp. b-Actin is a housekeeping gene and was studied as a control. Brie¯y, the RT product (2 ml) was added to 48 ml of the reaction mixture containing 1.5 mM MgCl2, 200 nM dNTP, 2.5 units of Taq DNA polymerase (GIBCO-BRL), 1 mCi [g- 32P] dCTP and forward/reverse primers (200 nM, for NP-1; 1 mM, for Sema3A; 2 pM, b-actin). After 3-min denaturation at 948C, ampli®cation was done in a DNA Thermal Cycler (Perkin-Elmer Cetus, Norwalk, CT) under the following conditions: denaturation at 948C for 30 s, annealing at 558C (for Sema3A and b-actin) or 658C (for NP-1) for 30 s, and extension at 728C for 90 s. We determined the optimum number of cycles for each primer set so that the speci®c product was ampli®ed during the exponential phase of ampli®cation. Based on the results of preliminary studies, 20 cycles were used for NP-1 and Sema3A, and b-actin was ampli®ed for 15 cycles. Negative control reactions performed without RT yielded no detectable fragments with any primer pair. The PCR product was separated by 4% acrylamide-7 M urea gel electrophoresis and quantitated by analysis on Bio-image Analyzer System BAS 2000 (Fuji Photo Film, Tokyo, Japan). PCR ampli®cation was performed under conditions of linearity for each individual primer set. 4.4. Immunoblot analysis The crude samples from fetal, neonatal and adult lung tissues were homogenized, and were separated by 10% SDS-PAGE (10 mg of protein/lane) and electrophoretically transferred to PVDF ®lter (Millipore, Bedford, MA, USA). The ®lter was preincubated in blocking buffer [10% non-fat dry skim milk in Tris-buffered saline (TBS), pH 7.6 and 0.3% (w/v) Tween-20] at 48C overnight, followed by incubation with anti-CRMP-2 antibody (Goshima et al. 1995) diluted 1:6000 in blocking buffer for 1 h at room temperature. After washing with TBS containing 0.3% Tween-20, the ®lter was incubated with horseradish peroxidase (HRP)conjugated anti- rabbit IgG (Kirkegaad and Perry Laboratories, Gathersburg, MD, USA) in blocking buffer for 1 h.

T. Ito et al. / Mechanisms of Development 97 (2000) 35±45

The blot was developed by enhanced chemiluminescence ECL (Amersham) and exposed to Hyper®lm ECL (Amersham). 4.5. Recombinant Sema3A and sNP-1 Recombinant Sema3A and sNP-1 were prepared as described previously (Goshima et al., 1999). Twenty micrograms of the expression vector encoding Sema3A, sNP-1, or vehicle control vector was transfected to human embryonic kidney 293T cells grown on 150 mm dishes by calcium phosphate co-precipitation. The transfection medium was replaced by serum free medium after a 4-h incubation. Conditioned medium (20 ml/plate) was harvested after 5 days of culture, and was concentrated 100-fold by membrane ®ltration using stirred cell (type 8050, Amicon). To obtain puri®ed Sema3A, expression vector DNA for Sema3A(His)6 was transfected into HEK293T cells, and the conditioned medium was applied to a 0.75-ml Nicontaining resin (Talon; CLONTECH). The column was washed with 5 ml of 20 mM Tris±HCl, pH 8.0, and 0.5 M NaCl. Sema3A activity eluted primarily in the 50 mM imidazole fraction. One unit of Sema3A was de®ned as the amount required in a 1-ml culture to collapse 50 % of growth cones of chick dorsal root ganglion neurons (Goshima et al., 1995). 4.6. Lung organ culture and immunohistochemistry At E11.5, eight pregnant mice were anesthetized, their gravid uteruses were excised, and all fetuses were removed. The lungs and tracheas were taken from the fetuses aseptically under a stereomicroscope for culture. Totally, 60 pairs of lung with trachea (15 pairs in each experimental group) were placed on Transwell collagen membrane (0.4 mm pore size and 24.5 mm diameter; Corning-Costar, Cambridge, MA, USA) and cultured in chemically de®ned medium (Fitton-Jackson modi®cation BGJb, GIBCO-BRL) containing 0.2 mg/ml ascorbic acid and 50 units/mg penicillin/ streptomycin with or without recombinant Sema3A (0, 3, 10 and 30 units/ml) for 2 days in culture. The concentrations of Sema3A used were slightly higher than but comparable to those used in DRG growth cone collapse assay (Goshima et al., 1995). Before harvesting the explants, bromodeoxyuridine (BrdU; Sigma) was added in the culture medium at the concentration of 50 mg/ml. Phase-contrast micrographs were taken at the start and at the end of culture. The number of terminal buds in the left lungs was counted before and at the end of cultivation, and the number of terminal buds (the number of buds at the end of the culture/the number of buds at the beginning of the culture) was calculated. After culturing, the tissues were ®xed with buffered paraformaldehyde for 1 h at room temperature. For BrdU immunohistochemistry for detecting DNA-synthesizing cells in the explants, the ®xed explants were embedded in OCT compound and frozen. Frozen sections, pretreated with 0.5 mg/ml pepsin in 4N HCl at room temperature for 10 min, were incubated

43

with a mouse monoclonal antibody against BrdU (Sigma) at 48C overnight, and labeled with FITC-conjugated antimouse IgG and with nuclear staining by propidium iodide (Sigma). BrdU-labeling index in the epithelium at the terminal bud was calculated as the percentage of positively stained nuclei after more than 500 nuclei were counted in the region. To investigate the possible involvement of NP-1 in the action of Sema3A on lung branching morphogenesis, we employed sNP-1 lacking the transmembrane and intracellular region as described previously (Goshima et al., 1999). Forty lung pairs from fetal day 11.5 (ten each in experimental groups) were cultured with and without Sema3A (at the concentration of 10 units/ml), with sNP-1 (at 1:20 dilution) and with both of them. After cultivation for 2 days, the relative number of terminal buds was calculated. 4.7. Lungs from np1 mutant mice As np1 homozygous null fetuses cannot survive as late as E12.5 (Kitsukawa et al., 1997), we obtained fetal lungs of the mice with np1 (1/1), (1/2) and (2/2) at E11.5 and the lung explants were cultured for 2 days with or without Sema3A (10 units/ml). The number of branches in the left lungs from the mice with various genotypes was evaluated before and after culture. Genotype analysis for np1 was performed by the PCR method (Kitsukawa et al., 1997). The numbers of the pairs of lungs examined was 19 (11 without Sema3A; eight with Sema3A) for np1 (1/1), 35 (20 without Sema3A; 15 with Sema3A) for np1 (1/2), and 14 (nine without Sema3A; ®ve with Sema3A) for np1 (2/2). For morphological analysis, fetuses of the mice with np1 (1/1), (1/2) and (2/2) at E12 were obtained, ®xed in paraformaldehyde, and embedded in paraf®n. Transverse section of the fetuses was stained with hematoxylin and eosin. 4.8. Statistics The values gained in these experiments were subject to un-paired t-test for analyzing signi®cance. A P value of below 0.05 was considered statistically signi®cant. Acknowledgements We wish to thank Dr Jonathan Raper for providing chick collapsin-1 cDNA clone. We thank Drs S.M. Strittmatter, F. Nakamura and T. Takahashi for helpful discussions and advice. We also thank Dr S.M. Strittmatter for anti-chick CRMP-2 antibody and critical reading of our manuscript. This work was supported in part by Grants-in-Aid from Ministry of Education, Science, Sports and Culture, Japan (T.I.) and in part by CREST of Japan Science and Technology Corporation (Y.G.), and Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists (M.K).

44

T. Ito et al. / Mechanisms of Development 97 (2000) 35±45

References Behar, O., Golden, J.A., Mashimo, H., Schoen, F.J., Fishman, M.C., 1996. Semaphorin III is needed for normal patterning and growth of nerves, bones and heart. Nature 383, 525±528. Byk, T., Dobransky, T., Cifuentes-Diaz, C., Sobel, A., 1996. Identi®cation and molecular characterization of unc-33-like phosphoprotein (Ulip), a putative mammalian homologue of the axonal guidance-associated unc33 gene product. J. Neurosci. 15, 688±701. Chen, H., Chedotal, A., He, Z., Goodman, C.S., Tessier-Lavigne, M., 1997. Neuropilin-2, a novel member of the neuropilin family, is a high af®nity receptor for the Semaphorins Sema E and Sema IV but not Sema III. Neuron 19, 547±559. Eckhardt, F., Meyerhans, A., 1998. Cloning and expression pattern of a murine Semaphorin homologous to H-Sema IV. NeuroReport 9, 3975± 3979. Eickholt, B.J., Mackenzie, S.L., Graham, A., Walsh, F.S., Doherty, P., 1999. Evidence for collapsin-1 functioning in the control of neural crest migration in both trunk and hindbrain regions. Development 126, 2181±2189. Encinas, J.A., Kikuchi, K., Chedotal, A., de Castro, F., Goodman, C.S., Kimura, T., 1999. Cloning, expression, and genetic mapping of Sema W, a member of the semaphorin family. Proc. Natl. Acad. Sci. USA 96, 2491±2496. Fan, J., Mans®eld, S.G., Redmond, T., Gordon-Weeks, P.R., Raper, J.A., 1993. The organization of F-actin and microtubules in growth cones exposed to a brain derived collapsing factor. J. Cell Biol. 121, 867±878. Giger, R.J., Urquhart, E.R., Gillespie, S.K.H., Levengood, D.V., Ginty, D.D., Kolodkin, A.L., 1998. Neuropilin-2 is a receptor for Semaphorin IV: insight into the structural basis of receptor function and speci®city. Neuron 21, 1079±1092. Goshima, Y., Hori, H., Sasaki, Y., Yang, T., Kagoshima-Maezono, M., Takenaka, T., Nakamura, F., Takahashi, T., Strittmatter, S.M., Misu, Y., Kawakami, T., 1999. Growth cone neuropilin-1 mediates collapsin1 facilitation of antero-and retrograde axoplasmic transport. J. Neurobiol. 39, 579±589. Goshima, Y., Kawakami, T., Hori, H., Sugiyama, Y., Takasawa, S., Hashimoto, Y., Kagoshima-Maezono, M., Takenaka, T., Misu, Y., Strittmatter, S.M., 1997. A novel action of collapsin: collapsin-1 increases antero- and retrograde axoplasmic transport independently of growth cone collapse. J. Neurobiol. 33, 316±328. Goshima, Y., Nakamura, F., Strittmatter, P., Strittmatter, S.M., 1995. Collapsin-induced growth cone collapse mediated by an intracellular protein related to UNC-33. Nature 376, 509±514. Goshima, Y., Sasaki, Y., Nakayama, T., Ito, T., Kimura, T., 2000. Functions of semaphorins in axon guidance and neuronal regeneration. Jpn. J. Pharmacol. 82, 273±279. Hamajima, N., Matsuda, K., Sakata, S., Tamaki, N., Sasaki, M., Nonaka, M., 1996. A novel gene family de®ned by human dihydropyrimidinase and three related proteins with differential tissue distribution. Gene 180, 157±163. He, Z., Tessier-Lavigne, M., 1997. Neuropilin is a receptor for the axonal chemorepellent Semaphorin III. Cell 90, 739±751. Hogan, B.L.M., Yingling, J.M., 1998. Epithelial/mesenchymal interactions and branching morphogenesis of the lung. Curr. Opin. Gene Dev. 8, 481±486. Ito, T., 1999. Differentiation and proliferation of pulmonary neuroendocrine cell. Prog. Histochem. Cytochem. 34, 247±322. Ito, T., Nogawa, H., Udaka, N., Kitamura, H., Kanisawa, M., 1997. Development of pulmonary neuroendocrine cells of fetal hamster in explant culture. Lab. Invest. 77, 449±457. Kagoshima-Maezono, M., Ito, T., Udaka, N., Sasaki, Y., Taniguchi, M., Kitamura, H., Fujisawa, H., Goshima, Y., Misu, Y., 1998. Inhibitory effect of SemaD /collapsin-1 on branching morphogenesis of mouse fetal lung. Jpn. J. Pharmacol. 76, 267. Kamata, T., Daar, I.O., Subleski, M., Copeland, T., Kung, H-F., Xu, R.-H.,

1998. Xenopus CRMP-2 is an early response gene to neural induction. Mol. Brain Res. 57, 201±210. Kawakami, A., Kitsukawa, T., Takagi, S., Fujisawa, H., 1996. Developmentally regulated expression of a cell surface protein, neuropilin, in the mouse nervous system. J. Neurobiol. 29, 1±17. Kawasaki, T., Kitsukawa, T., Bekku, Y., Matsuda, Y., Sanbo, M., Yagi, T., Fujisawa, H., 1999. A requirement for neuropilin-1 in embryonic vessel formation. Development 126, 4895±4902. Kitsukawa, T., Shimizu, M., Sanbo, M., Hirata, T., Taniguchi, M., Bekku, Y., Yagi, T., Fujisawa, H., 1997. Neuropilin-semaphorin III/Dmediated chemorepulsive signals play a crucial role in peripheral nerve projection in mice. Neuron 19, 995±1005. Kolodkin, A.L., Matthes, D.J., O'Connor, T.P., Patel, N.H., Admon, A., Bently, D., Goodman, C.S., 1993. The semaphorin genes encode a family of transmembrane and secreted growth cone guidance molecules. Cell 75, 1389±1399. Kolodkin, A.L., Levengood, D.V., Rowe, E.G., Tai, Y-T., Giger, R.J., Ginty, D.D., 1997. Neuropilin is a Semaphorin III receptor. Cell 90, 753±762. Li, W., Herman, R.K., Shaw, J.E., 1992. Analysis of the Caenorhabditis elegans axonal guidance and outgrowth gene unc-33. Genetics 132, 675±689. Luo, Y., Raible, D., Raper, J.A., 1993. Collapsin: a protein in brain that induces the collapse and paralysis of neuronal growth cones. Cell 75, 217±227. Miao, H-Q., Soker, S., Feiner, L., Alonso, J.L., Raper, J.A., Klagsbrun, M., 1999. Neuropilin-1 mediates collapsin-1/Semaphorin III inhibition of endothelial cell motility: functional competition of collapsin-1 and vascular endothelial growth factor-165. J. Cell Biol. 146, 233±241. Minturn, J.E., Fryer, H.J.L., Geschwind, D.H., Hock®eld, S., 1995. TOAD 64, a gene expressed early in neuronal differentiation in the rat, is related to unc-33, a C. elegans gene involved in axon outgrowth. J. Neurosci. 15, 6757±6766. Nogawa, H., Ito, T., 1995. Branching morphogenesis of fetal mouse lung epithelium in mesenchyme-free culture. Development 121, 1015±1022. Ohmichi, H., Koshimizu, U., Matsumoto, K., Nakamura, T., 1998. Hepatocyte growth factor (HGF) acts as a mesenchyme-derived morphogenic factor during fetal lung development. Development 125, 1315± 1324. Pasterkamp, R.J., De Winter, F., Holtmaat, A.J., Verhaagen, J., 1998. Evidence for a role of the chemorepellent semaphorin III and its receptor neuropilin-1 in the regeneration of primary olfactory axons. J. Neurosci. 18, 9962±9976. Puschel, A.W., Adams, R.H., Betz, H., 1995. Murine Semaphorin D/collapsin is a member of a diverse gene family and creates domains inhibitory for axonal extension. Neuron 14, 941±948. Roche, J., Boldog, F., Robinson, M., Varella-Garcia, M., Swanson, M., Waggoner, B., Fishel, R., Franklin, W., Gemmill, R., Drabkin, H., 1996. Distinct 3p21.3 deletions in lung cancer and identi®cation of a new human semaphorin. Oncogene 12, 1289±1297. Sekido, Y., Bader, S., Latif, F., Chen, J-Y., Duh, F-M., Wei, M-H., Albanesi, J.P., Lee, C.C., Lerman, M.I., Minna, J.D., 1996. Human semaphorins A(V) and IV reside in the 3p21.3 small cell lung cancer deletion region and demonstrate distinct expression patterns. Proc. Natl. Acad. Sci. USA 93, 4120±4125. Serra, R., Pelton, R.W., Moses, H.L., 1994. TGFb1 inhibits branching morphogenesis and N-myc expression in lung bud organ cultures. Development 120, 2153±2161. Shiratori, M., Oshika, E., Ung, L.P., Singh, G., Shinozuka, H., Warburton, D., Michalopoulos, G., Katyal, S.L., 1996. Keratinocyte growth factor and fetal rat lung morphogenesis. Am. J. Respir. Cell Mol. Biol. 15, 328±338. Soker, S., Takashima, S., Miao, H.Q., Neufeld, G., Klagsbrun, M., 1998. Neuropilin-1 is expressed by endothelial and tumor cells as an isoformspeci®c receptor for vascular endothelial growth factor. Cell 92, 735± 745. Takagi, S., Hirata, T., Agata, K., Mochii, M., Eguchi, G., Fujisawa, H.,

T. Ito et al. / Mechanisms of Development 97 (2000) 35±45 1991. The A5 antigen, a candidate for the neuronal recognition molecule, has homologies to complement components and coagulation factors. Neuron 7, 295±307. Takahashi, T., Nakamura, F., Strittmatter, S.M., 1997. Neuronal and non neuronal collapsin-1 binding sites in developing chick are distinct from other semaphorin binding sites. J. Neurosci. 17, 9183±9193. Takahashi, T., Fournier, A., Nakamura, F., Wang, L-H., Murakami, Y., Kalb, R.G., Fujisawa, H., Strittmatter, S.M., 1999. Plexin-Neuropilin1 complexes form functional semaphorin-3A receptors. Cell 99, 59± 69. Tamagnone, L., Artigiani, S., Chen, H., He, Z., Ming, G-I., Song, H-j, Chedotal, A., Winberg, M.L., Goodman, C.S., Poo, M-m, TessierLavigne, M., Comoglio, P.M., 1999. Plexins are a large family of receptors for transmembrane, secreted, and GPI-anchored semaphorins in vertebrates. Cell 99, 71±80. Taniguchi, M., Yuasa, S., Fujisawa, H., Naruse, I., Saga, S., Mishina, M.,

45

Yagi, T., 1997. Disruption of semaphorin III/D gene causes severe abnormality in peripheral nerve projection. Neuron 19, 519±530. Wang, L.H., Strittmatter, S.M., 1996. A family of rat CRMP gene is differentially expressed in the nerve system. J. Neurosci. 16, 6197±6207. Wright, D.E., White, F.A., Gerfen, R.W., Silos-Santiago, I., Snider, W.D., 1995. The guidance molecule semaphorin III is expressed in regions of spinal cord and periphery avoided by growing sensory axons. J. Comp. Neurol. 361, 321±333. Xiang, R.-H., Hensel, C.H., Garcia, D.K., Carlson, H.C., Kok, K., Daly, M.C., Kerbacher, K., Berg, A.V.D., Veldhuis, P., Buys, C.H.C.M., Naylor, S.L., 1996. Isolation of the human semaphorin III/F gene (SEMA3F) at chromosome 3p21.3, a region deleted in lung cancer. Genomics 32, 39±48. Zeng, X., Wert, S.E., Federici, R., Peters, K.G., Whitsett, J.A., 1998. VEGF enhances pulmonary vasculogenesis and disrupts lung morphogenesis in vivo. Dev. Dyn. 211, 215±217.