SEARCH

SEARCH BY CITATION

Keywords:

  • lung development;
  • mesenchyme;
  • axon guidance;
  • airway;
  • Slit;
  • Roundabout;
  • Cadherin-11;
  • in situ hybridization;
  • immunohistochemistry

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Mammalian lung development is mediated through complex interactions between foregut endoderm and surrounding mesenchyme. As airway branching progresses, the mesenchyme undergoes dramatic remodeling and differentiation. Little is understood about the mechanisms that direct mesenchymal organization during lung development. A screen for candidate genes mediating this process identified Slit, a ligand for the Roundabout (Robo) receptor previously associated with guidance of axonal projections during central nervous system development. Here, we demonstrate by in situ hybridization that two Slit genes (Slit-2 and Slit-3) and two Robo genes (Robo-1 and Robo-2) are expressed in fetal lung mesenchyme. Slit-2 and Robo-1 expression is present throughout mesenchyme at midgestation and is not detectable by newborn day 1. Slit-3 and Robo-2 expression is restricted to specific, complementary subsets of mesenchyme. Robo-2 is expressed in mesenchymal cells immediately adjacent to large airways, whereas Slit-3 expression predominates in mesenchyme remote from airway epithelium. The temporal and spatial distribution of Slit and Robo mRNAs indicate that these genes may direct the functional organization and differentiation of fetal lung mesenchyme. Developmental Dynamics 230:350–360, 2004. © 2004 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The mammalian lung is a complex organ incorporating at least 40 different cell types (Shannon and Deterding, 1997; Perl and Whitsett, 1999). Beyond this cellular complexity, the lung is organized along several structural paradigms intrinsic to its functional integrity. These include the growth and differentiation of foregut endoderm into a highly ordered branched airway structure with a large surface area and low surface tension requisite for efficient gas exchange. In tandem with airway development, mesoderm adjacent to the developing airway must contribute a functional low-resistance vascular system capable of receiving the entire cardiac output of the right ventricle with each cardiac cycle. During the early stages of lung development, a highly cellular mesenchyme derived from embryonic mesoderm surrounds each branch of the developing airway. As development proceeds, endothelial precursors within the mesenchyme organize to form an early vascular network. In concert with airway and vascular development, the mesenchyme thins and condenses, ultimately leading to alignment of airways and vessels in a precise geometric orientation to effect efficient exchange of oxygen and carbon dioxide during the respiratory cycle (DeMello and Reid, 2000).

While significant progress has been made toward identification of the factors required for airway and vascular development, little is understood about the factors that direct the functional alignment of these structures. As an initial approach to addressing this question, we developed a simple in vitro coculture system using an epithelial cell line (MLE-12) and an endothelial cell line (MFLM-91U) that recapitulates pulmonary endothelial–epithelial alignment. Microarray analysis demonstrated that, among the genes up-regulated during this endothelial–epithelial association, was a member of the Slit family of secreted ligands. These proteins are known to participate in guidance of axonal projections during central nervous system development and are also expressed in non-neuronal sites, including lung.

Slit homologues are found across a broad spectrum of invertebrate and vertebrate genomes (Chisolm and Tessier-Lavigne, 1999; Brose and Tessier-Lavigne, 2000). Higher vertebrates including mice, rats, and humans express three Slit genes. All are expressed in central nervous system structures and are reported to guide axonal projections through interaction with receptors encoded by two genes known as Roundabout (Robo)-1 and -2 (Whitford et al., 2002; Brose et al., 1999). Clues about Slit and Robo function were first obtained in Drosophila where mutants displayed phenotypes of chaotic axonal projections (Seeger et al., 1993; Kidd et al., 1998, 1999). A substantial body of literature now confirms that Slit–Robo interactions control where and when axonal projections cross the midline as well as other key aspects of axonal organization (Plump et al., 2002; Bagri et al., 2002; Whitford et al., 2002).

Several reports carefully analyze the developmental expression of various Slit and Robo genes in the central nervous system (Holmes et al., 1998; Yuan et al., 1999; Marillat et al., 2002). Collectively, they document distinct, complementary temporal–spatial expression patterns, reflecting the importance of directing axonal organization as the central nervous system grows and develops. These studies, along with others (Piper et al., 2000; Vargesson et al., 2001), also document non-neuronal expression of Slit and Robo in limb buds and kidney. Kramer et al. (2001) demonstrated the importance of Slit–Robo interactions in mesodermal migration and myogenesis. Three manuscripts (Xian et al., 2001; Clark et al., 2002; Anselmo et al., 2003) report pulmonary epithelial expression of Robo-1. We could not detect Robo-1 expression in our coculture system, prompting our expanded analysis of pulmonary Slit and Robo gene expression. Here, we present data demonstrating developmentally regulated Slit and Robo expression in lung. Pulmonary mesenchyme is the primary site of expression in a pattern suggesting that Slit and Robo may mediate organization and orientation of pulmonary mesenchymal derivatives, including endothelial structures and smooth muscle.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

We previously reported that MFLM-91U cells organize to form tubular networks when cultured on reconstituted basement membrane (Akeson, et al., 2000, Fig. 1A). When cultured under similar conditions, the pulmonary epithelial cell line MLE-12 does not organize to form higher ordered structures beyond small clusters of five to eight cells (Fig. 1B). In contrast, when cultured together, MLE-12 cells rapidly align with the tubular structures formed by MFLM-91U cells (Fig. 1C). MLE-12 cells are phenotypically similar to progenitors of type II pneumocytes. They express surfactant protein C but do not express other surfactant proteins. MLE-12 cells do not posses lamellar bodies or secrete surfactant and, therefore, correspond to airway epithelial cells characteristic of the pseudoglandular stage of pulmonary development.

thumbnail image

Figure 1. A: MFLM-91U cells organize into thick (arrowheads) and thin (arrow) multicellular tubular structures on Matrigel. Tube formation is established within 12 hr after initial plating and continues to evolve for at least 5 days. B: MLE-12 cells cultured alone on matrigel do not organize into higher-order structures when cultured under the same conditions as shown in A. C: Laser confocal microscopy depicting MFLM-91U cells expressing green fluorescent protein cocultured with MLE-12 cells labeled with the red fluorescent dye PKH 26. The MLE cells (asterisks) align with the tubular structure formed by the MFLM cells within 12 hr of plating. Scale bars = 10 μm in A,B, 3 μm in C.

Download figure to PowerPoint

We used this in vitro coculture system to identify gene products responsible for the alignment of airway epithelium with endothelium. Complementary DNA isolated from MLE-12 and MFLM-91U cells cultured separately or together for 90 min on reconstituted basement membrane was used to probe a microarray containing more than 7,500 mouse cDNAs. By using a competitive hybridization strategy, we identified genes preferentially expressed in each culture condition (Table 1).

Table 1. Expression Analysis of MLE-12 and MFLM-91U Cells Cultured on Reconstituted Basement Membrane (Matrigel)a
Highest expression in cocultureHighest expression in independent culture
  • a

    Equivalent numbers of each cell line were cultured in separate plates (independent culture) or in the same plate (coculture). After 90 minutes in culture on matrigel, RNA was extracted and submitted to the manufacturer for competitive hybridization with a Mouse Gem 1 microarray as described in the Experimental Procedures section. The table lists the 10 genes on the microarray having the highest expression in cocultured and independently cultured conditions. The identity of each expressed sequence tag (EST) on the microarray was determined with BLASTN (Altschul et al, 1997). Accession numbers for each EST and corresponding gene are indicated.

EST (AA 423126); 98% identity with Slit-3 (AF144629)EST (AA028346); 100% identity with Keratin complex 1, acidic, gene 19 (M28698)
EST (AA067890); 100% identity with SRB-7 (NP079591)EST (AA 475774); 100% identity with Cathepsin C (NM009982)
Angiopoietin-related protein (AF125176)EST (W98974); 100% identity with CD24a antigen (X72910)
EST (W42321); 100% identity with PTX3, Pentraxin-3 (X83601)EST (AA050030); 100% identity with CAPN6, Calpain-6 (NM007603)
EST (AA064307); 99% identity with CD34 antigen (NM133654)EST (AA 120432); 100% identity with Solute carrier family 21 (prostaglandin transporter), (AK009434)
EST (AA266002); 100% identity with BCL-3 (AF067774)EST (W10023); 100% identity with β-catenin (BC048153)
EST (AA068750); 100% identity with SDF-1β, stromal cell derived factor, (L12030)EST (W15720); 98% identity with Synaptogyrin 2 (NM009304)
EST (AA241281); 100% identity with AQP1, Aquaporin-1 (NM007472)EST (AA002481); 100% identity with mouse β-5A integrin (AF043257)
EST (AA265259); 100% identity with Oncostatin M, (AF058805)EST (AA002292); 100% identity with Mus musculus homolog of Drosophila eyes absent (NM010165)
EST (W89883); 100% identity with Procollagen type IIIα (X52046)EST (AA059687); 92% identity with β-galactosidase α-2,6 sialyltransferase (D16106)

We were intrigued to find increased expression of a mouse homologue of Drosophila Slit by cocultured MFLM-91U and MLE-12 cells. Given its known function as an axonal guidance protein, we postulated that it might play a similar role by mediating the alignment of airway and pulmonary vasculature during lung development. By using the limited amounts of RNA available from the independently and cocultured cells, we used RT-PCR to obtain qualitative evidence of increased Slit-3 expression in cocultured MFLM-91U and MLE-12 cells compared with their independently cultured counterparts (data not shown). We also verified the expression of Slit-2 and Slit-3 in MFLM-91U cells cultured on reconstituted basement membrane. Of interest, we did not detect expression of the Robo-1 or Robo-2 receptors for Slit on MLE-12 cells cultured alone on reconstituted basement membrane. To gain further understanding of Slit and Robo function in the context of pulmonary development, we sought to determine their temporal and spatial patterns of expression.

We next screened for expression of Slit-1, 2, and 3 in embryonic day (E) 15.5 fetal mouse lung RNA by reverse transcriptase-polymerase chain reaction (RT-PCR). Expression of Slit-2 and Slit-3, but not Slit-1, was found. The absence of Slit-1 expression in lung was confirmed in total RNA extracted from adult mouse lung as well as various time points during fetal lung development. Robo-1 and Robo-2 were also expressed in fetal and adult mouse lung (Fig. 2).

thumbnail image

Figure 2. Expression analysis of Slit and Robo in midgestation mouse lung. RNA from embryonic day 15.5 fetal mouse lung (Lu) and brain (Br) were analyzed for expression of Slit and Robo genes by reverse transcriptase-polymerase chain reaction. Slit -1 is not expressed in fetal lung.

Download figure to PowerPoint

To define the expression of Slit and Robo in the developing lung, we used in situ hybridization. RT-PCR was used to isolate partial cDNA containing coding sequences of Slit-2, Slit-3, Robo-1, and Robo-2. Each clone was then used as a probe for in situ hybridization of mouse sections corresponding to consecutive embryonic days from E11.5 through postnatal day 1. The nucleotide sequence of each probe was screened for the absence of redundant regions to ensure that each was specific.

Previous reports document developmentally regulated expression of Slit and Robo genes in kidney and within specific neuroanatomic locations in the brain (Holmes et al., 1998; Yuan et al., 1999; Piper et al., 2000; Bagri et al., 2002; Marillat et al., 2002). To verify our findings in lung, we first analyzed expression of Slit-2, Slit-3, Robo-1, and Robo-2 in these organs at midgestation. Our findings were consistent with these previous reports (see Supplemental Material, which is available at http://www.mrw.interscience.wiley.com/suppmat/1058-8388/suppmat/index.html).

In the developing lung, Slit-2 expression was most prominent in pseudoglandular fetal lung (E14.5–E16.5). A homogeneous signal was observed throughout the mesenchymal compartment and larger airway epithelium (Fig. 3).

thumbnail image

Figure 3. Analysis of Slit-2 expression in developing lung. A–C: Embryonic day 11.5. D–F: Embryonic day 14.5. The darkfield view in A and corresponding brightfield view in B demonstrate signal in mesenchyme and epithelial structures in lung (L) and trachea (Tr). There is strong Slit-2 expression in an adjacent vessel (v). A control section using a sense probe on an adjacent section is shown in C. In D (darkfield) and E (brightfield), a similar pattern is apparent with expression in mesenchyme and epithelial compartments. The asterisk in E denotes a large airway clearly demonstrating epithelial signal. Background signal generated by a sense probe on an adjacent section is shown in F. Scale bar = 50 μm in A (applies to A–F).

Download figure to PowerPoint

Expression of Slit-3 in lung was first detectable at E11.5 in mesenchyme surrounding nascent lung buds. By E14.5, intense mesenchymal expression is present in a gradient from proximal (highest level of expression) to distal (lowest level of expression). By E16.5 (late canalicular stage), mesenchymal expression is decreasing with the exception of large pulmonary vessels. Slit-3 expression is still detectable in postnatal day 1 lung where it is restricted to the endothelium of large vessels associated with conducting airways. Small vessels and airways do not express Slit-3 (Fig. 4).

thumbnail image

Figure 4. Slit-3 expression in during fetal and early postnatal development. A–C: Embryonic day (E) 11.5. D–F: E14.5. G–I: E16.5. J–L: Postnatal day 1. In A, Slit-3 expression detected with an antisense probe in early pseudoglandular lung (L in the brightfield view, B) is restricted to mesenchyme. In D, expression is particularly strong in mesenchyme surrounding proximal airways. By E16.5 (G), pulmonary vessels retain high expression. Some expression in other mesenchymal cells persists. In postnatal lung, Slit-3 expression is restricted to large pulmonary vessels. Scale bar = 50 μm in A (applies to A–L).

Download figure to PowerPoint

Robo-1 is first detectable in fetal lung at E13.5 in mesenchyme. By E14.5 (mid-pseudoglandular), obvious homogeneous expression of Robo-1 in lung mesenchyme is apparent. The mesenchymal expression pattern is diminished in intensity but maintained through E15.5. We cannot rule out the possibility of low-level epithelial expression at this stage of lung development. However, if present, it is at or near the limits of detection by our in situ methods. We were unable to detect expression of Robo-1 in lung beyond E15.5, despite prolonged (>4 weeks) exposure times (Fig. 5).

thumbnail image

Figure 5. Robo-1 expression during lung development. A–C: Embryonic day (E) 13.5. D–F: E14.5. G–I: E15.5. Antisense probes at each stage shown detect homogenous mesenchymal expression (A,D,G). No epithelial signal is seen. B, E, and H depict brightfield views of the adjacent antisense panels. Sequential sections hybridized with sense probes demonstrate background (C,F,I). Scale bar = 50 μm in A (applies to A–I).

Download figure to PowerPoint

Robo-2 expression is also restricted to mesenchyme in the developing lung. In contrast to Robo-1, mesenchymal expression of Robo-2 is heterogeneous with a particularly intense in situ signal detected in mesenchymal cells immediately adjacent to proximal airway epithelium. This pattern is most pronounced at the late pseudoglandular stage of development (Fig. 6). By postnatal day 1, Robo-2 expression is barely detectable above background (data not shown). At this developmental stage, expression appeared to be alveolar, although resolution limitations afforded by our in situ methods did not allow specific cellular localization of expression.

thumbnail image

Figure 6. Robo-2 expression during lung development. A: An antisense probe for Robo-2 detects expression in lung mesenchyme (L) immediately adjacent to developing airways. B: A sense control using a sequential section is shown. C,D: By E13.5 (darkfield, C; brightfield, D), mesenchymal expression is more intense, particularly in cells adjacent to airway epithelium in lung (L) and mainstem bronchus (Br). No expression is detected in aorta (Ao). In the brightfield section (D), silver grains appear black. E,F: At E16.5 (darkfield, E; brightfield, F), Robo-2 signal is still clearly present in lung (L) and mainstem bronchi (Br). There is also intense signal in the mantle layer of the spinal cord (SC). Vertebral body (Ver) and esophagus (Eo) do not express Robo-2. Scale bars = 50 μm in A (applies to A–D), 50 μm in E (applies to E,F).

Download figure to PowerPoint

To verify that Robo-2 expression is restricted to mesenchyme and not expressed in adjacent epithelium, we analyzed expression of Robo-2 in isolated mesenchymal and epithelial remnants recovered from E11.5 lungs by microsurgical techniques (Fig. 7). By RT-PCR, Robo-2 mRNA was detected only in mesenchyme. The quality of microsurgical separation was verified by analyzing each remnant for surfactant protein C (SP-C) expression. As expected, only the epithelial remnant expressed SP-C.

thumbnail image

Figure 7. Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of Robo-1 (406bp) and Robo-2 (823 bp) expression in isolated embryonic day 11.5 epithelium (E) and mesenchyme (M). Both genes are expressed in mesenchyme but not epithelium. The quality of the epithelial–mesenchymal separation is demonstrated by RT-PCR for surfactant protein C (SP-C; 337 bp). Signal is detected only in RNA extracted from epithelium. Amplification of actin (348 bp) was performed to verify RNA integrity. C, control.

Download figure to PowerPoint

Slit and Robo have been associated with muscle patterning during development (Kramer et al., 2001). We performed immunohistochemistry of midgestation lung sections for smooth muscle alpha actin (SMAA) to evaluate the relationship between Slit and Robo expression in cells destined to become smooth muscle. We found that, like Robo-2, mesenchymal cells expressing SMAA lie immediately adjacent to airway epithelial cells in the midgestation lung (Fig. 8). Thus, Robo-2 may define the subpopulation of mesenchymal cells destined to become airway smooth muscle. Rhee et al. (2002) have presented data suggesting that Robo signaling in neurologic development may be mediated through interactions with N-cadherin. We previously noted expression of cadehrin-11 in developing lung mesenchyme (Greenberg et al., 2002). Here, we find expression of cadherin-11 is most intense in cells located in the same region as cells expressing Robo-2 (Fig. 8).

thumbnail image

Figure 8. Expression of Robo-2, smooth muscle alpha actin, and cadherin-11 in periepithelial mesenchyme of pseudoglandular lung. A: Robo-2 expression as detected by in situ hybridization is restricted to mesenchyme immediately adjacent to epithelium (additional sections shown in Fig. 6). B,C: Cadherin-11 (black nickel-diaminobenzidine staining, B) and smooth muscle alpha actin (red 3-amino-9-ethyl-carbazole [AEC] staining, C) are also expressed in a similar subset of lung mesenchyme at this gestation.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Slit–Robo interactions can mediate repulsion or attraction of cells relative to specific targets (Kramer et al., 2001). Specific function appears to depend upon the cell type and the experimental system used for analysis. Mice deficient in Slit-1 and Slit-2 display multiple defects in axonal patterning involving retinal ganglionic cells, as well as axonal projections derived from thalamus and frontal cortex (Plump et al., 2002; Bagri et al., 2002). Careful analysis of these mice has suggested that Slit–Robo interactions are required to direct axons on specific pathways and define sites of decussation. Additional in vitro evidence suggests additional roles in cortical dendritic outgrowth (Whitford et al., 2002).

Several previous studies of Slit and Robo expression have documented expression in non-neuronal tissues. Piper et al. (2000) describe expression of Slit-1, Slit-2, Slit-3, Robo-1, and Robo-2 in developing kidney. Slit-1 mRNA was present in metanephric mesenchyme, whereas the other Slit and Robo genes were expressed in the tips of the invading ureteric tree and nascent glomeruli. Developmentally regulated Slit and Robo expression have also been documented in the developing limb (Vargesson et al., 2001). As with neuronal development, each Slit and Robo gene has a distinct, developmentally regulated expression pattern. In aggregate, the three Slit genes are expressed in a complementary relationship to Robo-1 and Robo-2. This pattern of expression is consistent with a functional ligand–receptor relationship directed toward positioning of specific tissue compartments of a developing organ.

We found expression of Slit and Robo in mouse lung is developmentally regulated and primarily restricted to mesodermal derivatives. This tissue compartment undergoes dramatic remodeling during lung development. During early stages, the volume of mesenchyme exceeds that of invading endoderm. As branching and alveolarization proceed, the mesenchyme thins and differentiates into smooth muscle, distal vasculature, and fibroblasts. This process incorporates apoptosis and cell movement in addition to differentiation. It is reasonable to postulate that Slit and Robo, which are known to direct cell movement and orientation in other systems, perform similar functions in pulmonary mesenchyme. Slit-3 and Robo-2 mark specific complementary subsets of mesenchyme in the developing lung. Smooth muscle alpha actin production colocalizes with mesenchyme expressing Robo-2. It will be interesting to determine whether Slit-3 and Robo-2 interactions determine the positioning of smooth muscle precursors in developing airways.

Xian and coworkers studied the consequences of a targeted deletion of exon 2 of Dutt1, an alternative splice form of Robo-1. Exon 2 is shared between both splice forms (Robo-1 and DUTT1; Xian et al., 2001). Mice homozygous for the deleted exon exhibited delayed lung maturation and increased neonatal mortality. Those surviving to adulthood had focal areas of epithelial hyperplasia in large airways. The authors suggest that mice lacking exon 2 of DUTT1 may prove to be a useful model for biological studies of premalignant neoplasia and neoplastic transformation in lung. These mice had normal nervous system anatomy and development, suggesting a significant degree of redundancy for Slit–Robo signaling in neuroectodermal development. By using a rabbit polyclonal antibody, they localized Dutt1 to airway epithelium. We did not detect Robo-1 transcripts in airway epithelium, perhaps due to the use of in situ hybridization rather than an antibody-based detection method. The pattern of Robo-1 expression we detected in developing brain and kidney by in situ hybridization does correspond with previously published reports (see Supplemental Material), demonstrating that the Robo-1 signal we detected in lung mesenchyme is specific.

Our studies have focused on Slit and Robo expression in the fetal lung with an emphasis on the late pseudoglandular period, a time of dramatic reorganization within the mesenchymal compartment. During this developmental stage, Slit and Robo expression appear restricted to mesenchyme. By using polyclonal antibodies directed against Robo-1, Clark et al. (2002) reported diffuse expression in mesenchyme and epithelium of fetal mouse lung, followed by more restricted expression in lung epithelium beyond fetal day E17.5. By using their antibody, we also detected diffuse staining in lung before E17.5 and sporadic staining of epithelium beyond that embryonic stage. Preimmune rabbit serum generated patchy staining in epithelium, suggesting nonspecific reactivity. Appropriate specificity controls were not available to further elucidate the explicit spectrum of its activity in our hands. Anselmo et al. (2003) also used immunohistochemical methods to detect Slit and Robo expression in developing mouse lung. They found Slit-2 and Slit-3 expression in mesenchyme and some epithelial sites at early stages of lung development. Immunofluorescence attributed to Robo-1 and Robo-2 expression was also detected in mesenchyme and airway epithelium as early as embryonic day 12. We used the same antibodies against Robo-1, Robo-2, Slit-2, and Slit-3 (Santa Cruz) to detect these gene products by immunohistochemical methods. At dilutions ranging from 1:50 to 1:200, with or without antigen retrieval, we detected only nonspecific signal in epithelium and mesenchyme that could not be inhibited with blocking peptide supplied by the manufacturer. In addition, the antibodies did not detect expression of Slit or Robo in brain or kidney.

These findings led us to question the utility of these antibodies for immunohistochemistry and prompted our focus on in situ hybridization as an alternative strategy. Our in situ studies of Robo-1 and Robo-2 did not detect expression of either gene before E13.5. We took advantage of the ability to separate E11.5 pulmonary epithelium from mesenchyme to evaluate Robo-1 and Robo-2 expression. As shown in Figure 8, we confirmed the absence of Robo-1 and Robo-2 expression in epithelium at E11.5 using RNA purified from isolated epithelium and mesenchyme in a sensitive RT-PCR assay. The lack of SP-C expression in the isolated mesenchyme confirms the complete segregation of epithelium from mesenchyme in this experiment. Thus, at early stages of pulmonary development, Robo-1 and Robo-2 are expressed by mesenchyme.

The colocalization of SMAA and Robo-2 in mesenchyme may reflect a role in defining organization of smooth muscle in the developing lung. Signaling through Robo-2, perhaps by Slit-2, may direct appropriate localization of smooth muscle precursors. Thus, disruption of Slit–Robo interactions in this cellular compartment may lead to inappropriate muscularization of airways. Inhibitors of Robo signaling such as Robo-N (Wu et al., 2001) may prove useful to test such a hypothesis.

Rhee et al. (2002) reported a mechanistic link between Slit–Robo signaling and axonal movement. When activated by Slit, Robo inactivates N-cadherin–mediated cell–cell adhesion. This process is mediated through increased phosphorylation of β-catenin and subsequent disruption of the linkage between N-cadherin and the actin cytoskeleton. Cadherin-11 is expressed in fetal mesenchyme where it is thought to mediate cell–cell interactions (Kimura et al., 1995). It is intriguing to consider that Slit–Robo interactions in developing lung mesenchyme might modulate cadherin-11 function, providing a mechanism for mesenchymal organization.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Animals and Tissue Preparation

All experiments involving animals were performed in accordance with National Institutes of Health Guidelines on the Care and Use of Laboratory Animals and were approved by the Animal Care Committee of the Cincinnati Children's Hospital Research Foundation. After death by CO2 asphyxiation, fetal mouse pups were removed from timed-pregnant FVB/N dams. The day of vaginal plugging was defined as E0.5. Lungs were removed intact and snap frozen in liquid nitrogen. E11.5 lung epithelium and mesenchyme were isolated as previously described (Shannon et al., 1998; Greenberg et al., 2002).

Cell Culture

Culture of MLE-12 cells (Wikenheiser et al., 1993) and MFLM-91U cells (Akeson et al., 2000) were performed as previously reported. For coculture experiments, 3 × 104 MLE-12 and 3 × 104 MFLM-91U cells were plated at a density of 3 × 104 cells/cm2 in separate or the same culture plates treated with reconstituted basement membrane (Matrigel). All experiments were performed with the same lot of Matrigel.

Expression Analysis

After 90 min of incubation at 37°C in 5% CO2 mRNA was isolated and quantified for microarray analysis (Mouse GEM 1, Incyte). This array included 7,801 unique genes derived from The Institute for Genomic Research (TIGR) Mus.ET database as well as expressed sequence tags from the GenBank mouse database. Differential gene expression was evaluated by submitting equimolar amounts of mRNA from MFLM-91U and MLE-12 cells cultured separately or together (cocultured) to the manufacturer for labeling and microarray hybridization. They prepared fluorescently labeled cDNA from the cocultured cells (Cy-3) and independently cultured cells (Cy-5). Equimolar amounts of cDNA from each culture condition were applied to the microarray for competitive hybridization. The 10 genes showing the highest levels of preferential hybridization with cDNA from cocultured cells or independently cultured cells are listed in Table 1.

RNA Isolation and Analysis by Reverse Transcriptase Polymerase Chain Reaction

Total RNA was isolated from lungs derived from FVB/N mice at various stages of development by using TRIzol (GibcoBRL). Poly A+ RNA was isolated from E11.5 epithelium and mesenchyme by using a MicroFastrak kit (Invitrogen). For RT-PCR analysis of Slit-1, -2 -3, Robo-1, and -2, and surfactant protein C (SP-C), first-strand cDNA was synthesized from 5 μg of total RNA by using Superscript reverse transcriptase (GibcoBRL). The cDNAs were purified with Qiaquick PCR purification system (Qiagen) and quantified before amplification. Amplifications were carried out by using 10 ng of first-strand cDNA. Conditions and primers for actin and SP-C were as previously described (Akeson et al., 2000). For Slit and Robo, the following primer pairs were used: Slit-1, 5′-AGCCGTGCCGAAACTCTACTG-3′ and 5′-AGCGGGCGTGTGTGGTCTGG-3′; Slit-2, 5′-CCTGCCCATCAATGCCTTCTCCTA-3′ and 5′-GCAGCCGCACTTCACCACTTTCTC-3′; Slit-3, 5′-ACGAGGTGCGCATCAACAACGAG-3′ and 5′-GC-GGCGGATGGCTTCACGGACTAT-3′; Robo-1, 5′-ACATGCCCCACCCACCAGACA-3′ and 5′-CCACTTTCAGGCCCGCATACTCC-3′; Robo-2, 5′-CAGGGCCGGACAGTGACATTC-3′ and 5′-CTTGGGGGTTGATCGCTCTGA-3′.

The PCR was carried out for each primer pair for 30 cycles. Each cycle included a denaturation step of 30 sec at 94°C and a polymerization step of 45 sec at 72°C. Annealing temperatures were 62°C for Slit-1, 58°C for Slit-2, 59°C for Slit-3, 59°C for Robo-1, and 58°C for Robo-2. The annealing time was 45 sec for all genes analyzed. Product sizes were as follows: Slit-2, 419 bp; Slit-3, 462 bp; Robo-1, 406 bp; Robo-2, 823 bp; β-actin, 348 bp; SP-C, 337 bp.

Preparation of Probes

In situ probe templates were generated by RT-PCR from E15.5 mouse lung mRNA by using the above primer pairs. PCR products were size-fractionated by agarose gel electrophoresis. Each PCR product was gel purified and sublconed into pCRII (Invitrogen). The identity and orientation of each clone was verified by nucleotide sequencing. After linearization of plasmid with the appropriate restriction enzyme, radiolabeled RNA probes were synthesized with T7 or SP6 RNA polymerase (Promega) using 300 μCi 35S-UTP (DuPont). Probes were reduced to an average size of 200 bp by alkaline hydrolysis.

In Situ Hybridization

Mouse tissues for in situ analysis were fixed in 4% paraformaldehyde for 18–24 hr at 4°C, dehydrated in ethanol, and paraffin embedded for sectioning. Whole embryos, sectioned along a parasagittal plane to include lung, brain, and kidney were used for embryonic days 11.5 through 14.5. From E15.5 through postnatal day 1, the thorax was isolated before fixation. Coronal sections, including lung, were subsequently cut for analysis. Two consecutive 5-μm sections were mounted on a single silane-coated glass slide. After deparaffination in xylene, the sections were rehydrated, post-fixed in 4% paraformaldehyde for 20 min, briefly treated with proteinase K (20 μg/ml), refixed, and dehydrated once again in ethanol. Antisense and sense probes for a given gene were each applied to consecutive sections on a single slide to facilitate comparison between antisense and sense signals. At least three slides were analyzed for each embryonic day. After an overnight hybridization at 55°C, free probe was removed by washing in 2–5× SSC at 65°C for 30–60 min. The sections were then dehydrated once again in ethanol, allowed to air-dry, and dipped in photographic emulsion (Kodak NTB2). Exposure times ranged from 7 to 21 days at 4°C. After development (Kodak D19), the sections were washed, refixed in 4% paraformaldehyde, dehydrated in ethanol, and cover-slipped. To facilitate localization of in situ signals, some sections were counterstained with hematoxylin.

Immunohistochemistry

Detection of smooth muscle alpha actin and cadherin-11 was performed as previously described (Akeson et al., 2000; Greenberg et al., 2002). Briefly, after fixation in 4% paraformaldehyde, tissues were washed in phosphate-buffered saline, dehydrated through a graded series of ethanol, and embedded in paraffin. Sections cut at 4 μm were mounted on polylysine-coated slides. Antigen retrieval was performed by microwaving at full power for 7.5 min in 10 mM sodium citrate, pH 6.0. For detection of cadherin-11, a rabbit polyclonal antibody was used (Zymed, South San Francisco, CA). Smooth muscle alpha actin was detected with a mouse monoclonal antibody (Sigma, St. Louis, MO). Biotinylated goat anti-rabbit (cadherin-11) or goat anti-mouse (smooth muscle alpha actin) secondary antibodies (Vector, Burlingame, CA) were used for detection followed by nickel–diaminobenzidine for cadherin-11 or 3-amino-9-ethyl-carbazole (AEC) for smooth muscle alpha actin.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

We thank Susan Wert, PhD, for her advice and comments regarding in situ hybridization. J.M.G. and J.M.S. were funded by the March of Dimes Birth Defects Foundation.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The Supplementary Material referred to in this article can be found at http://www.mrw.interscience.wiley.com/suppmat/1058-8388/suppmat/index.html

FilenameFormatSizeDescription
suppmat_fig1_greenberg.gif1897KSupporting Information file suppmat_fig1_greenberg.gif
suppmat_fig2_greenberg.gif2084KSupporting Information file suppmat_fig2_greenberg.gif

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.