Timeless in lung morphogenesis

Authors

  • Jing Xiao,

    1. Department of Pediatrics, Women's & Children's Hospital & Department of Medicine, Will Rogers Institute Pulmonary Research Center, School of Medicine, University of Southern California, Los Angeles, California
    Search for more papers by this author
  • Changgong Li,

    1. Department of Pediatrics, Women's & Children's Hospital & Department of Medicine, Will Rogers Institute Pulmonary Research Center, School of Medicine, University of Southern California, Los Angeles, California
    Search for more papers by this author
  • Nian-Ling Zhu,

    1. Department of Pediatrics, Women's & Children's Hospital & Department of Medicine, Will Rogers Institute Pulmonary Research Center, School of Medicine, University of Southern California, Los Angeles, California
    Search for more papers by this author
  • Zea Borok,

    1. Department of Pediatrics, Women's & Children's Hospital & Department of Medicine, Will Rogers Institute Pulmonary Research Center, School of Medicine, University of Southern California, Los Angeles, California
    Search for more papers by this author
  • Parviz Minoo

    Corresponding author
    1. Department of Pediatrics, Women's & Children's Hospital & Department of Medicine, Will Rogers Institute Pulmonary Research Center, School of Medicine, University of Southern California, Los Angeles, California
    • Women's & Children's Hospital, LAC+USC Medical Center, 1801 E. Marengo Street, Room 1G1, Los Angeles, CA 90033
    Search for more papers by this author

Abstract

The Clock gene, timeless, regulates circadian rhythm in Drosophila, but its vertebrate homolog is critical to embryonic development. Timeless was shown to be involved in murine urethral bud branching morphogenesis. We generated a polyclonal antibody to mouse TIMELESS (mTIM) and studied its distribution and its potential role during lung development, which also requires branching morphogenesis. In the early mouse embryo, TIM was localized to all organs, especially the neural epithelium. In embryonic day (E) 9.5 embryos, TIM was present in both epithelial and mesenchymal cells at the onset of lung morphogenesis. In E15 embryos, TIM decreased in the mesenchyme but remained pronounced in the epithelium of both large and small airways. Later, TIM was localized to a specific subset of epithelial cells with alveolar type 2 phenotype. This finding was verified by immunostaining of isolated alveolar type 2 cells. In the proximal airways, TIM was colocalized with CCSP to nonciliated columnar epithelial cells. Antisense oligonucleotides to mTim specifically inhibited branching morphogenesis of embryonic lungs in explant culture without affecting SpC expression an alveolar type 2 cell marker. In cultured lung cells, expression of TIM is independent of cell cycle and proliferation. These studies indicate that the function of Timeless is highly conserved in organs whose formation requires branching morphogenesis. Developmental Dynamics 228:82–94, 2003. © 2003 Wiley-Liss, Inc.

INTRODUCTION

Branching morphogenesis represents an evolutionarily conserved strategy used to increase the surface area of a given organ. In Drosophila, the tracheal network, a simple respiratory organ, is composed of a branched structure that delivers air from the spiracular openings to nearly all the cells of the fly. Classic, and as of recent, molecular genetic analyses of Drosophila tracheal development have uncovered several genes with key roles in the process of branching morphogenesis. For example, members of the FGF signaling pathway appear to be critical to the formation of the tracheal network (Min et al., 1998). Analysis of branching morphogenesis in mammalian organs, particularly in lung and kidney development, has also led to independent discoveries, some of which, such as the FGF pathway, are clearly evolutionary parallels of Drosophila tracheal morphogenesis (Min et al., 1998). Comparisons between the Drosophila simple tracheal network and the complex structure of the mammalian lungs suggest that branching morphogenesis entails specific common fundamental mechanisms.

In mammals, development of both kidneys and the lungs requires branching morphogenesis. Lung development in the mouse commences at around day 9.5 of gestation (E9.5) (Hogan, 1999; Cardoso, 2000; Warburton et al., 2000). The lung rudiment arises from a select group of endodermal cells that are thought to be set aside early in the development of the gut. The lung primordium grows by branching against the splanchnic mesenchyme to form the bronchial tree and the elaborate, highly branched network of the pulmonary distal airways and the prealveolar structures. Branching morphogenesis requires precise integration of several cellular processes that include proliferation, migration, and apoptosis. As such, many mediators, both transcription factors and signaling molecules are expressed during branching morphogenesis with precise spatial and temporal specificity (Minoo, 2000). For example, NKX2.1, a homeodomain transcriptional regulator is expressed in the lung epithelium and is absolutely necessary for progression of branching morphogenesis and, hence, formation of distal lung structures (Minoo et al., 1999). Amongst the signaling molecules, inhibition of Sonic hedgehog (SHH) or bone morphogenetic protein-4 (BMP-4) results in various degrees of branching inhibition (Pepicelli et al., 1998; Weaver et al., 1999). Abrogation of fibroblast growth factor-10 (FGF-10) leads to cessation of branching altogether (Min et al., 1998). Although all mediators of branching morphogenesis may in some way impact epithelial or mesenchymal cellular proliferation, specific mediators of the cell cycle have also been implicated in branching morphogenesis. Thus, abrogating platelet-derived growth factor (PDGF) signaling interferes with normal branching by inhibiting cellular proliferation (Lindahl et al., 1997).

Development of the mammalian kidney also is dependent on branching morphogenesis. Thus, several key functional parallels have been found between kidney and lung development. For example, both BMP and Wnt pathways, which play critical roles in branching morphogenesis of the ureteric buds (UB), also figure importantly in lung epithelial morphogenesis (Weaver et al., 1999; Li et al., 2002a). Recently, the vertebrate homolog of the Drosophila Timeless, a nuclear protein that participates in transcriptional regulation of several genes in the circadian clock was shown to be involved in the process of UB branching morphogenesis (Li et al., 2000). Mouse and human orthologs of the Drosophila timeless gene (dtim) were cloned by a database search for mammalian expressed sequence tags (ESTs) (Koike et al., 1998). Predicted translation of the cloned cDNAs showed that both human and mouse TIMELESS proteins (hTIM and mTIM, respectively) shared extensive sequence homology with Drosophila TIM. However, although the latter proteins fulfilled several functional criteria to be the ortholog of dTIM, unlike Drosophila, their mRNA levels did not oscillate in the superchiasmatic nucleus, the control center for the mammalian circadian rhythms. Human and mouse TIM proteins are approximately 1200 amino acids and share four regions of sequence conservation with dTIM, designated as TIM-homology domains (Zylka et al., 1998; Sangoram et al., 1998). These proteins include a nuclear localization signal and are thought to be involved in regulation of transcription. For example, in regulating the circadian rhythm, TIM, both in Drosophila and in mammals, is thought to dimerize with the product of the Period (per) locus, enter the nucleus and inhibit, through a negative feed-back loop, BMAL1/CLOCK-activated transcription, including their own (Saez and Young, 1996; Sangoram et al., 1998). In mouse adult tissues, a single mTim transcript of 2.4 kb was expressed in all organs, including heart, brain, spleen liver, kidney, and lung (Koike et al., 1998; Li et al., 2000). The relationship between PER and TIM is controversial. Whereas Sangoram et al. (1998) found interactions between TIM and mouse PER1 and PER2, such interactions could not be found in studies reported by Kume et al. (1999). The latter investigators also found no correspondence between mTIM and PER peripheral tissue distribution by Northern analysis. Thus, the role of TIM in regulation of the mammalian clock remains unknown.

The mouse Tim (mTim) locus has been inactivated by homologous recombination (Gotter et al., 2000). Homozygous Tim (-/-) embryos die early in utero with a general lack of cellular organization and necrotic cell death (Gotter et al., 2000). Developmental abnormalities were detected in embryos as early as E5.5, suggesting that mTim is important around implantation. The role of mTim in development of visceral organs has been addressed in kidney morphogenesis and, in particular, the branching morphogenesis of the urethral bud (Li et al., 2000). Tim expression has been found in both the mesenchyme and the epithelium of the developing rat kidney rudiment (Li et al., 2000). Antisense oligodeoxynucleotide (AS-ODN) interference with TIM synthesis resulted in inhibition of UB branching morphogenesis in explanted rat kidneys (Li et al., 2000). We have taken advantage of the similarities between kidney and lung development to address the question of whether TIM is a conserved component of the process of branching morphogenesis. Accordingly, we analyzed mTIM expression and function in mouse lung morphogenesis. The present study shows that selective repression of mTIM by AS-ODNs inhibits embryonic mouse lung branching morphogenesis without altering cell differentiation. Importantly, the function of mTIM does not appear to be related to the regulation of the cell cycle in lung epithelial cells in culture.

RESULTS

Characterization and Specificity of Anti-mTIM Antibody

The polyclonal anti-mTIM antibody was raised against a combination of three peptides, all of which were derived from domains that are highly conserved amongst TIM proteins from human, mouse, and rat. To establish the subcellular localization, cell type, and species-specificity of anti-mTIM, we performed Western blot analysis of both nuclear and cytoplasmic extracts of mouse NIH3T3 mesenchymal cells, the human lung epithelial carcinoma A549 and H441 cell lines, as well as the human cervical carcinoma cell line, HeLa. By using the anti-mTIM antibody as probe, the Western blot analysis showed that nuclear extracts of all cell lines contain a strong protein band of 120 kDa, consistent with the mRNA-based, predicted size of mTIM. In the cytoplasmic extracts of the same cell lines, a much weaker band of similar mobility was also detected, although the possibility that the latter represents nuclear contamination of cytoplasmic fractions could not be ruled out. Based on the strength of the antibody–antigen reaction, the highest level of TIM was found in the nuclear extracts of NIH3T3 cells. Thus, the generated antibody is highly specific in its detection of a major band of 120 kDa that is abundantly found in the nuclear extracts of both mouse and human cell lines. In addition, TIM is expressed in both mesenchymally and epithelially derived cell lines (Fig. 1).

Figure 1.

Specificity of mTIM antibody. Upper panel: Western blot analysis of mTIM in different cell lines. Nuclear (N) and cytoplasmic (C) extracts of A549, H441, NIH3T3, and HeLa cells were transferred onto Immobilon-P transfer membrane. The membrane was probed with the rabbit anti-mTIM antibody, which identified a protein band of 120 kDa in nuclear extracts of all cell lines, and a much weaker band of similar mobility in cytoplasmic extracts of the same cell lines. A–D: Characterization of mTIM antibody in 4% PFA-fixed mouse tissues. A,B: Sections of embryonic day 10 mouse whole embryo stained with hematoxylin and eosin and mTIM antibody, respectively. TIM immunostaining was apparent in most rudiments, including brain, lung, heart, liver, stomach, and urogenital ridge. MB, midbrain (higher magnification of nuclear staining shown in inset); HB, hindbrain; OS, optic stalk; TV, telencephalic vesicle; OP, olfactory epithelium; BA, bronchial arch; Tr, Trachea; H, heart; Lu, lung bud; Li, liver; UR, urogenital ridge; NT, neural tube. C,D: Positive nuclear staining of TIM in isolated PFA-fixed mouse E18 kidney and lung, respectively. Scale bars = 100 μm in B (applies to A,B), in D (applies to C,D).

To examine the utility of the anti-mTIM antibody for immunohistochemistry and to further verify its specificity, we first determined the distribution of mTIM in early, E10 whole mouse embryos. Paraformaldehyde-fixed and paraffin embedded E10 whole embryos were sectioned and prepared for immunostaining. Positive staining for mTIM was obtained in most organ rudiments, including brain, lung, heart, liver, stomach, and urogenital ridge. In particular, intense staining was observed in neural epithelium, such as in the roof of hindbrain, neuroepithelial wall of forebrain and midbrain, neural epithelium of neural tube, neural tube of tail, optic stalk, and olfactory epithelium (Fig. 1). In addition, paraformaldehyde-fixed, isolated tissues from later embryonic stages of mouse development were also examined. Nuclear staining of mTIM was observed in isolated kidney and lung tissues of E18 mouse embryos (Fig. 1C,D). Because paraformaldehyde is a strong fixative and not ideal for nuclear proteins, we verified the latter results by using E13 embryonic lungs fixed in Carnoy's fixative which has been shown to be ideal for nuclear antigens (Bassarova and Popov, 1998; Srinivasan et al., 2002). These results were identical to those obtained by using paraformaldehyde-fixed lung tissue. Positive staining was demonstrated in both mesenchymal and epithelial cells of E13 lungs (Fig. 2A). High-power magnification of the lung tissue showed that immunostaining in both mesenchymal and epithelial cells surrounding the airways was confined to the nucleus, with little if any staining in the cytoplasm (Fig. 2B). Specificity of the anti-mTIM antibody was further verified by either omission of the primary antibody or preabsorption with a mixture of peptides used in generation of the antibody. Under these conditions, positive staining in the epithelium or in the mesenchyme of the mouse embryonic lung was entirely abolished (Fig. 2C).

Figure 2.

Specificity of mTIM antibody in E13 embryonic mouse lung. A: Positive staining of mTIM in both mesenchymal and epithelial cells of Carnoy's-fixed E13 lungs. B: High-power magnification of (A), showing specific nuclear localization of TIM. Immunoreactivity for mTIM antibody was abolished by absorption with mTIM peptides (C) or elimination of mTIM antibody (D). The a designates airway, and bv designates blood vessels in the lung tissue. Scale bar = 50 μm in A,C,D and 12.5 μm in B.

Localization of mTIM in Early Mouse Embryo and During Lung Morphogenesis

Determination of spatial and temporal distribution as well as cell type specificity can provide valuable clues as to the potential role of a given protein. To examine the spatial and temporal pattern of mTIM distribution during lung development, we first determined its mRNA expression in total lung tissue excised from different embryonic stages as well as neonatal and adult animals. Northern blot analysis of poly (A)+RNA from different stage lung tissues was performed by using a cDNA probe for mTIM (GenBank accession no. AF126480; Li et al., 2000). The results showed a dynamic pattern of mTim expression during embryonic and postnatal lung development. The overall level of mTim transcripts decreased as lung development progressed. In E15 lungs, two mTim mRNA species of 2.4 kb and 4.4 kb were observed. As lung development progressed, the relative abundance of the two transcripts changed, so that, in E18 lungs, the 2.4-kb mRNA predominated over the 4.4-kb species (Fig. 3). In the lungs of 2-day-old neonatal mice, only a reduced 2.4-kb mRNA species was detectable by Northern analysis. In the adult lung, neither of the mTim transcripts was within the detection level of our Northern blot analysis. These data demonstrate a highly dynamic regulation of mTim expression during lung development.

Figure 3.

Expression of Timeless during mouse lung development. Upper panel: Northern blot analysis of poly (A)+ RNA isolated from embryonic day (E) 15, E18, PN2, and adult mouse lung. The blot was probed with a mouse Timeless cDNA (see Experimental Procedures section). The two mTim mRNA species of 2.4 kb and 4.4 kb are shown. A–F: Spatial distribution of mTIM during mouse lung development. Immunohistochemistry was performed to localize mTIM protein in embryonic lungs at various gestational ages, as designated. The a designates airway, and bv designates blood vessels. Scale bar = 40 μm in F (applies to A–F).

We next sought to determine the distribution and potential cell-type specificity of mTIM protein during lung development. Accordingly, immunohistochemical staining was performed on lung tissue derived from E11, E15, E18 embryos, and postnatal day 1 (PN1), PN5, and adult animals. In early embryonic stages, such as E11 and E15, mTIM was localized to the cell nucleus of both epithelial and mesenchymal compartments of the lung (Fig. 3A,B). The staining in the epithelial compartment appeared nearly uniform without a discernible proximodistal gradient. In E18 and PN1 peripheral lungs, the level of mTIM expression in the mesenchyme was reduced and nonuniformly distributed with clusters of cells showing little to no staining (Fig. 3C,D). In contrast to the mesenchyme, strong positive staining was observed in the epithelial layer of distal airways and saccular compartments in E18 and PN1. In addition, proximal lung structures in E18 embryos, including those of the trachea and the bronchi showed strong positive staining for mTIM (data not shown). Further postnatal development of the lung (PN5) was associated with nearly complete loss of mTIM staining in the mesenchyme (Fig. 3E). In PN5 and adult lung, only a subset of cells with alveolar epithelial type 2 cell phenotype surrounding the saccular or prealveolar structures were found to be positive for mTIM (Fig. 3E,F), and much more intense staining was observable in the trachea and the major airways. Positive staining of endothelial cells on blood vessels was also seen in embryonic lungs (Fig. 3C).

Cell Specificity of Timeless in the Mouse Lung

The pattern of staining observed in PN5 and adult lungs (Fig. 3E,F) suggested that mTIM may be specifically expressed in alveolar type 2 cells. To examine this possibility, we performed immunocytochemical staining on purified alveolar type 2 cells obtained from adult rat lungs. Furthermore, we also examined the expression of mTIM in alveolar type 1-like cells. The latter were derived by prolonged (8 days) in vitro culturing of purified alveolar type 2 cells (Borok et al., 1995). Figure 4 shows that purified type 2 cells uniformly exhibit the presence of nuclear localized mTIM (Fig. 4A). The use of a monoclonal antibody to NKX2.1 as a positive control demonstrates the purity of the isolated alveolar type 2 cell population (Fig. 4B). In contrast to the alveolar type 2 cells, little if any mTIM-positive staining was observed in alveolar type 1-like cells (Fig. 4C). The identity of the alveolar type 1-like cells was verified by the use of an antibody designated VIIIB2, which is a specific apical membrane marker for alveolar type 1 cell phenotype (Fig. 4D; Danto et al., 1992).

Figure 4.

Expression of mTIM in isolated alveolar epithelial cells. A: Rat alveolar type 2 cells immunostained with anti-mTIM antibody. B: Same cells as in A, immunostained with an anti-NKX2.1 antibody. C: Culture induced alveolar type 1-like cells immunostained with anti-mTIM. D: Same cells as in C, immunostained with VIIIb2 antibody. E,F: A549 and NIH3T3 fibroblasts, respectively, immunostained with anti-TIM antibody. Scale bar = 40 μm in F (applies to A–F).

Because mTIM was found to be expressed in the proximal lung, we also examined whether it is expressed in Clara, nonciliated bronchiolar epithelial cells. Immunohistochemical double-staining for mTIM and CCSP proteins was conducted in E18 lungs. CCSP is a well-documented marker for Clara cells (Zhou et al., 1996). mTIM and CCSP were found to be colocalized in bronchiolar epithelial cells as shown in Figure 5B. We also conducted double staining for alpha-smooth muscle actin and TIM in mouse embryonic lungs (E18). Alpha smooth muscle actin localized to the basal membrane of large blood vessels, around the major airways, and the endothelial cells and smooth muscle cells surrounding the same blood vessels showed positive nuclear staining for mTIM (Fig. 5D).

Figure 5.

Colocalization of mTIM with CCSP and alpha-smooth muscle actin (α-SMA) in bronchial epithelium and subendothelial smooth muscle precursors of embryonic day 18 lungs. A: Immunostaining with anti-CCSP antibody alone. B: Double immunostaining with anti-CCSP and mTIM antibodies. C: Immunostaining with anti–α-SMA alone. D: Double immunostaining with anti–α-SMA and mTIM. The a designates airway, and bv designates blood vessels. Scale bar = 20 μm in D (applies to A–D). Arrows indicate the basal membrane of bronchi in C and D.

Antisense Inhibition of mTim Abrogates Mouse Embryonic Lung Branching Morphogenesis

To determine the potential function of mTim during embryonic lung development, we used an AS-ODN strategy to repress the expression of endogenous TIM in murine embryonic lungs explanted in vitro (Li et al., 2000). This strategy has been used previously to ascertain the role of several key lung genes (Minoo et al., 1995). Accordingly, E11 mouse embryonic lungs were excised from embryos, divided into five groups and explanted separately in a serum-free organ culture medium BGJb (GIBCO, Grand Island, NY) supplemented with 50 units penicillin/ streptomycin. Two groups of lungs were treated with two separate antisense oligonucleotides, AS-ODN1 and AS-ODN2, each of which was designed to a separate region of the mTIM cDNA (see Experimental Procedures section). As control, we also treated E11 lungs with two ODNs whose sequence represented the sense strand of AS-ODN1 and AS-ODN2. The latter were designated as sense oligodeoxynucleotide 1 (S-ODN1) and S-ODN2, respectively (see Experimental Procedures section). In an additional control, a group of lungs was left untreated (medium control). All lungs were incubated identically for four days in culture. To examine the efficacy and specificity of the AS-ODNs, we compared the level of TIM protein in lungs treated with AS-ODNs with the level found in the two sets of control lungs (untreated or treated with S-ODNs). Figure 6 shows the results of the immunostaining for TIM protein for AS-ODN1 and its control S-ODN1. The lungs treated with S-ODN1 showed normal (compared with wild-type, not shown) level and distribution of TIM protein (Fig. 6A). In contrast, the lungs exposed to AS-ODN1 showed little if any positive mTIM immunohistochemical staining (Fig. 6B). To ascertain the specificity of AS-ODN1, we examined the expression of proliferating cell nuclear antigen (PCNA) in the lung sections by immunohistochemistry. Figure 6C,D shows that, in contrast to mTIM immunostaining, PCNA protein is detectable at approximately identical levels and pattern in all lung samples, regardless of whether they were treated with AS-ODN1 or S-ODN1. These data demonstrate that AS-ODN1 does not randomly repress gene expression in the explanted lungs. Nearly identical results were obtained when AS-ODN2 was used to inhibit mTIM expression (data not shown). Exposure to either AS-ODNs inhibited mouse embryonic lung morphogenesis, by up to 40%. These lungs appeared to have normal mesenchyme, but significantly reduced epithelial branching. In addition, as expected from reduced branching, the overall size of the lungs treated with AS-ODN1 and AS-ODN2 was also reduced compared with the controls (Fig. 7). In contrast, development of the equivalent embryonic lungs treated with the control S-ODN was similar or identical to the untreated (medium), control lungs. To ascertain whether abrogation of mTIM affected epithelial cell differentiation, we used reverse transcription-polymerase chain reaction to determine the expression of surfactant protein C (SP-C) mRNA in AS-ODNs– and S-ODNs–treated lungs. Figure 8 shows that SP-C mRNA is indeed expressed in lungs treated with mTIM AS-ODN1. These data demonstrate that abrogation of mTIM in the mouse embryonic lungs explanted in culture reduces epithelial branching without affecting cellular differentiation.

Figure 6.

Antisense oligodeoxynucleotides (AS-ODNs) to mTim mRNA inhibit mTIM protein synthesis. A: Cultured lungs treated with mTim sense oligodeoxynucleotide 1 (S-ODN1), showing normal level and pattern of mTIM protein. B: Cultured lungs treated with mTim AS-ODN1, showing little, if any mTIM immunostaining. C: Same lung as in A, immunostained with proliferating cell nuclear antigen (PCNA) antibody. D: Same lung tissue as in B stained for PCNA. Scale bar = 100 μm in D (applies to A–D).

Figure 7.

Morphology of embryonic lungs treated with mTIM antisense oligodeoxynucleotides (AS-ODNs; see Experimental Procedures section). A: An embryonic day (E) 11 lung before culture. B: E11 lungs treated with mTim AS-ODN1 for four days. C: E11 lung cultured with mTim sense oligodeoxynucleotide 1 (S-ODN1) for four days. D: E11 lung cultured in the absence of ODNs. The lower panel shows systematic quantification of branching activity in each of the experimental settings, Control, medium control; S, S-ODN1 (control); and AS-ODN1. Scale bar = 400 μm in D (applies to A–D).

Figure 8.

Antisense repression of mTIM does not affect surfactant protein C (SP-C) gene expression. Reverse transcriptase-polymerase chain reaction analysis of total RNA extracted from treated and control lungs in organ culture. cDNA from lungs cultured in presence of sense oligodeoxynucleotide 1 (S-ODN1; lane 1), 1.5 μM and 4.5 μM mTim antisense oligodeoxynucleotide 1 (AS-ODN1; lanes 2 and 3, respectively). Lane 4, cDNA from untreated E18 lungs (control) and H2O control. Neither SP-C nor Nkx2.1 mRNA levels was appreciably changed.

Timeless Is not Associated With Cell Proliferation

Previous characterization of another clock gene, Period (per), showed that its expression peaked at the same time of day as the peak in G1 phase of the cell cycle in human oral mucosal cells (Bjarnason et al., 2001). These data suggested a possible link between the circadian clock and the mammalian cell cycle. Because of its expression pattern during lung development, we also examined the possibility that mTIM may be involved in regulating the cell cycle in lung cells. Thus, we performed a flow cytometric assay by using fluorescein isothiocyanate (FITC) -conjugated antibody and propidium iodide (PI) staining in the lung carcinoma, A549 cells. Based on the DNA content, PI staining can separate cells at G1, M, and G2/S phases. We used Cyclin E as a specific marker of G1 phase, and an identical immuno-animal isotype IgG as negative control. As expected, the results showed that Cyclin E had a specific peak in the G1 phase of the cell cycle (not shown), but no differences were discernible for mTIM in G1, M, or G2/S phases (Fig. 9). These results suggest that expression of mTIM is independent of cell cycle progression in A549 cells. In support of the flow cytometric data, we also found no correspondence between the pattern of PCNA staining and mTIM on immunohistochemical analysis of serial sections of mouse embryonic lungs (Fig. 10). Therefore, the function of mTIM in lung cells does not appear to be related to regulation of cell cycle progression and proliferation.

Figure 9.

Flow cytometric analysis of mTIM in A549 cells. A: According to peak channel gated by propidium iodide (FL2-Height), A549 cells were divided into cell-cycle fractions of G1, S, and G2/M phases, showing that 65% of total cells were in G1 phase, 6% in S phase, and 19% in G2/M phase. B: Sorted G1, S, and G2/M phase A549 cells of (A) stained with primary antibody of mTIM and fluorescein isothiocyanate (FITC) -conjugated secondary antibody, respectively. mTIM (FL1-Height) -immunopositive staining is including in the right panel, and the FITC-conjugated isotype IgG control is in the left panel.

Figure 10.

Serial sections of postnatal day 1 lung stained for proliferating cell nuclear antigen (PCNA) and mTIM. A: Positive nuclear staining of PCNA. B: Hematoxylin counter staining of A, showing the cell nuclei that are negatively staining of PCNA. C: Positive nuclear staining of mTIM in the serial section of A. D: Hematoxylin counter staining of (C) showing the cell nuclei that are negatively staining of mTIM. Arrows point to saccular epithelial cells within the serial sections of the lung. Scale bar = 50 μm in D (applies to A–D).

DISCUSSION

The primary objective of the present study was to determine the distribution and function of the clock gene timeless during embryonic mouse lung morphogenesis. For this purpose, we generated a polyclonal antibody to the mouse Timeless protein and verified its specificity by several criteria. Immunohistochemical staining with the anti-mTIM antibody showed that mTIM is present throughout the early mouse embryo with primary expression localized to the neural epithelium (Fig. 1). In addition, however, TIM protein was found widely expressed in most organ rudiments, including brain, lung, heart, liver, stomach, and urogenital ridge (Fig. 2). These results are consistent with and extend previous observations derived from Northern blot and in situ hybridization analyses of rat TIM mRNA during early embryogenesis (Koike et al., 1998).

To determine the expression and distribution of mTIM during mouse embryonic lung morphogenesis, we first analyzed the profile of its mRNA in E15, E18, PN2, and adult lungs. Previously, Li et al. (2000) showed that two different size mTim mRNA species are detectable in the whole mouse embryo. The functional significance of the two transcripts remains unknown. Similarly, during lung morphogenesis, we found two mTim mRNAs of 2.4 and 4.4 kb that are dynamically expressed (Fig. 3). Northern blot analysis of mTim mRNA showed that both transcripts are most abundant in E15 lungs, a period coinciding with the highest branching activity during mouse lung morphogenesis. Also, in E18 lungs, both transcripts appear to be expressed, albeit at much lower levels (Fig. 3). Of interest, only a significantly weaker 2.4-kb mRNA was detectable in the adult lung.

Immunohistochemical analyses also revealed dynamic distribution of mTIM during the process of lung branching morphogenesis. It is important to note that detection and localization of mTIM in the nuclei of cells composing the various mouse tissues was conducted only after explicit determination of the specificity of the anti-mTIM antibody generated in this study. Thus, we used several important controls that included Western analysis of proteins, preabsorption of antibody by mTIM-specific peptides before use, and comparison between various fixatives to eliminate the possibility of fixative-based artifacts. The sum of the results obtained from the collection of the latter studies firmly established the specificity of the generated antibody for mTIM protein.

In early stages of lung development in E11 and E15 embryos, mTIM was localized to both epithelial and mesenchymal compartments (Fig. 3A,B). Subsequently, the intensity of mTIM staining in the mesenchyme progressively decreased as this compartment underwent condensation during later stages of lung development. By PN5, little if any staining for mTIM was discernible in the mesenchyme (Fig. 3E). Although we cannot comment on the precise functional significance of the changes observed in the expression of mTim mRNA, they are coincident with changes in the distribution of the mTIM protein, particularly the change from ubiquitous expression in both mesenchyme and the epithelium to increasing localization in the epithelial cells in E18 and postnatal lungs (Fig. 3C,D). Dynamic expression of timeless mRNA isoforms was also noted by Li et al. (2000), who found that the 4.4-kb transcript predominated in whole embryos, whereas the majority of adult tissues showed predominant expression of the smaller 2.4-kb transcript. This developmental isoform switch was also apparent in embryonic kidneys in which both isoforms were expressed in the embryo, whereas only the 2.4-kb transcript was detected in the adult (Li et al., 2000).

In the proximal lung, the tracheal and the bronchiolar epithelia were both strongly positive for mTIM in late development (E18 and postnatal samples). Localization of mTIM in nonciliated bronchiolar epithelial Clara cells was further substantiated by double immunocytochemical staining (Fig. 5). In the distal lung, mTIM was localized to cells with morphologic and phenotypic characteristics of alveolar type 2 cells (Fig. 3). To verify this finding, we demonstrated presence of mTIM immunoreactivity in isolated, purified alveolar type 2 cells (Fig. 4A). In addition to alveolar type 2 cells, the major lung cells in the distal lung epithelium, are alveolar type 1 cells, which form over 90% of the gas-exchange surface area. Studies in adult animals suggest that alveolar type 2 cells are the progenitors of alveolar type 1 cells and that, after injury, alveolar type 2 cells proliferate to differentiate into alveolar type 1 cells to restore the normal alveolar epithelium (Evans et al., 1975). Studies on alveolar type 1 cells have been hampered by technical difficulties to obtain purified populations of this critically important lung cell type. However, when maintained on inflexible substrata, alveolar type 2 cells rapidly lose their phenotypic hallmarks such as lamellar bodies and expression of surfactant apoproteins. The cells change morphologically to resemble alveolar type 1 cell in vivo, becoming flattened and developing long cytoplasmic processes and protuberant nuclei (Cheek et al., 1989). Concurrent with the loss of alveolar type 2 cell phenotypic characteristics, they increasingly express all known specific markers for in situ and freshly isolated type 1 cells examined to date (Borok et al., 2002). The latter include reactivity with the monoclonal antibody VIIIB2 (Danto et al., 1992), expression of caveolin-1 (Campbell et al., 1999), T1a (Borok et al., 1998a), and aquaporin 5 (Borok et al., 1998b). Together these findings have strongly suggested that type 2 cells in primary culture undergo transition toward the type 1 cell phenotype (type 1-like cells). Alveolar type 1-like cell monolayers, therefore, have been extensively used as a model with which to investigate the functional and biological properties of the alveolar epithelium. We, therefore, used this model to ascertain whether expression of mTIM occurs during in vitro transition of alveolar type 2 to alveolar type 1-like phenotype. Maintenance of purified alveolar type II cells in culture under well-controlled conditions that promote differentiation of alveolar type 1-like epithelial cells resulted in inhibition of mTIM as shown in Figure 4.

In the studies by Li et al. (2000) high-level expression of TIM was found in zones of active branching of ureteric bud (UB) and selective repression of mTim RNA expression in embryonic kidneys in culture showed a dramatic effect on kidney growth and branching morphogenesis. Likewise, the use of AS-ODNs in the current study inhibited lung branching morphogenesis in E11 mouse embryonic lungs. In comparison to the untreated and control S-ODNs–treated lungs, the AS-ODNs inhibited the production of mTIM as assayed by immunohistochemical analysis shown in Figure 6A,B. That inhibition through AS-ODNs strategy is highly selective for mTIM and not the consequence of a generalized inhibition of other nuclear proteins was verified by the demonstration that no differences exist in another nuclear protein, PCNA, between lungs treated with the AS-ODNs and control S-ODNs (Fig. 6C,D). In association with the decreased expression of mTIM in the AS-ODN–treated lungs, branching morphogenesis was inhibited by approximately 40 percent (Fig. 7). Inhibition occurred similarly when two independent AS-ODNs with different nucleotide sequence composition were used. In contrast, the use of control S-ODNs had no effect on branching morphogenesis, confirming the specificity of the AS-ODNs (Fig. 7). Of interest, inhibition of branching morphogenesis was not accompanied by changes in epithelial cell differentiation as assessed by SP-C gene expression, a marker for distal lung epithelial cell differentiation. Therefore, TIM does not appear to be involved in lung epithelial cell differentiation. These data also suggest that absence of TIM and differentiation of alveolar type 1-like cells in culture (Fig. 8), which is accompanied by loss of SP-C, are not causally related, but rather represent a convenient marker of alveolar type 1, compared with type 2 cells.

The mechanism by which inhibition of mTIM interferes with branching morphogenesis of lung and kidney are not understood. In Drosophila, two genes, per and tim, are essential in establishment and maintenance of circadian rhythm. per and tim mRNA levels cycle with the same period and phase, and a complex composed of PER and TIM regulates per and tim transcription through a negative loop. The PAS domain and CLD of PER are binding sites for TIM, so the encoded proteins PER and TIM physically interact (Saez and Young, 1996). Although most of the data regarding the potential role of PER and TIM are from the neural tissues, recent data have shown that both genes also oscillate in visceral organs (Zylka et al., 1998; Sangoram et al., 1998). However, the function of the latter remains unknown. Previous characterization of per showed its expression peaked at the same time of day as the peak in G1 phase in human oral mucosa (Bjarnason et al., 2001). These data suggested a possible link between the circadian clock and the mammalian cell cycle. We, therefore, examined the hypothesis that mTIM may also have a role in regulation of cellular proliferation during the process of branching morphogenesis. This hypothesis is rationally supported by our findings in the current study that differentiation of alveolar type I-like cells from isolated alveolar type 2 cells in culture (Borok et al., 1994) was coincident with loss of mTIM immunoreactivity (Fig. 4). Alveolar type 1 cells are considered to be terminally differentiated without the capacity to proliferate (Crapo et al., 1982). In addition, an essential component of lung branching morphogenesis is cellular proliferation in the absence or reduction of which it may be significantly slowed (Lindahl et al., 1997). However, a flow activated cell sorter analysis of TIM in the lung A549 cells demonstrated that TIM expression is independent of the cell cycle (Fig. 8). An important technical consideration pertains to the use of A549 cells to examine the relationship between cell proliferation and mTIM. We acknowledge that, although thought to be derived from adult “Clara” or alveolar type II cells, A549 cells are distinct from normal, and in this case embryonic pulmonary epithelial cells. Nevertheless, numerous published studies have been conducted in A549 cells for molecular, cellular, and biochemical analyses in the lung (Bui et al., 1993; Minemoto et al., 2003). In the current study, we provide evidence that, in A549 cells, mTIM does not correlate with the various phases of the cell cycle. In addition, we show that PCNA immunostaining, a reliable marker of proliferation also does not correlate with expression of mTIM in embryonic lung tissue (Fig. 10). Therefore, these data support the concept that expression and function of mTIM may not be related to cell proliferation. Finally, it is also important to point our that a recent finding indicates that mTIM may not represent the true mammalian homolog of Drosophila Timeless (Shearman et al., 2000) and that a more likely ortholog may be a newly identified fly gene designated as timeout (GenBank accession no. AE003698 and AE003699). The function of timeout remains unknown.

Branching morphogenesis is considered the most active period of epithelial–mesenchymal interactions. Analysis of lung and kidney branching morphogenesis have revealed that a small number of signaling molecules such as FGF, EGF, sonic hedgehog, TGF-beta, and Wnt may be involved (Hogan, 1999). Several transcription factors including NKX2.1, Gli, Smad are also necessary for normal branching morphogenesis and cellular differentiation within each organ. The studies presented in the current report show that mTIM participates in branching morphogenesis of the lung primordium. Combined with the findings of Li et al. (2000) in embryonic kidney development, these data suggest that TIM functionally contributes to a pathway that is highly conserved and, therefore, fundamental to development of both organs.

EXPERIMENTAL PROCEDURES

Generation of a Polyclonal Anti-Mouse Timeless Antibody

To generate a polyclonal antibody to mTIM protein, we used three synthetic peptides as antigen. The three peptides were chosen based on the translation of the mTim cDNA sequence (Takumi et al., 1999). The first 17-amino acid peptide has the sequence KEPDCLESVKDLIRYLR, which corresponds to the amino acid residues 27 to 43 found on the mTIM protein. The second peptide was derived from the corresponding amino acids 487 to 521 of mTIM and has the sequence RKFDERYHPRSFLRDLVE. This peptide is part of the highly conserved Drosophila dPER binding domain (Saez and Young, 1996; Sangoram et al., 1998). The third peptide had the amino acid sequence CGTPRVHRKKRFQIEDEDD, corresponding to amino acids 1179 to 1197 of the mTIM protein and is located in the carboxyl terminus of mTIM, which includes tetrapeptides of dTim CLD-DEDD. The polyclonal antibody was raised in rabbits and purified by affinity chromatography using all three antigenic peptides in combination.

Cell Isolation and Culture

The conditions for growth of H441 and A549 cells have been previously described (Li et al., 2002). The HeLa and the mouse 3T3 cells were grown in DMEM supplemented with 10% fetal bovine serum. Isolation of fresh alveolar type 2 cells was according to methods described previously (Dobbs et al., 1980). Alveolar type 2 cells were cultured and allowed to undergo differentiation into alveolar type 1-like cells according to methods described by Borok et al. (1995). The cells examined were derived from two independent sets of cell-isolation experiments.

Embryonic Lung Explant Culture

Whole embryonic lungs were dissected from E11 embryos under microscope and kept in autoclaved PBS on ice before ex vivo culturing. Dependent on the experiment, approximately 8 to 10 E11 embryos were isolated from each pregnant mouse. In each Grobstein Falcon dish, four to six lungs in duplicates were placed on filters (Millipore, Bedford, MA) that were in turn placed atop a stainless steel grid that was placed on top of the growth medium. In this way, the lungs are exposed to air on the top and to the growth medium on the bottom. The Falcon plates were kept under optimal humidity and maintained in 95% air and 5% CO2 for 4 days. The medium consisted of BGJb (GIBCO) supplemented with 50 units penicillin/streptomycin and changed every 2 days. Antisense and sense oligonucleotides were added to the culture medium at concentrations of 1.5 μM and 4.5 μM. Each experiment was repeated at least three or more times to minimize variability.

Oligodeoxynucleotides (ODNs) were designed as previously described (Minoo et al., 1995; Li et al., 2000). The sequence of the ODNs corresponded to the following domains of mouse Timeless mRNA (GenBank accession no. AB019001). Two AS-ODNs were designed as follows: AS-ODN1 had the sequence 5′-cac agt tca tca tgt aca agt cca t-3′, which is complementary to the region between nucleotides 176 and 152 in the mTim cDNA. The sequence of AS-ODN2 was 5′-aag ggcgct aca cgt ggc tag aag-3′, which is complementary to the region between nucleotides 202 and 179 in the mTim cDNA. Two sense oligodeoxynucleotides (S-ODNs) were used as control. S-ODN1 had the sequence 5′-atg gac ttg tac atg atg aac tgt g-3′, which corresponds to the nucleotides 152 to 176. S-ODN2 had the sequence 5′-cttcta gcc acg tgt agc gcc ctt-3′, which corresponds to nucleotides 179 to 202 in the mTim cDNA. To increase the lifetime and stability of the ODNs, three bases on the 5′ and 3′ ends of each ODN were thiosulfated. At completion of each study, lungs were photographed under a dissection microscope and the terminal branches were quantified by careful measurement that was repeated independently at least three times. The lungs were subsequently either flash frozen in liquid nitrogen for RNA extraction, or fixed in Carnoy's solution and paraffin embedded for histologic analyses. Antibodies to mouse Timeless and PCNA (Zymed Laboratories, South San Francisco, CA) were used for immunohistochemical analyses.

Preparation of Tissue and Cell Lines for Immunohistochemistry

To examine the distribution of mTIM protein during embryonic development, pregnant Swiss Webster mice were killed at different stages of gestation. The E10 whole embryos were fixed in 4% paraformaldehyde for 4 hours at 4°C. The embryonic lungs at various stages of development were quickly dissected and fixed in Carnoy's fixative, largely by previously described methods (Yuan et al., 2000). For postnatal and adult lungs, the killed mice were quickly dissected and the lungs were perfused with Carnoy's fixative by placing a catheter in the trachea. The lungs were then submerged in Carnoy's fixative and incubated overnight. Subsequent to fixation, all tissues were dehydrated through a series of ethanol solutions, and embedded in paraffin. Four-micron sections were prepared by the use of microtome for immunohistochemistry.

For preparation of cells, cultured alveolar epithelial cells, A549, H441, NIH 3T3, or HeLa cells were grown on Lab-Tek chamber slides. Freshly isolated alveolar type 2 cells, ranging from 0.5 × 105 to 1.5 × 105 cells in 100-μl volume in phosphate buffered saline (PBS) were cytospun onto slides at 500 RPM for 5 min. The cells were subsequently fixed in situ with 80% ethanol/PBS before routine immunohistochemical staining.

Immunohistochemistry

The antibodies used in immunohistochemistry were obtained from the following sources. The rabbit anti-CCSP was a generous gift of Dr. Franco Demayo, Baylor College of Medicine, Houston, TX. The mouse anti–α smooth muscle actin (α-SMA) was purchased from Sigma-Aldrich, St. Louis, MO. The anti-PCNA antibody was purchased from Zymed Laboratories. A commercially available rabbit immunohistochemistry kit (Zymed Laboratories) was used for immunohistochemistry. The tissue samples prepared for immunohistochemical analysis of mTIM were hydrated and treated with 1% H2O2 in methanol for 20 min and blocked with 10% normal goat serum, sections were incubated with mTIM antibody at 4°C overnight, then biotinylated secondary antibody and streptavidin–peroxidase conjugate were added and the slides were incubated for 10 min. Diaminobenzidine (DAB) substrate was used for color development. For determining the specificity of the anti-mTIM antibody, the antibody was incubated with the three peptides, used in generation of the antibody, in a 1:1 ratio, in PBS at 4°C overnight. The activity of the adsorbed antibody was then compared with no-adsorbed anti-mTIM antibody.

To detect cell specificity of mTIM in the lung, we used double-immunostaining. ZYMED LAB-SA kits (Zymed Laboratories) was used for this purpose. Rabbit anti-CCSP (generous gift of Dr. Franco Demayo, Baylor College of Medicine, Houston, TX) and mouse anti–α-SMA (Sigma-Aldrich) were used as the first antibodies. Secondary streptavidin-alkaline phosphatase–conjugated antibody was used with NBT/BCIP color (blue) development for CCSP and α-SMA staining. The same slides were next incubated with mTIM antibodies and subsequently treated with streptavidin–horseradish peroxidase secondary antibody. This procedure gave brown-color staining with DAB substrate.

Flow Cytometry

Proliferating A549 cells at low density were detached by trypsinization, and pelleted by centrifugation. The cells were washed once in growth medium containing serum and collected by centrifugation. The cell pellet was then resuspended in fresh PBS at ∼106 cells/ml. For fixation, approximately 1 ml of the cell suspension was transferred to a 15-ml tube containing 10 ml of methanol and incubated at 4°C for 4 hr. The cells were permeabilized in 0.25% Triton X-100/PBS on ice for 5 min and immunostained with antibodies to mTIM and Cyclin E (Pharmingen, San Diego, CA) as well as their nonimmune antibody controls at 4°C overnight. Subsequently, the cells were incubated for 30 min in FITC-conjugated goat anti-mouse IgG or anti-rabbit IgG in the dark at room temperature. After rinsing, the cells were centrifuged and resuspended in PI and incubated for 20 min in the dark at room temperature before measurement. A flow cytometry equipped for excitation with blue light (480-nm laser line) was set up for fluorescence detection and data collection. A 530- ± 20-nm bandpass filter was used for detection of antibody-associated green fluorescence of FITC, and a 620-nm long-pass filter was used for measuring DNA-associated red florescence of PI.

Western Blot Analysis

A549, H441, HeLa, and NIH3T3 cells were used for Western blot analyses. To prepare nuclear and cytoplasmic extracts, the cells were lysed in a buffer containing 10 mM HEPES, 1.5 mM MgCl2, 10 mM KCl, 0.1 mM ethylenediaminetetraacetic acid, 1 mM dithiothreitol, 0.2% (v/v) NP-40, and 0.5 mM phenylmethyl sulfonyl fluoride, mixed freshly, immediately before use. The cell lysate was then centrifuged at 3,000 RPM, and the supernatant was collected as cytoplasmic extract. The pelleted nuclei were washed once in the same lysis buffer and nuclear proteins were extracted by incubation in 400 mM NaCl in PBS buffer, according to established procedures (Bohinski et al., 1994). The protein samples were quantified with BCA Protein Assay Reagents (Pierce, Rockford, IL), then separated on 8.5% polyacrylamide gels and electrotransferred onto Immobilon-P transfer membranes (Millipore, Bedford, MA). After washing with TBS containing 5% milk and 0.1% Tween-20 several times at room temperature, the membranes were incubated with mTIM antibody overnight at 4°C. The filters were then incubated with an anti-rabbit secondary antibody conjugated to horseradish peroxidase (Pierce, Rockford, IL). The immunoreactive bands were developed by using ECL detection (Amersham, Buckinghamshire, UK).

Northern Blot Analysis

Poly (A)+ RNA from mouse E15, E18, postnatal day 2 (PN2), and adult lungs were isolated with a PolyATtract mRNA Isolation System III kit according to the manufacturer's protocol (Promega, Inc., Madison, WI). Northern blot was performed as described previously (Li et al., 2002b). In brief, 2 μg of poly (A)+ RNA was electrophoresed in 1% RNA formaldehyde agarose gel and blotted onto Nylon transfer membrane. Blots were hybridized with probe specific for mouse Timeless, and then autoradiographed. The probe was synthesized from a 0.8-kb fragment of the 3′ UTR of mTim cDNA (generous gift of Dr. Steven M. Reppert and Dr. Anthony L. Gotter, Massachusetts General Hospital, Boston, MA) according to previous publication.

Acknowledgements

We thank Dr. Qiuping Pan for her technical assistance.

Ancillary