SEARCH

SEARCH BY CITATION

Keywords:

  • branching morphogenesis;
  • cell shape change;
  • live imaging;
  • salivary gland;
  • time-lapse;
  • ultrastructure

Abstract

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

We cultured the rudimental submandibular gland (SMG) of mice with a non–cell-permeable fluorescent tracer, and observed cell behavior during epithelial branching morphogenesis using confocal time-lapse microscopy. We traced movements of individual cells as shadowgraph movies. Individual epithelial cells migrated dynamically but erratically. The epithelial cleft extended by wiggling and separated a cluster of cells into two buds during branching. We examined the ultrastructure of the clefts in SMG rudiments treated with the laminin peptide A5G77f, which induces epithelial clefting. A short cytoplasmic shelf with a core of microfilaments was found at the deep end of the cleft. We propose that epithelial clefting involves a dynamic movement of cells at the base of the cleft, and the formation of a shelf within a cleft cell. The shelf might form a matrix attachment point at the base of the cleft with a core of microfilaments driving cleft elongation. Developmental Dynamics 239:1739–1747, 2010. © 2010 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

Epithelial branching morphogenesis is a fundamental process that generates complex three-dimensional acinotubular structures in various exocrine glands. Branching requires dynamic epithelial reshaping, such as invagination or bending. At a cellular level, these processes must include dynamic shape changes and repositioning of individual cells. However, current understanding of branching is based mainly on analyses of fixed specimens of branching tissue at various developmental stages or under various experimental conditions. Therefore, there are considerable limitations in understanding the dynamic sequence of cellular events during branching.

The developing submandibular gland (SMG) is a well-known system for studying the mechanisms underlying branching morphogenesis (Kadoya and Yamashina,2005; Patel et al.,2006). On embryonic day (E) 12, the mouse SMG consists of an epithelial terminal cluster that connects to the distal end of an epithelial stalk. The branching starts with the formation of a cleft at the basal surface of the terminal cluster. Subsequent cleft elongation separates the cluster into hemispheres. Repeated epithelial branching results in an extensive arborizing duct system that terminates with many terminal end buds. Detailed observations by Nogawa (1983) showed that the epithelial clefts appeared as shallow indentations along the lobular surface of developing SMG explants, and that some of these indentions became stable clefts, whereas others disappeared. However, this observation was made by conventional light microscopy; hence, the detailed cellular events of this early step largely remain to be described. More recently, Larsen et al. (2006) succeeded in labeling individual SMG epithelial cells with an adenovirus carrying the gene for green fluorescent protein (ad-GFP) and revealed dynamic epithelial cell movements during SMG branching in time-lapse movies. However, that study was not able to elucidate the collective cell movements involved in branching, because the ad-GFP labeled only a subset of epithelial cells.

To depict cellular behavior and structural dynamics during epithelial clefting and branching, here we established a simple method to trace shape changes and migrations of living SMG cells using microscopy-based movies with a reasonable spatiotemporal resolution. The epithelial cleft appeared as a narrow gap between neighboring cells. In the movies, the cleft elongations were dynamic but their modes varied individually. Moreover, the deepest portion of the cleft was seen wriggling dynamically. However, none of the previously proposed mechanisms (Wessells et al.,1971; Bernfield and Banerjee,1972; Hieda and Nakanishi,1997; Alberts et al.,2008) for epithelial branching was readily applicable to explain these morphogenetic features.

We have documented that many clefts are induced within SMG explants when they are treated with the laminin peptide A5G77 or its active core sequence A5G77f (Kadoya et al.,2003). Therefore, we used this system and reexamined the epithelial clefts of A5G77f-treated SMG rudiments by electron microscopy. A short cytoplasmic projection with a core of microfilaments was frequently found at the deepest end of the individual cleft. The distal tip of a cleft was found invading the groove formed by the projection and sidewall of the cleft. Indeed, a similar structure was seen in the normal SMG rudiment in the absence of peptide treatment. Based on these findings, we propose a possible structural mechanism for cleft elongation during SMG branching.

RESULTS

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

Visualization of Living Cells Comprising the Rudimental SMG

The E13 SMG rudiments or fragments repeated typical branching morphogenesis on the glass-bottomed culture dish over 3 days (Fig. 1), which process was essentially the same as that observed in conventional filter top organ culture (Kadoya et al.,2003). When explants were incubated with a medium containing a non–cell-permeable fluorescent tracer, sulforhodamine B (SRB; 2.5 μM) and observed using confocal microscopy, SRB outlined the contours of individual cells within the rudimentary SMG (Fig. 2). The terminal cluster was seen to be composed of an outer layer of closely packed columnar cells (OCCs) and a central core of polymorphic epithelial cells (CPCs). Entire cellular contours of individual OCCs and CPCs were clearly traced and no luminal space could be seen. In contrast, the duct wall was seen to be composed of simple or pseudostratified columnar cells. SRB failed to penetrate into the ductal lumen. The contours of mesenchymal cells could be traced as well but their boundaries were vague.

thumbnail image

Figure 1. Branching morphogenesis of the mouse submandibular gland (SMG) rudiment on a glass-bottomed culture dish. Embryonic day (E) 13 SMG rudiments were placed on the glass-bottomed culture dish, covered with a small amount of culture medium, and cultured for 2 days. A: At the onset of organ culture. B: SMG explants after 2 days in culture. The epithelial terminal cluster (TC) of E13 SMG branched repeatedly and generated end buds (EB) and duct system (D). Scale bar = 0.5 mm.

Download figure to PowerPoint

thumbnail image

Figure 2. Shadowgraphs of living submandibular gland (SMG) cells as revealed by a fluorescent tracer in the culture medium. Embryonic day (E) 13 SMG explants were cultured with medium containing sulforhodamine B and observed using confocal microscopy. A: The epithelial terminal cluster is composed of an outer layer of closely packed columnar cells (OCCs, arrowheads) and a central core of polymorphic cells (CPCs). B: The duct is composed of simple or pseudostratified columnar epithelium. The tracer failed to penetrate into the lumen (L). Scale bar = 50 μm.

Download figure to PowerPoint

In confocal time-lapse movies of the living SMG rudiments, we found dynamic but distinct cell migrations and shape changes within the explants. In particular, both OCCs and CPCs changed their shapes quickly and we needed to acquire images every 75 sec for tracing the movement of individual cells accurately (Supp. Movie. S1, which is available online). In contrast, the duct cells seldom moved (Supp. Movie S2), indicating that the above dynamic cell movement was a distinct feature of the terminal cluster epithelium and not an artifact of the time-lapse microscopy such as focus drift. Although mesenchymal cells changed their shape substantially, they tended to stay in the same positions compared with the terminal cluster epithelial cells. Mitotic epithelial cells were seen frequently (Supp. Movie S1, marked M); however, we failed to see any direct association between the cell proliferation and clefting during the tracing periods in the present study.

Dynamics of Cleft Elongation

As shown in Figure 3 and Supp. Movie S3 (arrows), the clefting was quite dynamic and individual clefts extended or retracted in distinct modes. Shallow indentions (less than 1 μm deep) appeared occasionally along the basal surface of the terminal cluster. However, they were unstable and soon retracted. The dynamic processes of cleft elongation became more evident when clefts were observed at a higher magnification (Fig. 4; Supp. Movie S4). The distal end of the cleft wriggled dynamically and behaved as if it was ‘searching’ for the direction of invasion. These results suggested that dynamic and site-directed epithelial cell shape changes are involved directly in cleft elongation. A mesenchymal cell process sometimes appeared within the cleft but was distant from the distal tip of the invading cleft (Supp. Movie S1, arrow), suggesting a role for mesenchyme in stabilizing the cleft.

thumbnail image

Figure 3. Modes of cleft elongation. The lengths of clefts were measured at 12.5-min intervals (every 10 frames) for 4 hr and 10 min, and plotted. The elongation rate of the individual cleft was unsteady and changed during the tracing. Cleft #1 (triangle) stretched rather uniformly during the tracing period. Although there was a subtle expansion and contraction in length, cleft #2 (circle) stayed at almost the same depth during the tracing period. Clefts #3 (diamond) and #4 (square) stretched slowly (4–5 μm/hr) for 1–2 hr and then suddenly grew rapidly (25 μm/hr). These rapid expansions ceased to occur after half an hour, and the cleft growth returned to an initial slower rate or essentially halted.

Download figure to PowerPoint

thumbnail image

Figure 4. Dynamics of cleft elongation. An embryonic day (E) 13 submandibular gland (SMG) was cultured for 2 days and then examined by confocal time-lapse microscopy. Frames from every 2.5 min from a time-lapse movie (Supp. Movie S4) are shown. Note that the distal portion of the cleft (arrow) is wriggling dynamically.

Download figure to PowerPoint

Epithelial Cell Rearrangement During Cleft Elongation

Individual OCCs, CPCs, and mesenchymal cells were marked, and their migrations were traced in the movies. Although cell migrations in a rudiment were three-dimensional, we first analyzed the vertical migrations of individual cell populations by comparing their retention periods in a fixed optical plane. We found that the CPCs could be traced for longer than the OCCs (Fig. 5), indicating that the latter cells migrated more actively. The OCCs that had faced the cleft initially (Fig. 5, cells marked +1 and −1) tended to remain in the focal planes longer than the other OCCs. Mesenchymal cells were less motile.

thumbnail image

Figure 5. Retention of individual cell populations in a fixed optical plane. outer layer of closely packed columnar cells (OCCs) have been identified with Arabic numerals. Note that numbers indicate the distance from the cleft tip (arrow), so cells were numbered +1, +2, +3, … and −1, −2, −3, etc. Central core of polymorphic epithelial cells (CPCs) and mesenchymal cells have been marked with lower- and uppercase letters, respectively. We traced each cell in the movies. Some cells could be traced during an entire observation period (20 sec), while others disappeared earlier. Based on these data, we constructed a bar graph showing the cell retentions. Three independent tracings provided similar results.

Download figure to PowerPoint

Because the behavior of individual cells during cleft formation is largely unknown, we then examined the cell movements associated with cleft elongation. As shown in Figure 6, OCCs composing the lateral wall and/or the distal region of the cleft were rearranged considerably during cleft elongation. Moreover, some CPCs, which had been located near the bottom of a cleft but had not faced it (Fig. 6, yellow- and cyan-outlined cells), moved and integrated into a sidewall of the growing cleft.

thumbnail image

Figure 6. Movie (Supp. Movie S3) frames showing cell migration and shape changes during cleft elongation. Contours of five epithelial cells (shown in red, green, blue, cyan, and yellow) have been traced every 10 frames (12.5-min interval) for 212.5 min. The yellow- and cyan-outlined cells moved and made contact with the basement membrane. In contrast, the blue-outlined cell migrated but stayed at an area separate from the basement membrane. The green- and red-outlined cells changed their shapes but maintained contact with the basement membrane.

Download figure to PowerPoint

These results indicated that both OCCs and CPCs move nondirectionally or randomly and that OCCs migrate more actively than CPCs.

Transmission Electron Microscopy (TEM) of the Cleft in the Branching Salivary Epithelium

To elucidate the mechanical aspects of cleft elongation, we needed to learn more about their structure. However, systemic ultrastructural analysis of clefts has been greatly limited to date because there are too few clefts along the basal surface of the terminal cluster for transmission electron microscopy (TEM) studies. The laminin peptide A5G77f (LVLFLNHGH using the single letter amino acid code) has been found to induce many clefts along the contours of the terminal clusters in SMG rudiments in organ culture (Kadoya et al.,2003). Here we found that most, if not all, of these induced clefts retracted when the rudiments were further cultured for an additional day with medium containing 10% fetal calf serum (FCS) but not the laminin peptide (data not shown). These results suggested that the induced clefts are unstable, which is a feature of clefts at an early stage of formation (Nogawa,1983). Therefore, we examined the ultrastructure of clefts induced with A5G77f.

Figure 7 shows typical peptide-induced clefts of various lengths. The clefts invaded as a narrow fissure (∼0.3 μm wide) between adjacent cells. Occasionally, clefts were seen extending for greater than 10 μm. Epithelial cells facing the cleft were supported with a continuous basement membrane. Interstitial collagen fibers were frequently seen at the proximal or opening area of the clefts, but only faintly at their deeper or distal portions. At the deepest ends of the clefts, short cytoplasmic projections (Fig. 7B,C, arrows) were frequently seen protruding from either side of the epithelial walls surrounding the distal end of the clefts. A similar projection was also seen in very shallow cleft (Fig. 7A, dotted rectangle). Microfilaments were found to form the core of the projection. Cross-sectional views of the projection were rare, suggesting that it is not a finger-like protrusion but a narrow ridge. Moreover, it is likely that the ridges from the cells that compose the sidewall of a cleft are connected laterally and form a slab-like cytoplasmic extension, which is referred to as a shelf hereafter. As shown in Figure 8, shelves were noted in 73% of clefts (n = 70). The distal end of the cleft invaded the groove formed by the shelf and sidewall of the cleft. The electron density of the plasma membrane was slightly increased at the root of the shelf. Some anchoring filaments connecting the plasma membrane to the lamina densa of the basement membrane were also noted at the shelf, indicating that shelf might form an attachment point to the basement membrane. However, we were unable to see any structures resembling known subcellular structures such as hemidesmosomes. Although we failed to see shelves in 27% of cases, this might have been because we chose inappropriate cutting angles for the ultrathin sections. Shelves were also noted in the deep ends of epithelial clefts of intact E13 SMG and E13 SMG explant cultured in normal medium lacking the A5G77f peptide (Fig. 9).

thumbnail image

Figure 7. Electron micrographs showing various steps of the cleft formation in the submandibular gland (SMG) rudiments that were cultured with A5G77f laminin peptide for 3 days. A–C: Shelves are evident in a very shallow cleft (A, dotted rectangle) and at the bottom of clefts of various lengths (B,C, arrows). Note that actin filaments form the core of a shelf and that basement membranes line the cleft contours. To enhance the electron density of collagenous fibrils, specimens were stained en bloc with uranyl acetate (A,B) or tannic acid (C). However, collagenous fibrils are only faintly visible at the deep end of the clefts. Scale bars = 1 μm.

Download figure to PowerPoint

thumbnail image

Figure 8. Occurrence of shelves in clefts of various lengths. Embryonic day (E) 13 submandibular gland (SMG) rudiments were cultured with the A5G77f peptide for 3 days and were analyzed by the transmission electron microscopy. Seventy clefts were chosen randomly and examined for their lengths and shelf presence. Open and closed bars indicate clefts with and without a shelf, respectively.

Download figure to PowerPoint

thumbnail image

Figure 9. A,B: Electron micrographs showing clefts of normal embryonic day (E) 13 submandibular gland (SMG) (A) and an SMG explant that was cultured for 1 day without laminin peptide (B). Arrows show shelves. Scale bars = 0.5 μm.

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

These time-lapse movies clearly revealed dynamic cell movements during epithelial branching in the developing SMG. We also found that the epithelial cleft first appeared as a narrow fissure and extended between neighboring epithelial cells. Although a role of interstitial collagen fibers was previously suggested for cleft elongation in the SMG epithelium (Nogawa and Nakanishi,1987; Hieda and Nakanishi,1997), our observations showed that the mesenchymal components were distributed only barely at the distal portion of the cleft, where active cleft wiggling was noted. Interestingly, the elongation speeds of the cleft (4–25 μm/hr) were in the range similar to those recorded for the elongation of stress fibers (Hotulainen and Lappalainen,2006) but considerably slower than those for growth or shortening of microtubules in cultured cells (Shelden and Wadsworth,1993). We thus conclude that the epithelial cells essentially drive cleft elongation. Involvement of the epithelial cytoskeletal system, in particular actin filaments, is also likely.

Spooner and Wessells (1970) suggested that the contraction of basal actin bundles running from one cell to another along the basal surface might help the epithelial sheet to become convex. This model was further modified by Bernfield and Bernerjee (1972) and has been quoted in a recent textbook (Alberts et al.,2008). However, these models seem less applicable for the branching of the rudimental SMG epithelium, because the cleft is very thin and no apparent basal actin bundles that could support the model have been noted in the intact rudimental SMG epithelium undergoing extensive branching (Kadoya and Yamashina,1991,1993; Walker et al.,2008).

Here, we propose that the shelf is a structure that plays a role in the cleft elongation. It is a cytoplasmic ridge with a core of actin filaments. The distal tip of a cleft is covered with the basement membrane, invading the groove formed by the shelf and sidewall of the cleft. Thus, it is reasonable to believe that the shelf forms an attachment point to the basement membrane within an individual epithelial cell at the base of the cleft. It is also likely that the core of microfilaments generates a mechanical force that drives cleft extension (Fig. 10). Daley et al. (2009) have recently shown that epithelial cleft initiation and elongation are distinct processes, each of which is controlled by the actomyosin system differently. Because the shelf was noted in both the elongated and shallow clefts, it is very likely that the shelf is involved not only in the cleft elongation but also in its initiation. Although these authors proposed that cell proliferation would be required for cleft elongation, our movies failed to show its direct association to mitosis. The shelves were not mentioned in previous TEM studies (Coughlin,1975; Spooner and Faubion,1980; Fukuda et al.,1988). This might be because those authors mainly focused on the structures of well-established wider clefts, which might be stabilized already by mesenchymal cells and extracellular matrix, including fibronectin or collagen III (Nakanishi et al.,1988; Sakai et al.,2003).

thumbnail image

Figure 10. Schematic representation of cleft elongation during epithelial branching. The distal tip of a cleft invades a groove formed by the shelf and sidewall of the cleft. Retraction of the groove (green open arrows) pulls the tip of the cleft inward. Microfilaments associated with shelves are shown in blue lines. BM, basement membrane.

Download figure to PowerPoint

To induce sufficient numbers of clefts for TEM examinations, we used the A5G77f peptide, an active core sequence within the laminin peptide A5G77 (Kadoya et al.,2003). Laminin peptides are biologically active molecules screened by cell binding assays from libraries of short synthetic peptides derived from the sequences of laminins, a family of trimeric glycoproteins found in the basement membrane. The A5G77 sequence is derived from the EF-loop of the LG4 module of the laminin-α5 chain (Suzuki et al.,2003), one of the laminin-α chain isoforms that are expressed in the basement membrane of developing SMG epithelium (Kadoya et al.,2003), and is involved in the heparin/heparan sulfate binding to the laminin-α5 chain (Hozumi et al.,2009). Thus it is possible that the epithelial cell attachment to laminin-α5 chain through a receptor containing heparan sulfate is involved in the cleft formation. Epithelial cleft formation in the SMG is also controlled by the epidermal growth factor (EGF) and the fibroblast growth factor signaling pathways (Nogawa and Takahashi,1991; Kashimata and Gresik,1997; Patel et al.,2006). A recent study suggested that activation of the EGF receptor might stimulate epithelial cleft formation through the extracellular signal-regulated kinase 1/2 (ERK1/2) and phosphatidylinositol-3 kinase (PI3K) pathways in the developing SMG (Koyama et al.,2008). The mode of cleft induction by EGF in the rudiment differed morphologically from that seen in rudiments treated with A5G77f (Kadoya et al.,2003), suggesting the presence of signaling systems other than ERK1/2 or PI3K for regulating cleft formation.

Compared with OCCs, CPCs tended to remain within a fixed optical plane in the present study, indicating that the latter are less motile. This finding is in good agreement with that reported by Larsen et al. (2006), in which rudimental SMG epithelial cells started moving faster when they contacted the basement membrane. Although our findings suggest that cells facing the distal part of the cleft were less migratory, we were unable to reveal direct coordination between these distinct cell rearrangements and cleft elongation. Because the migration of OCCs is active and nondirected, individual cells comprising the cleft wall must have been rearranged gradually during cleft elongation. Moreover, some CPCs located near the bottom of a cleft, but not facing it, moved and integrated into a sidewall of the growing cleft. These findings suggest that there is no fixed group of cells that act to pull the cleft inward. Hence, it is likely that the pulling activity is communicated from one cell to its neighbor during cleft elongation. What is the mechanism for this communication? How do the adjoining cells cooperate to drive the cleft? These are challenging questions to be resolved.

The mechanisms driving cell rearrangements in epithelia have been extensively studied in Drosophila. Dynamic epithelial cell rearrangements, which required the modification of adhesive junctions between neighboring cells, were noted during germ band elongation in the Drosophila embryo (Butler et al.,2009). Furthermore, the remodeling of junctions was found to depend on local forces acting at the cell boundary generated by junction-associated myosin II (Bertet et al.,2004). We found that a specific myosin II inhibitor, blebbistatin, perturbed epithelial cell movements in the SMG rudiments and impaired epithelial branching morphogenesis (data not shown), suggesting the involvement of the actomyosin system for the dynamic and motile nature of epithelial cells comprising the developing SMG (Daley et al.,2009).

Our present method is based on confocal microscopy combined with a non–cell-permeable fluorescent tracer technique (Segawa,1999). The previous live cell imaging studies of developing organ explants or organ equivalents—including salivary gland (Larsen et al.,2006), mammary gland (Ewald et al.,2008), and kidney (Watanabe and Costantini,2004)—required tagging specific cells with GFP, which needs to be incorporated into cells either transgenically or through a viral vector. Our method does not require cell labeling with GFP. Instead, SRB, which just needs to be added to the culture medium, enables us to outline the shape of entire cells within the SMG rudiments. Hence, this simple procedure will be widely applicable for studies on cell dynamics in a variety of organs and tissues.

EXPERIMENTAL PROCEDURES

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

Organ Culture

Dated-pregnant female mice of the out-bred ICR strain were purchased (Oriental Yeast or CLEA Japan, both in Tokyo, Japan). The morning of the discovery of a vaginal plug was designated embryonic day (E) 0. Pregnant mice were euthanized by cervical dislocation, and embryonic SMGs were harvested from E13 embryos. Some rudiments were cut into three to five pieces, each containing one to three terminal clusters. All protocols involving animals were approved by the Animal Experimentation and Ethics Committee, Kitasato University School of Medicine and by the Committee for Animal Experiments, Kitasato University School of Allied Health Sciences.

For confocal time-lapse microscopy, SMG rudiments or their fragments were placed on glass-bottomed culture dishes (Iwaki, Asahi Techno Glass, Chiba, Japan), which were precoated with 2.5% growth factor-reduced Matrigel (BD Biosciences, San Jose, CA), and cultured with a small amount of Dulbecco's modified Eagle's medium/Nutrient mixture F12 (DME/F12; Invitrogen, Life Technologies Japan, Tokyo, Japan) supplemented with 10% FCS and 50 μg/ml gentamicin. The rudiments were then incubated in humid conditions (37°C, 5% CO2 in air) for 1–2 days before the confocal time-lapse microscopy. The medium was changed twice a day.

Some rudiments were cultured in a conventional filter-top system with serum-free DMEM/F12 as described (Kadoya et al.,2003). To stimulate clefting, the laminin-α5 chain-derived peptide A5G77f (LVLFLNHGH, a gift from Dr. M. Nomizu, Tokyo University of Pharmacy and Life Science) was added to the serum-free culture medium at a final concentration of 0.19 mM at the onset of organ culture (Kadoya et al.,2003). After three days without any medium change, the peptide-treated-explants were fixed for electron microscopy. For recovery experiments, some explants were transferred to fresh DMEM/F12 with 10% FCS but no A5G77f and cultured for an additional day.

Confocal Time-Lapse Microscopy and Image Processing

Sulforhodamine B (SRB; Molecular Probes, Tokyo, Japan), which has been applied to visualize secretion granules in conjunction with two-photon microscopy (Kasai et al.,2005), was dissolved in culture medium at 2.5 μM. Culture medium was replaced with SRB-containing medium at the beginning of time-lapse microscopy, which was performed using an MRC1024 or Radiance 2100 scan head (Bio-Rad, Tokyo, Japan) mounted on an inverted microscope (Nikon, Tokyo, Japan). An environmental chamber (Tokai Hit, Fujinomiya, Japan) was used to maintain the temperature at 37°C and to provide an atmosphere of humid 5% CO2 in air. Images were acquired using an oil immersion Plan Apo 60 NA 1.40 objective. For some experiments, images were magnified by zooming. Acquisition settings were optimized to obtain maximum resolution of individual cells while minimizing photo damage. Images were captured every 75 sec unless stated otherwise. The fluorochrome was excited using 514 nm (Ar laser) or 543 nm (Green He-Ne laser) wavelengths. Sequential confocal images were imported into Adobe Photoshop software (CS3, Adobe, San Jose, CA), adjusted for brightness and contrast, and assembled into a Quick Time (QT) format Movie (6 frames/sec). The movie was further compressed in Quick Time Pro 7.6 (Apple, Cupertino, CA) using H.264 codec.

Analysis of Cleft Elongation and Cell Migration

Depths of clefts were measured at 10-frame (12 min 30 sec) intervals for 4 hr. ImageJ software (NIH, Bethesda, MD; http://rsbweb.nih.gov/ij/) was used for calculating distances. As an index of the degree of cell migration, we examined the time during which a given cell could be traced in the movie. Because cell movement in the terminal cluster was found to be nondirected, we presumed that the cell was less migratory if it remained within a given optical plain for longer. Cell shapes were traced manually.

Transmission Electron Microscopy

Tissues were immersed in ice-cold 2.5% glutaraldehyde in 0.1 M phosphate buffer (PB; pH 7.2). To improve fixation, the specimens were then irradiated in a microwave oven 10 to 15 times, applying current 3 sec each time at 5-sec intervals as described (Kadoya and Yamashina,1991). Further fixation was conducted for 1 hr at 4°C. Tissues were washed with PB (5 min × 3), then post-fixed in 1% OsO4 in PB for 1 hr at room temperature, and rinsed with distilled water (DW). To enhance the electron density of collagenous fibrils, some specimens were further stained en bloc with either 0.5% uranyl acetate or 1% tannic acid in DW. They were then dehydrated with a series of increasing ethanol concentrations and embedded in an epoxy resin mixture. Ultrathin sections were cut, stained with aqueous uranyl acetate and lead citrate, and examined in an electron microscope (1200EX, JEOL, Tokyo, Japan) at an accelerating voltage of 80 kV.

Acknowledgements

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

Part of this work was done at the Bio-Imaging Research Center, Kitasato University School of Medicine.

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

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
DVDY_22312_sm_SuppMovie-S1.mov8845KSupporting Movie S1. Confocal time-lapse movie of cell dynamics within a terminal cluster of a developing submandibular gland (SMG). Images of a live SMG explant at a fixed optical plane were acquired every 75 sec for 5 hr 36 min and assembled to form a movie at 6 frames/sec. Two cells at the distal end of the cleft are marked with an asterisk and a triangle, respectively, and have been traced over 4 hr. Cells undergoing mitosis are marked (M). Arrow indicates a mesenchymal cell process that is seen distant from the invading cleft tip at the later stage of cleft elongation.
DVDY_22312_sm_SuppMovie-S2.mov1846KSupporting Movie S2. Confocal time-lapse movie of duct cells in the developing submandibular gland (SMG). Images of a live SMG duct at a fixed optical plane were acquired every 75 sec for 3 hr 45 min and assembled to a movie at 6 frames/sec.
DVDY_22312_sm_SuppMovie-S3.mov8331KSupporting Movie S3. Confocal time-lapse movie of cells dynamics during cleft elongation of the developing submandibular gland (SMG) epithelium. Images of a live SMG explant at a fixed optical plane were acquired every 75 sec for 5 hr 38 min and assembled to a movie at 6 frames/sec. The cleft grew slowly for 1–2 hr, and then suddenly extended rapidly. Arrows indicate the tip of a cleft during periods in which the cleft was extending rapidly.
DVDY_22312_sm_SuppMovie-S4.mov1079KSupporting Movie S4. Confocal time-lapse movie of cleft wriggling shown at high magnification. Images of live submandibular gland (SMG) explant at a fixed optical plane were acquired every 75 sec for 1 hr 30 min and assembled to a movie at 6 frames/sec. Arrows indicate the deepest end of an elongating cleft.

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.