Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Branching of the embryonic mouse lung epithelium results from a complex inductive interaction with its homologous mesenchyme (1, 2). The exact mechanisms mediating this process are unknown, however studies suggest that the liberation of specific growth factors and/or direct cell-cell contacts play a role by differentially regulating epithelial proliferation rates and basement membrane localization patterns within distal budding and cleft-forming regions (3, 4). It has therefore been postulated that, when compared with the more morphogenetically stable cleft regions, a relative thinning of the epithelial/mesenchymal basement membrane and relatively higher local epithelial proliferation rate directs or supports lung bud outgrowth (3). For example, budding lung epithelia demonstrate a relatively higher tritiated thymidine incorporation after their induction by either epidermal growth factor (EGF) or salivary mesenchyme (5). Furthermore, the addition of DNA synthesis inhibitors to in vitro cultured lung explants significantly reduces the epithelium's capacity to branch (9). Although specific epithelial cell cycle times within cleft and budding regions have not been quantitated in the whole lung during development, the important role of epithelial proliferation for normal bud formation has been further demonstrated by a number of recent transgenic studies. For example, targeted misexpression of bone morphogenetic protein-4 (BMP-4) inhibits epithelial proliferation and results in smaller lungs with distended air sacs (10). Similarly, EGF receptor-deficient lungs, when cultured in the presence of EGF, although displaying a similar size, contain approximately half as many peripheral buds as wild-type lungs (11).

A role for the differential deposition of basement membrane components at the epithelial/mesenchymal interface during lung branching morphogenesis has been implied by a variety of techniques. Both light and electron microscopy have demonstrated that a well-defined, intact basement membrane is associated with areas of morphogenetic stabilization, such as clefts, whereas thinnings and discontinuities are associated with areas of active bud outgrowth (12, 13). With regard to specific basement membrane components, sulfated glycoconjugates and fibronectin, on the one hand, demonstrate a greater immunolocalization to the clefts or areas of morphogenetic stabilization when compared with the more morphogenetically active budding regions (14, 15). On the other hand, immunohistochemical studies in the mouse have demonstrated a uniform collagen IV, nidogen, and laminin immunolocalization at the epithelial/mesenchymal interface (16). Interestingly, more recent studies of specific laminin , , and  messenger RNAs (mRNAs) in the developing mouse (20) and polypeptide chains in the developing human (21) have demonstrated their variable co-expression within distal budding and cleft-forming regions, thus suggesting that the immunolocalization pattern of the laminins may be more complex than initially thought. These latter results have also led to the suggestion that specific laminin isoforms play a complex role in the establishment of variations in epithelial/mesenchymal basement membrane assembly which are either instructive or permissive to cleft and bud formation (22, 23). Such a hypothesis is supported by the observed disruption to normal branching morphogenesis after the addition of both polyclonal and site-specific monoclonal anti-laminin antibodies to in vitro cultivated lung and salivary explants (19, 24, 25). In other studies, overexpression of the potent regulator of proliferation and extracellular matrix deposition, transforming growth factor-1 (TGF-1), in developing lung epithelia (26) and null mutation of integrin 3, a gene encoding a basement membrane receptor subunit known to complex with integrin 1 and thus bind fibronectin, laminin, and collagen (27), both resulted in decreased lung branching morphogenesis.

The aim of the present study was, therefore, to examine in detail the relationship between epithelial budding, epithelial proliferation, and immunolocalization patterns of the specific basement membrane components nidogen, laminin-1, fibronectin, and collagen IV in the developing mouse lung. The development of a high-resolution 5-bromo-2'-deoxyuridine (BrdU) technique, designed to quantitate epithelial proliferation rates within budding and cleft-forming regions of the lung, and a tissue culture technique designed to favor two-dimensional epithelial branching has enabled the demonstration of a clear correlation between these three parameters of lung morphogenesis. Furthermore, we show that epithelial proliferation is dependent upon epithelial cell-cell contact and that the potential to form a branching lung bud is restricted to the most distal budding regions of the lung and is associated with its ability to maintain and produce formative differences in basement membrane composition and epithelial proliferation rates.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Collection of Fetal Mouse Lungs

F1(CBA/Cah and C57BL/6J) mice were mated and the appearance of a vaginal plug was designated as embryonic day 0.5 (E0.5). Fetal lungs were dissected at E11.5 in phosphate-buffered saline (PBS) and transferred to M2 medium (30). Lungs were collected either intact or further dissected with tungsten needles to separate distalbud (DB) and interbud (IB) regions. Distalbud regions were defined as those which lay distal to the widest margin across the terminal bud (31). Interbud regions were defined as those which lay proximal to this margin. All tissue was explanted upon 35-mm plastic dishes coated with 0.1% swine-skin gelatin in 120 µl of Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum (batch No. 551; GIBCO, Burlington, ON, Canada).

Separation of Epithelium from Mesenchyme

Freshly dissected E11.5 left lungs were rinsed in PBS and then incubated in 1.5% pancreatin (Sigma, St. Louis, MO) and 2.5% trypsin (CSL, Melbourne, Australia) in PBS at 4°C for 15 min (20). The epithelium was then dissected from the mesenchyme with tungsten needles in M2 medium at 37°C and cultured as mesenchyme-free, intact epithelial tubes.

Proliferation Studies

Whole lung buds and dissected DB and IB regions were cultured in vitro overnight and then pulsed with BrdU (Amersham, Buckinghamshire, UK), undiluted, for periods ranging from 3 to 6 h (see RESULTS). Tissue was peeled from the culture plates and immediately frozen in Tissue Tek OCT embedding compound (Miles, Naperville, IL) cooled in isopentane on dry ice. Serial 5-µm sections were cut with a Reichert Jung Frigocut 2800E cryostat and selected for those containing areas where the epithelium formed terminal acini. Sections were dried onto slides coated with 0.5% gelatin and 0.5% potassium dichromate, fixed for 5 min in 4% paraformaldehyde, hydrolyzed with 0.1 M HCl for 10 min at room temperature, and washed with 0.3% Triton X-100. After three 5-min washes in PBS, sections were incubated for 1 h with an anti-BrdU mouse monoclonal antibody (Amersham) diluted 1:3 in PBS, washed again three times for 5 min each with PBS, and then incubated for 1 h with a peroxidase-conjugated antimouse immunoglobulin (Amersham) diluted 1:3 in PBS. The 2^o antibody was localized by the polymerization of 3,3'-diaminobenzidine in phosphate buffer in the presence of nickel and cobalt (DAB substrate; Amersham), thus producing a black stain after approximately 5 min. Tissue nuclei were counterstained with Harris' Haematoxylin (Sigma) to give a purple stain (32). Proliferation indices in the whole lung and dissected DB and IB regions were determined by counting the total number of BrdU-positive and BrdU-negative nuclei in the designated regions after each BrdU pulse.

The isolated epithelial explants were studied as whole mounts. Prior to detection of incorporated BrdU, as described above, tissue sections were fixed for 5 min in 70% ethanol and hydrolyzed with 0.1 M HCl for 10 min at room temperature. To determine the purity of cell populations, after fixation at 4°C in 70% ethanol overnight, isolated epithelial explants were exposed to a mouse anti-cytokeratin monoclonal antibody (Silenus, Hawthorn, Australia) at a dilution of 1:20 for 1 h at room temperature. Antigen-antibody complex formation was detected with the biotinylated goat antimouse IgG and alkaline phosphatase conjugated with streptavidin (Silenus) for 1 h at room temperature. A pink/red color was obtained by the addition of napthol AS-MX phosphate and fast red 0.1 M Tris, pH 8.2 salt (32). This is indicated by a grey color in the accompanying black-and-white photomicrographs. Lung mesenchymal-cell whole mounts served as negative controls.

Morphometric Analysis of Epithelial Isolates

Total cell number and proliferation rates were determined by one of two methods. At 0 and 24 h, total cell number was determined following trypsin (12.5%)/ethylenediaminetetraacetic acid (2%) treatment using a hemocytometer. The second method was used for cultured explants. Point counting was used to determine the area fraction of labeled cells after photographs of each sample had been taken (33). A double quadratic lattice of 0.5-cm² spacing was placed over each photomicrograph and the area fraction of labeled cells was determined as the number of points falling on the total cell population. Total cell number and the mean area fraction of BrdU-labeled cells were calculated for each time point. Average disc diameter was calculated using the average widest aspect of all discs from each respective time point.

Basement Membrane Localization

E11.5 whole lung explants and respective dissected regions were cultured in vitro overnight and prepared for serial sectioning in the same manner as those for the proliferation studies. Rabbit polyclonal antisera against laminin-1, diluted 1:100 in PBS (34); nidogen diluted 1:100 in PBS (35); fibronectin diluted 1:200 in PBS; and collagen IV diluted 1:200 in PBS (36) were incubated with unfixed sections for 1 h at room temperature. After several washes in PBS, bound antibody was detected with a peroxidase-conjugated donkey antirabbit IgG antibody (Silenus) diluted 1:100 in PBS which was applied to the sections for 1 h at room temperature. The 2^o antibody was detected with the DAB substrate as described above. Tissue sections were counterstained as described above. For controls, either the primary or secondary antibody were omitted and replaced with PBS.

Photography

Photomicrography was performed using an MC 80 microscope camera mounted on an Axioscope microscope (Zeiss, Oberkochen, Germany).

Statistical Analysis

One- and two-way analyses of variance, Student's t tests, and chi-squared test were used where appropriate (37).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Epithelial Proliferative and Branching Potential in the Developing Lung

Preliminary in vivo BrdU labeling studies of the E11.5 embryo demonstrated strong BrdU immunoreactivity of the developing intestinal epithelium, with an absence in the lung bud (data not shown). Proliferation studies were therefore performed in vitro. Explanted E11.5 lungs underwent progressive branching in vitro for a period of at least 96 h after dissection (data not shown). Individual lung buds were pulse-labeled in vitro with BrdU for 3, 4, 5, or 6 h, individually frozen, and orientated for sagittal serial sectioning. Sections were searched for those containing the entire lung epithelial tree and those in which the epithelium most closely approached the lung capsular mesenchyme. In this way, the most distal budding epithelia could be differentiated from epithelia constituting cleft-forming regions. Usually two or three serial sections from each frozen lung explant demonstrated this particular configuration. After a 3-h BrdU pulse, the majority of BrdU immunoreactive nuclei were observed within the epithelial layer of the DB regions (43% ± 13%, DB n = 58, lung n = 30; Figures 1a and 2a). Fewer BrdU-immunoreactive nuclei localized within the IB epithelium at this time (P < 0.05, 34% ± 17%, IB n = 48, lung n = 30). A greater number of BrdU-immunopositive epithelial cells were seen within the IB epithelium situated closer to the IB/DB border than were seen deep within the IB region proper. After a 4-h pulse, the difference between BrdU-immunoreactive epithelial nuclei in the DB and IB regions was 67% ± 16% compared with 35% ± 12%, respectively (P < 0.001, n = 51, lung n = 28; Figures 1b and 2a). Furthermore, in the mesenchyme, BrdU-immunoreactive cells were seen to preferentially localize to the DB regions where they additionally formed a distinct monolayer adjacent to the budding epithelium. After a 5-h pulse, a significantly greater number of BrdU-immunoreactive epithelial nuclei were observed in the DB regions when compared with the IB regions (P < 0.001, DB = 74% ± 14%, n = 63; IB = 32% ± 11%, DB n = 48, IB n = 36, lung n = 30; Figures 1c and 2a). After a 6-h pulse, approximately 2.5 times as many BrdU-immunoreactive epithelial nuclei were observed in the DB regions as compared with the IB regions. The difference in BrdU-immunoreactive epithelial nuclei between the DB and IB regions at this time was 87% ± 9% compared with 37% ± 14% (P < 0.001, DB n = 60, IB n = 48, lung n = 32; Figures 1d, 1e, and 2a). Distalbud epithelial cells, therefore, demonstrated a significantly greater BrdU incorporation rate at all times examined. The proportion of BrdU-immunoreactive cells within the epithelium always appeared higher than the proportion of BrdU-immunoreactive cells within the mesenchyme.

View larger version (112K): [in this window] [in a new window]   Figure 1.   Epithelial proliferation and branching within the E11.5 dissected lung. Interbud (IB) and distalbud (DB) BrdU-positive cells are brown/black (P); hematoxylin counterstain shows purple/blue nuclear staining (N). Dashed lines demarcate IB and DB regions. (a- e) Whole lung. Relatively more BrdU-immunopositive cells can be seen within the DB than the IB epithelium at each time. (a) Three-hour BrdU pulse. IB BrdU-immunopositive epithelial cells form a gradient of decreasing frequency away from the IB/DB border (G). (b) Four-hour BrdU pulse. The BrdU immunoreactivity of mesenchymal cells is associated with budding regions (R) and forms a monolayer immediately adjacent to the budding epithelium (M). (c) Five-hour BrdU pulse. (d) Six-hour BrdU pulse. (e) Higher magnification of (d). Nearly all epithelial nuclei of DBs demonstrate BrdU immunoreactivity. Relatively fewer IB epithelial nuclei are BrdU-positive. ( f ) The epithelium of isolated DBs shows numerous folds and indentations after 24 h of culture. (g) The epithelium of isolated IBs after 24 h of culture remains an unbranched tube. (h, i) Low and high magnifications, respectively, of isolated and cultured DBs after staining for BrdU after a 6-h pulse. The newly formed DB epithelium (NDB) contains a relatively higher proportion of BrdU-positive cells than the newly formed IB epithelium (NIB). The mesenchymal cells adjacent to the DB epithelium form a monolayer of BrdU immunoreactivity (L). A proximodistal gradient of BrdU immunoreactivity is seen within the NIB epithelium (G). (j) An isolated and cultured IB region stained for BrdU after a 6-h pulse. The explant remains tubular, and no new DBs or IBs form. Scale bar = 100 µm.

A strong linear relationship between BrdU immunoreactivity in the DB region and time was observed, indicating a continual and significant increase in BrdU uptake throughout the culture period (P < 0.001). A flattening from the pure linear relationship in the latter stages of culture suggests a reduced increase in BrdU incorporation (P < 0.05). This flattening may be due to the continued proliferation of BrdU-immunoreactive cells. Although greatest after the 6-h pulse, no significant differences in the percentage of BrdU-immunoreactive nuclei in the IB regions were seen during the culture period. Differences in BrdU incorporation did not appear to be due to a latency of penetration because BrdU-immunopositive mesenchymal cells were seen adjacent to immunonegative epithelial cells deep within the lung proper.

E11.5 DB and IB regions were isolated by dissection and placed into culture to investigate the autonomy of the mechanisms controlling epithelial proliferation and branching. Of 48 isolated DB regions examined after 24 and 48 h of culture, 44% and 83%, respectively, demonstrated numerous indentations and branch points (Figures 1f and 2b). In contrast, of 36 IB regions examined after 24 and 48 h of culture, 11% and 27%, respectively, demonstrated indentations and branch points (Figures 1g and 2b). These data demonstrate that isolated DB regions have a significantly greater capacity to undergo branching morphogenesis than isolated IB regions after both 24- and 48-h time periods in vitro (P < 0.05 and P < 0.001, respectively; n  36 per time point; Figure 2b). Although demonstrating a trend toward greater BrdU immunoreactivity within the newly formed distal budding epithelium after a 6-h pulse, immunoreactivity was not statistically different when compared with that within the epithelium of the newly formed clefts (P = 0.09, lung n = 9; Figures 1h, 1i, and 2c). A higher number of mesenchymal cells situated immediately adjacent to the newly formed buds appeared to show BrdU immunoreactivity when compared with those situated elsewhere within the explant. No clear patterns of regional differences in BrdU uptake were observed within the isolated IB explants (Figure 1j).

View larger version (21K): [in this window] [in a new window]   Figure 2.   BrdU incorporation in the in vitro cultivated whole lung and specifically dissected lung regions. (a) A comparison of the percentage of BrdU-immunoreactive epithelial nuclei in the DB versus IB regions after 3, 4, 5, and 6 h BrdU pulse times. DB and IB epithelial BrdU uptake is significantly different after a 3-h BrdU pulse (P < 0.05). This difference is maintained throughout the entire culture period (P < 0.001). BrdU uptake in the DB region increases with time (P < 0.001). BrdU uptake in the IB regions does not change (P = 0.69). The results obtained represent the combined data of three separate experiments. n  10 lung buds per time point per experiment. (b) The branching activity of each isolated region. After 24 and 48 h in vitro, a greater number of isolated DB regions than isolated IB regions have branched (P < 0.05 after 24 h, P < 0.001 after 48 h). (c) A comparison of the percentage of BrdU-immunoreactive epithelial nuclei in the newly formed DB versus the newly formed IB regions of the explanted DB regions after a 6-h BrdU pulse. A trend toward greater immunoreactivity in the newly formed DB epithelial regions is seen. Lung n = 9, NIB n = 6, DIB n = 8, P = 0.09.

Proliferation and Morphogenesis of Isolated Epithelial Aggregates In Vitro

In the absence of lung mesenchyme, lung epithelium fails to undergo normal morphogenesis (1). Epithelia were separated from their mesenchyme and, after a 24-h attachment period, cultured in the presence of BrdU for up to 72 h in order to investigate their capacity to retain regional differences in epithelial proliferation in the absence of mesenchymal morphogenetic cues (Figure 3). All cultured cells demonstrated cytokeratin immunoreactivity, thus demonstrating that these cell aggregates were not contaminated with mesenchyme. Whole mounts of isolated mesenchymal cells showed no cytokeratin immunoreactivity (data not shown). After the 12-h BrdU pulse, the lung epithelial tube collapsed to form an epithelial disc (Figure 3a). This disc comprised a central multilayered structure with a more peripheral monolayer. The central multilayered region showed a greater proportion of BrdU-immunoreactive cells than the more peripheral region. After 24- and 48-h pulses, a greater number of BrdU-immunoreactive epithelial cells appeared within the peripheral zone (Figures 3b and 3c). By 48 h, the central multilayered zone disappeared and the aggregate comprised a simple cell monolayer.

View larger version (50K): [in this window] [in a new window]   Figure 3.   Proliferative and branch-forming capacity of isolated epithelial cells. Epithelial cells are grey. BrdU-immunoreactive epithelial cells have a black nuclear stain (P). BrdU-negative epithelial cells show no nuclear stain (N). (a) A typical epithelial explant isolated from mesenchyme after a 12-h BrdU pulse. Note the central dense and multilayered region containing many BrdU-positive cells (C) and the more peripheral region showing few BrdU-positive cells (F). (b) After a 24-h BrdU pulse, more BrdU-positive cells appear within the peripheral region. (c) After a 48-h BrdU pulse, BrdU-positive cells are more dispersed and the original central and multilayered structure has collapsed. Scale bar = 100 µm.

Potential mechanisms responsible for the disappearance of the BrdU-immunoreactive and multilayered central zone were investigated. A continual increase in epithelial disc diameter throughout the entire culture period demonstrated radial epithelial cell migration (P < 0.001, n  8 per time point; Figure 4). Trypan blue exclusion tests demonstrated that the reduction in cell number during this migration was due to cell death rather than the detachment of living cells from the substrate.

View larger version (57K): [in this window] [in a new window]   Figure 4.   Epithelial migration after isolation and in vitro culture. Measurements of the epithelial disc after isolation from the mesenchyme. A progressive and significant increase in disc diameter is seen throughout the culture period (P < 0.001). At least eight individual explants were measured per time point.

A significant 2.6-fold increase in total epithelial cell number within the aggregates was observed after the 24-h attachment time by hemocytometric analysis (P < 0.01, number of cell aggregates = 15 at 0 h, n = 8 at 24 h; Figure 5a). BrdU was added at this time and cell number was then assessed by morphometric analysis after 12-, 24-, 48-, and 96-h pulses. No significant difference in total or BrdU-immunoreactive epithelial cell number was observed between the 12- and 24-h BrdU pulses (n = 13 after 12 h, n = 15 after 24 h; Figures 5b and 5c). Between the 24- and 48-h BrdU pulses and the formation of a monolayer, although no significant difference was observed in BrdU-immunoreactive cell number of the aggregates, total cell number decreased significantly (P < 0.05, n = 15 after 48 h; Figures 5b and 5c). Total epithelial cell number within the aggregates demonstrated a further decrease between the 48- and 96-h BrdU pulses (P < 0.01, n = 8 after 96 h; Figures 5b and 5c). This was accompanied by a significant increase in BrdU-immunoreactive epithelial cell number (P < 0.05, n = 10 at 96 h).

View larger version (20K): [in this window] [in a new window]   Figure 5.   Analysis of epithelial cell number and proliferation after isolation from the mesenchyme. (a) A significant increase in cell number is observed within the first 24 h of culture by hemocytometric analysis (P < 0.005, n  8 per time point). (b) A significant decrease in cell number is observed after the 48-h BrdU pulse, or 72 h of culture by point counting (P < 0.05, n  8 per time point). (c) No change is seen in the total number of BrdU-immunoreactive cells between the 12- and 48-h BrdU pulses by point counting. After the 96-h BrdU pulse, a significant increase in BrdU-immunoreactive cell number is observed (P < 0.05, n  10 per time point). Asterisks mark significantly different cell populations.

Localization Patterns of Specific Basement Membrane Components in the Developing Lung

In order to correlate directly specific basement membrane immunolocalization patterns with sites of epithelial BrdU incorporation and bud outgrowth, whole lung explants were first cultured overnight in the same fashion as those pulsed with BrdU. Once again, only sections containing an entire epithelial tree and those in which the epithelium most closely approached the lung capsular mesenchyme were assessed. Sections from different lungs rather than adjacent sections from the same lung were assessed due to the infrequency of epithelial sections which contained DB epithelia. A greater localization of nidogen was observed within the basement membrane of the bronchial epithelial cells comprising the developing cleft regions (Figure 6a). In contrast, a reduced immunoreactivity was seen in the basement membrane of the DB epithelia. A similar immunoreactivity pattern was observed for collagen IV, fibronectin, and laminin-1 (Figures 6b, 6c, and 6d, respectively). Due to our strict criteria of section selection for morphologic assessment, for each basement membrane component, areas of reduced immunoreactivity were seen to correlate directly with areas of higher epithelial BrdU immunoreactivity. Antibodies to collagen IV and laminin-1 also localized to small rings within the mesenchyme, possibly demarcating developing blood vessels (Figures 6b and 6d). In contrast, nidogen and fibronectin demonstrated diffuse staining throughout the mesenchyme (Figures 6a and 6c, respectively). It must be noted that a range of different concentrations of both primary and secondary antibodies were used and those reported demonstrated the clearest regional differences in immunoreactivity of each specific basement membrane component between the DB and IB regions. Furthermore, the anti-laminin-1 antibody employed demonstrates a far greater affinity for intact laminin when compared with reduced or alkylated laminin (34). No immunoreactivity for each of the basement membrane components examined was found in control sections from which the primary antibody was omitted (data not shown).

View larger version (99K): [in this window] [in a new window]   Figure 6.   Basement membrane localization patterns in the E11.5 mouse lung. Immunoreactive material is black. Laminin-1 and collagen IV immunoreactivity forms rings within the mesenchyme; (R) ring, (IBB) interbud basement membrane, (DBB) distalbud basement membrane, (M) mesenchyme, (E) epithelium. Thinning of nidogen (a), collagen IV (b), fibronectin (c), and laminin-1 (d) immunoreactivity is seen in the DBs when compared with the IB. Diffuse staining is observed within the mesenchyme for each component. Scale bar = 100 µm.

Basement Membrane Localization Patterns of Isolated DB Regions In Vitro

In order to investigate further the autonomy of the mechanisms controlling basement membrane deposition, E11.5 DB and IB regions were isolated by dissection and placed into culture, and the localization of specific basement membrane components was analyzed. Immunohistochemical analyses of serial sections obtained from isolated and in vitro cultivated DB regions demonstrated similar localization patterns for nidogen, collagen IV, fibronectin and laminin-1 in isolated DB explants after 24 h of culture as seen in the whole lung explants (Figures 7a-7d, respectively). Newly formed DB basement membranes demonstrated a relatively lower immunoreactivity for each component as compared with the newly formed IB regions. Reduced immunoreactivity of each basement membrane component correlated with the regions of relatively greater epithelial BrdU incorporation.

View larger version (143K): [in this window] [in a new window]   Figure 7.   Basement membrane localization patterns in isolated and in vitro cultivated DB regions. Immunoreactive material is black; (M) mesenchyme, (E) epithelium, (IBB) IB basement membrane, (DBB) DB basement membrane. Nidogen (a), collagen IV (b), fibronectin (c), and laminin-1 (d) localization is greater within the IB regions. Localization of each component within the DB basement membrane is comparatively low. Scale bar = 100 µm.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Using an in vitro culture technique designed to favor two-dimensional branching of the E11.5 mouse lung epithelial sheet, we have demonstrated a clear correlation between the processes of bud outgrowth, epithelial proliferation, and the localization of specific basement membrane components nidogen, laminin-1, fibronectin, and collagen IV. Our systematic quantitation of DB and IB epithelial proliferation rates by BrdU pulse-labeling extends previous tritiated-thymidine studies by demonstrating that relative to the clefts, budding of the lung epithelium is accompanied by an at least 2.5-fold increase in proliferation rate. These studies support the Basement Membrane Remodelling Hypothesis (3), thus implicating a relatively low immunolocalization of specific basement membrane components at the epithelial/mesenchymal interface and an accompanying relative increase in epithelial proliferation in bud outgrowth. Conversely, in the IB or cleft regions, a relatively higher immunolocalization of specific basement membrane components at the epithelial/mesenchymal interface and an accompanying relatively lower epithelial proliferation rate correlate with morphogenetic stabilization.

We have also demonstrated that isolated DB regions possess a greater propensity to undergo in vitro branching when compared with isolated cleft/IB regions. Furthermore, when compared with the newly established cleft/IB regions, a correlation between reduced nidogen, collagen IV, fibronectin, and laminin-1 immunoreactivity within the budding epithelial/mesenchymal basement membrane and a trend for greater epithelial BrdU incorporation within the newly established budding regions is maintained. These results further support the Basement Membrane Remodelling Hypothesis but also demonstrate an autonomous morphogenetic heterogeneity within the lung's presumptive bronchiolar/air-sac region. Previously, morphogenetic heterogeneity has been observed only when comparing presumptive bronchiolar/air-sac regions and presumptive tracheal regions (5, 8, 38, 39). In the former experiments, not repeated during this study, mesenchyme from presumptive bronchiolar/terminal air-sacs was able to induce supernumerary tracheal bud formation. Similarly, mesenchyme from presumptive trachea was able to inhibit branching of the presumptive bronchiolar/terminal air-sac regions. Thus the epithelium was suggested to represent a structure of uniform morphogenetic potential. We suggest, therefore, that the potential to form a lung bud is connected to a heterogeneity of the bronchiolar mesenchyme, independent of presumptive tracheal mesenchyme, which is compartmentalized according to specific positions within the lung's presumptive bronchiolar/air-sac region.

The precise nature of the diversity of cues necessary for epithelial bud outgrowth and cleft initiation/stabilization are unknown; however, our observation of a specialized, highly BrdU-immunoreactive monolayer immediately surrounding the budding distal tips of whole lungs pulsed for 3 and 4 h, and DB isolates pulsed for 6 h with BrdU, supports the notion of a reciprocal interaction involving paracrine and autocrine regional growth induction at the epithelial/mesenchymal interface (40). The relative increase in epithelial proliferation in areas of bud outgrowth as compared with areas of newly formed clefts may therefore be attributable to the action of specific growth factors and their receptors which, among many other candidates (41), include TGF-1 (26), the fibroblast growth factor 2 (FGF2), EGF receptors (11, 44), and BMP4 and BMP7 (10, 45), all of which have demonstrated important roles in normal branching morphogenesis in a number of recent transgenic studies.

Although the dispensability of direct epithelia-mesenchymal cell-cell contact for branching lung morphogenesis has been demonstrated by studies involving EGF, FGF, and salivary and lung epithelia embedded in Matrigel (42, 46), the indispensability of epithelial cell-cell signaling has been demonstrated by a synergistic inhibition of lung branching after the in vitro application of antibodies to E-and P-cadherin (47). Our findings of a gradual increase in BrdU immunoreactivity of epithelia within the DB region of the intact in vitro cultivated lung with time demonstrates asynchronous proliferation within this epithelial population. This implies that epigenetic signaling molecules play a more important role in regulating local epithelial proliferation rates during lung branching morphogenesis than communication via direct epithelial cell-cell contact.

Such a notion is supported by our studies of isolated lung epithelia. As previously described, in the absence of mesenchyme, lung epithelia collapse to form unbranched structures in which individual cells lose orientation with each other and die (1, 2, 48). Further, we show that such epithelia transiently retain an ability to proliferate, resulting in a net cell-number increase within the first 24 h of culture. Only epithelial cells constituting the central multilayered structure were BrdU-positive after a 12-h pulse, whereas those in the peripheral zone were BrdU-negative. Cells of the peripheral zone appeared to migrate from the central aggregate, lose cell-cell contacts, be incapable of proliferation, and die. In contrast, those cells remaining in the central zone retained cell-cell contact and proliferated. It therefore appears that direct epithelial cell-cell contact is a prerequisite for proliferation. Thus a hierarchy is established whereby permissibility to mesenchymal signaling is dependent upon direct cell-cell coupling. Such permissibility may then result in epithelial regions of differing proliferation rates and therefore contribute to bud outgrowth and cleft formation. Indeed, treatment of in vitro cultivated lungs with cytochalasin inhibited de novo bud formation, thus suggesting that an initial epithelial outpocketing produced by a localized alteration in cell shape is prerequisite to bud outgrowth (9).

Previous studies have demonstrated localization of nidogen and collagen IV throughout the entire epithelial/ mesenchymal basement membrane (16, 19). As previously described by other investigators for laminin-1 and fibronectin (15, 21, 22), we also demonstrated a clear gradient of immunoreactivity within the epithelial/mesenchymal basement membrane for nidogen and collagen IV whereby, when compared with cleft regions, immunolocalization is reduced in areas of bud outgrowth. Such differences in observations may be attributable to the relatively low concentrations of antibody, the age of the explant, the method of culture, and the precise plane of section used in our study, factors which, in combination, augmented the differences reported in the precise localization patterns (Mollard and Dziadek, personal observations). Interestingly, previous studies from our laboratory have demonstrated uniform expression of nidogen and collagen 1(IV) mRNA throughout the developing lung's mesenchyme by in situ hybridization (20). For a number of reasons, we speculate that differences between protein and mRNA localization patterns are attributable to changes in nidogen's susceptibility to proteolysis, according to differences in the sites of synthesis of the various chains of laminin-1 (23). First, mRNA-encoding laminin-1 and laminin-1 are uniformly expressed within the developing lung's epithelium and mesenchyme, whereas laminin-1 chain expression is restricted to the mesenchyme subjacent to the epithelium of the IB regions and within the epithelium of the most distal budding regions (20). Because laminin/nidogen complexes are rapidly assembled intracellularly after synthesis and laminin-1 protects nidogen from proteolytical degradation, nidogen synthesized in the DB regions may be more susceptible to proteolytic degradation than nidogen co-synthesized with laminin-1 in the IB regions (23, 49, 50). With its known ability to regulate the activity of proteolytic enzymes (51, 52) and its suggested role as a pivotal organizer of the basement membrane (23), nidogen may therefore orchestrate the required events which lead to a relative reduction of other components such as collagen IV in the DB basement membrane. That the interaction between nidogen and laminin does play a critical role during branching morphogenesis and in establishing basement membrane integrity has been directly demonstrated by the perturbation of lung and salivary gland branching morphogenesis and salivary gland basement membrane assembly in vitro after the addition of antibodies directed to the nidogen binding site on laminin (19, 25). Whether such alterations in basement membrane structure directly link relative changes in epithelial cell proliferation and morphogenetic processes is unknown; however, a reduced lung-branching potential in integrin 3 subunit null mutant mice supports the notion that the epithelium's underlying basement membrane plays a crucial role in determining its morphogenetic property (29).

In conclusion, our observations support the Basement Membrane Remodelling Hypothesis, which suggests that bud outgrowth is accomplished by a localized and relative reduction in basement membrane components and a corresponding relative increase in epithelial proliferation rates (3). We have also implicated nidogen as a determinant of morphogenetic potential by its protein localization pattern in the epithelial basement membrane, and have shown a dependency upon epithelial cell-cell contact for the permissibility to growth signaling. Our studies have directly identified and quantitated differences in epithelial proliferation rates within IB and DB compartments of the E11.5 lung. We have directly correlated these differences with changes in the localization of specific basement membrane components, i.e., nidogen, laminin-1, fibronectin, and collagen IV. We have also demonstrated that the potential for branching morphogenesis lies within the most distal growing tips of the developing lung's bronchial tree. Furthermore, we have shown that this potential is associated with the distal budding region's retained ability to instruct relatively higher epithelial proliferation rates in its newly formed buds as compared with its newly formed clefts. This potential is also connected to the ability of distal budding regions to instruct a higher localization of nidogen, laminin-1, fibronectin, and collagen IV to the basement membrane of newly formed clefts and a lower localization to the basement membrane of newly formed buds. With respect to a cause-and-effect relationship, it remains to be determined whether the demonstrated differences of these specific basement membrane components in the DB and IB regions signal alterations in the availability of mesenchymally derived growth factors to the associated epithelium, thus indirectly affecting epithelial proliferation, or whether they signal a direct effect by inducing specific changes to the epithelial cytoskeleton. Such questions may best be answered by in vitro cultivation of lungs made transgenic with inducible knockout constructs of each basement membrane component or, alternatively, their overexpression under the control of the surfactant promoter C promotor/enhancer (53). The subsequent effect on morphology, epithelial proliferation, and growth factor localization patterns and basement membrane assembly could then be determined by counting terminal buds, employing BrdU epithelial incorporation studies and immunohistochemistry.