Btbd7 Regulates Epithelial Cell Dynamics and Branching Morphogenesis

Abstract

During embryonic development, many organs form by extensive branching of epithelia through the formation of clefts and buds. In cleft formation, buds are delineated by the conversion of epithelial cell-cell adhesions to cell-matrix adhesions, but the mechanisms of cleft formation are not clear. Here we identify Btbd7 as a dynamic regulator of branching morphogenesis. Btbd7 provides a mechanistic link between the extracellular matrix and cleft propagation through its highly focal expression leading to local regulation of Snail2 (Slug), E-cadherin, and epithelial cell motility. Inhibition experiments show that Btbd7 is required for branching of embryonic mammalian salivary glands and lungs. Hence Btbd7 is a regulatory gene that promotes epithelial tissue remodelling and formation of branched organs.

Branching morphogenesis is a fundamental developmental process generating the branched epithelia of many organs, including lungs, kidneys, and the mammary and salivary glands (1–8). Branches are generated by local outgrowths or by formation of clefts that subdivide epithelia into buds. Bud or tubule extension is promoted by various growth factors (9–11), but the mechanism of cleft formation is poorly understood. One model for cleft formation involves the local loss of epithelial cell-cell adhesions and replacement by interactions with extracellular matrix accumulating within clefts (9, 12, 13).

The matrix protein fibronectin is required for salivary, kidney, and lung branching (13–15). Wedges of fibronectin translocate inward as clefts form between randomly motile epithelial cells, accompanied by loss of the cell-cell adhesion molecule E-cadherin in cells adjacent to the fibronectin (13, 16). How a matrix molecule can drive cleft formation and branching is unknown. Here we identify the activity of a regulatory gene termed Btbd7 at cleft-forming sites and provide mechanistic insight for dynamic cleft propagation in branching morphogenesis.

We focused initially on the mouse submandibular salivary gland, a classical model for analyzing mechanisms of branching morphogenesis (12, 17, 18). During branching, we observed local loss of the relatively columnar organization of cells in the outer layer of salivary epithelia (19) near the bottom of advancing clefts (Fig. 1A–C and S1). Time-lapse confocal microscopy of glands from EGFP-expressing transgenic mice suggested that transient intercellular gaps form at this site as clefts advance, which we confirmed directly using fluorescent dextran to document dynamic gap formation as clefts deepen (Fig. 1D and S2).

Figure 1.

Cleft formation and Btbd7. (A to C) Formation of clefts in a submandibular gland at E12.5 (A) or in glands cultured for 5 h (B), or 10 h (C). Fixed samples were stained with anti E-cadherin antibody (red) and DAPI (blue). Clefts (open triangles), basement membrane (white line), and E-cadherin (white dashed line) at cell boundaries of outer epithelial cells. (D) Cleft progression by transient gap formation as visualized by tetramethylrhodamine-labeled dextran; e, epithelium; m, mesenchyme. (E) Diagram of laser microdissected developing salivary gland (red, cleft; blue, bud region). (F and G) Btbd7 mRNA expression by qPCR in developing embryonic and postnatal salivary glands (F) and in other organs (G). Scale bar: 10 µm.

To search for a regulatory gene that might be involved in this process, we compared genes expressed in laser-microdissected epithelial cells from cleft regions versus end bud epithelial cells using T7 amplification and serial analysis of gene expression (SAGE) (Fig.1E). As expected (13), fibronectin was strongly differentially expressed in cleft epithelial cells compared to buds (124/20,053 tags versus 4/22,244 tags; P<0.001). Tags corresponding to Btbd7 (BTB (POZ) domain containing 7) were identified in cleft but not bud cells: 7/20,053 vs. 0/22,244 tags, respectively (P<0.01). This preliminary finding was confirmed by quantitative polymerase chain reaction with reverse transcription (qPCR; see below).

The deduced Btbd7 protein contains 1,130-amino acids with 2 putative BTB/POZ domains (Fig. S3). Regulatory proteins containing BTB domains are evolutionarily conserved from Drosophila to mammals (20, 21). Btbd7 mRNA expression in developing salivary glands was highest at embryonic day 13 (E13), a stage of particularly active salivary gland branching (Fig. 1F,G). Btbd7 expression by in situ hybridization showed striking concentration around the bottom and lower sides of forming clefts, whereas other salivary epithelial regions showed little or no expression (Fig. 2A–D, S4). The Btbd7-expressing cells surround sites of high fibronectin concentration at the base of forming clefts (13, 16), suggesting that Btbd7 might be induced by fibronectin or other matrix proteins. Because Btbd7 is also expressed in mesenchyme containing high levels of fibronectin (Fig. S4), we used mesenchyme-free explants of intact embryonic salivary gland epithelia. Fibronectin induced Btbd7 expression (P<0.001), whereas collagen I and IV had no effect (Fig. 2E). Thus, Btbd7 is a fibronectin-induced gene. This Btbd7 induction was accompanied by morphological conversion of the outer layer of columnar epithelial cells to a less organized pattern within 2 h (Fig. 2G), which mimicked the cell shape changes in vivo during cleft formation (e.g., Fig. 1A–C). This fibronectin-induced epithelial disorganization could be rescued by small interfering RNA (siRNA) knockdown of Btbd7 (Fig. 2H, S5).

Figure 2.

Btbd7 expression and regulation by fibronectin. (A–C) In situ hybridization of Btbd7 around the lower sides and bottom of clefts at various stages and magnifications. Open triangles and dotted lines show locations of clefts. Filled triangles highlight Btbd7 expression at the base of a cleft. (D Diagram of Btbd7 transcript pattern. (E) Btbd7 mRNA expression in salivary epithelial cells is induced by fibronectin as quantified by qPCR. (F–H) Exogenous fibronectin (Fn) reduces the organization of the outer layer of columnar epithelial cells; its effects are blocked by Btbd7 siRNA. Epithelia were stained for E-cadherin; traces below each image show basement membrane (white line) and E-cadherin (dashed line). Scale bars: (A–C) 50 µm; (F–H) 10 µm. ***P<0.001. Error bars, s.e.m.

To identify mechanisms of Btbd7 function, we turned to the well-characterized Madin-Darby canine kidney (MDCK) cell model for epithelial cell interactions (22). Transient Btbd7 transfection of MDCK cells or tetracycline-regulated expression in stably transfected cells (23) induced labile cell-cell adhesions and scattering of MDCK epithelial cell colonies to dispersing cell clusters in both 2D cultures and 3D collagen gels (Fig. S6, S7). This cell dispersal was associated with loss of E-cadherin from cell boundaries according to immunolocalization (Fig. 3A–F) and 77.1±2.5% decreased Ecadherin protein levels by immunoblotting (Fig. 3G). Btbd7 expression stimulated the velocity of cell movement in 2D cell culture two-fold, with increased random migration (Fig. S6). It induced outward cell dispersal from aggregates in 3D collagen gels (Fig. S7). Therefore, Btbd7 can induce loss of E-cadherin and increased cell dispersal in 2D and 3D, suggesting that the previously observed loss of E-cadherin in salivary gland cells adjacent to forming clefts in vivo (13) was due to the local induction of Btbd7.

Figure 3.

Btbd7 suppresses E-cadherin and induces Snail2. (A–C) Pooled, stably transfected MDCK cells were established for Tet-off regulation of expression of green fluorescent protein (GFP)-Btbd7 (A, B) or GFP alone (C) by doxycycline (Dox) removal to induce Btbd7. (D–F) The same cells were immunostained for E-cadherin. (G) Decrease of total E-cadherin protein by Western blot analysis after Btbd7 expression. (H) Btbd7 expression induces Snail2 expression by qPCR. Scale bar: 20 µm.

Epithelial cell dispersal is often regulated by transcription factors such as Snail, Snail2 (Slug), and Twist1. Btbd7 expression induced Snail2 mRNA in MDCK cells with little effect on the expression of Snail and Twist1 (Fig. 3H). Immunostaining confirmed that Btbd7 expression induces Snail2 expression and accumulation of this transcription factor in the nucleus (Fig. S8). This induction of Snail2 was not required for Btbd7-induced loss of E-cadherin (Fig. S9), analogous to the independent regulation of Snail2 and E-cadherin in HGF-treated MDCK cells (24).

After identifying this putative Btbd7 regulatory pathway in a model epithelial system, we searched for it in intact embryonic mouse tissues. Besides major increases of fibronectin and Btbd7 mRNA in cells adjacent to salivary clefts compared to bud epithelial cells of the same glands, Snail2 mRNA level was also elevated in cells near clefts (3.4-fold, P<0.02; Fig. 4A). Incubating isolated salivary gland epithelia with purified fibronectin induced Btbd7 rapidly within 20 min, followed by the induction of Snail2 at 60 min (Fig. 4B). These kinetics in intact salivary epithelia are consistent with the induction of Snail2 by Btbd7 observed in MDCK cells in vitro. To test this relationship more directly, we performed siRNA knockdowns of Btbd7 versus Snail2 in salivary epithelial cultures. Btbd7 knockdown inhibited fibronectin-induced Snail2 expression, whereas Snail2 knockdown did not affect Btbd7 expression, indicating that Btbd7 is upstream of Snail2 (Fig. 4C).

Figure 4.

Btbd7 expression induced by fibronectin regulates Snail2 expression and is required for branching morphogenesis. (A) mRNA expression of fibronectin, Btbd7, Snail2, and E-cadherin in cleft versus bud epithelia by qPCR. (B) Induction of Btbd7 and Snail2 expression after fibronectin addition to isolated salivary epithelium by qPCR. (C) Effects on epithelial expression of Btbd7 and Snail2 by siRNA knockdown of each gene. (D) Effect of Btbd7 siRNA knockdown in salivary gland organ culture and quantification of numbers of buds per gland (E; n=8). (F) Total bud volume after siRNA transfection. (G) Effect of Btbd7 siRNA knockdown in embryonic lung with quantification of branches (H; n=15). Scale bar, 100 µm. *P<0.05, **P<0.01, ***P<0.001. Error bars, s.e.m.

To test whether Btbd7 is necessary for branching morphogenesis, we suppressed Btbd7 expression in organ cultures of intact salivary glands using siRNA. Treatment with each of 4 different Btbd7 siRNA sequences produced substantial, morphologically similar suppression of cleft formation at 48–72 h (Fig. 4D,E and S10). Fewer (64% less) buds formed, but the buds were larger with the total volume of buds remaining about the same even though branching was inhibited (Fig. 4F). Cleft propagation speeds measured from time-lapse video movies (Movie S1) decreased from 6.2±1.1 µm/h in controls to 3.7±1.0 µm/h in Btbd7 knockdown glands, confirming Btbd7’s role in cleft propagation. We also tested whether Snail2 expression is required for branching morphogenesis using siRNA knockdown. Branching was inhibited ~67% by inhibition of Snail2 (Fig. S11), confirming its functional role in branching morphogenesis.

Although Btbd7 is normally expressed only focally in embryonic epithelia, we tested whether global over-expression by viral gene transfer to developing salivary epithelia could promote function. Ex vivo fragments of salivary epithelium transduced with adenoviral Btbd7 showed substantially enhanced tissue spreading (Fig. S12), consistent with the in vitro stimulation of MDCK cell dispersal by Btbd7. Virus microinjection and expression of GFP-Btbd7 versus GFP in intact embryonic salivary glands produced a small but statistically significant increase in branching morphogenesis (Fig. S13). Larger effects might be found if it were technically feasible to mimic the dynamic translocation of focal Btbd7 expression during cleft formation.

Branching morphogenesis of other organs, such as lungs and kidneys, can differ in the relative contributions to branching of bud outgrowth versus cleft formation. Nevertheless, fibronectin contributes to branching morphogenesis in these organs, especially lung (14, 15, 25), suggesting mechanistic commonalities. Although clefts in lung are broad curves rather than slits, Btbd7 is also expressed in lung at sites of epithelial curvature (Fig. S14). Branching morphogenesis of embryonic lungs was similarly inhibited by Btbd7 siRNA but not by control siRNA (Fig. 4G,H). We conclude that Btbd7 is needed for both mammaliansalivary gland and lung branching morphogenesis.

Our data establish a mechanism for cleft propagation involving Btbd7 as a mediator of epithelial dynamics and organ branching. A schematic model of this stepwise mechanism is shown in Fig. S15. The local accumulation of fibronectin in forming clefts rapidly induces the regulator Btbd7, which in turn induces local expression of the cell-scattering gene Snail2 and suppresses E-cadherin levels, thereby altering cell morphology and reducing cell-cell adhesion. This stimulates cell separation at the base of forming clefts by local, dynamic intercellular gap formation and promotes cleft progression. Inhibiting any of these regulatory steps (i.e., fibronectin accumulation or function, Btbd7 expression, or Snail2 expression) inhibits branching morphogenesis. This hierarchy of steps provides a mechanistic pathway for cleft formation in epithelial branching.