TGFβ signaling promotes matrix assembly during mechanosensitive embryonic salivary gland restoration

Abstract

Mechanical properties of the microenvironment regulate cell morphology and differentiation within complex organs. However, methods to restore morphogenesis and differentiation in organs in which compliance is suboptimal are poorly understood. We used mechanosensitive mouse salivary gland organ explants grown at different compliance levels together with deoxycholate extraction and immunocytochemistry of the intact, assembled matrices to examine the compliance-dependent assembly and distribution of the extracellular matrix and basement membrane in explants grown at permissive or non-permissive compliance. Extracellular matrix and basement membrane assembly were disrupted in the glands grown at low compliance compared to those grown at high compliance, correlating with defective morphogenesis and decreased myoepithelial cell differentiation. Extracellular matrix and basement membrane assembly as well as myoepithelial differentiation were restored by addition of TGFβ1 and by mechanical rescue, and mechanical rescue was prevented by inhibition of TGFβ signaling during the rescue. We detected a basal accumulation of active integrin β1 in the differentiating myoepithelial cells that formed a continuous peripheral localization around the proacini and in clefts within active sites of morphogenesis in explants that were grown at high compliance. The pattern and levels of integrin β1 activation together with myoepithelial differentiation were interrupted in explants grown at low compliance but were restored upon mechanical rescue or with application of exogenous TGFβ1. These data suggest that therapeutic application of TGFβ1 to tissues disrupted by mechanical signaling should be examined as a method to promote organ remodeling and regeneration.

Keywords: Basement membrane; Branching morphogenesis; Compliance; Extracellular matrix; Integrin beta 1; Organogenesis; Salivary gland; Transforming growth factor beta.

Figure 1. TGFβ signaling is required for rescue of mechanosensitive SMG development

(A–G) Representative single confocal images captured from the center of the organ explants cultured for 96 hours show representative proacinar morphology, as detected by DAPI staining and ICC for collagen IV (cyan) for basement membrane and SM α-actin (red) for myoepithelial cell differentiation, as detected by ICC in single confocal images. (A) Proacinar structures in SMG grown on permissive (0.5 kPa) gels show uniform epithelial buds with SM α-actin (red) expressed in the outer differentiating myoepithelial cell populations surrounded by basement membrane, indicated by collagen IV (cyan). White asterisks indicate regions where collagen IV in ingressing clefts is associated with SM α-actin. (B) SMG grown on non-permissive (20 kPa) gels exhibit non-uniform distorted epithelial buds with a profound loss of SM α-actin and clefts. (C) Mechanical rescue (MR) restores epithelial bud morphology, clefting, and SM α-actin (red) expressed in the outer differentiating myoepithelial cells. (D–E) Inhibition of TGFβ receptor signaling with SB431542 during the MR reduces levels of SM α-actin and clefts. (F–G) Control IgG does not prevent MR, but inhibition of TGFβ signaling with function blocking antibody during the MR reduces levels of SM α-actin and clefts. n=3 experiments. Scalebar = 50 µm.

Figure 2. TGFβ signaling promotes ECM and basement membrane assembly during rescue of mechanosensitive SMG development

(A–E) Glands were cultured on PA gels +/− rescue conditions, +/− TGFβ1, and +/− SB431542 for 96 hours and were subsequently subjected to DOC-extraction and ICC to detect assembled collagen I (green), collagen IV (cyan), and HSPG (red) in single confocal images. (A) Assembled collagen I (green) in the stroma surrounds the epithelial buds and secondary lobules, or groups of proacini, in glands cultured at permissive (0.5 kPa) compliance. Basement membrane proteins, HSPG and collagen IV, are assembled in uniform layers around proacinar buds. (B) Both primary and secondary networks of assembled collagen I and basement membrane are disrupted in glands cultured at non-permissive (20 kPa) compliance. (C) Addition of 5 ng/ml TGFβ1 to the media of glands grown on 20 kPa gels stimulates a rescue of ECM and BM proteins together with improved proacinar morphology. (D) Mechanical rescue (MR) of the glands demonstrates elevated levels of collagen I in the stroma and increased HSPG and collagen IV encasing proacini. (E) Inhibition of TFGβ receptor signaling with 60 µM SB431542 disrupts the collagen re-patterning, with clefts initiating but not progressing as well as with MR (D). n=3 experiments. Scalebar = 250 µm.

Figure 3. TGFβ signaling promotes cleft progression and elaboration of proacinar structures during rescue

(A–B) Glands were cultured on PA gels +/− rescue conditions and +/− TGFβ1 or SB431542 for 96 hours and subsequently subjected to DOC-extraction and ICC to detect assembled collagen I (green), collagen IV (cyan), and HSPG (red) in single confocal images, as in Figure 2. (A) Multiple clefts ingressing into proacini are apparent. (B) Assembled collagen I at non-permissive (20 kPa) compliance is greatly reduced relative to permissive (0.5 kPa) compliance. Basement membrane assembly around proacinar bud regions is less contiguous at 20 kPa, with clefts failing to progress and divide the enlarged bud region, shown by the white arrows. (C) Addition of 5 ng/ml TGFβ1 to the media of glands grown at kPa, increased levels of assembled stromal collagen I and HSPG and collagen IV in the basement membrane around buds and deeper clefts completing bud division into smaller bud units. (D) Mechanical rescue (MR) restores assembled collagen I in the stroma and HSPG and collagen IV in the basement membrane. (E) Inhibition of TGFβ1 with 60 µM SB431542 disrupts ECM and basement membrane assembly, with failure in cleft progression detected (white arrows). n=3 experiments. Scalebar = 50 µm.

Figure 4. TGFβ receptor signaling promotes progression of basement membrane proteins into bud regions to complete clefting

(A–E) Glands were cultured on PA gels +/− rescue conditions and +/− TGFβ1 or SB431542 for 96 hours under the following conditions: (A) 0.5 kPa, (B) 20 kPa, (C) 20 kPa + 5 ng/ml TGFβ1, (D) Mechanical rescue, defined as growth for 72 hours on 20 kPa gels followed by 24 hours growth on 0.5 kPa gels, and (E) Mechanical rescue + 60 µM SB431542. Glands were subsequently subjected to DOC-extraction and ICC to detect assembled collagen I (green), collagen IV (cyan), and HSPG (red) in sequential single section confocal images. Images are derived from Figure 3, with 1 µm distance between each slice. Dotted boxes surround equivalent clefting regions. Clefts progress and divide the bud into smaller buds when cultured on 0.5 kPa gels, when stimulated with TGFβ1, or when subjected to mechanical rescue (A, C, and D). Growth on 20 kPa gels or inhibition of mechanical rescue with 60 µM SB431542 results in thin, shallow clefts that fail to subdivide the bud (B and E). n=3 experiments. Scalebar = 50 µm.

Figure 5. TGFβ signaling promotes integrin β1 activation during rescue

(A–E) Explants were cultured on PA gels +/− rescue conditions and +/− SB431542 for 96 hours. DAPI (blue), active integrin β1 (cyan), and myoepithelial differentiation (SM α-actin,red) were detected by ICC in single confocal images. Asterisks indicate clefting regions with active integrin β1 and SM α-actin in the region. (A) Epithelial buds are uniform at permissive (0.5 kPa) compliance with SM α-actin (red) restricted to the outside of the proacini. (B) Explants cultured at non-permissive (20 kPa) compliance show less uniform structure with reduced active integrin β1 around bud peripheries and SM α-actin, with a reduction in clefting. (C) Addition of TGFβ1 to glands on 20 kPa gels partially rescues SM α-actin expression in the outer regions of more uniform proacinar buds, along with increased active integrin β1 throughout the epithelial buds. (D) Mechanical rescue (MR) restores robust SM α-actin expression in the outer regions of more uniform proacinar buds, along with increased active integrin β1 at the bud peripheries and in clefts. (E) Inhibition of TGFβ receptor signaling with 60 µM SB431542 during the 24 hour mechanical rescue conditions prevents rescue of the active integrin β1 around proacinar structures and in clefting regions, concomitant with reduction of SM α-actin in clefts. n=3 experiments. Scalebar = 50 µm.

Figure 6. Compliance-dependent activation of integrin β1 within the epithelium is modulated by TGFβ Signaling

(A–E) Explants were cultured on for 96 hours on 0.5 kPa gels, 20 kPa gels or 20 kPa gels + TGFβ1. Representative single confocal images of explants subjected to ICC to detect collagen IV (green) and DAPI (blue), total integrin β1 (red), and active integrin β1 (mAb9EG7) (cyan), with white asterisks denoting clefting regions in the merged panel. (A) In glands at permissive (0.5 kPa) compliance, total integrin β1 was detected in all cells, with active levels preferentially localizing to the bud periphery and clefts in proacini, demonstrated by white asterisks. (B) In explants grown at non-permissive (20 kPa) compliance, total integrin β1 is not obviously altered, while active integrin β1 is reduced around the bud peripheries and in clefting regions. (C) Total integrin β1 is not significantly altered by addition of 5 ng/ml TGFβ1 at 20 kPa, but levels of active integrin β1 are increased in the outer peripheries and in clefting regions. n=3 experiments. Scalebar = 50 µm. (D) Quantification of active versus total integrin β1 reveals that the ratio of active to total integrin β1 is significantly reduced at non-permissive compliance, and is restored by TGFβ rescue. A one-way ANOVA test indicates a significant difference between groups indicated (* p < 0.05), n=14 glands. (E) Western analysis of total integrin β1 from the experimental conditions depicted in A–C, with several bands representing glycosylated forms. Quantification of the western analysis to detect the total integrin β1 levels, with total pixel levels for the top single band and the bottom double bands separately normalized to GAPDH and expressed as arbitrary units (AUs). A one-way ANOVA test indicates no significant difference between any of the conditions (p > 0.05), n=3 experiments.