Arg kinase regulates epithelial cell polarity by targeting β1-integrin and small GTPase pathways

Abstract

Background: Establishment and maintenance of epithelial cell polarity is regulated in part by signaling from adhesion receptors. Loss of cell polarity is associated with multiple pathologies including the initiation and progression of various cancers. The β1-integrin adhesion receptor plays a role in the regulation of cell polarity; however, the identity of the signaling pathways that modulate β1-integrin function and connect it to the regulation of polarity pathways remains largely unknown.

Results: The present work identifies a role for Arg, a member of the Abl family nonreceptor tyrosine kinases, in the regulation of adhesive signals and epithelial cell polarity. In a three-dimensional cell culture model, activation of Arg kinase leads to a striking inversion of apical-basal polarity. In contrast, loss of Arg function impairs the establishment of a polarized epithelial cyst structure. Activated Arg kinase disrupts β1-integrin signaling and localization and impairs Rac1-mediated laminin assembly. Disruption of β1-integrin function by active Arg results in altered distribution of selected polarity complex components mediated in part by Rap1 GTPase signaling. Whereas polarity inversion is partially rescued by a constitutively active Rap1, Rac1-dependent laminin assembly is not, indicating that Rap1 and Rac1 signal independently during epithelial polarity.

Conclusions: These findings suggest that modulation of Arg kinase function may contribute not only to normal epithelial polarity regulation but also may promote pathologies associated with loss of cell polarity.

Figure 1

Active Arg inverts epithelial cyst polarity. (A) MDCKII cells expressing either empty vector or constitutively-active Arg (ArgPP) were grown in collagen gels for 6 days. The gels were fixed and stained for the apical polarity marker gp135 and the adherens junction marker E-cadherin, and visualized by confocal microscopy. Scale bars, 10μm. (B) Lines were drawn across representative cells in (A) and the fluorescence intensity distribution of the indicated markers along the lines are shown. (C) Western blots show Arg expression and activity by detection of p-CrkL levels. (D) Quantification of the percentage of cysts with inverted polarity at day 6 from 3 independent experiments; over 200 cysts from each group were analyzed by two-tailed unpaired Student's t-test; **, p<0.01. Error bars represent mean ± SD. (E, F) MDCKII cysts expressing either vector or ArgPP were stained for ZO-1 and actin and visualized by confocal microscopy. Fluorescence intensity distribution along the indicated lines are shown. Scale bars, 10μm.

Figure 2

Active Arg inhibits β1-integrin signaling. (A) MDCKII cysts expressing either vector or ArgPP at day 6 were stained for β1-integrin and actin, and visualized by confocal microscopy. Scale bars, 25μm. (B) MDCKII cells expressing vector or ArgPP were plated on collagen for 8 h and adherent cells were quantified. Data were analyzed by two-tailed unpaired Student's t-test; **, p<0.005. Error bars represent mean ± SEM. Vector, n=718; ArgPP, n=150. (C) Cells quantified in (B) were classified into three morphological groups (top). The percentage of cells in each group was analyzed by two-tailed unpaired Student's t-test (bottom); *, p<0.03; ***, p<0.001. Error bars represent mean ± SEM. (D) Cell lysates from MCDKII cells expressing either vector or ArgPP embedded in collagen for 6 to 12 h were subjected to active Rac1-GTP assay. Active Rac1 levels were detected by western blot (left panel) and analyzed by two-tailed unpaired Student's t-test (right panel); ***, p<0.001. Error bar represents mean ± SD. Lysates were also blotted for α3 integrin and tubulin (loading control). (E) 4-day MDCKII cysts expressing either vector or ArgPP were isolated from collagen gels. Expression of cyst-associated laminin was detected by western blotting. VDAC1 was used as a loading control. (F, G) MDCKII cysts expressing vector or ArgPP were stained for gp135 and laminin, and analyzed by confocal microscopy (F). Z-stack pictures were reconstructed to 3D models using the software “Volocity” (G). Scale bars, 10μm.

Figure 3

Disruption of the Par complex by Arg activation. (A, B) MDCKII cysts expressing either vector or ArgPP were stained for Par3 (A) or aPKC (B), and visualized by confocal microscopy. Scale bars, 10μm. (C) Lysates of cells expressing vector or ArgPP were incubated with anti-Par6 antibody, and co-immunoprecipitates were blotted for Par3 and aPKCζ. (D) Quantification of Par3 co-immunoprecipitated with Par6 by ImageJ and analyzed by two-tailed unpaired Student's t-test. *, p<0.04. Error bar represents mean ± SEM.

Figure 4

Loss of Arg function impairs establishment of epithelial cyst polarity while Arg activation inverts cyst polarity. (A, B) MDCKII cells overexpressing either Arg wild-type (WT), constitutively active (PP) or kinase dead (KR) mutants were grown in collagen gels for 6 days. The percentage of cysts with inverted polarity was analyzed by two-tailed unpaired Student's t-test (A); *, p<0.02. ArgWT, n=332; ArgPP, n=195; ArgKR, n=446. Protein expression was analyzed by western blotting (B). (C, D) MDCKII cells expressing either vector or ArgKR were embedded in collagen gels for 8~9 days. Cysts were stained for the indicated markers, followed by confocal microscopy. Scale bars, 10μm (C) and 25μm (D).

Figure 5

Active Arg suppresses collagen-induced β1-integrin protein levels. (A) MDCKII cells were plated on collagen-coated plates for the indicated times and β1-integrin protein levels were analyzed by western blotting and quantified with ImageJ. UN, undetectable. (B) Cells plated on collagen for 24 hours were lysed, subjected to protein fractionation, and the lysates from the cytosolic and membrane/cytoskeleton fractions were analyzed by western blotting with the indicated antibodies. Relative β1-integrin levels were quantified with ImageJ. UN, undetectable. (C) Quantification revealed no statistically significant difference in β1-integrin mRNA levels by real-time RT-PCR in cells expressing either vector or ArgPP before and after plating on collagen for 24 h. Data were analyzed by two-tailed unpaired Student's t-test. p>0.1. Error bars represent mean ± SD.

Figure 6

Regulation of β1-integrin and epithelial cyst polarity by active Arg are mediated by Rap1. (A) Active Arg inhibits Rap1 activation upon β1-integrin engagement. Control or ArgPP-expressing cells were left suspended or allowed to adhere to collagen for 30min. Rap1-GTP pull down assays were carried out and levels of active Rap1 were detected by western blotting (left panel), quantified with ImageJ and analyzed by two-tailed unpaired Student's t-test (right panel); *, p<0.02. Error bar represents mean ± SEM. (B, C) Active Rap1 (RapV12) partially rescues ArgPP-induced polarity inversion. MDCKII cells expressing the indicated proteins were plated on collagen-coated plates for 18h. Protein expression was analyzed by western blotting. Relative β1-integrin levels were quantified with ImageJ. Percentages of Cysts with inverted polarity were quantified and data were analyzed by two-tailed unpaired Student's t-test; *, p<0.04. Error bars represent mean ± SEM. Control, n=550; RapV12, n=600; ArgPP, n=197; ArgPP+RapV12, n=454. (D, E) Active Rap1 rescues polarity inversion but not laminin assembly in ArgPP-expressing cysts. MDCKII cells expressing the indicated proteins were grown in collagen for 7 days, followed by confocal imaging for the indicated markers. Scale bars, 10μm.

Figure 7

ArgPP-dependent regulation of Rap1 and β1-integrin signaling is mediated by CrkII. (A) Active Arg inhibits CrkII-C3G complex formation in response to β1-integrin engagement. Control and ArgPP-expressing MDCKII cells were either left in suspension or allowed to adhere to collagen-coated plates for 25min. The C3G-CrkII interaction was examined by co-immunoprecipitation and western blotting with the indicated antibodies (left panel). S: suspension; A: adhesion. Levels of C3G bound to CrkII in adherent cells were quantified with ImageJ and analyzed by two-tailed unpaired Student's t-test (right panel); **, p<0.01. Error bar represents mean ± SD. (B) CrkII Y221F mutant partially rescues Rap1 activity in ArgPP-expressing cells. MDCKII cells expressing the indicated proteins were plated on collagen for 10 h followed by Rap1-GTP pull-down assays and active Rap1 was detected by western blotting (left panel). The arrow marks the migration of the exogenous CrkII Y221F mutant protein. Active Rap1 levels were quantified with ImageJ and analyzed with two-tailed unpaired Student's t-test (right panel); *, p<0.05; **, p<0.01; NS, not statistically significant. Error bars represent mean ± SEM. (C, D) CrkII Y221F partially rescues ArgPP-induced polarity inversion. MDCKII cells expressing the indicated proteins were plated on collagen for 18h. Protein expression was examined by western blotting and β1-integrin protein levels were quantified by ImageJ. The percentages of cysts with inverted polarity in each experimental group were quantified and analyzed by two-tailed unpaired Student's t-test; **, p<0.01. Error bars represent mean ± SEM. CrkII WT + Vector, n=528; CrkII WT + ArgPP, n=433; CrkII Y221F + Vector, n=548; CrkII Y221F + ArgPP, n=214. (E) CrkII Y221F does not rescue laminin assembly in ArgPP-expressing cysts. Scale bars, 10μm.