ILK mediates the effects of strain on intestinal epithelial wound closure

Abstract

The intestinal epithelium is subjected to repetitive deformation during normal gut function by peristalsis and villous motility. Such repetitive strain promotes intestinal epithelial migration across fibronectin in vitro, but signaling mediators for this are poorly understood. We hypothesized that integrin-linked kinase (ILK) mediates strain-stimulated migration in intestinal epithelial cells cultured on fibronectin. ILK kinase activity increased rapidly 5 min after strain induction in both Caco-2 and intestinal epithelial cell-6 (IEC-6) cells. Wound closure in response to strain was reduced in ILK small interfering RNA (siRNA)-transfected Caco-2 cell monolayers when compared with control siRNA-transfected Caco-2 cells. Pharmacological blockade of phosphatidylinositol-3 kinase (PI3K) or Src or reducing Src by siRNA prevented strain activation of ILK. ILK coimmunoprecipitated with focal adhesion kinase (FAK), and this association was decreased by mutation of FAK Tyr925 but not FAK Tyr397. Strain induction of FAK Tyr925 phosphorylation but not FAK Tyr397 or FAK Tyr576 phosphorylation was blocked in ILK siRNA-transfected cells. ILK-Src association was stimulated by strain and was blocked by the Src inhibitor PP2. Finally, ILK reduction by siRNA inhibited strain-induced phosphorylation of myosin light chain and Akt. These results suggest a strain-dependent signaling pathway in which ILK association with FAK and Src mediates the subsequent downstream strain-induced motogenic response and suggest that ILK induction by repetitive deformation may contribute to recovery from mucosal injury and restoration of the mucosal barrier in patients with prolonged ileus. ILK may therefore be an important target for intervention to maintain the mucosa in such patients.

Fig. 1.

Integrin-linked kinase (ILK) is activated in response to strain. A and B: effect of cyclic strain on ILK kinase activity was assessed by in vitro kinase assay in cells subjected to cyclic strain for 0–60 min. Cells transfected with small interfering (si) RNA (siRNA) to ILK served as a negative control. ILK was activated by strain in a time-dependent fashion in both Caco-2 cells (A) and IEC-6 cells (B) (n = 7, *P < 0.05 for each). Total ILK served as a loading control. Representative Western blot analyses are shown above each graph. C: total ILK protein levels do not increase in response to strain (n = 3, P > 0.05).

Fig. 2.

ILK is required for strain-induced migration. A: cells were transfected with SMARTpool ILK siRNA targeted to human ILK in Caco-2 cells, siILK1, or siILK2 targeted to rat ILK in intestinal epithelial cell-6 (IEC-6) cells, or with control NT1 nontargeting siRNA before lysis and Western blot analysis for ILK protein. Transfection with the ILK siRNA sequence achieved 60–80% reduction in ILK protein levels in both Caco-2 cells and IEC-6 cells as shown by immunoblot. B: ILK reduction by siRNA inhibited strain-stimulated migration of Caco-2 cells. Circular wounds were made in monolayers of cells transfected with either NT1 or siRNA targeted to ILK, and these wounds were photographed. Cells were then cultured under static conditions or conditions of strain for 0–48 h before the holes were measured again and wound closure was calculated. Reduction of ILK reversed the motogenic effect of strain at all time points with maximum inhibition observed at 12 h (n = 17, *P < 0.05). C: a parallel study in IEC-6 cells confirmed that strain stimulated cell migration in cells transfected with control NT1 nontargeting siRNA but not in cells in which ILK was reduced by siRNA (n = 16, *P < 0.05).

Fig. 3.

Phosphatidylinositol-3 kinase (PI3K) is upstream of ILK, whereas Akt is downstream. A: blockade of PI3K with LY294002 prevents strain-induced ILK activation at 5 min (n = 3, *P < 0.05). B: strain stimulated a shift in the p85 subunit of PI3K from a cytosol-enriched fraction to a fraction enriched for cell membranes in an ILK-independent manner, which indicates that ILK does not modulate PI3K activation (n = 3, *P < 0.05). The membrane protein Na-K-ATPase (bottom panel) and the cytosolic protein Rho-GDI (bottom panel) were used to confirm fraction enrichment and equal loading. C: ILK reduction by siRNA inhibited strain-induced Akt phosphorylation. Top, typical blots; bottom, densitometric analysis summary.

Fig. 4.

ILK associates with focal adhesion kinase (FAK) and mediates FAKY925 phosphorylation. A: FAK coimmunoprecipitates with ILK independently of strain (n = 3, P > 0.05). Total Caco-2 cell lysate served as a control for the FAK band. An opposite coimmunoprecipitation of ILK using anti-FAK antibody was also performed, showing ILK coimmunoprecipitates with FAK independently of strain (n = 3, P > 0.05). Total Caco-2 cell lysate served as a control for the ILK band. B: Caco-2 cells were transfected with hemagglutinin (HA)-tagged wild-type or FAK point mutation constructs and subjected to static or cyclic strain conditions before immunoprecipitation of ILK and immunoblotting for HA. The FAK Tyr397Phe mutation did not alter ILK-FAK association (n = 3, *P < 0.05), but the FAK Tyr925Phe mutation reduced ILK-FAK association compared with association of ILK with wild-type FAK (n = 6, *P < 0.05). Equal transfection efficiency was confirmed by immunoblotting of total cell lysates for FAK and HA tag. Total lysates from cells transfected with KH3 control vector served as a control. Total lysates immunoprecipitated with normal mouse IgG also served as a control. C: ILK reduction selectively modulates strain-induced FAK phosphorylation. NT1 control siRNA or ILK siRNA transfected cells were subjected to static versus strain conditions. Lysates were immunoblotted for FAK phosphorylated at Tyr397, Tyr576, or Tyr925. ILK reduction by siRNA prevented strain-induced FAK Tyr925 phosphorylation, whereas strain-induced FAK Tyr397 and FAK Tyr576 phosphorylation were not prevented (n = 7, *P < 0.01). Top, typical blots; bottom, densitometric analysis summary.

Fig. 5.

Src associates with and is required for ILK activation by strain. ILK activation was assessed in the presence of Src reduction by siRNA (A) or Src inhibition by PP2 inhibitor (B). siRNA targeted to Src reduced Src protein by 70 ± 10% (n = 3, P < 0.05; A, bottom two blots, 1 of 3 similar experiments). 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo(3,4-d)pyrimidine (PP2) inhibitor significantly blocked basal as well as strain-stimulated phosphorylation or accumulation of the specific FAK 576 downstream target of the kinase (n = 3, bottom two blots for each, *P < 0.05). Both siRNA to Src and the Src inhibitor PP2 blocked the strain-induced activation of ILK (n = 5, *P < 0.01). C: cells were transfected with siRNA to ILK and immunoblotted for phosphorylated Src after being subjected to static or cyclic strain conditions. Reducing ILK did not alter strain-induced Src phosphorylation (n = 5, *P < 0.05) Top, typical blots; bottom, densitometric analysis summary. D: Caco-2 cells were infected with either Ad-GFP or Ad-GFP, subjected to static or strain conditions, and harvested for ILK in vitro kinase assay. Src activation by Ad-Src increased ILK activity by 435 ± 31% and 408 ± 24% in static cells and strained cells respectively (n = 3, P < 0.05; D). Ad-GFP infection served as control. Equal infection efficiency was confirmed by immunoblotting of total cell lysates for Src and GFP. GAPDH served as a control for equal loading.

Fig. 6.

Strain increased ILK-Src coimmunoprecipitation. However, the strain effect on ILK-Src association is blocked by PP2. Serum-starved cells were pretreated with PP2 or DMSO vehicle control and subjected to static or cyclic strain conditions. An equal amount of whole cell lysate for each condition was immunoprecipitated with anti-ILK antibody before immunoblotting for Src and for ILK. ILK coimmunoprecipitated more Src in strained cells compared with static control cells (bottom left two bars, n = 3, *P < 0.05). PP2 inhibition of Src abolished the strain effect on ILK-Src association (bottom right two bars, n = 3, *P < 0.05). Total lysates from cells transfected with siRNA targeted to Src or with nontargeting NT1 sequences served as controls (12 μg protein/lane). Total lysates immunoprecipitated with normal mouse IgG also served as a control. Top, typical blots; bottom, densitometric analysis summary.

Fig. 7.

Effect of ILK siRNA on myosin light chain (MLC) phosphorylation. Stain-stimulated MLC phosphorylation was prevented by siRNA reduction of ILK in Caco-2 cells (n = 6, *P < 0.05). Top, typical blots; bottom, densitometric analysis summary.

Fig. 8.

Inhibiting Akt, FAK, and MLC prevents strain stimulation of IEC-6 migration across fibronectin. Inhibiting these signals in IEC-6 cells with the Akt inhibitor Akt IV (500 nM), FAK inhibitor (2 μM), or MLC inhibitor ML-7 (10 μM) blocked the motogenic effect of strain compared with wound closure in static cells. Data summarizes results of 25–36 holes per experiment from at least three independent experiments (*P < 0.05).

Fig. 9.

Schematic diagram outlines a signaling pathway that may mediate strain-induced migration in intestinal epithelial cells consistent with our present and previous observations. Cyclic mechanical strain activates phosphatidylinositol 3-kinase (PI3K) and Src, which leads to the activation of ILK. ILK is then required for the activation of Akt, FAK, and MLC, which results in increased proliferation of intestinal epithelial cells.