Transmembrane collagen XVII modulates integrin dependent keratinocyte migration via PI3K/Rac1 signaling

Abstract

The hemidesmosomal transmembrane component collagen XVII (ColXVII) plays an important role in the anchorage of the epidermis to the underlying basement membrane. However, this adhesion protein seems to be also involved in the regulation of keratinocyte migration, since its expression in these cells is strongly elevated during reepithelialization of acute wounds and in the invasive front of squamous cell carcinoma, while its absence in ColXVII-deficient keratinocytes leads to altered cell motility. Using a genetic model of murine Col17a1⁻/⁻ keratinocytes we elucidated ColXVII mediated signaling pathways in cell adhesion and migration. Col17a1⁻/⁻ keratinocytes exhibited increased spreading on laminin 332 and accelerated, but less directed cell motility. These effects were accompanied by increased expression of the integrin subunits β4 and β1. The migratory phenotype, as evidenced by formation of multiple unstable lamellipodia, was associated with enhanced phosphoinositide 3-kinase (PI3K) activity. Dissection of the signaling pathway uncovered enhanced phosphorylation of the β4 integrin subunit and the focal adhesion kinase (FAK) as activators of PI3K. This resulted in elevated Rac1 activity as a downstream consequence. These results provide mechanistic evidence that ColXVII coordinates keratinocyte adhesion and directed motility by interfering integrin dependent PI3K activation and by stabilizing lamellipodia at the leading edge of reepithelializing wounds and in invasive squamous cell carcinoma.

Figure 1. Altered cell detachment in the absence of ColXVII.

A, Trypsin-based cell detachment assay. Confluent cell layers of keratinocytes derived from wild type (Ctrl) and Col17a1^{− /−} mice were treated with trypsin/EDTA (0.05%/0.02%) for indicated time points (cells of three individuals per genotype have been analyzed; number of independent measurements  = 5). B, For the centrifugal force-based cell detachment assay cells were allowed to adhere on laminin 332 (LN332) for 10 minutes before measuring the strength of the adhesion at indicated centrifugal forces (cells of three individuals per genotype have been analyzed; number of independent measurements  = 3). For both assays adherent cells were stained with 0.5% crystal-violet, lysed with 1% SDS, and the percentage of adherent cells was determined spectrophotometrically at 540 nm. Data are shown as mean ± SEM; **p<0.01.

Figure 2. Enhanced spreading and actin dynamics in Col17a1^−/− keratinocytes.

A, Keratinocytes derived from wild type (Ctrl) and Col17a1^{− /−} mice were grown on LN332 coated chamber slides for 30 minutes, fixed and processed for indirect immunofluorescence staining with an actin antibody. The insert represents one enlarged Col17a1^{− /−} cell. Arrowheads indicate the presence of multiple lamellipodia visible by tightly dense actin staining. Scale bar  = 10 µm. B, The graph shows the cell size in µm². The size of Col17a1^{− /−} keratinocytes was about 40% larger than that of Ctrl cells (n = 30; cells of two individuals per genotype have been analyzed). Data are shown as mean ± SEM; ***p<0.001. C, Semi-confluent Ctrl and Col17a1^{− /−} keratinocytes were fixed with 4% PFA and incubated with rhodamin-conjugated phalloidin for 1 hour. Phalloidin-staining indicated a higher number of stress fibers and reduced cortical actin in Col17a1^{− /−} keratinocytes.

Figure 3. Upregulation of α6β4 and β1 integrins in Col17a1^−/− skin.

A and B, Skin lysates from WT and Col17a1^{− /−} mice were immunoblotted with indicated antibodies. The graphs combine the quantification for β4 protein expression of four individuals per genotype (A) and β1 protein expression of three individuals (B). *p<0.05. C, Quantitative RT-PCR of primary wild type and Col17a1^{− /−} keratinocytes (cells of four individuals per genotype have been analyzed; number of independent measurements  = 3). Data are shown as mean ± SEM; *p<0.05.

Figure 4. β4 integrin subunit is functionally involved in cell adhesion and spreading of Col17a1^−/− cells.

Col17a1^{− /−} keratinocytes were transduced with empty pLKO vector (mock) and two different shRNA to β4 integrin subunit (β4kd#1 and β4kd#2). A, The cells were lysed and equal amounts of total protein were immunoblotted with indicated antibodies. B, Confluent layers of Col17a1^{− /−} cells were subjected to trypsin/EDTA detachment assay (as described in Figure 1). The percentage of adherent cells are shown as mean ± SEM (number of independent measurements  = 3); *p<0.05. C, Cells were grown on LN332 coated chamber slides for 30 minutes, fixed and processed for indirect immunofluorescence with an actin antibody. The graph shows the cell area in µm² (n = 35). The data are shown as mean ± SEM. *p<0.05. D, Col17a1^{− /−} keratinocytes were treated with DMSO (Ctrl) or different phospho-FAK inhibitors (PF 573228 [5 µM] and Inhibitor 14 [1 µM]) for 6 hours and thereafter subjected to trypsin/EDTA detachment assay. The data are shown as mean ± SEM (cells of three individuals have been analyzed; number of independent measurements  = 5).

Figure 5. Motility of Col17a1^−/− keratinocytes is enhanced via active Rac1.

A, Keratinocytes derived from wild type (Ctrl) and Col17a1^{− /−} mice were grown on glass-bottom culture dishes, and cell migration was recorded by time-lapse imaging. The distance migrated is indicated on the x-axis. The y-axis indicates the processive index (PI) which is defined as the ratio of the distance between the start and end position of a cell to the total distance actually travelled by the cell. B, G-LISA was used to determine Rac1 activity in protein extracts of sub-confluent Ctrl and Col17a1^{− /−} keratinocytes. Equal amounts of total protein were used. Symbols indicate the matched Ctrl/Col17a1^{− /−} littermates (number of independent measurements  = 4); *p<0.05.

Figure 6. Enhanced PI3K signaling in Col17a1^−/− keratinocytes.

A, Keratinocytes derived from wild type (Ctrl) and Col17a1^{− /−} mice were lysed and equal amounts of total protein were immunoblotted with phospho-Akt and total Akt antibodies. The graph shows quantification of phospho-Akt relative to total Akt (cells of three individuals per genotype have been analyzed in 3 independent measurements); *p<0.05. B, Col17a1^{− /−} keratinocytes were grown on glass-bottom culture dishes and treaded with either DMSO or LY294002 [50 µM]. Cell migration was recorded by time-lapse imaging every 5 minutes during 4 hours. The distance migrated is indicated on the x-axis, the processive index (PI) on the y-axis.

Figure 7. Loss of ColXVII leads to EGF-induced S1356 phosphorylation of the β4 integrin subunit.

A, Normal human keratinocytes (NHK) and ColXVII-deficient JEB keratinocytes were lysed and equal amounts of total protein were immunoblotted with the indicated antibodies. The graph combines the quantification of NHKs (four donors) and JEB keratinocytes (four patients). *p<0.05. B, This graph shows the quantification of β4 integrin subunit S1356 phosphorylation in NHKs and ColXVII-deficient JEB keratinocytes relative to total β4 integrin subunit expression levels. *p<0.05. C, Sub-confluent Col17a1^{− /−} keratinocytes with β4 integrin knockdown (β4 kd#1 and β4 kd#2) were treated with recombinant EGF [200 ng/ml] for 10 minutes, lysed and immunoblotted with indicated antibodies. Quantification is shown as mean ± SEM (number of independent measurements  = 3, including two clones with β4 knockdown and a mock control); *p<0.05.

Figure 8. Loss of ColXVII leads to phosphorylation of FAK.

A, Keratinocytes derived from wild type (Ctrl) and Col17a1^{− /−} mice were allowed to adhere to LN332 for 2 hours, lysed and immunoblotted with antibodies to phospho-FAK (Y397) and total FAK. B, Indirect immunofluorescence staining of Ctrl and Col17a1^{− /−} keratinocytes with phospho-FAK (Y397) antibody. Scale bar  = 10 µm. C and D, Col17a1^{− /−} keratinocytes were treated with DMSO (Ctrl) and two different phospho-FAK inhibitors (PF 573228 [5 µM] and Inhibitor 14 [1 µM]) for 6 hours. The cells were lysed and equal amounts of total protein were immunoblotted with antibodies to phospho-FAK (Y397) and total FAK (C, representative immunoblot) and to phospho-Akt and total Akt (D). The graph shows quantification of phospho-Akt relative to total Akt. Individuals are indicated by different symbols (number of independent measurements  = 2); **p<0.01. E, Col17a1^{− /−} keratinocytes with comprising β4 integrin subunit knockdown were treated with DMSO (Ctrl) or phospho-FAK inhibitor (PF 573228 [5 µM] for 6 hours. The cells were lysed and equal amounts of total protein were immunoblotted with antibodies to phospho-Akt and total Akt.

Figure 9. A schematic model of the migratory phenotype of Col17a1^−/− keratinocytes.

In wild type cells the endodomain of ColXVII binds to the intracellular domain of the β4 integrin subunit. In Col17a1^{− /−} keratinocytes, the genetic ablation of ColXVII leads to increased expression and phosphorylation (S1356) of the β4 subunit and to phosphorylation of FAK (Y397), which, in turn, enhance PI3K activity and induce undirected cell migration via Rac1 activation.