Keratinocyte-restricted deletion of ILK causes progressive hair loss. (A) ILK protein level in epidermal lysates of ILK Co and ILK-K5 mice. (B) Back skin of 2-wk-old ILK Co and ILK-K5 animals stained for ILK and α6 integrin. ILK is expressed in basal keratinocytes of the epidermis (E), ORS, HM, DP, arrector pili muscle (AP), and dermis (D). ILK-K5 skin retains ILK expression in DP and dermis but lacks ILK expression in epidermis, HM, and ORS. Bar, 25 μm. (C) Control and ILK-K5 animals at 8 wk of age.

ILK-K5 keratinocytes exhibit reduced lamellipodia stability. (A) Time-lapse microscopy of single control and ILK-K5 keratinocytes. ILK-K5 keratinocytes formed instable lamellipodia, leading to a constant change of direction (green arrows indicate the protrusion and red arrows the retraction of the cell). Single frames chosen from Videos 1 and 2 (available at http://www.jcb.org/cgi/content/full/jcb.200608125/DC1). Bar, 10 μm. (B) Quantification of lamella stability. ILK-K5 keratinocytes exhibit significantly reduced lamella stability. (C and D) Quantification of lamellipodia persistence (C) and lamellipodia frequency (D). Compared with control keratinocytes, ILK-K5 lamellipodia protrusions are significantly less stable and occur more frequently. Error bars in indicate 95% confidence interval of mean values. ***, P < 0.001.

ILK-K5 HFs show normal β-catenin stability and hair-specific differentiation. (A) Control and mutant 2-wk skin sections stained for Ki67 and Lef1 show an increased number of Ki67+ cells in the ORS (arrowheads), yet retained Lef1 expression in the HM and DP of ILK-K5 HFs. (B) Control and mutant 2-wk skin sections stained for β-catenin revealed nuclear β-catenin (arrowheads) in precortical HM and proximal HS cortex in both control and long (▴) and short (▪) ILK-K5 HFs. (C) BatGal reporter mice were intercrossed with ILK-K5 and control animals. LacZ activity is present in precortical HM and HS cortex of both control and ILK-K5 HFs. (D) Immunostaining of K14 for the ORS, of keratin K6irs1 for the IRS, and of keratin hHa1 for HS cortex and α6 integrin. ILK-K5 HFs revealed the presence of a multilayered K14+ ORS. K6irs1 and hHa1 were expressed but mislocalized in short ILK-K5 HFs (arrowheads). Bars, 50 μm.

Loss of ILK impairs migration. (A) Freshly isolated keratinocytes were subjected to migration on LM332 and invasion through Matrigel. ILK-K5 keratinocytes show impaired migration and invasion (mean + SD of three independent experiments; ***, P < 0.001). (B) Time-lapse microscopy of a scratch assay. ILK-K5 keratinocyte exhibit delayed wound closure. Bar, 100 μm. (C) Quantification of the wound closure in the scratch assay. ILK-K5 wound closure is retarded (error bars indicate 95% confidence interval of mean values). (D) Reduced directionality of single ILK-K5 keratinocytes in the leading front of keratinocytes after scratch induction (five representative cells selected out of 40 analyzed for each genotype).

Loss of ILK retards differentiation and disturbs polarity of epidermal keratinocytes. (A) Double immunostaining for K14, K10, or loricrin (Lor) and α6 integrin on back skin of 2-wk-old control and ILK-K5 animals. ILK-K5 epidermis shows several cell layers expressing K14 and loricrin, respectively. Integrin α6 expression is discontinuous in ILK-K5 skin and present on suprabasal cells (asterisks). Bar 25 μm. (B) Immunostaining for F-actin, β-catenin, and plakoglobin in 2-wk-old mouse skin. In control epidermis, F-actin and β-catenin are absent from the basal side of basal keratinocytes. In ILK-K5 epidermis, F-actin and β-catenin are found basally adjacent to nidogen (arrowheads). Plakoglobin localizes to the lateral–apical sides of basal keratinocytes of both control and ILK-K5 mice. Bar, 25 μm. (C) F-actin overlaps with α6 integrin in the mutant HF (arrowheads). Bars, 50 μm. (D) Western blot analysis reveals similar expression levels of E-cadherin and β-catenin in control and ILK-K5 epidermal lysates.

Keratinocyte-restricted deletion of ILK leads to epidermal hyperplasia and epidermolysis and perturbs HF development and growth. (A) Hematoxylin-eosin staining of sections derived from back skin of control and ILK-K5 mice. ILK-K5 mice display stunted HF morphogenesis leading to two HF types (▴, fully developed; ▪, shortened and prematurely arrested), progressive epidermal detachment (asterisks), and dermal pigment deposition (arrowheads). Bar, 100 μm. (B) Epidermis from 2-wk-old ILK-K5 mice is hyperplastic and detached from the underlying dermis (asterisks). ILK-K5 keratinocytes show a flattened morphology. Bar, 25 μm. (C) High magnification of hematoxylin-eosin–stained HFs from 9-d-old back skin. ILK-K5 HFs have multilayered ORS (▴) or show premature growth arrest with loosely attached, malformed DP (arrow; ▪). Bar, 50 μm. (D) ILK-K5 HF growth is perturbed during morphogenesis and cycling. HF lengths of a minimum of 100 HFs per time point are presented as histograms. PC, panniculus carnosum; E, epidermis; D, dermis.

Impaired formation of FAs but normal Rac1 activation in ILK-K5 keratinocytes. (A, left) Immunostaining of primary control and ILK-K5 keratinocytes for talin and F-actin. Control and ILK-K5 keratinocytes contain talin in FAs. ILK-K5 cells are enlarged twofold compared with control cells. Bars, 10 μm. (right) A close up demonstrates talin in FCs at the leading edge of control cells (arrowheads). ILK-K5 keratinocytes have fewer talin-positive FCs at the leading edge (arrowheads). Bars, 2.5 μm. (B) Immunostaining of control and ILK-K5 keratinocytes for paxillin, F-actin, and DAPI. Bars, 10 μm. (C) Quantification of paxillin-containing FAs. ILK-K5 keratinocytes exhibit reduced size and amount of paxillin containing FAs compared with control keratinocytes. Error bars indicate 95% confidence interval of mean values. *, P < 0.05; ***, P < 0.001. (D) Immunostaining of primary control and ILK-K5 keratinocytes for FAK, F-actin, and DAPI. Note that mutant keratinocytes form fewer FAs that are poorly linked to thin and disorganized actin fibers. Bars, 10 μm. (E) Western blot analysis of protein lysates from primary keratinocytes showing reduced FAK phosphorylation levels in the absence of ILK. (F) Western blot analysis of a GTPase pull-down assay showing normal activation of Rac1 in primary ILK-K5 keratinocytes after 30 min of adhesion on a LM322-rich matrix.

ILK ablation impairs keratinocyte adhesion, integrin expression, and BM integrity and alters proliferation in vivo. (A) Cell adhesion of ILK-K5 keratinocytes from 4-d-old mice on FN, Col I, Col IV, and LM332 is significantly reduced compared with control keratinocytes. Adhesion to poly-l-lysine (PLL) is not different (mean + SD of three independent experiments; ***, P < 0.001). (B) Integrin expression in epidermis from 2-wk-old mice. In control skin, β1 and β4 integrins are expressed in basal keratinocytes, whereas in ILK-K5 skin both integrins are also found on suprabasal keratinocytes (arrowheads). In ILK-K5 mice, β4 integrin shows discontinuous staining on basal keratinocytes, LM332 diffuses into the upper dermis (asterisks), and β1 integrin–expressing suprabasal cells retain K14 expression. Bars, 25 μm. (C) Electron micrographs of back skin sections of 2-wk-old control and ILK-K5 mice. Control skin exhibits a continuous lamina densa (asterisks) and hemidesmosomes (arrowhead), whereas ILK-K5 skin shows a discontinuous lamina densa, which is preserved at hemidesmosomes (arrowhead) but largely absent in between. Bar, 0.25 μm. (D) Ki67 staining revealed the presence of proliferating cells in ILK-K5 suprabasal layers. Suprabasal BrdU+ cells express β4 integrin. Bars, 25 μm.

ILK-deficient HFs accumulate proliferating cells in the ORS. (A) Fully developed ILK-K5 HFs display a hyperplastic ORS (++). Both types of ILK-K5 HFs (▴, ▪) show an elevated number of Ki67-positive ORS cells. Auber's line (green line; Auber, 1952) demarks the border between the proliferative and nonproliferative zones of the HM. Bar, 50 μm. (B) Quantification of BrdU+ cells in ILK Co and ILK-K5 HFs as a percentage of total cells in the ORS. The number of proliferating cells in the ORS is significantly increased in mutant HFs. (C) The percentage of proliferating cells in the HM is normal in 7-d HFs but significantly reduced in 14-d ILK-K5 HFs. (D) The overall number of cells is significantly reduced in the HM of 7- and 14-d ILK-K5 HFs. A minimum of 25 HFs were evaluated for B, C, and D at each time point (error bars indicate 95% confidence interval of mean values; *, P < 0.05; ***, P < 0.001). (E) Double immunostaining of CD34 and LM332 on skin sections of 24-d-old animals reveals the presence of a CD34+ bulge region (arrowheads) in ILK Co as well as ILK-K5 HFs. Note that ILK-K5 HFs display a severely abnormal morphology at this stage sometimes with detached DP (bracket). Bar, 25 μm. (F) Double immunostaining of Ki67 and LM332 on skin sections of 24-d-old animals. Proliferating hair germ is formed in HFs with DP and absent in ILK-K5 HFs without DP. HG, hair germ. Bar, 25 μm.