Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2007 Jun;170(6):2055-67.
doi: 10.2353/ajpath.2007.061277.

Focal Adhesion Kinase Controls pH-dependent Epidermal Barrier Homeostasis by Regulating Actin-Directed Na+/H+ Exchanger 1 Plasma Membrane Localization

Affiliations
Free PMC article

Focal Adhesion Kinase Controls pH-dependent Epidermal Barrier Homeostasis by Regulating Actin-Directed Na+/H+ Exchanger 1 Plasma Membrane Localization

Dusko Ilic et al. Am J Pathol. .
Free PMC article

Abstract

Ubiquitously expressed focal adhesion kinase (FAK), linked to multiple intracellular signaling pathways, has previously been shown to control cell motility, invasion, proliferation, and survival. Using mice with a keratinocyte-restricted deletion of fak (FAK(K5 KO)), we report here a novel role for FAK: maintenance of adult epidermal permeability barrier homeostasis. Abundant lacunae of unprocessed lipids in stratum corneum (SC) of FAK(K5 KO) mice and delayed barrier recovery pointed to malfunction of pH-dependent enzymes active in extracellular space of SC. Measuring the SC pH gradient showed significantly more neutral pH values in FAK(K5 KO) mice, suggesting the importance of FAK for acidification. Moreover, normal functions were restored when FAK(K5 KO) mice were exposed to a surface pH typical of mouse SC (pH = 5.5). Baseline levels and response to barrier disruption of secretory phospholipase A2 isoforms, enzymes that mediate generation of free fatty acids in epidermis, appeared similar in both FAK(K5 KO) and control littermates. We found that the critical SC acidification regulator Na(+)/H(+) exchanger 1 failed to localize to the plasma membrane in FAK-deficient keratinocytes both in vivo and in vitro. Thus, for plasma membrane localization in terminally differentiated keratinocytes, Na(+)/H(+) exchanger 1 requires an intact actin cytoskeleton, which is impaired in FAK-deficient cells.

Figures

Figure 1
Figure 1
The epidermal permeability barrier is abnormal in FAKK5 KO mice. A: Schematic drawing of mouse skin depicting epidermal permeability barrier within SC (inset). B: Barrier recovery, as measured by TEWL, in the first 6 hours after tape stripping is approximately threefold slower in FAKK5 KO mice. Asterisks show significantly different values. n = 5 animals/each time point. Data are presented ± SD.
Figure 2
Figure 2
Defective lipid processing into lamellar bilayers in FAKK5 KO SC is pH-dependent. A: FAKK5 KO mice demonstrate defective SC lipid processing. Electron micrographs of P7 FAKloxP/+ mouse skin compared with FAKK5 KO littermates. RuO4 postfixation. Arrowheads, normal structure of parallel lipid lamellae in normal littermates. Arrows, lacunae of unprocessed lipid in FAKK5 KO mice. B: Decreased lipid processing results in an abnormal increase in the substrate polar lipids. Nile Red staining demonstrates more abundant polar lipids (arrows) in SC of FAKK5 KO mice. C: Impaired lipid processing is due to decreased β-GlcCer’ase activity in FAKK5 KO mice, compared with their FAKloxP/+ littermates. Resorufin-β-d-glucopyranoside was used as a fluorescent substrate. As a control, fluorescence was blocked in the presence of conduritol B epoxide, a high-potency β-GlcCer’ase inhibitor. Decreased β-GlcCer’ase activity in SC of FAKK5 KO mice was reversed by exogenous acidification. D: Decreased β-GlcCer’ase activity is linked to impaired SC acidification. The pH gradient within SC of 4- to 6-month-old FAK K5 KO and FAK loxP/+ mice was determined by measuring surface pH after each of six consecutive strippings with 22-mm D-Squame disks. Comparison of SC surface pH showed significantly (*) higher values in FAKK5 KO mice after three to five strippings. E: Fluorescence lifetime imaging of epidermis in FAKloxP/+ and FAKK5 KO littermates. The decrease in acidity throughout the SC of FAKK5 KO mice confirms the importance of FAK for SC acidification. Fluorescence lifetime imaging measurements were obtained in paired FAKloxP/+ and FAKK5 KO littermates, and measurements of at least five extracellular domains were taken at each depth. Data are presented as the mean ± SEM. *P < 0.05, determined using a Student’s t-test.
Figure 3
Figure 3
Decreased numbers of SC desmosomes are linked with increased SP activity in FAKK5 KO mice. A: Serine protease activity was higher in FAKK5 KO versus FAKloxP/+ SC. BODIPY FL was used as a fluorescent substrate. Activity was blocked by acidification of tissue sections or serine protease inhibitors. B: Reduced number of corneodesmosomes (gray circles) led to corneocyte detachment and cleft formation (*) in SC of FAKK5 KO mice. Black circles, normal-appearing corneodesmosomes in SC of FAKloxP/+ littermates. C: SC of FAKK5 KO mice has significantly (*) fewer corneocyte layers and fewer corneodesmosomes in comparison with FAKloxP/+ littermates. Data are presented as the mean ± SD. *P < 0.05, determined using a Student’s t-test.
Figure 4
Figure 4
Exogenous acidification rescues defective barrier recovery in FAKK5 KO mice. Eight- to 12-week-old FAKloxP/+ or FAKK5 KO mice were subjected to barrier perturbation using tape stripping (see Materials and Methods) and then exposed to either unbuffered solution, as a control, or solution buffered to pH 5.5, to exogenously acidify the SC. TEWL was measured at 3 and 6 hours after tape stripping, and biopsies were obtained after the 6-hour measurement. Biopsies also were obtained at baseline (time 0) as a control. Time 0 animals were not used in subsequent experiments. A: TEWL as a measure of SC integrity showed comparable barrier recovery in FAKK5 KO and FAKloxP/+ mice 3 and 6 hours after disruption when the stripped area was acidified. n = 5 animals/each time point. Data are presented ±SD. Asterisks show significantly different values. B: Processing of lipid lamellae (arrows) was likewise normalized when SC of FAKK5 KO was exposed to HEPES buffer pH 5.5 after barrier break.
Figure 5
Figure 5
Real-time PCR analyses of sPLA2 isoforms expression. A: Baseline level of sPLA2 expressed in epidermis. Data are shown as fold increase/decrease of mRNA levels in skin samples of FAKK5 KO mice (n = 6) versus average of sPLA2 levels in FAKloxP/+ animals (n = 5) using log10 scale. Baseline levels of sPLA2 isoforms were similar in both backgrounds, except sPLA2-IIF and -V, which were increased in FAKK5 KO mice. B: Schematic drawing describing the experiment. C: Induction of sPLA2 isoforms and inflammatory response 9 hours after tape stripping is similar in epidermis of 4- to 6-month-old FAKloxP/+ and FAKK5 KO mice. Data are shown as fold increase/decrease of mRNA levels in samples from tape-stripped skin versus control sample from the same animal using log10 scale. n of FAKloxP/+ = 5; n of FAKK5 KO = 6. D: Real-time PCR analyses of IL-1β and IL-8 expression shown as in C. Induction of these cytokines is a sign of typical inflammatory response that follows the epidermal barrier disruption by tape stripping.
Figure 6
Figure 6
NHE1 plasma membrane localization, not protein levels, differs in epidermis of FAKK5 KO mice. A: NHE1 levels are either slightly higher (2-month-old) or similar (6-month-old) in epidermis of FAKK5 KO mice. Diffuse appearance of NHE1 bands interferes with accurate quantification by densitometry. B: Normal NHE1 plasma membrane localization in upper SG (arrows) of FAKloxP/+ mice is diminished in FAKK5 KO animals (arrowheads).
Figure 7
Figure 7
Deleting FAK in mouse keratinocytes in vitro mimics NHE1 mislocalization phenotype in vivo. A: Cre recombinase encoded by adenovirus excised floxed region of fak and depleted FAK expression 72 hours on transduction. Actin served as loading control. B: Primary p53−/− mouse keratinocytes with (FAKloxP/loxPp53−/−) or without (FAK−/−p53−/−) FAK cultured in 2 mmol/L Ca2+ for 5 days to induce differentiation synthesize similar levels of both an early (involucrin) and late (loricrin) differentiation markers. C: Deletion of FAK in vitro does not change overall NHE1 protein levels in differentiated keratinocytes. Results presented are from three independent experiments. Actin served as loading control. D: In terminally differentiated FAKloxP/loxPp53−/− keratinocytes, NHE1 is primarily localized on plasma membrane (arrows). Cre-mediated deletion of FAK resulted in diminished plasma membrane-localized NHE1 in FAK−/−p53−/− keratinocytes (arrowheads).
Figure 8
Figure 8
NHE1 mislocalization phenotype of FAKK5 KO mouse is reproduced in FAK-depleted primary human keratinocytes. Primary human keratinocytes were left untreated or transduced with a lentivirus encoding GFP (green) and either scrambled sequence or short hairpin RNA FAK (shFAK). Three days later, cells were exposed to 2 mmol/L Ca2+ for 5 days. NHE1 was localized mostly at plasma membrane in untreated (arrowheads) and scrambled sequence/GFP-transduced keratinocytes (short arrows), whereas shFAK/GFP-transduced keratinocytes had very little NHE1 left at the plasma membrane (long arrows).
Figure 9
Figure 9
Intact actin cytoskeleton is required for localization of NHE1 at the plasma membrane. A: Prominent actin stress fibers (arrows) present in primary p53−/− mouse keratinocytes with FAK (FAKloxP/loxPp53−/−) were not detected in keratinocytes without (FAK−/−p53−/−) cultured under the same conditions, in 2 mmol/L Ca2+ for 2 days. B: Disruption of actin stress fibers with 2 μmol/L cytochalasin D interfered with plasma membrane localization of NHE1, but not E-cadherin, in primary p53−/− mouse keratinocytes even when FAK is present (FAKloxP/loxPp53−/−). Colocalization of NHE1 and E-cadherin in FAKloxP/loxPp53−/− keratinocytes cultured for 2 days in medium supplemented with 2 mmol/L Ca2+ and 0.2% dimethylsulfoxide (arrows) or 2 μmol/L cytochalasin D (arrowheads).

Similar articles

See all similar articles

Cited by 15 articles

See all "Cited by" articles

Publication types

Substances

Feedback