Keratinocytes from APP/APLP2-Deficient Mice Are Impaired in Proliferation, Adhesion and Migration in Vitro

EXPERIMENTAL CELL RESEARCH 312 (2006) 1939– 1949

available at www.sciencedirect.com

www.elsevier.com/locate/yexcr

Research Article

Keratinocytes from APP/APLP2-deficient mice are impaired in proliferation, adhesion and migration in vitro

Christina Siemesa,1, Thomas Quasta,2, Christiane Kummera,3, Sven Wehnera,4, Gregor Kirfela, Ulrike Müllerb, Volker Herzoga,⁎ aInstitute of Cell Biology and Bonner Forum Biomedizin, University of Bonn, Ulrich-Haberlandstr. 61A, 53121 Bonn, Germany bMax Planck Institute for Brain Research, Deutschordenstr. 46, 60528 Frankfurt, Germany and Institute for Pharmacy and Molecular Biotechnology, University of Heidelberg, Im Neuenheimer Feld 364, 69120 Heidelberg, Germany

ARTICLE INFORMATION ABSTRACT

Article Chronology: Growing evidence shows that the soluble N-terminal form (sAPPα) of the amyloid precursor Received 4 January 2006 protein (APP) represents an epidermal growth factor fostering keratinocyte proliferation, Revised version received migration and adhesion. APP is a member of a protein family including the two mammalian 22 February 2006 amyloid precursor-like proteins APLP1 and APLP2. In the mammalian epidermis, only APP Accepted 23 February 2006 and APLP2 are expressed. APP and APLP2-deficient mice die shortly after birth but do not Available online 11 April 2006 display a specific epidermal phenotype. In this report, we investigated the epidermis of APP and/or APLP2 knockout mice. Basal keratinocytes showed reduced proliferation in vivo by Abbreviations: about 40%. Likewise, isolated keratinocytes exhibited reduced proliferation rates in vitro, APP, β-amyloid precursor protein which could be completely rescued by either exogenously added recombinant sAPPα,orby APLP1, β-amyloid precursor co-culture with dermal fibroblasts derived from APP knockout mice. Moreover, APP- like protein 1 knockout keratinocytes revealed reduced migration velocity resulting from severely APLP2, β-amyloid precursor compromised cell substrate adhesion. Keratinocytes from double knockout mice died like protein 2 within the first week of culture, indicating essential functions of APP-family members for KGF, keratinocyte growth factor survival in vitro. Our data indicate that sAPPα has to be considered as an essential epidermal sAPPα, soluble form of APP growth factor which, however, in vivo can be functionally compensated to a certain extent wt, wild type by other growth factors, e.g., factors released from dermal fibroblasts. EGF, epidermal growth factor © 2006 Elsevier Inc. All rights reserved. TGFα, transforming growth factor-α FGF, fibroblast growth factor

Introduction the outer barrier, protecting the organism against a large variety of environmental stress factors such as physical, chemical and The human epidermis represents the largest organ of the microbial noxious insults [1]. This protective role is accom- human organism covering an area of about 1.8 m2 and forms plished by a series of cooperating mechanisms involving the

⁎ Corresponding author. Fax:+49 228 735302. E-mail address: [email protected] (V. Herzog). 1 Current address: Institute for Cell Biology, ETH Zürich, Hönggerberg, HPM D27, 8039 Zürich, Switzerland. 2 Current address: Institute of Molecular Physiology and Developmental Biology, University of Bonn, Karlrobert-Kreiten-Str. 13, 53115 Bonn, Germany. 3 Current address: The Scripps Research Institute, Department of Cell Biology, 10550 North Torrey Pines Road, La Jolla, CA 92037, USA. 4 Current address: Department of Surgery, Medical School, Sigmund-Freud-Str. 25, 53105 Bonn, Germany.

0014-4827/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2006.02.025 1940 EXPERIMENTAL CELL RESEARCH 312 (2006) 1939– 1949 constant renewal of the epidermis by a continuous cycle of A pharmacological approach to analyze selectively the proliferation, differentiation and shedding of keratinocytes [2], function of sAPPα has become possible by the application of cornified envelope formation [3] and the epidermal resistance hydroxamic-acid-based zinc metalloproteinase inhibitors against physical stress based on the specific cytoskeleton of known to block the release of sAPPα [35]. This inhibition keratinocytes [4] as well as characteristic cell–cell [5] and cell– resulted in a 50–60% reduction of keratinocyte proliferation matrix interactions [6]. Important regulatory functions are and the normalization of the increased growth rates of mediated by ions, cytokines and growth factors [7] such as psoriatic keratinocytes [12]. Corresponding results have been epidermal growth factor (EGF) [8] derived from platelets, reported with progenitor cells of the subventricular zone of the transforming growth factor-α (TGFα) from basal keratinocytes lateral brain ventricle showing a significant decrease in their [8], keratinocyte growth factor (KGF) from dermal fibroblasts [9] growth rates [36]. and activin [10]. More recently, sAPPα, the soluble N-terminal Despite these findings, a potential epidermal phenotype secreted fragment of the amyloid precursor protein (APP), has due to the lack of APP or APLP2 has not been analyzed in detail. been described as an epidermal growth factor due to its In this study, we have, therefore, investigated the epidermis of mitogenic [11,12] and motogenic [13,14] properties. APP/APLP2-deficient mice in situ and in vivo, as well as APP is a member of a larger protein family including the isolated keratinocytes derived from these mice with respect to two amyloid precursor-like proteins APLP1 and APLP2 from their growth rate and motile characteristics. We provide mammals [15]. Both APLPs are highly homologous to APP and evidence for decreased rates in keratinocyte proliferation, proteolytically processed in a similar way, leading to the reduced adhesion to the extracellular matrix and defects in secretion of the large ectodomains (sAPPα and sAPLP) [16,17]. cell migration as a consequence of APP deficiency. These Using in situ hybridization and RT-PCR analysis, it has been functional deficits could be rescued by the addition of demonstrated that APP and APLP2 are expressed ubiquitously exogenous sAPPα. Thus, our observations strongly support in largely overlapping patterns during embryonic develop- the concept that sAPPα acts as an epidermal growth factor ment and in adult tissue [17]. In contrast, APLP1 is found involved in wound healing, reepithelialization and epidermal primarily in the nervous system [18]. Besides its function as a morphogenesis. growth factor, multiple other functions have been proposed for APP, mainly based on in vitro experiments (for review see [19]). According to its structural properties, APP has been Materials and methods suggested to serve as a cell surface receptor involved in signal transduction [20]. Moreover, the cytoplasmic domain which is Animals released by γ-secretase cleavage may function as a transcrip- − − − − − − tion factor [21]. APP has also been recognized as a matrix- Single and double knockout mice (APP / , APLP2 / , APP / / − − binding protein interacting with heparin [22], perlecan [23], APLP2 / ) [34] (Max Planck Institute for Brain Research, laminin [24], collagen type IV [25] and entactin [23]. It has been Frankfurt, Germany) and corresponding wild type control shown that APP is a copper binding protein regulating the mice with the same genetic background as the knockout mice homeostasis of Cu (II) and its reduction to Cu (I) [26]. APP has were used. The animals were housed in a special pathogen- also been proposed to function as a membrane receptor for free unit kept under optimal hygiene conditions. Timed kinesin-I, thereby mediating the axonal vesicular transport in matings were accomplished in the later afternoon followed neuronal cells [27]. Recent observations suggest, however, that by plug check on the next morning. The point of a detected APP does not directly bind kinesin-1 [28]. plug was accounted as embryonic day 0.5 (E0.5). To address the physiological functions of APP in APP- deficient null mutants [29,30] and mice carrying a hypomorphic Materials mutant of APP (APPΔ) [31] have been generated. The phenotypes of these mutants suggested that APP may play a role in Defined serum-free keratinocyte medium (SFM) was obtained neurite outgrowth and the formation of forebrain commis- from Invitrogen Corporation (Karlsruhe, Germany), Dulbecco's sures, postnatal somatic growth and neurobehavioral devel- minimal essential medium (DMEM), keratinocyte growth opment, locomotor activity and grip strength, copper medium (KGM) and trypsin/EDTA solutions were from Cam- homeostasis and the susceptibility to epileptic seizures and brex (Verviers, Belgium). The primary polyclonal rabbit anti- − − excitotoxic agents [29,30,32]. Whereas APP knockout (APP / ) sAPPα antibody 2189 was from Chemicon International − − and APLP2 knockout (APLP2 / ) mice showed only minor (Hofheim, Germany) and the monoclonal anti-actin antibody − − abnormalities, double mutants obtained by crossing APP / (clone C4) was from ICN (Eschwege, Germany). The secondary − − mice to APLP2 / mice were perinatally lethal, suggesting 5-(4.6-dichlorothiazin-2-yl)-amino-fluorescein hydrochloride functional redundancy [33,34]. It has also been shown that (DTAF) or peroxidase-labeled anti-mouse or anti-rabbit anti- APP family members serve essential, at least partially redun- bodies were from Dianova (Hamburg, Germany). dant functions by demonstrating early postnatal lethality for − − − − − − − − both APLP2 / /APLP1 / and APLP2 / /APP / double knockout PCR genotyping mutants. In addition, genetic evidence for distinct physiolog- − − − − − − − − ical roles of APP and APLP2 has been provided by showing that The APP / , APLP2 / and APP / /APLP2 / mice were generated − − crosses of the respective single knockouts with APLP1 / mice and genotyped by PCR as described before [34]. The PCR was result in double mutants of clearly different phenotypes, being performed on DNA obtained from tail biopsies as previously − − − − − − − − either lethal (APLP2 / /APLP1 / ), or viable (APP / /APLP1 / ) [34]. described [31]. For the APP knockout, a 3-primer PCR was used EXPERIMENTAL CELL RESEARCH 312 (2006) 1939– 1949 1941 with the primers UM44 (5′-GAGACGAGGACGCTCAGTCC- Keratinocyte and fibroblast co-cultures TAGGG-3′) and UM42 (5′-ATCACTGGTTCTAATCAGAGGCCC- 3′) against a 650 bp fragment of the wild type allele flanking For co-culture, wild type and knockout keratinocytes were exon 17. For the mutant allele (430 bp fragment), the primers seeded at a density of 1 × 106 cells on 15 mm cover slips coated UM42 located in intron 17 and P3-hygro (5′-CGAGATCAG- with 5 μg/ml collagen IV from human placenta (Sigma-Aldrich) CAGCCTCTGTTCCACA-3′) derived from the PGK-Hygro- and cultured for 3 days in SFM as described above. Fibroblasts cassette were applied. The amplification of the APLP2 knock- were seeded with a density of 1 × 106 cells in 6-well tissue out was performed using the primers APLP2 [1] (5′- culture plates (TPP) and also cultured for 3 days in DMEM. GCCAAGCTTGAGTCGGTGTATCCGTGCT-3′) and APLP2 [2] (5′- Thereafter, the medium of the fibroblasts was changed to KGM GCGACCGGAGGAGACGCACATCGGGAGCTCGCC-3′, 400 bp and the cover slips with the keratinocytes, placed on tissue- fragment) as well as the primers APLP2 [2] and APLP2 [3] (5′- culture inserts (25 mm, LAB-TEK Nalge Nunc International, CCATTGCTCAGCGGTGCTG-3′, 350 bp fragment). The DNA Naperville, USA) with a specific polycarbonate membrane was amplified for 38 cycles at 94°C for 45 s, at 55°C for 45 s (pore size of 3 μm), were put onto the cultured fibroblasts in and at 72°C for 1 min. PCR products were size-separated by the 6-well tissue culture plates. Care was taken that the 2% agarose gel electrophoresis and visualized by ethidium keratinocytes on the cover slips were completely immersed in bromide staining. the medium. This co-culture system was incubated for 24 or 48 h before the proliferation competent keratinocytes were Cell culture stained by BrdU assay as described below.

Keratinocytes from embryonic mouse skin were prepared 1 Preparation of recombinant sAPPα day before birth (day E19.5). The skin was freed of excess dermal tissue, cut into pieces of 5–10 mm2 and washed with RNA was isolated from human brain tissue and reversely serum- and EGF/BPE-free KGM (Cambrex). The skin pieces transcribed. APP cDNAs corresponding to the APP751 and were digested with 10 mg/ml dispase-II (Roche Molecular- APP770 isoforms were amplified by PCR using appropriate Biochemicals, Mannheim, Germany) in KGM for 15 h at 4°C, primers containing a His-tag at the N-terminus. The sAPPα- before epidermis and dermis were separated. Care was taken isoforms were expressed in EBNA293 cells (Invitrogen, Gro- that both the stratum basale of the isolated epidermis and ningen, The Netherlands) and cultured in serum-free DMEM, the dermal papillae of the dermis remained intact, which using the episomal pCEP4 expression vector system (Invitro- was controlled by scanning electron microscopy (not shown). gen) which included the signal peptide sequence from For separation of basal cells, the epidermis was dissociated osteonectin obtained from Dr. Rupert Timpl (Max-Planck with 0.05% trypsin for 6 min at 37°C. The cells were Institute for Biochemistry, Martinsried, Germany). The super- dispersed by pipetting, washed in completely supplemented natants of cells were collected and His-tagged sAPPα was SFM (Invitrogen) containing 10% FCS and 10 μg/ml tetracy- purified using Talon™ metal affinity chromatography (Clon- cline hydrochloride (Sigma-Aldrich Fine Chemicals, Tauf- tech, Heidelberg, Germany). The bound His-tagged isoform kirchen, Germany), separated from the fragments by was eluted by the application of a pH shift (50 mM Na2HPO4, filtration with a cell-strainer (70 μm pore size Nylon gauze; pH 5.0, in 300 mM NaCl) as described before [37] and fractions Becton Dickinson Labware, Heidelberg, Germany), suspended were tested for purity by Coomassie-stained SDS-PAGE and in freshly prepared medium and seeded in culture dishes gel-chromatography using standard procedures. The fractions coated with 5 μg/ml collagen IV from human placenta with highest degree of purity showing only a single band in the (Sigma-Aldrich). The keratinocytes were cultured in SFM SDS-PAGE were combined and dialyzed overnight against PBS with 10% FCS and 10 μg/ml tetracycline hydrochloride for prior to use. 24 h. After this time, the medium was replaced by serum- and tetracycline-hydrochloride-free SFM, which was Proliferation assays changed every 2 days. Under these conditions, the cells maintained their characteristic morphology and their nor- In vitro BrdU assay mal proliferation activity. Because of the decline of the To determine the proliferation rate of keratinocytes in culture, proliferation activity and the increased differentiation of incorporation of 10 mg/ml BrdU (Sigma-Aldrich) in DMEM − − − − keratinocytes from APP / and APLP2 / skin after several without FCS was allowed for 3 h [38]. Thereafter, the cells were passages, all experiments were performed in the zero fixed by immersion in absolute methanol for 10 min at 4°C and passages. air-dried. The DNA was denatured by incubating the cells in Fibroblasts were isolated from whole mount preparations 2N HCl (Roth GmbH, Karlsruhe, Germany) for 60 min at 37°C of the separated dermis. For this purpose, isolated dermis followed by the neutralization of the acid by immersing the pieces of the embryonic mouse skin were placed on 6-well slides for 3 × 10 min in 0.1 M borate buffer (pH 8.5, Sigma- tissue dishes (TPP Tissue Culture Labware, Trasadingen, CH) Aldrich) and 3 × 10 min in PBS. The cells were stained with the and cultured in DMEM with 10% FCS until the fibroblasts monoclonal anti-BrdU antibody (6 μg/ml, Roche-Diagnostics moved out of the dermis pieces. Subsequently, the dermis was GmbH, Mannheim, Germany) and subsequently with the removed and the fibroblasts were cultured until they were secondary DTF-labeled anti-mouse IgG antibody each for used for experiments. 60 min at 37°C. After washing in PBS (3 × 5 min), the cells For experiments, unsupplemented, serum- and EGF/BPE- were mounted in the presence of anti-fade reagent (Bio- free DMEM (DMEM*) or KGM* was used. meda Corporation, Foster City, CA) and viewed with a 1942 EXPERIMENTAL CELL RESEARCH 312 (2006) 1939– 1949 fluorescence microscope (Axiophot, Zeiss, Jena, Germany). the DTAF-labeled goat anti-mouse or anti-rabbit IgG anti- The proliferation-index was calculated from the ratio of BrdU- body in PBS for 60 min at 37°C. After washing 3 × 5 min positive cells and the total number of cells. For morphometric with PBS, the preparations were mounted with anti-fade quantification the imaging software ImageProPlus (Media reagent (Biomeda) and viewed using a LSM510 (Zeiss, Jena, Cybernetics, Silver Spring MD, USA) was applied. Germany).

In vivo BrdU assay Hematoxylin and eosin staining To analyze the proliferation rate of keratinocytes in vivo, − − − − pregnant APP / /APLP2 / and wild type mice were injected 1 To characterize differences in the morphology of the embry- − − − − day before birth and 4 and 2 h before preparation subcutane- onic mouse epidermis by comparing the APP / and APLP2 / ously with 0.5 ml BrdU solution (8 mg/ml in 0.9% NaCl) [39].In with the wild type skin, the isolated pieces of the skin were this assay, a transfer of the BrdU from the pregnant mouse to fixed in 4% paraformaldehyde in PBS for 2 h at room the embryonic DNA was taken advantage of. After 4 h of BrdU temperature, dehydrated in ethanol and embedded in paraf- incubation, the embryos were prepared and 7-μm cryostat fin. Seven-micrometer sections were stained with hematox- sections of the isolated skin were immunofluorescence ylin and eosin using standard procedures. labeled and quantitated as described above. To exclude regional differences in proliferative activity, only skin samples Computer-assisted analysis of keratinocyte migration velocity derived from the same areas of the back were excised. and lamella dynamics (live cell imaging)

Cell adhesion assay SACED assay Freshly isolated keratinocytes were placed at a density of For the analysis of cell-substrate adhesion, freshly isolated 1×105 cells/cm2 in two chambered cover glass culture dishes mouse keratinocytes were seeded at a density of 1 × 105 cells (LAB-TEK Nunc) coated with 5 μg/ml collagen IV from human per well in a 12-well uncoated tissue culture plate (TPP, placenta (Sigma-Aldrich) and cultured for 24 h as described Techno Plastic Products, Trasadingen, Switzerland), each well above. To analyze the effects of the knockout of APP and APLP2 containing a final volume of 500 μl KGM without FCS and EGF/ on lamella dynamics, the keratinocytes were cultured another BPE. After incubation at 37°C for the indicated times, the plates 24 h in KGM* before live cell imaging was started. Microscopy of were swirled three times and the media with the unattached living keratinocytes was performed using the standard soft- keratinocytes were collected. The number of cells was ware of an inverted LSM510 (Zeiss) equipped with a 63 × 1.4 NA determined using a Bürker counting chamber (Marienfeld plan apochromat oil immersion objective and an incubation GmbH, Laboratory Glassware, Lauda-Königshofen, Germany). chamber in which a constant temperature of 37°C was kept. For each time point, two samples were assayed and each Phase contrast images of motile keratinocytes were taken at an experiment was repeated three times. interval of 2 s for 5 min and the lamellopodia protrusion velocity was analyzed by SACED as described before [40]. Lines Immunoelectrophoretic blot analysis extending from inside the cell body to the substrate, crossing the lamella transversally to the cell edge, were placed over the To analyze the expression of APP in keratinocytes, cell- phase contrast image of a motile keratinocyte. Along these lysates were separated on 10% polyacrylamide SDS-gel and lines, distinct gray values of structures with a width of one blotted onto a nitrocellulose membrane (Protan-Nitrocellu- pixel visible in phase contrast in the area of interest were lose, Schleicher-Schuell, Dassel, Germany). Subsequently, the digitally recorded. A time space plot based on 150 images proteins were detected using the rabbit anti-APP antiserum resulted in a stroboscobic image, which allowed the analysis of 2189, visualized by chemiluminescence (ECL, Amersham, lamellopodia protrusion velocity. The assay was performed in Braunschweig, Germany), documented on XAR-5 films triplicate by analyzing 5 cells in each experiment. (Kodak, Stuttgart, Germany) and evaluated by digital image analysis. Time lapse microscopy To determine differences in the migration velocity between Immunocytochemistry knockout and wild type keratinocytes, the isolated cells were seeded in two-chambered cover slip culture dishes and To investigate the expression of APP in the epidermis of the cultured as described above. After incubation for 24 h in embryonic mouse skin and in cultured keratinocytes as well serum- and EGF/BPE-free KGM, phase contrast images of as to characterize the APP knockout, the cryostat sections of motile keratinocytes were captured at an interval of 10 s for the epidermis (7 μm thick) or the cells were fixed in 4% 1 h. The migration velocity of the cells was quantified by paraformaldehyde in PBS (45 min at room temperature) adding the images on top of each other and measuring the followed by treatment with 0.2% Triton X-100 (5 min at distances between the cell nucleus using the imaging software room temperature) and blocking with 3% BSA in PBS (1 h at ImageProPlus (Media Cybernetics, Silver Spring, MD, USA). 37°C). APP was detected by the polyclonal anti-APP antiserum 2189 (90 min at 37°C). To identify the basal lamina, the Statistical analysis cryostat sections were also stained with a polyclonal anti- laminin-5 antibody (Chemicon). For immunofluorescence The results were analyzed for statistical differences by a one- visualization, the sections or the cells were incubated with way analysis of variance (one-way ANOVA). In general, the EXPERIMENTAL CELL RESEARCH 312 (2006) 1939– 1949 1943

− − − − Despite their postnatal lethality, the APP / /APLP2 / embryos exhibited no specific phenotype. About 5–10% of these embryos revealed a dysplasia of the brain (exencephalia), but − − they were also viable until birth [34]. Furthermore, APP / / − − APLP2 / embryos showed no differences concerning the thickness and morphology of the skin compared to the wild type embryos (data not shown), indicating that the knockout of APP and APLP2 may be compensated for by other growth factors and, thus be morphogenetically not essential.

− − − − Decreased keratinocyte proliferation in APP / /APLP2 / epidermis in vivo − − − − Fig. 1 – Genotyping of APP / /APLP2 / mice by genomic PCR −/− of tissue from tail biopsies of the wild type and knockout Potential differences in the proliferation rate of the APP / −/− alleles. KO1 and KO2 (lanes 1 and 2) were heterozygous for APLP2 keratinocytes were analyzed by an in vivo BrdU assay APP, showing the wild type (upper lane) and the APP after subcutaneous injection of the pregnant mice at day E19.5 knockout allele (lower lane) and homozygous for APLP2 and sacrifice of pregnant females 4 h later. BrdU-positive cells − − − (APP+/ /APLP2 / ; lanes 5 and 6). KO3, an APP and APLP2 were detected in cryostat sections of the isolated embryonic − − − − double knockout mouse (APP / /APLP2 / ), exhibited only skin by anti-BrdU specific antibodies (Fig. 2). To localize the the APP and APLP2 knockout allele (lane 3 and 7) whereas basal cell layer, the basal lamina was immunolabeled with − − KO4 was an APLP2 single knockout mouse (APP+/+/APLP2 / ; anti-laminin-5 antibodies. A semi-quantitative analysis of lanes 4 and 8). immunofluorescence micrographs revealed a reduction by about 40% in the number of BrdU-positive keratinocytes in the − − − − skin of APP / /APLP2 / mice (green, Fig. 2b) as compared to wild differences were assumed to be statistically significant at type skin (green, Fig. 2a). As this result points to a role of APP in P < 0.05. the regulation of keratinocyte proliferation, it was of particular interest to study the proliferation rate of isolated keratinocytes. Despite optimal conditions during the separation, isolation − − − − Results and cultivation of the keratinocytes [41], the APP / /APLP2 / cells showed clearly reduced proliferation and died during the In this study, the mitogenic and motogenic effects of APP and first week of culture, which made reproducible experiments − − sAPPα on keratinocytes were determined using APP / and with these double knockout keratinocytes impossible. These − − APLP2 / knockout mice [29,33]. To obtain APP/APLP2 double observations may be explained by the early apoptosis of − − − − knockout (APP / /APLP2 / ) mice, animals homozygous for keratinocytes in culture and the absence of sAPPα as reported − − − APLP2 and heterozygous for APP (APP+/ /APLP2 / )were previously [42]. Therefore, all further studies were performed − − − − intercrossed and expected to yield 25% offspring deficient for on isolated keratinocytes from APP / or APLP2 / mice. These both proteins. Because of early postnatal lethality, embryos cells showed apparently normal growth upon plating and were obtained from the pregnant mice 1 day before birth, at reached confluence after 7 to 10 days. Differences in cell − − day E19.5 and upon tail biopsy genotyped by PCR as APP / / morphology between knockout and wild type keratinocytes − − − − − − − APLP2 / ,APP+/ /APLP2 / and APP+/+/APLP2 / (Fig. 1)as were not observed at this stage (Figs. 3a–c). Upon trypsinization described before [34]. and replating (first passage), knockout keratinocytes adhered,

− − − − Fig. 2 – In vivo BrdU assay of wild type and APP / /APLP2 / mice. The incorporation of BrdU in the embryonic skin resulted from the injection of pregnant mice at day E19.5. (a, b) Immunocytochemical detection of cryostat section showed a clearly reduced labeling of keratinocyte nuclei in the basal layer (BL) indicating a reduced proliferation rate in the double knockout skin (b) compared to the wild type skin (a). Semi-quantitative analysis of the proliferation competent keratinocytes in the sections − − − − revealed a reduction of proliferation by about 40% of the APP / /APLP2 / skin compared to the wild type skin (c) (E, epidermis; D, dermis; green, BrdU; red, laminin-5; n = 10 sections; N = 3 animals; scale bar, 50 μm). 1944 EXPERIMENTAL CELL RESEARCH 312 (2006) 1939– 1949

− − − − Fig. 3 – Keratinocyte culture of wild type and of APP / and of APLP2 / mice. In the zero passages, (P0) keratinocytes of the wild − − − − − − type (a), APLP2 / (b) and APP / (c) mice showed no morphological differences. After the first passages (P1) of the APLP2 / (e) and − − APP / (f) keratinocytes, the cells revealed a high degree of differentiation and only small areas of proliferation active keratinocytes with a high nucleus/cytoplasm ratio compared to the wild type keratinocytes were detectable (yellow arrows and dotted lines) (scale bars, 20 μm). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

but the number of proliferating keratinocytes with a high of sAPPα. Hence, both APP and APLP2 exert mitogenic effects nucleus/cytoplasm percentage was clearly reduced (Figs. 3e–f) on keratinocytes with APLP2 being a slightly weaker mediator. compared to the wild type keratinocytes (Fig. 3d). Since it was It should be pointed out that the effect of rec-sAPPα is not due not possible to separate the proliferation active keratinocytes to its His-tag as untagged sAPPα has been shown to induce from the differentiated cells, all experiments were conducted proliferation as well [43]. in the zero passages. − − Compensation of the APP / keratinocytes by fibroblasts − − Reduced proliferation of APP / keratinocytes is complemented by addition of rec-sAPPα To investigate possible compensatory effects of dermal − − growth factors, we co-cultured APP / keratinocytes with − − − − − − Keratinocytes isolated from APP / or APP2 / embryos were wild type or APP / fibroblasts (Fig. 5a). In this special co- characterized by Western blot analysis (Fig. 4a) and by culture system, the keratinocytes are only in contact with the immunofluorescence microscopy (Figs. 4b–d) using the anti- medium and the substances released by the fibroblasts. The body 2189, which allows to distinguish between APP and quantitation of the BrdU assay showed a significant increase − − − − APLP2. In APP / keratinocytes (Fig. 4c), no endogenous APP of the proliferation rate of the APP / keratinocytes in the − − was detectable, whereas in wild type (Fig. 4b) and APLP2 / presence of wild type fibroblasts (Fig. 5b; P = 0.001). The keratinocytes (Fig. 4d) it was mainly found in the perinuclear proliferation of the knockout keratinocytes was almost − − region displaying the typical staining pattern of the Golgi identically increased in the presence of APP / fibroblasts complex [11]. (P = 0.001), demonstrating that the effect of the APP knockout The proportion of proliferation competent keratinocytes on the proliferation of keratinocytes in vivo might be com- was determined by BrdU assay (Fig. 4e). A significant reduction pensated at least in part by other factors released by fibroblasts. − − (40%; P = 0.005) of BrdU-positive cells was observed in APP / − − − − − − and APLP2 / keratinocytes as compared to wild type cells. Effects of APP / or APLP2 / on migration velocity and This decrease could be abolished when the isolated keratino- lamellipodia dynamics − − − − cytes of the APP / , APLP2 / or wild type embryos were − − incubated in the presence of 10 nM rec-sAPPα for 24 h. APP / To determine the potential motogenic effect of the APP − − and APLP2 / keratinocytes showed significant increases in knockout on keratinocytes, we analyzed different parameters the proliferation rates after application of rec-sAPPα, reaching of the migratory process comprising migration velocity and statistically indistinguishable maximum levels, approximate- lamellipodia dynamics, i.e., their actin assembly-driven pro- ly 20% above the level reached by wild type cells in the absence trusion velocity at the leading edge and their persistence which EXPERIMENTAL CELL RESEARCH 312 (2006) 1939– 1949 1945

− − APLP2 / cells was reduced either but the difference to wild type − − cells was smaller than for APP / cells (Fig. 6c; P = 0.057). This SACED assay [13,40] allows to simultaneously analyze lamellipodia protrusion velocity (data not shown) and lamellipodia persistence (Fig. 7). Quantification of lamellipodia protrusion velocity revealed no significant differences be- − − − − tween the APP / , APLP2 / or wild type keratinocytes. In contrast to lamellipodia protrusion velocity, lamellipodia − − persistence was reduced in APP / keratinocytes, albeit the difference did not reach statistical significance. The lack of APLP2 had apparently no effect on the lifetime of membrane protrusions (Fig. 7). In contrast to wild type keratinocytes, APP- deficient cells formed several lamellar regions and displayed a more random, i.e., less directed, migration.

− − Effects of APP / on the efficiency of adhesion

Since the persistence of lamellipodia has been shown to depend on the efficiency of cell-substrate adhesion, we analyzed the adhesiveness of wild type and double knockout keratinocytes. After the indicated incubation times (30, 60, 120, 180 and 240 min), the number of the non-adherent cells in the supernatant was determined. Based on the seeded cell number, it was possible to deduce the number of adherent cells (Fig. 8). After 30 min, nearly 75% of the seeded wild type − − keratinocytes had adhered, whereas only 1% of the APP / cells became adherent. After 240 min, the number of adherent cells increased for all genotypes reaching almost identical levels. These results point to the importance of APP as an essential factor during the early stages of cell-substrate adhesion thus strongly supporting previous observations on the function of sAPPα in cell adhesion [42,45].

Discussion

It is well-established that the combined deficiency of APP and − − − − −/− −/− Fig. 4 – APP- or APLP2-expression of APP / and APLP2 / APLP2 in mice (APP /APLP2 ) is lethal clearly pointing to the keratinocytes. (a) APP expression was not detectable in the vital importance of these proteins [33,34]. In the normal − − APP / mice (upper blot). Protein loading was monitored by an human [11] and murine [46] epidermis, APP and APLP2 are anti-actin antibody (lower blot). (b–d) Immunocytochemical predominantly expressed in the basal keratinocyte layer, staining of APP confirmed the knockout of APP (c) and which represents the center of proliferation and of cell renewal demonstrated the localization of the endogenous APP in the and provides migratory cells during the process of epidermal − − α perinuclear region of wild type (b) and APLP2 / (d) wound healing. It has been shown that sAPP in human keratinocytes (scale bar, 10 μm). Reduced proliferation of keratinocytes acts as a potent growth factor due to its − − − − APP / keratinocytes is shown in panel e. Isolated APP / , mitogenic activity [11,12] and is also involved in mediating − − APLP2 / or wild type keratinocytes were incubated in cell adhesion [42] and stimulating keratinocyte migration [13]. −/− −/− serum- and EGF/BPE-free KGM with or without 10 nM Surprisingly, APP or APLP2 mice do not exhibit any rec-sAPPα for 24 h. The quantitation of the BrdU-positive obvious morphogenetic epidermal defects. A detailed analysis keratinocytes showed a significant reduction in the knockout of the epidermis in situ and of keratinocytes isolated from −/− −/− keratinocytes (light gray), which was compensated by the APP and APLP2 mice appeared, therefore, necessary. The addition of 10 nM rec-sAPPα (dark gray) (P < 0.05). key finding of our study is that the two epidermal members of the APP family, APP and APLP2, play crucial roles in keratinocyte proliferation, adhesion and migration in culture by serving as a source for the release of sAPPα and sAPLP2 in represents a direct measure for the adhesiveness between the vitro. Both sAPPα and sAPLP2 appear structurally closely cell and the substrate [44]. Evaluation of the locomotion related to each other and, hence, functionally redundant [47]. velocity revealed a significant reduction (P = 0.006) of cell Despite these structural homologies, the biological roles of APP − − migration for APP / keratinocytes (Fig. 6b) as compared to the and APLP2 appear to differ as shown by double mutants of − − − − − − − − wild type keratinocytes (Fig. 6a). The migration velocity of APLP2 / /APLP1 / which are lethal and of APP / /APLP1 / 1946 EXPERIMENTAL CELL RESEARCH 312 (2006) 1939– 1949

− − − − Fig. 5 – Compensation of the reduced APP / keratinocyte proliferation by co-culture with fibroblasts. Keratinocytes from APP / − − mice were placed on tissue culture inserts and cultured with wild type or APP / fibroblasts for 4 h in serum- and EGF/BPE-free − − KGM (a). (b) The quantitation of the BrdU assay showed a significant increase in proliferation of the APP / keratinocytes cultured − − with either wild type or APP / fibroblasts as compared to the keratinocytes without fibroblasts indicating that other growth factors may overcome the proliferation block (n = 3; * = significant to keratinocytes cultured without fibroblasts; P < 0.05).

− − which are viable [34]. Here, we report also on differences in the Constantly, significant results were obtained only for APP / biological roles between sAPPα and sAPLP2 in that sAPLP2 keratinocytes. The lack of sAPPα in keratinocytes may explain exerts in general the same but weaker physiological effects. the survival of keratinocytes for only a few days in culture. Hence, the supplementation with other growth factors including sAPLP2 can only partially compensate the lack of sAPPα. Our in situ observations show that keratinocytes in the − − − − epidermis of APP / /APLP2 / mice exhibit a nearly 40% reduced rate of proliferation. Isolated keratinocytes deficient in both proteins, however, undergo cell death within the first week of culture whereas keratinocytes deficient either in APP or APLP2 were able to survive in vitro for prolonged periods of time. Their growth rates were, nevertheless, strongly reduced − − − − by about 35% in APP / and about 28% for APLP2 / keratinocytes. Keratinocytes of single knockout mice were completely rescued in their growth rates by exogenously added recombinant sAPPα. Interestingly, the situation for double knockout keratinocytes is less severe in situ, as shown by the lack of a specific phenotype. A possible explanation is provided by the − − observation that APP / fibroblasts in co-culture are able to − − normalize the rate of proliferation in APP / keratinocytes,

− − Fig. 6 – Diminished migration velocity of APP / keratinocytes grown in two chambered cover slip culture − − dishes coated with collagen IV. APP / keratinocytes formed several lamellipodia and showed a reduced directional migration (b, arrows) compared to wild type (wt) (a, arrows) − − and APLP2 / keratinocytes (not shown). (c) Quantitation of the migrating cells revealed a significant reduction in the − − migration velocity of APP / and a strong but not significant Fig. 7 – Effect of APP deficiency on lamellipodia dynamics as − − diminishment of APLP2 / keratinocytes compared to wild analyzed by SACED assay. Lamellipodia persistence was − − type cells (n = 3; * = significant to wild type keratinocytes; markedly but not significantly reduced in APP / P < 0.05). keratinocytes (n =3;P < 0.05). EXPERIMENTAL CELL RESEARCH 312 (2006) 1939– 1949 1947

− − migration in APP / keratinocytes was more random, i.e., less directed. This is in great agreement with recent observations that sAPPα is able to induce a directed movement thereby showing its chemotactic potency [13]. The epidermal morphogenetic balance between cell proliferation and cell death is of crucial importance and might be − − − − disturbed in APP / /APLP2 / mice. A striking example of such an imbalance is the underlying cause of hyperproliferative skin diseases. Excessive production and exfoliation of epidermal cells are the hallmarks of such diseases, e.g., psoriasis. Psoriatic keratinocytes have been shown to have strongly increased growth rates which appear, at least in part, to be induced by the mitogenic overproduction of sAPPα [12].

− − Fig. 8 – Delayed adhesion of APP / keratinocytes compared Acknowledgments to wild type keratinocytes. When keratinocytes were seeded on uncoated culture dishes in KGM without FCS and EGF/BPE, We would like to thank Dipl. Biol. Robert Torka for his help the wild type cells (–♦–) became rapidly adherent within the with the time lapse microscopy and Dr. Wolfgang Neumüller first 30 min after seeding (arrow and circles) whereas the − − for critical reading of the manuscript. We also thank Deutsche APP / keratinocytes (–▲–) showed a delayed adhesion which Forschungsgemeinschaft (DFG Research Group Keratinocytes began not until 30 to 60 min after seeding (n = 3; * = significant – Proliferation and Differentiation in the Epidermis“), the to wild type keratinocytes; P < 0.05). “Bonner Forum Biomedizin” and the “Verband der Cher- mischen Industrie” for generous financial support. suggesting reduced apoptosis, enhanced cell size or delayed differentiation in the epidermis of these mice. It remains to be REFERENCES determined which of these possible factors contribute to this effect. We can also not exclude that sAPLP2 is able to [1] K.A. Holbrook, Ultrastructure of epidermis, The Keratinocytes compensate in part for the lack of sAPPα. However, as Handbook, Cambridge University, 1994. discussed before, sAPLP2 exerts minor biological effects and [2] J.C. Jones, A.E. Goldman, H.Y. Yang, R.D. Goldman, The α may support the effect of sAPP but does not appear to be organizational fate of intermediate filament networks in two essential for proliferation. Growth factors released by fibro- epithelial cell types during mitosis, J. Cell Biol. 100 (1985) blasts have been well characterized and include KGF and 93–102. fibroblast growth factor 10 (FGF10). KGF is, similar to sAPPα, [3] A.E. Kalinin, A.V. Kajava, P.M. Steinert, Epithelial barrier known to stimulate migration and to exert a strong cytopro- function: assembly and structural features of the cornified cell envelope, BioEssays 24 (2002) 789–800. tective effect in addition to its mitogenic effects on keratino- [4] P. Cowin, B. Burke, Cytoskeleton–membrane interactions, cytes [48] as well as on a variety of other cell types, e.g., lung Curr. Opin. Cell Biol. 8 (1996) 56–65. epithelial cells, which are protected from hyperoxide-induced [5] V.M. Braga, Cell–cell adhesion and signalling, Curr. Opin. Cell cell death in the presence of this growth factor [49]. The lack of Biol. 14 (2002) 546–556. a specific phenotype in vivo in spite of prominent defects [6] C. Jamora, E. Fuchs, Intercellular adhesion, signalling and the – being observed ex vivo has also been demonstrated for other cytoskeleton, Nat. Cell Biol. 4 (2002) E101 E108. [7] E. Fuchs, S. Raghavan, Getting under the skin of epidermal growth factors. Intriguingly, mice lacking KGF showed, similar morphogenesis, Nat. Rev., Genet. 3 (2002) 199–209. to APP/APLP2 double knockouts, normal epidermal morpho- [8] L.B. Nanney, J.P. Sundberg, L.E. King, Increased epidermal genesis, e.g., during wound repair [50]. growth factor receptor in fsn/fsn mice, J. Invest. Dermatol. 106 The cytoprotective role of sAPPα as shown by the use of (1996) 1169–1174. exogenously added recombinant sAPPα appears to be medi- [9] S. Werner, K.G. Peters, M.T. Longaker, F. Fuller-Pace, M.J. ated by supporting cell adhesion and by inhibition of Banda, L.T. Williams, Large induction of keratinocyte growth apoptosis [42]. Besides the cytoprotective effects of sAPPα, factor expression in the dermis during wound healing, Proc. Natl. Acad. Sci. U. S. A. 89 (1992) 6896–6900. the mechanisms involved in cell adhesion are well known to [10] R. Grose, S. Werner, An aPPArently protective mechanism for be essential for cell migration [13], in particular to establish keratinocytes in wounded skin, Trends Mol. Med. 8 (2002) contact between the newly formed lamellipodia and matrix 149–151. −/− constituents [51]. As a consequence, APP keratinocytes [11] J. Hoffmann, C. Twiesselmann, M.P. Kummer, P. Romagnoli, which display reduced lamellipodia persistence exhibit a V. Herzog, A possible role for the Alzheimer amyloid lowered migratory velocity, which in turn is presumably due precursor protein in the regulation of epidermal basal cell – to the strongly reduced sAPPα levels. These findings corre- proliferation, Eur. J. Cell Biol. 79 (2000) 905 914. [12] C. Siemes, T. Quast, E. Klein, T. Bieber, N.M. Hooper, V. Herzog, spond to former observations on the stimulatory role of sAPPα Normalized proliferation of normal and psoriatic on lamellipodia dynamics and migratory velocity in keratino- keratinocytes by suppression of alpha-secretase, J. Invest. cytes [13]. The importance of sAPPα as a motogenic growth Dermatol. 123 (2004) 556–563. factor is further supported by our findings that keratinocyte [13] G. Kirfel, B. Borm, A. Rigort, V. Herzog, The secretory 1948 EXPERIMENTAL CELL RESEARCH 312 (2006) 1939– 1949

beta-amyloid precursor protein is a motogen for human protein, kinesin-1, and the processing apparatus: revisited, epidermal keratinocytes, Eur. J. Cell Biol. 81 (2002) J. Neurosci. 25 (2005) 2386–2395. 664–676. [29] Z.W. Li, G. Stark, J. Gotz, T. Rulicke, M. Gschwind, G. Huber, U. [14] T. Quast, S. Wehner, G. Kirfel, K. Jaeger, M. De Luca, V. Herzog, Muller, C. Weissmann, Generation of mice with a 200-kb sAPP as a regulator of dendrite motility and melanin release amyloid precursor protein gene deletion by Cre recombinase- in epidermal melanocytes and melanoma cells, FASEB J. 17 mediated site-specific recombination in embryonic stem (2003) 1739–1741. cells, Proc. Natl. Acad. Sci. U. S. A. 93 (1996) 6158–6162. [15] R. Sandbrink, C.L. Masters, K. Beyreuther, Beta A4-amyloid [30] H. Zheng, M. Jiang, M.E. Trumbauer, D.J. Sirinathsinghji, R. protein precursor mRNA isoforms without exon 15 are Hopkins, D.W. Smith, R.P. Heavens, G.R. Dawson, S. Boyce, ubiquitously expressed in rat tissues including brain, but not M.W. Conner, K.A. Stevens, H.H. Slunt, S.S. Sisodia, H.Y. Chen, in neurons, J. Biol. Chem. 269 (1994) 1510–1517. L.H. Van der Ploeg, beta-Amyloid precursor protein-deficient [16] K. Paliga, G. Peraus, S. Kreger, U. Durrwang, L. Hesse, G. mice show reactive gliosis and decreased locomotor activity, Multhaup, C.L. Masters, K. Beyreuther, A. Weidemann, Cell 81 (1995) 525–531. Human amyloid precursor-like protein 1-cDNA cloning, [31] U. Muller, N. Cristina, Z.W. Li, D.P. Wolfer, H.P. Lipp, T. Rulicke, ectopic expression in COS-7 cells and identification of soluble S. Brandner, A. Aguzzi, C. Weissmann, Behavioral and forms in the cerebrospinal fluid, Eur. J. Biochem. 250 (1997) anatomical deficits in mice homozygous for a modified 354–363. beta-amyloid precursor protein gene, Cell 79 (1994) [17] H.H. Slunt, G. Thinakaran, C. Von Koch, A.C. Lo, R.E. Tanzi, S.S. 755–765. Sisodia, Expression of a ubiquitous, cross-reactive homologue [32] P. Tremml, H.P. Lipp, U. Muller, L. Ricceri, D.P. Wolfer, of the mouse beta-amyloid precursor protein (APP), J. Biol. Neurobehavioral development, adult openfield exploration Chem. 269 (1994) 2637–2644. and swimming navigation learning in mice with a modified [18] K. Lorent, L. Overbergh, D. Moechars, B. De Strooper, F. Van beta-amyloid precursor protein gene, Behav. Brain Res. 95 Leuven, H. Van den Berghe, Expression in mouse embryos (1998) 65–76. and in adult mouse brain of three members of the amyloid [33] C.S. von Koch, H. Zheng, H. Chen, M. Trumbauer, G. precursor protein family, of the alpha-2-macroglobulin Thinakaran, L.H. van der Ploeg, D.L. Price, S.S. Sisodia, receptor/low density lipoprotein receptor-related protein and Generation of APLP2 KO mice and early postnatal lethality in of its ligands apolipoprotein E, lipoprotein lipase, APLP2/APP double KO mice, Neurobiol. Aging 18 (1997) alpha-2-macroglobulin and the 40,000 molecular weight 661–669. receptor-associated protein, Neuroscience 65 (1995) [34] S. Heber, J. Herms, V. Gajic, J. Hainfellner, A. Aguzzi, T. Rulicke, 1009–1025. H. von Kretzschmar, C. von Koch, S. Sisodia, P. Tremml, H.P. [19] V. Herzog, G. Kirfel, C. Siemes, A. Schmitz, Biological roles of Lipp, D.P. Wolfer, U. Muller, Mice with combined gene APP in the epidermis, Eur. J. Cell Biol. 83 (2004) 613–624. knock-outs reveal essential and partially redundant [20] J. Rossjohn, R. Cappai, S.C. Feil, A. Henry, W.J. McKinstry, D. functions of amyloid precursor protein family members, Galatis, L. Hesse, G. Multhaup, K. Beyreuther, C.L. Masters, J. Neurosci. 20 (2000) 7951–7963. M.W. Parker, Crystal structure of the N-terminal, growth [35] S. Parvathy, I. Hussain, E.H. Karran, A.J. Turner, N.M. Hooper, factor-like domain of Alzheimer amyloid precursor protein, Alzheimer's amyloid precursor protein alpha-secretase is Nat. Struct. Biol. 6 (1999) 327–331. inhibited by hydroxamic acid-based zinc metalloprotease [21] X. Cao, T.C. Sudhof, A transcriptionally [correction of inhibitors: similarities to the angiotensin converting enzyme transcriptively] active complex of APP with Fe65 and histone secretase, Biochemistry 37 (1998) 1680–1685. acetyltransferase Tip60, Science 293 (2001) 115–120. [36] I. Caille, B. Allinquant, E. Dupont, C. Bouillot, A. Langer, U. [22] D. Schubert, M. LaCorbiere, T. Saitoh, G. Cole, Muller, A. Prochiantz, Soluble form of amyloid precursor Characterization of an amyloid beta precursor protein that protein regulates proliferation of progenitors in the adult binds heparin and contains tyrosine sulfate, Proc. Natl. Acad. subventricular zone, Development 131 (2004) 2173–2181. Sci. U. S. A. 86 (1989) 2066–2069. [37] C.U. Pietrzik, J. Hoffmann, K. Stober, C.Y. Chen, C. Bauer, D.A. [23] S. Narindrasorasak, R.A. Altman, P. Gonzalez-DeWhitt, B.D. Otero, J.M. Roch, V. Herzog, From differentiation to Greenberg, R. Kisilevsky, An interaction between basement proliferation: the secretory amyloid precursor protein as a membrane and Alzheimer amyloid precursor proteins local mediator of growth in thyroid epithelial cells, Proc. Natl. suggests a role in the pathogenesis of Alzheimer's disease, Acad. Sci. U. S. A. 95 (1998) 1770–1775. Lab. Invest. 72 (1995) 272–282. [38] H.G. Gratzner, Monoclonal antibody to 5-bromo- and [24] S. Narindrasorasak, D.E. Lowery, R.A. Altman, P.A. 5-iododeoxyuridine: a new reagent for detection of DNA Gonzalez-DeWhitt, B.D. Greenberg, R. Kisilevsky, replication, Science 218 (1982) 474–475. Characterization of high affinity binding between laminin [39] F. Dombrowski, L. Klotz, H.J. Hacker, Y. Li, D. Klingmuller, K. and Alzheimer's disease amyloid precursor proteins, Lab. Brix, V. Herzog, P. Bannasch, Hyperproliferative Invest. 67 (1992) 643–652. hepatocellular alterations after intraportal transplantation of [25] D. Beher, L. Hesse, C.L. Masters, G. Multhaup, Regulation of thyroid follicles, Am. J. Pathol. 156 (2000) 99–113. amyloid protein precursor (APP) binding to collagen and [40] B. Hinz, W. Alt, C. Johnen, V. Herzog, H.W. Kaiser, Quantifying mapping of the binding sites on APP and collagen type I, lamella dynamics of cultured cells by SACED, a new J. Biol. Chem. 271 (1996) 1613–1620. computer-assisted motion analysis, Exp. Cell Res. 251 (1999) [26] G. Multhaup, A. Schlicksupp, L. Hesse, D. Beher, T. Ruppert, 234–243. C.L. Masters, K. Beyreuther, The amyloid precursor protein [41] R. Caldelari, M.M. Suter, D. Baumann, A. De Bruin, E. Muller, of Alzheimer's disease in the reduction of copper(II) to Long-term culture of murine epidermal keratinocytes, copper(I), Science 271 (1996) 1406–1409. J. Invest. Dermatol. 114 (2000) 1064–1065. [27] A. Kamal, A. Almenar-Queralt, J.F. LeBlanc, E.A. Roberts, L.S. [42] S. Wehner, C. Siemes, G. Kirfel, V. Herzog, Cytoprotective Goldstein, Kinesin-mediated axonal transport of a membrane function of sAppalpha in human keratinocytes, Eur. J. Cell compartment containing beta-secretase and presenilin-1 Biol. 83 (2004) 701–708. requires APP, Nature 414 (2001) 643–648. [43] G.M. Popp, K.S. Graebert, C.U. Pietrzik, S.M. Rosentreter, P. [28] O. Lazarov, G.A. Morfini, E.B. Lee, M.H. Farah, A. Szodorai, S.R. Lemansky, V. Herzog, Growth regulation of rat thyrocytes DeBoer, V.E. Koliatsos, S. Kins, V.M. Lee, P.C. Wong, D.L. Price, (FRTL-5 cells) by the secreted ectodomain of beta-amyloid S.T. Brady, S.S. Sisodia, Axonal transport, amyloid precursor precursor-like proteins, Endocrinology 137 (1996) 1975–1983. EXPERIMENTAL CELL RESEARCH 312 (2006) 1939– 1949 1949

[44] J. Condeelis, Life at the leading edge: the formation of cell [48] S. Werner, Keratinocyte growth factor: a unique player in protrusions, Annu. Rev. Cell Biol. 9 (1993) 411–444. epithelial repair processes, Cytokine Growth Factor Rev. 9 [45] P. Soba, S. Eggert, K. Wagner, H. Zentgraf, K. Siehl, S. Kreger, A. (1998) 153–165. Lower, A. Langer, G. Merdes, R. Paro, C.L. Masters, U. Muller, S. [49] P. Ray, Y. Devaux, D.B. Stolz, M. Yarlagadda, S.C. Watkins, Kins, K. Beyreuther, Homo- and heterodimerization of APP Y. Lu, L. Chen, X.F. Yang, A. Ray, Inducible expression of family members promotes intercellular adhesion, EMBO J. 25 keratinocyte growth factor (KGF) in mice inhibits lung (2005) 3624–3634. epithelial cell death induced by hyperoxia, Proc. Natl. Acad. [46] C. Kummer, S. Wehner, T. Quast, S. Werner, V. Herzog, Sci. U. S. A. 100 (2003) 6098–6103. Expression and potential function of beta-amyloid precursor [50] L. Guo, L. Degenstein, E. Fuchs, Keratinocyte growth factor is proteins during cutaneous wound repair, Exp. Cell Res. 280 required for hair development but not for wound healing, (2002) 222–232. Genes Dev. 10 (1996) 165–175. [47] Q. Li, T.C. Sudhof, Cleavage of amyloid-beta precursor protein [51] G. Kirfel, V. Herzog, Migration of epidermal keratinocytes: and amyloid-beta precursor-like protein by BACE 1, J. Biol. mechanisms, regulation, and biological significance, Proto- Chem. 279 (2004) 10542–10550. plasma 223 (2004) 67–78.