Dispase-Mediated Basal Detachment of Cultured Keratinocytes Induces Urokinase-type Plasminogen Activator (uPA) and Its Receptor (uPA-R, CD87)

EXPERIMENTAL CELL RESEARCH 228, 246–253 (1996)ARTICLE NO. 0323

Dispase-Mediated Basal Detachment of Cultured KeratinocytesInduces Urokinase-type Plasminogen Activator (uPA)

and Its Receptor (uPA-R, CD87)

BIRGIT M. SCHAEFER, JEANNETTE REINARTZ, MICHAEL J. BECHTEL, STEFAN INNDORF,EKKEHARD LANG, AND MICHAEL D. KRAMER1

Laboratory for Immunopathology, University Institute for Immunology, Im Neuenheimer Feld 305, 69120 Heidelberg, Germany

4]. Urokinase-type plasminogen activator (uPA)2 isKeratinocytes synthesize and secrete urokinase- synthesized and secreted by keratinocytes [5] and is

type plasminogen activator (uPA), which is bound in bound in an autocrine manner to a specific receptoran autocrine manner to a specific receptor (uPA-R, (uPA-R, CD87) at their surface [6–9]. Plasminogen,CD87) at their surface. Plasminogen, which is also which is also bound to membrane binding sites [10, 11],bound to membrane binding sites, is readily activated is readily activated by uPA-R-bound uPA [9, 12]. Thus,by uPA-R-bound uPA. Thus, plasmin for proteolysis of plasmin for proteolysis of pericellular glycoproteins ispericullular glycoproteins is provided. While uPA-R provided. While uPA and uPA-R are at low to undetect-and uPA are at low to undetectable levels in keratino- able levels in keratinocytes of the normal epidermis [2,cytes of the normal epidermis, both compounds are 13], both compounds are found in migrating keratino-upregulated in migrating keratinocytes during reepi- cytes during reepithelialization of skin wounds [3, 14,thelialization of epidermal defects and in affected ker- 15] as well as in lesional keratinocytes of bullous der-atinocytes of various epidermal disorders, including matoses [13, 16]. Taken together, there is evidence thatbullous dermatoses. We have hypothesized that the uPA and uPA-R are not expressed in normal epidermisdisturbance of cell/matrix interactions—a common

but in diverse epidermal lesions. Since the factors thatfeature of these diverse pathological situations—in-upregulate these compounds remain elusive, we haveduces uPA/uPA-R. Accordingly, we explored whetherbegun to analyze potential stimuli.the dispase-mediated detachment of cultured kera-

In epidermal lesions, whether induced by mechanicaltinocytes, which have formed a multilayered epider-or physical wounding or by autoimmunological attack,mis-like structure in vitro, induced uPA and uPA-R.the regular pericellular interactions of the affected ker-We found increases in uPA secretion, cell-associatedatinocytes are disturbed. Pericellular interactions in-uPA activity, and uPA- and uPA-R-antigen in keratino-clude not only cell/cell binding but also the interactionscytes upon dispase-mediated detachment from theirbetween keratinocytes and the extracellular matrix.growth substratum. The increase was preceded by an

increase in uPA-R- and uPA-specific mRNA, which was The latter has been addressed in the present study.not observed when the proteinase inhibitor phos- Cell/matrix interactions have a profound influence onphoramidon was added together with dispase. In con- the proliferation and differentiation of epidermal kera-clusion, we present evidence that experimental de- tinocytes [17], while disturbance of the interactions in-tachment with dispase provides signals for the con- duces various alterations of keratinocyte morphologycomitant upregulation of uPA-R and uPA. The findings and function [18].support the hypothesis that cell/matrix interactions uPA and uPA-R are induced when keratinocytes aremay influence the expression of the cell surface-associ- placed into culture [19, 20] and are downregulated inated PA system in human keratinocytes. q 1996 Academic vitro upon differentiation into an epidermis-like struc-Press, Inc. ture [21]. At first approximation, the advanced kera-

tinocyte cultures in which uPA and uPA-R are down-regulated appear as an appropriate model for identi-INTRODUCTION

Plasminogen activation by epidermal keratinocytes2 Abbreviations used: GAPDH, glyceraldehyde-3-phosphate dehy-has been implicated in epidermal (patho)physiology [1– drogenase; PA, plasminogen activator; PAI, plasminogen activator

inhibitor; tPA, tissue-type plasminogen activator; uPA, urokinase-1 To whom correspondence and reprint requests should be ad- type plasminogen activator; uPA-R, urokinase-type plasminogen ac-

tivator-receptor (CD87).dressed. Fax: (//49)6221/564030.

2460014-4827/96 $18.00Copyright q 1996 by Academic Press, Inc.All rights of reproduction in any form reserved.

AID ECR 3353 / 6i15$$$361 10-21-96 02:52:42 ecl AP: Exp Cell

247UPREGULATION OF uPA/uPA-R IN DETACHED KERATINOCYTES

Enzyme-linked immunosorbent assay (ELISA). The enzyme-fying the factors that induce uPA/uPA-R in sedentarylinked immunosorbent assay for the detection and quantification ofkeratinocytes of the normal epidermis. By using suchuPA was performed as previously described [26]. The ELISA for thecultures we have tested whether the detachment of ker- detection of uPA-R was similarily performed: Flat-bottom microtiter

atinocytes from their growth substratum stimulates plates were coated with 100 ml/well of polyclonal goat anti-uPA-RIgG (5 mg/ml) in 50 mM Na2HCO3, 3 mM NaN3, pH 9.6, overnightthe upregulation of uPA and uPA-R. Cultured primaryat 47C. Nonspecific binding sites were blocked with 200 ml/well PBShuman epidermal keratinocytes, which had formed a0.2% (w/v) gelatin for 1 h at room temperature with continuous shak-multilayered epidermis-like structure, were detached ing. The precoated plates were then washed 21 with 200 ml/well of

from their growth substratum by using the proteinase PBS containing 0.05% Tween (PBS/Tween). The test samples diluteddispase, which is known to cause the detachment of in PBS/Tween were added (100 ml/well) and incubated for 1 h at

room temperature with continuous shaking. After the plates hadbasal keratinocytes from their underlying structuresbeen washed 41 using PBS/Tween the mAbs HD-uPA-R 13.1 andwithout disrupting intercellular contacts [22–24]. TheHD-uPA-R 15.4.1 (1/1) [21, 27] diluted to 1 mg/ml were added (100detached cells were analyzed for uPA and uPA-R ex-ml/well) and incubated for 1 h at room temperature with permanent

pression by immunohistological and molecular biologi- shaking. The plates were washed 41 using PBS/Tween, peroxidase-cal methods. labeled goat-anti-mouse IgG Fc diluted 1/5000 in PBS/Tween (100

ml/well) was added, and the plates were incubated for 1 h at roomtemperature. After the plates were washed 41 using PBS/Tween,MATERIALS AND METHODSbound peroxidase was detected by incubation with 100 ml/well of 1mg/ml o-phenylenediamine and 1 ml/ml H2O2 in 0.1 M KH2PO4 buffer,Materials. The plasmin-specific chromogenic substrate H-D-va-pH 6.0. The reaction was stopped after 5 min by adding 100 ml/welllyl-L-leucyl-L-lysine-p-nitroanilide dihydrochloride (S-2251; No.of 1.3 N H2SO4 and the reaction product was quantified by measuring41206) and plasminogen (20 U/mg; No. 41304) were from Haemo-absorbance at 492 nm (reference wavelength 404 nm) using an auto-chrom Diagnostika (Essen, FRG). Dispase II was obtained frommated ELISA reader [26].Boehringer (No. 165 859, Mannheim, FRG), and the dispase inhibitor

Immunofluorescence staining. The detached keratinocyte sheetsphosphoramidon was obtained from Calbiochem (No. 525275, Badwere snap frozen in liquid nitrogen and stored at 0807C until serialSoden, FRG). The monoclonal antibodies against uPA (HD-UK 1)frozen sections of 4–5 mm were cut. The sections were fixed in ace-and uPA-R (HD-UPAR 13.1) have previously been described [21].tone, air-dried, and double-stained with polyclonal and monoclonalBiotin-labeled goat anti-mouse IgG antibodies (No. 115-066-071),antibodies against uPA (HD-UK 1) and uPA-R (HD-UPA-R 13.1), asCy3-labeled rabbit anti-goat IgG antibodies (No. 305-166-046), Cy5-described previously [21]. In brief, the sections were incubated withlabeled streptavidin (No. 016-170-084), peroxidase-labeled goat anti-HD-UPA-R 13.1, which was detected by using biotin-labeled goatmouse IgG (No. 115-035-071), normal goat IgG (No. 005-000-003),anti-mouse IgG antibodies and Cy5-labeled streptavidin. Then, theand normal mouse IgG (No. 0571) were obtained from Dianova (Ham-sections were incubated with the goat anti-uPA antibodies whichburg, FRG). The BCA-Protein kit was obtained from Pierce (No.were visualized with Cy3-labeled rabbit anti-goat IgG antibodies.23235, Bradford, IL). The RNA isolation kit (No. 12163) was obtainedNegative controls were PBS, normal goat IgG, and isotype-matchedfrom Qiagen (Hilden, FRG), and the in vitro transcription kit (No.monoclonal mouse IgG instead of the first antibody. Cross-reactivity0450145) was obtained from Stratagene (Heidelberg, FRG). Protein-of Cy3-labeled rabbit anti-goat antibody with the biotin-labeled goatase K (No. 161519) was from Boehringer, t-RNA (No. R-8508), RNaseanti-mouse IgG sections was excluded by appropriate control experi-A (No. R-9009), and RNase T1 (no. R-7384) were from Sigma (Deisen-ments, i.e., double-stainings were performed with goat IgG or PBShofen, FRG). All other reagents for molecular biological techniquesinstead of the goat anti-uPA antibodies, followed by the rest of theand procedures were purchased from Boehringer. Buffer salts andstaining cascade. In these experiments cross-reactivity of Cy3-la-detergents of analytical grade were from Merck (Darmstadt, FRG).beled rabbit anti-goat with the biotin-labeled goat anti-mouse anti-DMEM and all medium supplements were from Seromed (Berlin,body was not observed. Sections were analyzed with a laser scanFRG). The serum-free HL-1 medium was from Paesel (Hanau, FRG).microscope (Zeiss, Oberkochen, FRG) working with a helium neonPhotographic materials, i.e., Ektachrome 64T color slide films (No.laser operating at a wavelength of 543 nm and a helium neon laser364 4325), NTB-2 emulsion (No. 165 4433), EDF/EDP photochemi-operating at a wavelength of 633 nm. The image documentation wascals (No. 807 7927), and XAR5 films (No. 165 1454), were from Kodakperformed with a screen-photographic system (Focus Graphics, Far-(New Haven, UK).chant, FRG).Cell culture. Human keratinocytes were obtained from skin biop-

Functional determination of PA. The functional activity of thesies by overnight trypsinization at 47C. Cells were cultivated usingplasminogen activators on frozen sections was studied by overlaythe feeder-layer technique according to Rheinwald and Green [25]zymography. The overlay mixture consisted of 0.5 ml 8% (w/v) low-under differentiating conditions in petri dishes (14 cm diameter) forfat dry milk powder solution, 0.75 ml PBS (0.9 mM Ca2/ and 1 mM8 days in DMEM containing 10% fetal calf serum and supplements.Mg2/), 0.7 ml 2.5% agar solution in water, and 20 ml of a 4 mg/mlConfluent cultures were treated for 30 min with dispase II (2.4solution of plasminogen. The overlay mixture was prepared at 507C;mg/ml in DMEM without FCS), washed twice in DMEM, and incu-80 ml was mounted on the cryostat section and spread evenly underbated for different intervals of time (0, 2, 4, 8, 16, and 24 h) inglass coverslips. Zymograms were allowed to develop for 24–48 h atcomplete DMEM. In control experiments the dispase inhibitor phos-377C and analyzed under dark-field illumination. To distinguish uPAphoramidon (100 mg/ml) was added during the dispase treatment.from tPA, amiloride (1 mM), a specific inhibitor of uPA [28], wasCell lysates were obtained by incubating the cells for 1 h at 47C inadded to the substrate. Alternatively, sections were incubated with10 mM Tris/0.01% Triton X-100 in 3 ml lysis buffer. Afterward cellmonoclonal anticatalytic anti-tPA or anti-uPA antibodies (100 mg/debris was removed by centrifugation. To estimate the amount ofml) [29] for 10 min prior to application of the overlay mixture.cellular protein the supernatant was tested by using a BCA Test,

which was performed according to the manufacturer’s instructions. In situ hybridization. The GAPDH, uPA, and uPA-R gene probes(Nos. 57090, 57328 and 65768; ATCC, MD) were separated fromCell lysates were then diluted to a concentration of 1 mg/ml, which

in control experiments was found to correspond toÉ107 cells/ml. The vector DNA by restriction enzyme digestion and preparative agarosegel electrophoresis. The insert DNA fragments were then purifiedsupernatants were collected and stored at 0807C.


248 SCHAEFER ET AL.

using the GeneClean kit (No. 1001-400; Bio 101, CA) according to structure with no major signs of intraepidermal dyshe-the manufacturer’s instructions. sion during the incubation period (Figs. 1A and 1B).

uPA and uPA-R gene fragments were subcloned in pBK-CMV and For immunofluorescence microscopy the frozen sec-used as templates for synthesis of antisense and sense RNA probes,tions were double-stained with antibodies for uPA andrespectively. RNA probes were generated by in vitro transcription

using a commercially available RNA transcription kit according to uPA-R. Immediately after detachment with dispase,the manufacturer’s instructions. Transcription reactions were con- the stainings for uPA and uPA-R were negative or onlyducted in the presence of 35S-UTP (1000 Ci/mmol). Unincorporated faint (Fig. 1C). After incubation of the keratinocyteradiolabel was removed by precipitation with ammonium acetate.

sheets for 4 to 24 h prominent stainings for uPA andFor in situ hybridization, detached keratinocyte sheets were snapuPA-R were observed (Fig. 1D). The most intensivefrozen and stored at 0807C until frozen sections (5–6 mm) were cut.

Frozen sections were fixed in 4% paraformaldehyde in PBS at room stainings were found after 8–16 h of incubation. uPAtemperature. Afterward, they were incubated in proteinase K solu- and uPA-R were stained in basal and in the adjacenttion (1 mg/ml, 377C) for 30 min, fixed once again in 4% paraformalde- suprabasal keratinocytes, whereas the uppermost cellhyde (room temperature, 20 min), and rinsed in PBS followed by

layers remained negative (Fig. 1D). uPA-R stainingacetylation (500 ml acetylanhydride in 200 ml 0.1 M triethanolamine,seemed to be more widespread than uPA-staining, i.e.,pH 8.0, 10 min, room temperature). Hybridizations were conducted

overnight at 457C in the presence of 35S-labeled RNA probes (2 1 more cells stained positive for uPA-R than for uPA.105 cpm/ml) and tRNA (5 mg/106 cpm). The hybridization buffer was According to the double-stainings, uPA and uPA-Rcomposed as follows: 50% formamide, 300 mM NaCl, 20 mM Tris (pH were partially co-localized (Fig. 1D).7.5), 5 mM ethylenediaminetetraacetic acid (pH 8.0), 11 Denhardt’ssolution, 10% dextrane sulfate, and 100 mM DL-dithiothreitol. After

Quantification of uPA and uPA-R in Secretions andhybridization, the sections were rinsed three times in washing buffer(50% formamide, 21 SSC, 1 mM ethylenediaminetetraacetic acid Lysates of Detached Keratinocytes by ELISA(pH 8.0), 10 mM DL-dithiothreitol). Nonspecifically bound RNA probe

Keratinocyte cultures were treated with dispase andwas removed by digestion with RNase A (20 mg/ml) and RNase T1(2 U/ml) for 30 min at 377C. Afterward, the sections were rinsed in the detached keratinocyte sheets were further incu-21 SSC, dehydrated, and exposed to autoradiographic NTB-2 emul- bated for different intervals of time (0–24 h). Condi-sion for 3 weeks. After being developed, sections were counterstained tioned media were collected and cellular lysates werewith calcium red (0.1%), analyzed, and photographed under dark-

prepared at various time points. In the supernatantsfield illumination by using a standard 16 Zeiss microscope and anuPA antigen was undetectable directly after dispaseEktachrome 64T slide film.detachment and increased to a level of 2.2 ng/ml afterNorthern blot analysis. Total cellular RNA was prepared by acid

guanidine thiocyanate–phenol–chloroform extraction [30]. RNA was 24 h. In the cell lysates uPA antigen increased fromquantified by measuring absorbance at 260 nm. RNA (5–10 mg/lane) 0.5 (0 h) to 31.6 ng/ml (24 h). In contrast to uPA, thewas size-fractionated on 1.2% agarose/2.2 M formaldehyde gels, uPA-R antigen was undetectable in the supernatantstransferred to nylon membranes (No. RPN 203N; Amersham,

at any time point tested, while the levels of uPA-R inBraunschweig, FRG) by vacuum blotting, and cross-linked to thethe cell lysates increased from 4.1 (0 h) to 8.8 ng/mlnylon membrane by exposure to ultraviolet light.

Membranes were prehybridized in Church buffer (0.5 M NaPi, pH (24 h) (Table 1).7.2, 0.5 M EDTA, 7% SDS) for 4 h. The DNA probes were labeled with[a-32P]dCTP by using the random primed labeling kit (No. 1004760; Detection of uPA Activity by Overlay ZymographyBoehringer) and hybridized to the immobilized RNA in Church bufferat 677C overnight. The membranes were washed at high stringency The function of uPA in frozen sections was analyzed(21 SSC/1% SDS; 0.21 SSC/1% SDS) and then exposed to Kodak by overlay zymography. Directly after detachment noXAR film at 0707C with an intensifier screen (Dr. Goos-Suprema, lysis of the indicator gel was observed (Fig. 2A). How-Heidelberg, FRG). Autoradiograms were scanned with an Image

ever, lysis was observed in frozen sections of keratino-master scanning system (No. 56-1153-24; Pharmacia, Freiburg,FRG). Densitometric measurements for uPA and uPA-R bands were cyte sheets, which had been further incubated for atnormalized to the corresponding GAPDH band. least 4 h after detachment (Fig. 2B). Lysis was plasmin-

ogen dependent and was inhibited by the uPA-specificinhibitor amiloride and by anticatalytic anti-uPA anti-RESULTSbodies, but not by anticatalytic anti-tPA antibodies.Taken together, overlay zymography revealed a time-

Histological and Immunohistological Analysis of dependent upregulation of uPA activity in detachedDispase-Detached Keratinocyte Sheets keratinocyte sheets. Thus, the data were in agreement

with the immunohistological demonstration of the uPANHEK were cultured for 8 days under differentiatingantigen.conditions and were then detached from the culture

substratum by dispase treatment and further incu- Detection of uPA- and uPA-R-Specific mRNA by inbated for different intervals of time (0–24 h). Frozen Situ Hybridizationsections of the keratinocyte sheets were prepared andanalyzed by conventional microscopy. The H&E- In situ hybridization of keratinocyte sheets directly

after detachment revealed only decent signals for uPA-stained sections revealed a multilayered epidermis-like



FIG. 1. Histology and immunofluorescence staining of detached keratinocyte sheets. Cultured keratinocytes were detached from theculture substratum by dispase treatment. Detached keratinocyte sheets were washed twice and snap frozen in liquid nitrogen instantlyafter dispase treatment or incubated for further 8 h in DMEM. H&E staining of frozen sections revealed a multilayered epidermal sheet(A) with no major sign of intraepidermal dyshesion after further incubation (B). Immunofluorescence staining for uPA (green color) anduPA-R (red color) directly after dispase treatment (C) or after 8 h of further incubation (D). After 8 h of further incubation the uPA-/uPA-R-specific stainings were confined to basal and adjacent keratinocytes of the sheet; comparing uPA staining with the uPA-R staining, moresuprabasal keratinocytes expressed uPA-R than uPA; uPA and uPA-R were partially co-localized (D). Original magnification (A–D) 2501;dotted lines indicate the basal pole of keratinocyte sheets.

specific mRNA and minor signals for uPA-R-specific Detection of uPA- and uPA-R-Specific mRNA byNorthern BlottingmRNA. The signals were mainly localized to the basal

cells (Fig. 2C and 2E). However, prominent signalsTotal RNA was isolated from untreated adherent andwere found in keratinocyte sheets which had been fur-

dispase-detached keratinocyte cultures after differentther incubated after detachment. The signals were lo-time points (0–24 h) and analyzed by Northern blottingcalized to basal and adjacent suprabasal keratinocytesfor the expression of uPA- and uPA-R-specific mRNA.(Figs. 2D and 2F; after 8 h of incubation). The upper-The specific mRNA levels were normalized for themost keratinocyte layers did not display uPA- or uPA-

R-specific signals. amount of GAPDH-specific mRNA. Low-level constitu-


250 SCHAEFER ET AL.

TABLE 1 induces uPA/uPA-R in keratinocytes adjacent to the‘‘wound’’ [6, 32]. Our findings extend these previousDetermination of Secreted and Cell-Associated uPA anddata by showing that ‘‘enzymatic wounding’’ of kera-uPA-R in Dispase-Detached NHEKs by ELISAtinocyte layers at their basal cell pole induces not only

Supernatant Lysate uPA but concomitantly also the uPA-R which is mostpertinent for directing plasminogen activation to theTime (h) uPA uPA-R uPA uPA-Rcell surface, a privileged site for generation and main-

0 — — 0.5 4.1 tenance of plasmin activity [6– 9, 20].2 0.3 — 5.0 4.8 In vivo, the expression of uPA and uPA-R is confined4 0.3 — 15.8 6.0 to migrating keratinocytes of the outgrowing epidermal8 0.4 — 22.8 5.7 sheet during reepithelialization, rather than to seden-16 1.8 — 24.6 6.5

tary keratinocytes of the normal epidermis [3, 14, 15].24 2.2 — 31.6 8.8uPA and uPA-R have also been demonstrated in lesions

Note. NHEK cultures were incubated for 30 min with dispase; the of various epidermal disorders, in particular bullousdetached sheets were washed twice and incubated further in DMEM. dermatoses [1, 4, 5, 13, 16]. Given the diverse natureAfter the indicated intervals of time, the conditioned media were

of these disorders it is unlikely that uPA and uPA-Rcollected and cell lysates were prepared. The supernatants and ly-are involved in lesion initiation; it is more likely thatsates were analyzed for uPA and uPA-R by using specific ELISAs.

The data represent the mean concentrations (ng/ml) of six replicate their expression represents a regulated epidermal reac-determinations. The standard deviations were consistently less than tion to the disease process. The key modulators of epi-10% of the mean value and are not included in the table. The detec- dermal uPA/uPA-R expression under pathological con-tion limits of the respective ELISAs wereÉ0.1 ng uPA/ml and É3 ng

ditions has remained elusive so far. According to ouruPA-R/ml). (—) Concentration below detection limit of the respectivedata, the proteolytic disturbance of cell/matrix bind-ELISA.ing—a common feature of epidermal disorders [33]—appears as an inducing event.

The role of uPA/uPA-R in the pathogenesis of epi-tive expression of uPA- and uPA-R-specific mRNA was dermal lesions, in particular in bullous dermatoses,observed in the untreated cultures (Fig. 3). Detach- remains unclear. Although our studies indicate thatment with dispase induced a substantial time-depen- expression of both compounds is unlikely to play adent increase in the relative amount of uPA- and uPA- role in lesion initiation, it may contribute to cuta-R-specific mRNA levels. The increase was already ap- neous damage and the maintenance of inflammatoryparent after 2 h (Fig. 3) and was maximal 8 h after activity through the generation of plasmin. This isdetachment. At this time point the uPA-specific mRNA also supported by our previous finding that plasminwas increased 77-fold and the uPA-R-specific mRNA generated at the cell surface interfered with the ad-18-fold. At later time points the amounts of uPA- and hesion of cultured keratinocytes to their growth sub-uPA-R-specific mRNA decreased. For control purposes, stratum [11]. Moreover, enzyme inhibition studieskeratinocyte cultures were treated with dispase in the [32] have indicated that uPA is involved in keratino-presence of the proteinase inhibitor phosphoramidon. cyte migration during wound healing; by uPA expres-In these culturesú90% keratinocytes remained adher- sion these keratinocytes appear to be enzymaticallyent (data not shown) and only minor or no increase of equipped to break their path through the glycopro-the relative amount of uPA- and uPA-R-specific mRNA tein-rich, in particular fibrin-rich, provisional matrixwas observed (Fig. 3). The latter findings indicated that of a wound bed [3, 6, 14, 32].the induction of uPA- and uPA-R-specific mRNA de- In epidermal lesions in vivo, the following sequencepended on the proteolytic activity of dispase. of events may be postulated: epidermal injury and

disturbance of the keratinocyte/matrix interaction,DISCUSSION whether caused by physical trauma, by autoimmuno-

logical attack, or by the genetic defect of cohesive struc-tures, provide a primary stimulus for the induction ofIn the present study we have shown that enzymatic

detachment of keratinocytes from their culture sub- an epidermal repair reaction that includes the concomi-tant upregulation of uPA and uPA-R. In wound healingstratum provides a strong stimulus for expression of

uPA and uPA-R (see Fig. 3). Since the protease inhibi- this may promote the migratory activity of affected ker-atinocytes, whereas in pathological situations this maytor phosphoramidon, a known inhibitor of dispase ac-

tivity [31], blocked the stimulation we concluded that add to the catabolic pathological events in the lesion.Dispase is a neutral protease prepared from Bacillusthe inducing signal depended on proteolysis. The pres-

ent results are reminiscent of the previous finding that polymyxa, which acts not only on glycoproteins, suchas fibronectin, but also on collagenous molecules, in-mechanical wounding of confluent keratinocyte layers



FIG. 2. Overlay zymography and in situ hybridization of detached keratinocyte sheets. Frozen sections of keratinocytes directly afterdispase treatment (A, C, E) and 8 h after further incubation in DMEM (B, D, F) were analyzed by overlay zymography (A, B) and in situhybridization (C–F). Overlay zymography did not reveal lysis of the casein gel in epidermal sheets directly after dispase treatment (A),whereas after 8 h of further incubation lysis was observed (B). In situ hybridization revealed weak signals for uPA (C) and uPA-R (E)mRNA directly after dispase treatment. After a further 8 h of incubation in DMEM strong signals for uPA (D) and uPA-R (F) mRNA wereobserved. Original magnification (A–F) 1001. A–F were photographed using dark-field illumination; C*–F* represent the photographscorresponding to C–F in transmission light illumination. Dotted lines indicate basal poles of keratinocyte sheets.

cluding the basement membrane type IV collagen [24] tinocyte/fibroblast co-cultures contain not only fibro-nectin but also type IV collagen and laminin which areand the hemidesmosomal protein HD4 [22]. The latter

is a major transmembrane glycoprotein with collagen produced to some extent by the keratinocytes them-selves but to a greater extent by the 3T3 feeder fibro-domains in its extracellular portion [22]. In the skin

dispase induces epidermo–dermal separation [24] by blasts. Marchisio and colleagues [36, 37] found thatfibronectin was located in apposition with the periph-acting on cohesive structures of the epidermal base-

ment membrane, while the desmosomes and stratified eral row of keratinocytes in exponentially growing colo-nies, but was missing from confluent colonies, whileepithelial characteristics stay intact [24]. Morphologi-

cal analyses revealed (i) that epidermo–dermal separa- laminin and type IV collagen were also found in con-fluent colonies underneath the ventral membrane oftion is induced at the lamina densa, most likely by

enzymatic dissolution of type IV collagen [24], and (ii) individual cells. Since laminin is resistant to cleavageby dispase, it might be the proteolytic activity on typethat the hemidesmosomes, the cellular organelles

which mediate firm epidermo–dermal contact [34], are IV collagen that accounts for detachment. Further-more, it has previously been reported that functionalaffected. The latter might reflect the ability of dispase

to degrade the extracellular domain of the hemidesmo- desmosomes are rather unlikely to be produced in clas-sical keratinocyte/fibroblast co-cultures [38]. There-somal protein HD4 [22].

Alitalo and co-workers [35] have shown that kera- fore, the hemidesmosome splitting activity of dispase


252 SCHAEFER ET AL.

Given the previous observations that expression ofuPA/uPA-R may aid in keratinocyte migration [3, 14,15, 32], it is tempting to speculate that uPA/uPA-Rmay also be involved in this epidermal self-remodelingprocess.

Finally, there is evidence for functions of the uPA/uPA-R in addition to surface-associated plasminogenactivation: the interaction between uPA and the uPA-R stimulates the proliferation of [39] and provides achemotactic stimulus [40] for epidermal cells. The vari-ous functions, i.e., increased proliferation and chemo-tactically guided migration, as well as plasmin-medi-ated breakdown of extracellular fibrin and other glyco-proteins, appear as part of the keratinocyte’s biologicalresponse to epidermal injury in vivo. Our findings allowone to hypothesize that impairment of cell/matrix bind-ing may trigger this cell biological program via induc-tion of uPA-R and uPA.

The authors are indebted to Dr. Herbert Spring (DKFZ, Heidel-berg) for help with the laser scan microscopy (KBF Grant 92011060to DKFZ) and to Mrs. Dorothea Schwarz and J. Reuland for helpwith the photographic work. We gratefully acknowledge the help ofDr. Christoph Mueller (Bern) in introducing the in situ hybridizationtechnique in our laboratory. This work was supported by DFG Kr

FIG. 3. Northern analysis of uPA- and uPA-R-specific mRNA 931/3-2 and PUG U 95001.levels in untreated, dispase-treated, or dispase/phosporamidon-treated NHEKs. NHEKs were incubated for 30 min with dispase (—)

REFERENCESor with dispase and phosphoramidon (-- -), a dispase inhibitor, whichblocked detachment of keratinocyte sheets almost entirely (ú90%).The detached and nondetached sheets were washed twice and incu- 1. Morioka, S., Singer, H. S., Hashimoto, K., Jensen, P. J., andbated further in DMEM. Total RNA was prepared after the indicated Lazarus, G. S. (1986) Rec. Adv. Dermatol. 7, 53–67.intervals of time. The blots were hybridized with [32P]uPA cDNA (l) 2. Grøndahl-Hansen, J., Ralfkiaer, E., Nielsen, L. S., Kristensen,or [32P]uPA-R cDNA (j), cleaned, and rehybridized with [32P]GAPDH P., Frentz, G., and Dano, K. (1987) J. Invest. Dermatol. 88, 28–cDNA. The autoradiograms of representative Northern blots are 32.shown at the top. The line graph was derived from the uPA (l) or

3. Grøndahl-Hansen, J., Lund, L. R., Ralfkiaer, E., Ottevanger,uPA-R (j) and GAPDH densitometric measurement for each timeV., and Dano, K. (1988) J. Invest. Dermatol. 90, 790–795.point and represents duplicate experiments. Data are depicted as

4. Jensen, P. J., Baird, J. M., Morioka, S., Lessin, S., and Lazarus,the ratio of mRNA content in dispase- or dispase/phosphoramidon-G. S. (1988) J. Invest. Dermatol. 90, 777–782.treated cultures and the mRNA level in untreated cultures.

5. Morioka, S., Jensen, P. J., and Lazarus, G. S. (1985) Exp. CellRes. 161, 364–372.

6. McNeill, H., and Jensen, P. J. (1990) Cell Res. 1, 843–852.does apparently not account for the detachment of cul- 7. Del Rosso, M., Fibbi, G., Dini, G., Grappone, C., Pucci, M., Cal-tured keratinocytes. However, the exact targets that dini, R., Magnelli, L., Fimiani, M., Lotti, T., and Panconesi, E.

(1990) J. Invest. Dermatol. 94, 310–316.are cleaved and that result in upregulation of uPA/uPA-R remain to be explored in future detailed studies. 8. Stoppelli, M. P., Tachetti, C., Cubellis, M. V., Corti, A., Hearing,

V. J., Cassani, G., Appella, E., and Blasi, F. (1986) Cell 45, 675–These studies should include more refined (possibly684.nonenzymatic) methods of interference with the kera-

9. Reinartz, J., Link, J., Todd, R. F., and Kramer, M. D. (1994)tinocyte growth substratum interaction.Exp. Cell Res. 214, 486–498.The disruption of epidermo–dermal binding induces

10. Isseroff, R. R., and Rifkin, D. B. (1983) J. Invest. Dermatol. 80,numerous alterations of the cellular phenotype and 297–299.function, which are thought to contribute to the repair

11. Reinartz, J., Batrla, R., Boukamp, P., Fusenig, N. E., andof the epidermal defect. Poumay and co-workers [23] Kramer, M. D. (1993) Exp. Cell Res. 208, 197–208.have analyzed detached epidermis obtained by dispase 12. Ellis, V., Behrendt, N., and Dano, K. (1991) J. Biol. Chem. 266,treatment of skin biopsies. They observed a spatial re- 12752–12758.organization of the detached epidermis which included 13. Schaefer, B. M., Jaeger, C., Drepper, E., and Kramer, M. D.

(1996) Autoimmunity 23, 155–164.the upward migration of basal cells and the subsequentdisappearence of the typical basal keratinocyte layer. 14. Rømer, J., Lund, L. R., Eriksen, J., Ralfkiaer, E., Zeheb, R.,



Gelehrter, T. D., Dano, K., and Kristensen, P. (1991) J. Invest. 27. Reinartz, J., Schaefer, B. M., Batria, R., Klein, C. E., andKramer, M. D. (1995) Exp. Cell Res. 220, 274–282.Dermatol. 97, 803–811.

28. Vassalli, J.-D., and Belin, D. (1987) FEBS 214, 187–191.15. Rømer, J., Lund, L. R., Eriksen, J., Pyke, C., and Dano, K.(1994) J. Invest. Dermatol. 102, 519–522. 29. Kramer, M. D., Vettel, U., Schmitt, M., Reinartz, J., Brunner,

G., and Meissauer, A. (1992) Fibrinolysis 6, 103–111.16. Schaefer, B. M., Jaeger, C., and Kramer, M. D. (1996) Br. J.Dermatol., in press. 30. Chomczynski, P., and Sacchi, N. (1986) Anal. Biochem. 162,

156–159.17. Yancey, K. B. (1995) J. Invest. Dermatol. 104, 1008–1014.31. Umezawa, H. (1976) Methods Enzymol. 45, 678–695.18. Poumay, Y., Boucher, F., Leclercq-Smekens, M., Degen, A., and

Leloup, R. (1993) Epithelial Cell Biol. 2, 7–16. 32. Morioka, S., Lazarus, G. S., Baird, J. M., and Jensen, P. J.(1987) J. Invest. Dermatol. 88, 418–423.19. Hashimoto, K., Singer, K. H., Lide, W. B., Shafran, K., Webber,

P., Morioka, S., and Lazarus, G. S. (1983) J. Invest. Dermatol. 33. Stanley, J. R. (1992) Adv. Immunol. 53, 291–325.81, 424–429. 34. Garrod, D. R. (1993) Curr. Opin. Cell Biol. 5, 30 –40.

20. Jensen, P. J., Michelle, J., and Baird, J. M. (1990) Exp. Cell Res. 35. Alitalo, K., Kuismanen, E., Myllyla, R., Kiistala, U., Asko-Selja-187, 162–169. vaara, S., and Vaheri, A. (1982) J. Cell Biol. 94, 497–505.

21. Schaefer, B. M., Stark, H. J., Fusenig, N. E., Todd, R. F., III, 36. Marchisio, P. C., Cancedda, R., and De Luca, M. (1990) Celland Kramer, M. D. (1995) Exp. Cell Res. 220, 415–423. Differ. Dev. 32, 355–360.

22. Nishizawa, Y., Uematsu, J., and Owaribe, K. (1993) J. Biochem. 37. Marchisio, P. C., Bondanza, S., Cremona, O., Cancedda, R., and113, 493–501. De Luca, M. (1991) J. Cell Biol. 112, 761–773.

23. Poumay, Y., Roland, I. H., Leclercq-Smekens, M., and Leloup, 38. Thacher, S. M., Malone, K. L., Dave, K., and Zhao, S. M. (1991)R. (1994) J. Invest. Dermatol. 102, 111 –117. Exp. Cell Res. 194, 238–247.

24. Stenn, K. S., Link, R., Moellmann, G., Madri, J., and Kuklinska, 39. Kirchheimer, J. C., Christ, G., and Binder, B. R. (1989) Eur. J.E. (1989) J. Invest. Dermatol. 93, 287–290. Biochem. 181, 103–107.

25. Rheinwalsd, J. G., and Green, H. (1975) Cell 6, 331–344. 40. Del Rosso, M., Anichini, E., Pedersen, N., Blasi, F., Fibbi, G.,Pucci, M., and Ruggiero, M. (1993) Biophys. Res. Commun. 190,26. Buessecker, F., Reinartz, J., Schimer, U., and Kramer, M. D.

(1993) J. Immunol. Methods 162, 193 –200. 347–352.

Received May 29, 1996Revised version received August 29, 1996


Documents

Dispase-Mediated Basal Detachment of Cultured Keratinocytes Induces Urokinase-type Plasminogen Activator (uPA) and Its Receptor (uPA-R, CD87)