Epidermal growth factor and keratinocyte growth factor differentially regulate epidermal migration,...

Epidermal growth factor and keratinocyte growth factordifferentially regulate epidermal migration, growth, anddifferentiation

SUSAN GIBBS, PhDa; ANNA NUBIA SILVA PINTO, PhDa; SEEMA MURLI, BSca; MARCEL HUBER, PhDb; DANIELHOHL, PhDb; MARIA PONEC, PhDa

Various growth factors such as epidermal growth factor and keratinocyte growth factor have been reported topromote wound closure and epidermal regeneration. In the present study epidermis reconstructed on de-epidermizeddermis was used to investigate the effects of epidermal growth factor and keratinocyte growth factor on keratinocyteproliferation, migration and differentiation. Our results show that epidermal growth factor supplemented cultures sharemany of the features which are observed during regeneration of wounded epidermis: a thickening of the entireepidermis, an enhanced rate of proliferation and migration, and an increase in keratin 6, keratin 16, skin-derivedantileukoproteinase, involucrin and transglutaminase 1 expression. The increase in transglutaminase 1 protein is accom-panied by an increase in the amount of active transglutaminase 1 enzyme. Surprisingly no increase in keratin 17 isobserved. Prolonging the culture period for more than two weeks results in rapid senescence and aging of the cultures.In contrast, keratinocyte growth factor supplemented cultures have a tissue architecture that is similar to healthynative epidermis and remains unchanged for at least 4 weeks of air-exposure. The rate of proliferation and theexpression of keratins 6, 16 and 17, skin-derived antileukoproteinase, involucrin and transglutaminase 1 is similar to thatfound in healthy epidermis and furthermore keratinocyte migration does not occur. When the culture medium issupplemented with a combination of keratinocyte growth factor and a low concentration of epidermal growth factor,skin-derived antileukoproteinase, involucrin and keratins 6, 16 and 17 expression is similar to that found in culturessupplemented with keratinocyte growth factor alone and in healthy epidermis. Only high transglutaminase 1 expressionremains similar to that observed in cultures supplemented with epidermal growth factor alone. Our results show thatthe regulation of keratinocyte growth, migration and differentiation depends on the availability of these growth factors.Epidermal growth factor may play a dominant early role in wound healing by stimulating keratinocyte proliferationand migration while keratinocyte growth factor may play a role later in the repair process by stabilizing epidermalturnover and barrier function. (WOUND REP REG 2000;8:192–203)

Wound repair involves a sequential program of events in BL Basal layerorder to close a wound as quickly as possible. An increase EGF Epidermal growth factorin the rate of migration of keratinocytes over the KGF Keratinocyte growth factor

PI Proliferation indexPMN Polymorphonuclear leukocyteFrom the Department of Dermatology,a Leiden UniversitySC Stratum corneumMedical Centre, The Netherlands, and Depart-SG Stratum granulosumment of Dermatology,b University Hospital Lau-SKALP Skin-derived antileukoproteinasesanne, Switzerland.SS Stratum spinosumReprint requests: SusanGibbs, PhD,Department ofDerma-TGF Transforming growth factortology, Leiden University Medical Centre, Building

3, Sylvius Laboratory, Wassenaarseweg 72, 2333AL Leiden, The Netherlands. Fax: `31 (071) 527-1910; Email: [email protected].

wounded area as well as an increase in the rate of prolif-Copyright q 2000 by The Wound Healing Society.ISSN: 1067-1927 $15.00 ` 0 eration of basal cells must occur. Simultaneously, prog-

192

WOUND REPAIR AND REGENERATIONVOL. 8, NO. 3 GIBBS ET AL. 193

eny cells must migrate into the upper epidermal layers, proliferation and migration, whereas KGF may play arole later in the repair process by stabilizing epidermalultimately terminally differentiating to form the stratum

corneum and repairing the barrier function of the skin. turnover and barrier function.During this process, the cells change dramatically bothmorphologically and biochemically. It is thought thatmany growth factors (e.g., epidermal growth factor MATERIALS AND METHODS[EGF], keratinocyte growth factor [KGF], transforming Cultures of normal human epidermal keratinocytes (ob-growth factor [TGF]-a, TGF-b, interleukin-1, the fibro- tained from the skin of healthy patients undergoing surgi-blast growth factor family, platelet-derived growth factor, cal corrections) were established as described earlier.12,13

and insulin-like growth factors)1–3 are involved during De-epidermized dermis was also prepared as describeddifferent stages of wound repair by regulating the synthe- earlier.14 Reconstructed epidermis was generated as fol-sis of specific proteins involved in different stages of lows. Secondary cultures of adult human keratinocytesepithelial proliferation and differentiation. were seeded onto de-epidermized dermis. The cultures

Upon injury, a strong induction of keratin 6, 16 and were grown submerged for 5 days in a medium consisting17 occurs in post-mitotic cells at the wound edge.4 It is of a mixture of 3 : 1 Dulbecco’s modified Eagle mediumthought that these keratins do not have a primary struc- and Ham’s F12 medium supplemented with 1% Hyclonetural function similar to other epidermal keratins but bovine serum (Greiner, Alphen a/d Rijn, Netherlands),rather that they may promote reorganization of the cyto- 1 mM hydrocortisone, 1 mM isoproterenol, 0.1 mM insulin,plasmic array of keratin filaments. This is an event that 1.0 2 10–5 M L-carnitine, 1.0 2 10–2 M L-serine, 1 mMprecedes the onset of keratinocyte migration into the dL-a-tocopherol-acetate. The medium was also enrichedwound site.4 Skin-derived antileukoproteinase (SKALP), with a lipid supplement as described by Boyce and Wil-also known as elafin, is another protein induced during liams15 containing 2.5 2 10–5 M palmitic acid, 2.5 2 10–5

the early stages of wound repair. Being a specific inhibi- M oleic acid, 1.5 2 10–5 M linoleic acid, 0.7 2 10–5 Mtor of polymorphonuclear leukocyte (PMN)-derived elas- arachidonic acid, and 2.4 2 10–5 M bovine serum albumintase and proteinase-3, it is a negative regulator of PMN (essential fatty acid free, as carrier). After 5 days humaninfiltration by providing a negative feedback mechanism recombinant KGF (Sigma Chemical Co., St. Louis, MO)to protect the tissue against excessive proteolysis by or mouse EGF (Sigma) was added to the medium. Afterproteinases.5,6 Therefore it is thought to play a role in an additional 2 days, cultures were lifted to the air–liquidthe acute, inflammatory phase of wound healing as well interface and the same culture medium was used exceptas being expressed in psoriasis and epidermal tumors.7–9 that serum was omitted and 50 mg/ml ascorbic acid wasIn addition to being involved in inflammatory responses, added.16 All additives were purchased from Sigma. TheSKALP has also been found to be an isodipeptide cross- medium was refreshed three times per week and all ex-linked component of the cornified envelope.10 For the periments were performed in triplicate.formation of a competent stratum corneum (SC) barrier,an increase in a number of cornified envelope compo- Keratinocyte migration measurementnents and cross-linking enzymes is required as well as Keratinocyte migration was examined by measuring thean increase in lipogenesis. Mansbridge and Knapp11 de- lateral expansion of air-exposed cultures. The keratino-tected induction of involucrin, filaggrin and transglutami- cytes were seeded into a metal ring with a diameter ofnase along with a number of differentiation specific 0.8 cm2 placed on de-epidermized dermis (4 cm2). Thekeratins (K1, K2 and K10) after tape stripping or suction increase in diameter of the culture was expressed asblister injury in humans in vivo. percentage of culture diameter compared to plating di-

The purpose of this study is to use reconstructed ameter. Three independent experiments were performed.epidermis as an in vitro model to investigate the rolesof EGF and KGF on keratinocyte migration, proliferation Morphology and immunohistochemistry

Cultures were harvested and washed in phosphate buf-and differentiation. These are potentially importantgrowth factors in the wound healing process. Recon- fered saline solution, fixed in 4% formaldehyde for 2

hours and then processed for conventional paraffin em-structed epidermis was generated by seeding human ker-atinocytes onto de-epidermized dermis and culturing at bedment. Sections (5 mm) were cut, deparaffinized and

rehydrated in preparation for morphological or immuno-the air–liquid interface in a serum free, fully definedkeratinocyte medium supplemented with either EGF or histochemical analysis of keratin 6, SKALP, involucrin

and Ki67.KGF. Our results suggest that EGF may play a dominantearly role in wound healing by stimulating keratinocyte Antibodies to human SKALP were a gift from Dr. J.

WOUND REPAIR AND REGENERATIONMAY–JUNE 2000194 GIBBS ET AL.

Schalkwijk, Nijmegen;8 antibodies to human involucrin dard lipids were used for quantitation. The quantitationwas performed after staining (copper acetate and copper(SY5) were a gift from Dr. F. Watt, London;17 antibodies

to keratin 16 (LL0025) were a gift from Dr. I. Leigh, sulfate in phosphoric acid) and charring, using a photo-densitometer (Shimadzu CS-9000) with automatic peakLondon;18 antibodies to human transglutaminase (BC1)

were purchased from Biomedical Technologies Inc.;19 an- integration (Shimadzu FDU-3).tibodies to human keratin 6 were purchased from PRO-GEN (Heidelberg, Germany); antibodies to keratin 17

RESULTSwere purchased from Sigma (St. Louis, MO) and antibod-ies to Ki67 were purchased from DAKO (Glostrup, Den- Reconstructed epidermis grown in the presence of 5 ng/mark). ml KGF gradually forms a multilayered tissue over a

For morphological observations, sections were period of 2 weeks of air exposure (Figure 1). The overallstained with hematoxylin and eosin. For immunohisto- tissue architecture closely resembles that of native epi-chemical analysis, the avidin biotin peroxidase complex dermis. The cultures consist of 6–8 living cell layers con-method was used essentially as described by the suppli- taining a well-defined basal layer (BL), stratum spinosumers (streptABCcomplex/HRP Code No. K 377, DAKO) (SS), and stratum granulosum (SG). The SG consists ofwith the following minor modifications: phosphate buf- 1–2 cell layers, and this tissue architecture is maintainedfered saline solution was used in place of Tris buffered for at least 4 weeks of air-exposure of the culture. Fur-saline solution; prior antigen retrieval was performed thermore, the SC is thin (as in native epidermis) and onlywith anti-Ki67 (DAKO) and keratins 6, 16 and 17 by im- increases slightly in thickness during the culture period.mersion of the slides in 0.01 M citrate buffer (pH 6.0) In contrast to KGF supplemented cultures, skinfor 30 minutes at 100 7C, followed by slow cooling to equivalents cultured in the presence of 5 ng/ml EGFroom temperature over a period of at least 2 hours. rapidly form a tissue consisting of 14–16 living cell layers

Immunohistochemical analysis of transglutaminase with a relatively thin SC after 1 week of air-exposureand transglutaminase activity assay was performed using (Figure 1). The SG is very prominent (in contrast to5 mm frozen sections as previously described.20 KGF supplemented cultures and native epidermis) and

consists of 6–10 cell layers. However, when the cultureEstimation of Ki67 proliferation index period is extended to 4 weeks of air-exposure the numberThe proliferation index is expressed as the number of of living cell layers rapidly decreases, reaching a thick-Ki67 positively stained nuclei/total number of basal epi- ness of about 5–7 cell layers, with a poorly defined SSdermal cells 2 100%. From 3 different regions of each consisting of prematurely flattened cells. The SC is verysection a minimum of 100 basal cells were counted at compact and consists of many cell layers.2002 magnification. From these 3 regions a mean value In contrast to cultures grown in the presence of EGFwas determined. This value was determined for 3 inde- or KGF, skin equivalents grown in the absence of growthpendent experiments each of which was performed in factors form a tissue with very few living cell layers (2–4)duplicate. The resulting value is expressed as mean with and a poorly defined tissue architecture throughout thestandard error of the mean (SEM) in the results. culture period (Figure 1).

EGF stimulates keratinocyte migration in contrastLipid extraction and separationCultures were harvested after 2 and 4 weeks of air- to KGF

The establishment of cell cultures with an initial areaexposure. The epidermis was separated from the dermisby heating for 1 minute at 40 7C and collected in chloro- of 0.8 cm2 on a dermal substrate of 4 cm2 enables cell

migration for up to 2 weeks of culture to be followed.form:methanol (1 : 2, v/v). The extraction and separationwas performed as previously described by Ponec and The lateral expansion of the cultures is directly related

to the extent of cell migration.23Weerheim.21 Total lipids were extracted according to themethod of Bligh and Dyer,22 with the addition of 0.25 M Keratinocyte migration is much more pronounced in

the presence of EGF than in the presence of KGF orKCl to ensure extraction of polar species and stored inchloroform : methanol (2 : 1, v/v) at – 20 7C under nitrogen control cultures (Figure 2). In the absence of either

growth factor the diameter of the cultures increases tountil use. The extracted lipids were separated by one-dimensional high performance chromatography (HP- only 104 5 6% of the original size after 2 weeks culture

at the air–liquid interface. Migration is strongly stimu-TLC) on 20 2 10 cm glass plates coated with silicagel (Kieselgel 60, Merck, Darmstadt, Germany) using the lated by the addition of 5 ng/ml EGF to the culture me-

dium (culture diameter increases to 193 5 22%) whereas‘‘total and ceramide development systems’’ as describedin detail previously.21 Serial dilutions of appropriate stan- the addition of 5 ng/ml KGF to parallel cultures does


Figure 1. KGF promotes normal differentiation in contrast to EGF. Human skin equivalents were cultured in the absence or presenceof 5 ng/ml KGF or EGF for a period of 1–4 weeks in air-exposed cultures. Hematoxylin and eosin staining is shown. The final magnificationof each photograph is 8002.

not significantly stimulate migration (culture diameter KGF (Figure 3). After 1 week of air-exposure the PI was30.4 5 8.0%. This value rapidly decreased over a periodincreases to 107 5 5%). As in EGF supplemented cultures

the reconstructed epidermis completely covered the der- of 3 weeks of culture at the air–liquid interface to a valueof 3.5 5 0.42% with no further significant change uponmal substrate by 3 weeks of culture at the air–liquid

interface (data not shown), migration studies were lim- prolongation of the culture period to 4 weeks of air-exposure.ited to 2 week air-exposed cultures only.

Cultures grown in the absence of growth factor (Fig-Ki67 proliferation index of cultures grown in the ure 3) had a PI similar to that of native epidermis forpresence of KGF or EGF the first 2 weeks of air-exposed culture (1 week air-The number of Ki67-positive cells in normal human epi- exposed: PI 4 15.2 5 2.1%; 2 weeks air-exposed: PI 4dermis was 11.3 5 2.1% of the BL (Figure 3). This value 12.9 5 1.5%) with senescence following rapidly hereafteris comparable to cultures grown in the presence of 5 (3 weeks air-exposed: PI 4 6.3 5 1.4%; 4 weeks air-ng/ml KGF (Figure 3). After 1 week of air-exposure, a exposed: PI 4 1.9 5 1.1%).proliferation index (PI) of 14.9% 5 0.8 was found. After2 weeks of air-exposure this value decreased to 8.64 5 Differential effects of KGF and EGF on terminal

differentiation processes in reconstructed epidermis0.69% with no further significant change during the restof the 4 week air-exposed culture period. As reconstructed epidermis grown for 1 week in the

presence of 5 ng/ml EGF was clearly hyperproliferativeThe PI of skin equivalents grown in the presence of5 ng/ml EGF showed very different kinetics compared when compared to normal epidermis or cultures supple-

mented with 5 ng/ml KGF, we were interested to seeto those observed for cultures grown in the presence of


whether expression of a number of markers for keratino-cyte differentiation also differs in these cultures. After2 weeks of air-exposure, in KGF supplemented cultures,keratins 6, 16 and 17 are not expressed (Figure 4A),similarly as in normal healthy epidermis (data notshown). In contrast in EGF supplemented cultures, kera-tin 6 is strongly expressed in all SB layers, keratin 16 isstrongly expressed in lower SB layers and no significantexpression of keratin 17 is observed. In the absence ofgrowth factor, keratins 6, 16 and 17 are expressed weaklyand intermittently in the SB layers (Figure 4A).

SKALP (Figure 4B) which is not present in detectableamounts in normal epidermis is only weakly detectablein KGF supplemented and control cultures (intermittentexpression occurring only in the SG). In EGF supple-mented cultures a regular expression in SG and upperSS is observed. Involucrin and transglutaminase 1 (Fig-ure 4B), which are expressed in the SG in normal healthyepidermis are also expressed in the SG in KGF supple-mented and control cultures. However, the expressionof both of these proteins extends deep into the SS ofEGF supplemented cultures. Keratin 10, loricrin and

Figure 2. EGF stimulates keratinocyte migration to a greater de- SPRR2 were not differentially expressed in KGF and EGFgree than KGF. Human skin equivalents were cultured in the supplemented cultures. The expression of these threeabsence or presence of 5 ng/ml KGF or EGF for a period of 2 proteins resembled that found in healthy native epider-weeks air-exposed culture. Increase in diameter of the culture

mis (data not shown).is expressed as percentage of culture diameter compared toIn order to determine which growth factor had aplating diameter. Each bar represents the mean (5 error) of

dominant effect on specific protein expression, culturesthree independent experiments.

were supplemented with a combination of KGF and EGF

Figure 3. EGF and KGF have different effects on epidermal proliferation. The PI of human skin equivalents grown over a period of1–4 weeks in the absence (m) or presence of 5 ng/ml KGF (m) or 5 ng/ml EGF (□) is shown. The PI of native epidermis (EPI) is shown(m). The PI was determined from the number of Ki67 positive stained nuclei. Each bar represents the mean 5 SEM of 3 independentexperiments.


Figure 4. Differential expression of pro-teins involved in keratinocyte differentia-tion in human skin equivalents. Epidermiswas cultured air-exposed for 2 weeks inthe absence or presence of 5 ng/ml KGFor EGF. Immunohistochemical stainingusing antibodies directed against humankeratins 6, 16 and 17 (A) and againsthuman involucrin, SKALP, and transglu-taminase 1 (TGase) (B) is shown. The finalmagnification of each photograph is8002.

(Figure 5). The concentration used was 2 ng/ml KGF and factor concentrations less than 2 ng/ml KGF and 0.5 ng/ml EGF are not sufficient to support cultures.0.5 ng/ml EGF, as this was the minimum effective growth

factor concentration that enabled the cultures to main- When the culture medium was supplemented witha combination of 2 ng/ml KGF and 0.5 ng/ml EGF fortain all the characteristics of an EGF or KGF supple-

mented culture (compare Figure 4 and Figure 5). Growth 2 weeks (Figure 5), keratins 6, 16 and 17 were absent;


Figure 5. Differential dominant effects ofEGF and KGF on keratinocyte differentia-tion. Epidermis was cultured air-exposedfor 2 weeks in the presence of 2 ng/mlKGF, 0.5 ng/ml EGF, or a combination of2 ng/ml KGF and 0.5 ng/ml EGF. Immuno-histochemical staining using antibodiesdirected against human keratins 6, 16and 17 (A) and human involucrin, SKALPand transglutaminase (TGase) (B) isshown. The final magnification of eachphotograph is 8002.

SKALP was sporadically expressed in SG and involucrin expressed in the SG and upper SS similar to thatobserved in EGF supplemented cultures (Figure 4: 5was regularly expressed in SG. The expression of

these proteins was similar to that observed in KGF ng/ml; Figure 5: 0.5 ng/ml). As the cultures were allharvested after 2 weeks of air-exposure, no significantsupplemented cultures (Figure 4: 5 ng/ml; Figure 5:

2 ng/ml). Transglutaminase 1, however, was strongly difference was observed in tissue architecture even


though the cultures expressed different levels of differ- most layers of SS as well as SG and SC. Active transglu-taminase 1 in both KGF supplemented cultures and inentiation markers.

Immunohistochemical staining was performed on native epidermis is detected only on the cell peripheryof keratinocytes in the SG. In contrast to KGF supple-frozen sections (data not shown) as well as on paraffin

sections. Due to the greater preservation of tissue archi- mented cultures, in the presence of EGF (Figure 6E: 0.5ng/ml; Figure 6F: 5 ng/ml) transglutaminase 1 protein istecture only staining on paraffin sections is shown. The

exception being for the transglutaminase antibody as this detected in all SS cell layers as well as in the SC andSG. In cultures supplemented with 5 ng/ml EGF, activeantibody only works on frozen sections.enzyme could be detected in all of the suprabasal livingcell layers whereas in cultures supplemented with 0.5Active transglutaminase 1 in EGF or KGFng/ml EGF, only in the upper SS and SG is the enzymesupplemented cultures and native epidermispresent in the active form. Lower SS layers contain theThe finding that the localization of transglutaminase 1protein in the inactive form. In control cultures the pro-protein in cultures supplemented with EGF was differenttein is expressed in the SC, SG, and SS whereas activewhen compared to cultures supplemented with KGF andtransglutaminase 1 is detected only in the SG and uppernative epidermis prompted us to determine whether allSS. In all cases, no active enzyme could be detected inof the enzyme was present in the active form (Figurethe SC.6). Transglutaminase 1 can be activated by incubating

sections in a high calcium buffer in the presence of dan-sylcadaverine. Active enzyme is then detected by incu- Lipid composition of reconstructed epidermis in EGF

or KGF supplemented culturesbating with antidansyl antibodies (see Materials andMethods). In cultures supplemented with 2 ng/ml KGF In order to determine whether KGF and EGF modulated

epidermal lipogenesis, the lipid composition of native(Figure 6C) and also in native epidermis (Figure 6A),transglutaminase 1 protein is expressed in the SC and epidermis was compared to that of the keratinocyte cul-

tures (Figure 7). Cultures grown for a period of 2 weeksSG. In cultures supplemented with 5 ng/ml KGF (Figure6D) transglutaminase 1 protein is detected in the upper- at the air–liquid interface in the presence of 5 ng/ml KGF

Figure 6. Transglutaminase is active in theliving cell layers of human skin equiva-lents. Healthy native epidermis is shownin (A). Human skin equivalents weregrown for 2 weeks under air-exposedconditions in the absence (B) or pres-ence of KGF (C, 2 ng/ml; D, 5 ng/ml) orEGF (E, 0.5 ng/ml; F, 5 ng/ml). Protein ex-pression is detected by red fluorescenceand protein activity by yellow fluores-cence. The final magnification of eachphotograph is 8002.


of KGF and EGF on epithelial proliferation and differenti-ation. The response to EGF is an early event resultingin an increased rate of proliferation and migration, andan increased keratin 6 and 16 expression. Keratins 6, 16and 17 are thought to promote reorganization of thecytoplasmic array of keratin filaments, an event that pre-cedes the onset of keratinocyte migration into the woundsite, with keratin 6 and 16 induction occurring prior tokeratin 17 induction.4 The specific induction of keratins6 and 16 by EGF has been shown to occur via nuclearproteins that bind to EGF responsive elements in thepromoter regions of these keratin genes.24 The observa-tion that keratin 17 expression was not increased in EGFsupplemented cultures indicates independent regulatorypathways for these keratins—a finding which is also sup-ported by Latkowski et al.25 Keratin 6 is expressed as atleast 6 isoforms26 in which keratin 6a is thought to forma pair with keratin 16 and keratin 6b is thought to pairwith keratin 17.27 Although it is not known which keratin6 isoform is recognized by the antikeratin 6 antibodyused in the present study, our findings indicate that EGFinduces specific keratin 6a expression in reconstructedepidermis. This may have strong implications for woundhealing and hyperproliferative disorders as has alreadybeen suggested in a number of studies which show differ-ential keratin 16 and 17 expression (e.g., in nonfollicularepidermal tumors,28 basal cell carcinomas,29 hypertro-

Figure 7. Lipid composition indicates enhanced aging in EGF phic scars,30 irritant contact dermatitis31 and after X-raysupplemented cultures but not in KGF supplemented cultures.

irradiation32).Lipids extracted from healthy native epidermis (lane 1); or cul-Prolonged supplementation of the culture mediumtures grown for 2 weeks (lanes 2 and 4) or 4 weeks (lanes 3 and

5) in the presence of 5 ng/ml EGF (lanes 2 and 3) or 5 ng/ml KGF with EGF results in an increased rate of aging of the(lanes 4 and 5) were separated by HP-TLC using the ceramide cultures due to senescence and also due to a deviationdevelopment system. CE: cholesterol esters; TG: triglyceride; in the terminal differentiation program (as concludedCHOL: cholesterol; FFA: free fatty acids; CER: ceramides; AGC:

by observation of tissue morphology, determination ofacylglucosylceramide; GSL: glucosphingolipids; CSO4: choles-

proliferation index and by lipid analysis).terol sulfate; PL: phospholipid.In EGF supplemented cultures, the increase in trans-

glutaminase 1 protein corresponds to an increase intransglutaminase 1 activity. This coincides with a rapid(lanes 4 and 5) or 5 ng/ml EGF (lanes 2 and 3) had aincrease in thickness of the SC observed over a periodlipid composition which closely resembles that of nativeof 4 weeks. The increase in transglutaminase activity,epidermis (lane 1). However, after 4 weeks of cultureinvolucrin and SKALP expression, and SC thickness areEGF supplemented cultures showed a decrease in theall responses which one would expect to occur in a tissuerelative amount of phospholipids and glucosphingolipidsduring the rapid assembly of cornified envelopes thatindicating a decrease in the number of living cell layersare required to repair the barrier function of the skin. As(lane 3). In contrast, the relative amount of both phos-SKALP is also a protease inhibitor, as well as a cornifiedpholipids and glucosphingolipids remained unchangedenvelope precursor, its induction by EGF may also befor the entire 4 week air-exposed culture period in KGFlinked to a negative feedback mechanism to protect thesupplemented cultures (lane 5).tissue against excessive proteolysis by proteinases se-creted during wound repair.9,33

DISCUSSION In contrast to EGF supplemented cultures, the prolif-eration index of KGF supplemented cultures is compara-The results from this study clearly indicate that it is

possible to functionally distinguish between the effects ble to that found in native epidermis and does not change


for at least 4 weeks of exposure of the cultures to the mitogen, a number of reports suggest that KGF is justas potent or even a stronger mitogen than EGF. Marcheseair. Furthermore, KGF did not stimulate migration. Kera-

tins 6, 16 and 17, SKALP, involucrin and transglutaminase et al.35 found that KGF was a potent mitogen in culture,equivalent to or more active than EGF. However, conven-1 expression showed a similar pattern of expression to

that observed in growth factor free cultures (even though tional submerged keratinocyte cultures grown under lowcalcium conditions were used for these studies. Whenthese cultures did not have such a well defined morphol-

ogy as KGF supplemented cultures) indicating that the these cultures were permitted to differentiate by the addi-tion of calcium to the culture medium, KGF (and alsoexpression of these proteins is not regulated by KGF.

SKALP was sporadically expressed, keratins 6, 16 and EGF) did indeed act as a poor mitogen. During skinregeneration in vivo, topical application of KGF (3 mg) to17 were absent and involucrin and transglutaminase 1

were expressed only in the SG. These are all characteris- partial thickness wounds in rabbits resulted in a normaldifferentiation of the regenerating epidermis but a signifi-tics of healthy native epidermis which has a low, steady

state of epithelial proliferation and differentiation and cant increase in the specific labelling of cycling cells inboth basal and suprabasal layers.42 This increase in thewhere cornified envelopes are generally only lost during

desquamation. rate of proliferation is most probably dose-dependent asa study performed by Staiano-Coico et al.37 with a lowerIn order to determine whether one growth factor was

dominant over the other as regards specific expression of concentration of KGF (1 mg) applied topically to fullthickness wounds in a porcine model reports that theproteins involved in keratinocyte differentiation, recon-

structed epidermis was generated in media supple- rate of wound closure was not increased but epidermalthickening and deep rete ridge formation occurred.mented with a combination of EGF and KGF at their

minimal effective concentrations. SKALP, involucrin and These treated wounds exhibited orthokeratotic matura-tion, and although cell proliferation was initially slightlykeratins 6, 16 and 17 expression was similar to that found

in cultures supplemented with KGF alone and in healthy higher than in control animals it was localized only inthe basal cells. This result was similar to our observationsepidermis. Only high transglutaminase 1 expression re-

mains similar to that observed in cultures supplemented in reconstructed epidermis: initially (after 1 week of airexposure), proliferation was also slightly higher than thatwith EGF alone. The results from these experiments indi-

cate that the effect of EGF on transglutaminase expres- found in healthy native epidermis and was also onlylocalized in basal cells.sion is dominant over the effects of KGF, whereas the

effect of KGF on keratins 6, 16 and 17, involucrin and A dramatic increase in KGF expression was observedin full thickness wounds created in mouse skin43 sug-SKALP expression is dominant over the effects of EGF.

Therefore we can conclude that the regulation of kera- gesting a direct role of KGF in wound healing. However,Marchese et al.44 found in re-epithelializing human skintinocyte growth, migration and differentiation during

wound healing might vary depending on the availability that the KGF receptor protein was undetectable in basalkeratinocytes during wound repair, whereas the cell PIof EGF and KGF receptors and their ligands.

In agreement with our findings with reconstructed was increased several-fold over that of basal cells incontrol skin. Interestingly, the lack of an essentialepidermis, Chih-Shan et al.34 found that in air-exposed

cultures grown on fibroblast populated collagen gels, involvement of KGF in wound healing was shown byGuo et al.45 who found KGF knockout mice had no abnor-EGF had a deleterious effect on morphogenesis and dif-

ferentiation whereas a number of other reports show malities in epidermal growth or wound healing. Howeveraccurate conclusions from these experiments has be-that KGF promotes the normal differentiation pheno-

type.35–37 A number of other researchers using conven- come complicated due to the recently identified FGF-10which is highly homologous to KGF and binds to thetional submerged cultures have also found that EGF

greatly stimulated keratinocyte migration whereas KGF same receptor (FGFR2-IIIb) with high affinity.3 FGF 10has been reported to exert significant effects on woundhad very little effect.38–40 Only Tsuboi et al.41 found that

KGF stimulated migration significantly (to the same ex- healing in both humans46 and rabbits47 and may thereforebe the major ligand for FGFR2-IIIb during wound healing.tent as TGF-a) in monolayer, proliferating keratinocytes

grown under submerged conditions in low Ca2` medium. This would also explain the results of Werner et al.48

who found that KGF receptor knockout mice did showThe discrepancy in these results is most probably dueto the different culture conditions used. a reduced rate of proliferation of keratinocytes at the

wound edge and a delayed rate of re-epithelialization ofHowever, data describing the effects of these twogrowth factors on keratinocyte proliferation are more the wound.

It has also been found that dominant negative mu-conflicting: while EGF is generally accepted as a potent


histochemical localization of SKALP/elafin in psoriatic epidermis.tants of the EGF receptor in the epidermis of transgenicJ Invest Dermatol 1993;100:390–3.

mice show marked hyperplasia, keratin 6 expression and 9. van Bergen BH, Andriessen MPM, Spruijt KIJ, van de Kerkhofincreased 5-bromo-2-deoxuridine incorporation.49 How- PCM, Schalkwijk J. Expression of SKALP/elafin during wound

healing in human skin. Arch Derm Res 1996;288:458–62.ever it should be noted that the EGF receptor is also a10. Steinert PM, Marekov LN. The proteins elafin, filaggrin, keratinreceptor for other ligands e.g., TGF-a.50 In the epidermis,

intermediate filaments, loricrin, and small proline-rich proteins 1the response to EGF or TGF-a is determined by the and 2 are isodipeptide cross-linked components of the human

epidermal cornified cell envelope. J Biol Chem 1995;270:17702–11.localization and the number of receptors and is modu-11. Mansbridge JN, Knapp AM. Changes in keratinocyte maturationlated by processes affecting the binding affinity, internal-

during wound healing. J Invest Dermatol 1987;89:253–63.ization and tyrosine-kinase activity of the receptor.51

12. Ponec M, de Haas C, Bachra B, Polano M. Effects of glucocorti-In conclusion, our findings support those of other coids on primary human skin fibroblasts. Arch Dermatol Res 1977;

259:117–23.researchers and indicate that wound healing involves a13. Ponec M, Kempenaar J, de Kloet E. Corticoids and cultured humancomplex series of interactions involving different cell epidermal keratinocytes: specific intracellular binding and clinical

types and the release of cytokines and growth factors. efficacy. J Invest Dermatol 1981;76:211–4.14. Ponec M, Weerheim A, Kempenaar J, Mommaas A, Nugteren D.In this study we show that EGF induces proliferation,

Lipid composition of cultured keratinocytes in relation to theirmigration and an enhanced expression of proteins in-differentiation. J Lipid Res 1988;29:949–61.

volved in epidermal barrier formation whilst KGF pro- 15. Boyce ST, Williams ML. Lipid supplemented medium induces la-mellar bodies and precursors of barrier lipids in cultured ana-motes normal differentiation observed in healthylogues of human skin. J Invest Dermatol 1993;101:180–4.epidermis. The regulation of growth, migration and dif-

16. Ponec M, Weerheim A, Kempenaar J, Mulder A, Gooris G, Bouws-ferentiation during wound healing may vary depending tra J, Mommaas A. The formation of competent barrier lipids inon the availability of EGF and KGF receptors and their reconstructed human epidermis requires the presence of vitamin

C. J Invest Dermatol 1997;109:348–55.ligands. Our results suggest that EGF may play an initial17. Hudson D, Weiland K, Dooley T, Simon M, Watt F. Characterizationrole in wound healing by stimulating keratinocyte prolif-

of eight monoclonal antibodies to involucrin. Hybridoma 1992;11:eration and migration in order to quickly close a wounded 367–79.

18. Purkis PE, Steel JB, Mackenzie IC. Antibody markers of basalarea. In contrast KGF may play a role later in the repaircells in complex epithelia. J Cell Sci 1990;97:39–50.process by stabilizing epidermal turnover and barrier

19. Thatcher SM, Rice RH. Keratinocyte specific transglutaminase offunction. cultured human epidermal cells: relation to cross linked envelope

formation and terminal differentiation. Cell 1985;40:685–95.20. Hohl D, Aeschlimann D, Huber M. In vitro and rapid in situ trans-

glutaminase assays for congenital ichthyoses—a comparativeACKNOWLEDGMENT study. J Invest Dermatol 1998;110:268–71.21. Ponec M, Weerheim A. Retinoids and lipid changes in keratino-This study was partially financially supported by Numico

cytes. In: Packer L, editor. Methods in enzymology. Retinoids, volResearch BV, The Netherlands.190, Part B. San Diego: Academic Press, 1990: 30–41.

22. Bligh EG, Dyer WJ. A rapid method of total lipid extraction andpurification. Can J Biochem Physio 1959;37:911–7.

23. Gibbs S, Boelsma E, Kempenaar J, Ponec M. Temperature sensitiveREFERENCES regulation of epidermal morphogenesis and the expression ofcornified envelope precursors by EGF and TGFa. Cell Tiss Res1. Breuing K, Andree C, Helo G, Slama J, Liu PY, Eriksson E. Growth1998;292:107–14.factors in the repair of partial thickness porcine skin wounds.

24. Jiang CK, Magnaldo T, Ohtsuki M, Freedberg IM, Bernerd F, Blu-Plastic Reconstr Surg 1997;100:657–64.menberg M. Epidermal growth factor and transforming growth2. Greenhalgh DG. The role of growth factors in wound healing. Jfactor alpha specifically induce the activation- and hyperprolifera-Trauma 1996;41:159–67.tion- associated keratins 6 and 16. Proc Natl Acad Sci USA 1993;3. Werner S. Keratinocyte growth factor: a unique player in epithelial90:6786–90.repair processes. Cytokine Growth Factor Rev 1998;9:153–65.

25. Latkowski JM, Freedberg IM, Blumenberg M. Keratinocyte growth4. Paladini RD, Takahashi K, Bravo NS, Colombe PA. Onset of re-factor and keratin gene regulation. J Dermatol Science 1995;9:epithelialization after skin injury correlates with a reorganization36–44.of keratin filaments in wound edge keratinocytes: defining a poten-

26. Takahashi K, Paladini RD, Coulombe PA. Cloning and character-tial role for keratin 16. J Cell Biol 1996;132:381–97.ization of multiple human genes and cDNAs encoding highly re-5. Schalkwijk J, Chang A, Janssen P, de Jongh GI, Mier PD. Skin-lated type II keratin 6 isoforms. J Biol Chem 1995;270:18581–92.derived antileukoproteinases (SKALPs): characterization of two

27. Smith FJD, Jonkman MF, van Goor H, Coleman CM, Covello SP,new elastase inhibitors from psoriatic epidermis. Br J DermatolUitto J, McLean WHI. A mutation in human keratin K6b produces1990;122:631–41.a phenocopy of the K17 disorder pachyonychia congenita type 2.6. Wiedow O, Ludemann J, Utrecht B. Elafin is a potent inhibitor ofHum Mol Genet 1998;7:1143–8.proteinase-3. Biochem Biophys Res Comm 1991;174:6–10.

7. Alkemade HA, Molhuizen HO, van Vlijmen-Willems IM, van Haelst 28. Yoshikawa K, Katagata Y, Kondo S. Relative amounts of keratin 17are higher than those of keratin 16 in hair-follicle-derived tumors inUJ, Schalkwijk J. Differential expression of SKALP/elafin in

human epidermal tumors. Am J Pathol 1993;143:1679–87. comparison with nonfollicular epithelial skin tumors. J InvestDermatol 1995;104:396–400.8. Scalkwijk J, van Vlijmen IM, Alkemade JA, de Jongh G. Immuno-


29. Yoshikawa K, Katagata Y, Kondo S. Biochemical and immunohisto- 40. Ando Y, Jensen PJ. Epidermal growth factor and insulin-likegrowth factor I enhance keratinocyte migration. J Invest Dermatolchemical analyses of keratin expression in basal cell carcinoma.

J Dermatol Science 1998;17:15–23. 1993;100:633–9.41. Tsuboi R, Sato C, Kurita Y, Ron D, Rubin JS, Ogawa H. Keratinocyte30. Machesney M, Tidman N, Waseem A, Kirby L, Leigh I. Activated

keratinocytes in the epidermis of hypertrophic scars. Am J Pathol growth factor (FGF-7) stimulates migration and plasminogen acti-vator activity of normal human keratinocytes. J Invest Dermatol1998;152:1133–41.

31. Willis CM, Reiche L, Wilkinson JD. Keratin 17 is expressed during 1993;101:49–53.42. Pierce GF, Yanagihara D, Klopchin K, Danilenko DM, Hsu E, Ken-the course of SLS-induced irritant contact dermatitis, but unlike

keratin 16, the degree of expression is unrelated to the density ney WC, Morris CF. Stimulation of all epithelial elements duringskin regeneration by keratinocyte growth factor. J Exp Med 1994;of dividing keratinocytes. Contact Dermatitis 1998;39:21–7.

32. Liu K, Kasper M, Trott KR. Changes in keratinocyte differentiation 179:831–40.43. Werner S, Peters KG, Longaker MT, Fuller-Pace F, Banda MJ,during accelerated repopulation of the irradiated mouse epider-

mis. Int J Radiat Biol 1996;69:763–9. Williams LT. Large induction of keratinocyte growth factor expres-sion in the dermis during wound healing. Proc Natl Acad Sci USA33. Bauman H, Gauldie J. The acute phase response. Immunol Today

1994;15:74–80. 1992;89:6896–900.44. Marchese C, Chedid M, Dirsch OR, Csaky KG, Santanelli F, Latini34. Chih-Shan JC, Lavker RM, Rodeck U, Risse B, Jensen PJ. Use of

a serum-free epidermal culture model to show deleterious effects C, LaRochelle WJ, Torrisi MR, Aaronson SA. Modulation of kera-tinocyte growth factor and its receptor in reepithelializing humanof epidermal growth factor on morphogenesis and differentiation.

J Invest Dermatol 1995;104:107–12. skin. J Exp Med 1995;182:1369–76.45. Guo L, Degenstein L, Fuchs E. Keratinocyte growth factor is re-35. Marchese C, Rubin J, Ron D, Faggioni A, Torrisi MR, Messina A,

Frati L, Aaronson SA. Human keratinocyte growth factor activity quired for hair development but not for wound healing. GenesDev 1996;10:165–75.on proliferation and differentiation of human keratinocytes: differ-

entiation response distinguishes KGF from EGF family. J Cell 46. Jimenez PA, Rampy MA. Keratinocyte growth factor-2 accelerateswound healing in incisional wounds. J Surg Res 1999;81:238–42.Physiol 1990;144:326–32.

36. Rubin JS, Bottaro DP, Chedid M, Miki T, Ron D, Cunha GR, Finch 47. Xia YP, Zhao Y, Marcus J, Jimenez PA, Ruben SM, Moore PA, KhanF, Mustoe TA. Effects of keratinocyte growth factor-2 (KGF-2) onPW. Keratinocyte growth factor as a cytokine that mediates mesen-

chymal–epithelial interaction. In: Goldberg ID, Rosen EM, editors. wound healing in an ischaemia-impaired rabbit ear model and onscar formation. J Pathol 1999;188:431–8.Epithelial–mesenchymal interactions in cancer. Basel, Switzer-

land: Birkhauser Verlag, 1995: 191–214. 48. Werner S, Smola H, Liao X, Longaker MT, Krieg T, HofschneiderPH, Williams LT. The function of KGF in morphogenesis of epithe-37. Staiano-Coico L, Krueger JG, Rubin JS, D’limi S, Vallat NP, Valen-

tino L, Fahey T, Hawes A, Kingston G, Madden MR, Mathwich M, lium and reepithelialization of wounds. Science 1994;266:819–22.49. Murillas R, Larcher F, Conti CJ, Santos M, Ullrich A, Jorcano JL.Gottlieb AB, Aaronson SA. Human keratinocyte growth factor

effects in a porcine model of epidermal wound healing. J Exp Expression of a dominant negative mutant of epidermal growthfactor receptor in the epidermis of transgenic mice elicits strikingMed 1993;178:865–78.

38. Zeigler ME, Krause S, Karmiol S, Varani J. Growth factor-induced alterations in hair follicle development and skin structure. EMBOJ 1995;14:5216–23.epidermal invasion of the dermis in human skin organ culture:

dermal invasion correlated with epithelial cell motility. Invasion 50. King L, Gates R, Stoscheck C, Nanney L. The EGF/TGFa receptorin skin. J Invest Dermatol 1990;94:164S–170S.Metastasis 1996;16:3–10.

39. Peehl DM, Wong ST, Rubin JS. KGF and EGF differentially regulate 51. Ebner R, Derynck R. Epidermal growth factor and transforminggrowth factor a: differential intracellular routing and processingthe phenotype of prostatic epithelial cells. Growth Reg 1996;6:

22–31. of ligand-receptor complexes. Cell Regul 1991;2:599–612.

Epidermal growth factor and keratinocyte growth factor differentially regulate epidermal migration,...

Documents

Transcript of Epidermal growth factor and keratinocyte growth factor differentially regulate epidermal migration,...