Keratinocyte growth factor stimulates migration and hyaluronan synthesis in the epidermis by activation of keratinocyte hyaluronan synthases 2 and 3.

Keratinocyte growth factor (KGF) activates keratinocyte migration and stimulates wound healing. Hyaluronan, an extracellular matrix glycosaminoglycan that accumulates in wounded epidermis, is known to promote cell migration, suggesting that increased synthesis of hyaluronan might be associated with the KGF response in keratinocytes. Treatment of monolayer cultures of rat epidermal keratinocytes led to an elongated and lifted cell shape, increased filopodial protrusions, enhanced cell migration, accumulation of intermediate size hyaluronan in the culture medium and within keratinocytes, and a rapid increase of hyaluronan synthase 2 (Has2) mRNA, suggesting a direct influence on this gene. In stratified, organotypic cultures of the same cell line, both Has2 and Has3 with the hyaluronan receptor CD44 were up-regulated and hyaluronan accumulated in the epidermis, the spinous cell layer in particular. At the same time the expression of the early differentiation marker keratin 10 was inhibited, whereas filaggrin expression and epidermal permeability were less affected. The data indicate that Has2 and Has3 belong to the targets of KGF in keratinocytes, and support the idea that enhanced hyaluronan synthesis acts an effector for the migratory response of keratinocytes in wound healing, whereas it may delay keratinocyte terminal differentiation.

Labware, Bedford, MA; 3.83 mg/ml) was mixed with Earle's balanced salt solution (10ϫ EBSS, Invitrogen), 7.5% sodium bicarbonate (Invitrogen), and 1 M sodium hydroxide solution, at a volume ratio of 8:1:0.3:0.2, respectively, and allowed to solidify on 24-mm diameter culture inserts (3.0-m pore size; Transwell®, Costar, Cambridge, MA). Recently confluent cultures of REKs were trypsinized, suspended in Dulbecco's minimum essential medium (high glucose) as above except with 10% fetal bovine serum (HyClone, Logan, UT), applied on the collagen gels, and grown for 3 days with culture medium present both beneath and above the insert. The upper medium was then removed to facilitate differentiation at the air-liquid interface.
KGF Treatment-KGF (Sigma) was used at 0.1-100 ng/ml final concentrations in complete medium. In the monolayer cultures, treatment times from 10 min to 72 h were used, whereas in the organotypic cultures, KGF (2 and 20 ng/ml) was added to the culture medium on the 4th culture day, and thereafter with each fresh medium (every other day for the first week and then daily). To study the early effects of KGF on intracellular hyaluronan, some of the monolayer cultures were incubated for 20 min at room temperature in 10 turbidity reducing units/ml Streptomyces hyaluronidase (Seikagaku, Tokyo, Japan), washed with HBSS, and incubation continued in new medium with KGF and 10 units/ml hyaluronidase at 37°C for 120 min.
Staining of Hyaluronan in Monolayer Cultures-REKs grown in 8-well chamber slides were fixed with 2% paraformaldehyde and 0.5% glutaraldehyde in 0.1 M sodium phosphate buffer, pH 7.4 (PB), for 20 min, washed with PB, and permeabilized with 0.3% Triton X-100 in 1% BSA-PB for 30 min. In a part of the slides, extracellular hyaluronan was enzymatically digested with Streptomyces hyaluronidase (10 turbidity reducing units/ml, 10 min at 37°C) prior to the permeabilization. After permeabilization, the cells were incubated overnight at 4°C with a biotinylated complex of hyaluronan binding region of bovine articular cartilage aggrecan G1 domain and link protein (bHABC) (26), diluted to 3-5 g/ml in 1% BSA. The bound probe was visualized by incubation with avidin-biotin-peroxidase complex (Vector Laboratories Inc., Burlingame, CA) for 1 h and with 0.05% 3,3Ј-diaminobenzidine (DAB, Sigma) and 0.03% H 2 O 2 for 5 min (26).
Electron Microscopy-REKs were fixed as described above and permeabilized with 0.05% saponin in 3% BSA-PB for 10 min on an ice bath. The staining for hyaluronan was done as above, except that all the incubation solutions and washing buffers contained 0.05% saponin, and the incubations were done at 4°C. The samples were postfixed with 1% reduced osmium tetroxide for 15 min, dehydrated in graded ethanol, and embedded in Spurr's resin. Thin sections were cut on Formvarcoated copper grids, stained with uranyl acetate and lead citrate, and viewed in a type 1200 EX microscope from JEOL (Tokyo, Japan).
Confocal Microscopy-The cells were fixed with 2% paraformaldehyde for 20 min, washed, and treated with Streptomyces hyaluronidase as described above. After permeabilization with 0.1% Triton X-100 in 1% BSA-PB for 10 min, the cells were incubated with bHABC (5 g/ml) and an anti-rat CD44 antibody (OX50, BIOSOURCE, Camarillo, CA, 1:100) overnight at 4°C, washed, treated with Texas Red streptavidin (Vector, 1:1000) and FITC-labeled anti-mouse IgG (Vector, 1:200) for 1 h at room temperature, washed, mounted with Vectashield (Vector), and viewed with an UltraView confocal scanner (PE-Wallac-LSR, Oxford, UK), built on a Nikon TE300 microscope. Ten optical sections from the top to the bottom of the cells were taken using a 100ϫ/1.3 numeric aperture objective. Further image processing was done with Photoshop software (Adobe, San Jose, CA).
Optical Density Measurements-Optical densities of the cells stained for hyaluronan using DAB were analyzed by a Leitz BK II microscope with 16ϫ/0.45 numeric aperture objective (Leitz, Wetzlar, Germany) and a digital camera (Photometrics CH 200, Roper Scientific Inc., Trenton, NJ) as described previously (27). Area-integrated mean optical density values for the DAB chromogen were calculated for each whole digitized area, excluding possible artifact areas. In the densitometric assays, the hyaluronan remaining after Streptomyces hyaluronidase treatment of non-permeabilized cells was designated as "intracellular," whereas in the chemical assays (see below) the term "intracellular" represents hyaluronan resistant to peeling of the pericellular hyaluronan with trypsin-EDTA. These two techniques give parallel results for the intracellular pool, but the values from the trypsin-EDTA method are somewhat higher (28).
Histological Analyses of the Organotypic Cultures-Two-week-old cultures were fixed overnight in 2% buffered paraformaldehyde containing 0.5% glutaraldehyde, or in Histochoice® (Amresco, Solon, OH). After fixation the cultures were washed with PB, pH 7.4, embedded in paraffin, and cut into 3-m-thick vertical sections.
Morphometry-Hematoxylin-eosin-stained sections were systemati-cally sampled by taking six digital images with a CoolSNAP camera (Roper Scientific) from each culture at constant intervals using a 20ϫ objective and a 1.25ϫ intermediate lens (Nikon Microphot FXA microscope). The heights of the basal cells, vital epidermis, and stratum corneum were each measured using the NIH Image 1.62/fat software for Macintosh (Wayne Rashband, National Institutes of Health, Bethesda, MD). Thresholding of areas exhibiting background intensity was used to exclude the areas between separated corneocytes in the stratum corneum measurements. Data from six cultures were analyzed for each group.
Hyaluronan Staining of the Sections-Aldehyde-fixed sections were rehydrated, treated for 5 min with 1% H 2 O 2 to block endogenous peroxidases, and stained for hyaluronan using bHABC, avidin-biotin, and DAB, as described above. The sections were counterstained with Mayer's hematoxylin for 1 min, washed, dehydrated, and mounted in DPX (Gurr ® , BDH Laboratory Supplies, Poole, UK).
Immunocytochemistry-Histochoice-fixed, deparaffinized sections were first incubated in target unmasking fluid (TUF TM , Monosan, Uden, The Netherlands) at 95°C, then for 5 min with 1% H 2 O 2 to block endogenous peroxidases, washed with PB, and incubated in 1% BSA in PB for 30 min to block nonspecific binding. The sections were incubated overnight at 4°C with monoclonal anti-keratin 10 (Monosan, 1:10 dilution in 1% BSA), polyclonal anti-filaggrin (a generous gift of Dr. Beverly Dale-Crunk, University of Washington, Seattle, WA; 1:5,000 dilution) or monoclonal anti-CD44 (OX50, 1:50), followed by an 1-h incubation with biotinylated anti-mouse IgG antibody (1:50, Vector) or biotinylated anti-rabbit antibody (1:70, Vector). The bound antibodies were visualized with the avidin-biotin-peroxidase and DAB, counterstained with hematoxylin, and mounted as described above. The controls included sections treated in the same way but with the primary antibody omitted.
In Vitro Hyaluronan Synthase Assay-The assay was done essentially as described (31). Subconfluent REK cells (16 -24 ϫ 10 6 ) were incubated for 14 h with 0 -100 ng/ml KGF, membrane fraction isolated and incubated with 0.5 mM UDP-GlcNAc and 0.05 mM UDP-GlcA (both from Sigma), the latter containing 2.5 Ci of UDP-[ 14 C]GlcA (PerkinElmer Life Sciences), for 2 h at 37°C. The samples were boiled with 1% SDS, incorporated activity separated by paper chromatography, quantified by liquid scintillation counting, and expressed as picomoles of GlcA incorporated/mg of protein in the membrane fraction (31).
Metabolic Labeling Assay-REKs were seeded into 6-well plates at 200,000 cells/well, and grown until subconfluent (2 days). Fresh medium containing [ 3 H]glucosamine (20 Ci/ml) and [ 35 S]SO 4 (10 Ci/ml) (Amersham Biosciences, Little Chalfont, UK), and the appropriate amounts of KGF (0, 1, 10, and 100 ng/ml) were added to the cells and incubated for 6 or 18 h. The medium and two 0.3-ml HBSS (EuroClone) washes of the cell layer were combined and designated "medium." Cell surface-associated hyaluronan was detached with 0.5 ml of 0.05% trypsin (w/v), 0.02% EDTA (w/v) for 10 min at 37°C, and the cells were pelleted and washed with 250 l of HBSS. The trypsin solution and the HBSS wash were combined and designated "pericellular," whereas the cell pellet was designated as the "intracellular" hyaluronan pool. Hyaluronan and other glycosaminoglycans were purified and quantitated from the different cellular compartments after determination of the specific activity of the hexosamines as described in detail previously (25,27).
Hyaluronan Disaccharide Analysis with Electrophoresis-Medium samples (400 l) were boiled for 10 min and digested with 40 l of proteinase K (Sigma, 600 g/ml in 100 mM ammonium acetate, pH 6.5) for 1.5 h at 60°C. After proteinase K inactivation by boiling for 10 min, 50 l of 50% trichloroacetic acid was added to precipitate proteins by centrifugation (15 min, 13,000 ϫ g). Dialyzed supernatants were evaporated, dissolved in 100 mM ammonium acetate, pH 6.5, and digested for 3 h at 37°C with 2 milliunits of Streptococcus hyaluronidase (Seikagaku), dried, and derivatized overnight at 37°C in 5 l of 0.1 M 2-aminoacridone (Lambda Fluoreszenztechnologie GmbH, Graz, Austria) in 3:17 (v/v) acetic acid:dimethyl sulfoxide, and 5 l of 1 M NaBH 3 CN. The 2-aminoacridone-derivatized disaccharides were stored at Ϫ20°C until electrophoresis as described (16,32), with the following modification; 30% polyacrylamide gels were cast in the laboratory in 100 mM Tris borate buffer, pH 8.9, and the same buffer was used as the running buffer. The intensities of the hyaluronan disaccharide bands derived from the samples and hyaluronan standards (Healon®, Amersham Biosciences) were digitized on a UV light box using a CCD camera. Quantitative image processing was done with NIH Image.
Hyaluronan Enzyme-linked Immunosorbent Assay-Organotypic cultures were changed into 1.5 ml of serum-free medium with KGF (0, 2, and 20 ng/ml) and continued for 24 h. The medium, epidermis, and collagen support were each analyzed separately, the latter two after extraction with 2 ϫ 2 ml of acetone at 4°C over 2 days, and digestion of the residue overnight at 60°C with 250 g/ml papain (Sigma) in 5 mM cysteine and 5 mM EDTA. After incubation, the samples were boiled for 10 min to inactivate the enzyme, centrifuged at 13,000 ϫ g, and the pellet discarded. Maxisorp Plates (Nunc, Roskilde, Denmark) were coated overnight at 4°C with 1 g/ml hyaluronan-binding complex (26), washed with PBS containing 0.5% Tween 20 (Tween-PBS), and blocked with 1% BSA for 1 h at 37°C. Standard hyaluronan (Provisc, Algon Laboratories Inc., Fort Worth, TX) at 150 ng/ml concentrations and samples diluted into 1% BSA in PBS (100 l) were added to the wells for 1 h at 37°C, the plates washed with Tween-PBS, and incubated with 1 g/ml bHABC for 1 h at 37°C, and washed with Tween-PBS. Horseradish peroxidase-streptavidin complex (Vector) was added for 1 h at 37°C, washed with Tween-PBS, and 1 mg/ml O-phenylenediamine dihydrochloride (Sigma) and 0.03% H 2 O 2 in 0.1 M phosphate citrate buffer, pH 5, were added at 37°C. The reaction was stopped after 15 min with 50 l of 4 M H 2 SO 4 and the absorbance read at 490 nm. Each sample and standard were done in triplicate.
Molecular Mass of Hyaluronan-Aliquots (0.5 ml) of radiolabeled culture medium, trypsin supernatant, and cell extract were subjected to gel filtration on a 1 ϫ 30-cm column of Sephacryl S-1000 (Amersham Biosciences), equilibrated, and eluted at 0.4 ml/min with 0.15 M sodium acetate, 0.1% CHAPS (Sigma), 0.05% Hibitane®, pH 6.8. From each fraction, one aliquot was incubated overnight at 37°C with 12.5 milliunits of Streptomyces hyaluronidase, whereas another received buffer only. Both aliquots were precipitated in 1% cetylpyridinium chloride (Sigma) with 5 g of carrier hyaluronan. The increase of [ 3 H]glucosamine in the supernatant of the hyaluronidase-treated aliquot was a specific measure of hyaluronan. The void volume of the Sephacryl S-1000 column (V 0 ) was considered to be in the first hyaluronan-positive fractions emerging in the chromatogram of Healon GV with the mean molecular mass of 6 ϫ 10 6 Da (Amersham Biosciences), and the total volume (V t ) in the elution position of glucuronic acid. The size distribution of hyaluronan in the samples was estimated from the K av values of known hyaluronan standards, provided by the resin manufacturer.
RT-PCR-Total RNA from monolayer cultures treated with KGF for 3, 6, 12, and 24 h, was isolated using the RNeasy ® Mini kit (Qiagen GmbH, Hilden, Germany), and from the 2-week-old organotypic cultures with the TRIzol® reagent (Invitrogen) according to the manufacturer's instructions, treated with DNase (Roche Molecular Biochemicals), quantitated with a spectrophotometer, and equal amounts taken to the RT-PCR reactions with the GeneAmp® Gold RNA PCR reagent kit (Applied Biosystems, Foster City, CA) using the primers, temperatures, and cycles shown in Table I. The RT-PCR products separated on 1.5% agarose gels were digitized by BioDocII TM video documentation system (Biometra, Göttingen, Germany) and quantitated by ethidium bromide fluorescence by using the NIH Image software.
Cell Proliferation-The REKs were seeded into 24-well plates at 60,000 cells/well, KGF added on the next day, and cells counted 24, 48, and 72 h later using a hemocytometer.
Two-week-old organotypic REK cultures were incubated with 5-bromo-2Ј-deoxyuridine for 1 h, washed with PBS, and fixed overnight with Histochoice®. The deparaffinized sections were first incubated in TUF TM as above, immunostained with an anti-BrdUrd antibody, and counterstained with propidium iodide (0.01 g/ml; Sigma) according to the instructions of the manufacturer (Roche Diagnostics Corp.). Ten fields per culture were counted for labeled nuclei using a 20ϫ objective and 1.25ϫ intermediate lens.
Cell Migration-The cells were seeded at 600,000 cells/well on 6-well plates, and grown for 24 h. A cell-free area was introduced by scraping the monolayer crosswise with a sterile 1000-l pipette tip, and the medium was replaced with new medium containing the appropriate concentration of KGF. The area covered by the cells in eight crossing areas in duplicate wells of each growth factor concentration was measured immediately after scraping and 24 h later using an Olympus CK2 inverted phase contrast microscope (Olympus Optical Co. Ltd. Tokyo, Japan), a Panasonic Wv CD 130-L video camera (Matsushita Electric Works, Tokyo, Japan), and NIH Image software.
Permeability Studies-Permeability was tested by clamping the organotypic cultures between two chambers filled with PBS and equilibrated at 37°C, as described previously (24). Tritiated corticosterone was added to the apical side of the epidermis, and aliquots were withdrawn repeatedly from the basal side for liquid scintillation counting. The permeability coefficient (P, cm/s) was calculated under sink conditions by dividing the steady state flux (dpm/s ϫ cm 2 ) through the epidermis by the concentration of corticosterone (dpm/cm 3 ) in the donor chamber.
Statistical Methods-The statistical significances of the differences between control and KGF-treated groups in the morphometric, migration, and hyaluronan-enzyme-linked immunosorbent assay measurements were tested using paired-samples t test. The data from bromodeoxyuridine labeling and permeability assays were analyzed by the nonparametric Mann-Whitney U test. A difference was considered statistically significant when the p value was less than 0.05.

KGF Changes the Morphology of Keratinocytes Grown as Monolayers, and Stimulates Migration-
The REK cells represent an exceptional keratinocyte cell line with a capacity for stratification and keratinization, but can also be maintained as monolayers if passaged before overt confluency (22,23,27). When cultured on plastic, REK cells of untreated control cultures had a flattened, epithelial morphology, whereas cells treated with KGF often showed a rounded (lift-up) appearance, with an elongated shape (Fig. 1, a and b), typical for migrating cells. Electron microscopy of such cells revealed numerous microvilli on the upper cell surface (Fig. 1g). The migratory phenotype of the KGF-treated cells was confirmed by an in vitro wounding test in which a standardized area in confluent REK cultures was scraped free of cells from ϳ1000-m-wide lanes.  Has1  5Ј-GCT CTA TGG GGC GTT CCT C-3Ј (left)  57°C  35-37  5Ј-CAC ACA TAA GTG GCA GGG TCC-3Ј (right)  Has2  5Ј-TCG GAA CCA CAC TGT TTG GAG TG-3Ј  62°C  33-35  5Ј-CCA GAT GTA AGT GAC TGA TTT GTC CC-3Ј  Has3  5Ј-ACT CTG CAT CGC TGC CTA CC-3Ј  66°C  3 3  5Ј-ACA TGA CTT CAC GCT TGC CC-3Ј  CD44  5Ј-TTG GGG ACT ACT TTG CCT CTT A-3Ј  55°C  3 3  5Ј-CCA CTG CTG ACA TCC TCA TCT A-3Ј  Profilaggrin  5Ј-CTC AGG GCA TCG CTC GTC A-3Ј  64°C  3 4  5Ј-GCT GGT GGC GGT CTT CGT G-3Ј  GAPDH  5Ј-TGA TGC TGG TGC TGA GTA TG-3Ј  60°C  3 2  5Ј-GGT GGA AGA ATG GGA GTT GC-3Ј KGF showed a dose-dependent stimulation of REK migration from the wound edges to the cleared area (Table II). On the other hand, cell proliferation, measured in similar (but nonwounded) cultures by cell counting, was not significantly affected (Table II), suggesting that migration is a major biological function regulated by KGF in these cells.
Hyaluronan Synthesis and Has2 Expression Are Rapidly Upregulated by KGF-Hyaluronan, isolated from the growth medium and assayed as fluorotagged disaccharides, increased in the presence of KGF in a dose-dependent way, and was more than doubled with the higher concentration, thus paralleling the enhanced migratory activity (Table II). The accumulation of hyaluronan in the culture medium was associated with a dose-dependent, 2-3-fold increase in Has activity, assayed in vitro from membrane preparations of REK cells treated with KGF (Fig. 2a). This suggested that the activity of one or more of the Has isoforms produced the higher level of hyaluronan in the REK cultures.
Isolation and analysis by RT-PCR of mRNA from cultures treated with KGF indicated that all three Has isoforms were expressed in the REK monolayer cultures (Fig. 2b). The change of fresh medium alone increased the expression of Has2 and Has3, whereas the signal of Has1 was not markedly affected, as indicated by comparison of the 0-and 3-h controls (Fig. 2b). The levels of Has1 and Has3 were not markedly affected by inclusion of KGF in the fresh medium, but Has2 levels were elevated at all time points (Fig. 2b). The up-regulation was confirmed in additional experiments, showing that Has2 mRNA levels started to increase within 3 h after introduction of KGF, and lasted at least for 24 h (Fig. 2c). This suggested that upregulation of Has2 was mainly responsible for the induction of Has activity and increased concentration of hyaluronan. That hyaluronan synthesis was one of the early targets of KGF in keratinocytes was also supported by quantitation of the newly synthesized hyaluronan after brief (6-h) metabolic labeling (Fig. 2d).  a, c, and f, whereas b, d, e, g, and h represent KGF-treated cultures. In c-e, Streptomyces hyaluronidase treatment after fixation was used to remove pericellular hyaluronan before permeabilization to specifically visualize intracellular hyaluronan. In a-d and f-h, DAB was used as a reporter, whereas in e FITC-avidin was used for hyaluronan (green), and Texas Red-labeled secondary antibody for CD44 (red). Arrowheads in f-h indicate the location of pericellular hyaluronan, and the arrow in h an intracellular hyaluronan-positive vesicle. Magnification bars in a-d, 25 m; in e, 10 m; in f-h, 500 nm. In panels i and j, REKs were treated for 4 -24 h with KGF (100 ng/ml), and optical densities were determined to estimate changes in the intracellular (j) and total cell-associated (pericellular ϩ intracellular) hyaluronan (i). Panel k shows experiments similar to those in j except with a shorter exposure to KGF. In addition, the contribution of extracellular hyaluronan to the KGF-induced intracellular hyaluronan accumulation was tested by keeping Streptomyces hyaluronidase in the culture medium during the incubation. Means Ϯ S.E. of four cultures, each representing the mean of eight randomly picked fields, are shown for treated and intact cultures in panels i and j, whereas panel k shows the means and ranges of two independent experiments.
KGF Increases Intermediate Size Hyaluronan-Because the biological functions of hyaluronan depend on its size, we analyzed the molecular mass of the radiolabeled hyaluronan in the different cellular compartments (Fig. 3). In control cultures, the hyaluronan released in the culture medium and present on the cell surface was largely excluded from the Sephacryl S-1000 gel filtration resin, indicating molecular mass at or above ϳ6 ϫ 10 6 Da. In contrast, the hyaluronan extracted from the intra-cellular sources of control cells contained a relatively low proportion of high molecular mass hyaluronan, most of the intracellular hyaluronan being small fragments, apparently below 90 kDa, as reported previously (28).
In the growth medium, KGF specifically increased hyaluronan chains in the intermediate size range, ϳ0.4 -2 ϫ 10 6 Da, as estimated using the calibration curves supplied by the resin manufacturer. KGF showed no marked influence on the size of cell surface hyaluronan. Intracellularly, a distinct peak of hyaluronan emerged in KGF-treated cultures at ϳ0.1-1.0 ϫ 10 6 Da, a size range almost absent in the controls (Fig. 3), and the low and high molecular mass chains increased as well.
Histochemical and Subcellular Localization of Hyaluronan in KGF-treated Keratinocytes-Histological staining confirmed the accumulation of total hyaluronan in the cell layer (Fig. 1, a  and b) and in the intracellular compartment (Fig. 1, c and d) during a 24-h incubation with 100 ng/ml KGF. The intracellular hyaluronan was specifically detected by removing cell surface-associated hyaluronan using Streptomyces hyaluronidase treatment prior to the permeabilization of the plasma membranes. The location of hyaluronan in the intracellular compartment was confirmed by confocal microscopy (Fig. 1e). The smaller hyaluronan-positive structures were similar to those found in control cultures, whereas the large accumulations were only present in KGF-treated cells (Fig. 1e). Transmission electron microscopy showed that the intracellular hyaluronan was localized in membrane coated vesicles (Fig. 1h). The pericellular hyaluronan that accumulated after 100 ng/ml KGF for 24 h was associated either with the microvilli on the upper surface of the rounded cells, or localized as small spots between the bottom of the cell and the substratum (Fig. 1, f-h,  arrowheads).
Hyaluronan Uptake in Keratinocytes Is Stimulated by KGF-Utilizing the metabolic double labeling with [ 3 H]glucosamine and [ 35 S]sulfate enabled the assay of newly synthesized hyaluronan that accumulated during a 24-h culture period, both intracellularly and on cell surface, compartments that are beyond the sensitivity of the direct chemical assays (Fig. 2e). Interestingly, the pericellular (trypsin-released) hyaluronan pool lacked the dose-dependent increase found in the medium pool, and showed a modest (25%) increase only at the highest (100 g/ml) dose (Fig. 2e). This was in a contrast to the quantity of the intracellular hyaluronan, which showed up to an 8.5-fold increase in the cultures treated with 100 ng/ml KGF, and had doubled already with 1 ng/ml (Fig. 2e). The accumulation of intracellular rather than cell surface hyaluronan suggests that KGF not only stimulates hyaluronan synthesis, but strongly promotes its uptake from the cell surface.
The time course of the response in cell-associated hyaluronan was followed more closely using densitometry of the histochemical stainings (Fig. 1, i-k). Quantitation by image anal-ysis showed that the increase of total cell-associated hyaluronan by KGF was evident after a 24-h treatment, but not after a 6-h treatment (Fig. 1i). In contrast, the intracellular hyaluronan was almost doubled after a 4-h treatment (Fig. 1j). The content of intracellular hyaluronan had actually reached an elevated level by 10 min after KGF addition (Fig. 1k), i.e. at a time point where it is little chance for deposition of newly synthesized hyaluronan by either transcriptional or translational regulation of the synthases. This supports the view that the increased intracellular hyaluronan represented pre-existing molecules endocytosed from plasma membrane as an immediate response to the signals raised by KGF receptors. Furthermore, the presence of hyaluronidase in the culture medium completely prevented the increase (Fig. 1k), indicating that the intracellular hyaluronan must have been exposed to the extracellular milieu before entering the vesicles.
Organotypic REK Cultures Grown in the Presence of KGF Accumulate Hyaluronan and Increase the Expression of Has2, Has3, and CD44 -REKs cultured at the air-liquid interface form a fully organized epidermis with all the strata present in normal epidermis (Fig. 4a). Such cultures also synthesize hyaluronan, a part of which is retained in the epidermis, whereas the rest diffuses down to the supporting collagen gel, and further to the culture medium (25). KGF treatment (20 ng/ml) caused a 2-5-fold increase in the amount of hyaluronan in the epidermis, and a 2-3-fold increase in that released into the underlying matrix (Fig. 4g). To more directly address the impact of KGF on hyaluronan synthesis in the organotypic cultures, we performed a 24-h metabolic labeling experiment that confirmed the increase of newly synthesized hyaluronan (data not shown).
Most of the KGF-induced hyaluronan accumulation resided in the spinous cell layers, whereas a smaller increase was found in the basal cell layer (Fig. 4, c and d). CD44, the main hyaluronan receptor in keratinocytes, was up-regulated by KGF in parallel with the synthesis of hyaluronan and was also increased both in the basal and suprabasal cell layers (Fig. 4, e and f).
In the organotypic cultures, RT-PCR showed not only an increased expression of Has2, but also a marked induction of Has3 (Fig. 4, h and i). Another finding specific for the organotypic cultures was the almost non-existent expression of Has1, whether KGF was present or not (Fig. 4h). In line with the histochemical findings, CD44 mRNA level was up-regulated by KGF, together with Has2 and Has3 (Fig. 4, h and i).
Influence of KGF on Epidermal Growth and Differentiation-KGF did not significantly change the bromodeoxyuridine labeling of the organotypic cultures, suggesting that there was a relatively minor stimulation of cell proliferation in the organotypic cultures (Table III), a finding in line with the lack of increase in cell numbers in monolayer cultures (Table II). Accordingly, whereas there was a trend for increased height of the basal cells, and increased thickness of the whole vital epider-mis in organotypic cultures, the influence of KGF was not statistically significant (Table III).
A marked reduction in the expression of the early epidermal differentiation marker keratin 10 was found in KGF-treated cultures, as indicated by the immunostaining of the tissue sections (Fig. 5, a and b), and Western blots (Fig. 5e). The expression of the late differentiation marker filaggrin showed less reduction than keratin 10 (Fig. 5, c-e), suggesting that KGF retarded but did not inhibit the terminal differentiation. The level of profilaggrin mRNA was also unchanged as ana- lyzed by RT-PCR (data not shown). In the organotypic cultures the epidermal permeability for corticosterone, a tracer often used in testing epidermal barrier function, showed a tendency to increase (not statistically significant) in the KGF-treated cultures (Fig. 5f).

DISCUSSION
The present study established that KGF enhances hyaluronan synthesis in keratinocytes, and that the increased hyaluronan synthesis is associated with an increased expression of the hyaluronan synthases Has2 and Has3. Stimulation of hyaluronan synthesis by KGF was seen both in monolayer cultures containing mainly proliferating and migrating basal cells, and in lifted organotypic cultures, which contain both proliferating basal cells and differentiating spinous and granular cells. The fact that the induction of hyaluronan synthesis was not dependent on the cellular environment or state of differentiation suggests that hyaluronan synthesis is an important target, perhaps one of the primary targets, of KGF. This idea is supported by the rapid up-regulation and maintenance of the elevated level of Has2 mRNA following KGF administration.
One of the main functions of KGF is its contribution to wound healing. The expression of KGF by dermal fibroblasts is rapidly and strongly up-regulated in wounded tissues (9), and the expression of KGF is associated with enhanced wound closure (10,11). Accordingly, animals missing the KGFR function in epidermis show retarded wound healing (33). Our data show that KGF has a strong effect on keratinocyte migration, leading to a 30% faster closure of the epithelium in an in vitro wounding assay, a result in line with experiments on normal human keratinocytes (5). The enhanced migratory response correlated with increased hyaluronan synthesis, suggesting that hyaluronan plays an essential role in the KGF-stimulated motility of keratinocytes. A similar association exists between stimulated hyaluronan synthesis and migration in keratinocytes treated with EGF (15). Furthermore, keratinocytes overexpressing Has2 migrate faster, whereas antisense inhibition of Has2 retards keratinocyte migration (16). Although similar findings with transfected Has genes have been done in other cell types (34,35), the influence of hyaluronan synthesis on motility depends on the cell type (36,37). Nevertheless, in  keratinocytes it is obvious that hyaluronan is an important determinant of the migratory activity, and KGF a major trigger of this response.
The most sensitive cellular targets of KGF in the organotypic cultures were the spinous cells, showing a very intense signal for hyaluronan. The stronger responsiveness of the spinous rather than basal cells is in line with the fact that the level of the KGFRs is highest in the spinous cell layer (3). However, more facile diffusion into the collagen support from the basal cell layer cannot be ruled out as a contributor to the apparently greater augmentation of the hyaluronan signal in the spinous layer. Concomitantly with the higher level of hyaluronan in the spinous cell layer, KGF-treated cultures showed a lower expression of an early differentiation marker (keratin 10), whereas expression of the late differentiation marker filaggrin and the diffusion barrier were less affected. These findings closely correspond to those in human keratinocyte organotypic cultures (6), confirming the general validity of the present culture model.
An inverse correlation between the content of hyaluronan in the spinous cell layer and the indicators of epidermal differentiation has also been noted with other effectors like vitamin A (38) and EGF (39), both of which stimulate hyaluronan synthesis and inhibit differentiation. Conversely, we have found that pharmacological concentrations of hydrocortisone enhance differentiation but inhibit hyaluronan synthesis (40) and Has2 expression. 2 This tight correlation between the status of epidermal differentiation and the synthesis of hyaluronan in the spinous cells may indicate that hyaluronan has a direct, inhibitory impact on keratinocyte entry into terminal differentiation, an issue that deserves further studies.
The fact that the greatest increase of hyaluronan following KGF treatment occurred in the intracellular compartment was somewhat surprising, and is in contrast to the pericellular hyaluronan pool that required a high KGF concentration and long treatment time to reach a more modest increase. The pericellular hyaluronan is assumed to contain the molecules under synthesis, and those associated with cell surface receptors (27). At a time when synthesis is increased, the most plausible explanation for the relatively low total cell surface hyaluronan content is reduction in the receptor-bound pool, perhaps because of a more rapid uptake into the cell. The smaller hyaluronan chains (found in the medium) may be readily endocytosed (rather than remaining resident) after binding to the receptors (41), and cause a backlog in the degradation pathway. Enhanced hyaluronidase activity at the cell surface (42), along with CD44, could enlarge the initial uptake compartment with intermediate size hyaluronan (28), and also reduce the average size of the hyaluronan released into medium. On the other hand, endocytosis generally enhances migration, possibly by facilitating membrane transport to the extending lamellipodium (reviewed in Ref. 43). Whether the increased intracellular hyaluronan content in the highly motile cells simply reflects an enhanced endocytosis activity, or whether it has a more specific function, remains open, but intracellular hyaluronan also accumulates in migratory fibroblasts and smooth muscle cells (44,45).
Interestingly, although hyaluronan synthesis was stimulated by KGF both in the monolayer and in organotypic cultures, the pattern of the Has isoenzymes was different in the two culture models. In the stratified, organotypic cultures both Has2 and Has3 were up-regulated, whereas only Has2 was induced in the monolayers. Previous studies on epidermal keratinocytes have indicated that both Has2 and Has3 are subject to regulation by growth factors and cytokines such as EGF (15,39), interferon ␥ (46), and transforming growth factor ␤ (39,46). The present study indicates that the type of cellular interactions or stage of differentiation clearly modifies the regulation of the Has3 gene.
Altogether, the present study shows that KGF, a growth factor highly induced in most wounds and an important mediator of re-epithelialization and healing, increases the synthesis and deposition of hyaluronan in the epidermis through increased expression of Has2 and Has3 genes. This connection further emphasizes the role of hyaluronan in the physiological regulation of keratinocytes and in the challenges to epidermal homeostasis.