βig-h3 Induces Keratinocyte Differentiation via Modulation of Involucrin and Transglutaminase Expression through the Integrin α3β1 and the Phosphatidylinositol 3-Kinase/Akt Signaling Pathway*

βig-h3 is an extracellular matrix protein whose expression is highly induced by transforming growth factor (TGF)-β1. Whereas βig-h3 is known to mediate keratinocyte adhesion and migration, its effects on keratinocyte differentiation remain unclear. In the present study, it was demonstrated that expression of both βig-h3 and TGF-β1 was enhanced during keratinocyte differentiation and that expression of the former was strongly induced by that of the latter. This study also asked whether changes in β-h3 expression would affect keratinocyte differentiation. Indeed, down-regulation of βig-h3 by transfection with antisense βig-h3 cDNA constructs effectively inhibited keratinocyte differentiation by decreasing the promoter activities and thus expression of involucrin and transglutaminase. The result was a ∼2-fold increase in mitotic capacity of the cells. Conversely, overexpression of βig-h3, either by transfection with βig-h3 expression plasmids or by exposure to recombinant βig-h3, enhanced keratinocyte differentiation by inhibiting cell proliferation and concomitantly increasing involucrin and transglutaminase expression. Recombinant βig-h3 also promoted keratinocyte adhesion through interaction with integrin α3β1. Changes in βig-h3 expression did not affect intracellular calcium levels. Subsequent analysis revealed not only induction of Akt phosphorylation by recombinant βig-h3 but also blockage of Akt phosphorylation by LY294002, an inhibitor of phosphatidylinositol 3-kinase. Taken together, these findings indicate that enhanced βig-h3, induced by enhanced TGF-β during keratinocyte differentiation, provoked cell differentiation by enhancing involucrin and transglutaminase expression through the integrin α3β1 and phosphatidylinositol 3-kinase/Akt signaling pathway. Lastly, it was observed that βig-h3-mediated keratinocyte differentiation was caused by promotion of cell adhesion and not by calcium regulation.

␤ig-h3 is an extracellular matrix protein whose expression is highly induced by transforming growth factor (TGF)-␤1. Whereas ␤ig-h3 is known to mediate keratinocyte adhesion and migration, its effects on keratinocyte differentiation remain unclear. In the present study, it was demonstrated that expression of both ␤ig-h3 and TGF-␤1 was enhanced during keratinocyte differentiation and that expression of the former was strongly induced by that of the latter. This study also asked whether changes in ␤ig-h3 expression would affect keratinocyte differentiation. Indeed, down-regulation of ␤ig-h3 by transfection with antisense ␤ig-h3 cDNA constructs effectively inhibited keratinocyte differentiation by decreasing the promoter activities and thus expression of involucrin and transglutaminase. The result was a ϳ2-fold increase in mitotic capacity of the cells. Conversely, overexpression of ␤ig-h3, either by transfection with ␤ig-h3 expression plasmids or by exposure to recombinant ␤ig-h3, enhanced keratinocyte differentiation by inhibiting cell proliferation and concomitantly increasing involucrin and transglutaminase expression. Recombinant ␤ig-h3 also promoted keratinocyte adhesion through interaction with integrin ␣3␤1. Changes in ␤ig-h3 expression did not affect intracellular calcium levels. Subsequent analysis revealed not only induction of Akt phosphorylation by recombinant ␤ig-h3 but also blockage of Akt phosphorylation by LY294002, an inhibitor of phosphatidylinositol 3-kinase. Taken together, these findings indicate that enhanced ␤ig-h3, induced by enhanced TGF-␤ during keratinocyte differentiation, provoked cell differentiation by enhancing involucrin and transglutaminase expression through the integrin ␣3␤1 and phosphatidylinositol 3-kinase/Akt signaling pathway. Lastly, it was observed that ␤ig-h3-mediated keratinocyte differentiation was caused by promotion of cell adhesion and not by calcium regulation.
Transforming growth factor-␤ (TGF-␤) 1 -inducible gene-h3 (␤ig-h3) was first cloned from A549 lung adenocarcinoma cells that had been stimulated with TGF-␤1 (1,2). ␤ig-h3 has since been shown to be an extracellular matrix protein that can be highly induced by TGF-␤ in several cell types, including mammary epithelial cells, keratinocytes, and lung fibroblasts (1,2). With structural homology to the insect protein fasciclin, ␤ig-h3 is a 76-to 78-kDa protein containing four repeat regions and 11 cysteine residues, mostly clustered in a distinct amino terminus. The ␤ig-h3 molecule appears to undergo partial processing at the carboxyl terminus to yield a 68 -70-kDa isoform (2). Although the ␤ig-h3 transcript has been detected in a variety of human and mouse tissues, including breast, heart, kidney, liver, stomach, and skeletal muscle (2), little information is available regarding the distribution of the protein in such human tissues as the arteries, eye, kidney, lung, and skin (3)(4)(5)(6).
It is known that ␤ig-h3 acts as a cell adhesion molecule in several cell types (7) and as a bifunctional linker protein to connect various matrix molecules to each other and to cells (8,9). ␤ig-h3 contains multiple cell adhesion motifs within its fasciclin-like domains capable of mediating interactions with a variety of cell types via integrins ␣3␤1 (10, 11), ␣1␤1 (12), and ␣v␤5 (7). It is known to mediate the migration and proliferation of normal human epidermal keratinocytes (NHEKs) through two integrin ␣3␤1-interacting motifs in the second and fourth fas-1 domains (10). It has also been shown to bind in vitro to a number of other matrix components, including fibronectin, laminin, and several collagen types (13,14). The precise roles of ␤ig-h3 in cell development are currently unknown, but it has been implicated in cell growth (2,10), osteoblast differentiation (15,16), and wound healing (6). During wound healing, for example, ␤ig-h3 is produced by a range of cell types including activated macrophages, neutrophils, fibroblasts, and keratinocytes (18).
␤ig-h3 has been reported in both the dermis and epidermis, with increased staining intensities in the papillary dermis and granular layer of the epidermis (4). Similarly, TGF-␤ has been localized in the normal dermis and epidermis. The investigators of the present study previously demonstrated that expression of TGF-␤, a potent inducer of differentiation for normal epithelial cells, is increased in and near terminal differentiation of mucosal keratinocytes (19). Moreover, the epidermis was found to display a highly coordinated program of sequential changes in gene expression coincident with the phenotypic evolution from a proliferating basal cell to the mature, nonviable cell (20).
Epidermal and mucosal keratinocytes undergo terminal differentiation when they migrate from the basal layer to the surface (21), and both cell types display a limited number of divisions in vitro. Serial subculture-induced keratinocyte differentiation mimics the physiological maturation process observed in the intact epidermis in vivo (19) and is more similar in some ways with this in vivo process than with calciuminduced differentiation (22). The in vitro differentiation system is thus useful for investigating the mechanisms of keratinocyte differentiation. Many proteins are known or suspected to be associated with the overall process of keratinocyte differentiation. In particular, the roles of involucrin and transglutaminase have been identified as the substrate and enzyme, respectively, required for cornified envelope formation.
Much evidence suggests that TGF-␤ plays important roles during wound healing and keratinocyte differentiation. Indeed, previous work by the present authors has shown that TGF-␤ and phospholipase C-␥1 promote keratinocyte differentiation (19,23). The importance of TGF-␤ suggests a possible role for ␤ig-h3 as well in the mediation of keratinocyte differentiation and in cell adhesion and spreading. This study sought to examine whether changes in ␤ig-h3 expression would affect keratinocyte differentiation, and to determine the molecular mechanism by which this occurred, in an effort to elucidate the role of ␤ig-h3 in the regulation of keratinocyte differentiation. The findings herein demonstrate that enhanced TGF-␤ during keratinocyte differentiation induced ␤ig-h3 expression and thus provoked cell differentiation by enhancing involucrin and transglutaminase expression through the integrin ␣3␤1 and phosphatidylinositol 3-kinase (PI3K)/Akt signaling pathway. Finally, the present study demonstrates that ␤ig-h3-mediated keratinocyte differentiation was caused by promotion of cell adhesion and not by calcium regulation.

EXPERIMENTAL PROCEDURES
Cell Culture-Normal human oral keratinocytes (NHOKs) were prepared and maintained as previously reported (19). Briefly, NHOKs were isolated from human gingival tissue specimens obtained from healthy volunteers (age range 20 to 30 years) who were undergoing oral surgery. Oral keratinocytes were isolated from separated epithelial tissue by trypsinization, and primary cultures were established in keratinocyte growth medium containing 0.15 mM calcium and a supplementary growth factor bullet kit (keratinocyte growth medium; Clonetics, San Diego, CA). Primary NHEKs were prepared in a manner similar to the NHOKs from human foreskins obtained from patients (1 to 3 years of age) undergoing surgery. Approximately 70% confluent primary NHOKs and NHEKs were plated at 1 ϫ 10 5 cells per 60-mm culture dish and cultured until the cells reached 70% confluence. Cells were then subcultured at every 70% confluence until they reached the post-mitotic stage of proliferation, at which time the culture was maintained for 12 days without further passage.
Determination of NHOK and NHEK Culture Population Doublings-Individual keratinocytes isolated from gingival epithelial specimens and foreskins were plated in 60-mm culture dishes. Three days after seeding, the number of cells that had been originally plated was determined by colony counting. Cells were cultured until they reached 70% confluence and then the cell numbers in the various dishes were counted in a hemocytometer. This count allowed determination of the number of population doublings (PDs) of the primary cultures. Harvested primary cultures were subcultured until they reached the postmitotic stage of proliferation. The PD was calculated at the end of each passage according to the formula 2 N ϭ (Cf/Ci), where N denotes the number of PDs; Cf, the total number of cells harvested at the end of a passage; and Ci, the total number of attached cells at seeding.
Transfection, Selection, and CAT Assay-NHOKs were transfected in suspension with antisense ␤ig-h3 or pcDNA3.1(ϩ) vector using a Polybrene/glycerol method (26) and then incubated in keratinocyte growth medium. Beginning 2 days after transfection, cells were incubated for 4 days in 70 M G418 (Invitrogen). Selected cells were serially subcultured until they reached the post-mitotic stage of proliferation and then harvested. The intracellular levels of ␤ig-h3, involucrin, transglutaminase, Akt, phospho-Akt, and calcium were quantitated by Western blotting and dual-wavelength fluorescence imaging, as described elsewhere.
In other experiments of the present study, exponentially proliferating keratinocytes, plated in 35-mm culture dishes, were co-transfected with antisense ␤ig-h3 or pcDNA3.1(ϩ) vector, involucrin-or transglutaminase-1-CAT chimeric plasmid promoter constructs, and 1.0 g of pZeoSVLacZ using the Polybrene/glycerol method (26). pZeoSVLacZ is a ␤-galactosidase expression vector that contains a ␤-galactosidase gene driven by a SV40 promoter and enhancer; this was used as an internal control for normalization of transfection efficiency. The cells were lysed 48 h after co-transfection and the cell extracts assayed for CAT and ␤-galactosidase activities. Detailed conditions for the enzyme activity assay are described elsewhere (27). A pCAT-control vector (Promega) containing a SV40 promoter and enhancer, shown to be unresponsive, was included as a control in each transfection experiment.
Adhesion Inhibition Assay-To identify the ␤ig-h3 receptor on NHOKs, 5 g/ml of mAbs to different types of integrins were incubated with exponentially proliferating (PD 13) or terminally differentiated (PD 20) NHOKs in 0.5 ml of incubation solution (2 ϫ 10 5 cells/ml) for 30 min at 37°C. These preincubated cells were transferred onto plates coated with 10 g/ml of recombinant ␤ig-h3 protein (plates had been protein-coated and stored overnight at 4°C) and then incubated for an additional 1 h at 37°C. After incubation, unattached cells were removed by two rinses with PBS. Attached cells were fixed with 10% formalin in PBS for 15 min, rinsed twice with PBS, and stained with 0.005% crystal violet for 1 h. The plates were gently rinsed with double distilled water three times. To ensure a representative count, each plate was divided into quarters and two fields per quarter were photographed using an Olympus BX51 microscope at ϫ100 magnification.
Cell Adhesion Assay-Cell adhesion was assayed as described previously (29). Briefly, 24-well culture plates (Nunc, Roskilde, Denmark) were coated with 10 g/ml recombinant ␤ig-h3 protein and held overnight at 4°C before being blocked with PBS containing 1% heat-inactivated bovine serum albumin (BSA; Sigma) for 1 h at 37°C. Cells were detached by treatment with 0.05% trypsin and 0.53 mM EDTA in PBS, resuspended in the culture media (1 ϫ 10 5 cells/500 l), added to each plate, and incubated for 1 h at 37°C. Removal of unattached cells and fixing, staining, and imaging of attached cells was identical as in the adhesion inhibition assay.
Effects of Exogenous TGF-␤1 and Recombinant ␤ig-h3 Protein on Keratinocyte Differentiation-To determine the effects of TGF-␤1 and recombinant ␤ig-h3 protein on keratinocyte differentiation and the PI3K/Akt signaling pathway, the expression of involucrin, transglutaminase, and ␤ig-h3; the promoter activities of involucrin and transglutaminase; the intracellular calcium level; and the phosphorylation status of Akt were determined in exponentially proliferating NHOKs, which were cultured for 4 days in the presence of 10 or 20 ng/ml TGF-␤1 or 10 g/ml recombinant ␤ig-h3 protein.
Measurement of Intracellular Calcium-After cells were transfected with either the antisense ␤ig-h3 cDNA constructs or ␤ig-h3 expression plasmids and exposed to recombinant ␤ig-h3 protein, the intracellular calcium level was measured by digital video microfluorimetry using an intensified CCD camera coupled to a microscope (Olympus 1X71S8F-2, Japan) supported by Metafluor software on a Pentium computer (Shutter Instrument, Novato, CA). This procedure has been described previously (23).
Western Blot Analysis-Western blot analysis was performed as in previous reports (19) using anti-human involucrin (SY5) mAb (Sigma), anti-human TGF-␤1 (sc-146) polyclonal antibody (Santa Cruz, Santa Cruz, CA), anti-human transglutaminase (Ab-1) polyclonal antibody (Oncogene, Uniondale, NY), polyclonal anti-␤ig-h3 antiserum against recombinant ␤ig-h3 protein (REGEN Biotech, Korea), and anti-␤-actin (20-33) mAb (Sigma). After probing with each antibody, the membrane was stained with 1ϫ Ponceau S stain for 10 min or subjected to immunoblot analysis using a specific antibody to ␤-actin for determination of the total protein loaded per lane. The relative protein levels were determined as follows. First, the densitometric intensities of the target protein or ␤-actin from the Western blots and the constitutive protein from the Ponceau S stain were measured with a LAS-1000 Plus (Fuji Photo film, Japan). The ratio of target protein to constitutive protein or ␤-actin for each sample was then calculated to correct for differences in protein loading. This ratio was normalized to the target protein/constitutive protein or ␤-actin ratio of the negative controls to obtain the relative levels of protein.
Akt Phosphorylation Assay-To determine the effects of TGF-␤1, recombinant ␤ig-h3 protein, and changes in ␤ig-h3 expression on the phosphorylation status of Akt, exponentially proliferating NHOKs were cultured for 4 days in the presence of 10 or 20 ng/ml TGF-␤1 or 10 g/ml recombinant ␤ig-h3 proteins. Exponentially proliferating keratinocytes were also transfected with either the antisense ␤ig-h3 cDNA constructs or ␤ig-h3 expression plasmids. The phosphorylation status of Akt was determined by Western blot analysis using anti-phospho-Akt (Ser 473 , 9271) and anti-Akt (9272) (Cell signaling Technology, Beverly, MA).

Expression of ␤ig-h3 and TGF-␤1 Is Enhanced during Keratinocyte Differentiation-Primary
NHOKs were subcultured at every 70% confluence until they reached the post-mitotic stage of proliferation, at which time the culture was maintained for 12 days without further passage. Primary NHOKs (PD 13.4) and cells with lower PDs (PD 15.8) displayed the typical keratinocyte morphology and retained an undifferentiated phenotype. Some cells with PD 16.9, meanwhile, maintained their squamous cell shape morphology. Cells with PD 17.3 displayed characteristics of terminal differentiation, including highly enlarged and elongated cytoplasm and balling and detaching from the culture dish (Fig. 1A). Expression of the protein involucrin, one of the molecular markers for keratinocyte differentiation (31,32), increased in NHOKs in conjunction with increasing PD (Fig. 1B). These results are consistent with these authors' previous report (19) and indicate that serial subculture of primary NHOKs induces keratinocyte differentiation.
Because TGF-␤ is known to induce both ␤ig-h3 expression in keratinocytes (1, 2) and cell differentiation (19), the expression of TGF-␤ and ␤ig-h3 was determined in serially subcultured NHOKs. Because TGF-␤ and ␤ig-h3 are secreted proteins, we determined their levels in cell lysates and the NHOK medium with different PDs. Western blotting of culture media showed that the exponentially proliferating NHOKs with PD 13.4 had very low levels of ␤ig-h3 and TGF-␤1 proteins. Expression of both of these proteins was notably increased, however, in the terminally differentiated NHOKs with PD 17.3 ( Fig. 1, C and D). Similarly, the level of intracellular 68-kDa ␤ig-h3 was significantly enhanced in terminally differentiated NHOKs (Fig. 1B). These results suggest that variations in ␤ig-h3 expression are closely correlated with keratinocyte differentiation.
TGF-␤ Induces ␤ig-h3 Expression and Promotes Keratinocyte Differentiation-␤ig-h3 is known to be induced by TGF-␤1 in many but not all cell types. The present study showed that TGF-␤ induces ␤ig-h3 expression in NHOKs. When exponentially proliferating NHOKs (PD 13.4) were cultured for 6, 12, 24, 48, 72, or 96 h in the presence of 10 ng/ml TGF-␤1, the levels of both ␤ig-h3 protein and ␤ig-h3 transcript were increased in a time-dependent manner compared with untreated cells (Fig. 2A). A second investigation into the effects of TGF-␤1 on keratinocyte differentiation and ␤ig-h3 expression had NHOKs with PD 13.4, cells that do not normally show a differentiated phenotype, being treated with 10 or 20 ng/ml TGF-␤1 for 96 h. TGF-␤1 treatment in keratinocytes resulted in loosening of cell-cell contact as well as cell spreading. Characteristics of differentiation were also exhibited (Fig. 2B). Western blotting showed that TGF-␤1 treatment induced ␤ig-h3 expression in cell lysates and culture media and led to significantly increased expression of involucrin and transglutaminase ( promotes their differentiation. This agrees with a previous study by these authors showing TGF-␤ expression to be significantly enhanced in and near terminally differentiated NHOKs and to be associated with keratinocyte differentiation (19).
Expression of ␤ig-h3 during Epidermal Differentiation, Response of NHEKs to TGF-␤, and the Role of TGF-␤ in Epidermal Differentiation Are Very Similar to That of NHOK Counterparts-There are some biological differences in the oral mucosa versus the skin. Therefore, we determined the expression of ␤ig-h3 during epidermal differentiation and tested the response of NHEKs to TGF-␤1 and its role in epidermal differentiation. Primary NHEKs were subcultured at 70% confluence until they reached the post-mitotic stage of proliferation, at which time the culture was maintained for 12 days without further passage. Like primary NHOKs, the serial subculture of primary NHEKs displayed characteristics of terminal differentiation, including cells with highly enlarged and elongated cytoplasm (data not shown). Involucrin expression increased in NHEKs in conjunction with increasing PD (Fig. 3A), indicating that serial subculture of primary NHEKs induces keratinocyte differentiation. Similarly, the level of intracellular ␤ig-h3 was enhanced in terminally differentiated NHEKs (Fig. 3A), suggesting that variations in ␤ig-h3 expression are also correlated with epidermal keratinocyte differentiation. When exponentially proliferating NHEKs (PD 13.5) were cultured for 6, 12, 24, 48, 72, or 96 h in the presence of 10 ng/ml TGF-␤1, the levels of both ␤ig-h3 protein and ␤ig-h3 transcript increased in a time-dependent manner compared with untreated cells (Fig. 3B). Western blot-ting showed that TGF-␤1 treatment induced ␤ig-h3 expression in cell lysates and culture media, which led to significantly increased expression of involucrin and transglutaminase (Fig. 3C). Such changes were very similar to that of NHOKs. These results indicate that TGF-␤1 treatment induces ␤ig-h3 expression in NHEKs and promotes their differentiation.
Suppression of ␤ig-h3 Expression Inhibits Keratinocyte Differentiation and Extends the Mitotic Capacity of Cells-To test whether ␤ig-h3 could mediate keratinocyte differentiation, and to study the underlying mechanisms of ␤ig-h3-mediated cell differentiation, expression of ␤ig-h3 was down-regulated by transfecting NHOKs with antisense ␤ig-h3 cDNA constructs. If ␤ig-h3 is required for the induction of keratinocyte differentiation, cells transfected with antisense ␤ig-h3 constructs should fail to induce keratinocyte differentiation. NHOK cultures (PD 11.3) were transfected with antisense ␤ig-h3 constructs, and transfectants were selected with G418 and further cultured for 12 days. The transfected cells showed a significant reduction in expression of ␤ig-h3 versus vector-transfected cells. Furthermore, NHOKs transfected with the antisense ␤ig-h3 retained an undifferentiated phenotype, whereas some cells transfected with the vector alone were enlarged and flattened (Fig. 4, A and  B). Similarly, the levels of involucrin and transglutaminase in the antisense-␤ig-h3-transfected cells were significantly reduced (Fig. 4B), indicating that suppression of ␤ig-h3 inhibited keratinocyte differentiation.
An explanation of how ␤ig-h3 expression causes inhibition of keratinocyte differentiation was sought by evaluation of involucrin and transglutaminase promoter activity following co-transfection of NHOKs with an antisense ␤ig-h3 construct and involucrin or transglutaminase promoter constructs. As predicted, involucrin and transglutaminase promoter activities were significantly inhibited in the antisense ␤ig-h3 construct-transfected cells compared with vector-transfected cells (Fig. 4C).
Because decreased ␤ig-h3 expression did inhibit keratinocyte differentiation, it was further investigated whether a decrease of ␤ig-h3 would also affect the mitotic ability of cells, in essence extending the in vitro lifespan of NHOKs. NHOK cultures (PD 12.1) were transfected with an antisense ␤ig-h3 construct, selected with G418, and subcultured until they reached the post-mitotic stage of proliferation. As shown in Fig.  4D, the in vitro lifespan of cells that were transfected with the antisense ␤ig-h3 construct was increased ϳ2-fold over vectortransfected cells. This increase in mitotic ability was presumably a consequence of decreased ␤ig-h3 expression. These results indicate that suppression of ␤ig-h3 expression inhibits keratinocyte differentiation by decreasing involucrin and transglutaminase promoter activities and also extends the mitotic capacity of the cells.
Overexpression of ␤ig-h3 and Recombinant ␤ig-h3 Protein Promote Keratinocyte Differentiation by Increasing Involucrin and Transglutaminase Promoter Activities-The role of ␤ig-h3 in cell differentiation was evaluated by examining the effect of overexpression of ␤ig-h3 in keratinocytes. For this, NHOK cultures (PD 11.8) were transfected with ␤ig-h3 expression plasmids to increase ␤ig-h3 expression and transfectants were selected with G418, further cultured for 12 days, and assayed for their levels of ␤ig-h3 and involucrin. Most cells transfected with vector displayed typical keratinocyte morphology and an undifferentiated phenotype, although some were enlarged and flattened. When NHOK cultures were transfected with the ␤ig-h3 expression plasmids, the majority of cells displayed the characteristics of terminal differentiation, principally a highly enlarged and elongated cytoplasm (Fig. 5A). Furthermore, the levels of involucrin as well as ␤ig-h3 were significantly increased in the cells transfected with ␤ig-h3 expression plasmids compared with vector-transfected cells (Fig. 5B), indicating that overexpression of ␤ig-h3 induces keratinocyte differentiation.
If ␤ig-h3 protein is required for the induction of involucrin and transglutaminase gene expression, their promoter activities should be increased in cells transfected with the ␤ig-h3 expression plasmids. As expected, transfection of cells with the ␤ig-h3 expression plasmids significantly increased both involucrin and transglutaminase promoter activities compared with vector-transfected cells (Fig. 5C).
A direct causal role for ␤ig-h3 in keratinocyte differentiation was further tested by examining the effect of recombinant ␤ig-h3 protein on NHOKs. When NHOKs (PD 12.4) were cultured for 4 days in the presence of 10 g/ml recombinant ␤ig-h3, the cells displayed a differentiated phenotype (Fig. 6A) and expressed higher levels of involucrin and transglutaminase than did untreated control cells (Fig. 6B). Generally, cell growth is vital to cell differentiation, so the influence of recombinant ␤ig-h3 on cell growth was also evaluated: treatment with 10 g/ml recombinant ␤ig-h3 decreased cell growth (Fig.  6C). Taken together, these observations indicate that overexpression of ␤ig-h3 and recombinant ␤ig-h3 promotes keratinocyte differentiation by increasing involucrin and transglutaminase expression and by inhibiting cell growth.
Changes in ␤ig-h3 Expression Do Not Affect Intracellular Calcium Levels-Calcium is well known to induce differentiation of human keratinocytes (34). To study the involvement of ␤ig-h3 in calcium regulation, changes in ␤ig-h3 expression were evaluated for their affects on intracellular calcium levels in NHOKs. Neither down-regulation of ␤ig-h3 (transfection with antisense ␤ig-h3) nor overexpression of ␤ig-h3 (transfection with ␤ig-h3 expression plasmids) changed the intracellular calcium level compared with vector-transfected cells (data not shown). Similarly, stimulation with 10 g/ml recombinant ␤ig-h3 for 6 min in serum-free medium did not affect the intracellular calcium level (Fig. 6D). In general, these results suggest that calcium is not involved in ␤ig-h3-mediated keratinocyte differentiation.
␤ig-h3 Mediates Keratinocyte Adhesion Through Integrin ␣3␤1-After establishing that ␤ig-h3 mediates keratinocyte FIG. 3. Expression of ␤ig-h3 during epidermal differentiation, response of NHEKs to TGF-␤, and the role of TGF-␤ in epidermal differentiation are very similar to that of NHOKs. Primary NHEKs (PD 13.5) were serially subcultured until they reached the postmitotic stage of cell proliferation, at which time the culture was maintained for 12 days without further passage as described in the legend to Fig. 1. A, levels of involucrin and ␤ig-h3 proteins in NHEKs with different PDs. B, levels of ␤ig-h3 transcript and protein in exponentially proliferating NHEKs treated with 10 ng/ml of TGF-␤1. Assay conditions were the same as described in the legend to Fig. 2A, except that NHEKs (PD 13.5) were used. A structural protein stained with Ponceau S served as the internal control to account for loading error. 28 S RNA was used as a control. C, levels of involucrin, transglutaminase, and ␤ig-h3 proteins in cell lysates and/or serum-free culture media from NHEKs cultured for 96 h in the presence of 10 or 20 ng/ml TGF-␤1. ␤ig-h3 Induces Keratinocyte Differentiation differentiation, this study turned to whether ␤ig-h3 induces said differentiation by promoting cell adhesion. Accordingly, a cell adhesion assay was devised. Exponentially proliferating (PD 13) and terminally differentiated (PD 20) NHOKs were seeded onto 24-well plates coated with recombinant ␤ig-h3 or BSA. As shown in Fig. 7A, ␤ig-h3 was promotive of cell adhesion of both types of NHOKs. These findings are in accordance with a previous study that showed ␤ig-h3 supported adhesion and spreading of endothelial epithelial cells as well as NHEKs (10,35). To further investigate whether restraint of ␤ig-h3 expression modifies cell adhesion, expression of ␤ig-h3 was reduced by transfecting NHOKs with antisense ␤ig-h3 cDNA constructs. NHOK cultures (PD 12) were transfected with antisense ␤ig-h3 constructs, and transfectants were selected with G418, then cultured for 12 days. When the transfected cells were seeded onto 96-well plates coated with or without recombinant ␤ig-h3, cell adhesion was significantly inhibited in antisense ␤ig-h3 construct-transfected cells compared with vector-transfected cells in both conditions (Fig. 7B). Next, in the absence of ␤ig-h3, the effects of TGF-␤ on keratinocyte adhesion were tested. Exponentially proliferating NHOKs (PD 11) were pretreated with different concentrations of TGF-␤1 for 24 h and then were used for cell adhesion assay. TGF-␤1treated cells adhered to ␤ig-h3-uncoated 24-well plates in a dose-dependent manner (Fig. 7C). Taken together, these results suggest that a change in cell adhesion properties is required for keratinocyte differentiation to proceed, and that ␤ig-h3, in response to TGF-␤1, mediates cell adhesion to allow expression of keratinocyte differentiation-related genes.
In conjunction with its role in cell adhesion, ␤ig-h3 has been shown to mediate several cellular functions through integrins (7, 10 -12). A pursuit was therefore set upon to identify the particular integrin or integrins by which adhesion, spreading, and differentiation of NHOKs are mediated by ␤ig-h3. The first step was to determine those integrins that are indeed expressed on the surface of NHOKs. This was accomplished by fluorescence-activated cell sorter using mAbs specific to each integrin type. Fig. 7 (D and E) shows that exponentially proliferating and terminally differentiated NHOKs expressed several integrins, including ␣3␤1 and ␣6␤4. The terminally differentiated NHOKs showed lower surface expressions of ␣ and ␤ integrin subunits than exponentially proliferating NHOKs (Fig. 7E). Next, monoclonal integrin function-blocking antibodies were employed with the result that only antibodies to integrin subunits ␣3 and ␤1 significantly inhibited ␤ig-h3-mediated adhesion of both types of NHOKs (Fig. 7, F and G). These findings suggest that ␤ig-h3 mediates adhesion of human oral keratinocytes and might mediate keratinocyte differentiation via integrin ␣3␤1.
␤ig-h3 Activates Phosphorylation of Akt-Currently, the identities of integrin ␣3␤1-associated signaling molecules that are responsible for mediating keratinocyte adhesion and/or differentiation in response to ␤ig-h3 are unclear. It has been shown that the activation of integrins upon cell adhesion to extracellular matrix protein leads to an increase in phosphorylation of focal adhesion kinase, which is the major tyrosine-phosphorylated protein (36 -38). In fact, the focal adhesion kinase pathway is activated by most integrins. However, the downstream signaling pathways that mediate integrin-focal adhesion kinase survival signaling are diverse, and the factors determining which pathway is utilized remain obscure. To determine the signaling pathways that contribute to NHOK adhesion and differentiation as induced by ␤ig-h3, an examination was conducted into the effects of ␤ig-h3 on the phosphorylation status of Akt and MAPKs (ERK, JNK, and p38). Overexpression of ␤ig-h3 or treatment with 10 g/ml recombinant ␤ig-h3 was found to induce phosphorylation of Akt (Fig. 8, A and B), although these had no effect on the phosphorylation of ERK, JNK, or p38 (data not shown). To further test whether PI3K/Akt is involved in ␤ig-h3-mediated NHOK adhesion and differentiation, NHOK cultures (PD 12.4) were adapted for 12 h, pretreated with 20 M LY294002, a synthetic inhibitor of the p110 catalytic subunit of PI3K, for 1 h, and then stimulated with 10 g/ml recombinant ␤ig-h3 for 5, 10, 20, 30, or 60 min. Whereas expression of total Akt did not appear to change, phospho-Akt was increased in cells treated with ␤ig-h3 without pretreatment by LY294002 compared with untreated cells (Fig. 8C). In contrast, LY294002 dramatically blocked the phosphorylation of Akt when paired with ␤ig-h3 treatment (Fig. 8D). Therefore, this study examined whether ␤ig-h3 affects keratinocyte differentiation when the phospho-Akt level is blocked by LY294002. For this, NHOK cultures (PD 11.6) were adapted for 12 h, pretreated with 20 M LY294002 for 1 h, stimulated with 10 g/ml recombinant ␤ig-h3 in the presence of 20 M LY294002 for 2 days, and then assayed for involucrin and transglutaminase expression and the phos-phorylation status of Akt. When LY294002 blocked the phosphorylation of Akt (data not shown), the levels of involucrin and transglutaminase expression were notably decreased in recombinant ␤ig-h3-treated cells pretreated with LY294002 compared with recombinant ␤ig-h3-treated cells without pretreatment (Fig. 8E). Taken together, these data suggest that the PI3K/Akt signaling pathway is associated with ␤ig-h3mediated keratinocyte differentiation. DISCUSSION The authors of the present study previously demonstrated that TGF-␤, a potent inducer of differentiation for normal epithelial cells known to localize in the normal epidermis (4), also induces keratinocyte differentiation (19). Here, it was observed that ␤ig-h3 and TGF-␤1 expression were significantly enhanced during keratinocyte differentiation and that TGF-␤1 exposure induced keratinocyte differentiation as well as ␤ig-h3 mRNA and protein expression. This study thus examined whether changes in ␤ig-h3 expression would affect keratinocyte differentiation.
The role of ␤ig-h3 in mediating the differentiation of keratinocytes was demonstrated as follows: (i) a decrease in ␤ig-h3 Adhesion of the exponentially proliferating (F) and terminally differentiated (G) NHOKs to ␤ig-h3 is blocked by antibodies to ␣3␤1 integrin. Cells were preincubated with 5 g/ml of the function-blocking mAbs to integrin ␣2, ␣3, ␣5, ␣6, ␣v, ␤1, and ␤4 subunits for 30 min at 37°C and then incubated for 1 h on plates precoated with 10 g/ml recombinant ␤ig-h3. Unbound cells were washed away and the adherent cells fixed, stained with crystal violet, and counted. Data are expressed as percentages of the value for cells preincubated without integrin antibodies (None; mean Ϯ S.D., n ϭ 3). expression led to inhibition of keratinocyte differentiation and a ϳ2-fold increase in mitotic capacity; (ii) down-regulation of ␤ig-h3 expression resulted in decreased promoter activities and thus expression of the involucrin and transglutaminase genes; (iii) overexpression of ␤ig-h3 significantly promoted keratinocyte differentiation and increased promoter activities and thus expression of the involucrin and transglutaminase genes. Together, these results demonstrate that ␤ig-h3 induces keratinocyte differentiation by increasing promoter activities and expression of involucrin and transglutaminase, markers of keratinocyte differentiation (31)(32)(33) that serve as substrate and enzyme for cornified envelope formation.
A growing body of reports identifies ␤ig-h3 as a cell adhesion substrate, regulator of cell growth, linker of matrix components, and transducer of TGF-␤-mediated signaling (1,4,7,8). Although the precise functions of ␤ig-h3 in cell development are currently unknown, it has been associated with inhibition of osteoblast differentiation (15,16,39). Reports show the level of ␤ig-h3 mRNA to be decreased in human bone marrow stromal cells treated with dexamethasone, a promoter of osteogenic cell differentiation (15). ␤ig-h3 is also reported to inhibit bone nodule formation of mouse osteoblasts (16) and human periodontal ligament cells in vitro (39). Nevertheless, the role of ␤ig-h3 on keratinocyte differentiation remains unknown.
Whereas one published study reports in vivo expression of ␤ig-h3 in the granular layer of the epidermis (4), another reports that in situ hybridization could not detect ␤ig-h3 expression in that tissue but rather detected ␤ig-h3 transcripts in basal keratinocytes (10). Numerous studies in various cell types, including keratinocytes, involve TGF-␤1 in the induction of ␤ig-h3 (1, 2). TGF-␤ has been localized in the normal dermis and epidermis and TGF-␤1 is constitutively expressed in suprabasal keratinocytes (30). Although there is no direct evidence showing that TGF-␤1 produced by suprabasal keratinocytes stimulates basal keratinocytes to synthesize ␤ig-h3 in normal skin and mucosa, expression of ␤ig-h3 might be regulated by TGF-␤ in the epidermis. Participation of ␤ig-h3 in keratinocyte differentiation is supported by our results, namely that overexpression of ␤ig-h3 by TGF-␤1 treatment promoted keratinocyte differentiation.
The present findings agree with these authors previous report that expression of TGF-␤ is significantly enhanced in and near terminally differentiated NHOKs and is associated with keratinocyte differentiation (19). These results were confirmed using recombinant ␤ig-h3 protein: ␤ig-h3 treatment resulted in both an increase of involucrin and transglutaminase expression and a differentiated phenotype in normal human keratinocytes. Here, it was also shown that ␤ig-h3 was capable of decreasing cell growth. As has been reported by others, overexpression of ␤ig-h3 was found to decrease cellular growth rate (2). This further suggests that ␤ig-h3 might be involved in differentiation.
One mechanism of action for ␤ig-h3 in keratinocyte differentiation would be to promote cell adhesion. In fact, ␤ig-h3 has been reported to mediate cell adhesion in several cell types such as skin fibroblasts (4), epidermal keratinocytes (10), corneal epithelial cells (6,7), and chondrocytes (12). In studies with NHOKs, ␤ig-h3 was also found to promote keratinocyte adhesion. To further investigate whether reduction of ␤ig-h3 expression and TGF-␤1 modify cell adhesion, NHOKs transfected with antisense ␤ig-h3 cDNA constructs and NHOKs pretreated with TGF-␤1 were examined. Cell adhesion was significantly inhibited in cells with reduced ␤ig-h3 compared with vector-transfected cells, but cell adhesion was significantly enhanced in TGF-␤1-pretreated cells. This suggests that ␤ig-h3-mediated keratinocyte differentiation is caused by the promotion of cell adhesion. The results presented here further indicate that ␤ig-h3 might mediate NHOK adhesion and differentiation by interacting with integrin ␣3␤1. Integrin ␣3␤1 is a known receptor for ␤ig-h3 (7, 10). Recently, TGF-␤1 has been shown to increase affinity of the integrin ␣3␤1 for ␤ig-h3, resulting in enhanced adhesion and migration of epidermal keratinocytes toward ␤ig-h3 (17).
Calcium action is another possible mechanism for keratinocyte differentiation, as it is well known to induce differentiation of human keratinocytes. Specifically, calcium concentrations higher than 0.1 mM in epidermal keratinocytes or 0.15 mM in oral keratinocytes induce terminal differentiation of human basal keratinocytes in vitro. Furthermore, a gradient of extracellular and intracellular calcium across the epidermis, with a lower concentration of calcium in the basal cell compartment than in the granular cell layer, is recognized as a principle regulator of keratinocyte maturation in vivo (33,34). These investigators have previously shown that intracellular calcium concentration gradually increases during NHOK differentiation (23) and that treatment with calcium induces said differentiation (22). In terms of intracellular calcium level, changes in ␤ig-h3 expression or exposure to recombinant ␤ig-h3 were FIG. 8. Overexpression of ␤ig-h3 by transfection with ␤ig-h3 expression plasmids and recombinant ␤ig-h3 activate the PI3K/ Akt pathway. A, the effect of overexpression of ␤ig-h3 on the phosphorylation status of Akt in keratinocytes. Assay conditions were the same as those described in the legend to Fig. 5A. Cell lysates were immunoblotted with antibodies for phospho-Ser 473 Akt or total Akt. B, recombinant ␤ig-h3 activates the phosphorylation of Akt in keratinocytes. NHOKs with PD 12.4 were cultured for 4 days in the presence of 10 g/ml recombinant ␤ig-h3. C, NHOK cultures (PD 13.8) were adapted for 12 h and then stimulated with 10 g/ml recombinant ␤ig-h3 for 5, 10, 20, 30, or 60 min in serum-free medium. D, pretreatment with LY294002, a PI3K inhibitor, blocks recombinant ␤ig-h3-induced phosphorylation of Akt in keratinocytes. NHOK cultures (PD 13.8) were adapted for 12 h, pretreated with 20 M LY294002 for 1 h, and stimulated with 10 g/ml recombinant ␤ig-h3 for 5, 10, 20, 30, or 60 min in serum-free medium. E, pretreatment with LY294002 blocks recombinant ␤ig-h3-induced expression of involucrin and transglutaminase in keratinocytes. Assay conditions were the same as in D, except that NHOKs with PD 11.6 were used and that LY294002-pretreated cells were stimulated with 10 g/ml recombinant ␤ig-h3 for 2 days in the presence of LY294002.
found to have no effect, suggesting that ␤ig-h3-mediated keratinocyte differentiation is not caused by calcium regulation.
The phosphorylation status of MAPKs (ERK, JNK, p38) and Akt, a downstream target of PI3K at Ser 473 , was examined in cells transfected with the wild-type ␤ig-h3 gene and in cells treated with recombinant ␤ig-h3 protein so as to elucidate something of the signaling pathway of ␤ig-h3-mediated keratinocyte differentiation. Both types of cells up-regulated for ␤ig-h3 showed a higher level of phosphorylated Akt than did vector-transfected and ␤ig-h3-untreated cells, respectively. The phosphorylation status of MAPKs in the ␤ig-h3 up-regulated cells was similar to that of the control (data not shown). Subsequent analysis revealed blocking of Akt phosphorylation by LY294002, an inhibitor of PI3K. The notion that exogenous calcium administration might induce keratinocyte differentiation by activating PI3K/Akt was tested and found not to be the case. Calcium did not affect Akt phosphorylation (data not shown).
It has been reported that, because the ␤ig-h3 gene is induced in several cell lines whose proliferation was affected by TGF-␤, ␤ig-h3 might be involved in mediating some of the signals of TGF-␤ (1). One recent study reports that TGF-␤1 activates focal adhesion kinase and Akt in epidermal keratinocytes and that PI3K could be a common downstream factor that transduces signals from TGF-␤1 and leads to activation of the integrin ␣3␤1 by ␤ig-h3 (17). Because overexpression of ␤ig-h3 or treatment with recombinant ␤ig-h3 was found to activate the PI3K/Akt pathway, decrease cell proliferation, and promote keratinocyte differentiation, we suggest that ␤ig-h3 serves as a downstream factor of TGF-␤ in keratinocyte differentiation. Whether TGF-␤1-induced ␤ig-h3 expression completely accounts for the increase in involucrin and transglutaminase levels observed with TGF-␤1 remains unknown. Nevertheless, exogenous recombinant ␤ig-h3 administration in exponentially proliferating keratinocytes, which do not normally show the differentiation phenotype, was sufficient to induce differentiation. We propose that ␤ig-h3 affects keratinocyte differentiation by promoting involucrin and transglutaminase expression through a molecular pathway that involves both integrin ␣3␤1 and PI3K/Akt (Fig. 9).
Considering that ␤ig-h3 and TGF-␤1 expression are significantly enhanced during keratinocyte differentiation, and that TGF-␤1 exposure induces said differentiation in addition to ␤ig-h3 mRNA and protein expression, ␤ig-h3 must be granted an important role in keratinocyte differentiation. This role is at least partially explained by the findings of the present study: (i) enhanced TGF-␤ during keratinocyte differentiation induced ␤ig-h3 expression and led to keratinocyte differentiation by enhancing involucrin and transglutaminase expression, (ii) involucrin and transglutaminase expression were mediated through the integrin ␣3␤1 and PI3K/Akt signaling pathway, and (iii) ␤ig-h3-mediated keratinocyte differentiation was caused by enhanced cell adhesion, not by calcium regulation.