Regulation of Transglutaminase Type II by Transforming Growth Factor-01 in Normal and Transformed Human Epidermal Keratinocytes*

This study examines the effect of transforming growth factor-B1 (TGF-01) on the expression of Type I and II transglutaminase in normal human epidermal keratinocytes (NHEK cells). Treatment of undifferen- tiated NHEK cells with 100 pM TGF-/31 caused a lo- to 15-fold increase in the activity of a soluble transglu- taminase. Based on its cellular distribution and im- munoreactivity this transglutaminase was identified as Type II (tissue) transglutaminase. TGF-Bl did not the of the Type I (ep-idermal) transglutaminase activity which is during squamous differentiation Type transglutaminase differen- Several

The transforming growth factor-p (TGF# family consists of an increasing number of related, but functionally distinct proteins (1,2). One member of this family, TGF-81, is produced and secreted as a latent, high molecular weight complex * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.&C. Section 1734 solely to indicate this fact. by many normal and neoplastic cells (l-3). TGF-B1 is synthesized as a 390-amino acid precursor polypeptide which is both glycosylated and phosphorylated (4,5). Proteolytic cleavage at the carboxyl terminus yields the mature 112-amino acid TGF-/!l monomer; active TGF-/31 consists of two identical monomer units linked by disulfide bridges (2). The complete amino acid sequence of normal and mouse TGF-fil, which are deduced from their cDNA sequences, are remarkably homologous (6,7). TGF-/32 has 72% amino acid homology with TGF$l and induces many of the same effects (2). TGF-/31 and TGF-P2 regulate cellular proliferation, differentiation, and other functions in many cell types (1,2). They can inhibit as well as stimulate cell proliferation, depending on the cell line/type and the growth conditions (8)(9)(10)(11)(12). TGF-,Bl is a promoter of chondrogenesis and a stimulator of angiogenesis (13,14). In tracheobronchial epithelial cells TGF-@l induces terminal cell division and expression of the squamous differentiated phenotype (15,16). However, TGF-81 inhibits adipogenic and myogenic differentiation (17,18). In NHEK cells, TGF$l induces reversible growth arrest (19,20). Many mesenchymal and epithelial cells which are affected by TGF-,Bl exhibit elevated expression of several extracellular matrix components and protease inhibitors (14,(21)(22)(23)(24)(25). The cellular responses to TGF-/31 and TGF-/?2 are mediated by specific cell-surface receptors (2,26). However, little is known about the signal transduction mechanism through which TGF-fi acts.
Our laboratory has been interested in the regulatory factors that modulate proliferation and differentiation of NHEK cells (27,28). In studies designed to examine the effect of TGF-&l on these processes we found that TGF-@I enhanced the level of transglutaminase activity in these cells.  (27,(32)(33)(34). When logarithmic cultures of NHEK cells were treated with 100 pM TGF-/31 for 2 days, a 7-fold stimulation in total cellassociated transglutaminase activity was observed (Table I). This increase in activity was due solely to a 13-fold increase in the transglutaminase activity associated with the soluble fraction (Type II transglutaminase).
In order to confirm the identity of the TGF-bl-stimulated transglutaminase as Type II transglutaminase, total cellular protein from NHEK cells of confluent, squamous differentiated cultures, logarithmic cultures treated for 2 days with 100 pM TGF$l, and undifferentiated logarithmic cultures were analyzed by immunoblot analysis using monoclonal antibodies B.Cl and Cub-7401 (32, 46). These two antibodies react specifically with either Type I or Type II transglutaminase which have been shown to migrate upon SDS-polyacrylamide gel electrophoresis as monomers with molecular weights of 90,000 and 82,000, respectively (32,30). Relatively little of either Type I or II transglutaminase could be identified in undifferentiated NHEK cells (Fig. 1). The monoclonal antibody Cub-7401, which reacts specifically with Type II transglutaminase, stained a 82-kDa protein from TGF-/31-treated cells, whereas no staining was observed with the monoclonal antibody B.Cl which reacts with Type I transglutaminase.
The antibody B.Cl stained specifically a 90-kDa protein from squamous differentiated cells, whereas no staining was observed with the antibody Cub-7401. These results confirm that TGF-@l-treated cells express mostly transglutaminase Type II. In contrast, squamous differentiated cells contain mostly transglutaminase Type I in agreement with previous findings (28, 32).
The lack of induction of Type I transglutaminase by TGF-/31 suggests that TGF-@1 does not induce squamous differentiation in NHEK cells. Additionally, TGF-@l did not induce cholesterol sulfate (Table I)  ,&-treated cells (data not shown). The increase in total and Type II transglutaminase activity was first measurable 8 h after the addition of TGF$l and reached a plateau after 3 days of treatment (Fig. 2B). Exposure of undifferentiated NHEK cells to 100 pM TGF-bl for 3 days stimulated Type II transglutaminase activity about l&fold; little effect was observed on the level of the transglutaminase activity associated with the particulate fraction (Type I transglutaminase). No significant changes in the level of transglutaminase activity were observed in untreated cells over this incubation period ( Fig. 2A). Moreover, no increase in transglutaminase Type II was observed when confluent cultures of differentiated NHEK cells were exposed to TGF-fil (Fig. 2C). These results show that undifferentiated and squamous differentiated NHEK cells respond differentially to TGF-Pl. No transglutaminase activity was detectable in the medium from cultures of either untreated or TGF-@l-treated NHEK cells (not shown). The increase in transglutaminase Type II activity was obtained in KGM as well as KDM medium and was observed consistently in epidermal cells isolated from human neonatal foreskin or human adult breast skin of different donors (not shown).
The increase in transglutaminase Type II activity was dependent on the TGF-Bl concentration (Fig. 3). Half-maximum stimulation occurred at a concentration of approximately 15 pM TGF-@l. The stimulation was optimal between 30 and 100 pM TGF-bl.
TGF-/32 was just as effective as TGFpl in either inhibiting proliferation of NHEK cells or inducing Type II transglutaminase activity (Fig. 4). The half-maximum concentration of TGF-/3 to inhibit proliferation was 1.5 PM, about lo-fold lower than for the induction of transglutaminase Type II. The dose-response was further examined via immu- NHEK cells were plated in 60-mm dishes at 5.0 x lo4 cells/dish and 2 days later incubated in the presence or absence of TGF-Pl (100 PM). At different time intervals after the addition of TGF-01, transglutaminase activity was determined. A and B, undifferentiated cells; C, squamous differentiated NHEK cells. Open symbols, untreated cells; closed symbols, TGF-@l-treated NHEK cells.
Squares, total transglutaminase activity associated with cells; triangles, transglutaminase activity associated with the particulate fraction; circles, transglutaminase activity associated with the soluble fraction. All transglutaminase activities were calculated as per mg of total cellular protein.
noblot analysis (Fig. 5). NHEK cells growing in the early exponential phase were treated with various concentrations of TGF-b1 for 3 days, then solubilized in SDS sample buffer and total cellular proteins examined by immunoblot analysis using the transglutaminase Type II-specific monoclonal antibody Cub-7401. The staining of a 80-kDa protein band, representing the Type II transglutaminase, was enhanced NHEK cells growing in the exponential phase were treated with the indicated concentration of TGF-01. After 24 h, total RNA was prepared as described under "Experimental Procedures." Equivalent amounts of RNA (10 pg) were probed with "'P-labeled pTG3400 following agarose gel electrophoresis and transfer to Nytran. After stripping, the blot was reprobed with 32P-labeled pGAD-28. GPDH, glyceraldehyde-3-phosphate dehydrogenase.  Type  II mRNA  accumulation in NHEK cells following the addition of TGF-81. Cells growing in the exponential phase were treated with TGFbl (100 PM) and at the indicated time intervals total RNA was prepared. Equivalent quantities of RNA (30 fig) were electrophoresed in agarose/formaldehyde gels, blotted to Nytran, and hybridized with "'P-labeled pTG3400. After stripping, the same blot was reprobed with "'P-labeled pGAD-28. CPDH, glyceraldehyde-3-phosphate dehydrogenase. NHEK cells were treated with 2.5 rg/ ml cyclohexrmrde, wrth TGF-01 (100 PM), or with cycloheximide and TGF-01 for 8 h. Total RNA was isolated and analyzed by the Northern blot procedure using "P-labeled pTG3400 and pGAD-28. Cycloheximide at 2.5 pg/ml inhibited protein synthesis by 97% and RNA synthesis by 48%. Lower concentrations of cycloheximide inhibited protein synthesis to a much smaller extent. GPDH, glyceraldehyde-3-phosphate dehydrogenase.
(2.5 pg/ml) abrogated the increase in Type II transglutaminase mRNA by TGF-@l (Fig. 9) indicating that the increase in transglutaminase Type II mRNA is dependent on protein synthesis. In contrast, glyceraldehyde-3-phosphate dehydrogenase mRNA expression was unaffected by either TGF-01 or cycloheximide.
We compared the action of TGF-@l on transglutaminase activity in normal epidermal keratinocytes with its effect on three human squamous carcinoma cell lines (XC-13, SCC-15, and SQCC/Yl) and two cell lines, NHEK-SV40-T8 and -Tll, which express the SV40 large T-antigen. The carcinoma-derived cell lines fail to undergo squamous differentiation in culture, whereas the SV40 T-antigen-transfected cells have an extended life span and are able to undergo squamous differentiation (27). NHEK-SV40-T8 and -Tll cells did not contain any detectable levels of transglutaminase Type II activity. Treatment of these cells with TGF-fil induced Type II activity more than 150-fold (Table II). The levels of Type II transglutaminase in TGF-/31-treated NHEK-SV40-T8 and -Tll were comparable to that of TGF-pl-treated NHEK cells. The larger apparent increase in the SV40 T-antigen-transformed cells was due to the presence of lower levels of this transglutaminase in untreated transformed cells. The carcinoma-derived cell lines SCC-13, SCC-15, and SCC/Yl were much less responsive to TGF-/I1 (Table II) than the NHEK or SV40-T-transformed cells. DISCUSSION We have been interested in those hormonal factors that regulate the growth and differentiation of normal human epidermal keratinocytes. TGF-fl is one factor that has important effects on the proliferation and differentiation of many types of cells including NHEK cells (2,19,20). NHEK cells in culture exhibit a high proliferative capacity (high colonyforming efficiency) during the exponential growth phase. At confluence, these cells undergo irreversible growth arrest (terminal cell division) and start to express a squamous differentiated phenotype (27). This induction of differentiation is characterized by an increase in the expression of several biochemical markers that include cholesterol sulfate, transglutaminase Type I, and specific keratins (27,32,50). During the exponential phase of cell growth, NHEK cells express low levels of both Type I (epidermal) transglutaminase, which is associated with the particulate fraction, and Type II (tissue) transglutaminase, which is localized in the cytosol. These two transglutaminases are distinct enzymes exhibiting different immunological properties and represent two different gene products (30)(31)(32)(33)(34)36,37). Treatment of NHEK cells with TGF-@l does not induce irreversible growth arrest or expression of a squamous differentiated phenotype as indicated by the retention of a relatively high colony-forming efficiency, and the low levels of expression of the differentiation markers transglutaminase Type I and cholesterol sulfate. These find- a Cells in the exponential phase were treated for 2 days with 100 PM TGF-/31 and then assayed for transglutaminase activity.
ings are in agreement with previous studies showing that TGF-61 does not induce terminal differentiation in NHEK cells (19,20). However, addition of TGF$l or TGF-P2 results in a lo-to 15-fold stimulation of transglutaminase Type II activity. This enhancement in transglutaminase activity appears to be related to increased synthesis of the transglutaminase Type 11 enzyme and to increased levels of corresponding mRNA. The increase in mRNA levels is cycloheximide sensitive indicating that protein synthesis is required for this action.
TGF-Bl has been shown to stimulate the synthesis of a number of gene products, such as fibronectin, collagen, and TGFj31 itself (2,21,22,24,51,52), in several cell systems. In NHEK ceils, TGF-/?l also induces the synthesis of fibronectin and to a lesser degree the synthesis of collagen Type IV.* TGF$l can regulate gene expression at the level of transcription as well as at the level of the stability of the mRNA (22)(23)(24)51). Like the increase in Type II transglutaminase mRNA, the stimulation of both fibronectin and collagen (~2(1) mRNA by TGF-@l have been shown to depend on protein synthesis. Binding sites for the transcriptional factor nuclear factor 1 in the promoter regions of the collagen c~2(1) and fibronectin gene have been implicated in the regulation of these genes by TGF-/31 (24, 51). However, other elements appear to be involved in the TGF-Pl inducibility as well (24). Although the full-length coding sequence of human and guinea pig transglutaminase Type II mRNA's have been established, no data are as yet available about the promoter sequence (35, 36). Efforts are under way to sequence the promoter region of human Type II transglutaminase and to determine its regulatory elements. It will be interesting to see whether nuclear factor l-binding sites are involved in the regulation of this enzyme by TGF-p.
TGF-81 did not induce transglutaminase Type II activity in certain epidermal carcinoma-derived cell lines. The nonresponsiveness of these cells to TGF$l appears not to be due to the absence of cell surface receptors since high affinity receptors have been demonstrated in these cells (52, 53)3 but could be related to an alteration in the TGF-@l-induced signal transduction at a level other than the TGF-P/receptor interaction or due to the activation of a signal that antagonizes the TGF-@ response.
Various other factors have been shown to induce Type II transglutaminase in different cell systems. Sodium butyrate induces Type II transglutaminase activity in human fibroblasts WI-38 (54) and in PC12 pheochromocytoma cells (55). Retinoic acid has been shown to stimulate Type II transglutaminase in human myeloblastic leukemia HL60 cells and in mouse peritoneal macrophages (30,31). In the latter case, the induction is further stimulated by CAMP and inhibited by pertussis toxin. Type II transglutaminase activity was not further modulated by sodium butyrate, CAMP, or pertussis toxin in either TGF-p-treated or untreated NHEK cells, indicating differences in the mechanism of transglutaminase induction in the various cell systems. In human epidermal keratinocytes grown on 3T3 feeders and in the presence of serum, as well as in mouse epidermal cells, retinoic acid has been shown to increase Type II transglutaminase activity (33, 56); however, in serum-free medium and in the absence of feeder cells, retinoic acid was unable to induce transglutaminase Type II activity under a wide variety of conditions (27) Type II reported in epidermal cells by retinoic acid is related to increased synthesis of TGF-/3, or to the activation of TGF-/3 produced by the epidermal or 3T3 feeder cells or to the TGF-@ present in serum. It has been recently reported that retinoic acid stimulates the synthesis of TGF-/32 in mouse epidermal keratinocytes (57). This finding with the results presented in this study appear to suggest that the induction of transglutaminase Type II by retinoic acid is mediated by TGF-P2.
Very little is known about the function of transglutaminase Type II. It has been suggested that this enzyme is involved in the regulation of cellular proliferation, cellular morphology, and programmed cell death (apoptosis) (55,58,59). Transglutaminase Type II may play an important role in stabilizing the cytoskeletal network of developing myotubes (60). Other studies have implicated Type II transglutaminase in the covalent cross-linking of microtubules to other cellular components (61) and in the cross-linking of neurotilaments in the brain (62). In hepatocytes, it has been shown that transglutaminase Type II is able to cross-link cytokeratins (63). Our preliminary studies3 indicate that also in NHEK cells, cytokeratins can function as substrates for transglutaminase Type II. Cross-linking of cytokeratins may influence various aspects of cellular behavior such as cell growth and morphology. Recently, a very different activity of transglutaminase Type II has been discovered, Lee et al. (64) demonstrated that this protein contains GTPase activity. Whether this GTPase activity has any function in the mediation of certain actions of TGF-fil has to be established.