Protein C Inhibitor is Expressed in Keratinocytes of Human Skin

Protein C inhibitor is a member of the serpin family that inhibits a variety of serine proteases. Protein C inhibitor is present in numerous body fluids and is produced in the liver and by various epithelial cells. To determine if this epithelial serpin is present in skin, immunohistochemical studies were performed that showed strong staining for protein C inhibitor antigen in the epidermis. Protein C inhibitor mRNA was detected in the keratinocyte cell line HaCaT and the epidermoid carcinoma cell line A431 using reverse transcription–polymerase chain reaction suggesting that also in normal skin protein C inhibitor is derived from keratinocytes. Conditioned media from these cell lines were analyzed on immunoblots, which revealed a protein C inhibitor-antigen band that comigrated with protein C inhibitor derived from the hepatoma cell line HepG2. Using an enzyme-linked immunosorbent assay specific for total protein C inhibitor antigen the accumulation of protein C inhibitor in the cell culture supernatants of HaCaT keratinocytes was found to be 0.3 ng per h per 1 million cells. This is similar to the amount of plasminogen activator inhibitor-1 produced by these cells, which also produce tissue plasminogen activator and urokinase. Fluorescence-activated cell sorter analysis revealed similar expression of intracellular protein C inhibitor antigen in proliferating and confluent HaCaT cells. These findings demonstrate that protein C inhibitor antigen is present in the normal epidermis and that protein C inhibitor is constitutively expressed by keratinocytes in culture. Therefore, protein C inhibitor may provide protease inhibitory activity not only to internal, but also to the external surface of the body. Additionally, protein C inhibitor could contribute to the regulation of retinoid supply in the epidermis, as we have shown recently that retinoic acid binds specifically to protein C inhibitor. Protein C inhibitor is a member of the serpin family that inhibits a variety of serine proteases. Protein C inhibitor is present in numerous body fluids and is produced in the liver and by various epithelial cells. To determine if this epithelial serpin is present in skin, immunohistochemical studies were performed that showed strong staining for protein C inhibitor antigen in the epidermis. Protein C inhibitor mRNA was detected in the keratinocyte cell line HaCaT and the epidermoid carcinoma cell line A431 using reverse transcription–polymerase chain reaction suggesting that also in normal skin protein C inhibitor is derived from keratinocytes. Conditioned media from these cell lines were analyzed on immunoblots, which revealed a protein C inhibitor-antigen band that comigrated with protein C inhibitor derived from the hepatoma cell line HepG2. Using an enzyme-linked immunosorbent assay specific for total protein C inhibitor antigen the accumulation of protein C inhibitor in the cell culture supernatants of HaCaT keratinocytes was found to be 0.3 ng per h per 1 million cells. This is similar to the amount of plasminogen activator inhibitor-1 produced by these cells, which also produce tissue plasminogen activator and urokinase. Fluorescence-activated cell sorter analysis revealed similar expression of intracellular protein C inhibitor antigen in proliferating and confluent HaCaT cells. These findings demonstrate that protein C inhibitor antigen is present in the normal epidermis and that protein C inhibitor is constitutively expressed by keratinocytes in culture. Therefore, protein C inhibitor may provide protease inhibitory activity not only to internal, but also to the external surface of the body. Additionally, protein C inhibitor could contribute to the regulation of retinoid supply in the epidermis, as we have shown recently that retinoic acid binds specifically to protein C inhibitor. protein C inhibitor plasminogen activator inhibitor-1 tissue plasminogen activator urokinase type plasminogen activator Protein C inhibitor (PCI) is a sngle chain glycoprotein with a molecular weight of 57 kDa. PCI is a relatively unspecific, heparin-binding serine protease inhibitor (serpin), originally described in plasma as an inhibitor of the anticoagulant protease activated protein C (Marlar and Griffin, 1980Marlar R.A. Griffin J.H. Deficiency of protein C inhibitor in combined factor V/VIII deficiency disease.J Clin Invest. 1980; 66: 1186-1189Crossref Scopus (134) Google Scholar). Serine proteases inhibited by PCI include thrombin, factor Xa, factor XIa, tissue and plasma kallikreins, acrosin, tissue plasminogen activator (tPA), and urokinase (for review seeSuzuki et al., 1989Suzuki K. Deyashiki Y. Nishioka J. Toma K. Protein C. inhibitor. Structure and function.Thromb Haemostas. 1989; 61: 337-342Google Scholar;Geiger et al., 1996Geiger M. Zechmeister-Machhart M. Uhrin P. et al.Protein C inhibitor (PCI).Immunopharmacology. 1996; 32: 53-56Crossref PubMed Scopus (30) Google Scholar, Geiger et al., 1997Geiger M. Krebs M. Jerabek I. Binder B.R. Protein C inhibitor (PCI) and heparin cofactor II (HCII): possible alternative roles of these heparin-binding serpins outside the hemostatic system.Immunopharmacology. 1997; 36: 279-284Crossref PubMed Scopus (9) Google Scholar). PCI antigen is present in various body fluids (Laurell et al., 1992Laurell M. Christensson A. Abrahamsson P. Stenflo J. Lilja H. Protein C inhibitor in human body fluids.J Clin Invest. 1992; 89: 1094-1101Crossref PubMed Scopus (135) Google Scholar), and the tubular epithelium of human kidney (Radtke et al., 1994Radtke K.P. Fernández J.A. Greengard J.S. et al.Protein C inhibitor is expressed in tubular cells of human kidney.J Clin Invest. 1994; 94: 2117-2124Crossref PubMed Scopus (44) Google Scholar) and epithelial cells in the male reproductive tract (Laurell et al., 1992Laurell M. Christensson A. Abrahamsson P. Stenflo J. Lilja H. Protein C inhibitor in human body fluids.J Clin Invest. 1992; 89: 1094-1101Crossref PubMed Scopus (135) Google Scholar) have been shown to produce PCI. Plasma PCI is thought to be synthesized in the liver, as the human hepatoma-derived cell line HepG2 secretes significant amounts of PCI in vitro (Morito et al., 1985Morito F. Saito H. Suzuki K. Hashimoto S. Synthesis and secretion of protein C inhibitor by the human hepatoma-derived cell line, Hep G2.Biochem Biophys Acta. 1985; 884: 209-215Crossref Scopus (20) Google Scholar;Fair and Marlar, 1986Fair D.S. Marlar R.A. Biosynthesis and secretion of factor VII, protein C, protein S, and the protein C inhibitor from a human hepatoma cell line.Blood. 1986; 67: 64-70Crossref PubMed Google Scholar) and as PCI plasma levels are decreased in patients with liver diseases (Francis and Thomas, 1984Francis R.B. Thomas W. Behaviour of protein C inhibitor in intravascular coagulation and liver disease.Thromb Haemostas. 1984; 52: 71-74PubMed Google Scholar). Plasma PCI is elevated in survivors of myocardial infarction (Carroll et al., 1997Carroll V.A. Griffith M.R. Geiger M. Merlo C. Furlan M. Lämmle B. Binder B.R. Plasma protein C inhibitor is elevated in survivors of myocardial infarction.Arteriosclerosis Thromb Vasc Biol. 1997; 17: 114-118Crossref PubMed Scopus (31) Google Scholar). Activated protein C-PCI complexes have been shown in plasma samples from patients with disseminated intravascular coagulation (Espana et al., 1990Espana F. Vicente V. Tabernero D. Scharrer I. Griffin J.H. Determination of plasma protein C inhibitor and of two activated protein C-inhibitor complexes in normals and in patients with intravascular coagulation and thrombotic disease.Thromb Res. 1990; 59: 593-608Abstract Full Text PDF PubMed Scopus (73) Google Scholar), and urokinase-PCI complexes in plasma samples from patients receiving urokinase therapy (Geiger et al., 1989Geiger M. Huber K. Wojta J. Stingl L. Espana F. Griffin J.H. Binder B.R. Complex formation between urokinase and plasma protein C inhibitor in vitro and in vivo.Blood. 1989; 74: 722-728Crossref PubMed Google Scholar). These findings suggest that PCI could play a part in the regulation of hemostasis. Most coagulation and fibrinolytic enzymes, however, are inactivated more efficiently by other serpins. Therefore, the precise physiologic role of PCI is still unclear. Our group has previously shown that tissue kallikrein is efficiently inhibited by PCI (Ecke et al., 1992Ecke S. Geiger M. Resch I. Jerabeck I. Stingl L. Maier M. Binder B.R. Inhibition of tissue kallikrein by protein C inhibitor.J Biol Chem. 1992; 267: 7048-7052Google Scholar), suggesting that PCI may be a physiologically important endogenous inhibitor of this serine protease. Many components of the kallikrein-kinin system including tissue kallikrein are expressed in human skin, where they are potentially involved in the regulation of proliferation and in the mediation of inflammatory processes (Poblete et al., 1991Poblete M.T. Reynolds N.J. Figueroa C.D. Burton J.L. Muller-Esterl W. Bhoola K.D. Tissue kallikrein and kininogen in human sweat glands and psoriatic skin.Br J Dermatol. 1991; 124: 236-241Crossref PubMed Scopus (33) Google Scholar;Schremmer-Danninger et al., 1995Schremmer-Danninger E. Heinz-Erian P. Töpfer-Petersen E. Roscher A.A. Autoradiographic localization and characterization of bradykinin receptors in human skin.Eur J Pharmacol. 1995; 283: 207-216Crossref PubMed Scopus (22) Google Scholar). Google Scholar,Google Scholar So far no epidermal inhibitor for tissue kallikrein has been described. Furthermore, we have recently shown that PCI also belongs to the group of hormone-binding serpins which includes corticosteroid-binding globulin (Hammond et al., 1987Hammond G.L. Smith C.L. Goping I.S. et al.Primary structure of human corticosteroid binding globulin, deduced from hepatic and pulmonary cDNAs, exhibits homology with serine protease inhibitors.Proc Natl Acad Sci USA. 1987; 84: 5153-5157Crossref PubMed Scopus (228) Google Scholar) and thyroxine-binding globulin (Flink et al., 1986Flink I.L. Bailey T.J. Gustafson T.A. Markham B.E. Morkin E. Complete amino acid sequence of human thyroxine-binding globulin deduced from cloned DNA. Close homology to the serine antiproteases.Proc Natl Acad Sci USA. 1986; 83: 7708-7712Crossref PubMed Scopus (159) Google Scholar): We were able to show that 3H-all-trans-retinoic acid bound in a concentration-dependent and specific manner to PCI but not to other inhibitory serpins and binding could be competed by unlabeled all-trans-retinoic acid, 9-cis-retinoic acid and retinol. Binding of estradiol, progesterone, testosterone, cortisol, and aldosterone to PCI was not detected. We have also preliminary data that PCI might be involved in the delivery of retinoic acid to target cells.Google Scholar The facts that (i) tissue kallikrein is expressed in the epidermis, that (ii) the epidermis is an important target organ for retinoids (Fisher and Voorhees, 1996Fisher G.J. Voorhees J.J. Molecular mechanisms of retinoid actions in skin.FASEB J. 1996; 10: 1002-1013Crossref PubMed Scopus (327) Google Scholar), and that (iii) PCI is expressed in various kinds of epithelial cells lining internal surfaces, prompted us to investigate the possible presence of PCI in human skin and human keratinocyte cell lines. We used the epidermoid cell line A431 and spontaneously immortalized human keratinocytes (HaCaT) in these experiments. HaCaT cells maintain full epidermal differentiation capacity and have been widely used as representative of normal keratinocytes as they share many features with primary cells (Boukamp et al., 1988Boukamp P. Dzarliewa-Perusevska R.T. Breitkreutz D. Hornung J. Markham A. Fusenig N.E. Normal keratinization in a spontaneously immortalized aneuploid human keratinocyte cell line.J Cell Biol. 1988; 106: 761-771Crossref PubMed Scopus (3258) Google Scholar;Reinartz et al., 1994Reinartz J. Link J. Todd R.F. Kramer M.D. The receptor for urokinase-type plasminogen activator of a human keratinocyte line (HaCaT).Exp Cell Res. 1994; 214: 486-498Crossref PubMed Scopus (22) Google Scholar, Reinartz et al., 1996Reinartz J. Schaefer B. Bechtel M. Kramer M.D. Plasminogen activator inhibitor type-2 (PAI-2) in human keratinocytes regulates pericellular urokinase-type plasminogen activator.Exp Cell Res. 1996; 223: 91-101Crossref PubMed Scopus (20) Google Scholar;Bechtel et al., 1996Bechtel M.J. Reinartz J. Rox J.M. Inndorf S. Schaefer B.M. Kramer M.D. Upregulation of cell surface associated plasminogen activation in cultured keratinocytes by interleukin-1β and tumor necrosis factor-α.Exp Cell Res. 1996; 223: 395-404Crossref PubMed Scopus (47) Google Scholar;Breitkreutz et al., 1998Breitkreutz D. Schoop V.M. Mirancea N. Baur M. Stark H.J. Fusenig N.E. Epidermal differentiation and basement membrane formation by HaCaT cells in surface transplants.Eur J Cell Biol. 1998; 75: 273-286Crossref PubMed Scopus (114) Google Scholar). Normal human trunk skin was obtained from patients who underwent plastic surgery. The samples were embedded in Tissue-Tek OCT Compound (Sakura, Tokyo, Japan), snap frozen in liquid nitrogen, and sectioned at 8 μm thickness using a cryostat. The sections were mounted on to Histobond microscope slides (Marienfeld, Germany) and air-dried for 1 h at room temperature. Sections were then fixed for 15 min in 0.5% paraformaldehyde in phosphate-buffered saline (PBS; 2 mmol KH2PO4 per liter, 8 mmol Na2HPO4 per liter, 140 mmol NaCl per liter, pH 7.4) and permeabilized by sequential treatment with 0.2% and 1% Triton X-100 in PBS (1 × 10 min each). Slides were then rinsed with PBS and incubated with 1% bovine serum albumin (BSA; Behring, Marburg, Germany) in PBS for 30 min. Incubations with 10 μg monoclonal anti-PCI IgG (4PCI) per ml (Ecke et al., 1992Ecke S. Geiger M. Resch I. Jerabeck I. Stingl L. Maier M. Binder B.R. Inhibition of tissue kallikrein by protein C inhibitor.J Biol Chem. 1992; 267: 7048-7052Google Scholar) or nonimmune mouse IgG diluted in primary antibody diluting buffer (Biomeda, Foster City, CA) were carried out in a humid chamber at 4°C over night. The slides were then rinsed with PBS and washed with automation buffer (Biomeda) (2 × 5 min). Thereafter the slides were treated sequentially with biotinylated sheep antimouse IgG (Amersham, Little Chalfont, Buckinghamshire, U.K.) diluted 1:100 in PBS and streptavidin–peroxidase conjugate (Amersham) diluted 1:300 in PBS. After each incubation step the slides were washed as described above. Bound peroxidase was visualized using the Liquid DAB-Plus Substrate Kit (Zymed, San Francisco, CA). Thereafter slides were washed with distilled water, tissue sections were counterstained with hematoxylin, rinsed with water, and mounted in Aquatex (Merck, Darmstadt, Germany). Controls included the substitution of the first antibody with monoclonal anti-heparin cofactor II IgG, monoclonal anti-plasminogen activator inhibitor 1 IgG (5PAI) (Technoclone, Vienna, Austria), and monoclonal anti-complement factor D IgG (D10/4) (Connex, Martinsried, Germany). Sets of slides were processed together in duplicates. Human hepatoma cells (HepG2) and the human epidermoid cell line A431 were obtained from American Type Culture Collection (ATCC; Rockville, MD). The spontaneously transformed human keratinocyte cell line HaCaT was kindly provided by Dr. N.E. Fusenig [Deutsches Krebsforschungszentrum (DKFZ), Heidelberg, Germany] (Boukamp et al., 1988Boukamp P. Dzarliewa-Perusevska R.T. Breitkreutz D. Hornung J. Markham A. Fusenig N.E. Normal keratinization in a spontaneously immortalized aneuploid human keratinocyte cell line.J Cell Biol. 1988; 106: 761-771Crossref PubMed Scopus (3258) Google Scholar). The cells were grown under culture conditions (37°C, 5% CO2, 95% air) in Dulbecco’s modified Eagle’s media (Sigma, St. Louis, MO) containing 10% supplemented calf serum (Hyclone, Logan, UT), 1 mmol glutamine per liter, and antibiotics (100 μg streptomycin per ml, 100 IU penicillin per ml, 250 ng amphotericin B per ml) (all: JHR Biosciences, Lenaxa, KS). The cells received fresh medium every 3 d and were subcultured using a split ratio of 1:5 as soon as they reached confluence. Cells were either grown in 75 cm2 tissue culture flasks (Iwaki, Japan) in 10 ml medium or in six-well plates (Iwaki) in 2 ml medium per well. After reaching confluence the monolayers were washed twice with Hanks’ balanced salt solution (Sigma) and incubated in serum free medium supplemented with 0.5% BSA (Sigma). After appropriate incubation times conditioned media were collected, centrifuged to remove cell debris, and stored at –70°C until analyzed. Cells were grown to confluence in 75 cm2 tissue culture flasks, washed twice with Hanks’ balanced salt solution (Sigma), and total cellular RNA was isolated using Trizol reagent (Gibco, Grand Island, NY) according to the manufacturer’s instructions. Total RNA was subjected to reverse transcription and subsequent polymerase chain amplification to show the presence of PCI mRNA in analyzed cells. For reverse transcription 1 μg of total RNA from HaCaT and A431 cells and 0.2 μg or 0.1 μg of total RNA from HepG2 cells were used. The RNA was preheated at 65°C for 15 min to remove possible secondary structures and then cooled down to 4°C. Reverse transcription reaction was made using a first-strand cDNA synthesis kit for reverse transcription–polymerase chain reaction (Boehringer Mannheim, Mannheim, Germany) at 42°C for 60 min. Following components were included: an anti-sense human PCI sequence specific primer 5′-CCT GTT GAA CAC TAG CCT CTG AGA G-3′ at final concentration of 0.2 μmol per liter, 50 mmol Tris–HCl buffer per liter, pH 9.0, 1.5 mmol MgCl2 per liter, 0.2 mmol of each deoxyribonucleoside triphosphate per liter, 50 U of RNase inhibitor, and 20 U of avian myeloblastosis virus–reverse transcriptase. Subsequently, 10 μl of the reaction mixture were taken and after addition of a sense human PCI sequence specific primer 5′-GGA TCA GTA TCA CTA CCT CCT GGA C-3′, polymerase, and other components, the reaction mixture contained 50 mmol Tris–HCl buffer per liter, pH 9.0, 1.5 mmol MgCl2 per liter, 0.2 mmol of each deoxyribonucleoside triphosphate per liter, 0.2 μmol of each primer per liter, and 1.0 U of DNA polymerase (Dynazyme; Finnzymes Oy, Finland) in a total volume of 25 μl. The mixture was predenaturated for 2 min at 94°C and then subjected to 35 step cycles of 94°C (35 s), 58°C (30 s), 72°C (20 s). Amplified DNA fragments exhibited the expected size (442 base pairs) as judged from 1% agarose gel electrophoresis. Their identities were confirmed by sequencing. Prior to electrophoresis samples from keratinocytes and HepG2 cells were concentrated and partially purified by affinity chromatography on heparin-Sepharose CL-6B (Pharmacia, Uppsala, Sweden). Ten milliliters of conditioned media from HaCaT, A431, or HepG2 cells, respectively, grown in 75 cm2 tissue culture flasks were incubated with 0.5 ml heparin-Sepharose equilibrated in 50 mM Tris–HCl, 100 mM NaCl, pH 7.5 (loading buffer) for 1 h at room temperature under constant shaking. The Sepharose was filled in a small column and washed with loading buffer. Thereafter bound protein was eluted with 1 ml 1 M NaCl in loading buffer. The samples were concentrated to 100 μl using a 10 kDa molecular weight cut-off Ultra Spin Microfilter (Roth, Karlsruhe, Germany), electrophoresed on a 10% sodium dodecyl sulfate–polyacrylamide gel, and electrophoretically transferred to a Hybond-P PVDF membrane (Amersham) using standard protocols. The membrane was blocked with 1.5% skim milk powder in 0.02 mol Tris–HCl per liter, 0.5 mol NaCl per liter, pH 7.5 (blocking buffer) over night at 4°C followed by a 2 h incubation with 10 μg per ml of specific rabbit anti-PCI IgG diluted in blocking buffer. The antibody used was prepared following standard protocols by immunizing a New Zealand white rabbit with highly purified urinary PCI (Ecke et al., 1992Ecke S. Geiger M. Resch I. Jerabeck I. Stingl L. Maier M. Binder B.R. Inhibition of tissue kallikrein by protein C inhibitor.J Biol Chem. 1992; 267: 7048-7052Google Scholar). The blot was washed with 0.02 mol Tris–HCl per liter, 0.5 mol NaCl per liter, pH 7.5, containing 0.05% Tween 20 (3 × 5 min) and twice with the same buffer without Tween 20 followed by a 1 h incubation with peroxidase linked anti-rabbit IgG (Amersham, Little Chalfont, Buckinghamshire, UK) diluted 1:1000 in blocking buffer. The membrane was washed as described and bound peroxidase was detected using chemiluminescence (ECL; Amersham). Prestained sodium dodecyl sulfate-standards (BioRad, Richmond, CA) were used for estimation of Mrs. PCI antigen in the conditioned media was quantitated by a specific enzyme-linked immunosorbent assay (ELISA) as described previously (Priglinger et al., 1994Priglinger U. Geiger M. Bielek E. Vanyek E. Binder B.R. Binding of urinary protein C inhibitor to cultured human epithelial kidney tumor cells (TCL-598).J Biol Chem. 1994; 269: 14705-14710Google Scholar), using acid-treated monoclonal anti-PCI IgG (4PCI) as catching antibody and peroxidase-labeled immunopurified rabbit anti-PCI IgG as detecting antibody. PAI-1 in conditioned media was quantitated by a specific ELISA based on monoclonal antibodies, which allows determination of active and latent PAI-1 as well as PAI-1 in complex with tPA (Technoclone, Vienna, Austria). tPA antigen and uPA antigen were measured with specific ELISA according to the instruction of the manufacturer (Technoclone, Vienna, Austria). HaCaT cells were grown in six-well plates in Dulbecco’s modified Eagle’s media supplemented with 10% supplemented calf serum. Forty-eight hours after reaching confluence the cells were harvested with trypsin-ethylenediamine tetraacetic acid, washed with PBS, pH 7.4, and centrifuged at 450 × g for 5 min. Alternatively, HaCaT cells were seeded at low concentrations to obtain scattered single cells, incubated for 48 h in Dulbecco’s minimal essential media containing 10% supplemented calf serum, and trypsinized as described above. One-half of the pelleted cells was fixed with 0.5% paraformaldehyde in PBS, permeabilized with 0.025% saponin (Sigma) in PBS. The other half was used unpermeabilized. Unspecific binding sites were blocked by incubating the cells with 1% BSA in PBS. Thereafter cells were washed with PBS, pH 7.4, containing 1% BSA and 0.025% saponin and centrifuged at 450 × g for 5 min. Subsequently cells were incubated with rabbit anti-PCI IgG at a concentration of 20 μg per ml in PBS containing 1% BSA and 0.025% saponin for 2 h at room temperature. After washing as described above the cells were incubated with fluorescein-labeled goat anti-rabbit IgG (Vector Laboratories, Burlingame, CA) diluted 1:50 in PBS containing 1% BSA and 0.025% saponin. Control cells were only incubated with the fluorescein-labeled antibody. Then the cells were washed once as described above and once with PBS, pH 7.4. Analysis was performed by a FACS Scan flow cytometer using CellQuest 3.1 software (Becton Dickinson, San Jose, CA). An immunohistochemical analysis with monoclonal anti-PCI IgG was performed to determine whether or not PCI is present in normal adult skin. Pronounced uniform staining for PCI antigen was detected in the keratinocytes throughout the epidermis (Figure 1a). In control sections stained with nonimmune mouse IgG (Figure 1b) and monoclonal antibodies to heparin cofactor II, plasminogen activator inhibitor 1 (PAI-1), or complement factor D no signal was observed (data not shown). In situ hybridization studies using specific sense and anti-sense digoxigenin labeled RNA-probes for PCI were performed. No PCI mRNA was apparent in human skin (data not shown), however, presumably because the method was not sensitive enough to detect the low levels of PCI mRNA present in normal keratinocytes (see below). In order to confirm the results obtained with the immunohistochemical studies of normal skin, PCI synthesis by keratinocytes was analyzed. PCI mRNA expression in spontaneously immortalized human keratinocytes (HaCaT) and epidermoid carcinoma cells (A431) was analyzed using reverse transcription–polymerase chain reaction. RNA prepared from human hepatoma cells (HepG2) was used as a positive control. Figure 2 demonstrates that PCI mRNA was present in all three cell lines. A strong signal, that was dependent on the amount of RNA used, was obtained with HepG2 cells. In RNA prepared from A431 cells a pronounced band for PCI mRNA was detected, whereas the signal for PCI mRNA in HaCaT cells was rather weak. To determine whether the PCI mRNA produced in these cells is translated and the protein is secreted into the supernatant, conditioned media were analyzed in immunoblots and ELISA. Figure 3 demonstrates that PCI antigen accumulated over time in the cell culture supernatants from HaCaT, A431, and HepG2 cells. PCI in conditioned media was partially purified and concentrated using heparin-Sepharose and analyzed in immunoblots (Figure 3a), developed with specific polyclonal antibodies. These blots revealed a protein band with a Mr of ≈60 kDa in the conditioned media of HaCaT and A431 cells, which comigrated with PCI derived from HepG2 cells. Specificity of the reactions was validated with control blots incubated only with the second, peroxidase-labeled antibody. No bands were detected in these blots (data not shown). The amount of PCI obtained from conditioned media of A431 and HepG2 cells was much higher than that of HaCaT cells, reflecting the rate of PCI secretion. These data were confirmed by a specific antigen ELISA (Figure 3b), showing that PCI antigen accumulated much faster in the supernatants from A431 and HepG2 as compared with HaCaT cells. The concentrations of PCI antigen in conditioned media of HaCaT, A431, and HepG2 cells were 33 ng per ml, 475 ng per ml, and 650 ng per ml, respectively, after a conditioning time of 24 h. The accumulation of PCI was compared with that of uPA, tPA, and PAI-1 in conditioned media of HaCaT cells. Figure 4, demonstrates that the secretion rates of the serpins (PCI and PAI-1) were similar (approximately 0.3 ng per h per 1 million cells), whereas considerably more uPA than tPA was produced. FACS analysis was used to compare the expression of intracellular PCI antigen in proliferating and confluent HaCaT cells. As can be seen from Figure 5, no difference in PCI antigen between proliferating and nonproliferating cells was observed. The PCI detected in this assay was located mainly intracellularly as only a weak signal for PCI antigen was observed in nonpermeabilized cells. In the present study we have shon that PCI is present in normal human epidermis and that PCI mRNA and antigen are expressed by cultured keratinocytes. The immunohistochemical localization experiments for PCI revealed that PCI was expressed in the basal, proliferating as well as in the more superficial, differentiating keratinocytes of the epidermis. The notion that PCI is constitutively expressed by keratinocytes is supported by the FACS data showing that intracellular PCI is equally present in proliferating and confluent, growth arrested HaCaT. It has been shown that keratinocytes can produce the plasminogen activators uPA and tPA as well as the plasminogen activator inhibitors PAI-1 and PAI-2 in culture and that the pattern of expression of these proteins depends on the state of differentiation (Jensen et al., 1990Jensen P.J. John M. Baird J. Urokinase and tissue type plasminogen activators in human keratinocyte culture.Exp Cell Res. 1990; 187: 162-169Crossref PubMed Scopus (46) Google Scholar, Jensen et al., 1995Jensen P.J. Wu Q. Janowitz P. Ando Y. Schechter N.M. Plasminogen activator inhibitor type 2: an intracellular keratinocyte differentiation product that is incorporated into the cornified envelope.Exp Cell Res. 1995; 217: 65-71Crossref PubMed Scopus (77) Google Scholar;Chen et al., 1993Chen C.S. Lyons-Giordano B. Lazarus G.S. Jensen P.J. Differential expression of plasminogen activators and their inhibitors in an organotypic skin coculture system.J Cell Sci. 1993; 106: 45-53PubMed Google Scholar;Reinartz et al., 1996Reinartz J. Schaefer B. Bechtel M. Kramer M.D. Plasminogen activator inhibitor type-2 (PAI-2) in human keratinocytes regulates pericellular urokinase-type plasminogen activator.Exp Cell Res. 1996; 223: 91-101Crossref PubMed Scopus (20) Google Scholar). HaCaT cells secrete twice as much PAI-2 into the supernatant as compared with PAI-1 (Reinartz et al., 1996Reinartz J. Schaefer B. Bechtel M. Kramer M.D. Plasminogen activator inhibitor type-2 (PAI-2) in human keratinocytes regulates pericellular urokinase-type plasminogen activator.Exp Cell Res. 1996; 223: 91-101Crossref PubMed Scopus (20) Google Scholar). This is in accordance to previous studies showing that in contrast to PAI-1, PAI-2 can be detected in normal epidermis by mRNA, antigen, and activity assays (Lyons-Giordano et al., 1994Lyons-Giordano B. Loskutoff D. Chen C.S. Lazarus G. Keeton M. Jensen P.J. Expression of plasminogen activator inhibitor type 2 in normal and psoriatic epidermis.Histochemistry. 1994; 101: 105-112Crossref PubMed Scopus (52) Google Scholar). These results indicate that PAI-2 (for review seeBelin, 1993Belin D. Biology and facultative secretion of plasminogen activator inhibitor-2.Thromb Haemostas. 1993; 70: 144-147PubMed Google Scholar) is the predominant PAI in keratinocytes in vivo as well as in vitro (Jensen et al., 1995Jensen P.J. Wu Q. Janowitz P. Ando Y. Schechter N.M. Plasminogen activator inhibitor type 2: an intracellular keratinocyte differentiation product that is incorporated into the cornified envelope.Exp Cell Res. 1995; 217: 65-71Crossref PubMed Scopus (77) Google Scholar). The distribution of PCI in the epidermis is similar to that of PAI-2: PAI-2 mRNA and antigen have been detected throughout the epidermis, whereas PAI-1 was localized predominantly to the keratinocytes of the basal layer (Chen et al., 1993Chen C.S. Lyons-Giordano B. Lazarus G.S. Jensen P.J. Differential expression of plasminogen activators and their inhibitors in an organotypic skin coculture system.J Cell Sci. 1993; 106: 45-53PubMed Google Scholar;Lyons-Giordano et al., 1994Lyons-Giordano B. Loskutoff D. Chen C.S. Lazarus G. Keeton M. Jensen P.J. Expression of plasminogen activator inhibitor type 2 in normal and psoriatic epidermis.Histochemistry. 1994; 101: 105-112Crossref PubMed Scopus (52) Google Scholar). Receptor-bound urokinase is thought to play an important part in proliferation and differentiation of keratinocytes (Reinartz et al., 1994Reinartz J. Link J. Todd R.F. Kramer M.D. The receptor for urokinase-type plasminogen activator of a human keratinocyte line (HaCaT).Exp Cell Res. 1994; 214: 486-498Crossref PubMed Scopus (22) Google Scholar), and PAI-2 has been implicated in the regulation of uPA activity in the epidermis (Reinartz et al., 1996Reinartz J. Schaefer B. Bechtel M. Kramer M.D. Plasminogen activator inhibitor type-2 (PAI-2) in human keratinocytes regulates pericellular urokinase-type plasminogen activator.Exp Cell Res. 1996; 223: 91-101Crossref PubMed Scopus (20) Google Scholar). The second order rate constants for the interaction of PCI with urokinase (1–8 × 103 per M per s without heparin and 4 × 103–9 × 104 per M per s with heparin) are much lower than those for the interaction of urokinase with PAI-1 (>107 per M per s) or PAI-2 (2 × 106 per M per s) (Thorsen et al., 1988Thorsen S. Philips M. Selmer J. Lecander I. Astedt B. Kinetics of inhibition of tissue-type and urokinase-type plasminogen activator by plasminogen-activator inhibitor type 1 and type 2.Eur J Biochem. 1988; 175: 33-39Crossref PubMed Scopus (115) Google Scholar;Geiger et al., 1989Geiger M. Huber K. Wojta J. Stingl L. Espana F. Griffin J.H. Binder B.R. Complex formation between urokinase and plasma protein C inhibitor in vitro and in vivo.Blood. 1989; 74: 722-728Crossref PubMed Google Scholar). Therefore, it is unlikely that PCI plays a major part in the regulation of epidermal urokinase activity. It has been postulated that keratinocytes elaborate proinflammatory mediators including bradykinin in response to diverse stimuli and thereby initiate cutaneous inflammation (Barker et al., 1991Barker J.N. Mitra R.S. Griffiths C.E. Dixit C.E. Nickoloff B.J. Keratinocytes as initiators of inflammation.Lancet. 1991; 337: 211-214Abstract PubMed Scopus (607) Google Scholar). The main components of the kallikrein–kinin system (reviewed inScicli and Carretero, 1986Scicli A.G. Carretero O.A. Renal kallikrein-kinin system.Kidney Int. 1986; 29: 120-130Crossref PubMed Scopus (169) Google Scholar), including tissue kallikrein, are present in human skin (Poblete et al., 1991Poblete M.T. Reynolds N.J. Figueroa C.D. Burton J.L. Muller-Esterl W. Bhoola K.D. Tissue kallikrein and kininogen in human sweat glands and psoriatic skin.Br J Dermatol. 1991; 124: 236-241Crossref PubMed Scopus (33) Google Scholar). As PCI is a potent inhibitor of tissue kallikrein (Ecke et al., 1992Ecke S. Geiger M. Resch I. Jerabeck I. Stingl L. Maier M. Binder B.R. Inhibition of tissue kallikrein by protein C inhibitor.J Biol Chem. 1992; 267: 7048-7052Google Scholar) and as no other inhibitor of this serine protease has been described in skin, PCI might be a physiologically important inhibitor of tissue kallikrein in the epidermis. PCI seems to be a protein, associated with internal as well as external body surfaces, as it has been identified in numerous body fluids and in various epithelial cells (Laurell et al., 1992Laurell M. Christensson A. Abrahamsson P. Stenflo J. Lilja H. Protein C inhibitor in human body fluids.J Clin Invest. 1992; 89: 1094-1101Crossref PubMed Scopus (135) Google Scholar;Radtke et al., 1994Radtke K.P. Fernández J.A. Greengard J.S. et al.Protein C inhibitor is expressed in tubular cells of human kidney.J Clin Invest. 1994; 94: 2117-2124Crossref PubMed Scopus (44) Google Scholar). This protease inhibitor is unusual in that it inhibits chymotrypsin as well as or even better than trypsin, suggesting that it has a broad reactivity with various serine proteases (Suzuki et al., 1984Suzuki K. Nishioka J. Kusumoto H. Hashimoto S. Mechanism of inhibition of activated protein C inhibitor.J Biochem. 1984; 95: 187-195Crossref PubMed Scopus (100) Google Scholar;Cooper and Church, 1995Cooper S.T. Church F.C. Reactive site mutants of recombinant protein C inhibitor.Biochim Biophys Acta. 1995; 1246: 29-33Crossref PubMed Scopus (23) Google Scholar). It has been suggested that PCI in the male reproductive tract might function as a scavenger of prematurely activated acrosin, a serine protease stored as proacrosin in the acrosome of spermatozoa, and thereby protect the male reproductive system from proteolytic damage by this protease (Zheng et al., 1994Zheng X. Geiger M. Ecke S. et al.Inhibition of acrosin by protein C inhibitor and localization of protein C inhibitor to spermatozoa.Am J Physiol. 1994; 267: C466-C472PubMed Google Scholar). Similarly PCI might protect other body surfaces including the skin from proteolysis by providing unspecific protease inhibitory activity. PCI is a member of the family of serine protease inhibitors (serpins) (for review, seeHarper and Carrell, 1994Harper P.L. Carrell R.W. The serpins.in: Bloom A.L. Forbes C.D. Thomas D.P. Tuddenham Eg.D. Haemostasis and Thrombosis. Vol. 3. Churchill Livingstone, Edinburgh1994: 641-653Google Scholar). During the course of evolution, two members of this group, corticosteroid-binding globulin (Hammond et al., 1987Hammond G.L. Smith C.L. Goping I.S. et al.Primary structure of human corticosteroid binding globulin, deduced from hepatic and pulmonary cDNAs, exhibits homology with serine protease inhibitors.Proc Natl Acad Sci USA. 1987; 84: 5153-5157Crossref PubMed Scopus (228) Google Scholar) and thyroxine-binding globulin (Flink et al., 1986Flink I.L. Bailey T.J. Gustafson T.A. Markham B.E. Morkin E. Complete amino acid sequence of human thyroxine-binding globulin deduced from cloned DNA. Close homology to the serine antiproteases.Proc Natl Acad Sci USA. 1986; 83: 7708-7712Crossref PubMed Scopus (159) Google Scholar), have apparently exchanged their proteinase inhibitory properties for a new role as carriers of insoluble hormones. We have recently shown that all-trans-retinoic acid bound in a specific and concentration-dependent way to PCI but not to other serpins, suggesting an additional role of PCI as retinoic acid delivering serpin.3 The growth and differentiation of the epidermis and other epithelia is critically dependent upon undisturbed delivery of retinoids (Fisher and Voorhees, 1996Fisher G.J. Voorhees J.J. Molecular mechanisms of retinoid actions in skin.FASEB J. 1996; 10: 1002-1013Crossref PubMed Scopus (327) Google Scholar). Therefore epidermal PCI might not only serve as a protease inhibitor being involved in the regulation of the kallikrein–kinin system and/or the protection of the skin from various proteases, but might also take part in the regulation of retinoic acid supply in the skin. This work was supported in part by grants from the Austrian Science Foundation to M.G. (P 10823-Med and P 12308-Gen). The authors thank Dr. N.E. Fusenig (Deutsches Krebsforschungszentrum, Heidelberg, Germany) for providing them with the HaCaT cell line and Mr. T. Nardelli for artwork contributions.