Binding of Recombinant Human Cytokeratin 19 to Laminin

Binding of Recombinant Human Cytokeratin 19 to Laminin: A Possible Role in Interaction between Intermediate Filament Derived from Epithelial Cells and Extracellular Matrixes

Naomi Dobashi1, Jiro Fujita1,*, Masayuki Murota2, Yuji Ohtsuki3, Shuji Bandoh1, Yutaka Ueda1, Kazutaka Dohmoto1, Satoko Hojo1, Mikio Nishioka2, Toshihiko Ishida, and Jiro Takahara1 1First Department of Internal Medicine, Kagawa Medical University, Kagawa 2Third Department of Internal Medicine, Kagawa Medical University, Kagawa 3Department of Pathology, Kochi Medical School, Kochi, Japan

ABSTRACT. Cytokeratin 8 (CK8) and cytokeratin 19 (CK19) is a specific cytoskeletal component of simple epithelia, including bronchial epithelial cells. We hypothesized that CK8 or CK19 released from epithelial cells may bind to and cause damage to extracellular matrixes through binding of anti-CK8 or anti-CK19 autoantibodies. In the present study, bindings of recombinant human CK8 and CK19 to laminin (both derived from mouse sarcoma cells and human), collagen, gelatin, and fibronectin were evaluated by a modified enzyme-linked immunosorbent assay (ELISA). In addition, binding of CK19 to laminin was also confirmed by inhibition assay. As a result, CK19 strongly bound to mouse laminin as well as human laminin. Pretreatment with laminin significantly reduced the binding of CK19 to laminin. However, binding of recombinant CK19 to laminin was not demonstrated by Western immunoblot, suggesting that SDS treatment of laminin diminished the binding. These results suggest that released CK19 from epithelial cells may have played a role in the damage of basement membrane by accumulation of an immune complex composed by CK19 and anti-CK19 autoantibody.

Key words: cytokeratin 8/cytokeratin 19/laminin/extracellular matrix/epithelial cells

Cytokeratins (Cks) are a multigene family of polypeptides the aid of monoclonal antibodies in patients with lung can- which form the cytoskeleton of epithelial cells. On the basis cer (Pujol et al., 1993). In addition, a CYFRA 21-1 assay of differences in molecular weight and pH, at least 20 differ- has been developed. Combines 2 monoclonal antibodies di- ent types of CKs have been identified. These are expressed rected specifically against the degradation product of CK19 in certain sets along various routes of epithelial differentia- (Pujol et al., 1993). Clinical investigations have suggested tion in normal epithelium, tumor cells, and culture cells that serum CYFRA 21-1 is one of the most important tumor (Moll et al., 1982; Fuchs, 1988; Coulombe, 1993; Steinert, markers in the follow-up of non-small cell lung cancer, es- 1993). Cytokeratin 8 (CK8) and cytokeratin 19 (CK19) are pecially squamous cell lung cancer (Pujol et al., 1993). Fur- specific cytoskeletal structures of simple epithelia, includ- thermore, we have previously demonstrated that anti-CK8 ing bronchial epithelial cells (Moll et al., 1982; Fuchs, 1988; and anti-CK19 autoantibodies are formed in patients with Coulombe, 1993; Steinert, 1993). Although the precise idiopathic pulmonary fibrosis (Dobashi et al., 1998; Fujita mechanism of their generation remains unknown, fragments et al., 1999). of cytokeratin are soluble in serum and can be detected with In this study, we hypothesized that recombinant CK8 or recombinant CK19 may bind to extracellular matrixes and we demonstrate new evidence that recombinant CK19 has * To whom correspondence should be addressed: Jiro Fujita, M.D., First the ability to interact with the major basement membrane Department of Internal Medicine, Kagawa Medical University, 1750-1, protein, laminin. Miki-cho, Kita-gun, Kagawa, 761-0793, Japan. Tel: +81-87-891-2145, Fax: +81-87-891-2147 E-mail: [email protected] Abbreviations: CK8, cytokeratin 8; CK19, cytokeratin 19; ELISA, enzyme-linked immunosorbent assay; CKs, cytokeratins; SDS, sodium dodecyl sulfate.

171 N. Dobashi et al.

Materials and Methods Germany) was incubated in wells for 60 min. After washing the plate, bound CK8 or CK19 was incubated with mouse monoclonal anti-human CK8 (clone 18.7, Progen Biotechnik GMBH, Binding of cytokeratin 19 to laminin, collagen, gelatin Heidelberg, Germany, at 1:1000 dilution) or with mouse mono- and fibronectin clonal anti-human CK19 (clone 19.2, Progen Biotechnik GMBH, Tests for binding of CK8 or CK19 to various extracellular matrixes Heidelberg, Germany, at 1:1000 dilution). After incubation and were performed using modified enzyme-linked immunosorbent washing, the solid phase-bound mouse monoclonal anti-human assays (ELISA). For binding assays, polystyrene microtiter plates CK8 or CK19 antibody was further incubated with peroxidase- (ELISA plates, Cooke Laboratory Products Div., Dynatech conjugated goat anti-mouse IgG antibody (Sigma ImmunoChemi- Laboratories Inc., Alexandria, VA, USA) were coated with various cals, lot 38H4856, St Louis, MO, USA, diluted 1:1000) for 60 min. concentrations of mouse laminin (from basement membrane of After further washes, TMB Peroxidase EIA Substrate Kit (Bio- Engelbreth-Holm-Swarm Mouse sarcoma, lot 79H4000, Sigma Rad Laboratories, Hercules, CA, USA) was used to measure Chemical Co., St Louis, MO, USA), human laminin (derived from the amount of solid phase-bound antibodies. All incubations were human placenta, purified by immunoaffinity from pepsinized performed at room temperature. Data were expressed as placenta, lot 19H4052, Sigma Chemical Co.), collagen (Type I, mean±standard deviation from triplicate determinations. derived from calf skin, lot 87H2303, Sigma Chemical Co.), gelatin (Bio-Rad Laboratories, Hercules, CA, USA), and fibronectin (de- Avidity of binding of cytokeratin 19 to extracellular rived from human plasma, lot 38683, CAPPELTM Reseach Prod- matrix ucts, Durham, NC, USA). The purity of laminin was checked The binding of CK19 to extracellular matrixes was further charac- by Coomassie blue staining of a sodium dodecyl sulfate (SDS)- terized by testing the capacity of the extracellular matrixes to polyacrylamide gel. inhibit binding of CK19 to laminin. For inhibition assays, the After washing, commercially available recombinant human microtiter plate wells were coated with 0.5 mg/ml of laminin. CK8 (Progen Biotechnik GMBH, lot 65138, Heidelberg, Germa- Recombinant human CK19 was preincubated with several concen- ny) or CK19 (Progen Biotechnik GMBH, lot 503209, Heidelberg,

Fig. 1. Binding of human recombinant cytokeratin 8 and cytokeratin 19 to mouse laminin (A, B) and human laminin (C, D). Left bars represent the presence of recombinant cytokeratin 8 (A, C) or cytokeratin 19 (B, D) and right bars represent the absence of recombinant cytokeratin 19 (negative control). Cytokeratin 8 does not bind to mouse laminin (A) or human laminin (C). In contrast, human recombinant cytokeratin 19 binds to mouse laminin (B) as well as human laminin (D).

172 Binding of Cytokeratin 19 to Laminin

19.2, at 1:1000 dilution) and peroxidase conjugated goat anti- mouse IgG antibody (Sigma ImmunoChemicals, lot 38H4856, St. Louis, MO, USA), and stained with 4CN PLUS for chromogenic detection of horse radish peroxidase (NEMTM Life Science Prod- ucts, Boston, MA, USA). In addition, recombinant human CK19, recombinant human CK8, and albumin were transferred to nitrocellulose membrane and incubated with mouse laminin (2 mg/ ml) for 60 min. Mouse laminin was then detected by immunoblot- ting using rabbit anti-mouse laminin antibody (LSL Co., Ltd., Tokyo, Japan, lot 832011, at 1:100 dilution) and peroxidase conjugated goat anti-rabbit IgG antibody (Sigma ImmunoChemicals, lot 68H4875, St Louis, MO, USA), and stained with 4CN PLUS for chromogenic detection of horse radish peroxidase (NEMTM Life Science Products, Boston, MA, USA).

Results Human recombinant CK19 bound dose dependently to mouse laminin (Fig. 1B). In addition, human recombinant CK19 bound dose dependently to human laminin (Fig. 1D). In contrast, cytokeratin 8 did not bind to mouse laminin or

Fig. 2. Binding of human recombinant cytokeratin 8 (A) and cytokeratin 19 (B) to collagen. Left bars represent the presence of recombinant cytokeratin 8 (A, C) or cytokeratin 19 (B, D) and right bars represent the absence of recombinant cytokeratins (negative control). Human recombinant cytokeratin 19 slightly binds to collagen. In contrast, cytokeratin 8 does not bind to collagen. trations of laminin, and the amount of cytokeratin 19 bound to the wells was detected with enzyme conjugated antibodies as described above. Data were expressed as mean±standard deviation from triplicate determinations, and results were given as a percent- age of inhibition.

SDS-PAGE electrophoresis and Western blotting SDS-polyacrylamide gel electrophoresis was performed according to Laemmli’s method (Laemmli, 1970) with a slight modification. Commercially available laminin was mixed with SDS (2.0%) and heated (100°C, 5 min). The samples were then applied to a 10/20% SDS polyacrylamide gel, electrophoresed (60 mA, 120 min.), fixed in 50% methanol, 10% acetic acid, and stained with Coomassie Blue. Proteins were then electrophoretically transferred onto Fig. 3. Binding of human recombinant cytokeratin 8 (A) and cytokeratin nitrocellulose membrane by the method described previously 19 (B) to fibronectin. Left bars represent the presence of recombinant et al (Towbin ., 1979). Laminin transferred to nitrocellulose mem- cytokeratin 19 and right bars represent the absence of recombinant brane was incubated with recombinant human CK19 (2 mg/ml) for cytokeratin 19 (negative control). Human recombinant cytokeratin 19 60 min. Recombinant human CK19 was then detected by immu- slightly binds to fibronectin. In contrast, cytokeratin 8 does not bind to noblotting using the mouse monoclonal anti-human CK19 (clone fibronectin.

173 N. Dobashi et al.

fibronectin (Fig. 3B). In contrast, CK8 did not bind to fibronectin (Fig. 3A). The binding of CK19 to mouse laminin was inhibited by preincubation with mouse laminin (Fig. 4). However, the binding of CK19 to mouse laminin was not inhibited by the preincubation with human albumin or human transferrin (data not shown). By Western immunoblot analysis, we could not demonstrate that recombinant human CK19 bound to mouse or human laminin transferred to nitrocellulose membrane (Fig. 5A). However, we clearly demonstrated that mouse laminin bound to CK19 transferred to nitrocellulose membrane (Fig. 5B).

Discussion In our present study, we first demonstrated that recombinant CK19 bound to extracellular matrix, laminin. Laminin, the major glycoprotein in all basement membranes, consists of the thin extracellular matrices that under- lie epithelial cells and endothelial cells and surround nerve, muscle, and fat cells (Beck et al., 1990; Timpl et al., 1987). Laminin consists of two different B chains (B1 and B2) of about the same size (230 and 220 kDa) and a larger A chain (400 kDa), which are assembled into a cross-shaped structure stabilized by disulfide bonds (Beck et al., 1990; Timpl Fig. 4. Inhibition of binding of human cytokeratin 19 to microtiter plates coated with laminin by laminin. The binding of cytokeratin 19 to laminin is et al., 1987). This glycoprotein was first isolated from the inhibited by preincubated laminin. Engelbreth-Holm-Swarm tumor because it can readily be extracted and purified in intact form under nonreducing and nondenaturing conditions. This laminin exhibits an overall human laminin (Fig. 1A, 1C). Human recombinant CK19 identity of 74% in G2 to G5 regions with human laminin bound slightly to collagen (Fig. 2B). In contrast, cytokeratin and 90% in other parts of the molecule (Beck et al., 1990). 8 did not bind to collagen (Fig. 2A). Neither human recom- Numerous studies have indicated that laminin has diverse binant CK19 or recombinant CK8 bound to gelatin (data not biological activities which include stimulating adhesion, shown). Human recombinant CK19 bound slightly to migration, growth, and differentiation of various cells

Fig. 5. Western immunoblot analysis to demonstrated binding of CK19 to laminin.(A) Detection of recombinant human cytokeratin 19 bound to mouse or human laminin transferred to nitrocellulose membrane. M: molecular weight marker (dalton), lane 1: recombinant human cytokeratin 19 (positive control, arrow), lane 2: mouse laminin, lane 3: human laminin. Bindings of recombinant human cytokeratin 19 to mouse or human laminin are not demonstrated in this experiment. (B) Detection of mouse laminin bound to cytokeratin 19 transferred to nitrocellulose membrane. Lane 1: 5 mg of recombinant cytokeratin 19, lane 2: 2.5 mg of recombinant cytokeratin 19, lane 3: 5 mg of recombinant cytokeratin 8, lane 4: 5 mg of bovine albumin, M: molecular weight marker. Binding of mouse laminin to recombinant human cytokeratin 19 are clearly demonstrated (arrow).

174 Binding of Cytokeratin 19 to Laminin

(Sasaki et al., 1988; Aumailley et al., 1987; Graf et al., beginning a new era. Curr. Opin. Cell Biol., 5: 17–29. 1987; Kubota et al., 1992). It binds cells through cell sur- Dobashi, N., Fujita, J., Ohtsuki, Y., Yamadori, I., Yoshinouchi, T., Kamei, face receptors (integrins and nonintegrin proteins) and inter- T., Tokuda, M., Hojo, S., Okada, H., and Takahara, J. 1998. Detection of acts with other components of basement membranes includ- anti-cytokeratin 8 antibody in the serum of patients with cryptogenic fibrosing alveolitis and pulmonary fibrosis associated with collagen ing collagen IV, heparan sulfate proteoglycan, and nidogen/ vascular disorders. Thorax, 53: 969–974. entactin (Mecham, 1991; Clement et al., 1990; Languino et Fuchs, E. 1988. Keratins as biochemical markers of epithelial differentia- al., 1989; Shimizu et al., 1990). tion. Trends Genet., 4: 277–281. In the present study, we first demonstrated the binding of Fujita, J., Dobashi, N., Ohtsuki, Y., Yamadori, I., Yoshinouchi, T., Kamei, recombinant CK19 to mouse laminin as well as human T., Tokuda, M., Hojo, S., Okada, H., and Takahara, J. 1999. Elevation of laminin. We previously demonstrated the formation anti- anti-cytokeratin 19 antibody in sera of the patients with idiopathic CK19 autoantibody in patients with idiopathic pulmonary pulmonary fibrosis and pulmonary fibrosis associated with collagen vascular disorders. Lung, 177: 311–319. fibrosis (Fujita et al., 1999) and autoimmune hepatitis (in Garcia Lerma, J.G., Moneo, I., Ortiz de Landazuri, M., and Sequi Navarro, preparation). It has been reported that CK19 fragment is J. 1995. Comparison of the anti-laminin antibody response in patients released from epithelial cells by the digestion of protease with systemic lupus erythematosus (SLE) and parasitic diseases (filaria- (Nakamura et al., 1997). Therefore, there is a possibility sis). Clin. Immunol. Immunopathol., 76: 19–31. that this released CK19 fragment accumulates in the extra- Graf, J., Iwamoto, Y., Sasaki, M., Martin, G., Kleinman, H.K., Robey, F.A., cellular matrix, laminin. If CK19 or its fragments are re- and Yamada, Y. 1987. Identification of an amino acid sequence in lami- leased, it may accumulate in laminin, at which point the nin mediating cell attachment, chemotaxis and receptor binding. Cell, anti-CK19 autoantibody, which is formed in patients with 48: 989–996. Kubota, S., Tashiro, K., and Yamada, Y. 1992. Signaling site of laminin certain diseases, may attack the basement membrane. with mitogenic activity. J. Biol. Chem., 267: 4285–4288. Therefore, binding of CK19 to laminin suggests the possi- Laemmli, U.K. 1970. Cleavage of structural proteins during the assembly bility of damage to the basement membrane by the autoanti- of the head of bacteriophage T4. Nature, 227: 680–685. body against CK19 derived from epithelial cells. This evi- Languino, L.R., Gehlsen, K.R., Wayner, E., Carter, W.G., Engvall, E., and dence may be clinically important since anti-CK19 antibody Ruoslahti, E. 1989. Endothelial cells use a2b1 in tegrin as a laminin can react with CK19 as well as the basement membrane receptor. J. Cell Biol., 109: 2455–2462. through the binding of CK19 to laminin. Mecham, R.P. 1991. Receptors for laminin on mammalian cells. FASEB J., In the present study, we could not prove the binding of 5: 2538–2546. Moll, R., Franke, W.W., and Schiller, D.L. 1982. The catalog of human CK19 to mouse or human laminin transferred to the cytokeratins: patterns of expression in normal epithelia, tumors and cul- nitrocellulose membrane. It has been suggested that the tured cells. Cell, 31: 11–24. recognition of laminin as the autoantigen in SLE patients Nakamura, H., Abe, S., Shibata, Y., Yuki, H., Suzuki, H., Saito, H., Sata, depends on laminin conformation, which is related to the M., Kato, S., and Tomoike, H. 1997. Elevated levels of cytokeratin 19 in lack of reactivity observed by Western immunoblot in the bronchoalveolar lavage fluid in patients with chronic airway inflam- patients’ sera (Garcia Lerma et al., 1995). This is in agree- matory diseases- a specific marker for bronchial epithelial injury. Am. J. ment with previous observations in which the binding of Respir. Crit. Care Med., 155: 1217–1221. H241, a monoclonal anti-DNA antibody with anti-laminin Pujol, J.L., Grenier, J., Daures, J.P., Daver, A., Pujol, H., and Michel, F.B. 1993. Serum fragment of cytokeratin subunit 19 measured by CYFRA activity derived from an MRL-lpr/lpr mice, can be abol- 21-1 immunoradiometric assay as a marker of lung cancer. Cancer Res., ished with the exposure of laminin to sodium dodecyl sul- 53: 61–66. fate, whereas a polyclonal anti-laminin antibody also reacts Sabbaga, J., Line, S.R., Potocnjak, P., and Madaio, M.P. 1989. A murine with denatured laminin (Sabbaga et al., 1989). nephritogenic monoclonal anti-DNA autoantibody binds directly to In conclusion, our data demonstrate that CK19 specifical- mouse laminin, the major non-collagenous protein component of the ly bound to laminin. These results suggest that released glomerular basement membrane. Eur. J. Immunol., 19: 137–143. CK19 from epithelial cells may play a role in the damage of Sasaki, M., Kleinman, H.K., Huber, H., Deutzmann, R., and Yamada, Y. 1988. Laminin, a multidomain protein. J. Biol. Chem., 263: 16536– basement membrane by accumulation of the immune com- 16544. plex composed by CK19 and anti-CK19 autoantibody. Shimizu, Y., Van Seventer, G.A., Horgan, K.J., and Shaw, S. 1990. Regu- lated expression and binding of three VLA (b1) integrin receptors on T References cells. Nature, 345: 250–253. Steinert, P.M. 1993. Structure, function, and dynamics of keratin interme- Aumailley, M., Nurcombe, V., Edgar, D., Paulsson, M., and Timpl, R. diate filaments. J. Invest. Dermatol., 100: 729–734. 1987. The cellular interactions of laminin fragments. J. Biol. Chem., Timpl, R., Paulsson, M., Dziadek, M., and Fujiwara, S. 1987. Basement 262: 11532–11538. membranes. Methods Enzymol., 145: 363–391. Beck, K., Humter, I., and Engel, J. 1990. Structure and function of lami- Towbin, H., Staehelin, T., and Gordon, J. 1979. Electrophoretic transfer of nin: anatomy of a multidomain glycoprotein. FA SE B J., 4: 148–160. proteins from polyacrylamide gels to nitrocellulose sheets. Proc. Natl. Clement, B., Sequi-Real, B., Savagner, P., Kleinman, H.K., and Yamada, Y. Acad. Sci. USA, 76: 4350–4354. 1990. Hepatocyte attachment to laminin is mediated through multiple receptors. J. Cell Biol., 110: 185–192. (Received for publication, February 22, 2000 Coulombe, P.A. 1993. The cellular and molecular biology of keratins: and in revised form, April 20, 2000)

175