- Email: [email protected]
Effect of O-Glycosylated Mucin on Invasion and Metastasis of HM7 Human Colon Cancer Cells
JOBNAME: BBRC 222#3 PAGE: 1 SESS: 10 OUTPUT: Thu Jun 6 18:49:14 1996 /xypage/worksmart/tsp000/71605g/14
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
222, 694–699 (1996)
0806
Effect of O-Glycosylated Mucin on Invasion and Metastasis of HM7 Human Colon Cancer Cells Wan-Hee Yoon,*,1 Hae-Duck Park,* Kyu Lim,† and Byung-Doo Hwang† *Department of Surgery and †Department of Biochemistry, Chungnam National University, School of Medicine, Taejon, Korea Received April 15, 1996 Mucinous colorectal cancers have a poorer prognosis than colorectal cancers which produce a low amount of mucin, but the exact mechanism is not well understood. The present study was undertaken to elucidate the possible mechanisms of invasion and metastasis of colon cancer cells producing high levels of mucin using mucin glycosylation inhibitor, benzyl-a-N-acetylgalactosamine. The binding activity of treated HM7 cells to endothelial leukocyte adhesion molecule (ELAM-1) was significantly decreased and fixed cell binding of MoAb SNH-3 and 19-9 (specific for sialyl Lex and sialyl Lea, respectively) was also significantly decreased. Metalloproteinase activity in conditioned medium and invasion of matrigel-coated porous filters by treated HM7 cells were decreased. However, there was no difference between control and treated HM7 cells in terms of matrix protein binding. These results suggest that O-glycosylated mucin is important in the invasive and metastatic properties of HM7 human colon cancer cells. © 1996 Academic Press, Inc.
Mucinous colorectal cancers have a poorer prognosis than that of colorectal cancers which produce low amount of mucin (1–4). Recently, a study reported that mucin production by human colon cancer cells correlated with their metastatic potential and affected their ability to colonize in the liver in experimental model system (5), but the mechanism of higher metastatic potential was not well elucidated. Mucins are the major secreted glycoproteins of the colon, and quantitative or qualitative changes in these high molecular weight O-linked glycoproteins could alter the biological behavior of colon cancer cells (6). Benzyl-a-N-acetylgalactosamine (benzyl-a-GalNAc) is an analogue of N-acetylgalactosamine, the first sugar in the mucin core region. This compound can compete as an acceptor for the elongation of oligosaccharides by glycosyltransferases and inhibit the formation of fully glycosylated mucin in a dose-dependent manner without affecting cell growth or viability (7). In the present investigations, we have examined the various properties on invasion and metastasis in vitro and the effect of inhibition of mucin glycosylation on these parameters was also compared. MATERIALS AND METHODS Cell line and culture condition. The HM7, a high mucin producing variant of LS174T human colon cancer cells (6,8) were grown in DMEM (Sigma, St. Louis, MO) containing 10% fetal calf serum (GibcoBRL, Life Technologies, Grand Island, NY), penicillin (100 U/ml) and streptomycin (100 mg/ml). Cultures were maintained at 37°C in a humidified 5% CO2 atmosphere. Inhibition of mucin glycosylation. The medium of 50% confluent cells was discarded and replaced with fresh medium containing 2 mM benzyl-a-N-acetylgalactosamine (Sigma, St. Louis, MO). The cells and the medium were harvested after incubation for 48 hrs and proceeded for use. Endothelial leukocyte adhesion molecule (ELAM-1) binding assay. Affinity-purified goat anti-human IgG (Sigma, St. Louis, MO) at a concentration of 10 mg/ml was adsorbed to the microtiter plates for 2 hrs at room temperature, and remaining protein binding sites were blocked by overnight incubation with 0.1% BSA in PBS. The coated plates were then incubated with 1 mg/ml ELAM-Rg or control chimeric protein of CD7 (Oncogen, Seatle, WA) in 0.1% BSA in PBS for 1 hr at 22°C and washed three times with PBS. To each well, 105 HM7 cells (benzyl-a-GalNAc treated or untreated) were
1 To whom all correspondence should be addressed. Department of Surgery, Chungnam National University, School of Medicine, 640 Daesa-dong, Joong-ku, Taejon, 301-040, Korea. Fax: 82-42-257-8024.
694 0006-291X/96 $18.00 Copyright © 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
JOBNAME: BBRC 222#3 PAGE: 2 SESS: 10 OUTPUT: Thu Jun 6 18:49:14 1996 /xypage/worksmart/tsp000/71605g/14
Vol. 222, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
added, and the plates were incubated for 20 min at 22°C with rotation at 100 rpm. After gentle washing twice with PBS, the attached cells were quantitated with a tetrazolium (MMT, Sigma, St. Louis, MO)—based colorimetric assay (9), reading absorbance at 540 nm. Fixed cell ELISA. The fixed cells in 96-well microtiter plates with 4% paraformaldehyde were assayed. Monoclonal antibodies, SNH-3 (Gift of Dr. A. Singhal, Seatle, WA) against sialyl Lex and 19-9 (Centocor, Malvern, PA) against sialyl Lea were used for evaluation of ELAM-1 determinants, whereas peroxidase-labeled peanut agglutinin (PNA, specific for T antigen), Vicia villosa agglutinin B4 (VVA, specific for Tn antigen) and Concanavalin A (Con A, irrelevant lectin) purchased from Sigma Chemical (St. Louis, MO) were used for detection of mucin core region antigens. Matrix-degrading metalloproteinases (MMPs) and urokinase-plasminogen activator (uPA) assay. MMPs activity of the culture medium was measured by thiopeptolide substrate assay (10,11) using a modified peptide thiopeptolide (Bachem, Philadelphia, PA) and DTDP (Sigma, St. Louis, MO). And uPA activity was determined by a method of Zimmerman et al (12). In vitro invasion and motility assay. Transwell cell culture chambers containing 6.5 mm diameter polycarbonate filters with 8 mm pores (Costar, Cambridge, MA), basement membrane matrigel (87 mg/filter, Collaborative Research, Bedford, MA) and 2 × 105 cells/well were used for the assay using a method described by Reich et al. (13) and Schlechte et al. (14), but with some modifications. After 72 hrs incubation of cultures, percent of invaded cells was calculated by MTT method. For the motility assay, the same system without matrigel was used. Cell-matrix protein binding assay. Basement membrane matrigel (6 mg/well) was added to microtiter plates. Cell suspension (105 cells/100 ml) was added and incubated for 1 hr at 37°C. After washing with serum free media gently, amounts of attached cells were detected by MTT colorimetric method. Statistical analysis. Significance of difference between the values was assessed using Student’s t-test. Statistical significance was assigned when P < 0.05.
RESULTS Effect of benzyl-a-GalNAc on endothelial leukocyte adhesion molecule (ELAM-1) binding and its determinants of HM7 cells. Colon cancer cells can bind to E-selection (ELAM-1), which recognizes both sialyl Lex and sialyl Lea on cell surface (15). To determine the effect of glycosylated mucin of cell surface on ELAM-1 binding, we performed the ELAM-1 binding assay with or without treatment of benzyl-a-GalNAc. The nonspecific binding activity of control and treated HM7 cells to CD7 protein was quite low, whereas ELAM-1 binding activities of both cells were significantly higher than that of CD7 protein (Table 1). The specific ELAM-1 binding activity (excluded with nonspecific binding) of treated HM7 cells were significantly inhibited by 63.7% as compared to control. Therefore, it is believed that mucin glycosylation inhibition of HM7 cells with benzyl-a-GalNAc can inhibit cell surface ELAM-1 binding epitopes. In order to determine whether the ELAM-1 binding epitopes were altered by mucin glycosylation inhibition, we examined the cell surface sialyl Lex and sialyl Lea as ELAM-1 binding determinants and peripheral mucin antigens, as well as mucin core region antigens using a fixed cell ELISA. Binding of monoclonal antibodies SNH-3 for sialyl Lex and 19-9 for sialyl Lea was decreased by about 25% (Fig. 1), whereas cell TABLE 1 ELAM-1 Binding Activities of HM7 Colon Cancer Cells with or without 2 mM Benzyl-a-GalNAc Treatment Cells
Protein
Binding Acitivtya
Control HM7
ELAM-1 CD7
0.440 ± 0.034 0.090 ± 0.008
Treated HM7
ELAM-1 CD7
0.217 ± 0.059* 0.090 ± 0.008
a Numbers represent the mean optical densities ± SD of five determinations. The values correlate with number of viable cells adherent to the adhesion molecules as determined by MTT colorimetric assay. Absorbance was read at 540 nm wavelength. * P < 0.01 vs. ELAM-1 binding of control HM7 cells.
695
JOBNAME: BBRC 222#3 PAGE: 3 SESS: 10 OUTPUT: Thu Jun 6 18:49:14 1996 /xypage/worksmart/tsp000/71605g/14
Vol. 222, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Binding of monoclonal antibodies to fixed cells, absorbance, 414 nm, HM7 cells treated for 48 hr with (■, v) or without (u, V) 2 mM benzyl-a-GalNAc were fixed with paraformaldehyde and assayed different dilutions of monoclonal antibodies SNH-3 (u, ■) and 19-9 (V, v).
surface binding of PNA for T antigen and of VVA for Tn antigen was increased (7.1-fold, and 2.6-fold, respectively; Fig. 2). However, the binding of irrelevant lectin, Con A, which recognize mannose was similar. Thus, benzyl-a-GalNAc treatment decreases the expression of peripheral carbohydrate antigens such as sialyl Lex and sialyl Lea, and increases the expression of core carbohydrate antigens of surface mucin of HM7 cells. Effect of benzyl-a-GalNAc on matrix-degrading metalloproteinases (MMPs) and urokinaseplasminogen activator (uPA). MMPs activity of HM7 conditioned medium was significantly inhibited by 34% with benzyl-a-GalNAc treatment (Table 2). The activity of uPA in control and treated HM7 culture medium were quite lower than 10% fetal calf serum, and no difference was observed between control and treated HM7 culture medium (Data not shown). Effect of benzyl-a-GalNAc on in vitro invasion and motility. To assess the invasive ability of HM7 cells by benzyl-a-GalNAc treatment, we performed in vitro invasion assay using a reconstituted basement membrane matrix. In this assay, like in vivo invasion, the cells in the top well must adhere to, hydrolyse and migrate through matrigel in order to enter the lower chamber. In this experiment, the treated HM7 cells showed a significant decreased invasion as compared to control cells by 37.2% (Fig. 3A). On the other hand, the percent of motility (Fig. 3B) and adhesion to matrigel (data not shown) were not different between control and treated cells. Since there was no difference in terms of matrix protein binding and cellular locomotion, as well as a decreased MMPs
FIG. 2. Binding of lectins to fixed cells, absorbance 414 nm. HM7 cells treated 48 hr with (m, v, ■) or without (n, V, u) 2 mM benzyl-a-GalNAc were fixed with paraformaldehyde and assayed different dilutions of lectins PNA (n, m), VVA (V, v) and Con A (u, ■). 696
JOBNAME: BBRC 222#3 PAGE: 4 SESS: 10 OUTPUT: Thu Jun 6 18:49:14 1996 /xypage/worksmart/tsp000/71605g/14
Vol. 222, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS TABLE 2 Matrix-Degrading Metalloproteinase (MMPs) Activity of HM7 Colon Cancer Cells with or without Treatment of 2 mM Benzyl-a-GalNAc Cells
MMPs Activitya (DOD/hr/106 cells)
Control HM7 Treated HM7
0.303 ± 0.022 0.200 ± 0.007*
Note. The enzyme activity was obtained in serum containing 24 hr spent medium by a thiopeptolide substrate assay. Activity was expressed as the change of optical densities adjusted to cell number at 324 nm for 1 min with 1 sec interval. a Mean ± SD of nine determinations. * P < 0.01 vs. untreated cells.
activity of treated cells in MMPs assay, the result of inhibited invasion of treated HM7 cells may be due to decreased MMPs activity by mucin glycosylation inhibition. DISCUSSION Cell membrane glycoproteins may play an important role in several stages of invasion and metastasis (16–21) including cell growth, motility during invasion, homotypic and heterotypic interactions between cells, platelet, and the immune system, and the adherence of metastatic cells to endothelium and the extracellular matrix. Mucins are the major secreted glycoprotein of the colon, and quantitative changes in these high-molecular weight O-linked glycoproteins could alter the biological behavior of tumors arising from this epithelium (6). A number of mucin peripheral carbohydrate antigens, including sialyl Lea and sialyl Lex, and mucin core region antigens, such as T and Tn, are known to present on colon cancer mucins (22–24). Endothelial leukocyte adhesion molecule (ELAM-1) is an endothelial cell adhesion molecule that allows myeloid cells or certain tumor cells to attach to the walls of blood vessels (25). ELAM-1
FIG. 3. Percentage invasion and motility of HM7 colon cancer cells with or without 2 mM benzyl-a-GalNAc treatment through transwell cell culture chambers of 8 mm pore size. See Materials and Methods. Percentage of control invaded cells or motile cells was normalized to 100%. The values are based on three experiments with triplicate samples performed in each experiments. The percentage invasion of benzyl-a-GalNAc treated HM7 cells was significantly different from control HM7 cells. (A) In the presence of matrigel, at 72 hr. Mean ± SD of 8.45 and 0.99% for control HM7 cells and 5.31 and 0.16% for treated cells (P < 0.05). (B) In the absence of matrigel, at 72 hr. Mean ± SD of 8.19 and 0.77% for control HM7 cells and 7.92 and 0.12% for treated cells (P > 0.05). 697
JOBNAME: BBRC 222#3 PAGE: 5 SESS: 10 OUTPUT: Thu Jun 6 18:49:14 1996 /xypage/worksmart/tsp000/71605g/14
Vol. 222, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS x
recognizes the sialyl Le and sialyl Lea determinants that expressed on the surface of some tumor cells as peripheral mucin antigens (15,25). In order to determine whether the ELAM-1 binding activity of HM7 human colon cancer cells is altered by inhibition of mucin glycosylation, we examined the ELAM-1 binding activity with benzyl-a-GalNAc treatment. There was a marked decrease in the activity of treated HM7 cells as much as 63.7% compared to control HM7 cells. The likely explanation for this is that there is inhibition of cell surface mucin peripheral antigens. To confirm this result, we performed fixed cell ELISA. Sialyl Lex and sialyl Lea, the ELAM-1 binding epitopes, were decreased by about 25% in treated HM7 cells, whereas mucin core region antigens, such as T and Tn were increased markedly. These results suggest surface mucin of colon cancer cell is critical for attachment to vascular endothelium for succeeding distant metastasis, and inhibition of mucin synthesis could be one way to inhibit distant metastasis due to decrease ELAM-1 binding. In order to complete the complex invasion and metastatic process, tumor cells must attach and traverse basement membrane barriers and penetrate vascular structures, both at primary and distant organs. This involves adhesion to basement membrane matrix, production of proteolytic enzymes and locomotion of the invasive cells (13,14,26–28). The adhesion of treated HM7 colon cancer cells to reconstituted basement membrane protein, matrigel was found to be not different when compared to adhesion by control cells. The important enzymes that have been shown to be closely associated with invasive and metastatic potential are MMPs and uPA (13,14,26,28). While the relationship between mucin and protease secretion remains unclear and somewhat speculative, it is possible that mucin-type glycoproteins may stimulate protease enzyme. The basement membrane glycoproteins laminin and vitronectin receptors for example, stimulate MMPs production (29,30). We have demonstrated that HM7 colon cancer cells produced MMPs especially type IV collagenases and serine proteinases (28). According to our results, inhibition of mucin production by benzyl-a-GalNAc decreased MMPs activity by 34%. While it is possible that the enzymatic activity of the MMPs was itself affected by benzyl-a-GalNAc, this is unlikely since these proteins have N-glycosylation sites (31), whereas benzyl-a-GalNAc is a specific inhibitor of O-glycosylation. In contrast to MMPs, secreted uPA activity of HM7 cells was quite low and not affected by benzyl-a-GalNAc. Among the various in vitro invasion assay models, transwell chambers are useful for rapid quantitation of invasive or motile cells (14). The results of in vitro motility and invasion assay showed that inhibition of mucin production decreased HM7 cell invasion by 37.2%, whereas it did not affect HM7 locomotion. Since there was no difference in terms of matrix protein binding and cellular locomotion, as well as decreased MMPs activity of HM7 cells inhibited mucin production, the result of decreased invasion of treated HM7 cells might be due to decreased MMPs activity by inhibition of mucin glycosylation. In summary, the present study indicates that glycosylated mucin modulates ELAM-1 binding and MMPs activity, therefore enhances invasive and metastatic properties of HM7 human colon cancer cells. ACKNOWLEDGMENT This work was supported by a grant from a Non-Directed Research Fund, Korea Research Foundation (1993).
REFERENCES 1. 2. 3. 4. 5.
Umpleby, H. C., Ranson, D. L., and Williamson, R. C. N. (1985) Br. J. Surg. 72, 715–718. Symond, D. L., and Vickery, A. L. (1976) Cancer 37, 1891–1900. Pihl, E., Nairn, R. C., Hughes, E. S. R., Cuthbertson, A. M., and Rollo, A. J. (1980) Pathology 12, 439–447. Wolfman, E. F., Astler, V. B., and Coller, F. A. (1957) Surgery 42, 846–852. Bresalier, R. S., Niv, Y., Byrd, J. C., Duh, Y., Toribara, N. W., Rockwell, R. W., Dahiya, R., and Kim, Y. S. (1991) J. Clin. Invest. 87, 1037–1045. 6. Kuan, S. F., Byrd, J. C., Basbaum, C. B., and Kim, Y. S. (1987) Cancer Res. 47, 5715–5724. 7. Kuan, S. F., Byrd, J. C., Basbaum, C., and Kim, Y. S. (1989) J. Biol. Chem. 264, 19271–19277. 698
JOBNAME: BBRC 222#3 PAGE: 6 SESS: 10 OUTPUT: Thu Jun 6 18:49:14 1996 /xypage/worksmart/tsp000/71605g/14
Vol. 222, No. 3, 1996 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Tom, B. H., Rutzky, L. P., Jakstys, M. M., Oyasu, R., Kaye, C. I., and Kahan, B. D. (1976) In Vitro 12, 180–191. Carmichael, J., De Graff, W. G., Gazdar, A. F., Minna, J. D., and Mitchell, J. B. (1987) Cancer Res. 47, 936–942. Weingarten, H., and Feder, J. (1985) Analytical Biochem. 147, 437–440. Weingarten, H., Martin, R., and Feder, J. (1985) Biochemistry 24, 6730–6734. Zimmerman, M., Quigley, J. P., Dorn, C., Goldfarb, R., and Troll, W. (1978) Proc. Natl. Acad. Sci. USA 75, 750–753. Reich, R., Thompson, E. W., Iwamoto, Y., Martin, G. R., Deason, J. R., Fuller, G. C., and Miskin, R. (1988) Cancer Res. 48, 3307–3312. Schlechte, W., Brattain, M., and Boyd, D. (1990) Cancer Commun. 2, 173–179. Takada, A., Ohmori, K., Yoneda, T., Tsuyuoka, K., Hasegawa, A., Kiso, M., and Kannagi, R. (1993) Cancer Res. 53, 354–361. Bresalier, R. S., Rockwell, R. W., Dahiya, R., Duh, Q. Y., and Kim, Y. S. (1990) Cancer Res. 50, 1299–1307. Hakomori, S. I. (1989) Adv. Cancer Res. 52, 257–331. Yamori, T., Kimura, H., Steward, K., Ota, D. M., Cleary, K. R., and Irimura, T. (1987) Cancer Res. 47, 2741–2747. Bresalier, R. S., Boland, C. R., and Kim, Y. S. (1984) Gastroenterology 87, 115–122. Kellokumpu, I. H. (1986) Cancer Res. 46, 4620–4625. Irimura, T., Carlson, D. A., Price, J., Yamori, T., Gianazzi, R., Ota, D. M., and Cleary, K. R. (1988) Cancer Res. 48, 2353–2360. Matsushita, Y., Nakamori, S., Seftor, E. A., Hendrix, M. J. C., and Irimura, T. (1991) Exp. Cell Res. 96, 20–25. Yamori, T., Ota, D. M., Cleary, K. R., Hoff, S., Hager, L. G., and Irimura, T. (1989) Cancer Res. 49, 887–894. Itzkowitz, S. H., Yuan, M., Fukushi, Y., Lee, H., Shi, Z., Zurawski, V., Hakomori, S., and Kim, Y. S. (1988) Cancer Res. 48, 3834–3842. Walz, G., Aruffo, A., Kolanus, W., Bevilacqua, M., and Seed, B. (1990) Science 250, 1132–1135. Liotta, L. A., and Stetler-Stevenson, W. G. (1991) Cancer Res. 51, 5054s–5059s. Dahiya, R., Yoon, W. H., Boyle, B., Schoenberg, S., Yen, T. B., and Narayan, P. (1992) Biochem. Int. 27, 567–577. Yoon, W. H., Sul, J. Y., Park, H. D., Lim, K., Hwang, B. D., Chang, E. S., Ho, J. J. L., and Kim, Y. S. (1995) Proc. Amer. Assn. Cancer Res. 36, 618. Kanemoto, T., Reich, R., Gratorex, D., Adler, S. H., Shiraishi, N., Martin, G., Yamada, Y., and Kleinman, J. K. (1990) Proc. Natl. Acad. Sci. USA 87, 2279–2283. Seftor, R. E. B., Seftor, E. A., Gehlsen, K. R., Stetler-Stevenson, W. G., Brown, P. D., Rouslahti, E., and Hendrix, M. J. C. (1992) Proc. Natl. Acad. Sci. USA 89, 1557–1561. Wilhelm, S. M., Collier, J. E., Marmer, B. C., Eisen, A. Z., Grant, G. A., and Goldberg, G. I. (1989) J. Biol. Chem. 264, 17213–17221.
699
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS ARTICLE NO.
222, 694–699 (1996)
0806
Effect of O-Glycosylated Mucin on Invasion and Metastasis of HM7 Human Colon Cancer Cells Wan-Hee Yoon,*,1 Hae-Duck Park,* Kyu Lim,† and Byung-Doo Hwang† *Department of Surgery and †Department of Biochemistry, Chungnam National University, School of Medicine, Taejon, Korea Received April 15, 1996 Mucinous colorectal cancers have a poorer prognosis than colorectal cancers which produce a low amount of mucin, but the exact mechanism is not well understood. The present study was undertaken to elucidate the possible mechanisms of invasion and metastasis of colon cancer cells producing high levels of mucin using mucin glycosylation inhibitor, benzyl-a-N-acetylgalactosamine. The binding activity of treated HM7 cells to endothelial leukocyte adhesion molecule (ELAM-1) was significantly decreased and fixed cell binding of MoAb SNH-3 and 19-9 (specific for sialyl Lex and sialyl Lea, respectively) was also significantly decreased. Metalloproteinase activity in conditioned medium and invasion of matrigel-coated porous filters by treated HM7 cells were decreased. However, there was no difference between control and treated HM7 cells in terms of matrix protein binding. These results suggest that O-glycosylated mucin is important in the invasive and metastatic properties of HM7 human colon cancer cells. © 1996 Academic Press, Inc.
Mucinous colorectal cancers have a poorer prognosis than that of colorectal cancers which produce low amount of mucin (1–4). Recently, a study reported that mucin production by human colon cancer cells correlated with their metastatic potential and affected their ability to colonize in the liver in experimental model system (5), but the mechanism of higher metastatic potential was not well elucidated. Mucins are the major secreted glycoproteins of the colon, and quantitative or qualitative changes in these high molecular weight O-linked glycoproteins could alter the biological behavior of colon cancer cells (6). Benzyl-a-N-acetylgalactosamine (benzyl-a-GalNAc) is an analogue of N-acetylgalactosamine, the first sugar in the mucin core region. This compound can compete as an acceptor for the elongation of oligosaccharides by glycosyltransferases and inhibit the formation of fully glycosylated mucin in a dose-dependent manner without affecting cell growth or viability (7). In the present investigations, we have examined the various properties on invasion and metastasis in vitro and the effect of inhibition of mucin glycosylation on these parameters was also compared. MATERIALS AND METHODS Cell line and culture condition. The HM7, a high mucin producing variant of LS174T human colon cancer cells (6,8) were grown in DMEM (Sigma, St. Louis, MO) containing 10% fetal calf serum (GibcoBRL, Life Technologies, Grand Island, NY), penicillin (100 U/ml) and streptomycin (100 mg/ml). Cultures were maintained at 37°C in a humidified 5% CO2 atmosphere. Inhibition of mucin glycosylation. The medium of 50% confluent cells was discarded and replaced with fresh medium containing 2 mM benzyl-a-N-acetylgalactosamine (Sigma, St. Louis, MO). The cells and the medium were harvested after incubation for 48 hrs and proceeded for use. Endothelial leukocyte adhesion molecule (ELAM-1) binding assay. Affinity-purified goat anti-human IgG (Sigma, St. Louis, MO) at a concentration of 10 mg/ml was adsorbed to the microtiter plates for 2 hrs at room temperature, and remaining protein binding sites were blocked by overnight incubation with 0.1% BSA in PBS. The coated plates were then incubated with 1 mg/ml ELAM-Rg or control chimeric protein of CD7 (Oncogen, Seatle, WA) in 0.1% BSA in PBS for 1 hr at 22°C and washed three times with PBS. To each well, 105 HM7 cells (benzyl-a-GalNAc treated or untreated) were
1 To whom all correspondence should be addressed. Department of Surgery, Chungnam National University, School of Medicine, 640 Daesa-dong, Joong-ku, Taejon, 301-040, Korea. Fax: 82-42-257-8024.
694 0006-291X/96 $18.00 Copyright © 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.
JOBNAME: BBRC 222#3 PAGE: 2 SESS: 10 OUTPUT: Thu Jun 6 18:49:14 1996 /xypage/worksmart/tsp000/71605g/14
Vol. 222, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
added, and the plates were incubated for 20 min at 22°C with rotation at 100 rpm. After gentle washing twice with PBS, the attached cells were quantitated with a tetrazolium (MMT, Sigma, St. Louis, MO)—based colorimetric assay (9), reading absorbance at 540 nm. Fixed cell ELISA. The fixed cells in 96-well microtiter plates with 4% paraformaldehyde were assayed. Monoclonal antibodies, SNH-3 (Gift of Dr. A. Singhal, Seatle, WA) against sialyl Lex and 19-9 (Centocor, Malvern, PA) against sialyl Lea were used for evaluation of ELAM-1 determinants, whereas peroxidase-labeled peanut agglutinin (PNA, specific for T antigen), Vicia villosa agglutinin B4 (VVA, specific for Tn antigen) and Concanavalin A (Con A, irrelevant lectin) purchased from Sigma Chemical (St. Louis, MO) were used for detection of mucin core region antigens. Matrix-degrading metalloproteinases (MMPs) and urokinase-plasminogen activator (uPA) assay. MMPs activity of the culture medium was measured by thiopeptolide substrate assay (10,11) using a modified peptide thiopeptolide (Bachem, Philadelphia, PA) and DTDP (Sigma, St. Louis, MO). And uPA activity was determined by a method of Zimmerman et al (12). In vitro invasion and motility assay. Transwell cell culture chambers containing 6.5 mm diameter polycarbonate filters with 8 mm pores (Costar, Cambridge, MA), basement membrane matrigel (87 mg/filter, Collaborative Research, Bedford, MA) and 2 × 105 cells/well were used for the assay using a method described by Reich et al. (13) and Schlechte et al. (14), but with some modifications. After 72 hrs incubation of cultures, percent of invaded cells was calculated by MTT method. For the motility assay, the same system without matrigel was used. Cell-matrix protein binding assay. Basement membrane matrigel (6 mg/well) was added to microtiter plates. Cell suspension (105 cells/100 ml) was added and incubated for 1 hr at 37°C. After washing with serum free media gently, amounts of attached cells were detected by MTT colorimetric method. Statistical analysis. Significance of difference between the values was assessed using Student’s t-test. Statistical significance was assigned when P < 0.05.
RESULTS Effect of benzyl-a-GalNAc on endothelial leukocyte adhesion molecule (ELAM-1) binding and its determinants of HM7 cells. Colon cancer cells can bind to E-selection (ELAM-1), which recognizes both sialyl Lex and sialyl Lea on cell surface (15). To determine the effect of glycosylated mucin of cell surface on ELAM-1 binding, we performed the ELAM-1 binding assay with or without treatment of benzyl-a-GalNAc. The nonspecific binding activity of control and treated HM7 cells to CD7 protein was quite low, whereas ELAM-1 binding activities of both cells were significantly higher than that of CD7 protein (Table 1). The specific ELAM-1 binding activity (excluded with nonspecific binding) of treated HM7 cells were significantly inhibited by 63.7% as compared to control. Therefore, it is believed that mucin glycosylation inhibition of HM7 cells with benzyl-a-GalNAc can inhibit cell surface ELAM-1 binding epitopes. In order to determine whether the ELAM-1 binding epitopes were altered by mucin glycosylation inhibition, we examined the cell surface sialyl Lex and sialyl Lea as ELAM-1 binding determinants and peripheral mucin antigens, as well as mucin core region antigens using a fixed cell ELISA. Binding of monoclonal antibodies SNH-3 for sialyl Lex and 19-9 for sialyl Lea was decreased by about 25% (Fig. 1), whereas cell TABLE 1 ELAM-1 Binding Activities of HM7 Colon Cancer Cells with or without 2 mM Benzyl-a-GalNAc Treatment Cells
Protein
Binding Acitivtya
Control HM7
ELAM-1 CD7
0.440 ± 0.034 0.090 ± 0.008
Treated HM7
ELAM-1 CD7
0.217 ± 0.059* 0.090 ± 0.008
a Numbers represent the mean optical densities ± SD of five determinations. The values correlate with number of viable cells adherent to the adhesion molecules as determined by MTT colorimetric assay. Absorbance was read at 540 nm wavelength. * P < 0.01 vs. ELAM-1 binding of control HM7 cells.
695
JOBNAME: BBRC 222#3 PAGE: 3 SESS: 10 OUTPUT: Thu Jun 6 18:49:14 1996 /xypage/worksmart/tsp000/71605g/14
Vol. 222, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Binding of monoclonal antibodies to fixed cells, absorbance, 414 nm, HM7 cells treated for 48 hr with (■, v) or without (u, V) 2 mM benzyl-a-GalNAc were fixed with paraformaldehyde and assayed different dilutions of monoclonal antibodies SNH-3 (u, ■) and 19-9 (V, v).
surface binding of PNA for T antigen and of VVA for Tn antigen was increased (7.1-fold, and 2.6-fold, respectively; Fig. 2). However, the binding of irrelevant lectin, Con A, which recognize mannose was similar. Thus, benzyl-a-GalNAc treatment decreases the expression of peripheral carbohydrate antigens such as sialyl Lex and sialyl Lea, and increases the expression of core carbohydrate antigens of surface mucin of HM7 cells. Effect of benzyl-a-GalNAc on matrix-degrading metalloproteinases (MMPs) and urokinaseplasminogen activator (uPA). MMPs activity of HM7 conditioned medium was significantly inhibited by 34% with benzyl-a-GalNAc treatment (Table 2). The activity of uPA in control and treated HM7 culture medium were quite lower than 10% fetal calf serum, and no difference was observed between control and treated HM7 culture medium (Data not shown). Effect of benzyl-a-GalNAc on in vitro invasion and motility. To assess the invasive ability of HM7 cells by benzyl-a-GalNAc treatment, we performed in vitro invasion assay using a reconstituted basement membrane matrix. In this assay, like in vivo invasion, the cells in the top well must adhere to, hydrolyse and migrate through matrigel in order to enter the lower chamber. In this experiment, the treated HM7 cells showed a significant decreased invasion as compared to control cells by 37.2% (Fig. 3A). On the other hand, the percent of motility (Fig. 3B) and adhesion to matrigel (data not shown) were not different between control and treated cells. Since there was no difference in terms of matrix protein binding and cellular locomotion, as well as a decreased MMPs
FIG. 2. Binding of lectins to fixed cells, absorbance 414 nm. HM7 cells treated 48 hr with (m, v, ■) or without (n, V, u) 2 mM benzyl-a-GalNAc were fixed with paraformaldehyde and assayed different dilutions of lectins PNA (n, m), VVA (V, v) and Con A (u, ■). 696
JOBNAME: BBRC 222#3 PAGE: 4 SESS: 10 OUTPUT: Thu Jun 6 18:49:14 1996 /xypage/worksmart/tsp000/71605g/14
Vol. 222, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS TABLE 2 Matrix-Degrading Metalloproteinase (MMPs) Activity of HM7 Colon Cancer Cells with or without Treatment of 2 mM Benzyl-a-GalNAc Cells
MMPs Activitya (DOD/hr/106 cells)
Control HM7 Treated HM7
0.303 ± 0.022 0.200 ± 0.007*
Note. The enzyme activity was obtained in serum containing 24 hr spent medium by a thiopeptolide substrate assay. Activity was expressed as the change of optical densities adjusted to cell number at 324 nm for 1 min with 1 sec interval. a Mean ± SD of nine determinations. * P < 0.01 vs. untreated cells.
activity of treated cells in MMPs assay, the result of inhibited invasion of treated HM7 cells may be due to decreased MMPs activity by mucin glycosylation inhibition. DISCUSSION Cell membrane glycoproteins may play an important role in several stages of invasion and metastasis (16–21) including cell growth, motility during invasion, homotypic and heterotypic interactions between cells, platelet, and the immune system, and the adherence of metastatic cells to endothelium and the extracellular matrix. Mucins are the major secreted glycoprotein of the colon, and quantitative changes in these high-molecular weight O-linked glycoproteins could alter the biological behavior of tumors arising from this epithelium (6). A number of mucin peripheral carbohydrate antigens, including sialyl Lea and sialyl Lex, and mucin core region antigens, such as T and Tn, are known to present on colon cancer mucins (22–24). Endothelial leukocyte adhesion molecule (ELAM-1) is an endothelial cell adhesion molecule that allows myeloid cells or certain tumor cells to attach to the walls of blood vessels (25). ELAM-1
FIG. 3. Percentage invasion and motility of HM7 colon cancer cells with or without 2 mM benzyl-a-GalNAc treatment through transwell cell culture chambers of 8 mm pore size. See Materials and Methods. Percentage of control invaded cells or motile cells was normalized to 100%. The values are based on three experiments with triplicate samples performed in each experiments. The percentage invasion of benzyl-a-GalNAc treated HM7 cells was significantly different from control HM7 cells. (A) In the presence of matrigel, at 72 hr. Mean ± SD of 8.45 and 0.99% for control HM7 cells and 5.31 and 0.16% for treated cells (P < 0.05). (B) In the absence of matrigel, at 72 hr. Mean ± SD of 8.19 and 0.77% for control HM7 cells and 7.92 and 0.12% for treated cells (P > 0.05). 697
JOBNAME: BBRC 222#3 PAGE: 5 SESS: 10 OUTPUT: Thu Jun 6 18:49:14 1996 /xypage/worksmart/tsp000/71605g/14
Vol. 222, No. 3, 1996
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS x
recognizes the sialyl Le and sialyl Lea determinants that expressed on the surface of some tumor cells as peripheral mucin antigens (15,25). In order to determine whether the ELAM-1 binding activity of HM7 human colon cancer cells is altered by inhibition of mucin glycosylation, we examined the ELAM-1 binding activity with benzyl-a-GalNAc treatment. There was a marked decrease in the activity of treated HM7 cells as much as 63.7% compared to control HM7 cells. The likely explanation for this is that there is inhibition of cell surface mucin peripheral antigens. To confirm this result, we performed fixed cell ELISA. Sialyl Lex and sialyl Lea, the ELAM-1 binding epitopes, were decreased by about 25% in treated HM7 cells, whereas mucin core region antigens, such as T and Tn were increased markedly. These results suggest surface mucin of colon cancer cell is critical for attachment to vascular endothelium for succeeding distant metastasis, and inhibition of mucin synthesis could be one way to inhibit distant metastasis due to decrease ELAM-1 binding. In order to complete the complex invasion and metastatic process, tumor cells must attach and traverse basement membrane barriers and penetrate vascular structures, both at primary and distant organs. This involves adhesion to basement membrane matrix, production of proteolytic enzymes and locomotion of the invasive cells (13,14,26–28). The adhesion of treated HM7 colon cancer cells to reconstituted basement membrane protein, matrigel was found to be not different when compared to adhesion by control cells. The important enzymes that have been shown to be closely associated with invasive and metastatic potential are MMPs and uPA (13,14,26,28). While the relationship between mucin and protease secretion remains unclear and somewhat speculative, it is possible that mucin-type glycoproteins may stimulate protease enzyme. The basement membrane glycoproteins laminin and vitronectin receptors for example, stimulate MMPs production (29,30). We have demonstrated that HM7 colon cancer cells produced MMPs especially type IV collagenases and serine proteinases (28). According to our results, inhibition of mucin production by benzyl-a-GalNAc decreased MMPs activity by 34%. While it is possible that the enzymatic activity of the MMPs was itself affected by benzyl-a-GalNAc, this is unlikely since these proteins have N-glycosylation sites (31), whereas benzyl-a-GalNAc is a specific inhibitor of O-glycosylation. In contrast to MMPs, secreted uPA activity of HM7 cells was quite low and not affected by benzyl-a-GalNAc. Among the various in vitro invasion assay models, transwell chambers are useful for rapid quantitation of invasive or motile cells (14). The results of in vitro motility and invasion assay showed that inhibition of mucin production decreased HM7 cell invasion by 37.2%, whereas it did not affect HM7 locomotion. Since there was no difference in terms of matrix protein binding and cellular locomotion, as well as decreased MMPs activity of HM7 cells inhibited mucin production, the result of decreased invasion of treated HM7 cells might be due to decreased MMPs activity by inhibition of mucin glycosylation. In summary, the present study indicates that glycosylated mucin modulates ELAM-1 binding and MMPs activity, therefore enhances invasive and metastatic properties of HM7 human colon cancer cells. ACKNOWLEDGMENT This work was supported by a grant from a Non-Directed Research Fund, Korea Research Foundation (1993).
REFERENCES 1. 2. 3. 4. 5.
Umpleby, H. C., Ranson, D. L., and Williamson, R. C. N. (1985) Br. J. Surg. 72, 715–718. Symond, D. L., and Vickery, A. L. (1976) Cancer 37, 1891–1900. Pihl, E., Nairn, R. C., Hughes, E. S. R., Cuthbertson, A. M., and Rollo, A. J. (1980) Pathology 12, 439–447. Wolfman, E. F., Astler, V. B., and Coller, F. A. (1957) Surgery 42, 846–852. Bresalier, R. S., Niv, Y., Byrd, J. C., Duh, Y., Toribara, N. W., Rockwell, R. W., Dahiya, R., and Kim, Y. S. (1991) J. Clin. Invest. 87, 1037–1045. 6. Kuan, S. F., Byrd, J. C., Basbaum, C. B., and Kim, Y. S. (1987) Cancer Res. 47, 5715–5724. 7. Kuan, S. F., Byrd, J. C., Basbaum, C., and Kim, Y. S. (1989) J. Biol. Chem. 264, 19271–19277. 698
JOBNAME: BBRC 222#3 PAGE: 6 SESS: 10 OUTPUT: Thu Jun 6 18:49:14 1996 /xypage/worksmart/tsp000/71605g/14
Vol. 222, No. 3, 1996 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Tom, B. H., Rutzky, L. P., Jakstys, M. M., Oyasu, R., Kaye, C. I., and Kahan, B. D. (1976) In Vitro 12, 180–191. Carmichael, J., De Graff, W. G., Gazdar, A. F., Minna, J. D., and Mitchell, J. B. (1987) Cancer Res. 47, 936–942. Weingarten, H., and Feder, J. (1985) Analytical Biochem. 147, 437–440. Weingarten, H., Martin, R., and Feder, J. (1985) Biochemistry 24, 6730–6734. Zimmerman, M., Quigley, J. P., Dorn, C., Goldfarb, R., and Troll, W. (1978) Proc. Natl. Acad. Sci. USA 75, 750–753. Reich, R., Thompson, E. W., Iwamoto, Y., Martin, G. R., Deason, J. R., Fuller, G. C., and Miskin, R. (1988) Cancer Res. 48, 3307–3312. Schlechte, W., Brattain, M., and Boyd, D. (1990) Cancer Commun. 2, 173–179. Takada, A., Ohmori, K., Yoneda, T., Tsuyuoka, K., Hasegawa, A., Kiso, M., and Kannagi, R. (1993) Cancer Res. 53, 354–361. Bresalier, R. S., Rockwell, R. W., Dahiya, R., Duh, Q. Y., and Kim, Y. S. (1990) Cancer Res. 50, 1299–1307. Hakomori, S. I. (1989) Adv. Cancer Res. 52, 257–331. Yamori, T., Kimura, H., Steward, K., Ota, D. M., Cleary, K. R., and Irimura, T. (1987) Cancer Res. 47, 2741–2747. Bresalier, R. S., Boland, C. R., and Kim, Y. S. (1984) Gastroenterology 87, 115–122. Kellokumpu, I. H. (1986) Cancer Res. 46, 4620–4625. Irimura, T., Carlson, D. A., Price, J., Yamori, T., Gianazzi, R., Ota, D. M., and Cleary, K. R. (1988) Cancer Res. 48, 2353–2360. Matsushita, Y., Nakamori, S., Seftor, E. A., Hendrix, M. J. C., and Irimura, T. (1991) Exp. Cell Res. 96, 20–25. Yamori, T., Ota, D. M., Cleary, K. R., Hoff, S., Hager, L. G., and Irimura, T. (1989) Cancer Res. 49, 887–894. Itzkowitz, S. H., Yuan, M., Fukushi, Y., Lee, H., Shi, Z., Zurawski, V., Hakomori, S., and Kim, Y. S. (1988) Cancer Res. 48, 3834–3842. Walz, G., Aruffo, A., Kolanus, W., Bevilacqua, M., and Seed, B. (1990) Science 250, 1132–1135. Liotta, L. A., and Stetler-Stevenson, W. G. (1991) Cancer Res. 51, 5054s–5059s. Dahiya, R., Yoon, W. H., Boyle, B., Schoenberg, S., Yen, T. B., and Narayan, P. (1992) Biochem. Int. 27, 567–577. Yoon, W. H., Sul, J. Y., Park, H. D., Lim, K., Hwang, B. D., Chang, E. S., Ho, J. J. L., and Kim, Y. S. (1995) Proc. Amer. Assn. Cancer Res. 36, 618. Kanemoto, T., Reich, R., Gratorex, D., Adler, S. H., Shiraishi, N., Martin, G., Yamada, Y., and Kleinman, J. K. (1990) Proc. Natl. Acad. Sci. USA 87, 2279–2283. Seftor, R. E. B., Seftor, E. A., Gehlsen, K. R., Stetler-Stevenson, W. G., Brown, P. D., Rouslahti, E., and Hendrix, M. J. C. (1992) Proc. Natl. Acad. Sci. USA 89, 1557–1561. Wilhelm, S. M., Collier, J. E., Marmer, B. C., Eisen, A. Z., Grant, G. A., and Goldberg, G. I. (1989) J. Biol. Chem. 264, 17213–17221.
699