inhibition of selectin-dependent tumor cell adhesion to endothelial cells and platelets by blocking...
TRANSCRIPT
Vol. 182, No. 3, 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
February 14, 1992 Pages 1288-1295
INHIBITION OF SELECTIN-DEPENDENT TUMOR CELL ADHESION TO ENDOTHELIAL CELLS AND PLATELETS BY BLOCKING
0-GLYCOSYLATION OF THESE CELLS
Naoya Kojimal, Kazuko Handal, Walter Newman2, and Sen-itiroh Hakomoril
lThe Biomembrane Institute, 201 Elliott Ave.W.,Seattle, WA 98119; and Dept. of Pathobiology, University of Washington, Seattle, WA
20tsuka Pharmaceutical Co., Maryland Research Laboratories, 9900 Medical Center Dr.,Rockville, MD 20850
Received January 3, 1992
SUMMARY: Expression of sialosyl-LeX (SLeX) and sialosyl-Lea (SLen) on tumor cell lines HL60, Colo205, and U937 was greatly suppressed by application of benzyl-cr- GalNAc for inhibition of O-linked carbohydrate chain extension, which resulted in reduced adhesion of tumor cells to activated endothelial cells or platelets mediated by ELAM-1 (E-selectin) or GMP-140 (P-selectin). Inhibitors or modifiers of N-glycosyla- tion had no effect on expression of SLeX or SLea in these tumor cells. These findings suggest the possibility that targeting of 0-glycosylation inhibitors or modifiers to tumor cells may effectively suppress metastatic potential. 0 1992 Academic Press, Inc.
Selectins, a family of adhesion molecules which includes ELAM-1 (E-selectin),
GMP-140 (CD62; P-selectin), and LECAM-1 (L-selectin), show a common structural
motif, and mediate adhesion of leukocytes and other cells, expressing a defined
carbohydrate (CHO ) epitope, to activated endothelial cells (ECs) or platelets (see for
reviews l-3). The lectin domains of ELAM-1 and GMP-140 have been shown to bind
to the CHO structures SLex (4-6) and SLe” (7-9). Since SLeX and SLen are known to
be major tumor-associated antigens (lo), these selectins are likely to play important
roles in defining tumor cell adhesion to platelets and ECs, and thus may be involved in
initiation of metastasis (11-13). Since MAbs directed to SLex or SLes, or oligosaccha-
The abbreviations used are: BzaGalNAc, benzyl-a-GalNAc; CHO, carbohydrate; EC, endothelial cell; GSL, glycosphingolipid; HUVEC, human umbilical vein endothelial cell; IL-l, interleukin-lp; MAb, monoclonal antibody; PBS, phosphate-buffered saline; PMA, phorbol 1Zmyristate 13-acetate; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis: SLea, sialosyl-Lea; SLex, sialosyl-LeX.
0006-291x/92 $1.50 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved. 1288
Vol. 182, No. 3, 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
ride sequences of these epitopes, block adhesion of neutrophils or tumor cells to
platelets or to ECs (4-9), these MAbs or oligosaccharide sequences are expected to be
useful in suppressing inflammatory processes or blocking metastatic deposition of tumor
cells.
Two additional approaches can be considered for inhibition of selectin-dependent
adhesion of leukocytes or tumor cells to platelets and ECs: (i) blocking expression of
selectin at the surface of ECs and platelets by appropriate inhibitors of transmembrane
signaling; (ii) blocking expression of SLeX, LeX, SLe”, or Lea at the surface of leuko-
cytes or tumor cells. Approach (i) has been successfully accomplished via application of
N,N-dimethyl- or N,N,N-trimethylsphingosine; these compounds block transmembrane
signaling via protein kinase C or other mechanisms (14). The approach described in
the present report falls in category (ii), specifically, blocking expression of SLeX or SLe”
through inhibition of extension of 0-glycosylation, and thereby inhibiting adhesion of
tumor cells to ECs or platelets via ELAM-1 or GMP-140 binding.
MATERIALS AND METHODS
Inhibition or modification of N- or 0-glvcosvlation of tumor cells. Promyelocytic leukemia HL60, colonic tumor Colo205, or histiocytic lymphoma U937 cells were cultured in RPMI-1640 supplemented with 10% fetal calf serum. For inhibition or modification of N-glycosylation, cells were cultured in medium containing swainsonine (10 pg/ml), castanospermine (100 pg/ml), l-deoxymannojirimycin (100 pg/ml) (Genzyme, Cambridge, MA), or tunicamycin (0.5 &ml) (Sigma Chemical Co., St. Louis, MO) for various durations (maximum 72 hr). Effects of these reagents on N-glycosylation have been summarized by Elbein (15). Although no true inhibitor of 0-glycosylation has yet been discovered, it is known that extension of 0-glycosylation can be blocked by culturing cells in the presence of 2 mM benzyl-cr-GalNAc (BzaGalNAc) for various durations (maximum 72 hr) (16). The accumulation of O-linked cr-GalNAc at the cell surface, resulting from this blocking, was quantified through increased reactivity with FITC-labeled Helix pomatia lectin (Sigma) as previously described (16). BzcrGlcNAc (2 mM) (Sigma) was used as a control reagent.
Surface SLea. SLeX. and plvcoprotein expression of cells treated with modifiers of N- or 0-alvcosvlation. Cells grown with or without glycosylation modifiers were incubated with anti-SLeX MAbs SNH3 and -4 (17), or anti-SLe” MAb CA-19-9 (IS), subsequently incubated with fluorescent F(ab’)z anti-mouse IgG, and analyzed by flow cytometry with EPICS (Coulter, Hialeah, FL). To assess the change of glycoproteins containing SLeX epitope by glycosylation modifiers, cells were lysed with 1% NP-40, 150 mM NaCl, 5 mM EDTA, 10 mM Tris, pH 7.4, containing phenylmethylsulfonyl fluoride (2 mM), aprotinin (100 pclg/ml), and leupeptin (100 r.&ml). Lysates were run on 6% SDS-PAGE gels under reducing conditions. Proteins were “Western blotted” with SNH4 and detected by Enhanced Chemiluminescence (ECL) Western blotting detection system (Amersham, Arlington Heights, IL) as per manufacturer’s instructions.
Selectin-dependent adhesion of HL60, Colo205. and U937 cells on activated HUVECs or platelets. HUVECs (obtained from Cell Systems, Kirkland, WA) were cultured to confluency in 48-well plates (Costar, Cambridge, MA) and stimulated with
1289
Vol. 182, No. 3, 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
1 U/ml IL-l for 4 hr. Non-stimulated HUVECs were used as a control. Expression of ELAM-1 on IL-l-stimulated HUVECs was confirmed by reactivity with anti-EL&I-l MAb 3B7 (IgGz,) (19). HL60, Colo205, and U937 cells were metabolically labeled by culture in the presence of [3H]thymidine after pretreatment with glycosylation modifier, and added to HUVEC-coated plates. After 15 min incubation, plates were washed with PBS, and adherent cell number estimated by conversion from radioactivity count. In another set of experiments, 96-well plates (Falcon, Lincoln, NJ) were coated with 1 &ml of a truncated, recombinant ELAM-1 lacking transmembrane and cytoplasmic domains for 18 hr. Plates were then coated with 1% BSA, washed with PBS, and coated with metabolically-labeled, glycosylation-modified cells as described above. After 60 min incubation, plates were washed with PBS and adherent cell number estimated by conversion from radioactivity count.
Assays of cell adhesion to activated or native platelets coated and fixed on 48- well plates were performed as previously described (14). HL60 and U937 cells were pretreated with 2 mM BzaGalNAc for 72 hr and labeled with [3H]thymidine. After washing with PBS, 1~10~ HL60 or U937 cells were added to each well, and plates were incubated for 30 min at room temp. After washing to remove unbound cells, bound cells were detached with trypsin and counted by liquid scintillation counter. Platelets bound on plates were incubated with anti-GMP-140 MAb IOP-62 (1:2, 1:6 dilution) (Immunotech, Marseille, France) at room temp for 30 min, followed by addition of HL60 cells, in order to evaluate dependence of adhesion on GMP-140 expression. Non-specific mouse IgG was used as control.
RESULTS
Inhibition of 0-alvcosvlation extension, but not of N-alvcosvlation, results in
blockinp of SLex or SLea exoression in HL60, Colo205. and U937 cells. Culturing of
these cells for 72 hr in the presence of 2 mM BzaGalNAc resulted in complete or
nearly-complete blocking of SLeX expression as measured by reactivity with anti-SLeX
MAb SNH4 for HI.60 cells, or blocking of SLe? expression as measured by reactivity
with anti-&$ MAb CA19-9 for Co10205 cells (data not shown). The flow cytometric
pattern for HL60 cells reflecting the effect of BzaGalNAc is shown as an example in
Fig. 1. Only BzcuGalNAc-treated cells showed greatly reduced reactivity with SNH4
(Fig. 1A) and significant reactivity with H. pomatia lectin (which recognizes a-GalNAc)
(Fig. 1B); i.e., BzaGalNAc blocks chain elongation of O-linked CHOs at the cell
surface, leading to accumulation of O-linked a-GalNAc without chain elongation.
In contrast, reagents affecting N-glycosylation, such as swainsonine and castano-
spermine (which blocks formation of N-linked complex-type structures), had no
inhibitory effect on SLex expression (Fig. lC,D). Note that whereas swainsonine had
negligible effect on SLex expression, castanospermine actually enhanced SLex expression
(Fig. lC, reaction by SNH3; Fig. lD, reaction with SNH4).
Glycoprotein patterns of HI-60 cells immunoblotted with SNH4 were similar for
swainsonine-treated vs. untreated control cells, i.e., there were two major broad bands
1290
Vol. 182, No. 3, 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
0 1 LOG FLUORESCENCE INTENSITY
200 K
’ 45 I:
66 K
Fieure 1. Flow cvtometric nattern of HL60 cells with or without BzaGalNAc treatment, probed bv anti-SLeX MAb SNH4 (Panel Al or bv Helix Domatia lectin (Panel B), or stained with MAbs SNH3 (Panel Cl or SNH4 (Panel D)
Panel A: -, control HL60 cells. -a-, cells treated with BzaGalNAc. -----, cells treated with BzcrGalNAc and stained with non-specific mouse IgG.
Panel B: -, control HL60 cells. -.-, cells treated with BzaGalNAc. Note that BzcrGalNAc-treated cells showed strong staining with H. pomatiu lectin relative to controls.
Panels C and D: -, control HL60 cells. -.-, cells treated with castano- spermine. -a . a-, cells treated with swainsonine. -----9 cells treated with castano- spermine and stained with non-specific mouse IgG.
Fieure 2. Patterns of SLeX-bearing glvcouroteins for HL60 cells treated with dvcosvla- tion modifiers
Glycoproteins were separated by SDS-PAGE on 6% gel and immunoblotted with anti-SLeX MAb SNHA Lane 1, untreated control cells. Lane 2, BzaGalNAc, 2 mM, 72 hr. Lane 3, swainsonine, 10 pg/ml, 72 hr.
.
with Mr ~200 and x60-70 kDa. These two bands were undetectable for BzcrGalNAc-
treated cells (Fig. 2).
Effect of 0- or N-alvcosvlation modification on adhesion of HL60, Colo205, and
U937 cells to HUVECs (through ELAM-1). These results are shown in Fig. 3. Only
those cells treated with BzcrGalNAc showed clear inhibition of adhesion on stimulated
HUVECs (Fig. 3A). This effect is clearly due to a loss of ELAM-l-dependent adhe-
sion, since only BzaGalNAc-treated cells showed significantly reduced adhesion on
ELAM-l-coated plastic plates (Fig. 3B).
Culturing of cells in the presence of various N-glycosylation inhibitors (castano-
spermine, swainsonine, deoxymannojirimycin) did not inhibit the expression of SLex or
SLe” in HL60 or Co10205 cells, nor affect adhesion of these cells to HUVECs (Fig.
4A). Although castanospermine actually increased SLex expression in HL60 cells (Fig.
1291
Vol. 182, No. 3, 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
4 6
1.5
1.0
0.5
0 I
T
1.1 In3, h-i Bs-CE@INAC - + - + - + Cells H L60 Co10 205 U937
-+ -+ -+ HL60 colops u937
Figure 3. Adhesion of HL60. Colo205. and U937 cells on HUVECs (Panel A) and on ELAM-l-coated plastic mates (Panel B)
Panel A: Solid columns, non-stimulated HUVECs. Open columns, HUVECs stimulated by IL-1 for 4 hr to induce ELAM-1 expression. - and +, without or with pretreatment by BzcuGalNAc. Panel B: Plastic plates were coated with ELAM-1 (10 &ml), washed, and incubated for 1 hr with HL60, Colo205, or U937 cells. Number of adherent cells was counted. - and +, without or with pretreatment by BzaGalNAc.
r A
Figure 4. Adhesion of HL60 (Panel A) and Co10205 (Panel B) cells to HUVECs after pretreatment with O-glvcosvlation and N-Plvcosvlation inhibitors
Solid columns, control HUVECs. Open columns, IL-l-stimulated HUVECs. Inhibitors used are indicated on the ordinate. “Medium,” no inhibitor added. Cells were incubated for 3 days, labeled with [3H]thymidine as described in the Fig. 5 legend, and added to HUVEC-coated wells. Adherent cells were estimated by radioactivity count. Tunicamycin was highly cytotoxic to HL60 cells and therefore not tested.
1292
Vol. 182, No. 3, 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
lC), it slightly decreased adhesion of these cells to HUVECs (Fig. 4A). Adhesion of
SLea-expressing Co10205 cells to activated HUVECs was also not affected by various N-
glycosylation inhibitors including tunicamycin (Fig. 4B). Adhesion of both HL60 and
Co10205 on stimulated HUVECs was only inhibited by BzcrGalNAc but not by its
isomer BzcrGlcNAc (Fig. 4A,B). These results suggest that SLeX or SLea epitope
present on N-glycosylated side chains or GSLs is not appropriately recognized by
selectin (see Discussion).
Inhibitory effect of BzaGalNAc on adhesion of HL60 or U937 cells to platelets
(GMP140-deoendent adhesion). Adhesion of both HL60 and U937 cells on activated
platelets was significantly reduced when cells were pretreated with BzaGalNAc. This
was observed with both thrombin-activated (Fig. 5A) and PMA-activated platelets (Fig.
5B). However, the degree of reduction was less striking than for ELAM-1 adhesion.
Adhesion of control cells to activated platelets is assumed to be dependent on GMP-
140 expression, since adhesion was greatly reduced by preincubation with anti-GMP-140
MAb IOP-62 (data not shown).
15 -3
‘0
X
-0 5 10
2 In
=
s 5
E3z-o-GalNAc - + - + Cells HL-60 u 937
Fieure 5. Effect of BzcuGalNAc oretreatment of HL60 and U937 cells on adhesion to activated platelets
Solid columns, control platelets. Open columns, thrombin-activated platelets. Hatched columns, PMA-activated platelets. - and +, without or with pretreatment by BzcrGalNAc. Values shown are mean f S.D. from triplicate wells, from a representative experiment with similar results obtained from two separate donors. Since pretreatment of platelets with anti-GMP-140 MAb IOP-62 reduced HL60 cell adhesion to activated platelets by x:85-90%, such adhesion is assumed to be due to GMP-140 expression.
1293
Vol. 182, No. 3, 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
DISCUSSION
Results of this study indicate that the CHO epitopes SLeX and SLe” in HL60
(promyelocytic leukemia cells), Co10205 (colon cancer cells), and U937 (histiocytic
lymphoma cells), capable of being recognized by selectins, are present mainly in the
form of O-linked chains. SLex or SLea may not be present as N-linked chains in these
tumor cells, although the presence of N-linked SLex structure in normal granulocytes
has been reported (20). SLex and SLe” epitopes are present in GSLs of these cells
(data not shown). GSLs may not be efficiently recognized by selectins, particularly by
ELAM-1. Although the exact rationale for discriminatory recognition of the same
epitope present in O-linked chain, N-linked chain, or GSL remains unclear, we know
that O-linked structures tend to be clustered, as observed in glycophorin A (21) and
leukosialin (22), and may be efficiently recognized by ELAM-1.
GMP-140-mediated cell adhesion was less affected by BzaGalNAc treatment than
was ELAM-l-mediated adhesion. A possibility remains that the presence of a second
receptor for GMP-140, such as sulfatide (23) or SLex or SLea GSL receptor, is better
recognized by GMP-140 than by ELAM-1.
In this study, the only 0-glycosylation modifier used was BzcrGalNAc, which
inhibits extension of O-linked CHO chains. However, other 0-glycosylation inhibitors
could be equally or even more effective in blocking selectin-dependent adhesion of
myelogenous cells or tumor cells. BzcrGalNAc is probably not the ideal reagent for this
purpose, since systemic administration of GalNAc is known to be hepatotoxic, causing
hepatitis (24). Therefore, other techniques for inhibiting 0-glycosylation are obviously
desirable. To block selectin-dependent adhesion of Lex- or SLeX-expressing myelogen-
ous or tumor cells, liposomes containing high concentrations of BzcrGalNAc could be
conjugated to MAbs, thereby targeting the cells in question and leading to inhibition of
inflammatory or metastatic processes. This approach, in addition to blocking selectin
expression or epitope binding, could play a part in development of “anti-adhesion”
therapy for various disease processes which involve selectin-dependent cell adhesion.
ACKNOWLEDGMENTS. The authors thank Dr. Anil Singhal for donation of MAbs SNH3 and -4, Dr. Stephen Anderson for scientific editing and preparation of the manuscript, and Ms. Tammi Keeler for technical assistance. This study was supported by funds from The Biomembrane Institute, in part under a research contract with Otsuka Pharmaceutical Co. S.H. is supported by National Cancer Institute Outstanding Investigator Grant CA42505
1294
Vol. 182, No. 3. 1992 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
1. 2. 3. 4.
5.
6.
7.
8.
9.
10. 11. 12.
13. 14.
15. 16.
17.
18.
19.
20.
21.
22. 23.
24.
REFERENCES
Stoolman, L.M. (1989) Cell 56, 907-910. Brandley, B.K., Swiedler, S.J., and Robbins, P.W. (1990) Cell 63, 861-863. Springer, T.A. and La&y, L.A. (1991) Nature 349, 196-197. Lowe, J.B., Stoolman, L.M., Nair, R.P., Larsen, R.D., Berhend, T.L., and Marks, R.M. (1990) Cell 63, 475-484. Phillips, M.L., Nudelman, E.D., Gaeta, F.C.A., Perez, M., Singhal, A.K., Hako- mori, S., and Paulson, J.C. (1990) Science 250, 1130-1132. Polley, M.J., Phillips, M.L., Wayner, E.A., Nudelman, E.D., Singhal, A.K., Hakomori, S., and Paulson, J.C. (1991) Proc. Natl. Acad. Sci. USA 88, 6224-6228. Berg, E.L., Robinson, M.K., Mansson, O., Butcher, E.C., and Magnani, J.L. (1991) J. Biol. Chem. 266, 14869-14872. Takada, A., Ohmori, K., Takahashi, N., Tsuyuoka, K., Yago, A., Zenita, K., Hasegawa, A., and Kannagi, R. (1991) Biochem. Biophys. Res. Commun. 179, 713-719. Handa, K., Nudelman, E.D., Stroud, M.R., Shiozawa, T., and Hakomori, S. (1991) Biochem. Biophys. Res. Commun. 181, 1223-1230. Hakomori, S. (1989) Adv. Cancer Res. 52, 257-331. Rice, G.E. and Bevilacqua, M.P. (1989) Science 246, 1303-1306. Walz, G., Aruffo, A., Kolanus, W., Bevilacqua, M.P., and Seed, B. (1990) Science 250, 1132-1135. Hakomori, S. (1991) Curr. Opin. Immunol. 3, 646-653. Handa, K., Igarashi, Y., Nisar, M., and Hakomori, S. (1991) Biochemistry 30, 11682-11686. Elbein, A.D. (1987) Ann. Rev. Biochem. 56, 497-534. Kuan, S.-F., Byrd, J.C., Basbaum, C., and Kim, Y.S. (1989) J. Biol. Chem. 264, 19271-19277. Muroi, K., Suda, T., Nojiri, H., Ema, H., Amemiya, Y., Miura, Y., Nakauchi, H., Singhal, A.K., and Hakomori, S. (1992 in press) Blood. Magnani, J.L., Nilsson, B., Brockhaus, M., Zopf, D., Steplewski, Z., Koprowski, H., and Ginsburg, V. (1982) J. Biol. Chem. 257, 14365-14369. Graber, N., Gopal, T.V., Wilson, D., Beall, L.D., Polte, T., and Newman, W. (1990) J. Immunol. 145, 819-830. Fukuda, M., Spooncer, E., Oates, J.E., Dell, A., and Klock, J.C. (1984) J. Biol. Chem. 259, 10925-10935. Tomita, M., Furthmayr, H., and Marchesi, V.T. (1978) Biochemistry 17, 4756-4770. Fukuda, M. (1991) Glycobiology 1, 347-356. Aruffo, A., Kolanus, W., Walz, G., Fredman, P., and Seed, B. (1991) Cell 67, 35-44. Liehr, H., Grun, M., Seelig, H.-P., Seelig, R., Reutter, W., and Heine, W.-D. (1978) Virchows Arch. B Cell. Path. 26, 331-344.
1295