Proc. Nadl. Acad. Sci. USA Vol. 87, pp. 1541-1545, February 1990
Cell Biology
Carcinoembryonic antigen functions as an accessory adhesion molecule mediating colon epithelial cell-collagen interactions (extracellular matrix/receptor/Arg-Gly-Asp/integrin)
MASSIMO PIGNATELLI, HELGA DURBIN, AND WALTER F. BODMER Director's Laboratory, Imperial Cancer Research Fund, Lincoln's Inn Field, London WC2A 3PX, United Kingdom
Contributed by Walter F. Bodmer, November 20, 1989
ABSTRACT We have previously shown that a human colon carcinoma cell line (SW1222) expresses a collagen receptor recognizing the Arg-Gly-Asp tripeptide sequence found in collagen. This receptor mediates the cellular attachment to collagen and, subsequently, the glandular differentiation seen in a three-dimensional collagen gel culture. In a search to identify cell surface molecules mediating the adhesion and differentiation of SW1222 cells, we have screened a panel of monoclonal antibodies recognizing epithelial cell surface determinants for their ability to inhibit the collagen binding of SW1222 cells. We have found that four monoclonal antibodies recognizing the 180-kDa carcinoembryonic antigen (CEA) glycoprotein and other members of the CEA family inhibited (up to 87%) the binding of SW1222 cells to type I collagen matrix. Using a cell attachment assay, we have not detected any direct collagen binding of either purified CEA or another CEAexpressing human colon carcinoma cell line (LS174T). These data suggest that CEA is not a collagen-binding protein itself but is likely to be associated with the functional Arg-Gly-Asp collagen receptor expressed by SW1222 cells. We suggest that CEA may function as an accessory molecule, controlling the functional activity of the SW1222 collagen receptor.
which is the binding site of several extracellular matrix cellular receptors (9). To characterize further this and other cell adhesion molecules mediating epithelial cell-collagen interaction, we screened a panel of monoclonal antibodies (mAbs) recognizing epithelial determinants for their ability to inhibit the binding of SW1222 cells to type I and type IV collagens. PR3B1O mAb, which has been shown to react with CEA (10), inhibited the binding of SW1222 cells only to type I collagen matrix. This finding coupled with the known homology between CEA and other cell adhesion molecules (5, 7) led us to investigate whether CEA could also be a cell adhesion molecule mediating the collagen binding of colon epithelial cells. Here we show that four different mAbs recognizing the classical 180-kDa CEA glycoprotein and other members of the CEA family have similar inhibitory activity on the binding of SW1222 cells to type I collagen. In addition, we show that CEA is not a collagen-binding protein itself but is likely to be associated with the Arg-Gly-Asp-directed collagen receptor complex expressed by the SW1222 human colon carcinoma cell line.
MATERIALS AND METHODS
Carcinoembryonic antigen (CEA), first described by Gold and Freedman (1) in 1965, is a highly glycosylated 180-kDa glycoprotein, which was thought to be an oncofetal protein specific for the colon. It is now known to be one member of a family of at least 10 related gene products that share structural and sequence homology and are found on a variety of malignant and normal tissues (2). Recently cDNA sequencing of the gene for the protein core of CEA has revealed a highly significant homology with members of the immunoglobulin gene family (3-5). A key functional feature of the immunoglobulin gene family molecules is their role in cell surface recognition in immunity (e.g., major histocompatibility antigens, CD2, and lymphocyte function-associated antigen 3) and in controlling the behavior of cells in various tissues (e.g., neural cell adhesion molecule) (6). The striking homology between CEA and the members of the immunoglobulin gene family has led to the suggestion that CEA may share some of their functional characteristics (5, 7). We have shown (8) that type I collagen, an extracellular matrix protein, stimulates a human colon carcinoma cell line (SW1222) to express a nearly normal differentiated phenotype in vitro. Thus, when grown in a three-dimensional type I collagen gel, these cells form gland-like structures analogous to normal intestinal crypts, with cells organized around a central lumen and nuclei polarized to the periphery. Furthermore, we have shown that this interaction between collagen and the colon carcinoma cells is mediated by a cell surface collagen receptor that recognizes the Arg-Gly-Asp tripeptide sequence,
Cells. The SW1222 and LS174T human colon carcinomaderived cell lines (11, 12) were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% (vol/vol) fetal calf serum at 37°C in 10% C02/90% air. mAbs. PR3B10, PR4D6, PR5C5, and PR6B5 are part of a panel of mAbs recognizing antigens expressed by normal and transformed colorectal epithelial cells (13) and were selected for their strong reactivity with the SW1222 human colon carcinoma cell line. Their production and characterization have been described (13). MP4F10, MP6F3, MP6D6, and MP3D9 mAbs were obtained by immunizing BALB/c mice with live SW1222 colon carcinoma cells and were produced by standard techniques (14). The initial screening for antibody production from the fusion was performed by using a f3-galactosidase/anti-/3-galactosidase enzyme-linked immunosorbent assay (GAG-ELISA) (15) performed on the original immunizing cell line (SW1222) (M.P. and W.F.B., unpublished data). Three purified mouse mAbs reacting either selectively with CEA (601, Sera-Lab, Crawley Down, Sussex, U.K.) or also with other members of the CEA family (C46, Amersham; Oxoid anti-CEA, Oxoid, Basingstoke, U.K.) were also used. Synthetic Peptides. The Ser-Arg-Gly-Asp-Thr-Gly and SerArg-Gly-Glu-Thr-Gly peptides were synthesized by using a peptide synthesizer (model 430A, Applied Biosystems) by the solid-phase technique (16) and were assayed for purity by amino acid analysis and analytical high-performance liquid
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Abbreviations: CEA, carcinoembryonic antigen; mAb, monoclonal antibody; GAG-ELISA, f-galactosidase/anti-,p-galactosidase enzyme-linked immunosorbent assay. 1541
1542
Cell Biology: Pignatelli et al.
chromatography with an Aquafore RP-300 reverse-phase column (17). Cell Adhesion Assay. The adhesion of SW1222 and LS174T human colon carcinoma cell lines to human type I or type IV collagen (Sigma) was assayed as described (8). The ability of a series of mAbs reacting with members of the CEA family and other cell surface epithelial determinants to inhibit the collagen binding of SW1222 cells was assayed by including the mAbs in the incubation medium (DMEM containing 2.5 mg of bovine serum albumin per ml and 25 mM Hepes, pH 7.4). The PR3B10, PR4D6, PR5C5, PR6B5, MP4F10, MP6F3, MP6D6, and MP3D9 mAbs were used as undiluted supernatant from culture or 10 times concentrated and were diluted 1:1 with the incubation medium. The CEA mAbs (C46, 601, and Oxoid) were obtained in purified form and used at 10 and 100 ,ug/ml (diluted with the incubation medium). Cells were also incubated with BBM1 mAb, which recognizes 92-microglobulin (18), and with incubation medium alone as negative controls. Binding of Human CEA to Type I Collagen. Ninety-six-well microtiter plates (Dynatech) were coated with both 20 and 40 tkg/ml of either human type I collagen or bovine serum albumin as described (8). Stock solution of purified CEA from liver metastases (Sigma) was diluted with phosphatebuffered saline (PBS) composed of NaCl at 10 mg/ml, KCI at 250 gg/ml, Na2HPO4 at 1.43 mg/ml, and KH2PO4 at 250 ,tg/ml (pH 7.3) and added to the previously coated plates at 20, 2, and 0.2 ,ug/ml for 3 hr at room temperature. The SW1222 and LS174T human colon carcinoma cells were also included in the cell attachment assay (105 cells per well) as positive and negative controls, respectively. After two washes with PBS, the unoccupied sites were blocked by adding PBS containing 0.2% gelatin and 0.2% sodium azide for 1 hr at room temperature. Plates were then washed twice with PBS containing 0.2% casein (Oxoid). Fifty microliters of PR3B10 mAb was added in duplicate for 1 hr, and the binding of antibody to collagen-bound CEA was detected by a GAG-ELISA (15). Immunoprecipitation. Purified human CEA (Sigma) was labeled with 1251I (Amersham) by using iodogen (Pierce) (19) and was then precipitated with the following mAbs reacting with members of the CEA family: PR3B10 (13), 601, C46, and Oxoid anti-CEA. The DH12 mouse mAb recognizing the P, chain (20), shared by the integrin ,31 subfamily of cell surface heterodimeric extracellular matrix receptors (21), was used as negative control. The immunoprecipitated fractions were then analyzed on a NaDodSO4/7.5% polyacrylamide gel (22).
RESULTS Screening of mAbs Recognizing Epithelial Cell Surface Determinants in a Cell Attachment Assay. A series of mAbs recognizing epithelial cell surface determinants was analyzed for the ability to inhibit the binding of SW1222 human colon carcinoma cell line to type I and to type IV collagen by a cell attachment assay. Only PR3B10 mAb specifically inhibited (50%) the binding of SW1222 cells to type I collagen (Fig. 1A). The binding to type IV collagen was not altered by any mAb tested (Fig. 1B). Reactivity of PR3B10 mAb with Human CEA. To define the specificity of PR3B10 mAb and its reactivity with CEA, 50 ,l of PR3B10 tissue culture supernatant was added to 96-well microtiter plates previously coated with serial concentrations of purified human CEA (Sigma). The binding of PR3B10 to CEA was then determined by using the GAG-ELISA (15). Three monoclonal antibodies reacting with CEA (601, C46, and Oxoid) were included in the assay as positive controls. W6/32, a mAb recognizing the major histocompatibility antigens A, B, and C associated with the light chain /82-
Proc. Natl. Acad. Sci. USA 87 (1990) PR 3B10 PR 5C5 PR 4D6 PR 6B5 MP 6D6 MP 3D9 MP 6F3 BBMI MP 4F10 CONTROL C2 2
ID
4C
'
PR 3B10 PR 5C5
PR 4D6 PR 6B5
_
MP 6D6
_
MP 3D9
MP 6F3
BBM1 MP 4F10
CONTROL 0
20
40
62
Arbitrary fluorescence units FIG. 1. Screening of a panel of mAbs (PR3B10, PRC5, PR4D6, PR6B5, MP6D6, MP3D9, MP6F3, MP4F10, and BBM1) recognizing
epithelial cell surface determinants for their ability to inhibit the binding of SW1222 colon carcinoma cells to type I (A) and type IV (B) collagen. * , Specific binding to collagens (20 gg/ml); , binding to bovine serum albumin (20 ,ug/ml). Cells (5 x 104) were resuspended with 50 ,I of each mAb (culture supernatant concentrated 10 times and diluted 1:1 with DMEM containing 2.5 mg of bovine serum albumin per ml and 25 mM Hepes, pH 7.4) and were allowed to attach to the coated plates for 2 hr at room temperature. The unattached cells were washed away with PBS, and the number of attached cells was estimated by measuring alkaline phosphatase (8). Values are expressed in arbitrary fluorescence units (mean ± SD of three determinations).
microglobulin (23), was used as negative control. PR3B10 mAb showed a specific"" reactivity with human CEA comparable to the other three anti-CEA mAbs (Fig. 2A). In the second experiment, human CEA was 125I-labeled by the iodogen method and was immunoprecipitated with PR3B10 and the three other anti-CEA mAbs (601, C46, and Oxoid). The immunoprecipitated fractions were then examined by NaDodSO4/7.5% PAGE. PR3B10 and the other three anti-CEA mAbs immunoprecipitated a band with an approximate molecular weight of 180 kDa, consistent with the published size of CEA (24). Inhibition of SW1222 Collagen Binding by Other mAbs Reacting with CEA and CEA-Related Glycoproteins. To evaluate whether the inhibitory activity of PR3B10 mAb on SW1222 cells was specifically due to its CEA activity, three other mAbs (601, C46, and Oxoid) recognizing CEA and other members of the CEA family were also tested in a cell
Cell
Biology: Pignatelli et al.
Proc. Natl. Acad. Sci. USA 87 (1990) 192
3 A
5
1543
B
200U)
cn .
0 a)
92.5-
a) CD 0
U) 0 cys
.,
67-
4610
1
CEA concentration,jug/ml FIG. 2. (A) Reactivity of PR3B10 (m), 601 (n), C46 (o), and Oxoid (x) mAbs with human CEA. mAbs were added to 96-well microtiter plates previously coated with serial concentrations of human purified CEA (0.2-20 Ag/ml) for 3 hr at room temperature, and their binding to CEA was then determined by GAG-ELISA (15). W6/32 mAb (A), which recognizes the major histocompatibility antigens A, B, and C associated with the light chain 832-microglobulin (23), was used as negative control. (B) Immunoprecipitation of 251I-labeled human CEA with DH12 (lane 1), PR3B10 (lane 2), Oxoid (lane 3), 601 (lane 4), and C46 (lane 5) mAbs. The DH12 mAb, which reacts with the ,81 integrin chain (20), was used as negative control.
attachment assay. The MP6F3 mAb, which also recognizes an epitope expressed by SW1222 cells (M.P. and W.F.B., unpublished data) and has been shown not to be involved in the collagen binding in our preliminary screening (Fig. 1A), was used as negative control. All three anti-CEA mAbs inhibited the binding of SW1222 colon carcinoma cells to type I collagen. They showed inhibition of cell adhesion ranging from 72% (mAb 601) to 87% (mAb C46). These results were comparable to the inhibitory effect of mAb PR3B10 (78%) SRGETG
SRGDTG I3B10
OXOID
DISCUSSION
60Q
C46
'SI P6F3 0
20
40
60
80
1 00
1 20
% cell attachment FIG. 3. Inhibition of the binding of SW1222 cells to collagencoated plates by anti-CEA mAbs and the Arg-Gly-Asp-containing peptide. Cells (5 x 104) were allowed to attach for 2 hr at room temperature in the presence of PR3B10 at a concentration 10 times that in the culture supernatant, Oxoid at 100 Ag/ml, 601 at 100 ,ug/ml, C46 at 100 ,Ag/ml, and MP6F3 at 10 times the culture supernatant concentration mAbs and the Ser-Arg-Gly-Asp-Thr-Gly (SRGDTG in single-letter code) and Ser-Arg-Gly-Glu-Thr-Gly (SRGETG) synthetic peptides at 2 mg/ml. The unattached cells were washed away with PBS, and the number of attached cells was estimated by measuring alkaline phosphatase activity (8). Values are expressed as a percentage of the maximal binding (mean ± SD of three determi-
nations).
and an Arg-Gly-Asp-containing peptide (75%) (Fig. 3) as previously shown (8). Type I Collagen Binding Activity of CEA. A cell attachment assay was used to test whether CEA had the ability to bind specifically type I collagen matrix. Purified human CEA was applied to type I collagen plates, and the CEA molecules bound were detected by using PR3B10 mAb. The antibody binding to collagen-bound CEA was then detected by using the GAG-ELISA (15). No binding of CEA to type I collagen was found (Fig. 4A), suggesting that CEA is not a collagenbinding protein itself. Since both Arg-Gly-Asp-containing peptides and the anti-CEA mAbs block SW1222 collagen binding, it is likely that CEA (or CEA-related antigens) is a subunit of the SW1222 collagen receptor. LS174T human colon carcinoma cells, which are known to express the 180-kDa CEA glycoprotein and other members of the CEA family (13, 25) as do SW1222 cells (Fig. 4B), did not bind type I collagen-coated plates (Fig. 4A), supporting our hypothesis that CEA does not have the ability to bind collagen by itself.
In this study we have shown that several mAbs recognizing the 180-kDa CEA glycoprotein and other members of the CEA family were able to inhibit the binding of a CEAexpressing human colon carcinoma cell line (SW1222) to type I collagen matrix. Using a cell attachment assay, we failed to detect any direct collagen binding of either purified human CEA or another human colon carcinoma cell line (LS174T) that expresses high levels of CEA and CEA-related antigens (25). These data suggest that CEA is not a collagen-binding protein itself but is likely to be associated with a functional collagen receptor complex expressed by SW1222 cells. Therefore, we propose that CEA may function as an accessory molecule, controlling the functional activity of the SW1222 collagen receptor. We have shown (8) that the collagen receptor complex expressed by SW1222 cells mediates the cellular attachment to collagen and subsequently the glandular differentiation seen in three-dimensional collagen gel (8). This receptor recognizes the Arg-Gly-Asp tripeptide sequence present in collagen and requires the presence of divalent cations, such
1544
Cell Biology: Pignatelli et al.
'W
TYPE COLL
Proc. Natl. Acad. Sci. USA 87 (1990)
BSA
0
4000 D~~~~~~~~~ 3000
2000
1000
0
PBS
MP 6F3
PR
3B10
601
OXOID
FIG. 4. (A) Type I collagen binding activity of CEA (in), bovine albumin (BSA) (0), and SW1222 (ml) and LS174T (u) colon carcinoma cells. Purified human CEA (20 jug/ml) was applied to type I collagen- or bovine serum albumin-coated plates (20 jug/ml) for 3 hr at room temperature. SW1222 and LS174T cells (105 cells per well) were included as positive and negative controls, respectively. The CEA molecules bound were detected by using PR3B10 mAb in GAG-ELISA (15). (B) Expression of MP6F3, PR3B1O, 601, and Oxoid mAbs on SW1222 ( ) and LS174T ( *) colon carcinoma cells serum
by GAG-ELISA (15).
Ca2+ or Mg2+ or both, for its functional activity (26). Several extracellular matrix receptors have been now characterized by affinity chromatography and by inhibition of cell adhesion to extracellular matrix proteins by specific mAbs. Some of them belong to the integrin family of cell surface receptors, which are a,8-heterodimeric transmembrane proteins divided into three subfamilies (,1, /32, and 33), based on the sharing of a common p chain (21). The Arg-Gly-Asp tripeptide sequence constitutes the cell binding site of some extracellular matrix receptors that belong to the 81 (27, 28) and integrin subfamilies (29). The majority of integrin receptors also require divalent cations for their binding (9, 21). However, some reports have also shown that some of the Arg-Gly-Asp-directed extracellular matrix receptors do not share structural and functional similarities with the integrin as
33
family (30, 31). The inhibition of the SW1222 collagen binding by anti-CEA mAbs suggests that CEA may control the function and/or the affinity of the Arg-Gly-Asp-directed collagen receptor on SW1222 cells. It is likely that the Arg-Gly-Asp-directed collagen receptor provides the primary specificity of attach-
ment of SW1222 cells to collagen, whereas CEA may exert a modulatory influence on the cell-collagen interaction. This hypothesis is supported by the fact that we were unable to show any direct binding of human CEA to collagen-coated plates. In addition, a human colon carcinoma cell line (LS174T), which expressed high levels of the 180-kDa CEA glycoprotein and other members of the CEA family, did not specifically bind collagen. Cheresh et al. (32, 33) have shown that cell surface gangliosides, which are sialic acid-containing glycolipids, can similarly modulate the function of some of the Arg-Gly-Asp-directed extracellular matrix receptors. They showed that mAbs directed specifically to the carbohydrate moiety of the digangliosides GD3 and GD2 inhibited melanoma and neuroblastoma cell attachment to collagen, fibronectin, laminin, and vitronectin (32). As we show here for CEA, the role of gangliosides in cell attachment also does not involve their direct binding with the extracellular matrix proteins, but they seem to be part of the divalent cationdependent functional complex involving the Arg-GlyAsp-directed receptor (33). The accessory function of several cell surface molecules that are involved in immune recognition has been described (34-36). The functional interaction between T-cell antigen receptor-CD3 antigen complex and the CD2 molecule, which are also members of the immunoglobulin gene family, has been shown to be due to a physical association of these T-lymphocyte cell surface structures (37). CEA may function similarly as an associated cell surface molecule that interacts with the collagen receptor expressed by colon epithelial cells. Whether the 180-kDa CEA glycoprotein itself or other members of the CEA family are involved in cell-collagen interaction is unclear. The 601 mAb was the only mAb with no known cross-reactivity with other CEA-related glycoproteins (nonspecific cross-reactive antigen and biliary glycoprotein 1) (38, 39) according to the manufacturer. This mAb showed a similar inhibitory activity on the SW1222 collagen binding, suggesting that CEA itself is the cell adhesion molecule. However, further functional studies using mAbs specific for other members of the CEA family will be required to answer this issue fully. Recently Benchimol et al. (7) have shown that CEA can affect the homotypic sorting of cells in heterogeneous populations of aggregating cells by functioning as a Ca2+_ independent homotypic intercellular adhesion molecule. Since CEA is highly expressed basolaterally during gut embryogenesis, they proposed that CEA could play a role as an intercellular adhesion molecule in intestinal organization (7). Here we show that CEA also can function as an accessory cell adhesion molecule, mediating cell-matrix interaction, which is also a fundamental process necessary for the development of a polarized epithelial cell phenotype (40). CEA shows, therefore, a strong structural and functional similarity with the neural cell adhesion molecule, which is also a member of the immunoglobulin gene family (6) and is involved in cell-matrix and cell-cell interaction among neurons, glia, and muscle cells during embryogenesis (41). The role played by CEA in colon carcinogenesis is rather unclear. In colorectal tumors, CEA is normally found expressed on the entire surface of the colon epithelial cells, showing, however, no relationship with the pathological extent of the disease, clinical prognosis, or the degree of tumor differentiation (42, 43). We suggest, in fact, that the detection of high levels of CEA in colorectal cancers may be only an epiphenomenon resulting from the disruption of cell-matrix interactions because of the loss of the CEA-associated extracellular matrix receptors by tumor cells. We have recently shown that loss of two integrin collagen receptors (very late antigen 2 and very late antigen 3) occurs relatively frequently in colorectal cancer (44), and this allows tumor cells to escape the controlling mechanisms on cell differentiation induced by
Cell
Proc. Nati. Acad. Sci. USA 87 (1990)
Biology: Pignatelli et al.
collagen matrix and those soluble factors that enhance differentiation and inhibit proliferation through effects on collagen binding (26). We are grateful to Dr. B. De Strooper for the generous gift of DH12 mAb and to Dr. M. J. Crumpton for helpful comments. 1. Gold, P. & Freedman, S. 0. (1965) J. Exp. Med. 121, 439-462. 2. Rogers, G. T. (1983) Biochim. Biophys. Acta 695, 227-249. 3. Zimmermann, W., Ortlieb, B., Friedrich, R. & von Kleist, S. (1987) Proc. Nati. Acad. Sci. USA 84, 2960-2964. 4. Thompson, J. A., Pande, H., Paxton, R. J., Shively, L., Padma, A., Simmer, R. L., Todd, C. W., Riggs, A. D. & Shively, J. E. (1987) Proc. Natl. Acad. Sci. USA 84, 29652969. 5. Barnett, T. R., Kretschmer, A., Douglas, A. A., Goebel, S. J., Hart, J. T., Elting, J. J. & Kamarck, M. E. (1989) J. Cell Biol. 108, 267-276. 6. Williams, A. F. & Barclay, A. N. (1988) Annu. Rev. Immunol. 6, 381-405. 7. Benchimol, S., Fuks, A., Jothy, S., Beauchemin, N., Shirota, K. & Stanners, C. P. (1989) Cell 57, 327-334. 8. Pignatelli, M. & Bodmer, W. F. (1988) Proc. Natd. Acad. Sci. USA 85, 5561-5565. 9. Ruoslahti, E. & Pierschbacher, M. D. (1987) Science 238, 491-497. 10. Richman, P. I. & Bodmer, W. F. (1988) J. Pathol. 156, 197211. 11. Leibovitz, A., Stinson, J. C., McCombs, W. B., III, McCoy, C. E., Mazur, K. C. & Mabry, N. D. (1976) Cancer Res. 36, 4562-4569. 12. Tom, B. H., Rutzky, L. P., Jakstys, M. M., Oyasu, R., Kaye, G. I. & Kahan, B. D. (1976) In Vitro 12, 180-191. 13. Richman, P. I. & Bodmer, W. F. (1987) Int. J. Cancer 39, 317-328. 14. Galfrd, G., Howe, S. C., Milstein, C., Butcher, G. W. & Howard, J. C. (1977) Nature (London) 266, 550-552. 15. Durbin, H. & Bodmer, W. F. (1987) J. Immunol. Methods 97, 19-27. 16. Barany, G. & Merrifield, R. (1979) in The Peptides, eds. Gross, E. & Meienhofer, J. (Academic, New York), pp. 1-284. 17. Rothbard, J., Lechler, R., Howland, K., Bal, V., Eckels, D., Sekaly, R., Long, E., Taylor, W. & Lamb, J. (1988) Cell 52,
515-523. 18. Brodsky, F. M., Bodmer, W. F. & Parham, P. (1979) Eur. J. Immunol. 9, 536-545. 19. Fraker, P. J. & Speck, J. C. (1978) Biochem. Biophys. Res. Commun. 80, 849-857. 20. Jaspers, M., De Strooper, B., Spaepen, M., van Leuven, F.,
21. 22. 23. 24. 25. 26. 27.
28. 29. 30. 31.
32.
1545
David, G., van den Berghe, H. & Cassiman, J. J. (1988) FEBS Lett. 231, 402-406. Hynes, R. 0. (1987) Cell 48, 549-554. Laemmli, U. K. (1970) Nature (London) 227, 680-685. Barnstable, C. J., Bodmer, W. F., Brown, G., Galfre, G., Milstein, C., Williams, A. F. & Ziegler, A. (1978) Cell 14,9-20. Slayter, H. S. & Coligan, J. E. (1975) Biochemistry 14, 23232330. Toribara, N. W., Sack, T. L., Gum, J. R., Ho, S. B., Shively, J. E., Willson, J. K. V. & Kim, Y. S. (1989) Cancer Res. 49, 3321-3327. Pignatelli, M. & Bodmer, W. F. (1989) Int. J. Cancer 44, 518-523. Pierschbacher, M. D. & Ruoslahti, E. (1984) Proc. Natl. Acad. Sci. USA 81, 5985-5988. Bourdon, M. A. & Ruoslahti, E. (1989) J. Cell Biol. 108, 1149-1155. Pytela, R., Pierschbacher, M. D. & Ruoslahti, E. (1985) Proc. Natl. Acad. Sci. USA 82, 5766-5770. Dedhar, S., Ruoslahti, E. & Pierschbacher, M. (1987) J. Cell Biol. 104, 585-593. Lu, M. L., Becham, D. A. & Jacobson, B. C. (1989) J. Biol. Chem. 264, 13546-13558. Cheresh, D. A., Pierschbacher, M. D., Herzig, M. A. & Mu-
joo, K. (1986) J. Cell Biol. 102, 688-696. 33. Cheresh, D. A., Pytela, R., Pierschbacher, M. D., Klier, F. G., Ruoslahti, E. & Reisfeld, R. A. (1987) J. Cell Biol. 105, 1163-1173. 34. Alcover, A., Alberini, C., Acuto, O., Clayton, L. K., Transy, C., Spagnoli, G. C., Moingeon, P., Lopez, P. & Reinherz, E. L. (1988) EMBO J. 7, 1973-1977. 35. Bockenstedt, L. K., Goldsmith, M. A., Dustin, M., Olive, D., Springer, T. A. & Weiss, A. (1988) J. Immunol. 141, 19041911. 36. Altmann, D. M., Hogg, N., Trowsdale, J. & Wilkinson, D. (1989) Nature (London) 338, 512-514. 37. Brown, M. H., Cantrell, D. A., Brattsand, G., Crumpton, M. J. & Gullberg, M. (1989) Nature (London) 339, 551-553. 38. von Kliest, S., Chavanel, G. & Burtin, P. (1972) Proc. Natl. Acad. Sci. USA 69, 2492-2494. 39. Svenberg, T. (1976) Int. J. Cancer 17, 588-596. 40. Boulan, E. R. & Nelson, W. J. (1989) Science 245, 718-725. 41. Rutishauser, U. (1984) Nature (London) 310, 594-597. 42. Ahnen, D. J., Kinoshita, K., Nakane, P. K. & Brown, W. R. (1987) Gastroenterology 93, 1330-1338. 43. Shirota, K., Minassian, H. & Jothy, S. (1988) Exp. Mol. Pathol. 49, 305-315. 44. Pignatelli, M., Smith, M. E. F. & Bodmer, W. F. (1990) Br. J. Cancer, in press.
Cell Biology
Carcinoembryonic antigen functions as an accessory adhesion molecule mediating colon epithelial cell-collagen interactions (extracellular matrix/receptor/Arg-Gly-Asp/integrin)
MASSIMO PIGNATELLI, HELGA DURBIN, AND WALTER F. BODMER Director's Laboratory, Imperial Cancer Research Fund, Lincoln's Inn Field, London WC2A 3PX, United Kingdom
Contributed by Walter F. Bodmer, November 20, 1989
ABSTRACT We have previously shown that a human colon carcinoma cell line (SW1222) expresses a collagen receptor recognizing the Arg-Gly-Asp tripeptide sequence found in collagen. This receptor mediates the cellular attachment to collagen and, subsequently, the glandular differentiation seen in a three-dimensional collagen gel culture. In a search to identify cell surface molecules mediating the adhesion and differentiation of SW1222 cells, we have screened a panel of monoclonal antibodies recognizing epithelial cell surface determinants for their ability to inhibit the collagen binding of SW1222 cells. We have found that four monoclonal antibodies recognizing the 180-kDa carcinoembryonic antigen (CEA) glycoprotein and other members of the CEA family inhibited (up to 87%) the binding of SW1222 cells to type I collagen matrix. Using a cell attachment assay, we have not detected any direct collagen binding of either purified CEA or another CEAexpressing human colon carcinoma cell line (LS174T). These data suggest that CEA is not a collagen-binding protein itself but is likely to be associated with the functional Arg-Gly-Asp collagen receptor expressed by SW1222 cells. We suggest that CEA may function as an accessory molecule, controlling the functional activity of the SW1222 collagen receptor.
which is the binding site of several extracellular matrix cellular receptors (9). To characterize further this and other cell adhesion molecules mediating epithelial cell-collagen interaction, we screened a panel of monoclonal antibodies (mAbs) recognizing epithelial determinants for their ability to inhibit the binding of SW1222 cells to type I and type IV collagens. PR3B1O mAb, which has been shown to react with CEA (10), inhibited the binding of SW1222 cells only to type I collagen matrix. This finding coupled with the known homology between CEA and other cell adhesion molecules (5, 7) led us to investigate whether CEA could also be a cell adhesion molecule mediating the collagen binding of colon epithelial cells. Here we show that four different mAbs recognizing the classical 180-kDa CEA glycoprotein and other members of the CEA family have similar inhibitory activity on the binding of SW1222 cells to type I collagen. In addition, we show that CEA is not a collagen-binding protein itself but is likely to be associated with the Arg-Gly-Asp-directed collagen receptor complex expressed by the SW1222 human colon carcinoma cell line.
MATERIALS AND METHODS
Carcinoembryonic antigen (CEA), first described by Gold and Freedman (1) in 1965, is a highly glycosylated 180-kDa glycoprotein, which was thought to be an oncofetal protein specific for the colon. It is now known to be one member of a family of at least 10 related gene products that share structural and sequence homology and are found on a variety of malignant and normal tissues (2). Recently cDNA sequencing of the gene for the protein core of CEA has revealed a highly significant homology with members of the immunoglobulin gene family (3-5). A key functional feature of the immunoglobulin gene family molecules is their role in cell surface recognition in immunity (e.g., major histocompatibility antigens, CD2, and lymphocyte function-associated antigen 3) and in controlling the behavior of cells in various tissues (e.g., neural cell adhesion molecule) (6). The striking homology between CEA and the members of the immunoglobulin gene family has led to the suggestion that CEA may share some of their functional characteristics (5, 7). We have shown (8) that type I collagen, an extracellular matrix protein, stimulates a human colon carcinoma cell line (SW1222) to express a nearly normal differentiated phenotype in vitro. Thus, when grown in a three-dimensional type I collagen gel, these cells form gland-like structures analogous to normal intestinal crypts, with cells organized around a central lumen and nuclei polarized to the periphery. Furthermore, we have shown that this interaction between collagen and the colon carcinoma cells is mediated by a cell surface collagen receptor that recognizes the Arg-Gly-Asp tripeptide sequence,
Cells. The SW1222 and LS174T human colon carcinomaderived cell lines (11, 12) were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% (vol/vol) fetal calf serum at 37°C in 10% C02/90% air. mAbs. PR3B10, PR4D6, PR5C5, and PR6B5 are part of a panel of mAbs recognizing antigens expressed by normal and transformed colorectal epithelial cells (13) and were selected for their strong reactivity with the SW1222 human colon carcinoma cell line. Their production and characterization have been described (13). MP4F10, MP6F3, MP6D6, and MP3D9 mAbs were obtained by immunizing BALB/c mice with live SW1222 colon carcinoma cells and were produced by standard techniques (14). The initial screening for antibody production from the fusion was performed by using a f3-galactosidase/anti-/3-galactosidase enzyme-linked immunosorbent assay (GAG-ELISA) (15) performed on the original immunizing cell line (SW1222) (M.P. and W.F.B., unpublished data). Three purified mouse mAbs reacting either selectively with CEA (601, Sera-Lab, Crawley Down, Sussex, U.K.) or also with other members of the CEA family (C46, Amersham; Oxoid anti-CEA, Oxoid, Basingstoke, U.K.) were also used. Synthetic Peptides. The Ser-Arg-Gly-Asp-Thr-Gly and SerArg-Gly-Glu-Thr-Gly peptides were synthesized by using a peptide synthesizer (model 430A, Applied Biosystems) by the solid-phase technique (16) and were assayed for purity by amino acid analysis and analytical high-performance liquid
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Abbreviations: CEA, carcinoembryonic antigen; mAb, monoclonal antibody; GAG-ELISA, f-galactosidase/anti-,p-galactosidase enzyme-linked immunosorbent assay. 1541
1542
Cell Biology: Pignatelli et al.
chromatography with an Aquafore RP-300 reverse-phase column (17). Cell Adhesion Assay. The adhesion of SW1222 and LS174T human colon carcinoma cell lines to human type I or type IV collagen (Sigma) was assayed as described (8). The ability of a series of mAbs reacting with members of the CEA family and other cell surface epithelial determinants to inhibit the collagen binding of SW1222 cells was assayed by including the mAbs in the incubation medium (DMEM containing 2.5 mg of bovine serum albumin per ml and 25 mM Hepes, pH 7.4). The PR3B10, PR4D6, PR5C5, PR6B5, MP4F10, MP6F3, MP6D6, and MP3D9 mAbs were used as undiluted supernatant from culture or 10 times concentrated and were diluted 1:1 with the incubation medium. The CEA mAbs (C46, 601, and Oxoid) were obtained in purified form and used at 10 and 100 ,ug/ml (diluted with the incubation medium). Cells were also incubated with BBM1 mAb, which recognizes 92-microglobulin (18), and with incubation medium alone as negative controls. Binding of Human CEA to Type I Collagen. Ninety-six-well microtiter plates (Dynatech) were coated with both 20 and 40 tkg/ml of either human type I collagen or bovine serum albumin as described (8). Stock solution of purified CEA from liver metastases (Sigma) was diluted with phosphatebuffered saline (PBS) composed of NaCl at 10 mg/ml, KCI at 250 gg/ml, Na2HPO4 at 1.43 mg/ml, and KH2PO4 at 250 ,tg/ml (pH 7.3) and added to the previously coated plates at 20, 2, and 0.2 ,ug/ml for 3 hr at room temperature. The SW1222 and LS174T human colon carcinoma cells were also included in the cell attachment assay (105 cells per well) as positive and negative controls, respectively. After two washes with PBS, the unoccupied sites were blocked by adding PBS containing 0.2% gelatin and 0.2% sodium azide for 1 hr at room temperature. Plates were then washed twice with PBS containing 0.2% casein (Oxoid). Fifty microliters of PR3B10 mAb was added in duplicate for 1 hr, and the binding of antibody to collagen-bound CEA was detected by a GAG-ELISA (15). Immunoprecipitation. Purified human CEA (Sigma) was labeled with 1251I (Amersham) by using iodogen (Pierce) (19) and was then precipitated with the following mAbs reacting with members of the CEA family: PR3B10 (13), 601, C46, and Oxoid anti-CEA. The DH12 mouse mAb recognizing the P, chain (20), shared by the integrin ,31 subfamily of cell surface heterodimeric extracellular matrix receptors (21), was used as negative control. The immunoprecipitated fractions were then analyzed on a NaDodSO4/7.5% polyacrylamide gel (22).
RESULTS Screening of mAbs Recognizing Epithelial Cell Surface Determinants in a Cell Attachment Assay. A series of mAbs recognizing epithelial cell surface determinants was analyzed for the ability to inhibit the binding of SW1222 human colon carcinoma cell line to type I and to type IV collagen by a cell attachment assay. Only PR3B10 mAb specifically inhibited (50%) the binding of SW1222 cells to type I collagen (Fig. 1A). The binding to type IV collagen was not altered by any mAb tested (Fig. 1B). Reactivity of PR3B10 mAb with Human CEA. To define the specificity of PR3B10 mAb and its reactivity with CEA, 50 ,l of PR3B10 tissue culture supernatant was added to 96-well microtiter plates previously coated with serial concentrations of purified human CEA (Sigma). The binding of PR3B10 to CEA was then determined by using the GAG-ELISA (15). Three monoclonal antibodies reacting with CEA (601, C46, and Oxoid) were included in the assay as positive controls. W6/32, a mAb recognizing the major histocompatibility antigens A, B, and C associated with the light chain /82-
Proc. Natl. Acad. Sci. USA 87 (1990) PR 3B10 PR 5C5 PR 4D6 PR 6B5 MP 6D6 MP 3D9 MP 6F3 BBMI MP 4F10 CONTROL C2 2
ID
4C
'
PR 3B10 PR 5C5
PR 4D6 PR 6B5
_
MP 6D6
_
MP 3D9
MP 6F3
BBM1 MP 4F10
CONTROL 0
20
40
62
Arbitrary fluorescence units FIG. 1. Screening of a panel of mAbs (PR3B10, PRC5, PR4D6, PR6B5, MP6D6, MP3D9, MP6F3, MP4F10, and BBM1) recognizing
epithelial cell surface determinants for their ability to inhibit the binding of SW1222 colon carcinoma cells to type I (A) and type IV (B) collagen. * , Specific binding to collagens (20 gg/ml); , binding to bovine serum albumin (20 ,ug/ml). Cells (5 x 104) were resuspended with 50 ,I of each mAb (culture supernatant concentrated 10 times and diluted 1:1 with DMEM containing 2.5 mg of bovine serum albumin per ml and 25 mM Hepes, pH 7.4) and were allowed to attach to the coated plates for 2 hr at room temperature. The unattached cells were washed away with PBS, and the number of attached cells was estimated by measuring alkaline phosphatase (8). Values are expressed in arbitrary fluorescence units (mean ± SD of three determinations).
microglobulin (23), was used as negative control. PR3B10 mAb showed a specific"" reactivity with human CEA comparable to the other three anti-CEA mAbs (Fig. 2A). In the second experiment, human CEA was 125I-labeled by the iodogen method and was immunoprecipitated with PR3B10 and the three other anti-CEA mAbs (601, C46, and Oxoid). The immunoprecipitated fractions were then examined by NaDodSO4/7.5% PAGE. PR3B10 and the other three anti-CEA mAbs immunoprecipitated a band with an approximate molecular weight of 180 kDa, consistent with the published size of CEA (24). Inhibition of SW1222 Collagen Binding by Other mAbs Reacting with CEA and CEA-Related Glycoproteins. To evaluate whether the inhibitory activity of PR3B10 mAb on SW1222 cells was specifically due to its CEA activity, three other mAbs (601, C46, and Oxoid) recognizing CEA and other members of the CEA family were also tested in a cell
Cell
Biology: Pignatelli et al.
Proc. Natl. Acad. Sci. USA 87 (1990) 192
3 A
5
1543
B
200U)
cn .
0 a)
92.5-
a) CD 0
U) 0 cys
.,
67-
4610
1
CEA concentration,jug/ml FIG. 2. (A) Reactivity of PR3B10 (m), 601 (n), C46 (o), and Oxoid (x) mAbs with human CEA. mAbs were added to 96-well microtiter plates previously coated with serial concentrations of human purified CEA (0.2-20 Ag/ml) for 3 hr at room temperature, and their binding to CEA was then determined by GAG-ELISA (15). W6/32 mAb (A), which recognizes the major histocompatibility antigens A, B, and C associated with the light chain 832-microglobulin (23), was used as negative control. (B) Immunoprecipitation of 251I-labeled human CEA with DH12 (lane 1), PR3B10 (lane 2), Oxoid (lane 3), 601 (lane 4), and C46 (lane 5) mAbs. The DH12 mAb, which reacts with the ,81 integrin chain (20), was used as negative control.
attachment assay. The MP6F3 mAb, which also recognizes an epitope expressed by SW1222 cells (M.P. and W.F.B., unpublished data) and has been shown not to be involved in the collagen binding in our preliminary screening (Fig. 1A), was used as negative control. All three anti-CEA mAbs inhibited the binding of SW1222 colon carcinoma cells to type I collagen. They showed inhibition of cell adhesion ranging from 72% (mAb 601) to 87% (mAb C46). These results were comparable to the inhibitory effect of mAb PR3B10 (78%) SRGETG
SRGDTG I3B10
OXOID
DISCUSSION
60Q
C46
'SI P6F3 0
20
40
60
80
1 00
1 20
% cell attachment FIG. 3. Inhibition of the binding of SW1222 cells to collagencoated plates by anti-CEA mAbs and the Arg-Gly-Asp-containing peptide. Cells (5 x 104) were allowed to attach for 2 hr at room temperature in the presence of PR3B10 at a concentration 10 times that in the culture supernatant, Oxoid at 100 Ag/ml, 601 at 100 ,ug/ml, C46 at 100 ,Ag/ml, and MP6F3 at 10 times the culture supernatant concentration mAbs and the Ser-Arg-Gly-Asp-Thr-Gly (SRGDTG in single-letter code) and Ser-Arg-Gly-Glu-Thr-Gly (SRGETG) synthetic peptides at 2 mg/ml. The unattached cells were washed away with PBS, and the number of attached cells was estimated by measuring alkaline phosphatase activity (8). Values are expressed as a percentage of the maximal binding (mean ± SD of three determi-
nations).
and an Arg-Gly-Asp-containing peptide (75%) (Fig. 3) as previously shown (8). Type I Collagen Binding Activity of CEA. A cell attachment assay was used to test whether CEA had the ability to bind specifically type I collagen matrix. Purified human CEA was applied to type I collagen plates, and the CEA molecules bound were detected by using PR3B10 mAb. The antibody binding to collagen-bound CEA was then detected by using the GAG-ELISA (15). No binding of CEA to type I collagen was found (Fig. 4A), suggesting that CEA is not a collagenbinding protein itself. Since both Arg-Gly-Asp-containing peptides and the anti-CEA mAbs block SW1222 collagen binding, it is likely that CEA (or CEA-related antigens) is a subunit of the SW1222 collagen receptor. LS174T human colon carcinoma cells, which are known to express the 180-kDa CEA glycoprotein and other members of the CEA family (13, 25) as do SW1222 cells (Fig. 4B), did not bind type I collagen-coated plates (Fig. 4A), supporting our hypothesis that CEA does not have the ability to bind collagen by itself.
In this study we have shown that several mAbs recognizing the 180-kDa CEA glycoprotein and other members of the CEA family were able to inhibit the binding of a CEAexpressing human colon carcinoma cell line (SW1222) to type I collagen matrix. Using a cell attachment assay, we failed to detect any direct collagen binding of either purified human CEA or another human colon carcinoma cell line (LS174T) that expresses high levels of CEA and CEA-related antigens (25). These data suggest that CEA is not a collagen-binding protein itself but is likely to be associated with a functional collagen receptor complex expressed by SW1222 cells. Therefore, we propose that CEA may function as an accessory molecule, controlling the functional activity of the SW1222 collagen receptor. We have shown (8) that the collagen receptor complex expressed by SW1222 cells mediates the cellular attachment to collagen and subsequently the glandular differentiation seen in three-dimensional collagen gel (8). This receptor recognizes the Arg-Gly-Asp tripeptide sequence present in collagen and requires the presence of divalent cations, such
1544
Cell Biology: Pignatelli et al.
'W
TYPE COLL
Proc. Natl. Acad. Sci. USA 87 (1990)
BSA
0
4000 D~~~~~~~~~ 3000
2000
1000
0
PBS
MP 6F3
PR
3B10
601
OXOID
FIG. 4. (A) Type I collagen binding activity of CEA (in), bovine albumin (BSA) (0), and SW1222 (ml) and LS174T (u) colon carcinoma cells. Purified human CEA (20 jug/ml) was applied to type I collagen- or bovine serum albumin-coated plates (20 jug/ml) for 3 hr at room temperature. SW1222 and LS174T cells (105 cells per well) were included as positive and negative controls, respectively. The CEA molecules bound were detected by using PR3B10 mAb in GAG-ELISA (15). (B) Expression of MP6F3, PR3B1O, 601, and Oxoid mAbs on SW1222 ( ) and LS174T ( *) colon carcinoma cells serum
by GAG-ELISA (15).
Ca2+ or Mg2+ or both, for its functional activity (26). Several extracellular matrix receptors have been now characterized by affinity chromatography and by inhibition of cell adhesion to extracellular matrix proteins by specific mAbs. Some of them belong to the integrin family of cell surface receptors, which are a,8-heterodimeric transmembrane proteins divided into three subfamilies (,1, /32, and 33), based on the sharing of a common p chain (21). The Arg-Gly-Asp tripeptide sequence constitutes the cell binding site of some extracellular matrix receptors that belong to the 81 (27, 28) and integrin subfamilies (29). The majority of integrin receptors also require divalent cations for their binding (9, 21). However, some reports have also shown that some of the Arg-Gly-Asp-directed extracellular matrix receptors do not share structural and functional similarities with the integrin as
33
family (30, 31). The inhibition of the SW1222 collagen binding by anti-CEA mAbs suggests that CEA may control the function and/or the affinity of the Arg-Gly-Asp-directed collagen receptor on SW1222 cells. It is likely that the Arg-Gly-Asp-directed collagen receptor provides the primary specificity of attach-
ment of SW1222 cells to collagen, whereas CEA may exert a modulatory influence on the cell-collagen interaction. This hypothesis is supported by the fact that we were unable to show any direct binding of human CEA to collagen-coated plates. In addition, a human colon carcinoma cell line (LS174T), which expressed high levels of the 180-kDa CEA glycoprotein and other members of the CEA family, did not specifically bind collagen. Cheresh et al. (32, 33) have shown that cell surface gangliosides, which are sialic acid-containing glycolipids, can similarly modulate the function of some of the Arg-Gly-Asp-directed extracellular matrix receptors. They showed that mAbs directed specifically to the carbohydrate moiety of the digangliosides GD3 and GD2 inhibited melanoma and neuroblastoma cell attachment to collagen, fibronectin, laminin, and vitronectin (32). As we show here for CEA, the role of gangliosides in cell attachment also does not involve their direct binding with the extracellular matrix proteins, but they seem to be part of the divalent cationdependent functional complex involving the Arg-GlyAsp-directed receptor (33). The accessory function of several cell surface molecules that are involved in immune recognition has been described (34-36). The functional interaction between T-cell antigen receptor-CD3 antigen complex and the CD2 molecule, which are also members of the immunoglobulin gene family, has been shown to be due to a physical association of these T-lymphocyte cell surface structures (37). CEA may function similarly as an associated cell surface molecule that interacts with the collagen receptor expressed by colon epithelial cells. Whether the 180-kDa CEA glycoprotein itself or other members of the CEA family are involved in cell-collagen interaction is unclear. The 601 mAb was the only mAb with no known cross-reactivity with other CEA-related glycoproteins (nonspecific cross-reactive antigen and biliary glycoprotein 1) (38, 39) according to the manufacturer. This mAb showed a similar inhibitory activity on the SW1222 collagen binding, suggesting that CEA itself is the cell adhesion molecule. However, further functional studies using mAbs specific for other members of the CEA family will be required to answer this issue fully. Recently Benchimol et al. (7) have shown that CEA can affect the homotypic sorting of cells in heterogeneous populations of aggregating cells by functioning as a Ca2+_ independent homotypic intercellular adhesion molecule. Since CEA is highly expressed basolaterally during gut embryogenesis, they proposed that CEA could play a role as an intercellular adhesion molecule in intestinal organization (7). Here we show that CEA also can function as an accessory cell adhesion molecule, mediating cell-matrix interaction, which is also a fundamental process necessary for the development of a polarized epithelial cell phenotype (40). CEA shows, therefore, a strong structural and functional similarity with the neural cell adhesion molecule, which is also a member of the immunoglobulin gene family (6) and is involved in cell-matrix and cell-cell interaction among neurons, glia, and muscle cells during embryogenesis (41). The role played by CEA in colon carcinogenesis is rather unclear. In colorectal tumors, CEA is normally found expressed on the entire surface of the colon epithelial cells, showing, however, no relationship with the pathological extent of the disease, clinical prognosis, or the degree of tumor differentiation (42, 43). We suggest, in fact, that the detection of high levels of CEA in colorectal cancers may be only an epiphenomenon resulting from the disruption of cell-matrix interactions because of the loss of the CEA-associated extracellular matrix receptors by tumor cells. We have recently shown that loss of two integrin collagen receptors (very late antigen 2 and very late antigen 3) occurs relatively frequently in colorectal cancer (44), and this allows tumor cells to escape the controlling mechanisms on cell differentiation induced by
Cell
Proc. Nati. Acad. Sci. USA 87 (1990)
Biology: Pignatelli et al.
collagen matrix and those soluble factors that enhance differentiation and inhibit proliferation through effects on collagen binding (26). We are grateful to Dr. B. De Strooper for the generous gift of DH12 mAb and to Dr. M. J. Crumpton for helpful comments. 1. Gold, P. & Freedman, S. 0. (1965) J. Exp. Med. 121, 439-462. 2. Rogers, G. T. (1983) Biochim. Biophys. Acta 695, 227-249. 3. Zimmermann, W., Ortlieb, B., Friedrich, R. & von Kleist, S. (1987) Proc. Nati. Acad. Sci. USA 84, 2960-2964. 4. Thompson, J. A., Pande, H., Paxton, R. J., Shively, L., Padma, A., Simmer, R. L., Todd, C. W., Riggs, A. D. & Shively, J. E. (1987) Proc. Natl. Acad. Sci. USA 84, 29652969. 5. Barnett, T. R., Kretschmer, A., Douglas, A. A., Goebel, S. J., Hart, J. T., Elting, J. J. & Kamarck, M. E. (1989) J. Cell Biol. 108, 267-276. 6. Williams, A. F. & Barclay, A. N. (1988) Annu. Rev. Immunol. 6, 381-405. 7. Benchimol, S., Fuks, A., Jothy, S., Beauchemin, N., Shirota, K. & Stanners, C. P. (1989) Cell 57, 327-334. 8. Pignatelli, M. & Bodmer, W. F. (1988) Proc. Natd. Acad. Sci. USA 85, 5561-5565. 9. Ruoslahti, E. & Pierschbacher, M. D. (1987) Science 238, 491-497. 10. Richman, P. I. & Bodmer, W. F. (1988) J. Pathol. 156, 197211. 11. Leibovitz, A., Stinson, J. C., McCombs, W. B., III, McCoy, C. E., Mazur, K. C. & Mabry, N. D. (1976) Cancer Res. 36, 4562-4569. 12. Tom, B. H., Rutzky, L. P., Jakstys, M. M., Oyasu, R., Kaye, G. I. & Kahan, B. D. (1976) In Vitro 12, 180-191. 13. Richman, P. I. & Bodmer, W. F. (1987) Int. J. Cancer 39, 317-328. 14. Galfrd, G., Howe, S. C., Milstein, C., Butcher, G. W. & Howard, J. C. (1977) Nature (London) 266, 550-552. 15. Durbin, H. & Bodmer, W. F. (1987) J. Immunol. Methods 97, 19-27. 16. Barany, G. & Merrifield, R. (1979) in The Peptides, eds. Gross, E. & Meienhofer, J. (Academic, New York), pp. 1-284. 17. Rothbard, J., Lechler, R., Howland, K., Bal, V., Eckels, D., Sekaly, R., Long, E., Taylor, W. & Lamb, J. (1988) Cell 52,
515-523. 18. Brodsky, F. M., Bodmer, W. F. & Parham, P. (1979) Eur. J. Immunol. 9, 536-545. 19. Fraker, P. J. & Speck, J. C. (1978) Biochem. Biophys. Res. Commun. 80, 849-857. 20. Jaspers, M., De Strooper, B., Spaepen, M., van Leuven, F.,
21. 22. 23. 24. 25. 26. 27.
28. 29. 30. 31.
32.
1545
David, G., van den Berghe, H. & Cassiman, J. J. (1988) FEBS Lett. 231, 402-406. Hynes, R. 0. (1987) Cell 48, 549-554. Laemmli, U. K. (1970) Nature (London) 227, 680-685. Barnstable, C. J., Bodmer, W. F., Brown, G., Galfre, G., Milstein, C., Williams, A. F. & Ziegler, A. (1978) Cell 14,9-20. Slayter, H. S. & Coligan, J. E. (1975) Biochemistry 14, 23232330. Toribara, N. W., Sack, T. L., Gum, J. R., Ho, S. B., Shively, J. E., Willson, J. K. V. & Kim, Y. S. (1989) Cancer Res. 49, 3321-3327. Pignatelli, M. & Bodmer, W. F. (1989) Int. J. Cancer 44, 518-523. Pierschbacher, M. D. & Ruoslahti, E. (1984) Proc. Natl. Acad. Sci. USA 81, 5985-5988. Bourdon, M. A. & Ruoslahti, E. (1989) J. Cell Biol. 108, 1149-1155. Pytela, R., Pierschbacher, M. D. & Ruoslahti, E. (1985) Proc. Natl. Acad. Sci. USA 82, 5766-5770. Dedhar, S., Ruoslahti, E. & Pierschbacher, M. (1987) J. Cell Biol. 104, 585-593. Lu, M. L., Becham, D. A. & Jacobson, B. C. (1989) J. Biol. Chem. 264, 13546-13558. Cheresh, D. A., Pierschbacher, M. D., Herzig, M. A. & Mu-
joo, K. (1986) J. Cell Biol. 102, 688-696. 33. Cheresh, D. A., Pytela, R., Pierschbacher, M. D., Klier, F. G., Ruoslahti, E. & Reisfeld, R. A. (1987) J. Cell Biol. 105, 1163-1173. 34. Alcover, A., Alberini, C., Acuto, O., Clayton, L. K., Transy, C., Spagnoli, G. C., Moingeon, P., Lopez, P. & Reinherz, E. L. (1988) EMBO J. 7, 1973-1977. 35. Bockenstedt, L. K., Goldsmith, M. A., Dustin, M., Olive, D., Springer, T. A. & Weiss, A. (1988) J. Immunol. 141, 19041911. 36. Altmann, D. M., Hogg, N., Trowsdale, J. & Wilkinson, D. (1989) Nature (London) 338, 512-514. 37. Brown, M. H., Cantrell, D. A., Brattsand, G., Crumpton, M. J. & Gullberg, M. (1989) Nature (London) 339, 551-553. 38. von Kliest, S., Chavanel, G. & Burtin, P. (1972) Proc. Natl. Acad. Sci. USA 69, 2492-2494. 39. Svenberg, T. (1976) Int. J. Cancer 17, 588-596. 40. Boulan, E. R. & Nelson, W. J. (1989) Science 245, 718-725. 41. Rutishauser, U. (1984) Nature (London) 310, 594-597. 42. Ahnen, D. J., Kinoshita, K., Nakane, P. K. & Brown, W. R. (1987) Gastroenterology 93, 1330-1338. 43. Shirota, K., Minassian, H. & Jothy, S. (1988) Exp. Mol. Pathol. 49, 305-315. 44. Pignatelli, M., Smith, M. E. F. & Bodmer, W. F. (1990) Br. J. Cancer, in press.
