Human Tissues and Tumors - NCBI

3 downloads 422 Views 3MB Size Report
Jefferson Medical College,* ThomasJefferson University, and the ... California, San Francisco, California* .... For photography, slideswere re- mounted in 90% ...
American Journal ofPathology, Vol. 135, No. 1, July 1989 Copyright © American Association ofPathologists

Expression of the Cell-Cell Adhesion Glycoprotein Cell-CAM 120/80 in Normal Human Tissues and Tumors

Shmuel Eidelman,* Caroline H. Damsky,t Margaret J. Wheelock,t and Ivan Damjanov* From the Department ofPathology and Cell Biology,

Jefferson Medical College,* ThomasJefferson University, and the WistarInstitute,t Philadelphia, Pennsylvania, andfrom the Departments ofStomatology andAnatomy, Schools ofDentistry and Medicine, University of California, San Francisco, California*

Polyclonal and monoclonal antibodies raised to the 80 kdglycoprotein component of the cell to cell adhesion molecule cell-CAM 120/80 were used to map its distribution immunohistochemically in normal human tissues and in benign and malignant tumors. Cell- CAM 120/80 was found in all normal epithelial tissues, but was not expressed on neural, lymphoid, smooth, striated and cardiac muscle, connective tissue, or thegerm cells in either sex. The expression of this adhesion molecule was polarized in ductal and glandular epithelia and evenly circumferential in squamous and transitional epithelia. Some organs, such as the kidney, liver and endocrine glands, showed unique organ to tissue specific patterns. Maturation-dependent loss of cell-CAM 120/80 was noticed in superficial layers of squamous epithelium and the placenta. Benign epithelial tumors expressed cell-CAM 120/ 80 in a manner comparable with their tissue of origin. Malignant tumors expressed cell-CAM 120/80 either in a manner similar to the tissue of their origin or assumed a less polarizedphenotype. Overall, the immunoreactivity in many malignant tumors appeared weaker and the polarization was less pronounced. Thus, cell-CAM 120/80 is a universal marker of human epithelial cells, but its mode of expression differs in various anatomic sites, and may be influenced by maturation or malignant transformation of cells. (Am J Pathol 1989, 135:

101-110) Cell adhesion molecules (CAMs) have been shown to play an important role in embryonic development and

morphogenesis, as well as in the maintenance of the normal structure and function of adult tissues.1' 2 One of these molecules, known as cell-CAM 120/80,3 uvomorulin,4 cadherin,5 or Arc-1,6 has been studied in our laboratories and its function and distribution in mouse embryonic development have been documented.7-9 Cell-CAM 120/80 has also been demonstrated in adult murine10 and some human tissues7 but systematic studies on the distribution of this adhesion molecule in adult mammalian tissues have not yet been performed. Likewise, there are no comprehensive data on the expression of cell-CAM 120/80 in various benign and malignant tumors, although it has been shown that it is present in some tumor cell lines and not in others.7 In this article we report results on the immunohistochemical localization of cell-CAM 120/80 in normal adult human tissues and tumors. The studies were performed with polyclonal and monoclonal antibodies prepared against the 80 kd glycoprotein fragment of this molecule derived from a human breast carcinoma cell line.3'11 We found that cell CAM 120/80 is expressed in essentially all epithelial tissues except syncytotrophoblast of the placenta. Several patterns of cell CAM 120/80 expression were identified indicative of quantitative and qualitative difference in the expression of this molecule in various tissues. We also found that most, if not all, benign and malignant epithelial tumors express cell-CAM 120/80 like the tissues of their origin, although the cell surface distribution pattern may be altered especially in anaplastic malignant tumors.

Materials and Methods Normal human tissues and various tumors were obtained at the time of surgery. Small tissue specimens were covSupported in part by Public Health Service Grants AA07-186 and HD21355, and by grants from the W. W. Smith Charitable Trust and The Comprehensive Rectal Cancer Center of Thomas Jefferson University, Philadelphia, Pennsylvania. Accepted for publication March 30, 1989. Address reprnt requests to 1. Damjanov, Department of Pathology, Jefferson Medical College, Philadelphia, PA 19107.

101

102 Eidelman et al AJPJuly 1989, Vol. 135, No. 1

Table 1. Cell-CAM 120/80 in Normal Human Tissues Pattern of expression Polarized

Nonpolarized

Other

Trachea and Kidney Skin Vagina and exocervix bronchus Liver Tongue, pharynx, Gastrointestinal tract Pancreatic islets esophagus (stomach, small Lung Adrenal and large Urinary bladder and intestine) Parathyroid ureter Renal pelvis Exocrine pancreas Placenta and pancreatic Mesothelium ducts Gallbladder and bile ducts Prostate and epididymis Fallopian tube and endometrium Breast Thyroid

ered with OCT (Lab Tek, Naperville, IL) and snap frozen in methyl-butane precooled in liquid nitrogen. Frozen blocks were sectioned on a cryostat at 5 ,um. Sections were fixed in cold absolute methanol for 10 minutes at 4 C. After fixation the specimens were washed in phosphatebuffered saline (PBS) and incubated with protein A-purified IgG of rabbit antiserum raised against the 80 kd glycoprotein fragment of cell-CAM 120/80 (anti GP-80).3 The primary antibody was used at a dilution of 1:200 for 1 hour at room temperature or overnight at 4 C. A rat IgG monoclonal antibody recognizing the same glycoprotein11 was also used by diluting the ascites 1:100 or 1: 10,000. Optimal staining was obtained with the antibody diluted to a final concentration of 2 to 3 gg protein per ml (1 :1000). The specificity of both antibodies was confirmed by Western blotting human placental extracts electrophoresed in 7% polyacrilamide gel and electroblotted to nitroTable 2. Normal Human Organs and Tissues in which Cell-CAM 120/80 Could Not be Demonstrated Central nervous system Peripheral nerves Muscle: Striated Heart

Smooth Vascular endothelium Fibrous connective tissue Adipose tissue Lymph nodes

Spleen Peripheral blood cells

Ovary Seminiferous tubules of testis Placental syncytiotrophoblast Umbilical cord Decidua Gastric parietal cells Visceral layer of Bowman capsule

cellulose paper. The polyclonal antibody diluted to 50 jig/ ml gave two major bands of 120 kd and 80 kd. The monoclonal antibody could be diluted to 10 ng/ml, and it consistently reacted with a single band of 120 kd as reported previously.1 After three washes in PBS, the specimens were incubated with fluorescein isothiocyanate (FITC): conjugated goat-anti-rabbit, IgG, or FITC rabbit-anti-rat IgG (Cappel, Malvern, PA) diluted 1:100. The secondary antibodies were absorbed on methanol permeabilized cultured mouse mammary tumor cells to remove nonspecific reactivity. After the incubations, the slides were rinsed in PBS, mounted in 10% glycerol, and examined with a Zeiss or Olympus microscope equipped for phase contrast and UV-epi-illumination. For photography, slides were remounted in 90% glycerol containing 0.1% p-phenylenediamine to prevent fading of fluorescence. The control sections were incubated with PBS or preimmune rabbit serum followed by the secondary antibody.

Results Normal Tissues At least three samples of each tissue were studied. Essentially identical results were obtained with the polyclonal and the monoclonal antibody to cell-CAM 120/80. The findings are summarized in Tables 1 and 2. Cell-CAM 120/80 was immunohistochemically demonstrated in all epithelial cells containing tissues (Table 1). No cell-CAM 120/80 was seen in the seminiferous tubules of the testes or in the Graafian follicle cells in the ovary. Connective tissues, lymphoid organs, the cardiovascular system, and the central and peripheral nervous system were all devoid of this adhesion molecule (Table 2). No immunoreactive cell-CAM 120/80 could be demonstrated around the parietal cells in the stomach mucosa. Two distinct patterns of cell-CAM 120/80 expression were noted: nonpolarized and polarized. However, several immunoreactive tissues did not express cell-CAM 120/80 in either of these two patterns and thus they will be described under a separate heading. Cells that were evenly immunoreactive along their entire circumference were considered to express the cell-CAM 120/80 in a nonpolarized manner. Polarized expression of cell-CAM 120/80 was, on the other hand, characterized by the accentuated immunoreactivity on peripheral intercellular borders or immunoreactivity limited to parts of the intercellular contact areas.

Cell-Cell Adhesion Glycoprotein 103 AJPJuly 1989, Vol. 135, No. 1

Figure 1. Circumferential nonpolarized distribution of cell-CAM 120/80 along the cell membrane of epidermal cells. A: Note the lack of immunoreactivity in the superficial stratum corneum. B: The arrow marks the lack of immunoreactivity along the cell surface in contact with the basal membrane (immunofluorescence microscopy, X320).

Nonpolarized Expression This pattern was characterized by uniform expression of cell-CAM 120/80 along the entire circumference of each cell without any signs of polarity, accentuation of parts of the cell membrane, or aggregation of immunoreactive material at any particular portion of the cell surface (Figure 1). Nonpolarized expression of cell-CAM 120/80 was seen in the squamous epithelium of the skin, upper gastrointestinal and respiratory tracts, the lower part of the female genital tract (vagina and cervix), and in the transitional epithelium of the urinary tract. In these tissues most cells showed circumferential immunoreactivity of their plasma membrane, although the basal layer of these epithelia did not express cell-CAM 120/80 along the surface that was in contact with the underlying basal membrane. The superficial keratinized layer of the skin and tongue were devoid of cell-CAM 120/80. The transition from the immunoreactive-to-nonreactive superficial cells occurred gradually in the upper third of the epithelium but varied from one anatomic site to another.

Polarized Expression This pattern was seen as lateral cell border reactivity in simple cuboidal and columnar epithelia of the stomach,

small and large bowel, gallbladder, extrahepatic and intrahepatic bile ducts, pancreatic acini and ducts, endometrial and endocervical glands, fallopian tube, prostate, epididymis, thyroid follicles, and mammary glands and ducts (Figure 2). In these epithelia, cell-CAM 120/80 was found along the entire lateral border, but appeared accentuated on the lateral sides of the apical pole at the site corresponding to intermediate junctions.12 The basal portions of the cell membranes were devoid of cell-CAM 120/80.

Other Patterns On pancreatic islet cells the cell-CAM 120/80 was seen as uniform, faint dot-like or short, linear fluorescence along the lateral border of juxtaposed islet cells (Figure 3). Similar lateral cell reactivity of lower intensity was seen between hepatocytes. The free border of hepatocytes facing the Disse's space was unreactive (Figure 4). Similar intercellular polarization of cell-CAM 120/80 was seen along the interparenchymal cell junctions of pulmonary mesothelium and alveolar cells. The cells in the adrenals cortex and parathyroid glands also showed limited cell surface reaction along the lateral cell membranes. Adrenal medulla was unreactive.

104 Eidelman et al AJPJuly 1989, Vol. 135, No. I

Figure 2. A: Polarized reactivit in the gastric glands. B: Polarized staining ofthe gallbladder epithelium. C: Colonic crypts, sectioned transversely, showedperiluminal reactivity with the antibody to the cell-CAM 120/80 (X280).

iA_ Figure 3. A: Pancreatic acina cell show strong reactivity along the lateral cell membranes, which is even more pronounced along the apical border. B: Ducts in the right side of the photograph show strong polarized reactivity whereas the islet cells show dot-like reaction at the intercellular contact points (X480).

Cell-Cell Adhesion Glycoprotein 105 AJPJuly 1989, Vol. 135, No. 1

In the kidney, the parietal epithelial cells lining the Bowman's capsule showed circumferential immunoreactivity with the antibody to cell-CAM 120/80, whereas the glomerular tufts were unreactive (Figure 5). The proximal renal tubules expressed cell-CAM 120/80 along the entire cell circumference except for the apical luminal surface. The unique curvilinear fluorescence of proximal tubular cell membrane was on a weaker order of magnitude than the fluorescence of the Bowman's capsule cells suggesting quantitative differences in the amount of cell-CAM 120/80 in these two structures. The loops of Henle were weakly reactive in contrast to distal tubules and the collecting ducts that expressed cell-CAM 120/80 strongly along the basilateral plasma membrane. The fluorescence along the plasma membrane of distal and collecting tubules was strong, ie, of the same intensity as the fluorescence of Bowman's capsular cells. In the first trimester and term placenta the cytotrophoblastic cells expressed cell-CAM 120/80 circumferentially along the entire surface, whereas the syncytiotrophoblastic cells were unreactive with either the polyclonal or monoclonal antibody (Figure 6). Other components of the placenta, including the umbilical cord, were unreactive. Maternal decidua was also negative.

.

7

-

Tumors More than 60 human tumors were examined, the findings of which are summarized in Tables 3 and 4. Overall, the distribution of cell-CAM 120/80 in tumors was similar to the expression of this molecule in corresponding normal tissues. All epithelial tumors expressed cell-CAM 120/80, whereas all mesenchymal tumors were negative. Polarity of cells with periluminal accentuation of the lateral cell membrane fluorescence was maintained in all benign tumors, as well as in differentiated neoplasms arising from

_ NA

3.

Figure 5. A: Low power magnification of the kidney shows in different portions different intensities ofanimmunoreactivity of the nephron. Note unreactive glomerulus and immunoreactive ducts corresponding to distal tubules and collecting ducts (X 160). B: Higherpower view of the glomerulus showing the immunoreactivity of the Bowman 's capsular epithelium (arrow) and the strong reactivity of the distal convoluted tubules (X280). C: The weak curvilinear immunoreactivity of lateral cell borders of theproximal renal tubules are indicated by the arrow 0X320). Figure 4. The liver cells show weak reactivity along the-lateral cell borders. The dots indicate the perisinusoidal surfaces of hepatocytes that are unreactive. Arrow points out the polarized stronger staining of the bile ducts (X320).

cells that under normal conditions expressed cell-CAM

120/80 in a molarized manner. Thus, all adenomas and well-differentiated adenocarcinomas of the gastrointesti-

106 Eidelman et al AJPJuly 1989, Vol. 135, No.

1

Table 3. Human Tumors Expressing Cell-CAM 120/80 Polarized Nonpolarized Tumor (no.)

-

t.~~~~~~~~~~~~~~~qv

Figure 6. Midterm placental villi. The cytotrophoblastic cells show circumferential immunoreactivity, whereas the syncytiotrophoblastic cells (arrows) are unreactive (X280).

nal tract, as well as those arising in other sites such as the breast, ovary, or lung, presented with polarized expression of cell-CAM 120/80 (Figure 7). In contrast, poorly differentiated tumors were composed of cells displaying no polarity with cell-CAM 120/80 evenly distributed along the entire cell circumference. Adenocarcinomas labeled as nonpolarized formed clearly visible glands showing back-to-back arrangement and glands within glands, but were also composed of solid tumor cell nests. Squamous cell carcinomas, papillary carcinomas of the thyroid, the epithelial portion of a mixed tumor of the parotid gland, liver cell carcinomas, islet cell carcinomas, renal cell carcinomas, and the embryonal carcinomas of the testis did not show distinct polarized expression of cell-CAM 120/ 80 (Figure 8). Because many of these tumors contained cells arranged into distinct histologic structures such as cords, papillae, or gland-like formatious, nonpolarized expression of cell-CAM 120/80 was not exclusively limited to undifferentiated carcinomas. Although we have not performed systematic measurements to compare the amount of cell-CAM 120/80 in normal tissues with that in malignant tumors, our overall impression, based on subjective estimates of fluorescence, was that the tumors reacted more weakly than the corresponding normal tissues. However, completely unreactive epithelial neoplasms were not identified, and the focal areas in some tumors that appeared unreactive corre-

Tongue Squamous cell carcinoma (1) Thyroid Follicular adenoma (1) Follicular carcinoma (2) Papillary carcinoma (1) Breast Fibroadenoma (3) Infiltrating duct carcinoma (5) Lung Squamous cell carcinoma (2) Adenocarcinoma (3) Parotid Mixed tumor (1) Colon Villous adenoma (4) Tubular adenoma (5) Adenocarcinoma (5) Liver-hepatocellular carcinoma (3) Pancreas Adenocarcinoma (2) Islet cell tumor (1) Kidney-renal cell carcinoma (3) Bladder-transitional cell carcinoma (2) Ovary-endometroid carcinoma (2) Uterus-adenocarcinoma (2) Testis Embryonal carcinoma (2)

0

0

2

0

0

1

3 2 0

2

0

3 2 1

0

4 5 4 0

0

0

1 3

2 0 0

3

0

3

2

0

0

2

0

2

sponded to necrotic portions of these tumors. In addition to various benign and malignant tumors of mesenchymal origin, cell-CAM 120/80 negative malignancies included seminomas, pheochromocytomas, melanomas, and lymphomas.

Discussion Cell-CAM 120 is one of several cell-cell adhesion molecules that have been characterized during the past few years.2 Antibodies to human cell-CAM 120/80 react with the equivalent molecule in the mouse known as uvomorulin4 or E-cadherin.13 Molecular comparison of human and murine cell-CAM 120/80 (uvomorulin) has disclosed extensive homology of 82% for the nucleotide and 83% for Table 4. Human Tumors Devoid of Cell-CAM 120/80 Malignant fibrous histiocytoma (1) Leiomyoma (3) Leiomyosarcoma (1) Rhabdomyosarcoma (1) Seminoma (5) Pheochromocytoma (2) Melanoma (3) Lymphoma (2) Number of tumors examined is given in parentheses.

Cell-Cell Adhesion Glycoprotein 107 AJPJuly 1989, Vol. 135, No. I

the deduced amino acid sequence."4 Uvomorulin exhibits a 65% identity in an end-to-end alignment with the L-CAM, the equivalent avian molecule mediating cell-to-cell adhesion in chickens.15-17 All these data indicate that this calcium-dependent cell adhesion molecule is conserved and only slightly modified during evolution. In view of the conserved nature of cell-CAM 120/80, we expected that our immunohistochemical findings would correspond to those described for L-CAM in chickens, the only species in which this molecule has been systematically mapped in adults.18 Also, the expression of human cell-CAM 120/80 in adult tissues could be roughly predicted from previous studies on mouse embryos, placenta, and fetuses.9'19 Overall, these predictions were confirmed in the present study, and only some minor differences were noted that were probably related to anatomic variability between humans and other species and to the minor differences between human cell-CAM 120/ 80 and the equivalent molecules in mice and chickens. As in the chicken and mouse tissues, cell-CAM 120/80 was present only in epithelial tissues and was not expressed in mesenchymal or neural tissues. Although LCAM appeared along the entire circumference of chicken liver cells, human cell-CAM 120/80 was limited to hepatocyte-hepatocyte junctions. Human placenta differed considerably from chicken and mouse extraembryonic membranes, but due to interspecies differences,20 any comparison would be inappropriate. Chicken kidney did not show the peculiar pattern described in this article for the human kidney, and because the article by Nose and Takeichi19 does not contain a description of E-cadherin expression in the mouse kidney, it cannot be concluded whether the differences among the various parts of the human nephron are unique to humans or a feature of other mammalian species as well. Because the two previous articles18 19 do not contain descriptions of adult endocrine glands in mice and chickens, it was not possible to compare the present findings in humans with those in other species. In a previous study performed on mouse embryos,9 we showed that cell-CAM 120/80 is lost from the cell surface of epithelial cell derivatives on their delamination from the compact embryonic cell layer. In the present study, we found that the expression of cell-CAM 120/80 may be modified within the layers of a single epithelium. Thus, in the stratified squamous epithelium, the cell membrane of the basal layer contacting the basal membrane did not express the adhesion molecule; the cells in the suprabasal layers expressed it circumferentially; and, finally, the molecule was completely lost from the keratinized superficial layers. Similar differentiation- or maturation-related changes in the expression of the cell-CAM 120/80 were seen in placental epithelium, in which the cell CAM 120/

Figure 7. A: Polarized periluminal staining in a well differentiated adenocarcinoma of the colon (X 160). B: Higher power view ofa moderately differentiated adenocarcinoma of the colon showspolarizedperiluminal immunoreactivity (X320).

80 negative syncytiotrophoblastic cells presumably form through the fusion of circumferentially positive cytotrophoblastic cells. From the present data, it is not possible to determine whether this was due to the loss of cell surface molecules or to the downregulation of their synthesis. In addition to the differentiation-maturation related modulation of cell-CAM 120/80 expression, as seen in squamous epithelia and in the placenta, this molecule may be differentially distributed in anatomically distinct portions of a single organ. Thus, whereas intrahepatic bile duct express cell-CAM 120/80 in a polarized manner and

108 Eidelman et al AJP July 1989, Vol. 135, No. 1

Figure 8. A: Well-differentiated adenocarcinoma of the colon shows loss ofpolarization and strong immunoreactivity (X280). B: differentiated adenocarcinoma of Poorlj differentiated adenocarcinoma of the colon shows lack ofpolarization (X280). C: Poorly colon shows circumferential immunoreactivity and no polarization (X280). D: Carcinoma of the colon metastatic to the livershows no polarization of immunoreactivity (X480).

very strongly, the hepatocytes show a different pattern of cell surface polarization and a much lower intensity. Similarly, the initial portion of the nephron, the parietal layer of Bowman's capsule, shows strong reactivity with the antibodies to cell-CAM 120/80 contrasting with the low intensity of the staining in the proximal tubules. On the other hand, the terminal portions of the nephron express cell-CAM 120/80 as intensely as the epithelia of the Bowman's capsule. Similar anatomic variation, although not in the same sequence and pattern, has been reported for a calcium-independent molecule cell-CAM 105.21 Hence, the cellular requirements for adhesion molecules

apparently vary from one anatomic site to another and depend on many factors, all of which are essentially unknown. As with the normal tissue from which they originated, the benign and malignant tumors either expressed or were negative for the cell-CAM 120/80. Adenomas and carcinomas, ie, tumors of epithelial cells, were positive for cell-CAM 120/80, whereas the benign mesenchymal tumors, sarcomas, and lymphomas were negative. Melanomas, which express vimentin, a marker of mesenchymal tumors, rather than keratin, the marker of epithelial tumors, 22 were also devoid of cell-CAM 120/80. Pheochro-

Cell-Cell Adhesion Glycoprotein 109 AJPJuly 1989, Vol. 135, No. 1

mocytomas and seminomas, as expected originating from cell-CAM 120/80 negative cells in the medulla of the adrenal and the seminiferous tubules, were also cell-CAM 120/80 negative. Embryonal carcinoma, which is also of germ cell origin, expressed cell-CAM 120/80, in accordance with the epithelial nature of this neoplasm.23'24 Because malignant cells show reduced mutual adhesiveness25 and because some epithelial malignant epithelial tumor cell lines cultured in vitro do not express cellCAM 120/80,7 we hypothesized that this adhesion molecule may not be present on highly anaplastic carcinoma cells. However, despite an extensive survey of more than 60 human tumors, we could not identify a single case of epithelial malignancy lacking cell-CAM 120/80. The only difference between undifferentiated anaplastic carcinomas and well-differentiated or benign epithelial tumors was that the latter neoplasms often retained the polarity and showed accentuated accumulation of cell-CAM 120/ 80 on the apical lateral side, whereas the former expressed the cell adhesion molecule circumferentially and without obvious aggregation along the lateral sides. Similar uniform circumferential expression of L-CAM26 or cadherin E27 was noticed on transfection of mouse fibroblasts that do not normally express this adhesion molecule. Similarity between the transfected cells and the highly anaplastic carcinoma cells suggests that this adhesion molecule facilitates the adherence of cells but is probably not responsible for the maintenance of their polarization. In conclusion, cell-CAM 120/80 appears to be an almost universal human epithelial cell marker, one that can under specific maturation-defined conditions disappear from terminally differentiated cells, ie, cells from the squamous epithelium or trophoblast. The pattern of expression of this molecule is tissue specific, and this pattern of expression is retained in well-differentiated tumor cells originating from those tissues. The amount of cell-CAM 120/ 80 may be reduced in malignant cells but is not completely lost, a finding in keeping with the primordial and evolutionarily conserved nature of this adhesion molecule.

References 1. Edelman GM: Expression of cell adhesion molecules during embryogenesis and regeneration. Exp Cell Res 1984, 161: 1-16 2. Edelman GM: Cell adhesion molecules in the regulation of animal form and tissue pattern. Annu Rev Cell Biol 1986, 2:

81-96 3. Damsky CH, Richa J, Solter D, Knudsen K, Buck CA: Identification and purification of a cell surface glycoprotein mediat-

4.

5.

6. 7.

8.

9.

10.

ing intercellular adhesion in embryonic and adult tissue. Cell 1983,34:455-466 Hyafil F, Morello D, Babinet C, Jacob F: A cell surface glycoprotein involved in the compaction of embryonal carcinoma cells in cleavage stage embryos. Cell 1980, 21:927-934 Yoshida C, Takeichi M: Teratocarcinoma cell adhesion: Identification of a cell surface protein involved in calcium-dependent cell aggregation. Cell 1982, 28:217-224 Imhof BA, Vollmers HP, Goodman SL, Birchmeir W: Cell-cell interaction and polarity of epithelial cells: Specific perturbation using a monoclonal antibody. Cell 1983, 35:667-675 Damsky CH, Richa J, Wheelock M, Damjanov I, Buck CA: Characterization of cell-CAM 120/80 and the role of surface membrane adhesion glycoproteins in early events in mouse embryo morphogenesis, The Cell in Contact: Adhesions and Junctions as Morphogenetic Determinants. Edited by GM Edelman, J-P Thiery. New York, John Wiley & Sons, 1985, pp 233-255 Richa J, Damsky CH, Buck CA, Knowles BB, Solter D: Cell surface glycoproteins mediate compaction, trophoblast attachment, and endoderm formation during early mouse development. Dev Biol 1985,108:513-521 Damjanov I, Damjanov A, Damsky CH: Developmentally regulated expression of the cell-cell adhesion glycoprotein cellCAM 120/80 in peri-implantation mouse embryos and extraembryonic membranes. Dev Biol 1986,16:194-202 Vestweber D, Kemler R: Rabbit antiserum against a purified surface glycoprotein decompacts mouse preimplantation embryos and reacts with specific adult tissues. Exp Cell Res

1984,152:169-178 11. Wheelock MJ, Buck CA, Bechtol KB, Damsky CH: Soluble 80-kd fragment of cell-CAM 120/80 disrupts cell-cell adhesion J Cell Biochem 34:187-202 12. Boller K, Vestweber D, Kemler D: Cell adhesion molecule uvomorulin is localized in the intermediate junctions of adult intestinal epithelial cells. J Cell Biol 1985, 100:327-332 13. Yoshida-Noro C, Suzuki N, Takeichi M: Molecular nature of the calcium-dependent cell-cell adhesion system in mouse teratocarcinoma and embryonic cells studied with a monoclonal antibody. Dev Biol 1984,101:19-27 14. Mansouri A, Spurr N, Goodfellow PN, Kemler R: Characterization and chromosomal localization of the gene encoding the human cell adhesion molecule uvomorulin. Differentiation 1988, 38:67-71. 15. Schuh R, Vestweber D, Riede J, Rosenberg U, Jackle H, Kemler R: Molecular cloning of the mouse cell adhesion molecule uvomorulin: cDNA contains a Bi -related sequence. Proc Natl Acad Sci USA 1986, 83:1364-1368 16. Ringwald M, Schuh R, Vestweber D, Eistetter H, Lottspeich F, Engel J, Dolz R, Jahnig F, Epplen J, Mayer S, Muller C, Kemler R: The structure of cell adhesion molecule uvomorulin: Insights into the molecular mechanisms of Ca2+-dependent cell adhesion. EMBO J 1987, 6:3647-3652 17. Gallin WJ, Sorkin BC, Edelman GM, Cunningham BA: Sequence analysis of a cDNA clone encoding the liver cell adhesion molecule L-CAM. Proc Natl Acad Sci USA 1987, 84: 2808-2812

11 0 Eidelman et al AIPJul5y 1989, Vol. 135, No.

1

18. Thiery JP, Delouvee A, Gallin WJ, Cunningham BA, Edelman GM: Ontogenetic expression of cell adhesion molecules: LCAM is found in epithelia derived from the three primary germ layers. Dev Biol 1984,102:61-78 19. Nose A, Takeichi M: A novel cadherin cell adhesion molecule: Its expression patterns associated with implantation and organogenesis of mouse embryos. J Cell Biol 1986, 103:2649-2658 20. Mossman HW: Vertebrate Fetal Membranes: Comparative Ontogeny and Morphology; Evolution; Phylogenetic Significance; Basic Functions; Research Opportunities. New Brunswick, NJ, Rutgers University Press, 1987 21. Odin P, Asplund M, Busch C, Obrink B: Immunohistochemical localization of cell CAM 105 in rat tissues: Appearance in epithelia, platelets and granulocytes. J Histochem Cytochem 1988, 36:729-739 22. Osborn M, Weber K: Tumor diagnosis by intermediate filament typing: a novel tool for surgical pathology. Lab Invest 1983, 48:372-394

23. Miettinen M, Virtanen I, Talerman A: Intermediate filament proteins in human testis and testicular germ cell tumors. Am J Pathol 1985,120:402-410 24. Damjanov I, Andrews PW: Ultrastructural differentiation of a clonal human embryonal carcinoma cell line in vitro. Cancer Res 1983, 43:2190-2198 25. Coman DR: Decreased mutual adhesiveness: A property of cells from squamous cell carcinomas. Cancer Res 1944, 4: 625-629 26. Edelman GM, Murray BA, Mege R-M, Cunningham BA, Gallin WJ: Cellular expression of liver and neural cell adhesion molecules after transfection with their cDNAs results in specific cell-cell binding. Proc Natl Acad Sci USA 1987, 84: 8502-8506 27. Nagafuchi A, Shirayoshi Y, Okazaki K, Yasuda K, Takeichi M: Transformation of cell adhesion properties by exogenously introduced E-cadherin cDNA. Nature 1987, 329:341 346