Soluble Factor in Normal Tissues That Stimulates High-Molecular-Weight Sialoglycoprotein Production by Human Colon Carcinoma Cells1

(CANCER RESEARCH 50, 3331-3338. June I, I990| Soluble Factor in Normal Tissues That Stimulates High-Molecular-Weight Sialoglycoprotein Production by Human Colon Carcinoma Cells1

Tatsuro Irimura,2 Andrew M. Mclsaac, Debora A. Carlson, Masato Vagita, Elizabeth A. Grimm, David G. Menter, David M. Ota, and Karen R. Cleary Departments of Tumor Biology IT. I., A. M., D. A. C., M. Y., E. A. G., D. G. M.], General Surgery /E. A. G., D. M. O.J, and Pathology [K. R. C.J, The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030

ABSTRACT linked carbohydrate chains. Secreted mucins are believed to function as protective molecules and as lubricants on tissue The stimulation of high molecular Â»eight Sialoglycoprotein synthesis surfaces. Despite previous attempts by several laboratories to by a soluble factor derived from normal colon tissues was studied in vitro with human colon carcinoma cell lines, HT-29 P and a metastatic variant classify and characterize colorectal mucins, it remained unclear HT-29 LMM. The synthesis of all three high-molecular-weight sialo- whether a specific biological function is associated with unique glycoproteins (approximate M, 900,000, 740,000, and 450,000) by HT- carbohydrate structures in these mucins (10,13-22). Our recent 29 P cells or HT-29 LMM cells growing in vitro was enhanced by studies utilizing pathological specimens of colorectal carcinoma supplementing the culture medium with a conditioned medium of fresh suggested that at least four mucin-like glycoproteins with dif human colon organ culture. Changes were detected by polyacrylamide ferent carbohydrate chains were independently regulated and gel electrophoresis of lysates from [3H)glucosamine-labeled cells on 3% either directly or inversely correlated with the metastatic poten gels followed by fluorography, or by electrophoresis of lysates from tial of these malignant tumors (23-27).3 Thus, these changes unlabeled cells followed by incubation with '"I-labeled wheat germ were apparently associated with the progression of colon car agglutinin and autoradiography. No changes were detected in the major cinoma to the metastatic phenotype, but not simply with malig protein components or in glycoproteins at M, < 200,000 as revealed by polyacrylamide gel electrophoresis. The treated cells did not change their nant transformation. Determining a positive correlation be growth rate or morphology. The connective tissue portions of the colon tween the expression of high molecular weight sialoglycopro tissues were apparently responsible for the production of this stimulatory teins and the progression of colorectal carcinoma to the substance. The stimulatory activity was preserved at 56Â°C but was metastatic phenotype as we demonstrated seemed to be partic inactivated by heating at 100Â°C. The substance was eluted from a ularly significant because increased production of sialoglyco- Sephacryl S-200 column at a position between the elution positions of conjugates has been implicated in a variety of experimental ovalbumin and trypsinogen. The colon carcinoma cells treated with the mÃ©tastasesin rodents (28-34). conditioned medium and producing increased amounts of high-molecular- Two important basic biological questions remained unan weight sialoglycoproteins were less sensitive to the cytolytic effects of recombinant interleukin 2-activated human peripheral blood lymphocytes swered after these studies. The first was what molecular and than untreated cells were. The treated colon carcinoma cells induced cellular mechanisms are involved in the increased expression of stronger platelet aggregation than their untreated counterparts did. sialoglycoproteins during the progression of colorectal carci Therefore, this substance may represent one of the normal host tissue noma to the metastatic phenotype. This question is not simple factors that can influence and modulate malignant behavior of carcinoma because the amount of sialoglycoproteins produced by colon cells growing in vivo. carcinoma cells growing in vivo appears to be influenced by the cellular origin of the carcinoma as well as by microenvironmen- tal factors (26). The second question was whether sialoglyco INTRODUCTION proteins account for the metastatic behavior of colon carcinoma Changes in the production and structures of glycoconjugates cells. In experimental animal models, various cell surface char upon malignant transformation have been described for a wide acteristics have been proposed as being potentially involved in variety of cells (1-5). In colon epithelial cells, changes in the metastatic phenotypes. For example, siah lai ion of cell surface carbohydrate chains on mucin-like, very high-molecular-weight adhesion molecules may modulate the interaction of tumor glycoprotein molecules during carcinogenesis have been re cells with extracellular matrices (35). The effects of cell surface ported since the 1960s (6, 7). A variety of antigenic molecules glycosylation on invasion of embryonic muscle tissues by tumor associated with normal and malignant human colon epithelium cells have also been reported (36). Furthermore, sialoglycocon- were reported as mucin-like, high-molecular-weight glycopro jugates on tumor cell surfaces are known to interfere with the teins or mucin associated, but the epitope structures and spec cytotoxic effects of lymphocytes through nonspecific or specific ificities of the corresponding antibodies were not clear (8-12). mechanisms (37-41). In rat renal carcinoma, sialoglycoconju- Furthermore, the biochemical basis for changes in the expres gates are proposed to play an important role in tumor cell- sion of such tumor-associated antigens was not well understood. induced platelet aggregation (42). Mouse colon carcinoma cells Particularly, the effects of host microenvironments on the pro that were treated with an inhibitor of sialyltransferase showed duction of these molecules were not elucidated. Mucins are significant reduction in their ability to induce platelet aggrega major products of various epithelial cells and are defined by tion (43). their large molecular size and high content of serine/threonine- We describe a soluble factor from normal colon tissues that appears to stimulate the production of cell-associated high- Received 8/16/89; revised 2/9/90. The cosls of publication of this article Â«eredefrayed in part by the payment molecular-weight sialoglycoproteins by colon carcinoma cells of page charges. This article must therefore be hereby marked advertisement in in vitro. The present study is carried out with organ culture- accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1This work was supported by USPHS Grant RO1-CA39319. CA45225. and conditioned media because we did not find a similar stimulatory CA50231: Texas Higher Education Program Grant 1549: and the National factor when an aqueous extract of normal colon tissues or Foundation for Cancer Research. ! To whom requests for reprints should be addressed, at Department of Tumor 3S. D. Hoff. Y. Matsushita. D. M. Ota. K. R. Cleary. T. Yamori, S. Hakomori. Biology. Box 108. The University of Texas M. D. Anderson Cancer Center, 1515 and T. Irimura. Increased expression of sialyl-dimeric Le" antigen in liver Holcombe Boulevard. Houston. TX 77030. mÃ©tastasesofcolorectal carcinoma. Cancer Res., 49: 6883-6888. 1989. 3331

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1990 American Association for Cancer Research. MUCOMODULIN EFFECTS ON COLON CARCINOMA CELLS partial fractionation products of the extract were tested at by measuring absorbance at 590 nm after being dissolving in 1% SDS. various concentration. The results should eventually lead to The morphology of these cells during incubation with NCCM was answers to important questions regarding the function and examined under a phase contrast microscope. regulation of carcinoma mucins. Electrophoretic Analysis of Colon Carcinoma Cell Lysates. Cellular glycoproteins were extracted with non-ionic detergent and treated with SDS and 2-mercaptoethanol as previously described (31). Electropho MATERIALS AND METHODS retic analysis of high-molecular-weight sialoglycoproteins was per formed by using 3% polyacrylamide gels according to previously estab Preparation of Conditioned Media from Normal Colon Tissues. Hu lished methods (25, 26). Sialoglycoproteins were identified by direct man colon tissues used in this study were derived from sigmoid or left binding of I2Ã•I-WGAto polyacrylamide gels. Lysates from ['Hjgluco- colon tissue resected during surgery for colon carcinoma. Normal samine-labeled and ['H]threonine-labeled cells were analyzed by fluo- mucosa at least 5 cm from the carcinoma was used. For the preparation rography with Enhance (DuPont NEN, Wilmington, DE). Lysates from of NCCM,4 the colon epithelium with attached submucosal connective ["Sejselenomethionine-labeled cells were directly autoradiographed tissues and muscularis propria was made free of fat tissues. Fragments after electrophoretic separation. of tissue (approximately 1.0 x 0.3 cm) were then rinsed five times with QuantitÃ¤ten of Changes in High-Molecular-Wcight Sialoglycoprotein gentle agitation and dÃ©cantationwith DPBS supplemented with 50 Production. In most of the experiments, the relative amounts of high- units/ml penicillin, 50 ng/m\ streptomycin, and 1.25 pg/m\ amphoter- molecular-weight sialoglycoprotein production were estimated by the icin B (Fungizone; Sigma, St. Louis, MO). The tissue fragments were binding of 125I-WGA(10 Mg/ml) to the M, 740,000 sialoglycoprotein then incubated with DME/F12 supplemented with antibiotics as de on 3% polyacrylamide gels. This was achieved by removing and count scribed for DPBS. The incubation volume was 10 ml of medium/3.3 g of tissue. After a 48-h incubation at 37Â°Cin a humidified atmosphere ing the radioactivity associated with the appropriate area of the gels. Alternatively, densitometry scanning of the autoradiographs was used with 5% carbon dioxide, supernatants were removed. The supernatants for estimating relative amounts of the M, 740,000 sialoglycoprotein. were either centrifuged at 5000 x g or filtered through Whatman No. Treatments and Kractionation of Normal Colon-conditioned Media. 1 filter paper to remove insoluble materials and then filter sterilized The heat stability of the stimulatory activity for sialoglycoprotein with 0.2-^m nitrocellulose filter membranes. production by colon carcinoma cells was tested by heating NCCM at Conditioned media from different portions of normal colon tissues 56, 80, or IOOÂ°Cfor 30 or 60 min. To determine the size of the were obtained as follows. One tissue fragment free of fat (approximately substance in NCCM responsible for the stimulatory activity, fraction 2x6 cm) was prepared from normal colon wall tissue, cut into three ation by gel filtration Chromatograph) was carried out. NCCM (I ml) portions of equal size, and rinsed with DPBS containing antibiotics as was fractionated on a Sephacryl S-200 column (0.8 x 20 cm) equili described above. One portion was incubated whole in DME/F12. The brated with DME/F12 medium, and l-ml fractions were collected. To second was cut between epithelial cell layer and muscularis through the further study the biochemical nature of this stimulatory activity, 1 ml connective tissue portion, and yielded a muscularis tissue with attached of NCCM was applied to a l-ml column of Con A-Sepharose (Phar connective tissues and epithelial cell layers with attached connective macia). The pass-through fraction, which was eluted with DPBS, and tissues. These tissue fragments were rinsed with antibiotics again and another fraction, which bound to Con A-Sepharose and was eluted with incubated separately. The third portion was scraped with a glass slide 0.5 M Â«-methylmannoside, were pooled separately. These fractions to obtain epithelial cells and to prepare epithelial cell-conditioned were dialyzed against water, lyophilized, and reconstituted in DME/ media. F12 medium. Cultures of Human Colon Carcinoma Cells. HT-29 P cells and HT- Cytolytic Assays. NCCM-treated and untreated HT-29 P cells and 29 LMM cells (obtained from Dr. I. J. Fidler, The University of Texas HT-29 LMM cells were labeled with 5lCr by incubation with 100 MCi/ M.D. Anderson Cancer Center, variant cell lines of HT-29; see Refs. ml of sodium[MCr]chromate for 2 h (46) and then compared for their 26 and 44 for details) were used to examine the effect of NCCM on sensitivity to lymphokine-activated killer lymphocytes. The cells were sialoglycoprotein production. The cells were seeded at 5 x IO5 cells/ washed extensively, detached from the dish, and then suspended in 10-cm plastic tissue culture dish in 10 ml of DME/F12 with 10% FBS. DME/F12 at a final concentration of 5 x IO4cells/ml. Fresh human One day later, the culture media were further supplemented with 0.5 peripheral blood lymphocytes were obtained from healthy donors and ml of NCCM (treated or fractionated), other test substances, or incu subjected to Ficoll-Hypaque centrifugation and depletion of plastic- bation medium containing FBS for control. The substances tested and nylon-adherent cells. The cells were cultured for 4 days in the include serially diluted TGF-ff, (R & D Systems, Minneapolis, MN) or presence of 100 units/ml of recombinant interleukin 2 (Cetus Corp., IL-6 (Genzyme, Boston, MA). In another experiment, various concen Emeryville, CA) and were then harvested, washed by centrifugation, trations of a polyclonal antibody against TGF-ff, + ÃŸ2(neutralizing and resuspended in RPMI 1640 medium containing 5% heat-inacti antibody from R & D Systems) were included. In some experiments, vated human AB serum (Hazelton, Kansas City. MO), 300 mg/ml of NCCM-treated and untreated cells were also incubated with [3H]glu- L-glutamine, and lOmM 4-(2-hydroxyethyl)-l-piperazineethanesulfonic cosamine at a final concentration of 5 ^Ci/ml for 24 h before harvesting acid buffer (lymphokine-activated killer lymphocytes). Large granular to metabolically label cellular glycoproteins. [75Se]-Selenomethionine at a final concentration of 50 nCi/ml and 10 nC\/m\ of ['H]threonine lymphocytes from fresh blood were separated by Percoli (Pharmacia, Uppsala. Sweden) discontinuous gradient centrifugation and were used were used for metabolic labeling of polypeptides. as natural killer cells. Standard 4-h 5lCr-release assays performed in U- Growth Rate and Morphology of HT-29 Cells Grown in Presence or bottomed 96-well plates were used to test the sensitivity of the cells to Absence of Normal Colon-conditioned Media. During the incubation of the cytotoxic effects of lymphokine-activated killer lymphocytes as well HT-29 P cells and HT-29 LMM cells with NCCM, the number of cells as natural killer cells. In these assays, serially diluted effector cells (1.25 in cultures were estimated with direct cell counting on a cell counter x IO4 to 4 x IO5 cells/100 n\) were added with target cells (5 x IO3 after detaching the cells by trypsin treatment. The effects of NCCM on cells/100 ^1). centrifuged at 500 rpm for 5 min, and incubated for 4 h the proliferation rate of these cells were also determined by a crystal at 37Â°Cunder humidified conditions in the presence of 5% carbon violet staining method (45) by using 96-well micro tissue culture plates. In this experiment, the cells were seeded at 5 x lO'/well in the absence dioxide and 95% air. At the end of the assay, the plates were centrifuged again, and the supernatants were harvested by using a Skatron super or presence of 5% NCCM. The cells were fixed in 0.05% glutaraldehyde natant collection system. (Maximum release of 51Cr was produced by at days 1, 2, 3, 4, and 5, stained with 0.1% crystal violet, and rinsed addition of 0.1 M sodium hydroxide.) The percentage of specific lysis repeatedly with water. The relative number of the cells were estimated was calculated by the formula: *The abbreviations used are: NCCM, normal colon conditioned medium: cpm with effector cells - cpm spontaneously released DPBS. Dulbccco's phosphate-buffered saline: DME/F12, 1:1 mixture of Dulbec- x 100. co's minimum essential medium and Ham's F-12 medium: FBS, fetal bovine cpm maximum - cpm spontaneously released serum: Con A, eoncanavalin A: TGF-/J. transforming growth factor {]'.WGA. wheat germ agglutinin: SDS, sodium dodecyl sulfate. All determinations were performed in quadruplicate. 3332

Platelet Aggregation Assays. Platelets were prepared from freshly .o\Â° drawn heparinized blood from normal healthy donors. A suspension of platelets at a concentration of 4 x 106/rnl in Hanks' balanced salt solution was placed in an aggregometer and preincubated for 15 min at 37Â°C.An aggregation reaction was initiated by adding 5 x IO5 untreated or NCCM-treated cells suspended in 0.1 ml of Hanks' bal anced salt solution. The degree of light scattering was determined by measuring the absorbance at 550 nm (47). 900,000 RESULTS 740,000 Increase in Levels of High-Molecular-Weight Sialoglycopro- teins in Colon Carcinoma Cells Induced by NCCM in Vitro. Because we had previously observed that a greater amount of 450,000 high-molecular-weight sialoglycoproteins was associated with HT-29 P cells and HT-29 LMM cells growing in vivo in nude mice than with the same cells growing in vitro, we attempted to mimic the alterations in the sialoglycoproteins in cell cultures. The cells grown in the absence or presence of NCCM were extracted with non-ionic detergent solution (lysis buffer) and 190,000 the extracts were separated by polyacrylamide gel electrophoresis in the presence of SDS. High-molecular-weight sialogly coproteins were examined after they were incubated with I25I- WGA and autoradiographed as previously described (25, 26). Fig. 1 shows electrophoretic profiles of high-molecular-weight [3H]GlcN [3H]Thr [75Se]Seleno-Met sialoglycoproteins of HT-29 P and HT-29 LMM cells incubated with 5% NCCM for 4 days. An increase in the binding of I25I- Fig. 2. Results of electrophoretic analysis in the presence of SDS on 3% polyacrylamide gels of lysates of HT-29 LMM cells grown in the absence or WGA to high-molecular-weight glycoproteins (M, ~900,000, presence of 5% NCCM for 4 days. The molecular weights (ordinale) of the major 740,000, and 450,000) was seen in NCCM-treated cells. No bands were estimated from the positions of intact laminin and subunits of laminin as described previously (25, 26). Left, autoradiograms of metabolica!!} [3H]- significant changes were observed in WGA binding to the glucosamine-labeled components. Center, autoradiograms of metabolically [JHJ- components with M, < 200,000. A slight change seen on M, threonine-labeled components. Right, autoradiograms of metabolically [75Se]se- 540,000 component was not reproducible. Fig. 2, left, shows lenomethionine-labeled components. similar changes as detected by the incorporation of ['H]glucosamine into HT-29 LMM cells. The incorporation of [3H]- glucosamine into the M, 190,000 component did not signifi cantly change after NCCM treatment. As seen in Fig. 2, center 1234 and right, incorporation of ['Hjthreonine or [75Se]selenomethionine into components with molecular weights ranging be tween 400,000 and 1,000,000 was very low relative to that to components M, < 200,000. The incorporation did not seem to significantly change after NCCM treatment, but this was not 900,000 conclusive because the levels of the incorporation were below the ranges of detection by these methods. No alterations were 740,000 detected in the profiles of proteins with molecular weights as low as 30,000, as revealed by Coomassie blue staining or by the incorporation of ['Hjthreonine or [75Se]selenomethionine (data not shown). Mild acid treatment (50 mM sulfuric acid treatment 450,000 at 80Â°Cfor 60 min) of glycoproteins in polyacrylamide gels prior to incubation with WGA eliminated WGA reactivity from all the components, indicating that the WGA-identified mole cules are sialoglycoproteins. The intensity of the bands revealed by fluorography based on [3H]glucosamine incorporation was not significantly altered after mild acid treatment (data not shown), indicating that the increase in WGA reactivity and ['H]glucosamine incorporation seen in NCCM-treated cells was not simply due to enhanced sialylation. The electrophoretic 190,000 migration distance of high-molecular-weight sialoglycoproteins did not change after treatment with NCCM. At this time, it is not clear whether the changes in WGA binding or ['Hjgluco- samine incorporation is due to increased production of entire sialoglycoprotein molecules or increased glycosylation. Fig. 1. Results of electrophoretic analysis in the presence of SDS on 3% polyacrylamide gels of lysates of HT-29 P cells (Lanes I and 2) and HT-29 LMM Growth Rate and Morphology of HT-29 Cells Grown in Pres cells (Lanes 3 and 4) grown in the absence (Lanes ! and 3) or presence (Lanes 2 ence or Absence of Normal Colon-conditioned Media. There was and 4) of 5% NCCM for 4 days. The molecular weights (ordinale) of the major no detectable change in the growth rate of these cells during bands were estimated from the positions of intact laminin and subunits of laminin as described previously (25, 26). Autoradiograms of the gels incubated with 12!I- incubation in NCCM, and the saturation densities of these cells WGA, rinsed, and dried are shown. grown in the presence or absence of NCCM were very similar. 3333

Although these cells in culture always have heterogeneous mor The fraction containing the activity was eluted from the column phology, the proportion of cells with flat appearances, com at a position between ovalbumin (M, -45,000) and trypsinogen pared with other cells that were tightly associated with each (M, -25,000), approximately corresponding to M, -30,000 other, did not appear to differ between untreated and NCCM- (data not shown). treated cells (data not shown). Time Course and Reversibility of Effects of Normal Colon- <Â£V

Fig. 4. Reversibility of the effects of NCCM on the levels of high-molecular- 900,000 weight (ordinate) sialoglycoproteins by HT-29 LMM colon carcinoma cells. HT- 29 LMM cells were cultured for 8 days in 10-cm culture dishes. During the period, 5% NCCM was added to the media at days 1-3, days 2-4, days 3-5, days 4-6, days 5-7, or days 6-8. and the media were changed to plain DME/F12 after 740,000rinsing the cells with 10 ml of the same medium three times. The cells from these cultures and a control culture were harvested by brief trypsinization and lysed. Lysates were separated by SDS-polyacrylamide gel electrophoresis in 3% gels. The gels were incubated with I251-WGA,rinsed, dried, and autoradiographed. '' 450,000 Untreated 56Â°C30 min 56Â°C60 min 80Â°C30 min 80'C 60 min 100Â°C30min j Con A-Sepharose/ pass Con A-Sepharose/ bound 190,000 40 60 80 100 120 % stimulation Fig. 5. Effects of NCCM on '"I-WGA binding to an M, 740.000 component. NCCM was heat treated or fractionated on the columns as indicated. Treated or fractionated NCCM at the equivalent dilution was added to separate cultures of HT-29 LMM cells and incubated for 4 days. Lysates of the cells were analyzed by SDS-polyacrylamide gel electrophoresis in 3% gels, followed by incubation Fig. 3. Time dependency of the effects of NCCM on the levels of high- with II5I-\VGA. The gels were rinsed, dried, and autoradiographed. The radioac molecular-weight (ordinate) sialoglycoproteins by HT-29 LMM colon carcinoma tive materials in the gels corresponding to the M, 740.000 sialoglycoprotein were cells. Separate cultures of HT-29 LMM cells were incubated with 5% NCCM for counted with a gamma counter, and the relative increase of the radioactivity of 1-5 days. Lysates from the cells were separated by SDS-polyacrylamide gel bound '"I-WGA was calculated based on the counts from HT-29 LMM cells electrophoresis in 3% gels, incubated with '"I-WGA, rinsed, dried, and autora grown in the absence of NCCM (0%) or presence of 5% NCCM (100%), on the diographed. same gel. 3334

Cells Producing the Factor and Their Viability. Previously we molecular-weight sialoglycoproteins. Neutralizing polyclonal have shown that cultures of normal human colon tissues main antibodies against TGF-ÃŸi+ ft> did not change the effects of tain their metabolic activity such as production of sulfated NCCM. Human IL-6 (5-5000 units/ml) did not have signifi mucins and production and secretion of collagenolytic enzymes cant effects on high-molecular-weight sialoglycoprotein pro for at least 48 h (24, 48). Therefore, we prepared conditioned duction, as revealed by I25I-WGA binding after electrophoretic media from normal colon tissues after a 48-h incubation. We separation. further tested whether the cells in normal colon tissues maintain Interaction of Activated Lymphocytes with Colon Carcinoma their metabolism during this incubation period. When protein Cells Cultured in Presence or Absence of Normal Colon-condi profiles with metabolically incorporated [3H]threonine at pe tioned Media. We hypothesized that the increased sialoglyco riods of 0-24 h, 24-48 h, and 48-72 h were analyzed and proteins would influence and modulate the interaction of colon compared by polyacrylamide gel electrophoresis, difference was carcinoma cells with activated lymphocytes. The NCCM- not seen as judged by Coomassie brilliant blue staining or by treated and untreated cells previously labeled with sodium fluorography (data not shown). HistolÃ³gica! observation of the [5'Cr]chromate were examined for their sensitivity to human tissues at days 0, 1, and 2 indicated that connective tissues and peripheral blood lymphocytes activated with interleukin 2 in muscularis remained intact during the incubation period, al short-term killing assay. HT-29 P cells or HT-29 LMM cells though some damage was seen in epithelial cells. with or without previous treatment with NCCM were sensitive To identify the tissue compartments producing this stimula to natural killer cells, as assayed by the use of Percoll-separated tory factor for sialoglycoprotein production by colon carcinoma large granular lymphocytes as effectors. Lymphokine-activated cells, we separated normal colonie mucosa into the epithelial killer lymphocytes showed marked cytolytic activity against cell layer and the connective tissues with muscularis propria. these cells. Fig. 7 shows representative data demonstrating that Epithelial cells with submucosa were also prepared. We then HT-29 P cells or HT-29 LMM cells treated with NCCM are obtained conditioned media from the cultures of these tissue more resistant to the cytotoxic effect of activated lymphocytes segments and compared them for their ability to stimulate than untreated counterparts are. sialoglycoprotein production. The results shown in Fig. 6 in Induction of Platelet Aggregation by Colon Carcinoma Cells dicate that a higher concentration of this soluble factor seemed Cultured in Presence or Absence of Normal Colon-conditioned to be produced and secreted from tissue fragments containing Media. It has been reported that metastatic mouse colon car exposed connective tissues. Thus, we tested whether a condi cinoma cells possess the capacity to induce strong platelet tioned medium of human colon fibroblasts elicited activity aggregation and that tumor cell surface sialic acid might be similar to that of NCCM. Human colon fibroblast cells CCD- important in this process (43). Therefore, we hypothesized that 18Co (ATCC-CRL 1459) from the American Type Culture HT-29 P and HT-29 LMM human colon carcinoma cells Collection, were grown in DME/F12 with 10% FBS. The producing different amounts of high-molecular-weight sialog conditioned media of these cells at a concentration as low as lycoproteins might have different platelet-aggregating activities. 1% showed similar effects on sialoglycoprotein production, Untreated and NCCM-treated cells were mixed with suspen suggesting that fibroblasts are the cells responsible for the sions of human platelets, and the changes in light transmission production of the soluble factor in NCCM. under constant gentle mixing were monitored. As shown in Fig. Could Effects of Normal Colon-conditioned Media be Ex 8, the rate of platelet aggregation induced by NCCM-treated plained by Previously Known Soluble Mediators? Because fibro blasts are known to produce TGF-0 and IL-6 which have apparent molecular weights within the range of 25,000 to 80 45,000, we examined whether these known soluble mediators 70 had similar biological effects on HT-29 P cells and HT-29 60 LMM cells as those seen with NCCM. TGF-/3, (0.1-10 /ig/ml) 50 did not have any effect on the binding of WGA to high- 40 30 20 .cÃ¯Ã¯ 10 D>2 0 O Â°Q -10 1 10 100

IlS 80 70 60 50

Ã§ao-*: tr 40 Ã¼^ o Ã”j 30 Â£5 Sir .01 .1 1 10 20 % Conditioned Media Added 10 Fig. 6. Comparison of different portions of normal colon tissues for produc 0 tion of the factor that stimulates sialoglycoprotein production by HT-29 LMM cells. Whole colonie wall (A); epithelial cell layer (O); epithelial cells with attached -10 connective tissues (A); muscularis with attached connective tissues (â€¢),were 1 10 100 compared. Serially diluted conditioned media from these cultures were added E/T ratio to separate cultures of HT-29 LMM cells and incubated for 4 days. Lysates of Fig. 7. Sensitivity of NCCM-treated and untreated human colon carcinoma the cells were analyzed by SDS-polyacrylamide gel electrophoresis in 3% gels, cells to cytolysis by interleukin 2-activated human peripheral blood lymphocytes followed by incubation with '"I-WGA, rinsing, drying, and autoradiography. (A, â€¢)and by large granulated lymphocytes (A. O). HT-29 P cells (a) and HT-29 The radioactive materials in the gels containing the M, 900.000. 740,000, and LMM cells (b) before (A. A) or after (â€¢,O)treatment with NCCM were used as 450,000 component were counted with a gamma counter, and the relative increase target cells. The tumor cells were previously labeled with "Cr, and the percentage of bound WGA was calculated. kD, molecular weight in thousands. of cytolysis was determined after a 4-h incubation. 3335

10Ch incorporation, the observed changes may be due to increased glycosylation or to increased synthesis of whole glycoprotein molecules. Increased sialylation does not account for the in creased incorporation of ['Hjglucosamine because a very similar NCCM difference in the radioactivity between untreated and NCCM- treated cells was observed after specific removal of sialic acid from these sialoglycoproteins by mild acid hydrolysis. Struc Control tural characterization of purified sialoglycoproteins synthesized in the presence or absence of NCCM along with identification and quantitation of the core polypeptide can answer the ques tions about the biochemical basis of the stimulation of sialo glycoprotein synthesis caused by NCCM treatment. Neverthe less, this provides a unique system with which we can elucidate the regulation of high-molecular-weight sialoglycoprotein pro duction by colon carcinoma cells and the biological functions of these molecules. pHJGlucosamine-labeled high-molecular -weight sialoglycoproteins that were either treated or not treated with NCCM were examined for the sensitivity of the sugar chains to alkaline degradation, and the results strongly sug 5 10 Time in Minutes gested that these high-molecular-weight sialoglycoproteins are mucin-like molecules. A similar stimulatory effect was seen on Fig. 8. Time course of platelet aggregation as indicated by increase of light transmission induced by untreated and NCCM-treated HT-29 P cells. HT-29 other human colon carcinoma cell lines such as HCC 1410, LMM cells induced more intense platelet aggregation as shown by higher light scattering than untreated HT-29 P cells did. It was further augmented by NCCM DiFi, andKM12C. treatment (data not shown). We have also tested other cellular and molecular changes in the NCCM-treated HT-29 P cells and HT-29 LMM cells. No cells was significantly higher than that induced by untreated changes were seen in the growth rate or the morphological cells. appearance of the cells under the same conditions. These find ings indicated that these cells were not undergoing differentia tion toward goblet-like cells or any other cells, in contrast with DISCUSSION the differentiation seen in other sublines of HT-29 cells grown We previously showed that the production of high-molecular- in glucose-depleted medium (49). Foss et al. (50) have reported weight sialoglycoproteins by human colon carcinoma HT-29 P that protein kinase activity associated with srcpp60is much lower cells and HT-29 LMM cells was apparently stimulated when in colon carcinoma cells expressing differentiated phenotypes the cells were grown in vivo in nude mice (26). We hypothesized than those with undifferentiated phenotypes. However, the that the levels of high-molecular-weight sialoglycoproteins on levels of phosphorylation of srcpp60and epidermal growth factor colon carcinoma cells were influenced by the intrinsic nature of receptor did not change after NCCM treatment of HT-29 P the tumor cells (clonal variation) as well as by host microenvi- cells or HT-29 LMM cells. Furthermore, the increase in sial ronmental factors that have not yet been characterized. In the oglycoprotein production was reversible. These results strongly present report, such a factor which stimulates sialoglycoprotein suggest that the effect of NCCM is not due to selective growth production is shown in vitro using culture-conditioned media of a carcinoma cell subpopulation expressing higher contents of normal colon tissues. We obtained NCCM by incubating of high-molecular-weight sialoglycoproteins or to irreversible fresh human colon tissues in culture. The changes in sialogly differentiation. However, whether other cellular changes may coproteins in human colon carcinoma cells were revealed by be associated with increased sialoglycoprotein production in the binding of 125I-WGA to high-molecular-weight components duced by NCCM remains unknown. after electrophoretic separation of the cell lysates or by meta Our initial experiments were performed with NCCM, con bolic incorporation of [3H]glucosamine followed by electropho ditioned media of organ-cultured colon tissues containing epi retic separation of the cell lysates. Production in vitro of all of thelial cells, connective tissues, and muscularis. To examine the the three high-molecular-weight sialoglycoproteins (approxi metabolic activity of the tissue during such incubations, meta bolic labeling of the tissues with [3H]threonine, followed by mate M, 900,000, 740,000, and 450,000) was enhanced by the addition of this conditioned medium to the culture, whereas electrophoretic analysis of the lysate of the tissues was per differences in the production of glycoproteins with M, < formed. The results of this experiment and morphological 200,000 or in total protein profiles were not detected. These examinations demonstrated that the tissues maintained their results suggested that the conditioned media from human colon metabolic activity and their original morphology during the 48- tissues specifically enhanced high-molecular-weight sialogly h incubation period. To identify which portion of the tissue coprotein production compared with the synthesis of other produced and secreted the active substance responsible for the glycosylated and nonglycosylated proteins. NCCM-treated cells stimulation of high-molecular-weight sialoglycoprotein produc secreted an increased amount of a M, 900,000 sialoglycoprotein tion, we obtained conditioned media from different portions of but did not secrete the M, 740,000 or M, 450,000 components. large bowel wall tissues, and tested their ability to induce the This may explain why the M, 900,000 component appeared to change in I25I-WGA binding to high-molecular-weight sialo be predominant in tumors grown in vivo, which maintained glycoproteins produced by HT-29 LMM cells. The results secretion products (26). suggested that the cells within connective tissues were respon Because the increase in high-molecular-weight sialoglycopro sible for the production of this factor. Furthermore, we found tein production was detected by means of increased sialylated that fibroblasts are probably the cells that produce this factor carbohydrate chains, i.e., I25I-WGA binding or ['H]glucosamine in normal colon tissues. Substances in serum were apparently 3336

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1990 American Association for Cancer Research. MUCOMODULIN EFFECTS ON COLON CARCINOMA CELLS necessary for fibroblasts to produce and/or secrete the factor in 4Glc 1-lceramide), and Altevogt et al. (54) showed in the ESb/ vitro. Normal colon tissues did not require serum for the pro Eb T-lymphoma system that the terminal TV-acetylgalactosa- duction of the factor, probably because the active principle in mines recognized by some lectins (soybean agglutinin and Vicia the serum was endogenously provided. villosa agglutinin) are concealed by sialylation on the surface of The sialoglycoprotein-stimulatory activity in NCCM was rel ESb cells highly metastatic to the liver. In the stimulation of atively stable after mild heating or freeze-thawing. Preliminary mouse thymocyte proliferation by allogeneic B-lymphoma cells, fractionation and characterization experiments suggested that sialic acid on the surface of the tumor cells serves as a negative the factor has a molecular weight of 30,000 and is likely to be signal reducing thymocyte response (39). Similar concealing a protein or glycoprotein. Therefore, we have tentatively named effects of high-molecular-weight mucin-like glycoproteins were the factor mucomodulin, referring to its ability to modulate already reported in mouse and rat mammary carcinoma systems mucin synthesis. Based upon our finding that this molecule is (38, 41). Our study demonstrated that the expression of high- produced by fibroblasts, we have tested whether known poly- molecular-weight mucin-like sialoglycoproteins was induced by peptides produced by normal fibroblasts such as TGF-0 and IL- normal tissue-derived factors and was associated with reduced 6 possess a similar activity and have concluded that these sensitivity of human colon carcinoma cells to lymphokine- substances did not stimulate the production of high-molecular- activated killer activity. We examined the incidence of conju weight sialoglycoproteins by colon carcinoma cells. Therefore, gate formation of activated lymphocytes with HT-29 P and the factor(s) may specifically regulate cell-cell interaction by HT-29 LMM target cells (55, 56), based on the hypothesis that changing the architecture of the cell surface. Our observation a greater amount of high-molecular-weight sialoglycoproteins is unique because the effect seems very specific to mucin-like concealed the sites on the tumor cells required for recognition high-molecular-weight sialoglycoproteins. Rearick et al. (51) by lymphokine-activated killer lymphocytes. A smaller propor reported that mucin production by cultured rabbit trachea! cells tion of lymphocytes adhered to NCCM-treated cells than to was stimulated by mouse 3T3 fibroblast-conditioned media in untreated cells. In contrast, only a small difference was found the presence of retinoic acid. In the human colon carcinoma in the proportion of dead conjugated target cells among effec experimental systems we used, the effects of retinoic acid were tor-target conjugates detected by trypan blue dye exclusion tests inhibitory against high-molecular-weight sialoglycoprotein pro in semisolid agar. These results suggested that high-molecular- duction. weight sialoglycoproteins apparently influence tumor cell-lym Examination of a large number of fresh tumor specimens phocyte conjugations. A more systematic approach is necessary previously showed that WGA-binding high-molecular-weight to prove this hypothesis. sialoglycoproteins are often associated with advanced stages of The other well-known biological activity of tumor cell surface colorectal carcinoma with high metastatic potential (26). There sialoglycoproteins is their ability to induce platelet aggregation. fore, the cellular mechanisms for the regulation of these mole Pearlstein et al. (42, 57) reported that a substance that induces cules may have direct relevance to the progression and malig platelet aggregation present in extracts of renal carcinoma and nant behavior of human colorectal cancer. Many experimental other cells can be destroyed by sialidase treatment. Kijima-Suda and clinical studies have demonstrated that the biological be et al. (43) demonstrated that the amount of mouse colon havior of tumor cells directly isolated from tumor tissues was carcinoma cell surface sialoglycoconjugates decreased after different from that of cells growing in culture. Sialoglycopro treatment of the cells with an inhibitor of sialyltransferase and teins and stimulators of sialoglycoprotein production may me that there was a concomitant decrease in platelet aggregation diate such alterations. Endogenous factors that stimulate the and in metastatic potential. Later, they suggested that mucin- production of sialoglycoproteins have not been documented type sugar chains are important for the induction of platelet previously. We assessed whether colon carcinoma cells treated aggregation (58). Our results clearly demonstrated that the with NCCM was altered in their sensitivity to natural immune ability of human colon carcinoma cells to induce platelet aggre effectors, such as natural killer lymphocytes and lymphokine- gation was enhanced by prior treatment of the cells with activated killer lymphocytes. The cells expressing increased NCCM. It is plausible, therefore, that the changes in platelet sialoglycoproteins were shown to be much less sensitive to aggregation by colon carcinoma cells can be attributed to the cytolysis by human lymphokine-activated killer lymphocytes increase in sialoglycoproteins. Thus, an induction of high- suggesting that the higher levels of cell surface sialoglycoprotein molecular-weight sialoglycoprotein production probably pro resulted in increased chances of survival and metastatic co vides colorectal carcinoma cells with an enhanced capacity to lonization of colon carcinoma cells. metastasize by means of their abilities to evade immune defense The importance of sialoglycoconjugates in immune effector mechanisms and to form emboli with platelets. Purification cell recognition has been suggested for various tumor cells. For and characterization of this substance, which we are tentatively example, Dennis and LafertÃ©(37) demonstrated that WGA- calling mucomodulin, will greatly facilitate future studies of resistant MDAY leukemia-like mouse cells lacked cell surface tumor cell-host cell interaction in metastasis formation. sialoglycoproteins and were more sensitive to natural killer cell- mediated cytolysis. They showed that nonsialylated sugar REFERENCES chains can be recognized by syngeneic natural killer lympho cytes. A similar finding was reported by Ahrens and Ankel 1. Warren. L., Buck, C. A., and Tuszynski, G. P. Glycopeptide changes and using Chinese hamster ovary cells (52). In contrast. Van Rin- malignant transformation: a possible role for carbohydrate in malignant behavior. Biochim. Biophys. Acta. 516: 97-127. 1978. sum et al. (53) found that sialylated carbohydrate chains were 2. Smets. L. A. Cell transformation as a model for tumor induction and crucial structures in a post-binding event in natural killer lym neoplastic growth. Biochim. Biophys. Acta, 605: 93-111, 1980. phocyte-mediated cytolysis. Fukuta et al. (40) demonstrated 3. Hakomori. S. Aberrant glycosylation in tumors and tumor-associated car bohydrate antigens. Adv. Cancer Res., 52: 257-331, 1989. that cytolysis of melanoma mediated by human anomalous 4. Sanier, U. V., Gilbert, F., and Click, M. C. Change in glycosylation of killer cell (similar or equivalent to lymphokine-activated killer membrane glycoproteins after transfection of N1H 3T3 with human tumor cells) was blocked by a monoclonal antibody specific for disi- DNA. Cancer Res., 44: 3730-3735, 1984. 5. Reading, C. L., and Hutchins, J. T. Carbohydrate structure in tumor immu aloganglioside GD2 (GalNAc l-4[NeuAc 2-8NeuAc 2-3]Gal 1- nity. Cancer Metastasis Rev.. 4: 221-260, 1985. 3337

6. Filipe, M. I. Value of histochemical reactions for mucosubstances in the with cell surface sialylation of cultured murine tumor cell lines. Science diagnosis of certain pathological conditions of the colon and rectum. Gut, (Wash. DC), 212: 1514-1516, 1981. 10: 577-586, 1969. 34. Schirrmacher, V., Altevogt, P., Fogel, M., Dennis, J., Waller, C. A.. Barz, 7. Kim, Y. S., Issacs, R., and Perdomo, J. M. Alterations of membrane glyco- D., Schwartz, R., Cheingsong-Popov, R., Springer, G., Robinson. P. J., peptides in human colonie adenocarcinoma. Proc. Nati. Acad. Sci. USA, 71: Nebe. T., Brossmer, W., Vlodavsky, 1., Paweletz, N., Zimmerman, H-P., and 4869-4873, 1974. Uhlenbruck, G. Importance of cell surface carbohydrates in cancer cell 8. Burtin, P., von Kleist, S., and Sabine, M-C. Loss of a normal colonie adhesion, invasion and metastasis. Invasion Metastasis, 2: 313-360, 1982. membrane antigen in human cancers of the colon. Cancer Res., 31: 1038- 35. Dennis, J., Waller, C., Timple, R., and Schirrmacher, V. Surface sialic acid 1041, 1971. residues attachment of metastatic tumor cells to collagen type IV and 9. Goldenberg, D. M., Pant, K. D., and Dahlman, H. L. Antigens associated fibronectin. Nature (Lond.), 300: 274-276, 1980. with normal and malignant gastrointestinal tissues. Cancer Res., 36: 3455- 36. Bolscher, J. M., Schallier, D. C. C., van Rooy. H., Storme, G. A., and Smets, 3463, 1976. L. A. Modification of cell surface carbohydrates and invasive behavior by an 10. Gold. D. V., and Miller, F. Comparison of human colonie mucoprotein alkyl lysophospholipid. Cancer Res., 48: 977-982, 1988. antigen from normal and neoplastic mucosa. Cancer Res., 38: 3204-3211, 37. Dennis, J. W.. and LafertÃ©,S.Recognition of asparagine-linked oligosaccha 1978. rides on murine tumor cells by natural killer cells. Cancer Res., 45: 6034- 11. Hughes, N. R., Walls, R. S., Newland, R. C, and Payne, J. E. Antigen 6040, 1985. expression in normal and neoplastic colonie mucosa. Three tissue-specific 38. Codington, J. F., Klein, G., Cooper, A. G., Lee, N., Brown, R. W., and antigens using monoclonal antibodies to isolated colonie glands. Cancer Res., Jeanloz, R. W. Further studies on the relationship between large glycoprotein 46:2164-2171, 1986. molecules and allotransplantability in the TA3 tumor of the mouse: studies 12. Das, K. M., Sakamaki, S., Vecchi, M., and Diamond, B. The production and on segregating TA3-Ha hybrids. J. Nati. Cancer Inst., 60: 811-818. 1978. characterization of monoclonal antibodies to a human colonie antigen asso 39. Powell, L. D., Bause, E., Legier, G., and Hart, G. W. Influence of asparagine- ciated with ulcerative colitis: cellular localization of the antigen by using the linked oligosaccharides on tumor cell recognition in the mixed lymphocyte monoclonal antibody. J. luminimi.. 139: 77-84, 1987. reaction. J. Immunol., 135: 714-724, 1985. 13. Tsao. D., and Kim, Y. S. Glycoprotein from human colonie adenocarcinoma. 40. Fukuta, S., Werkmeister, J. A.. Burns. G. F., Ginsburg, V., and Magnani, J. Isolation and characterization of cell surface carcinoembryonic antigen from L. Monoclonal antibody Leo Mel 3, which inhibits killing of human mela a cultured human tumor cell line. J. Biol. Chem., 253: 2271-2278, 1978. noma cells by anomalous killer cells, binds to a sugar sequence in GD2 14. Slomiany, B. L., Varahabhotla, L. N., and Slomiany, A. Isolation and (II3(NeuAc)2-GgOse,Cer) and several other gangliosides. J. Biol. Chem., 262: characterization of oligosaccharides from rat colonie mucus glycoprotein. J. 4800-4803, 1987. Biol. Chem., 255: 9719-9723. 1980. 41. Sherblom, A. P., and Moody, C. E. Cell surface sialomucin and resistance to 15. LaMont, J. T., and Ventla, A. S. Purification and composition of colonie natural cell mediated cytotoxicity of rat mammary tumor ascites cells. Cancer epithelial mucin. Biochim. Biophys. Acta, 626: 234-243, 1980. Res., 46: 4543-4546, 1986. 16. Boland, C. R., Montgomery, C. K., and Kim, Y. S. Alterations in colonie 42. Pearlstein, E., Salk, P. L., Yogeeswaran, G., and Karpatkin. S. Correlation mucin occurring with cellular differentiation and malignant transformation. between spontaneous metastatic potential, platelet-aggregating activity of cell Proc. Nati. Acad. Sci. USA, 79: 2051-2055, 1982. surface sialylation in 10 metastatic-variant derivatives of rat renal sarcoma 17. Podolsky, D. K., and Isselbacher, K. J. Glycoprotein composition of colonie cell line. Proc. Nati. Acad. Sci. USA, 77: 4336-4339, 1980. mucosa. Specific alterations in ulcerative colitis. Gastroenterology, 87: 991- 43. Kijima-Suda, I., Miyamoto. Y.. Toyoshima. S., Itoh, M., and Osawa, T. 998, 1984. Inhibition of experimental pulmonary metastasis of mouse colon adenocar 18. Podolsky, D. K. Oligosaccharide structures of human colonie mucin. J. Biol. cinoma 26 subline by a sialic acid:nucleoside conjugate having sialyltransfer- Chem., 260: 8262-8271. 1985. ase inhibiting activity. Cancer Res., 46: 858-862, 1986. 19. Podolsky, D. K. Oligosaccharide structures of isolated human colonie mucin 44. Price, J. E., Daniels. L. M.. Campbell, D. E., and Giavazzi, R. Organ species. J. Biol. Chem., 260: 15510-15515, 1985. distribution of experimental mÃ©tastasesof a human colorectal carcinoma 20. Arends, J. W., Bosnian. F. T., and Hilgers, J. Tissue antigens in large-bowel injected in nude mice. Clin. Exp. Metastasis, 7: 55-68, 1989. carcinoma. Biochim. Biophys. Acta, 780: 1-19, 1985. 45. Gillies, R. J., Didier. N., and DentÃ³n. M. Determination of cell number in 21. Wesley, A., Mantle, M., Man, D.. Qureshi, R., Fonstein, G., and Forster, J. monolayer cultures. Anal. Biochem.. 159: 109-113. 1986. Neutral and acidic species of human intestinal mucin. Evidence for different 46. Grimm, E. A., Mazumder, A., Zhang, H. Z., and Bonavida. B. Lymphokine- core polypeptides. J. Biol. Chem., 260: 7955-7959, 1985. activated killer cell phenomenon lysis of natural killer-resistant fresh solid 22. Boland, C. R., Roberts, J. A., Siddiqui. B., Byrd, J., and Kim, Y. S. Cancer- tumor cells by interleukin 2-activated autologous human peripheral blood associated colonie mucin in cultured tumor cells and athymic (nude) mouse lymphocytes. J. Exp. Med.. 155: 1823-1841, 1982. xenografts. Cancer Res., 46: 5724-5729, 1986. 47. Menter, D. G., Onoda, J. M., Taylor. J. D., and Honn, K. V. Effects of 23. Irimura, T., Ota, D. M., and Cleary, K. R. Ulex europeus ag%\\Amml-reactive prostacyclin on tumor cell-induced platelet aggregation. Cancer Res., 44: high molecular weight glycoproteins of adenocarcinoma of distal colon and 450-456, 1984. rectum and their possible relationship with metastatic potential. Cancer Res., 48. Irimura, T., Yamori, T., Bennett, S. C, Ota, D. M., and Cleary, K. R. The 47:881-889, 1987. relationship of collagenolytic activity to stage of human colorectal carcinoma. 24. Yamori, T., Kimura, H.. Stewart, K., Ota, D. M., Cleary, K. R., and Irimura, Int. J. Cancer, 40: 24-31, 1987. T. Differential production of high molecular weight sulfate glycoproteins in 49. Pinto, M., Appay, M. D., Simon-Assmann, P., Chevalier, G., Dracopoli, N., normal colonie mucosa, primary colon carcinoma and mÃ©tastases.Cancer Fogh, J., and Zweibaum, A. Enterocytic differentiation of cultured human Res., 47: 2741-2747, 1987. colon cancer cells by replacement of glucose by galactose in the medium. Biol. Cell, 44: 193-196, 1982. 25. Irimura, T., Carlson, D. A., and Ota, D. M. Electrophoretic analysis of four 50. Foss, F., Bolen, J., Veilette, A., Bostick. F., and Rosen, N. Regulation of c- high molecular weight sialoglycoproteins produced by metastatic human colon carcinoma cells. J. Biol. Chem., 37: 1-9, 1988. src expression in colon carcinoma. Proc. Am. Assoc. Cancer Res., 29: 41, 26. Irimura, T.. Carlson, D. A., Yamori, T., Price, J., Giavazzi, R., Ota, D. M., 1988. and Cleary, K. R. Differential expression of a sialoglycoprotein with an 51. Rearick, J. I., Deas, M., and Jetten. A. M. Synthesis of mucous glycoproteins by rabbit trachÃ©alcells in ritro. Modulation by substratum, retinoids and approximate molecular weight of 900,000 on metastatic human colon carci cyclic AMP. Biochem. J., 242: 19-25, 1987. noma cells growing in culture and tumor tissues. Cancer Res., 48: 2353- 52. Ahrens, P. A., and Ankel, H. The role of asparagine-linked carbohydrate in 2360, 1988. natural killer cell-mediated cytolysis. J. Biol. Chem., 262: 7575-7579. 1987. 27. Yamori. T., Ota, D. M., Cleary, K. R., Hoff. S., Hager, L. G., and Irimura, 53. Van Rinsum. J., Smets, L. A., Van Rooy, H., and Van den Eijnden, D. H. T. Monoclonal antibody against human colonie sulfomucin: immunochemi- Specific inhibition of human natural killer cell-mediated cytotoxicity by sialic cal detection of its binding sites in colonie mucosa, colorectal primary acid and sialo-oligosaccharides. Int. J. Cancer, 38: 915-922, 1986. carcinoma, and mÃ©tastases.Cancer Res., 49: 887-894, 1989. 54. Altevogt, P., Fogel, M., Cheinsong-Popov, R., Dennis, J., and Schirrmacher, 28. Finne, J., Tao, T. M., and Burger, M. M. Carbohydrate changes in glycopro V. Related high and low metastatic tumor lines show different patterns of teins of a poorly metastasizing wheat germ agglutinin-resistant melanoma lectin binding and cell surface sialylation. Cancer Res., 43: 5138-5144,1983. clone. Cancer Res.. 40: 2580-2587, 1980. 55. Grimm, E., and Bonavida, B. Mechanism of cell-mediated cytotoxicity at the 29. Dennis, J. W., and LafertÃ©,S. Tumor cell surface carbohydrate and the single cell level. 1. Estimation of cytotoxic T lymphocyte frequency and metastatic phenotype. Cancer Metastasis Rev., 5: 185-204, 1987. relative lytic efficiency. J. Immunol., 'l23: 2861-2869, 1979. 30. Chan, W. S.. Jackson, A., and Turner. G. A. Differences in surface expression 56. Yagita, M., Sugiyama, H., and Takahashi, T. Depression of natural killer of WGA-binding proteins of cells from a lymphosarcoma and its liver cell activity in patients with aplastic anemia. Acta Haematol. Jpn., 48: 893- mÃ©tastases.Br.J. Cancer, 49: 181-191. 1984. 903, 1985. 31. Irimura, T., and Nicolson, G. L. Carbohydrate chain analysis by lectin binding 57. Pearlstein, E., Cooper, L. B., and Karpatkin, S. J. Extraction and character to electrophoretically separated glycoproteins from murine B16 melanoma ization of a platelet aggregating material from SV40-transformed mouse 3T3 sublines of various metastatic properties. Cancer Res., 44: 791-798, 1984. fibroblasts. J. Lab. Clin. Med.. 93: 332-344. 1979. 32. Irimura. T., Tressler, R. J., and Nicolson, G. L. Sialoglycoproteins of murine 58. Kijima-Suda, L, Miyazawa, T., Itoh, M.. Toyoshima, S., and Osawa, T. RAW 117 large cell lymphoma/lymphosarcoma sublines of various metastatic Possible mechanism of inhibition of experimental pulmonary metastasis of colonization properties. Exp. Cell Res., 765: 403-416, 1986. mouse colon adenocarcinoma 26 subline by a sialic acid:nucleoside conjugate. 33. Yogeeswaran, G., and Salk, P. L. Metastatic potential is positively correlated Cancer Res., 48: 3728-3732, 1988.

3338

Downloaded from cancerres.aacrjournals.org on September 24, 2021. © 1990 American Association for Cancer Research. Soluble Factor in Normal Tissues That Stimulates High-Molecular-Weight Sialoglycoprotein Production by Human Colon Carcinoma Cells

Tatsuro Irimura, Andrew M. McIsaac, Debora A. Carlson, et al.

Cancer Res 1990;50:3331-3338.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/50/11/3331

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/50/11/3331. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.